Aircraft Accident Report

NATIONAL

TRANSPORTATION

SAFETY BOARD

WASHINGTON, D.C. 20594

AIRCRAFT ACCIDENT REPORT

PB2001-910402

NTSB/AAR-01/02

DCA99MA060

RUNWAY OVERRUN DURING LANDING

AMERICAN AIRLINES FLIGHT 1420

MCDONNELL DOUGLAS MD-82, N215AA

LITTLE ROCK, ARKANSAS

JUNE 1, 1999

7195A

this page intentionally left blank

Aircraft Accident Report

Runway Overrun During Landing

American Airlines Flight 1420

McDonnell Douglas MD-82, N215AA

Little Rock, Arkansas

June 1, 1999

NTSB/AAR-01/02

PB2001-910402 National Transportation Safety Board

Notation 7195A 490 L’Enfant Plaza, S.W.

Adopted October 23, 2001 Washington, D.C. 20594

National Transportation Safety Board. 2001. Runway Overrun During Landing, American Airlines

Flight 1420, McDonnell Douglas MD-82, N215AA, Little Rock, Arkansas, June 1, 1999. Aircraft

Accident Report NTSB/AAR-01/02. Washington, DC.

Abstract: This report explains the accident involving American Airlines flight 1420, a McDonnell

Douglas MD-82, which crashed after it overran the end of runway 4R during landing at Little Rock

National Airport in Little Rock, Arkansas. Safety issues discussed in this report focus on flight crew

performance, flight crew decision-making regarding operations in adverse weather, pilot fatigue, weather

information dissemination, emergency response, frangibility of airport structures, and Federal Aviation

Administration (FAA) oversight. Safety recommendations concerning these issues are addressed to the

FAA and the National Weather Service.

The National Transportation Safety Board is an independent Federal agency dedicated to promoting aviation, railroad, highway, marine,

pipeline, and hazardous materials safety. Established in 1967, the agency is mandated by Congress through the Independent Safety Board

Act of 1974 to investigate transportation accidents, determine the probable causes of the accidents, issue safety recommendations, study

transportation safety issues, and evaluate the safety effectiveness of government agencies involved in transportation. The Safety Board

makes public its actions and decisions through accident reports, safety studies, special investigation reports, safety recommendations, and

statistical reviews.

Recent publications are available in their entirety on the Web at <http://www.ntsb.gov>. Other information about available publications also

may be obtained from the Web site or by contacting:

National Transportation Safety Board

Public Inquiries Section, RE-51

490 L’Enfant Plaza, S.W.

Washington, D.C. 20594

(800) 877-6799 or (202) 314-6551

Safety Board publications may be purchased, by individual copy or by subscription, from the National Technical Information Service. To

purchase this publication, order report number PB2001-910402 from:

National Technical Information Service

5285 Port Royal Road

Springfield, Virginia 22161

(800) 553-6847 or (703) 605-6000

The Independent Safety Board Act, as codified at 49 U.S.C. Section 1154(b), precludes the admission into evidence or use of Board reports

related to an incident or accident in a civil action for damages resulting from a matter mentioned in the report.

iii Aircraft Accident Report

Contents

Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

1. Factual Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 History of Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Injuries to Persons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.3 Damage to Airplane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.4 Other Damage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5 Personnel Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5.1 The Captain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.5.2 The First Officer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.5.3 The Flight Attendants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.6 Airplane Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.6.1 Maintenance Records. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.6.2 Spoiler System Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

1.6.2.1 Testimony on Spoiler System Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

1.6.2.2 Other Spoiler Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

1.6.3 Braking System Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

1.6.4 Weight and Balance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1.6.5 N215AA’s Previous Flights on the Day of the Accident. . . . . . . . . . . . . . . . . . . . . 22

1.6.6 MD-80 Demonstrated Landing Distance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.7 Meteorological Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.7.1 Airport Weather Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.7.2 National Weather Service Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1.7.3 American Airlines Weather Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

1.7.4 Additional Weather Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1.7.4.1 Lightning Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1.7.4.2 Witness Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1.7.4.3 Windshear Hazard Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

1.8 Aids to Navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1.9 Communications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1.10 Airport Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1.10.1 Runway 4R/22L Safety Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

1.10.2 Runway 22L Approach Lighting System Support Structure . . . . . . . . . . . . . . . . . . 37

1.10.3 Runway 4R Assessments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

1.10.3.1 Tire Marks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

1.10.3.2 Runway Surface Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

1.10.4 Air Traffic Control Tower Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

1.11 Flight Recorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

1.11.1 Cockpit Voice Recorder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

1.11.2 Flight Data Recorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

1.12 Wreckage and Impact Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

1.12.1 General Wreckage Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Contents iv Aircraft Accident Report

1.12.2 Spoiler System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

1.12.3 Engines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

1.12.4 Landing Gear and Brake Assemblies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

1.13 Medical and Pathological Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

1.14 Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

1.15 Survival Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

1.15.1 General. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

1.15.2 Evacuation of Passengers and Crewmembers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

1.15.3 Emergency Response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

1.15.4 Passenger Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

1.16 Tests and Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

1.16.1 Spoiler System Ground Tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

1.16.1.1 Testing Conducted After the Palm Springs Incident . . . . . . . . . . . . . . . . . . . . . . 56

1.16.2 Main Landing Gear Tire Examination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

1.16.3 Cockpit Voice Recorder Sound Spectrum Study. . . . . . . . . . . . . . . . . . . . . . . . . . . 58

1.16.4 Airplane Performance Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

1.16.4.1 Calculated Ground Trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

1.16.4.2 Ground Deceleration Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

1.16.5 Engineered Materials Arresting System Computer Model . . . . . . . . . . . . . . . . . . . 63

1.17 Organizational and Management Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

1.17.1 Aviation Safety Action Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

1.17.2 Flight Crew Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

1.17.2.1 Simulator Flight Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

1.17.2.2 Observations of Simulator Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

1.17.2.3 Human Factors and Safety Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

1.17.3 Approach Procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

1.17.3.1 Approach Briefing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

1.17.3.2 Before Landing Checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

1.17.3.2.1 Manufacturer’s Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

1.17.3.3 Crew Coordination Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

1.17.3.4 Stabilized Approach Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

1.17.3.4.1 Federal Aviation Administration Guidance . . . . . . . . . . . . . . . . . . . . . . . . . 75

1.17.3.5 Thunderstorm and Windshear Avoidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

1.17.3.6 Continuation of an Approach Below the Decision Height . . . . . . . . . . . . . . . . . 77

1.17.3.7 Missed Approach Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

1.17.4 Landing Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

1.17.4.1 Wind Landing Limits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

1.17.4.1.1 Manufacturer’s Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

1.17.4.2 Spoiler Deployment Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

1.17.4.2.1 Manufacturer’s Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

1.17.4.3 Braking Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

1.17.4.3.1 Manufacturer’s Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

1.17.4.4 Use of Reverse Thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

1.17.4.4.1 Manufacturer’s Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

1.17.4.4.2 MD-80 All Operators Letter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

1.17.5 Postaccident Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

1.17.5.1 DC-9 Operating Manual Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

1.17.5.2 Flight Manual Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

1.17.6 Federal Aviation Administration Oversight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

1.17.6.1 Principal Operations Inspector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Contents v Aircraft Accident Report

1.17.6.2 Aircrew Program Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

1.18 Additional Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

1.18.1 Runway Friction Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

1.18.2 Study on Thunderstorm Penetration in the Terminal Area . . . . . . . . . . . . . . . . . . . 94

1.18.3 Studies on Flight Crew Decision-Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

1.18.3.1 Safety Board Study of Flight Crew Involvement in Major Accidents. . . . . . . . . 95

1.18.3.2 National Aeronautics and Space Administration Study on Flight Crew

Decision Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

1.18.4 Technologies to Detect and Locate Downed Airplanes. . . . . . . . . . . . . . . . . . . . . . 96

1.18.5 Previous Weather-Related Accidents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

1.18.5.1 Previous Weather-Related Safety Recommendations . . . . . . . . . . . . . . . . . . . . . 98

1.18.6 Previous Fatigue-Related Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

1.18.6.1 Previous Fatigue-Related Safety Recommendation. . . . . . . . . . . . . . . . . . . . . . 103

1.18.6.2 Additional Fatigue Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

1.18.7 Other Previous Related Accidents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

1.18.7.1 Accidents Involving Rescue Efforts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

1.18.7.2 Accidents Involving Delayed Emergency Response. . . . . . . . . . . . . . . . . . . . . 108

1.18.8 Other Previous Related Safety Recommendations . . . . . . . . . . . . . . . . . . . . . . . . 109

1.18.8.1 Aircraft Operations in the Airport Environment . . . . . . . . . . . . . . . . . . . . . . . . 109

1.18.8.2 Stabilized Approach Guidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

1.18.8.3 Manufacturer’s Operating Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

2. Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.2 Accident Scenario. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.2.1 The Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

2.2.1.1 Descent Into the Terminal Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

2.2.1.2 Maneuvering to the Airport for Final Approach . . . . . . . . . . . . . . . . . . . . . . . . 115

2.2.1.3 Final Approach Segment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

2.2.1.4 Summary of the Flight Crew’s Performance During the Approach . . . . . . . . . 126

2.2.2 The Landing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

2.2.2.1 Lack of Spoiler Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

2.2.2.1.1 Autospoiler Arming. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

2.2.2.1.2 Checklist Design Regarding Spoiler Arming . . . . . . . . . . . . . . . . . . . . . . . 132

2.2.2.1.3 Manual Spoiler Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

2.2.2.2 Use of Reverse Thrust Above 1.3 Engine Pressure Ratio . . . . . . . . . . . . . . . . . 135

2.2.2.3 Use of Manual Braking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

2.2.2.4 Summary of the Landing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

2.2.3 Human Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

2.2.3.1 The Role of Situational Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

2.2.3.1.1 Industry Standards and Practices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

2.2.3.2 The Role of Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

2.3 Meteorological Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

2.3.1 Weather Information Provided by the Local Controller . . . . . . . . . . . . . . . . . . . . 146

2.3.1.1 Weather Information Depiction on Air Traffic Control Radar Systems . . . . . . 147

2.3.2 Additional En Route Weather Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

2.3.2.1 Dispatch Office Weather Radar. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

2.3.2.2 Center Weather Service Unit Staffing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

2.3.2.3 Automated Surface Observing System Lockout Period . . . . . . . . . . . . . . . . . . 149

2.3.3 Airport Weather Equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Contents vi Aircraft Accident Report

2.3.3.1 Runway Visual Range System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

2.3.3.2 Low Level Windshear Alert System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

2.4 Emergency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

2.4.1 Aircraft Rescue and Fire Fighting Staffing Levels . . . . . . . . . . . . . . . . . . . . . . . . 154

2.4.2 Crash Detection and Location Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

2.4.3 Interagency Emergency Response Critique. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

2.5 Airport Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

2.5.1 Runway Safety Areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

2.5.2 Nonfrangible Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

2.6 American Airlines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

2.6.1 Stabilized Approach Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

2.6.2 Spoiler System Training. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

2.6.3 Spoiler and Braking Systems Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

2.7 Federal Aviation Administration Oversight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

3. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

3.1 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

3.2 Probable Cause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169

4. Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

5. Appendixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

A: Investigation and Hearing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

B: Cockpit Voice Recorder Transcript . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

C: Automated Surface Observing System Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

vii Aircraft Accident Report

Figures

1. Flight 1420’s Approach Path to the Airport and Key CVR Comments. . . . . . . . . . . . . . . . . . . . . . . . 7

2. Spoiler Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3. Spoiler Hydraulic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4. Spoiler Handle in the Unarmed Position. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

5. Spoiler Handle in the Armed Position. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6. Position of the Crank Arm and Roller With the Spoiler Handle in the

Unarmed and Armed Positions (View From Forward Right to Aft Left) . . . . . . . . . . . . . . . . . . . . . . 19

7. 0.4° Base Reflectivity Scan at 2334:27. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

8. 0.4° Base Reflectivity Scan at 2340:28. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

9. 0.4° Base Reflectivity Scan at 2345:57. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

10. 0.4° Base Reflectivity Scan at 2351:59. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

11. Radial Velocity Image Surrounding the Time of the Accident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

12. Aerial Photograph of Runway 4R/22L, Airplane Wreckage, and

Runway 22L Approach Lighting System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

13. Airplane Wreckage and Runway 22L Approach Lighting System. . . . . . . . . . . . . . . . . . . . . . . . . . . 45

14. View of Left Side of Airplane Wreckage and Runway 22L Approach Lighting System. . . . . . . . . . 45

15. Interior Airplane Configuration and Occupant Injuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

16. Effects of Spoilers, Brakes, and Reverse Thrust on Stopping Distance . . . . . . . . . . . . . . . . . . . . . . . 62

17. Categories of the Postaccident Audit Recommendations According to

Expected Completion Date and Safety Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

18. Weather and Flight Information for 2339:12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115

19. Weather and Flight Information for 2339:45 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .116

20. Weather and Flight Information for 2342:55 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117

21. Weather and Flight Information for 2344:30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .118

Figures viii Aircraft Accident Report

22. Weather and Flight Information for 2345:47 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119

23. Weather and Flight Information for 2346:52 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

24. Weather and Flight Information for 2347:08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

25. Weather and Flight Information for 2349:12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

26. Weather and Flight Information for 2350:02 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

ix Aircraft Accident Report

Abbreviations

AC advisory circular

ACARS aircraft communication addressing and reporting system

AD airworthiness directive

afl above field level

agl above ground level

AIM Aeronautical Information Manual

APM aircrew program manager

ARFF aircraft rescue and fire fighting

ARTCC air route traffic control center

ASAP American Airlines Safety Action Program

ASOS Automated Surface Observing System

ASR Airport Surveillance Radar

ATC air traffic control

ATC T air traffic control tower

ATI S automatic terminal information service

ATOS Air Transportation Oversight System

CAMI Civil Aerospace Medical Institute

CFR Code of Federal Regulations

CVR cockpit voice recorder

CWSU center weather service unit

DA decision altitude

D-BRITE Digital Bright Radar Indicator Terminal Equipment

dBz decibel

DEVS Driver’s Enhanced Vision System

DH decision height

DOT Department of Transportation

Abbreviations x Aircraft Accident Report

ELT emergency locator transmitter

EMAS Engineered Materials Arresting System

EPR engine pressure ratio

FAA Federal Aviation Administration

FCOM flight crew operating manual

FDR flight data recorder

fpm feet per minute

GPWS ground proximity warning system

HBAT Flight Standards Handbook Bulletin for Air Transportation

Hg mercury

HIWAS Hazardous In-flight Weather Advisory Service

Hz hertz

IFR instrument flight rules

ILS instrument landing system

IOE initial operating experience

LLWAS Low Level Windshear Alert System

LOFT line-oriented flight training

MDA minimum descent altitude

MEL minimum equipment list

MEMS Metropolitan Emergency Medical Services

METAR meteorological aerodrome report

mHz megahertz

msl mean sea level

NASA National Aeronautics and Space Administration

nm nautical mile

NPRM notice of proposed rulemaking

NWS National Weather Service

POI principal operations inspector

psi pounds per square inch

Abbreviations xi Aircraft Accident Report

PTRS Program Tracking and Reporting System

RVR runway visual range

SDR service difficulty report

SIGMEC significant meteorological condition

SIGMET significant meteorological information

SPECI special weather observation

TAF terminal aerodrome forecast

TDWR Terminal Doppler Weather Radar

TRACON terminal radar approach control

TSA time since awakening

UTC coordinated universal time

ref

reference airspeed for landing

VFR visual flight rules

VOR very high frequency omnidirectional radio range

WSP Weather Systems Processor

WSR-88D Weather Surveillance Radar 1988 Doppler system

xii Aircraft Accident Report

Executive Summary

On June 1, 1999, at 2350:44 central daylight time, American Airlines flight 1420,

a McDonnell Douglas DC-9-82 (MD-82), N215AA, crashed after it overran the end of

runway 4R during landing at Little Rock National Airport in Little Rock, Arkansas. Flight

1420 departed from Dallas/Fort Worth International Airport, Texas, about 2240 with 2

flight crewmembers, 4 flight attendants, and 139 passengers aboard and touched down in

Little Rock at 2350:20. After departing the end of the runway, the airplane struck several

tubes extending outward from the left edge of the instrument landing system localizer

array, located 411 feet beyond the end of the runway; passed through a chain link security

fence and over a rock embankment to a flood plain, located approximately 15 feet below

the runway elevation; and collided with the structure supporting the runway 22L approach

lighting system. The captain and 10 passengers were killed; the first officer, the flight

attendants, and 105 passengers received serious or minor injuries; and 24 passengers were

not injured. The airplane was destroyed by impact forces and a postcrash fire. Flight 1420

was operating under the provisions of 14 Code of Federal Regulations Part 121 on an

instrument flight rules flight plan.

The National Transportation Safety Board determines that the probable causes of

this accident were the flight crew’s failure to discontinue the approach when severe

thunderstorms and their associated hazards to flight operations had moved into the airport

area and the crew’s failure to ensure that the spoilers had extended after touchdown.

Contributing to the accident were the flight crew’s (1) impaired performance resulting

from fatigue and the situational stress associated with the intent to land under the

circumstances, (2) continuation of the approach to a landing when the company’s

maximum crosswind component was exceeded, and (3) use of reverse thrust greater than

1.3 engine pressure ratio after landing.

The safety issues in this report focus on flight crew performance, flight crew

decision-making regarding operations in adverse weather, pilot fatigue, weather

information dissemination, emergency response, frangibility of airport structures, and

Federal Aviation Administration (FAA) oversight. Safety recommendations concerning

these issues are addressed to the FAA and the National Weather Service.

1 Aircraft Accident Report

1. Factual Information

1.1 History of Flight

On June 1, 1999, at 2350:44 central daylight time,

American Airlines flight 1420,

a McDonnell Douglas DC-9-82 (MD-82), N215AA, crashed after it overran the end of

runway 4R during landing at Little Rock National Airport in Little Rock, Arkansas.

Flight 1420 departed from Dallas/Fort Worth International Airport, Texas, about 2240

with 2 flight crewmembers, 4 flight attendants, and 139 passengers aboard and touched

down in Little Rock at 2350:20. After departing the end of the runway, the airplane struck

several tubes extending outward from the left edge of the instrument landing system (ILS)

localizer array, located 411 feet beyond the end of the runway; passed through a chain link

security fence and over a rock embankment to a flood plain, located approximately 15 feet

below the runway elevation; and collided with the structure supporting the runway 22L

approach lighting system. The captain and 10 passengers were killed; the first officer, the

flight attendants, and 105 passengers received serious or minor injuries; and 24 passengers

were not injured.

The airplane was destroyed by impact forces and a postcrash fire.

Flight 1420 was operating under the provisions of 14 Code of Federal Regulations (CFR)

Part 121 on an instrument flight rules (IFR) flight plan.

Flight 1420 was the third and final leg of the first day of a 3-day sequence for the

flight crew. The flight sequence began at O’Hare International Airport, Chicago, Illinois.

According to American Airlines company records, the captain checked in for the flight

about 1038, and the first officer checked in about 1018. Flight 1226, from Chicago to Salt

Lake City International Airport, Utah, departed about 1143 and arrived about 1458 (1358

mountain daylight time). Flight 2080, from Salt Lake City to Dallas/Fort Worth, departed

about 1647 (1547 mountain daylight time) and arrived about 2010, 39 minutes later than

scheduled because of an airborne hold during the approach resulting from adverse

weather in the airport area. The captain was the flying pilot for flight 1226, and the first

officer was the flying pilot for flight 2080.

Flight 1420, from Dallas/Fort Worth to Little Rock, was scheduled to depart

about 2028 and arrive about 2141. However, before its arrival at Dallas/Fort Worth, the

flight crew received an aircraft communication addressing and reporting system

(ACARS)

message indicating a delayed departure time of 2100 for flight 1420. After

deplaning from flight 2080, the flight crew proceeded to the departure gate for flight

1420. The flight crew then received trip paperwork for the flight, which included an

American Airlines weather advisory for a widely scattered area of thunderstorms along

Unless otherwise indicated, all times in this report are central daylight time, based on a 24-hour clock.

See section 1.2 for an injury chart.

ACARS is a data link system that, among other things, enables airline dispatchers and flight

crews to communicate via text messages while an airplane is in flight.

Factual Information 2 Aircraft Accident Report

the planned route and two National Weather Service (NWS) in-flight weather advisories

for an area of severe thunderstorms

along the planned route.

The airplane originally intended to be used for the flight was delayed in its arrival

to Dallas/Fort Worth because of the adverse weather in the area. After 2100, the first

officer notified gate agents that flight 1420 would need to depart by 2316 because of

American’s company duty time limitation.

The first officer then telephoned the flight

dispatcher to suggest that he get another airplane for the flight or cancel it.

Afterward, the

accident airplane, N215AA, was substituted for flight 1420. The flight’s 2240 departure

time was 2 hours 12 minutes later than the scheduled departure time. The captain was the

flying pilot, and the first officer was the nonflying pilot.

About 2254, the flight dispatcher sent the flight crew an ACARS message

indicating that the weather around Little Rock might be a factor during the arrival. The

dispatcher suggested that the flight crew expedite the arrival to beat the thunderstorms if

possible, and the flight crew acknowledged this message. The first officer indicated, in a

postaccident interview, that “there was no discussion of delaying or diverting the

landing” because of the weather. According to the predeparture trip paperwork, two

alternate airports—Nashville International Airport, Tennessee, and Dallas/Fort Worth—

were specified as options in case a diversion was needed.

Beginning about 2258, flight 1420 was handled by controllers from the Fort

Worth Air Route Traffic Control Center (ARTCC). About 2304, the Fort Worth center

broadcast NWS Convective SIGMET [significant meteorological information] weather

advisory 15C for an area of severe thunderstorms that included the Little Rock airport

area. The cockpit voice recorder (CVR) indicated that the flight crew had discussed the

weather and the need to expedite the approach. At 2325:47, the captain stated, “we got to

get over there quick.” About 5 seconds later, the first officer said, “I don’t like

that…that’s lightning,” to which the captain replied, “sure is.” The CVR also indicated

that the flight crew had the city of Little Rock and the airport area in sight by at 2326:59.

About 2327, the Fort Worth center cleared the flight to descend to 10,000 feet

mean sea level (msl) and provided an altimeter setting of 29.86 inches of mercury (Hg).

The flight was transferred about 2328 to the Memphis ARTCC, which provided the same

altimeter setting.

A severe thunderstorm has winds measuring 50 knots (58 mph) or greater, 3/4-inch or larger

hail, or tornadoes and can produce torrential rain and frequent lightning.

See section 1.7 for specific details on weather advisories discussed in this section.

American Airlines’ maximum pilot duty time (by contractual agreement with the Allied Pilots

Association) is 14 hours from the scheduled time of check-in for the first flight leg to the time

that the last flight leg is scheduled to land.

The flight dispatcher stated that he received the first officer’s call between 2150 and 2200.

The Center Weather Service Unit (CWSU) of the Memphis center was staffed for 16-hour

operation and had closed about 2130.

Factual Information 3 Aircraft Accident Report

According to the CVR, the flight crew contacted the Little Rock Air Traffic

Control Tower (ATCT) at 2334:05. The controller advised the flight crew that a

thunderstorm located northwest of the airport was moving through the area and that the

wind was 280º at 28 knots gusting to 44 knots. The first officer told the controller that he

and the captain could see the lightning. The controller told the flight crew to expect an

ILS approach to runway 22L.

The first officer indicated in a postaccident interview that,

during the descent into the terminal area, the weather appeared to be about 15 miles away

from the airport and that he and the captain thought that there was “some time” to make

the approach.

The CVR indicated that, between at 2336:04 and at 2336:13, the captain and first

officer discussed American Airlines’ crosswind limitation for landing. The captain

indicated that 30 knots was the crosswind limitation but realized that he had provided the

limitation for a dry runway. The captain then stated that the wet runway crosswind

limitation was 20 knots, but the first officer stated that the limitation was 25 knots. In

testimony at the National Transporation Safety Board’s public hearing on this accident,

the first officer stated that neither he nor the captain checked the actual crosswind

limitation in American’s flight manual. The first officer testified that he had taken the

manual out but that the captain had signaled him to put the manual away because the

captain was confident that the crosswind limitation was 20 knots.

At 2339:00, the controller cleared the flight to descend to an altitude of 3,000 feet

msl. The controller then asked the flight crew about the weather conditions along the

runway 22L final approach course, stating his belief that the airplane’s weather radar was

“a lot better” than the weather radar depiction available in the tower. At 2339:12, the first

officer stated, “okay, we can…see the airport from here. We can barely make it out but

we should be able to make [runway] two two…that storm is moving this way like your

radar says it is but a little bit farther off than you thought.” The controller then offered

flight 1420 a visual approach to the runway, but the first officer indicated, “at this point,

we really can’t make it out. We’re gonna have to stay with you as long as possible.”

At 2339:45, the controller notified flight 1420 of a windshear alert,

reporting that

the centerfield wind was 340º at 10 knots, the north boundary wind was 330º at 25 knots,

The ILS is a precision approach system consisting of a localizer and a glideslope, which provide

lateral and vertical guidance, respectively, to a runway. The system uses ground-based radio transmitters

that provide both the localizer and the glideslope signals.

The Safety Board held a public hearing on this accident from January 26 to 28, 2000, in

Little Rock (see appendix A). The Board may hold a public hearing as part of its investigation

into an accident to supplement the factual record of the investigation. The Board calls technical

experts and material witnesses to testify, and Board investigative staff and designated representatives

from the parties to the investigation ask questions to obtain additional factual information. The hearing

is not intended to analyze factual information for cause.

The captain correctly indicated that 20 knots was the crosswind limitation for landing on a

wet runway. See section 1.17.4.1 for additional information on American’s crosswind limitations.

Windshear is generally defined as a change in wind direction and/or speed over a short distance.

Windshear alerts for Little Rock airport are issued by a Low Level Windshear Alert System (LLWAS),

which is explained in section 1.7.1.

Factual Information 4 Aircraft Accident Report

and the northwest boundary wind was 010º at 15 knots. The flight crew then requested

runway 4R so that there would be a headwind, rather than a tailwind, during landing. At

2340:20, the controller instructed the flight crew to fly a heading of 250º for vectors to the

runway 4R ILS final approach course. After reaching the assigned heading, the airplane

was turned away from the airport and clear of the thunderstorm that had previously been

reported by the controller. The CVR indicated that, between 2340:46 and 2341:31, the first

officer stated the localizer frequency and course, the decision altitude, the minimum safe

altitude, and a portion of the missed approach procedure for runway 4R.

Between 2342:19 and 2342:24, the CVR indicated that the captain asked the first

officer, “do you have the airport? Is that it right there? I don’t see a runway.” At 2342:27,

the controller told the flight crew that the second part of the thunderstorm was apparently

moving through the area and that the winds were 340º at 16 knots gusting to 34 knots. At

2342:40, the first officer asked the captain whether he wanted to accept “a short

approach” and “keep it in tight.” The captain answered, “yeah, if you see the runway.

‘cause I don’t quite see it.” The first officer stated, “yeah, it’s right here, see it?” The

captain replied, “you just point me in the right direction and I’ll start slowing down here.”

At 2342:55, the first officer said, “it’s going right over the…field.” At 2342:59, the first

officer told the controller, “well we got the airport. We’re going between clouds. I think

it’s right off my, uh, three o’clock low, about four miles.” The controller then offered a

visual approach for runway 4R, and the first officer accepted. At 2343:11, the controller

cleared flight 1420 for a visual approach to runway 4R and indicated “if you lose it, need

some help, let me know please.”

At 2343:35, the first officer stated, “you’re comin’ in. There’s the airport.” Three

seconds later, the captain stated, “uh, I lost it,” to which the first officer replied, “see it’s

right there.” The captain then stated, “I still don’t see it...just vector me. I don’t know.”

At 2343:59, the controller cleared flight 1420 to land and indicated that the winds were

330º at 21 knots. At 2344:19, the captain stated, “see we’re losing it. I don’t think we can

maintain visual.” At 2344:30, the first officer informed the controller that visual contact

with the airport had been lost because of a cloud between the airplane and the airport. The

controller then cleared the airplane to fly a heading of 220º for radar vectors for the ILS

approach to runway 4R and directed the flight to descend to and maintain 2,300 feet msl.

At 2345:47, the first officer told the controller “we’re getting pretty close to this storm.

we’ll keep it tight if we have to.” The controller indicated to the flight crew that, “when

you join the final, you’re going to be right at just a little bit outside the marker if that’s

gonna be okay for ya.” The captain stated, “that’s great,” and the first officer told the

controller, “that’s great with us.” At 2346:39, the controller advised the flight crew that

the airplane was 3 miles from the outer marker.

The outer marker is located 5.9 miles from the airport. In a postaccident interview, the first

officer stated that, at this point, there was urgency to land because the weather was “up against” the airport.

Factual Information 5 Aircraft Accident Report

At 2346:52, the captain stated, “aw, we’re goin’ right into this.” At the same time,

the controller reported that there was heavy rain at the airport, the automatic terminal

information service (ATIS) information in effect at the time was no longer current,

the

visibility was less than 1 mile, and the runway visual range (RVR)

for runway 4R was

3,000 feet. The first officer acknowledged this information. At 2347:08, the controller

again cleared flight 1420 to land and indicated that the wind was 350º at 30 knots gusting

to 45 knots. The first officer then read back the wind information as 030º at 45 knots. At

2347:22, the captain stated, “three thousand RVR. We can’t land on that.” Four seconds

later, the first officer indicated that the RVR for runway 4R was 2,400 feet, and the

captain then said, “okay, fine.”

At 2347:44, the captain stated, “landing gear down,” and the CVR recorded a

sound consistent with the landing gear being operated. About 5 seconds later, the captain

stated, “and lights please.” At 2347:53, the controller issued a second windshear alert for

the airport, reporting that the centerfield wind was 350º at 32 knots gusting to 45 knots,

the north boundary wind was 310º at 29 knots, and the northeast boundary wind was 320º

at 32 knots. This transmission was not acknowledged by the flight crew. At 2348:10, the

captain stated, “add twenty [knots],” to which the first officer replied, “right.”

At 2348:12, the controller reported that the runway 4R RVR was now 1,600 feet.

About 2348:18, the captain indicated that the flight was established on final approach;

6 seconds later, the first officer informed the controller that the flight was established on

the inbound portion of the ILS. The controller repeated the clearance to land; stated that

the wind was 340º at 31 knots, the north boundary wind was 300º at 26 knots, and the

northeast boundary wind was 320º at 25 knots; and repeated the RVR. At 2348:41, the

first officer acknowledged this information. The controller did not receive any further

transmissions from flight 1420. At 2349:02, the first officer asked the captain, “want

forty flaps?” The captain indicated that he thought he had already called for the landing

flaps, after which the first officer stated, “forty now.” At 2349:10, the controller informed

the flight crew that the wind was 330º at 28 knots. Two seconds later, the captain stated,

“this is a can of worms.”

According to the CVR, the first officer stated, “there’s the runway off to your

right, got it?” at 2349:24. The captain replied, “no,” to which the first officer stated, “I

ATIS information Romeo, which was issued about 2326, was based on a 2322 special weather

observation indicating (among other things) a thunderstorm with frequent in-cloud and cloud-to-cloud

lightning located west through northwest and moving northeast, winds from 190º at 14 knots, a

visibility of 7 miles, and an altimeter setting of 29.88 inches of Hg.

RVR is a measurement of the visibility near a runway’s surface. This measurement represents

the horizontal distance that a pilot should be able to see down a runway from the approach end.

According to the Jeppesen Sanderson approach plate for the ILS approach to runway 4R (dated

February 20, 1998), the lowest authorized RVR for that runway was 2,400 feet.

If the weather is reported to be below published minimums, American Airlines and the Federal

Aviation Regulations (14 CFR 121.651) allow airplanes that are established on the final approach

segment to continue the approach to the appropriate decision height (DH) or minimum descent altitude

(MDA) and land in accordance with the conditions for the type of approach being conducted. For

this Category I ILS approach, the DH was 200 feet.

Factual Information 6 Aircraft Accident Report

got the runway in sight. You’re right on course. Stay where you’re at.” The captain then

stated, “I got it. I got it.” At 2349:32, the controller reported the wind to be 330º at

25 knots. At 2349:37, an unidentified voice in the cockpit stated, “wipers,” and the CVR

then recorded a sound consistent with windshield wiper motion. (This sound continued

throughout the rest of the flight.) At 2349:53, the controller reported the wind to be 320º

at 23 knots.

The CVR indicated that, at 2349:57, an unidentified voice in the cockpit stated,

“aw…we’re off course” and that, 1 second later, an unintelligible comment was made by

an unidentified voice in the cockpit. In a postaccident interview, the first officer stated

that he thought the approach was stabilized until about 400 feet above field level (afl), at

which point the airplane drifted to the right. The first officer also stated that he said “go

around” about that time but not in a very strong voice. The first officer indicated that he

had looked at the captain to see if he had heard him but that the captain was intent on

flying and was doing “a good job.”

The CVR indicated that, at 2350:00, the first officer said, “we’re way off.” Flight

data recorder (FDR) information indicated that the localizer deviation value was about

one dot to the right at that point.

About 1 second later, the captain stated, “I can’t see it.”

About 3 seconds afterward, the first officer asked, “got it?” to which the captain replied,

“yeah I got it.” At 2350:13 and :14, the CVR recorded the sound of the ground proximity

warning system (GPWS) radio altitude callout “sink rate.”

Calculations based on FDR

data indicated that the airplane was descending through an altitude of about 70 feet afl at

the time of the first sink rate warning and about 50 feet afl at the time of the second

warning. Figure 1 shows flight 1420’s flightpath to Little Rock and runway 4R along with

key CVR comments and the airplane’s location when the comments were made.

FDR and CVR data indicated that the airplane touched down on the runway about

2350:20. About 2350:22, the first officer stated “we’re down;” about 2 seconds later, he

stated, “we’re sliding.” FDR data also indicated that, over a 7-second period after

touchdown, both thrust reversers were deployed and the left and right engines’ engine

pressure ratios (EPR) reached settings of 1.89 and 1.67, respectively.

The thrust

reversers were subsequently moved to the unlocked status (neither deployed nor stowed).

Cockpit instrumentation shows the airplane’s location relative to the glideslope and localizer

signals. Displacement is shown in terms of the airplane’s angular deviation above or below the

glideslope and left or right of the localizer. Pilots can judge the amount of displacement by needle

deflections that reference “dots” on the face of the instruments. Because the dots represent an angular

measurement of an airplane’s deviation from an ILS component, the amount of feet of displacement

depends on the airplane’s distance from the ILS antennas.

The sink rate alert indicates a rate of descent that exceeds predetermined thresholds.

Thrust reversers redirect engine exhaust to help slow an airplane. EPR is a measurement of

engine power as a ratio of gases in the exhaust pipe compared with air entering the inlet.

Factual Information 7 Aircraft Accident Report

According to the FDR, the flight spoilers did not deploy symmetrically at touchdown,

but a momentary 8° deflection of the left outboard flight spoiler concurrent with a left

aileron deflection was recorded.

Figure 1. Flight 1420’s Approach Path to the Airport and Key CVR Comments

A spoiler is a device located on an airplane wing’s upper surface that, when activated, provides

increased drag and decreased lift by disrupting the flow of air over the wing. After touchdown,

the spoilers increase the load on the landing gear tires, which is essential for maximizing braking

and obtaining the shortest stopping distance.

FDR data also showed that the flight spoilers extended symmetrically for 55 seconds, beginning

at 2336:42, during the descent into Little Rock.

-7-6-5-4-3-2-1012345678

-15

-14

-13

-12

-11

-10

-9

-8

-7

-6

-5

-4

-3

-2

-1

2350:05/430 ft.

Captain: "yeah I got it."

2350:01/472 ft.

Captain: "I can't see it."

2349:31/940 ft.

Captain: "I got it, I got it."

2349:12/1,138 ft.

Captain:

"this is a can of worms."

2346:52/2,300 ft.

Captain: "we're

goin' right into this."

2345:47/2,546 ft.

F/O to controller: "we're getting

pretty close to this storm."

2339:12/5,325 ft.

F/O to controller: "we can

see the airport from here."

Radar returns

Runway 4R

Selected CVR information

Distance north of runway 4R threshold, nm

Distance east of runway 4R threshold, nm

2344:19/3,300 ft.

Captain: "see we're

losing it. I don't think

we can maintain visual."

2342:55/3,300 ft.

F/O: "it's going

right over the field."

2344:30/3,300 ft.

F/O to controller: "there's

a cloud between us and the airport.

We just lost the field."

2347:08/2,300 ft.

Controller to crew: "runway

4R, cleared to land. The wind

350 at 30, gusts 45."

Factual Information 8 Aircraft Accident Report

FDR data indicated that the right and left brake pedals began to move at 2350:25

and :30, respectively, and both pedals reached full travel at 2350:31. About the time that

the brakes were applied, the thrust reversers were deployed again. At 2350:32, the CVR

recorded an unidentified voice in the cockpit stating “on the brakes.”

The left engine

reached a maximum setting of 1.98 reverse EPR, and the right engine reached a setting of

1.64 reverse EPR. The left brake pedal was relaxed at 2350:34 before returning to its full

position 2 seconds later. About the time that the left brake pedal was relaxed, the

reversers were returned to the unlocked status. As the right thrust reverser was being

moved to the unlocked status, the right engine reached a maximum setting of 1.74 reverse

EPR.

At 2350:36, FDR data indicated a full 60º deployment of the right inboard flight

spoiler, concurrent with a full aileron deflection.

At 2350:40, the left thrust reverser was

moved back to the deployed position, but the right reverser moved briefly to the deployed

position and then moved to the stowed position. According to FDR data, the left thrust

reverser remained deployed, and the right thrust reverser remained stowed, for the

remainder of the flight. About 1 second later, the CVR recorded expletives stated by an

unidentified voice in the cockpit, which were followed by the sounds of initial impact at

2350:44 and several additional impacts beginning at 2350:47. The CVR stopped

recording at 2350:48. The airplane came to rest about 800 feet from the departure end of

runway 4R, 34° 44.18 minutes north latitude and 92° 11.97 minutes west longitude. The

accident occurred during the hours of darkness.

1.2 Injuries to Persons

Table 1. Injury chart

Note: Title 14 CFR 830.2 defines a serious injury as any injury that (1) requires hospitalization for more than 48 hours,

starting within 7 days from the date that the injury was received; (2) results in a fracture of any bone, except simple fractures

of fingers, toes, or the nose; (3) causes severe hemorrhages or nerve, muscle, or tendon damage; (4) involves any internal

organ; or (5) involves second- or third-degree burns or any burns affecting more than 5 percent of the body surface. A minor

injury is any injury that does not qualify as a fatal or serious injury.

In a postaccident interview, the first officer indicated that he did not help with the flight

controls until the captain said “brakes” as the airplane was nearing the end of the runway, at which

time the first officer helped with the brakes.

The FDR recorded only the left aileron position, which indicated a full trailing edge-down

deflection at the time.

Injuries Flight Crew Cabin Crew Passengers Other Total

Fatal

1 0 10 0 11

Serious

1 3 41 0 45

Minor

0 1 64 0 65

None

0 0 24 0 24

Total

2 4 139 0 145

Factual Information 9 Aircraft Accident Report

1.3 Damage to Airplane

According to American Airlines, the damage to the airplane was estimated at

$10.7 million.

1.4 Other Damage

The airplane destroyed several tubes extending outward from the left edge of the

runway 4R ILS localizer array, part of a chain link security fence, and approximately

250 feet of the runway 22L approach lighting system support structure and walkway. The

damage was estimated at $325,000.

1.5 Personnel Information

1.5.1 The Captain

The captain, age 48, was hired by American Airlines in July 1979. He held an

Airline Transport Pilot certificate and a Federal Aviation Administration (FAA) First

Class medical certificate dated February 9, 1999, with no restrictions. The captain was

type rated on the Boeing 727 (727) and the MD-80. He qualified as a 727 flight engineer

on August 24, 1979; first officer on March 15, 1985; and captain on September 21, 1988.

He qualified as an MD-80 captain on July 31, 1991. The captain was also a lieutenant

colonel in the U.S. Air Force Reserves.

The captain began his aviation career with the U.S. Air Force in 1972. He flew

T-33 and EB-57 airplanes and was a command flight examiner and instructor pilot for the

B-57 airplane. He left active military service in 1979 at the rank of captain and began

working for American Airlines afterward. The captain was furloughed after 1 year with

American but was recalled by the company 3 1/2 years later.

In July 1998, the captain was promoted to check airman on the MD-80. In a

postaccident interview, the MD-80 Fleet Manager stated that the captain was

recommended for this position by the Chicago-O’Hare base manager and another check

airman because of his technical competence, performance as a line pilot, and ability and

desire to instruct. In January 1999, the captain was promoted to chief pilot at the

Chicago-O’Hare base.

The base manager indicated that the captain wanted to be a chief

pilot because he had been flying the MD-80 for a long time and wanted a change. The

base manager also indicated that the captain was selected for a chief pilot position

because of his flying background, company achievements, and leadership skills.

During his furlough from American Airlines, the captain worked as a nuclear engineer on

submarine propulsion plants.

American Airlines policy requires chief pilots to fly 1 month per year as line pilots. The

Chicago-O’Hare base manager encouraged chief pilots to fly once a week as line pilots.

Factual Information 10 Aircraft Accident Report

The Chicago-O’Hare base manager, who flew with the captain twice in May

1999, said that he was “extremely comfortable flying with the captain” and that the

captain had “a great deal of common sense.” A first officer who had flown with the

captain from Chicago to Dallas/Fort Worth indicated that he was “a knowledgeable pilot

who was not intimidating.”

American Airlines records indicated that the captain had accumulated

10,234 hours total flying time, including 7,384 hours as a company pilot-in-command

(5,518 of which were on the MD-80). His last recurrent ground training and proficiency

check occurred on July 19, 1998, and his last line check occurred July 26, 1998. He had

flown approximately 54, 46, 14, and 12 hours in the 90, 60, 30, and 7 days, respectively,

preceding the accident. FAA records indicated no accident, incident, or enforcement

action.

The captain’s wife described his activities in the 3 days before the accident as

routine, adding that the captain had no scheduled events during that time. She stated that

the captain went to sleep about 2200 the night before the accident and slept until between

0700 and 0730. She further stated that, on nonflying days, the captain would typically go

to sleep between 2130 and 2200, wake up about 0515, and leave for work about 0600.

The captain’s wife indicated that he slept later than usual on the morning of the accident

because the timing of the first day of the trip did not necessitate an early rising.

The captain’s wife indicated that his health was good, he was a nonsmoker, and

he consumed a minimal amount of alcohol. At the time of the accident flight, the captain

was not taking any prescription drugs. No significant life events occurred in the weeks

before the accident, and his finances and personal situation were reported to be stable.

Records at the National Driver Register showed no indication of driver’s license

revocation or suspension for the captain.

The captain and first officer had not flown together before the day of the accident.

According to two flight attendants who were working aboard the two legs before the

accident flight (flights 1226 and 2080), the captain and first officer seemed to have a

good working relationship with each other. The first officer for flight 1420 stated, in a

postaccident interview, that the captain had made him feel comfortable in the cockpit.

Also, the first officer testified at the public hearing on this accident that his interaction

with the captain was not affected by the fact that he was a chief pilot.

1.5.2 The First Officer

The first officer, age 35, was hired by American Airlines in January 1999. He held

an Airline Transport Pilot certificate and an FAA First Class medical certificate dated

November 12, 1998, with no restrictions. He qualified as a first officer on the MD-80 on

February 22, 1999. He was serving a 1-year probation period required of new company

hires. The first officer was type rated on the Learjet and the Boeing 737.

Factual Information 11 Aircraft Accident Report

The first officer received his private pilot’s license in 1983. He began his career in

1988 with the U.S. Navy and completed primary flight training. The first officer had been

selected for advanced jet training but was given an honorable discharge in 1991 because

of a reduction in force. Before he was hired by American Airlines, the first officer

worked as a corporate pilot, flying C-210, Learjet 35, and King Air E-90 airplanes. He

was also the director of operations and the chief pilot for an air charter company and a

flight instructor.

A captain who flew with the first officer in May 1999 stated that he was an “above

average new hire who was very competent and knowledgeable.” Another captain who

flew with the first officer in May 1999 stated that he was an “experienced pilot with good

cockpit discipline.”

According to American Airlines records, the first officer had accumulated

4,292 hours of flying time, 182 of which were as a company MD-80 pilot. He had flown

approximately 176, 112, 65, and 7 1/2 hours in the 90, 60, 30, and 7 days, respectively,

preceding the accident. His proficiency check occurred on February 22, 1999, and his

line check occurred on March 10, 1999. FAA records indicated no accident, incident, or

enforcement action.

On May 30, 1999, the first officer traveled from his home outside Los Angeles,

California, to Chicago. The first officer indicated that he had been commuting from his

home to the Chicago-O’Hare base for about 3 months and that, as a result, he was

adjusted to the central time zone. The first officer indicated that he was involved in

routine activities while in the Chicago area. He went to bed between 2000 and 2200 the

night before the accident and woke up about 0730.

The first officer indicated that he was a nonsmoker and that he was not taking any

prescription medications at the time of the accident flight. Records at the National Driver

officer.

1.5.3 The Flight Attendants

Flight 1420 was staffed with four flight attendants hired by American Airlines

between June 1987 and August 1992. All of the flight attendants were qualified on

MD-80 series airplanes. The flight attendants completed the company’s initial training,

which included instruction in emergency procedures and evacuation drills, and their most

recent company recurrent emergency procedures training was completed in either 1998

or 1999.

On the day before the accident flight, three of the flight attendants began the same

3-day trip sequence. On the day of the accident flight, they had worked three trip

segments before flight 1420. The fourth flight attendant began a 2-day trip sequence on

the day of the accident flight. She had worked two trip segments before flight 1420.

Factual Information 12 Aircraft Accident Report

Flights 1226 and 2080 (the two legs before the accident flight) were not among the trip

segments worked by any of the flight attendants.

1.6 Airplane Information

The MD-80 airplane is a derivative model of the DC-9 airplane. As a result, much

of the MD-80’s structure and many of its systems, components, and installations are

similar to the earlier DC-9 model. According to Boeing,

the Douglas DC-9 airplane

entered service in December 1965; the final DC-9 was delivered in October 1982. The

MD-80 airplane’s first flight occurred in October 1979. The FAA certified the MD-80

series airplane in August 1980, and the airplane entered service in November 1980. The

MD-80 model airplanes—the MD-81, -82, -83, -87, and -88—were in production through

1999. The DC-9 family of airplanes also includes the MD-90 and the Boeing 717.

The accident airplane, N215AA, serial number 49163, was delivered new to

American Airlines on August 1, 1983. At the time of the accident, the airplane had

accumulated 49,136 flight hours and 27,103 cycles.

A review of American Airlines’ Air

Carrier Certificate, which included the standards, terms, conditions, and limitations

contained in the FAA-approved Operations Specifications, revealed no discrepancies.

The FAA’s Type Certificate Data Sheet, which prescribes the conditions and limitations

under which airplanes meet airworthiness requirements, noted no discrepancies for DC-9

or MD-80 series airplanes. The FAA’s Program Tracking and Reporting System (PTRS)

indicated no discrepancies for the accident airplane from January 1998 to May 1999.

N215AA was equipped with two Pratt & Whitney JT8D-217C turbofan engines.

The No. 1 (left) engine, serial number 718427, was installed on N215AA on September

7, 1997; the No. 2 (right) engine, serial number 725712, was installed on July 30, 1998.

American Airlines’ records indicated that, for engine No. 1, the time since new

was 29,734 hours (15,711 cycles),

the time since overhaul was 11,216 hours

(5,189 cycles), and the time since installation was 5,256 hours (2,447 cycles). The

records also indicated that, for engine No. 2, the time since new was 25,131 hours

The Boeing Company and McDonnell Douglas Corporation merged in August 1997. Douglas

Aircraft Company and McDonnell Aircraft Company merged in April 1967.

An airplane cycle is one complete takeoff and landing sequence.

The PTRS is an FAA computer tracking system that includes information on inspection and

surveillance activities by FAA inspectors. In 1997, the FAA began developing the Air Transportation

Oversight System (ATOS) as the FAA’s new oversight system and the eventual replacement for

PTRS. The Manager of the Flight Standard Division for the FAA’s Western Pacific Region, who

testified at the Safety Board’s public hearing on this accident, stated that ATOS was intended to

replace PTRS because that data repository system had reached a plateau and there was no way

to make further safety gains with PTRS. Only surveillance data for 10 major U.S. air carriers (including

American Airlines) are currently recorded under ATOS; certification activities and data for the 10

carriers are still recorded under PTRS.

An engine cycle is one complete startup and shutdown sequence.

Factual Information 13 Aircraft Accident Report

(13,216 cycles), the time since overhaul was 11,658 hours (5,421 cycles), and the time

since installation was 2,618 hours (1,229 cycles).

The accident airplane was equipped with two AlliedSignal VOR

/ILS receivers

and an AlliedSignal GPWS computer. The accident airplane was also equipped with a

forward-looking X-band airborne weather radar unit that depicted three levels of

reflectivity in green, yellow, and red (according to intensity from lightest to heaviest).

X-band airborne weather radar systems are subject to attenuation, that is, the scattering or

absorption of electromagnetic energy with heavy precipitation. The accident airplane’s

airborne weather radar did not have attenuation alerts to warn the flight crew of any

masking of weather resulting from heavy precipitation, and the radar was not able to

detect lightning strikes.

The airborne weather radar had a power output of approximately 125 watts and

was integrated with a 30-inch antenna that provided a 3.4º beam width. The radar had a

stabilization feature that helped to keep the radar beam steady during turns and pitch

changes. The radar had range selections of 10, 20, 40, 80, 160, and 320 miles, a 180º

horizontal azimuth display, and a vertical tilt control of ±15º. In a postaccident interview,

the first officer stated that the airborne weather radar was being operated in the 10-, 20-,

and 40-nautical mile (nm) ranges and that the tilt was set at the maximum up (+15º)

setting. In public hearing testimony, the first officer indicated that he saw only green

radar returns depicted on the weather radar unit.

1.6.1 Maintenance Records

American’s engineering specification maintenance intervals for MD-82 airplanes

include “Periodic Service”; “A,” “B,” and “C” checks; and “HC” [heavy C] checks.

Periodic Service checks are to be accomplished a maximum of 2 flying days from the last

periodic service or higher check. A and B checks are to be accomplished every 65 and

470 flight hours, respectively. The first C check is to be accomplished within 5,000 flight

hours, and all subsequent C checks are to be accomplished within 4,200 flight hours of

the last C check. The first HC check is to be accomplished within 14,000 flight hours, the

second one within 12,000 flight hours since the previous HC check or a total time of

24,000 flight hours, and the third and subsequent HC checks within 12,000-flight hour

intervals.

The accident airplane’s last Periodic Service and A checks were performed on

May 31, 1999. All tires and wheels were checked for airworthiness. Tire inflation and

brake wear were also checked. No maintenance items were deferred. The last B check

was performed on April 21, 1999. All tires and wheels were checked for airworthiness.

Tire pressure, flight and ground spoilers, and hydraulic subsystems were also checked. In

addition, the requirements of FAA Airworthiness Directive (AD) 98-11-10 were

performed. This AD mandated an inspection of the spoiler handle latching lever pin and

actions to prevent it from jamming. (According to the AD, a jammed spoiler handle pin

VOR stands for very high frequency omnidirectional radio range.

Factual Information 14 Aircraft Accident Report

can result in retraction of the spoilers and full advancement of the left throttle during a

go-around.) The last C check was performed on January 6, 1999, and the last HC check

was performed on January 22, 1994.

The No. 1 (left) and No. 2 (right) nose gear tire and wheel assemblies were last

replaced on May 27 and February 4, 1999, respectively. The No. 1 (outboard) and

2 (inboard) left main landing gear tire and wheel assemblies were last replaced on May 7,

1999. The No. 3 (inboard) and 4 (outboard) right main landing gear tire and wheel

assemblies were last replaced on May 15 and April 7, 1999, respectively. The No. 1 and 2

left main landing gear brakes were last replaced on April 3 and April 2, 1999,

respectively. The No. 3 and No. 4 right main landing gear brakes were last replaced on

December 29, 1998, and April 3, 1999, respectively. No discrepancies were noted after

these replacements. All but one of the replacements were nonroutine discrepancies

generated from Periodic Service inspections; the No. 1 left nose gear tire and wheel

assembly replacement resulted from a pilot report of excessive nose wheel vibration at

liftoff.

Discrepancies recorded in American Airlines airplane maintenance logbooks are

entered into the company’s Field Maintenance Reliability System. Entries from May 15,

1998, to June 1, 1999, in the accident airplane’s maintenance logbook were reviewed for

discrepancies that referenced flight controls/spoilers, antiskid control, wheels, brakes,

rain protection, engine controls, and engine reversing. Also, Field Maintenance

Reliability reports, which include descriptions of any mechanical discrepancy, any

corrective maintenance actions, and any minimum equipment list (MEL) deferrals, were

generated for the accident airplane from June 1, 1998, to May 31, 1999. Selected

discrepancies involving autothrottle/speed control, landing, spoilers/drag devices,

windows/windshields, brakes, fuselage, engine fuel and control, engine controls, and

thrust reversers were reviewed for corrective maintenance actions. No discrepancies were

noted.

FAA service difficulty reports (SDR) were reviewed from all DC-9-82 airplane

operators regarding the airplane’s flight controls/spoilers. Between January 1984 and

August 1999, 49 SDRs were submitted regarding flight controls/spoilers; no maintenance

trends or discrepancies were found. Also, SDRs for all systems on the accident airplane

were reviewed. Between September 1985 and August 1999, 14 SDRs were submitted, but

none were relevant to the circumstances of the flight 1420 accident.

In addition, SDRs were reviewed from all DC-9 type-certificated airplanes

regarding the airplane’s drag control system and drag control actuator. Between

January 1995 and October 23, 2001, there were 62 reports regarding the drag control

system and 21 reports regarding the drag control actuator, including 13 reports submitted

after the flight 1420 accident. Most of the reports were inspection related, false

indications, or adjustment or chaffing problems. No maintenance trends or discrepancies

were noted; however, one report, which involved a DC-9-32, was noted as being relevant

to the circumstances of this accident. Specifically, the discrepancy report stated the

following: “Discrepancy: unable to arm ground spoilers on approach, and spoilers did not

deploy manually on landing. Corrective Action: Replace spoiler control actuator and

Factual Information 15 Aircraft Accident Report

spoiler control box. Spoilers adjusted and checked serviceable. Note: the report does not

state if the autospoiler ‘do not use’ light was illuminated.”

1.6.2 Spoiler System Information

The MD-80 series airplane has one ground spoiler, one inboard flight spoiler, and

one outboard flight spoiler on each wing. Each of the flight spoilers is extended and

retracted by its own hydraulic actuator. The left hydraulic system, through the left spoiler

bypass valve and a 1,500-pounds per square inch (psi) pressure reducer valve, supplies

hydraulic power for the two inboard flight spoiler actuators. The right hydraulic system,

through the right spoiler bypass valve and a 1,500-psi pressure reducer valve, supplies

hydraulic power for the two outboard flight spoiler actuators.

Both hydraulic systems

supply hydraulic power, through the two ground spoiler control valves, to the two ground

spoiler actuators, which are at full system pressure (3,000 psi) and do not have pressure

reducer valves. Two position sensors—one mounted on the right inboard flight spoiler

panel and the other on the left outboard flight spoiler panel—provide information to the

FDR on the position of those spoilers. The FDR records each spoiler position alternately

at a sampling rate of two times per second.

The flight spoilers are manually operated through the aileron control system by

either the control wheel or the spoiler handle in the cockpit. A control wheel input can

supplement lateral control (provided by the ailerons) by extending the flight spoilers on

the downward-moving wing to a maximum of 60º. An input from the spoiler handle

extends the flight spoilers symmetrically on both wings to a maximum of 35º in flight

(referred to as the speed brake function) and a maximum of 60º on the ground. A

spring-loaded torsion bar mechanically holds the flight spoilers in the retract position

when they are not extended.

The ground spoilers are manually operated by an input from the spoiler handle on

the cockpit center pedestal and electrical signals from the proximity system electronic

unit via several spoiler control relays. The ground spoilers are extended to 60º only

during landing or a rejected takeoff. The ground spoilers are locked down by hydraulic

power and a mechanical overcenter link during all other phases of flight.

The ground and flight spoilers can be automatically operated by the autospoiler

system.

This system consists of two ground spoiler control valves, the autospoiler

switching unit, four wheel spin-up transducers (by way of a ground spoiler control box),

and two ground control nose oleo switches. To use the autospoiler system for landing, a

If one of the hydraulic pumps were to fail or was rendered inoperable for maintenance reasons,

a power transfer unit would allow the remaining hydraulic system to supply power to the inoperative

system. The MD-82 is also equipped with an auxiliary hydraulic pump to power the right hydraulic

system on the ground.

The MD-80 can be configured with a spoiler lockout mechanism to prevent in-flight spoiler

deployment when the flaps are extended. None of American’s MD-80s are configured with the mechanism.

MD-80 airplanes can be dispatched with the autospoiler system inoperative. Under that circumstance,

the pilot would be required to manually deploy the spoilers upon touchdown.

Factual Information 16 Aircraft Accident Report

pilot raises the spoiler handle up to the ARM position (before touchdown), which reveals

a red ARM indicator stripe and positions the roller on the spoiler handle in front of an

autospoiler crank arm. (The red indicator stripe provides a visual cue to the flight crew

that the autospoiler system is armed.) When the autospoiler switching unit commands the

autospoiler actuator to move the crank arm from the retract to the extend positions, the

crank arm pushes the spoiler handle fully aft and extends all of the flight and ground

spoilers. After touchdown, the autospoiler switching unit commands the autospoiler

actuator to move the crank arm, which in turn moves the spoiler handle fully aft when

either the wheel spin-up transducers signal main wheel spin-up or the ground control

nose oleo switches signal nose gear touchdown (in case of a failure of the spin-up

transducers). The autospoiler system activates the autospoiler actuator on each landing

regardless of whether the handle is in the ARM position.

The MD-80 is configured with a spoiler autoretract mechanism that retracts the

ground and flight spoilers if the left throttle is advanced above idle (about 1 3/4 to

2 inches). The left throttle arm has a crank that “knocks down,” or dislodges, the spoiler

handle from the latching mechanism so that the handle return spring can return the

spoilers to the stowed position. (The spoiler handle can also be manually dislodged.) The

right throttle does not have a crank to knock down the spoiler handle.

An amber AUTO SPOILER DO NOT USE light illuminates on the overhead

annunciator panel in the cockpit if the autospoiler system detects certain failures.

Specifically, the light will illuminate if (1) the ground spoiler actuator fails to change

positions within 10 seconds after being so commanded, (2) the spoiler control relay

circuit or the ground spoiler control box circuit has an internal short to ground, (3) only

one takeoff or land relay channel is energized, or (4) the two weight-on-wheel sensors

(proximity sensors on the main landing gear that indicate when the struts are compressed)

and the landing gear handle are inconsistent. If this light appears for one of the first three

causes, the flight crew must manually deploy the ground spoilers upon landing. If the

light appears for the last cause, the autospoiler actuator will operate as intended as long as

it receives normal inputs.

Figures 2 through 6 show the MD-80 spoiler system. Figure 2 shows the spoiler

control system, including the locations of the flight and ground spoilers. Figure 3 shows

the spoiler hydraulic system. Figures 4 and 5 are photographs showing the location of the

spoiler handle in the unarmed and armed positions, respectively. Figure 6 shows the

position of the crank arm and roller with the spoiler handle in unarmed and armed

positions.

Factual Information 17 Aircraft Accident Report

Figure 2. Spoiler Control System

Figure 3. Spoiler Hydraulic System

Factual Information 18 Aircraft Accident Report

Figure 4. Spoiler Handle in the Unarmed Position

Figure 5. Spoiler Handle in the Armed Position

Factual Information 19 Aircraft Accident Report

Figure 6. Position of the Crank Arm and Roller With the Spoiler Handle in the Unarmed

and Armed Positions (View From Forward Right to Aft Left)

1.6.2.1 Testimony on Spoiler System Operation

According to the Boeing Company’s Engineering Manager for Landing Gear,

Brake, and Hydraulic Systems, who testified at the Safety Board’s public hearing on this

accident, the autospoiler system provides “quick, timely, and full 60º deployment” of all

spoiler panels with “minimal flight crew input” during a rejected takeoff or after a landing

touchdown. The Engineering Manager stated that the only human factor involved in the

system’s operation is the arming of the spoilers to ensure that the autospoiler system can

successfully engage and actuate the spoiler system. The Engineering Manager also

testified that the spoiler system is very reliable. For example, if the spoilers were armed

and tire spin-up did not occur upon touchdown, the spoilers would still automatically

deploy because the system was designed to use nose gear compression as the backup

signal for spoiler deployment.

The Engineering Manager indicated that the two spoiler parameters on the

accident airplane’s FDR appeared to be functioning normally. He explained that a full turn

of the steering yoke in the cockpit (that is, a full aileron deflection) would have caused the

right inboard flight spoiler to fully deflect momentarily, as indicated by FDR data. The

Engineering Manager testified that the right flight spoilers reduced lift on that wing for

only about 2 seconds, which would not have made much difference in the airplane’s

braking.

SPOILER HANDLE

CRANK ARM

(ACTUATOR

EXTENDED)

UNARMED ARMED

SPOILER HANDLE

ROLLER

CRANK

ARM

Factual Information 20 Aircraft Accident Report

The Engineering Manager said that there are no design features in the autospoiler

system to warn pilots that the spoiler handle has not been armed. He added that some

changes have been made to the system since it was designed in the 1960s, but most of

these changes have been to ensure that the ground spoilers do not inadvertently deploy

during flight. According to the Engineering Manager’s testimony, the electrical signal

required to deploy the ground spoilers requires all of the following: the landing gear

handle switch is in the down position, the left or right main landing gear

weight-on-wheels signal is received from the proximity system electronic unit, and the

left throttle switch is in the idle position.

The Engineering Manager further testified that there are no aural or visual

warnings in the cockpit to indicate that the spoilers did not deploy. However, there are

some aural and visual cues to indicate to the flight crew that the spoilers have been

deployed, including the extensive motion of the spoiler handle as it is moving into the

extend position and the associated “clanking” sound.

An aerodynamics engineer from Boeing presented information on the effect of the

spoilers on the weight on the wheels for a 127,000-pound MD-80 series airplane at

1 second after touchdown (the point during the landing roll when the airplane is at its

highest speed and is developing the highest lift). The weight distribution of the airplane

was estimated according to the percent of total landing weight supported by the wings,

the main landing gear, and the nose gear.

With the spoilers deployed, about 20 percent of the airplane’s total landing weight

is supported by the wings, about 77 percent of the total weight is supported by the main

gear, and about 3 percent of the total weight is supported by the nose gear. Without

spoilers, about 70 percent of the airplane’s total landing weight is supported by the

wings, about 27 percent of the total weight is supported by the main gear, and about

3 percent of the total weight is supported by the nose gear.

Without spoilers and with an

additional 20 knots of speed, about 90 percent of the airplane’s total landing weight is

supported by the wings, about 7 percent of the total weight is supported by the main gear,

and about 3 percent of the total weight is supported by the nose gear. According to the

Boeing engineer, this finding is important because, when less weight is applied on the

main gear, it has less braking force and produces less cornering force in a skid.

1.6.2.2 Other Spoiler Events

During the public hearing on this accident, the Safety Board requested that

Boeing determine whether it had received any reports of armed autospoilers not

deploying on MD-80 or DC-9 aircraft. Boeing sent two letters, dated April 20 and

October 24, 2000, to the Safety Board that indicated that the company was able to find

nine records of such reports. Most of these events could be explained by either

The estimated weight distribution for landings without spoilers assumed a 10º nose-down elevator.

The Boeing engineer stated that applying nose-down elevator reduces the weight on the main gear

by transferring it to the nose gear. He also stated that, because the nose gear does not have any

brakes, the braking force is reduced and the distance to stop is increased.

Factual Information 21 Aircraft Accident Report

component or procedural failures. For example, Boeing stated that an autospoiler actuator

on a DC-9 was removed and replaced because the spoiler handle did not come all of the

way back and latch. In another example, Boeing indicated that the autospoilers on a DC-9

did not deploy on landing but that an inspection revealed that the autospoiler circuit

breaker was open. For the remaining events, the data were insufficient to verify the

details of the event or determine the cause.

American Airlines reported two instances in which the spoiler handle aboard an

MD-80 series airplane extended and then retracted after main landing gear touchdown.

The events occurred at Dallas/Fort Worth on June 15, 2000, aboard flight 497 and on

September 17, 2000, aboard flight 787. The FDR from the flight 497 airplane only

recorded the right outboard spoiler position at a sampling rate of once per second, and the

FDR indicated that the spoiler deployed at touchdown and then retracted about 1 second

later. The FDR from the flight 787 airplane showed that both the left outboard and right

inboard flight spoilers deployed at touchdown and then retracted about 1 second later.

According to American, both flight crews could feel the spoilers retract, after which the

crews manually deployed the spoilers and safely decelerated the airplanes. American also

indicated that the flight crews believed the throttles were idle at touchdown.

1.6.3 Braking System Information

The MD-80 series airplane braking system can be operated manually or

automatically. The automatic braking (autobrake) system on the MD-80 is optional; the

accident airplane was equipped with the system. According to Boeing’s Engineering

Manager for Landing Gear, Brake, and Hydraulic Systems, the MD-80 autobrake system

provides “rapid and full application of the brakes” in the event of a rejected takeoff and

“timely and consistent brake application” during landing. The first officer stated in a

postaccident interview that the captain elected to use only manual brakes for landing.

The autobrake system control panel is on the aft right portion of the center

pedestal. To arm the system before landing, a pilot selects one of three deceleration

levels—minimum, medium, and maximum—and moves the arm switch to the ARM

position. After landing, the spoiler handle is moved aft either manually or automatically

by the autospoiler system, and the autobrake switches then initiate the automatic

application of the brakes. (The autobrakes will operate only if the spoiler handle is in the

extended position.) With the minimum and medium levels, brake application begins

about 3 seconds after landing; with the maximum level, brake application begins about

1 second after landing.

The autobrake system can be unarmed so that the braking function is returned to

the pilots. The system is unarmed by manually depressing any brake pedal more than

25 percent of its maximum travel, advancing any throttle out of the near-idle range,

The CVR indicated that, at 2331:22, the first officer stated, “manual brakes?” to which the

captain replied, “uh, manual’s fine.”

Factual Information 22 Aircraft Accident Report

moving the arm switch to the DISARM position, or returning the spoiler handle to the

stowed position.

MD-80 series airplanes are equipped with an antiskid braking system. This

system adapts braking pressure (applied manually or automatically) to runway conditions

by sensing an impending skid condition and adjusting the brake pressure on each

individual wheel to allow for maximum braking performance. The inboard brakes have

touchdown protection to prevent them from being applied before or at touchdown. The

outboard brakes do not have touchdown protection.

1.6.4 Weight and Balance

According to the trip paperwork, the takeoff center of gravity for N215AA was

16.7 percent of mean aerodynamic chord, which was within the approved limits of the

airplane. The trip paperwork also included the following information for the airplane:

basic operating weight, 83,123 pounds; passenger weight, 25,020 pounds; baggage

weight, 3,475 pounds; zero fuel weight, 111,618 pounds; fuel, 24,500 pounds; ramp

weight, 136,118 pounds; taxi fuel burn, 2,080 pounds; takeoff weight, 134,038 pounds;

estimated fuel burn, 6,289 pounds;

and estimated landing weight, 127,749 pounds. The

zero fuel, ramp, takeoff, and landing weight maximums were 122,000; 150,500; 136,300;

and 130,000, respectively.

1.6.5 N215AA’s Previous Flights on the Day of the Accident

On the day of the accident, N215AA flew from Dallas/Fort Worth to Denver and

back to Dallas/Fort Worth. The captain of those flights indicated, in a postaccident

interview, that the spoilers were armed and deployed normally for each landing. He also

indicated that the weather radar worked well during deviations around weather in the

Denver area. The first officer of those flights stated, in a postaccident interview, that the

spoilers, thrust reversers, and manual brakes worked normally. He also stated that the

airplane’s weather radar operated normally in the 20-, 40- 80- and 160-mile ranges.

FDR data for the accident airplane’s previous landing at Dallas/Fort Worth

showed that the flight spoilers deployed symmetrically immediately after touchdown to

their full 60° position and remained that way for about 33 seconds until they returned to

their stowed position. Boeing’s Engineering Manager for Landing Gear, Brake, and

Hydraulic Systems indicated that the FDR data for the airplane’s previous landing

“definitely” showed normal spoiler deployment. FDR data also showed that both thrust

reversers deployed after the airplane touched down at Dallas/Fort Worth.

The flight dispatcher sent the flight crew an ACARS message about 2311, which revised the

fuel burn amount to allow for en route weather deviations. The message also included a reminder

to the flight crew that Nashville and Dallas/Fort Worth airports were the flight’s two alternate airports.

According to FDR data, both thrust reversers also deployed when the airplane powered back

from the gate during its departure from Dallas/Fort Worth to Little Rock.

Factual Information 23 Aircraft Accident Report

1.6.6 MD-80 Demonstrated Landing Distance

As part of an airplane’s certification, a manufacturer must demonstrate the

distance to land from a 50-foot height to a complete stop (14 CFR 25.125). The

aerodynamics engineer from Boeing, in public hearing testimony, stated that the MD-80

“demonstrated landing distance” was measured in two parts—the air distance, from

50 feet to touchdown, and the ground distance, from touchdown to stop—and on a dry,

hard-surfaced runway. According to the aerodynamics engineer, the operating

requirements in Part 121 provide for additional safety margins beyond the demonstrated

landing distance, specifically, variations in the speed, touchdown point, runway surface

condition, tire condition, temperature, and runway slope. The Part 121 minimum dry

runway length is the demonstrated landing distance plus 67 percent. The Part 121

minimum wet runway length is the minimum dry runway length plus 15 percent.

The Boeing engineer testified that the MD-80’s landing performance was

demonstrated to the FAA using the following test conditions: forward center of gravity,

ref

(1.3 times the stalling speed) at 50 feet, 40º flaps, autospoilers, pilot-actuated

(manual) antiskid braking, no reverse thrust, and a weight range of 109,200 to

149,500 pounds. The engineer presented information regarding the required landing

distances for an MD-80 with a landing weight of 127,000 pounds. The actual

demonstrated landing distance was 2,830 feet, so the Part 121 minimum dry and wet

runway lengths were 4,715 and 5,425 feet, respectively. Runway 4R/22L at Little Rock

National Airport is 7,200 feet in length, so there is about an 1,800-foot margin between

the required minimum wet runway length and the end of the runway.

1.7 Meteorological Information

1.7.1 Airport Weather Information

Weather observations at Little Rock National Airport are made by an Automated

Surface Observing System (ASOS), which is maintained by the NWS. The ASOS records

continuous information on wind speed and direction, cloud cover, temperature,

precipitation, and visibility. The ASOS wind anemometer is installed 32 feet afl. The

ASOS transmits an official meteorological aerodrome report (known as a METAR) at

53 minutes past each hour and a special weather observation (known as a SPECI) as

conditions warrant; such conditions include a wind shift, change in visibility, and change

in ceiling (cloud cover or height). The system is prevented from issuing any reports

between 47:20 and 53:20 after the hour (known as the lockout period) so that the hourly

observation can be prepared, edited, and transmitted.

ASOS observations at the Little Rock airport are augmented by certified weather

observers under contract with the FAA. The augmentation station is located in the

airport’s terminal building, and the ASOS unit is located near the approach end of

runway 4L. The NWS inspected the ASOS unit on May 21, 1999, and found it to be

working properly.

Factual Information 24 Aircraft Accident Report

The ASOS special weather observation for 2323 was as follows:

Special weather observation for Little Rock at 0423Z, winds from 180º at

09 knots, visibility 7 miles with thunderstorms, a few clouds at 7,000 feet in

cumulonimbus clouds, ceiling broken at 10,000 feet, temperature 25º C, dew

point temperature 23º C, altimeter 29.86 inches of Hg. Remarks: ASOS

observation, thunderstorm began at 23 minutes past the hour, frequent lightning

in-cloud and cloud-to-cloud located from the west through the northwest,[

]

thunderstorm west through northwest, moving northeast.

The ASOS edit log indicated that, at 2347:22, a special observation was canceled

because it occurred during the lockout period. The ASOS edit log also indicated that the

following observation was recorded about the time of the accident but was not

disseminated because of the lockout period:

Little Rock weather observation at 0450:31Z, winds from 290º at 16 knots

gusting to 28 knots, visibility 1 1/2 miles in thunderstorm and heavy rain[

] and

mist, a few clouds at 3,700 feet, ceiling overcast 5,000 feet, temperature 18.9º C,

dew point 16.7º C, altimeter 29.94 inches of Hg. Remarks: ASOS observation,

peak wind from 290º at 35 knots at 0433Z, wind shift at 0431Z, thunderstorm

began at 0423Z, rain began at 0424Z, sea level pressure 1014.0 mb [millibars],

frequent lightning in-cloud, and cloud-to-cloud, west through northwest,

occasional lightning in-cloud, cloud-to-cloud, and cloud-to-ground east,

thunderstorm west through northwest, thunderstorm east moving east,

precipitation since last hourly observation 0.37 inches.

The hourly observation about 2353 indicated that the wind was from 280º at

18 knots gusting to 26 knots and that visibility was 1 mile in thunderstorms and heavy

rain and mist. A special observation was issued about 2355, indicating that the wind was

from 290º at 13 knots gusting to 26 knots and that visibility was 3/4 mile in

thunderstorms and heavy rain and mist. Another special observation, issued about 2358,

indicated that the wind was from 290º at 10 knots gusting to 76 knots; the winds were

varying from 210 to 030º; the visibility was 1/2 mile in thunderstorms, small hail, heavy

rain, and mist; and a peak wind from 320º at 76 knots was measured about 2356.

ASOS also provides precipitation measurements in 1- and 15-minute increments.

Between 2331 and 2345, the ASOS measured 0.14 inch of rain, with the first measurable

rain greater than trace amounts (less than 0.01 inch) about 2336. About 2350

(immediately before the time of the accident), the system measured a total of 0.37 inch of

rain. The 15-minute measurement for 2346 to 0000 indicated that 1.09 inches of rain had

Weather observations are transmitted in coordinated universal time (UTC). The “Z” designation

that follows the time in the weather observation stands for Zulu, which indicates UTC time. Central

daylight time is 5 hours behind UTC time.

The Federal Meteorological Handbook number 1 defines frequent lightning as one to six flashes

per minute.

The NWS defines heavy rain as 0.03 inch of rain within 6 minutes or more than 0.30 inch

of rain per hour.

Factual Information 25 Aircraft Accident Report

fallen. Appendix C shows detailed ASOS wind information and precipitation amounts

surrounding the time of the accident.

ATIS information is based on ASOS observations provided to the tower by the

official airport weather observer.

ATIS information Romeo, which was current

beginning about 2326, stated:

Good evening Little Rock Adams Field[

] information Romeo zero four two two

Zulu special observation, wind one niner zero at one four, visibility seven,

thunderstorm, few clouds at seven thousand, cumulonimbus, ceiling one zero

thousand broken, temperature two five, dew point two three, altimeter two niner

eight eight, frequent lightning in-cloud, cloud-to-cloud, west through northwest,

moving northeast. ILS runway two two left approach in use. Notices to Airmen,

runway two two right, four left ILS out of service. Attention all aircraft,

hazardous weather information for the Little Rock area available on HIWAS

[Hazardous In-flight Weather Advisory Service], flight watch, or flight

service...advise on initial contact you have Romeo.

In addition to ASOS, Little Rock National Airport is equipped with an FAA Type

FA-10240 Low Level Windshear Alert System (LLWAS), which was operational at the

time of the accident. The LLWAS uses six wind sensors at remote stations located around

the airport to collect wind speed, direction, and gust data.

One of the six sensors is the

centerfield wind sensor, which is installed 70 feet afl. Readings from this sensor are used

by tower controllers as the source of real-time wind data for pilots. LLWAS alerts are

displayed in the ATCT so that controllers can warn flight crews of potentially hazardous

windshear conditions. (As discussed in section 1.1, the controller transmitted windshear

alerts to the 1420 flight crew about 2339 and 2347.) The FAA performed a sensor

performance evaluation after the accident, which determined that all of the LLWAS

sensors were working properly at the time of the accident.

The Safety Board requested that the Massachusetts Institute of Technology’s

Lincoln Laboratory review the LLWAS data. According to public hearing testimony by

Lincoln Laboratory’s Source Scientist for the LLWAS and TDWR

algorithms, the

FAA Order 7110.65, section 2-9-2, states that towers are required to make a new ATIS recording

whenever the following situations occur: any new official weather is received (even if there has

not been a change in values), runway braking action reports indicate that runway braking is worse

than the description included in the current ATIS broadcast, and any other pertinent information

has changed (for example, the runway or the instrument approach in use and new or canceled notices

to airmen, pilot reports, or HIWAS updates).

Little Rock National Airport is also known as Adams Field.

According to the FAA’s Program Manager for LLWAS, the FAA plans to upgrade the current

system in May 2002 to an LLWAS-RS, which will involve adding and relocating sensors and upgrading

software. Other airports will have their LLWAS systems upgraded to either the LLWAS-RS or

LLWAS-NE version, which will involve a network expansion. LLWAS-NE versions will be integrated

with Terminal Doppler Weather Radar (TDWR) systems or the Weather Systems Processor (WSP,

see section 1.18.5.1).

TDWR is a C-band radar installed at 41 major U.S. airports. TDWR provides timely and

accurate detection of hazardous windshear in and near airport terminal approach and departure areas.

It also provides microburst, gust front, wind shift, and precipitation intensity information.

Factual Information 26 Aircraft Accident Report

LLWAS alerts issued on the night of the accident “were very credible and gave a good

interpretation of what was going on.” This witness also testified that, after the airplane

landed on the runway, the LLWAS centerfield sensor surged to 41 knots and that a

microburst had begun to impact the airport by 2352.

(An LLWAS alert related to this

event was recorded at 2352:10.)

The Safety Board reviewed the tapes made by the airport’s surveillance cameras

for additional information on the weather conditions before the time of the accident. The

surveillance cameras recorded heavy rain, strong gusting winds, lightning, and

deteriorating visibility.

1.7.2 National Weather Service Information

The NWS prepared several weather products describing the conditions

surrounding the time of the accident. A terminal aerodrome forecast (TAF), prepared by

the North Little Rock Forecast Office, was issued about 1830 and was amended about

2258.

The amended TAF, which was valid starting about 2300, stated, in part, the

following:

Beginning at 0400Z, winds forecasted from 200º at 12 knots gusting to 20 knots,

visibility greater than 6 miles, scattered clouds at 2,500 feet, ceiling overcast at

6,000 feet. Temporary condition between 0400Z and 0600Z, winds variable at

25 knots gusting to 40 knots, visibility 1 mile in thunderstorm, heavy rain and

mist, ceiling overcast at 1,500 feet in cumulonimbus clouds.

NWS in-flight weather advisories notify pilots en route of the possibility of

encountering hazardous flying conditions that may not have been forecast at the time of

their preflight briefing. Two of the five NWS in-flight weather advisory categories—

Convective SIGMET and Severe Weather Forecast Alert—were in effect at the time of

the accident.

Convective SIGMET

No. 15C, prepared by the NWS Aviation Weather Center in

Kansas City, Missouri, was issued about 2255 and was valid until 0555 on June 2, 1999.

The advisory was broadcast in its entirety to the flight crew about 2304 and stated the

following:

A microburst is a severe localized wind blasting down from a thunderstorm. A microburst

usually covers an area of less than 2 1/2 miles in diameter and lasts less than 20 minutes.

TAFs are normally issued every 6 hours with amendments issued as conditions warrant.

The NWS Aviation Forecaster in North Little Rock indicated that he amended the 1830 TAF

because, by 2250, he realized that the line of thunderstorms and heavy rain would be impacting

the airport within the next hour. The 1830 TAF was included in the preflight weather package

to the flight crew (see section 1.7.3).

The other three NWS in-flight weather advisory categories are SIGMETs, center weather advisories,

and airman’s meteorological information (better known as AIRMETs).

A Convective SIGMET implies severe or greater turbulence and microburst/windshear activity.

Factual Information 27 Aircraft Accident Report

Attention all aircraft, Convective SIGMET 15 Central valid until 0555 Zulu for

Arkansas and Oklahoma...area of severe thunderstorms moving from 300 at

20 knots, tops above flight level 450, hail to two inches, and wind gusts to seven

zero knots possible. Additional hazardous weather information for Arkansas and

Oklahoma available from flight service, flight watch, or HIWAS frequencies.

Severe Weather Forecast Alert No. 357, prepared by the NWS Storm Prediction

Center in Norman, Oklahoma, was issued about 2123 and was valid until 0300 on June 2,

1999. The advisory, which encompassed portions of northern Texas, northwest

Louisiana, Arkansas (including Little Rock), and southeast Oklahoma, warned of a few

severe thunderstorms with hail to 2 inches, extreme turbulence, and surface wind gusts to

70 knots; a few cumulonimbus clouds with maximum tops to 50,000 feet; and a mean

storm motion from 280º at 20 knots.

A Weather Surveillance Radar 1988 Doppler (WSR-88D) system located in North

Little Rock (6 miles north-northwest of the airport) provides a three-dimensional volume

scan of the atmosphere at varying degrees of elevation and within a range of 240 miles.

The five lowest elevation angles for this WSR-88D are 0.4, 1.5, 2.4, 3.3, and 4.3º.

The

volume scan process takes 6 minutes.

The WSR-88D’s composite reflectivity image of all elevation scans from 2345

depicted a northeast-to-southwest-oriented band of weather, with several large areas

indicating reflectivities of level 6 (extreme) activity,

encompassing the Little Rock

airport area. Reflectivities over the airport ranged from 50 to 64 decibels (dBz), or level 5

(intense) to 6 (extreme) activity. The 2351 composite reflectivity image continued to

depict the large weather band surrounding the airport and the 50- to 64-dBz reflectivity

range over the airport.

The WSR-88D’s base reflectivity images provided the radar reflectivities at the

individual elevation scans. The images documented the line of thunderstorms moving

across the Little Rock airport area during flight 1420’s approach. The WSR-88D 0.4º

elevation scan of base reflectivity that was completed at 2334:27 depicted reflectivities

of 50 dBz, or NWS level 5 (intense) activity over the northwest section of the airport. By

2340:28, the 0.4º base reflectivity image depicted a large area of activity over the airport

with reflectivities of 45 dBz, or level 4 (very strong) activity, and greater. The two

strongest areas of activity were located approximately 5 miles west-northwest and

northeast at this time. Another area of activity located approximately 3 miles southwest

of Little Rock was beginning to move toward the east-southeast with reflectivities

reaching 54 dBz, or level 6 (extreme) activity. Figures 7 and 8 show the base reflectivity

products for 2334:27 and 2340:28, respectively.

The 0.4º elevation scan covered conditions near the surface of the airport. The 1.5º elevation

scan covered conditions encountered in the area approaching the airport.

The NWS categorizes reflectivity according to a six-level intensity scale; level 6 is the highest,

measuring greater than 54 decibels. The other levels are 1, very light; 2, light to moderate; 3, strong;

4, very strong; and 5, intense.

Factual Information 28 Aircraft Accident Report

Figure 7. 0.4° Base Reflectivity Scan at 2334:27

Figure 8. 0.4° Base Reflectivity Scan at 2340:28

Factual Information 29 Aircraft Accident Report

The WSR-88D’s 0.4º elevation scan of base reflectivity that was completed at

2345:57 indicated that the area over the airport and the approach end of runway 4R

reached reflectivities of 54 dBz, or level 5 (intense) activity, and that the area

approximately 5 miles west of runway 4R reached maximum reflectivities of 61.5 dBz, or

level 6 (extreme) activity. Figure 9 shows the base reflectivity product for this time

period.

Figure 9. 0.4° Base Reflectivity Scan at 2345:57

At 2351:59, the 0.4º base reflectivity image depicted a maximum of 52 dBz, or

level 5 (intense) activity, along the flight track, with maximum reflectivities of 60 dBz, or

level 6 (extreme) activity, located 1 1/2 miles to the northwest. Figure 10 shows the base

reflectivity product for this time period. Also, the 1.5º elevation scan at 2353:03 indicated

returns of 58 dBz, or level 6 activity, with maximum reflectivities of 62.5 dBz 1/2 mile

from the approach end of runway 4R.

The WSR-88D’s radial velocity image (showing components of the wind speed

that are coming directly toward or away from the radar) for 2346:29 indicated that winds

from the northwest at 15 meters per second (30 knots) were over runway 4R. Figure 11

shows the radial velocity image for this time period with the airplane’s flight track. At

2352:30, the radial velocity image depicted winds from the northwest at 18 meters per

second (36 knots) over the runway with winds reaching 30 meters per second (60 knots)

within 1/2 mile of the approach end of the runway.

Factual Information 30 Aircraft Accident Report

Figure 10. 0.4° Base Reflectivity Scan at 2351:59

Figure 11. Radial Velocity Image Surrounding the Time of the Accident

Factual Information 31 Aircraft Accident Report

The 1900 upper air sounding (that is, an evaluation of the conditions supporting

the development of severe storms) from the NWS Forecast Office in North Little Rock

was valid beginning about 2000. The upper air sounding included several stability

indexes, which indicated that the atmosphere was unstable with a high potential for

severe thunderstorms.

In addition, the NWS issued two public weather warnings for severe

thunderstorms in the Little Rock area. The first warning was issued about 2156 and was

valid until 2245, and the second warning was issued about 2317 and was valid until 0020

on June 2, 1999. Both warnings indicated the threat of strong winds and the potential for

hail, and the general public was warned to stay indoors until the storms passed. These

weather warnings were made available to the local Little Rock area, and the tower

received these warnings through its direct line with the NWS. The FAA does not require

controllers to provide pilots with public weather warnings because they are not

considered aviation products and contain no aviation references. The local controller did

advise a pilot of a light multiengine airplane, which departed the airport about 2328, of

the second public weather warning because that airplane was traveling northbound

toward the counties covered by the warning. Flight 1420 was approaching the airport

from the south and was not on the tower’s frequency at the time.

1.7.3 American Airlines Weather Information

At Dallas/Fort Worth, the flight crew viewed a graphical display of the weather

radar and received a preflight weather package that contained weather information issued

about 2205. The information included reports, forecasts, and notices to airmen for the

departure, destination, and alternate airports and the current in-flight weather advisories

for the routes. A thunderstorm SIGMEC [significant meteorological condition] stated the

following:

En route thunderstorm SIGMEC. Valid from 0255Z through 0800Z on June 2,

1999. Over Texas, Louisiana, Arkansas, and Oklahoma. Coverage widely

scattered area of thunderstorms located from 10 miles northeast of Fayetteville,

AR, to Little Rock, AR, to Texarkana, AR, to Paris, TX, to Fayetteville, AR, then

back to 10 miles northeast of Fayetteville. Thunderstorms moving to the east at

20 knots. Maximum tops at and above 50,000 feet. Outlook, thunderstorms

increasing through 0600Z and then decreasing.

Other in-flight weather advisories included in the preflight weather package were

a brief of NWS Severe Weather Forecast Alert No. 357 and NWS Convective SIGMET

No. 11C, which indicated an area of severe thunderstorms moving from 300º at 20 knots

with cloud tops above 45,000 feet and the possibility of hail to 2 inches and wind gusts to

70 knots. The SIGMET was valid over portions of Arkansas (including Little Rock),

Oklahoma, and Texas until 2355, but Convective SIGMET 15C replaced it about 2255.

The preflight weather information also included the 2153 ASOS observation, the TAF

A SIGMEC is an American Airlines-issued weather advisory about conditions that may influence

the safety of flight operations.

Factual Information 32 Aircraft Accident Report

issued at 1830,

and airport field conditions for Little Rock (all runways were open and

wet, with 0 inch of water and no reports of braking action problems).

The American Airlines flight dispatcher was using a high-resolution Weather

Services International radar mosaic product to track the weather in relation to an

airplane’s flightpath. The radar mosaics eliminate false echoes and ground clutter, but the

product is delayed by several minutes.

The flight dispatcher indicated, during public

hearing testimony, that he received updates on the weather every 15 minutes but that the

radar data were 5 to 15 minutes old by that time. The dispatcher also indicated that he did

not have access to real-time WSR-88D single-site weather data or TDWR.

The flight dispatcher’s 2254 ACARS message to the flight crew stated the

following:

Right now on radar there is a large slot to Little Rock. Thunderstorms are on the

left and right, and Little Rock is in the clear. Sort of like a bowling alley

approach.[

] Thunderstorms are moving east-northeastward toward Little Rock

and they may be a factor for our arrival. I suggest expediting our arrival in order

to beat the thunderstorms to Little Rock if possible.

About 2257, the flight crew sent an ACARS message, requesting weather

information for Little Rock airport. The flight crew received the 2153 ASOS observation,

which was already included in the preflight weather package. The 2253 ASOS

observation was not available to the flight crew until after 2300.

The flight dispatcher stated that he did not see the 2258 amended TAF (which

changed the wind direction and was more specific regarding the impact period of the

thunderstorm). The dispatcher did not receive the amended TAF because it was issued

when American’s primary weather circuit from the FAA was preparing to receive hourly

weather observation information.

This TAF, which was valid beginning about 1900, indicated that the visibility beginning about

2300 would be greater than 6 miles and that the probability of a thunderstorm between 2300 and

0300 on June 2 was greater than 50 percent.

To eliminate ground clutter, a computer program compares the images with previous observations

and local ground clutter patterns and then overlays the images with the other data to correct for

beam height and distance errors.

The flight dispatcher indicated that the Internet has sites that use Doppler weather radar technology

to depict weather information but that he did not have access to such information.

In public hearing testimony, the flight dispatcher indicated that the “bowling alley” message

was meant to be as brief and concise as possible but give the pilots an image of what to expect

because the airborne weather radar was not able to show a full area of coverage. According to

the CVR, the captain stated at 2326:52, “this is the bowling alley right here,” and at 2332:31,

“down the bowling alley.”

Factual Information 33 Aircraft Accident Report

1.7.4 Additional Weather Information

1.7.4.1 Lightning Data

American Airlines provided the Safety Board with a display from the National

Lightning Detection Network

depicting 1,177 lightning strikes over the state of

Arkansas between 2345 and 2400. The Board used lightning verification reports from

this network to determine that, between 2346 and 2351, 903 cloud-to-ground lightning

strikes were detected within 20 miles of the center of the airport and 46 cloud-to-ground

lightning strikes were detected within 5 miles of the center of the airport.

1.7.4.2 Witness Statements

The Safety Board interviewed two witnesses to the weather conditions

surrounding the time of the accident. One witness, a cross-country truck driver in the

vicinity of the airport, indicated that “torrential rain” was occurring when he saw the

accident airplane “coming in cocked with the wings tilted to the right.” He also stated that

“strong gusty winds,” “intermittent golf ball size hail,”

and “almost continuous”

lightning were occurring after he saw the airplane. The other witness was waiting in the

airport terminal for his wife. He indicated that hail had started just before the airplane

landed and that the hail was “really hard” when the airplane touched down. He stated that

thunder and lightning were occurring simultaneously. He further stated that the thunder

vibration “felt like it would break the glass [on the windows in the terminal]” and that the

hail had cracked the glass.

1.7.4.3 Windshear Hazard Study

A research scientist with the National Aeronautics and Space Administration’s

(NASA) Langley Research Center in Hampton, Virginia, examined available weather

data and conducted modeling simulations to determine the turbulence, windshear, and

crosswind hazards surrounding the time of the accident. The draft report on this research

indicated that a strong “bow echo” squall line system approached the airport during the

time of the accident and produced hazardous crosswinds.

According to the report, a bow

echo refers to a radar echo that appears to undergo a forward acceleration at its midpoint,

thus forming a bulge in the radar signature, and is known to harbor severe weather.

The

report also indicated that the accident airplane might not have encountered hazardous

levels of windshear.

The report suggested that crosswind and windshear hazards were not synonymous

because each affected airplane control differently. The report indicated that the hazards

Global Atmospherics, Inc., in Tucson, Arizona, operates the National Lightning Detection Network.

The leading edges of the airplane’s wings showed no damage that would be consistent with

a heavy hail encounter.

Proctor, Fred H. 1999. Investigation of the Storm Associated With the 1 June 1999 Aircraft

Accident at Little Rock, Arkansas

. NASA Langley Research Center.

The report added that a bow echo could be detected with conventional and Doppler radar.

Factual Information 34 Aircraft Accident Report

associated with a crosswind threat were “collision with obstacles, lack of control

authority on touchdown resulting in damage to aircraft and injury to passengers, [and]

impaired directional control on the runway.” The report further indicated that the hazard

associated with a windshear threat was “flight into terrain.” In addition, the report stated

that hazardous crosswinds and windshear could affect airplanes at low altitudes during

the approach and departure phases of flight but that crosswind could only affect airplanes

during the takeoff and landing flight phases. The report concluded that advisories and

alerts for hazardous crosswinds could be developed and implemented into existing

LLWAS systems.

1.8 Aids to Navigation

The Little Rock VOR is on the 113.9 megahertz (mHz) radio frequency and has

distance measuring equipment associated with it. The runway 4R/22L ILS is on the

111.3 mHz radio frequency. No problems with these navigational aids were reported.

1.9 Communications

No communications problems were reported between the flight crew and any of

the air traffic control (ATC) facilities that handled flight 1420.

1.10 Airport Information

Little Rock National Airport is located immediately south of the Arkansas River

and approximately 2 miles east of metropolitan Little Rock at an elevation of 260 feet

msl. The airport is owned by the city of Little Rock and is operated by the Little Rock

Municipal Airport Commission.

Little Rock National Airport has three concrete transverse grooved runways:

4L/22R, 4R/22L, and 18/36. Runway 4R/22L is 7,200 feet long and 150 feet wide and is

equipped with high-intensity runway edge lights

and centerline lights. (Runway 4L/22R

is 8,273 feet long and 150 wide; runway 18/36 is 5,124 feet long and 150 feet wide.)

Runway 4R is equipped with a medium-intensity approach lighting system with runway

alignment indicator lights. Runway 22L has a medium-intensity approach lighting system

with sequenced flashing lights and a precision approach path indicator.

There are published ILS approaches for runways 4L, 4R, 22L, and 22R. At the

time of the accident, the ILS equipment for runway 4L/22R was out of service because of

The RVR system log, which records runway edge light settings once per hour indicated that,

about 2344, the runway 4R edge lights were set at step 3. (Step 1 is the lowest edge light setting;

step 5 is the highest.) The RVR log is the only available recorded information on the light setting

before the accident. In a postaccident interview, the controller could not recall whether he changed

the light setting after 2344 in response to the decreasing visibility.

Factual Information 35 Aircraft Accident Report

the installation of upgrades. (Although the upgrade work had been completed, this ILS

equipment remained unusable because it had not yet been flight checked.) Equipment

monitor logs showed that the ILS equipment for runway 4R was operating normally. A

postaccident flight check could not be conducted because the ILS localizer antenna was

damaged during the accident sequence.

The FAA’s Lead Systems Engineer for Navigation and Landing, in testimony at

the Safety Board’s public hearing on this accident, stated that a new-generation RVR

system was installed for runway 4R/22L in August 1996. The new-generation system

consists of an infrared transmitter source and an infrared transmitter receiver, which can

detect rain, mist, or snow, and a runway sensor light, which computes the RVR.

The

digital readout is transmitted to the tower, and local controllers use this information to

indicate to pilots upward, downward, or steady trends. The display can be updated every

2 seconds based on the average for the preceding minute.

The RVR sensors are mounted about 18 feet above the ground and are located at

both ends of the runway, near the ILS glideslope antennas and the painted touchdown

zone markers, to detect the touchdown and rollout RVRs.

The system has a quarterly

maintenance period involving calibration and certification. The system’s last

maintenance before the accident was on May 17, 1999. According to the Lead Systems

Engineer, there were no indications of problems with the RVR system on the night of the

accident. Also, an inspection was done on June 2, 1999, during which the system was

recertified.

An archiving function within the RVR data processor unit can retain 12 hours of

RVR data (the previous 2 hours and the following 10 hours). The unit must be set to start

an event log within 2 hours of an occurrence, or the data will be lost. The 1-minute RVR

data surrounding the time of this accident were not retained because an event log was not

started within 2 hours.

RVR data at Little Rock are not reported directly to the airport’s ASOS unit; the

contract weather observer must obtain the 10-minute average visibility reading from the

tower for inclusion in an observation.

The manager for the contract weather observers at

Little Rock stated in a postaccident interview that the observers would not delay the

transmission of an ASOS observation for an RVR reading.

The earlier system used transmissometers—light sources that transmitted beams to receivers

mounted on stanchions—to determine the RVR. The new system has been installed at about 150

U.S. airports.

Runway 4R/22L was not equipped with sensors to detect the midpoint RVR.

Annex 3 to the Convention on International Civil Aviation (Chicago Convention), Chapter 4

(dated May 11, 1998), includes recommended practices for Contracting States regarding meteorological

observations and reports. Paragraph 4.1.8 states that “[weather] observation systems should include

automated equipment for measuring or evaluating, as appropriate, and for monitoring and remote

indicating of surface wind, runway visual range, cloud height, and…other meteorological parameters

affecting landing and take-off operations.” The United States is 1 of 185 countries that are signatories

to the Chicago Convention.

Factual Information 36 Aircraft Accident Report

Little Rock National Airport is certified by the FAA as an aircraft rescue and fire

fighting (ARFF) index C facility.

At the time of the accident, the FAA’s most recent

annual airport certification inspection was from July 29 to 31, 1998. The last full-scale

airport disaster drill was held in October 1996.

According to the Airport Manager, Little Rock’s Airport Certification Manual

was approved by the FAA on May 10, 1999. Title 14 CFR 139.205 states that the manual

must include “a grid map or other means of identifying locations and terrain features on

and around the airport which are significant to emergency operations.” The Manager of

the Airport Safety and Certification Branch, FAA Office of Airport Safety and Standards,

indicated that the airport’s Emergency Access Plan (a diagram that shows emergency

vehicle access points) was considered by the FAA to be a satisfactory alternative means

of identifying locations and terrain features. The access plan was part of the airport’s

FAA-approved Airport Emergency Plan and was included in the Airport Certification

Manual.

1.10.1 Runway 4R/22L Safety Areas

Runway 4R/22L was opened for aircraft operations in September 1991.

According to an FAA November 23, 1999, memorandum, runway 4R/22L was built to

abate noise over the communities located southeast of the airport. Runway 4R (oriented

southwest to northeast) was intended to be used primarily for takeoffs, and runway 22L

(oriented northeast to southwest) was intended to be used primarily for landings.

The memorandum detailed the site constraints on the design and construction of

the runway, including a flood plain of a creek and rising terrain to the southwest and the

flood plain of the Arkansas River to the northeast. The total length available for the

runway and its associated safety areas was 8,650 feet. A runway length of 7,200 feet was

needed,

so 1,450 feet was available for the safety areas. The runway was designed so

that a 1,000-foot safety area extended from the southwest (runway 22L) and a 450-foot

safety area extended from the northeast (runway 4R).

The specifications for runway safety areas are contained in 14 CFR 139.309,

“Safety Areas,” which became effective on January 1, 1988. The specifications state the

following:

(a) To the extent practicable, each certificate holder shall provide and maintain

for each runway and taxiway which is available for air carrier use—

According to 14 CFR 139.315 and 139.317, an ARFF index C facility is to have two or

three firefighting vehicles with a total of at least 3,000 gallons of water and aqueous film forming

foam. The Little Rock airport has three ARFF vehicles, each with the capacity to carry 1,500 gallons

of water and 200 gallons of aqueous film forming foam.

The FAA’s memorandum did not specify how the 7,200-foot length requirement was established.

The FAA’s memorandum stated that accident data indicated that runway overruns during landing

are twice as likely to occur than underruns, which explains the decision to placed 1,000-foot runway

safety area at the southwest end of the runway.

Factual Information 37 Aircraft Accident Report

(1) If the runway or taxiway had a safety area on December 31, 1987, and

if no reconstruction or significant expansion of the runway or taxiway

was begun on or after January 1, 1988, a safety area of at least the

dimensions that existed on December 31, 1987; or

(2) construction, reconstruction, or significant expansion of the runway or

taxiway began on or after January 1, 1988, a safety area which

conforms to the dimensions acceptable to the Administrator at the time

construction, reconstruction, or expansion began.

Paragraph (c) of this section states that “FAA Advisory Circulars in the 150 series

contain standards and procedures for the configuration and maintenance of safety areas

acceptable to the Administrator.” Table 3-3 in FAA Advisory Circular

(AC) 150/5300-13, “Airport Design,” dated June 5, 1991, indicates that the runway

safety area length should be 1,000 feet.

The FAA’s memorandum stated that the Little Rock airport operator was required

to comply with Section 139.309(a)(1) because, according to FAA records, five grants for

the construction of runway 4R/22L were issued before January 1, 1988 (with the first one

in 1982). The memorandum also stated that the runway safety areas for runway 4R/22L

met the regulatory requirements.

In July 2001, the Little Rock Airport Manager indicated that the airport was

working with the Army Corps of Engineers, the Federal Emergency Management

Agency, Pulaski County, and the cities of Little Rock and North Little Rock to extend the

runway safety area at the departure end of runway 4R to 1,000 feet. The airport manager

also indicated that, if all hydraulic studies and permits are approved, the runway safety

area extension could be completed by July 2002.

1.10.2 Runway 22L Approach Lighting System Support Structure

The runway 22L approach lighting system support structure, which is located in a

flood plain area of the Arkansas River, is not considered frangible. (The top portion of the

structure—its walkway—is considered frangible.) According to AC 150/5300-13, a

frangible navigational aid “retains the structural integrity and stiffness up to a designated

maximum load, but on impact from a greater load, breaks, distorts, or yields in such a

manner as to present the minimum hazard to aircraft.”

The FAA’s Lead Systems

Engineer for Navigation and Landing stated that a structure mounted in a flood plain,

such as the runway 22L approach lighting system structure, cannot be frangible because

of the possibility of moving water, ice, and floating debris.

The Lead Systems Engineer stated that there are benefits to placing or

establishing frangible structures in runway safety areas, citing the November 12, 1995,

American Airlines flight 1572 accident in East Granby, Connecticut, as an example. In

AC 150/5300-13 indicates that frangible navigational aids include electrical and visual air

navigational aids, lights, signs, and associated supporting equipment.

Factual Information 38 Aircraft Accident Report

that accident, a McDonnell Douglas MD-83 was on final approach, at night and in strong,

gusty wind conditions, to runway 15 at Bradley International Airport in Windsor Locks,

Connecticut, when it collided with trees. The airplane landed short of the runway and

then struck an ILS localizer array (a frangible structure), which properly broke into pieces

and thus minimized further damage to the airplane.

The Lead Systems Engineer also

indicated that the FAA’s effort to replace nonfrangible structures with frangible ones

(where appropriate) has been a relatively slow process but that work at about

three-fourths of the 450 sites with such structures has been accomplished.

1.10.3 Runway 4R Assessments

1.10.3.1 Tire Marks

Tire marks

consistent with those from the left main landing gear began 5,228 feet

before the departure end of runway 4R, about 1 foot to the right of the runway’s centerline,

and continued for 149 feet (5,079 feet before the departure end). Tire marks were not

present on the runway’s next 207 feet but began again 4,872 feet from the departure end of

the runway, about 13 feet to the right of the centerline, and continued until the edge of a

gravel downslope located 459 feet beyond the end of the runway. The tire marks were

approximately 98 feet to the left of the centerline at the end of the runway surface. Also,

the grass to the left of the runway showed tire marks consistent with those from the left

main landing gear beginning about 710 feet before the departure end of the runway.

Tire marks consistent with those from the right main landing gear began

4,303 feet before the departure end of runway 4R, about 47 feet to the right of the

runway’s centerline, and continued until the edge of the gravel downslope located

459 feet beyond the end of the runway. The tire marks were approximately 82 feet to the

left of the centerline at the end of the runway surface. In addition, the grass to the left of

For more information, see National Transportation Safety Board. 1996.

Collision With Trees

on Final Approach, American Airlines Flight 1572, McDonnell Douglas MD-83, N566AA, East Granby,

Connecticut, November 12, 1995.

Aircraft Accident Report NTSB/AAR-96/05. Washington, DC.

One airport with nonfrangible structures that need to be replaced is Chattanooga Metropolitan

Airport, Tennessee. The Safety Board received a copy of a June 3, 1999, letter from the President

of the Chattanooga Metropolitan Airport Authority to the FAA Administrator. The letter expressed

concern about nonfrangible poles in the airport’s runway safety area and cited a 1973 accident involving

a DC-9 airplane that struck approach lights mounted on such poles while landing during a thunderstorm.

No one was killed in that accident, but the airport authority recognized that the nonfrangible poles

were a hazard. The letter further indicated that the approach lighting system was scheduled for replacement

but that the date continued to be postponed. The Lead Systems Engineer stated that the replacement

lighting system has been delivered by the manufacturer and is in storage but has not yet been installed

because of a lack of funding.

The tire marks were more whitish in color than the surrounding off-white concrete surface.

At those points where the tire marks crossed white runway paint markings, the white paint was

cleaner and whiter than the surrounding paint. At those points where the tire marks crossed black

runway paint markings, the black paint was cleaner and darker than the surrounding paint, and there

were no white marks. For information on the relationship between the tire marks and runway friction,

see section 1.18.1.

Factual Information 39 Aircraft Accident Report

the runway showed tire marks consistent with those from the right main landing gear

beginning about 465 feet before the departure end of the runway.

Tire marks consistent with those from the nose gear began 5,079 feet before the

departure end of runway 4R, about 6 feet to the right of the runway’s centerline, and

continued for 207 feet (4,872 feet before the departure end). Tire marks were not present

on the runway’s next 119 feet but began again 4,753 feet before the departure end of the

runway, about 18 feet to the right of the centerline, and continued with occasional

interruption until the left edge of the ILS localizer array located 411 feet beyond the end

of the runway. The tire marks were approximately 67 feet to the left of the centerline at

the end of the runway surface.

Section 1.12.1 contains a photograph showing the tire marks off the end of the

runway.

1.10.3.2 Runway Surface Information

The following assessments of runway 4R were made starting the day after the

accident:

• A visual inspection revealed several small holes (about 4 inches in diameter)

in the pavement on the approach end of runway 4R. Some of the holes had

been filled with epoxy. There was evidence of light and medium rubber

deposits on the runway surface. No evidence of structural pavement failure

was present.

• Friction survey tests, using an Airport Surface Friction Tester, were conducted

at 40 to 60 mph in both runway directions. AC 150/5320-12C, “Measurement,

Construction, and Maintenance of Skid-Resistant Airport Pavement Surfaces”

(dated March 18, 1997), states that the maintenance planning friction levels at

40 and 60 mph are 0.60 and 0.47, respectively. The average friction readings

for runway 4R were 0.69 at 40 mph and 0.55 at 60 mph.

• The groove specifications for runway 4R were 1/4-inch deep by 1/4-inch wide

and spaced 2 inches apart (center to center). An inspection of selected grooves

found that all met the 1/4-inch width and 2-inch spacing specifications.

Thirteen panels (19 feet in length) along the full length of runway 4R,

including several located by the marks left by the main landing gear tires,

were selected for an inspection of groove depth. The average depth was

1/4 inch.

• Field measurements for the runway’s transverse slope (from crown to

shoulder) averaged 1.42 percent. The construction drawings indicated a

transverse slope of 1.5 percent.

Factual Information 40 Aircraft Accident Report

• The average surface texture depth measurements for runway 4R were

0.055 inch (clean grooved concrete) and 0.015 inch (clean ungrooved

concrete).

The average surface texture depth for the rubber-coated, grooved

touchdown area of runway 4R was 0.050 inch. (The average surface texture

depth for the rubber-coated, grooved touchdown area of runway 22L was

0.055 inch.) AC 150/5320-12C states that, when the average texture depth

measurement falls below 0.045 inch (the recommended average texture depth

for newly constructed pavements), the airport operator should conduct texture

depth measurements each time a runway friction survey is conducted. The AC

indicates that corrective actions need to be taken when the average texture

depth is below 0.030 inch.

In addition, the transverse water flow characteristics of runway 4R were measured

by water drainage tests performed on November 16, 1999. The tests involved the release

of water from a tanker truck hose onto the centerline of the runway at 100- and 500-feet

increments, starting at 5,608 feet before the departure end of the runway (just before the

initial left main landing gear tire marks). Dry and wet runway drainage tests were

performed, and the winds were calm at the time of the tests.

The test data indicated that the average flow rates from the left to right shoulder

edges when the surface was wet were about 10 percent higher than the rates when the

surface was dry. Also, a senior research engineer from NASA’s Langley Research Center

(who was a member of the Airplane Performance Group) determined that, with no winds

and the cross (or transverse) slope and surface texture depth values measured after the

accident, runway 4R was capable of handling rainfall rates up to 1.4 inches per hour

before surface flooding (that is, water depths reaching 0.1 inch and greater) would occur

at 15 feet from the centerline. Crosswinds from the left side of the runway, which existed

at the time of the accident, would result in deeper water and more flooding on the left side

of the runway and shallower water and less flooding on the right side of the runway. The

NASA Langley engineer testified that the measurements taken in November 1999 of

runway 4R indicated that the cross slope was “very uniform” and provided good drainage

of the water from the centerline to the shoulder.

1.10.4 Air Traffic Control Tower Information

The Little Rock ATCT is located on the terminal building and includes a terminal

radar approach control (TRACON). The tower cab has positions for cab coordinator,

local control-1 and -2, ground control, and flight data. The local control-1 and ground

control positions have panels showing readings for the LLWAS system sensors. The

local control-1 position also has a Digital Bright Radar Indicator Tower Equipment

(D-BRITE) radar display. A System Atlanta Information Display System–4 monitor,

located to the right of the flight data position, contains airport-related information and

displays hourly and special ASOS observations.

The transverse grooves end about 13 feet from the runway’s left and right shoulder edges.

Factual Information 41 Aircraft Accident Report

ATC radar data are provided by an Airport Surveillance Radar (ASR)-8 sensor

located on the airport between runways 4L and 4R. Radar data processing is performed

by an Automated Radar Terminal System IIE system that is linked with the TRACON

and D-BRITE radar display.

At the time of the accident, the ATCT was operating with midnight shift

staffing—one local controller and one controller-in-charge. All approach and tower

control positions were combined at the local control-1 position.

The local controller who was handling flight 1420 at the time of the accident was

initially certified as a control tower operator in December 1986. The controller served as

an air traffic controller for Mather Air Force Base in California for 3 years before

beginning work with the FAA. According to his training records, the controller started at

the Midway ATCT in Chicago in October 1988 and became a certified professional

controller there in June 1990. He transferred to the Little Rock ATCT in November 1992

and has been a certified professional controller there since September 1993. His last

medical certification was in October 1997.

On the day of the accident, the local controller had worked the 0600 to 1400 shift.

Afterward, he went home, slept for about 4 hours, and returned to the ATCT about 2250

for the 2300 to 0700 shift. He received a position relief briefing from the evening shift

controller and then called the TRACON to combine the radar positions in the tower cab.

The controller-in-charge was in the TRACON performing administrative duties.

Flight 1420 was the first air carrier operation of the local controller’s shift. The

controller stated that he first saw the airplane when it was about 1 mile out during final

approach and that the landing appeared “normal” and “within the touchdown zone.”

Because of the reduced visibility, the controller lost sight of the airplane during its

rollout. The ATCT transcript indicated that, at 2350:54, the controller requested that

flight 1420 report clear of the runway. The controller attempted to contact flight 1420

five more times, between at 2351:16 and 2353:22, and called the ARFF units on the crash

phone at 2352:00.

Section 1.15.3 provides information on the emergency response.

The local controller also called the controller-in-charge, asking for his assistance.

When the controller-in-charge arrived in the tower cab, the local controller informed him

about the possibility of an accident. The ATCT transcript indicated that, about 0003:16,

the ARFF units reported that they had located the airplane off the end of the runway. The

controller-in-charge then began administrative notification activities. About 0015, the

controller-in-charge relieved the local controller, at which time the local controller

continued the notification process.

In a postaccident interview, ARFF personnel indicated that the call on the crash phone was

received about 2355—3 minutes later than the controller reported initiating the call. Because the

ATC and CVR times could be fairly well correlated, the ARFF response times in this report are

based on ATC times.

Factual Information 42 Aircraft Accident Report

1.11 Flight Recorders

1.11.1 Cockpit Voice Recorder

The accident airplane was equipped with a Fairchild model A-100A CVR, serial

number 53282. The exterior of the CVR showed no evidence of structural damage but

was coated with soot. The interior of the CVR and the tape sustained no apparent heat or

impact damage.

The CVR was sent to the Safety Board’s audio laboratory in Washington D.C., for

readout and evaluation. The CVR data started at 2319:44 and continued uninterrupted

until 2350:48.1 when electrical power to the CVR ceased. The recording consisted of

four channels of “good quality” audio information.

The four channels contained the

cockpit area microphone, the captain’s audio panel, the first officer’s audio panel, and the

interphone and public address system. A transcript was prepared of the entire 31-minute

4-second recording (see appendix B).

No sounds that were consistent with the arming or the deployment of the spoilers

could be detected on the CVR tape. Two flight tests were conducted on August 27, 1999,

to determine whether such sounds could be detected on a CVR recording. Both flight

tests were conducted on an American Airlines revenue passenger flight, and both

airplanes were MD-82 models equipped with a Fairchild model A100A CVR that was

similar to the one installed on the accident airplane. As part of the flight tests, the

nonflying pilot verbally confirmed when the spoilers were armed and deployed.

The first flight test was conducted on American flight 1829 from Ronald Reagan

Washington National Airport to Chicago-O’Hare. The second flight test was conducted

on American flight 154 from Chicago to Washington, D.C. Both tests revealed that the

spoiler arming and automatic deployment could be clearly heard on the CVR recordings.

In fact, the captain on the second test flight attempted to arm the spoiler handle very

slowly to make minimum noise, but a definite “click” sound was recorded on the CVR as

the spoiler handle was lifted.

1.11.2 Flight Data Recorder

The accident airplane was equipped with an L3 model FA2100 FDR, serial

number 00718. The FDR used solid-state flash memory technology as the recording

medium and was configured to digitally record a minimum of 25 hours of operational

data before the oldest data were overwritten.

The Safety Board ranks the quality of CVR recordings in five categories: excellent, good,

fair, poor, and unusable. For a recording to be considered good quality, most of the crew conversations

need to be accurately and easily understood. The transcript developed from the recording might indicate

several words or phrases that were not intelligible; such losses are attributed to minor technical

deficiencies or momentary dropouts in the recording system or simultaneous cockpit/radio transmissions

that obscure one another.

Factual Information 43 Aircraft Accident Report

The FDR was sent to the Safety Board’s FDR laboratory in Washington, D.C., for

readout and evaluation.

The exterior of the FDR showed evidence of fire and smoke

damage. The interior of the FDR showed no signs of damage, and the recording was

retrieved from the crash-survivable storage unit. The FDR contained more than 62 hours

of data, and American Airlines provided conversion formulas for the data. Examination

of the recovered data indicated that the FDR operated normally. Data transcribed

included flight 1420’s pushback from the gate at Dallas/Fort Worth and the accident

airplane’s previous landing at Dallas/Fort Worth.

1.12 Wreckage and Impact Information

1.12.1 General Wreckage Description

The Safety Board performed a complete survey of the accident site and airplane

structure. The airplane was found approximately 800 feet beyond the departure end of

runway 4R. Wreckage was found throughout the flood plain located approximately

15 feet below the runway elevation and down a rock embankment. Wreckage was also

found up to 150 feet laterally from the runway 22L approach lighting system and

approximately 500 to 850 feet from the end of runway 4R. No fluid markings or airplane

components were found on the runway surface.

The fuselage had separated into three main sections (forward, center, and aft). The

forward and center fuselage sections were oriented on a magnetic heading of

approximately 115º; the rear fuselage section and the empennage were oriented on a

magnetic heading of approximately 205º. No evidence of fire damage was found in the

forward and center sections of the fuselage, but a postcrash fire had completely

consumed the passenger cabin in the rear fuselage section. The left wing was fractured

and was completely severed near its root and wing tip. The right wing was found attached

to the fuselage. Both engines were attached to their pylons, which were attached to the

fuselage. The nose gear and right main landing gear were sheared from their attachments,

and the left main landing gear had folded into its main gear wheel well.

The airplane’s collision with the approach lighting system crushed the nose of the

airplane rearward and destroyed the left side of the fuselage from the airplane’s nose to

the cockpit’s rear bulkhead and from the beginning of the first-class section aft to the

second row of the coach section. Large sections of the approach lighting system were

intermingled with fuselage structure that had been peeled away from the airplane. The

collision with the approach lighting system also created a hole in the left side of the cabin

that extended from the overhead stowage bins to the cabin floor in the first-class and

coach sections.

For a listing of the parameters recorded by the FDR, see the FDR Group Chairman’s Factual

Report in the public docket for this accident. Data from the accident flight and the airplane’s previous

landings indicated that two parameters—brake pressure left and brake pressure right—were inactive

on N215AA.

Factual Information 44 Aircraft Accident Report

Figures 12 through 14 are photographs of the airplane wreckage. Figure 12 shows

the airplane wreckage, the tire tracks off the end of runway 4R, and the damage to the

runway 22L approach lighting system.

Figure 13 shows the airplane wreckage and the

runway 22L approach lighting system. Figure 14 shows a closer view of the airplane and

the runway 22L approach lighting system. Additional details about the airplane wreckage

are presented in sections 1.12.2 through 1.12.4.

Figure 12. Aerial Photograph of Runway 4R/22L, Airplane Wreckage, and Runway 22L

Approach Lighting System

The runway 22L approach lighting system is supported by steel columns or platform assemblies

that include steel columns. Five steel columns and two platform assemblies were struck by the accident

airplane during its overrun. Each of these seven support structures was spaced 50 feet apart along

a line that coincided with the centerline of runway 4R/22L. The first support structure was located

about 530 feet from the end of the runway, and the seventh was located 830 feet from the end

of the runway.

Factual Information 45 Aircraft Accident Report

Figure 13. Airplane Wreckage and Runway 22L Approach Lighting System

Figure 14. View of Left Side of Airplane Wreckage and Runway 22L Approach

Lighting System

Factual Information 46 Aircraft Accident Report

1.12.2 Spoiler System

The left flight spoilers and left ground spoiler were found in the retracted position.

No damage was noted on either of the flight spoilers. The left ground spoiler panel was

fractured inboard of the actuator hinge attachment. The inboard half of the left ground

spoiler was heavily sooted on its entire upper surface; no soot was found on the outboard

half. The ground spoiler actuator exhibited heat and fire damage and moved freely. The

right flight spoilers and right ground spoiler were found in the retracted position. No

damage was noted on either of the flight spoilers. The right ground spoiler panel was

fractured outboard of the actuator hinge. The mechanical overcenter link was latched.

The cockpit center pedestal had considerable displacement and deformation in the

left downward direction. The spoiler handle was found fully aft. About one-half of the

autospoiler system ARM red indicator stripe was visible, and the handle guide was

resting on the pedestal surface.

Tests performed on the accident airplane’s spoiler system are discussed in

section 1.16.1.

1.12.3 Engines

Both engines showed no evidence of any uncontainments, case ruptures, or

precrash fires. The engines’ low pressure rotors rotated freely, and the fan blades had

minor impact damage on the leading edges of the airfoils. The left engine thrust reverser

was partially deployed, and the right engine’s thrust reverser was completely stowed. The

left engine’s fuel control was in the reverse thrust range, and the right engine’s fuel

control was in the forward thrust range.

On June 8, 1999, the two engines, thrust reversers, and EPR transmitters were

examined and tested at American Airlines Maintenance and Engineering Center. The

engines were able to produce normal-rated takeoff thrust without exceeding operating

limitations. The thrust reversers were able to cycle from the stowed-to-deployed and

deployed-to-stowed positions. The EPR transmitters were found to function normally.

1.12.4 Landing Gear and Brake Assemblies

The left main landing gear was lodged into its wheel well, and the wheel assembly

was oriented 90º clockwise from its normal position. The outboard (No. 1) and inboard

(No. 2) tires were found deflated and exhibited deformation. Both tires had large cuts and

lacerations on the tread surface and sidewalls. The tread depths for the No. 1 and 2 tires

ranged from 0.156 to 0.250 inch. No flat spots were present on either of the tires.

According to Michelin Aircraft Tire Corporation (the manufacturer of the accident airplane’s

tires), flat spots result from skidding without tire rotation, for example, during brake lockup.

Factual Information 47 Aircraft Accident Report

The right main landing gear and its rear spar attachment fitting was found

separated from the rear spar and aft of the right inboard wing. The inboard tire (No. 3)

was found deflated, and the outboard tire (No. 4) was found pressurized at about 195 psi

(unloaded). The tires had large cuts and lacerations on both the tread surfaces and

sidewalls. The tread depths for the No. 3 and 4 tires ranged from 0.094 to 0.313 inch. No

flat spots were present on either of the tires.

The nose gear strut and wheel assembly were found protruding from beneath the

right side of the fuselage belly. Both tires were found deflated, and both inboard wheel

halves were cracked. The tread depths for the left and right tires ranged from 0.125 to

0.188 inch. The nose-wheel steering cylinders were found in the debris field with both

their pistons extended almost equal amounts.

The left main landing gear brake assemblies showed no evidence of overheat

damage but showed evidence of impact damage. The No. 1 brake had approximately

1 inch of wear pin remaining, and the No. 2 brake had 11/16 inch of wear pin remaining.

The right main landing gear brake assemblies showed no evidence of impact or overheat

damage. The No. 3 brake had approximately 3/4 inch of wear pin remaining, and the

No. 4 brake had about 1 inch of wear pin remaining.

The antiskid control box was tested at Crane Hydro-Aire (the manufacturer),

Burbank, California, on August 3, 1999. The control box passed all functional tests

described in the manufacturer’s TP42-807 test procedure. Tests performed on the

accident airplane’s main landing gear tires are discussed in section 1.16.2.

1.13 Medical and Pathological Information

According to the Pulaski County Coroner, the captain and the passengers in seats

3A, 8A, 17B, 18A, and 18B died as a result of traumatic injuries, and the passengers in

seats 19A, 19B, 19D, 27E, and 28D died from smoke and soot inhalation and/or thermal

injuries. Two of the passengers (seats 8A and 28D), died on June 10 and June 16, 1999,

respectively.

Tissue and fluid specimens from the captain were transported to the FAA’s Civil

Aerospace Medical Institute (CAMI) in Oklahoma City, Oklahoma, for toxicological

analysis. The CAMI laboratory performed its routine analysis for major drugs of abuse

and prescription and over-the-counter medications, and the results were negative. The

analysis detected no ethanol in the captain’s blood and tissue specimens.

American Airlines’ Area Medical Director indicated in a September 1, 1999,

letter to the Safety Board that postaccident drug testing was not performed on the first

officer because of his medical condition after the accident and his inability (because of

sedation) to comprehend the documentation requiring knowledge and consent. The letter

According to 14 CFR 830.2, fatalities that occur within 30 days after an accident are to be

included in the total number of fatal injuries.

Factual Information 48 Aircraft Accident Report

also stated that a representative of the FAA’s Office of Drug Testing agreed with the

decision not to test the first officer.

1.14 Fire

A fuel-fed fire erupted between the center and aft fuselage sections after the

impact sequence. The fire spread and eventually consumed the interior of the aft fuselage

section. The ARFF trucks arrived at the accident scene about 0008. The firefighters

applied water and aqueous film forming foam to the fire and extinguished the exterior fire

within 60 seconds. Firefighters then suppressed smaller fires, including one under the left

wing, by applying aqueous film forming foam for another hour. The Safety Board’s

investigation revealed no evidence of an in-flight fire.

1.15 Survival Aspects

1.15.1 General

The accident airplane’s interior was original equipment installed in 1983. The

airplane was configured with 139 passenger seats, 14 in first class and 125 in coach class.

The cockpit contained two flight crew seats and one observer seat. An aft-facing,

double-occupancy flight attendant jumpseat was located by the 1L exit door; an

aft-facing, single-occupancy flight attendant jumpseat was located by the 2L exit door;

and a forward-facing, double-occupancy flight attendant jumpseat was mounted on the

tailcone exit door. Figure 15 shows the interior airplane configuration and the injuries

sustained by the passengers and crewmembers according to seat location.

The airplane was equipped with an overhead emergency lighting system and a

floor proximity escape path lighting system. The wiring and lamps on both systems

forward of row 17 were tested after the accident using an alternate electrical power

source. All undamaged lamps in both systems operated normally.

The remaining battery

packs and control units from both systems were tested at American Airlines facilities in

Dallas according to the manufacturers’ test procedures. All of the batteries and control

units performed as designed.

Some ARFF and Metropolitan Emergency Medical Service (MEMS) personnel reported that

they found the cabin floor lights illuminated.

Factual Information 49 Aircraft Accident Report

Figure 15. Interior Airplane Configuration and Occupant Injuries

Factual Information 50 Aircraft Accident Report

1.15.2 Evacuation of Passengers and Crewmembers

The first officer could not evacuate the airplane on his own because his left femur

had been fractured during the accident sequence. He was removed from the airplane

wreckage by rescue workers, who had to cut through metal and step on the center

pedestal to extricate him.

The on-scene commander indicated that rescue workers also

had to remove some surviving passengers from the first-class section. The flight

attendants seated on the forward jumpseat were seriously injured in the crash and could

not assist with passenger evacuations. The flight attendants in the aft cabin were able to

assist with passenger evacuations.

Passengers that were forward of the fuselage separation at row 18 escaped

through a large hole on the left side of the first-class section and through a separation in

the fuselage at row 12. The forward entrance (1L) and forward galley (1R) doors could

not be used because of structural deformation of the fuselage. Six passengers seated on

the left side of the first class section (seats 3A and B, 4A and B, and 5A and B) were

ejected in their seats through the large hole. (Five of these six passengers survived.) The

flight attendant seated on the inboard forward jumpseat was carried out of the airplane by

a passenger through the large hole in first class. The flight attendant seated on the

outboard forward jumpseat also left the airplane through the large hole in the first-class

section.

Seven passengers (seated in 17A and B, 18A and B, and 18D through F) were

ejected in their seats into the area between the fuselage sections—aft of row 16 on the

left, aft of row 17 on the right, and forward of row 19. (Four of these seven passengers

survived.) One passenger reportedly exited the airplane at the fuselage break aft of

row 17 on the right side. Two passengers exited the airplane through the fuselage

separation directly forward of row 19.

All four overwing emergency exits were opened by passengers from inside the

cabin. Several passengers who were seated aft of row 19 and forward of row 29 used

three of the four overwing emergency exits to escape. The passengers seated next to the

left and right forward overwing exits reported that they had trouble opening the

respective doors but that someone else was able to open these doors. A fire was outside of

the left forward overwing exit, and no passengers reported or were observed escaping the

airplane through that exit. Four passengers used the right forward overwing exit to

escape. The passenger seated next to the left aft overwing exit was able to open the hatch,

reporting that the door “seemed to pop out easily and quickly.” The passenger seated next

to the right aft overwing exit initially had trouble opening the hatch but was able to get it

open. Four passengers used the left aft overwing exit to escape, and at least 26 passengers

used the right aft overwing exit.

Because emergency personnel had to step on the center pedestal to extricate the first officer

from the wreckage, the documented positions of cockpit instruments after the accident might not

indicate their positions at the time of the accident. In addition, the emergency personnel indicated

that they had turned switches off when they went through the airplane.

Factual Information 51 Aircraft Accident Report

The aft galley (2L) door could not be used because of impact damage from the

runway 22L approach lighting system. The flight attendant seated in the aft galley

jumpseat and four passengers used the gap between the fuselage and the top of the door to

escape.

The flight attendant seated in the aft cabin jumpseat opened the aft bulkhead door

(leading to the tail cone exit) with the assistance of passengers. The flight attendant and

several passengers entered the tail cone area, but the tail cone did not fall away from the

airplane after the flight attendant and at least one passenger pulled the release handle. The

flight attendant and passengers then kicked and jumped on the tail cone and created a gap

between the fuselage and the tail cone that 12 people used to escape from the airplane.

1.15.3 Emergency Response

As discussed in section 1.10.4, the local controller indicated that he called the

ARFF units on the crash phone about 2352. According to ARFF personnel, the controller

stated that an American Airlines airplane was down on runway 4R but did not specify the

approach or departure end of the runway. The ARFF station responded with all available

assets—four firefighters (including a fire captain) and three fire trucks.

The driver of fire

truck No. 2 indicated that the fire trucks had departed the station within 1 minute of the

local controller’s call. The driver of fire truck No. 3 reported that he drove into “blinding

rain and wind.”

All three fire trucks proceeded toward the approach end of runway 4R. (The Little

Rock Fire Department District Chief testified at the Safety Board’s public hearing that all

three units went in the same direction because they were trained to work as a team.)

ARFF personnel indicated that the trucks proceeded slowly (estimated at 15 to 20 mph)

because of the restricted visibility (estimated to be about 100 feet) and unknown location

of the airplane. The fire captain notified Little Rock Central Communications about 2355

that ARFF vehicles were responding to the report of an American Airlines airplane down

on runway 4R.

The ATCT transcript indicated that, at 0000:11, fire truck No. 2 indicated that the

airplane was not at the approach end of runway 4R and asked the controller whether the

fire trucks should “sweep the runway.” Five seconds later, the controller stated that the

airplane was at the departure end of the runway and cleared the fire trucks to proceed in

Title 14 CFR 139.319(j) requires that “sufficient rescue and firefighting personnel are available

during all air carrier operations to operate the vehicles, meet the response times, and meet the minimum

agent discharge rates required by this part.”

Factual Information 52 Aircraft Accident Report

the other direction.

The ATCT transcript also indicated that, at 0001:08, the controller

informed fire truck no. 2 that “I saw him [the airplane] as he went past midfield.”

In postaccident interviews, the firefighters indicated that visibility improved once

past the midfield point on their way to the departure end of runway 4R. Fire truck no. 2

indicated that it experienced sliding on the pavement and noted airplane tire tracks

leaving the runway surface and a missing runway light. When the trucks arrived at the

“22L” painted on the runway, ARFF personnel saw a glow and blowing smoke. The

ATCT transcript indicated that, at 0003:16, fire truck No. 1 reported that the airplane was

off the end of the runway and on fire and stated “this is an alert three” twice.

The fire

captain informed Little Rock Central Communications of the alert 3 status and indicated

that the airplane was off the northern end of runway 4R outside of the airport. The fire

captain requested a “full response” from Central Communications.

The fire trucks were unable to proceed directly to the airplane because of the

slope at the end of the runway. As a result, the trucks had to travel in the opposite

direction to an access road and then turn onto a perimeter road back in the direction of the

accident site. The fire captain and another firefighter indicated that they had to stop to

open a locked perimeter security gate, which took about 20 seconds, before continuing on

the perimeter road to the accident site. The three ARFF vehicles reached the accident

scene about 0008, and firefighters began extinguishing the fire immediately upon arrival.

The Little Rock Fire Department District Chief was monitoring his radios at

Central Communications when he overheard a report from fire truck No. 2 that the tower

had lost communication with an American Airlines airplane. The district chief advised

Central Communications that he would report to the accident scene as the on-scene

commander. The chief arrived on scene and assumed command from the fire captain,

who had set up a command post by that time. The chief reported that most of the fire had

been put out by the time he arrived on scene. The district chief testified that the rain

helped to extinguish the fire and keep it abated. The chief also testified off-airport help

began arriving on scene about 2 to 3 minutes after the ARFF units had arrived.

Metropolitan Emergency Medical Services (MEMS) records indicated that

dispatch received notification of the alert 3 status about 0005 from an off-duty dispatcher

who overheard ARFF radio transmissions on his scanner.

MEMS dispatch confirmed the

The Little Rock Fire Department Chief testified at the public hearing on this accident that

the standard phraseology between the fire department and the control tower regarding the location

of an airplane normally involves runway and taxiway numbers. He indicated that, when the controller

stated that an airplane was down on runway 4R, the firefighters assumed that the airplane was

located at the approach end. The chief further indicated that, if the controller had known that the

airplane was at the departure end of runway 4R, he would have “more than likely” told the rescue

personnel to go to runway 22L.

“Alert 3” is defined in the Airport Emergency Plan as an aircraft accident that has occurred

on or in the vicinity of the airport. It is the most serious of the airport’s three alert categories.

The district chief stated that ARFF personnel’s first priority is fire control so that an escape

path can be provided. The chief also stated that, once the fire is controlled, ARFF personnel can

assume rescue responsibilities and begin treating victims.

Factual Information 53 Aircraft Accident Report

alert 3 status with Little Rock Central Communications about 0006, and dispatch

contacted the MEMS supervisor about 0008. Central communications informed MEMS

about 0011 that an American Airlines airplane was down and that “a big response” was

needed. The MEMS supervisor departed for the accident site about 0012. The supervisor

arrived in the accident area about 0017 but could not report on scene because he

encountered a locked gate adjacent to the airport’s United Parcel Service facility. He

contacted the on-scene commander by radio and was directed to an open gate.

The MEMS supervisor reported on scene about 0022 and set up a triage area. A

MEMS unit (comprising two emergency medical technicians and a paramedic) arrived

2 minutes later and began triage activities. Another MEMS unit arrived on scene

afterward and continued triage activities. Some ambulatory survivors were transported on

a bus to a fire station before MEMS personnel could assess them, so a separate triage area

was established at the fire station.

The Little Rock Fire Department District Chief indicated that, to check for

passengers and crewmembers after the accident, firefighters went inside the airplane and

did “line abreast searches” about 50 to 100 feet on each side of the airplane and from the

airplane to the Arkansas River. The district chief also indicated that, to be sure that

everyone was accounted for, the firefighters repeated this process once daybreak

occurred.

The Little Rock Fire Department reported that 13 engine companies, one ladder

company, 1 heavy rescue unit, 1 hazardous materials unit, and 9 staff vehicles were

involved during the peak of the emergency response. MEMS estimated that

19 ambulances and a number of other medical support or supply vehicles participated in

the emergency response. Also, a medical helicopter made two flights to local hospitals,

transporting four people. Survivors of the crash were taken to Arkansas Children’s

Hospital, Arkansas Heart Hospital, Baptist Medical Center, Baptist Memorial Medical

Center, Southwest Regional Medical Center, St. Vincent Hospital, and the University of

Arkansas Medical Center.

The Little Rock Fire Department Chief testified that, since the accident, six more

ARFF personnel were hired and that the number of personnel for each shift increased to

six. He also indicated that the fire department has looked into getting a Driver’s

Enhanced Vision System (DEVS). (See section 1.18.4 for information about this system.)

At the time of the public hearing on this accident, Little Rock National Airport

had not conducted a formal debriefing with all of the parties involved in the emergency

response on the night of the accident. In July 2001, the Little Rock Airport Manager

stated that the airport conducted individual critiques in February and March 2000 with

the Little Rock Fire Department, the Little Rock ATCT, the Little Rock Office of

The Little Rock Fire Department indicated that its recorded times were ahead of MEMS times

by 2 minutes 46 seconds. Because the fire department’s times appeared to correlate with the original

ARFF times, which were adjusted back by about 3 minutes to correlate with ATC times, no time

adjustments were necessary for MEMS times.

Factual Information 54 Aircraft Accident Report

Emergency Services, airlines that fly into the airport, MEMS, and the Little Rock Police

Department. According to the airport manager, the purpose of these critiques was to

review the airport’s emergency plan. The airport manager also indicated that the airport

conducted a group critique on March 15, 2000. All of the emergency response agencies

were invited to attend and provide comments. Present at the group critique were

representatives from the airport, the Little Rock ATCT, FAA Flight Standards District

Office, six airlines, and three local hospitals. The agenda for the group critique indicated

various issues to be discussed, including communications, access to an accident site,

triage and treatment areas, grid map parameters, and water rescue capabilities. The

Deputy Airport Manager sent a memorandum, dated March 16, 2000, to the emergency

response agencies and group critique participants. The memorandum documented the

recommendations, observations, and concerns that the hospital representatives expressed

during the group critique.

1.15.4 Passenger Statements

The Safety Board interviewed 56 surviving passengers in person or by telephone

after the accident. Also, the Board sent questionnaires to all 129 surviving passengers,

and 110 questionnaires were returned.

Passengers reported that the flight was bumpy and that a pilot had announced over

the public address system the possibility of some rough weather. Passengers also stated

that they saw lightning outside the airplane during the final descent and that a pilot had

indicated over the public address system that there was a “light show” outside the

airplane. Passengers reported that the rain became harder, and some passengers described

turbulence and hail, as the final descent continued.

Passengers reported that the touchdown was very hard and that the airplane did

not slow down. One passenger indicated that the wheels were “shuddering.” Another

passenger stated that he heard a sound similar to car brakes and that the airplane

“fishtailed.” Other passengers indicated that they heard the thrust reversers come on and

off. Passengers also reported that the flight attendants yelled “brace” to prepare for the

impending impact.

Passengers indicated that most people evacuated the airplane in an orderly

manner (although there were a few reports of passengers pushing or jumping over others)

and that some passengers helped others get out of the airplane and away from the

wreckage. Passengers also indicated that the fire did not enter the aft fuselage interior

until after the passengers had evacuated from that part of the cabin but that smoke entered

the aft cabin immediately after the fire began. Passengers further indicated that, once

outside the airplane, they encountered heavy rain, strong winds, and hail. Some

passengers reported huddling in groups, and others reported sheltering themselves from

the weather with bales of hay in a field, until emergency workers arrived.

Two passengers reported information regarding the passenger in seat 27E (a

21-year-old male), who was killed in the accident. The passenger in seat 21D indicated in

Factual Information 55 Aircraft Accident Report

a postaccident interview that, when the back of the airplane had apparently cleared of

passengers, he stuck his head inside the cabin, through the right aft overwing exit, and

yelled, “is anybody else in there?” three times, to which someone answered, “that’s

everyone—that’s all.” The passenger in seat 21D identified the voice inside the airplane

as that of the passenger in seat 27E. (The passenger in seat 21D was the director of a

church choir group, and the passenger in seat 27E was a member of the group.) In

addition, the passenger in seat 27D stated in a postaccident interview that he saw the

passenger in seat 27E “scoot” toward the aisle and stand up after the crash. The body of

the passenger from seat 27E was found in the area by rows 32 and 33.

1.16 Tests and Research

1.16.1 Spoiler System Ground Tests

To better understand MD-80 spoiler system operation, the Safety Board, while on

scene in Little Rock, requested that American conduct ground tests on an MD-82 airplane

similar to the accident airplane. These tests, which were conducted at American Airlines’

Maintenance and Engineering Center in Tulsa, Oklahoma, included verification of the

maximum spoiler deflection with a full right roll input and the response of the system to

the spoiler autoretract mechanism. These tests were intended to assist the Board in

gathering on-scene evidence; as a result, the tests were not witnessed by the Board, and

the test data were not recorded by American. After a subsequent incident at Palm Springs,

California, involving an American Airlines MD-80 series airplane that did not experience

autospoiler extension at touchdown, the tests were repeated and additional tests were

conducted so that spoiler panel positions could be recorded and CVR and FDR data could

be obtained. These tests and the Palm Springs incident are discussed further in

section 1.16.1.1.

Other tests conducted on the accident airplane’s spoiler system were as follows:

• The autospoiler switching unit and ground spoiler control box were tested at

American Airlines’ Maintenance and Engineering Center on July 1, 1999.

Both passed functional tests. The tests were done in accordance with

American Airlines Engineering Specification Order 80503 functional test

procedures chapter 10 and American’s Engineering Specification Order

80340 functional test procedure chapter 8.

• The two ground control nose oleo switches were tested at the Safety Board’s

laboratory in Washington, D.C., on July 10, 1999. Both switches operated

smoothly, and electrical continuity was verified in the air (open) and ground

(closed) switch positions.

• The center pedestal was examined at the Safety Board’s laboratory on July 15,

1999. No significant marks were found on the spoiler control handle or handle

slot that would indicate the handle’s position at impact. The autospoiler crank

arm was found in the extended position and above the roller on the handle.

Factual Information 56 Aircraft Accident Report

The left throttle autoretract crank was measured to be 1 3/4 inches from the

idle stop. No evidence of preexisting failures was found on any center

pedestal switch or control.

• The autospoiler actuator was examined at Telair International (the

manufacturer of the actuator), Oxnard, California, on August 6, 1999, and at

the Safety Board’s laboratory. The actuator had sustained impact damage.

Members of the Systems Group who participated in the teardown agreed that

the actuator was most likely capable of functioning within its operational

parameters.

During the disassembly, the wire to the direct current brake was

found broken, and the brake plate was found to have excessive wear.

1.16.1.1 Testing Conducted After the Palm Springs Incident

On February 16, 2000, about 0708 Pacific standard time, American Airlines

flight 9503, an MD-83, N597AA, departed the left edge of runway 13R while landing at

Palm Springs International Airport, Palm Springs, California. The airplane was on a

positioning flight from Los Angeles International Airport to Palm Springs. The captain

and first officer were not injured, and the airplane received minor damage.

Tests on the

Palm Springs incident airplane were initially conducted at American Airlines’

Maintenance and Engineering Facility on August 24, 2000, but were repeated on

October 12, 2000, because of a problem with data recording. The tests and their results

are as follows:

• Spoiler knockdown: The spoilers were extended, and the left throttle was

advanced until the spoiler handle automatically retracted. The spoiler handle

was observed to slowly depress until the point at which it was knocked down

and automatically retracted. The throttle was measured to be 1 3/4 inches

above idle.

• Touchdown retract: The crank arm on the left throttle knocked down the

spoiler handle when the throttle was advanced about 1 3/8 inch above idle

(about 1.16 EPR) before the handle was extended. Six of these “handle

knockdown” operations were conducted; for each operation, the FDR

recorded an input for both spoiler positions. The recorded left outboard flight

spoiler positions ranged from 1.0° and 9.8°, and the recorded right inboard

flight spoiler positions ranged from 1.3° and 10.5°. (The recorded positions

depended on when the sample was taken in relation to spoiler extension.) The

right spoiler panels were observed during some of the operations. The

Systems Group estimated that the right flight spoilers extended to about 8° to

The actuator passed all functional tests except one. The actuator failed to operate when it

was positioned at mid-stroke, a weight was attached to the arm, and a reduced voltage was applied.

This failure was not significant to the overall outcome of the testing.

Because of the successful brake test and the characteristics of the damage to the wire, the

Systems Group determined that the damage to the wire occurred most likely during the accident

sequence and that the wire was completely separated during the disassembly.

The description for this incident, DCA00IA027, can be found on the Safety Board’s Web site

at <http://www.ntsb.gov>.

Factual Information 57 Aircraft Accident Report

10° before immediately retracting. The time interval for the spoiler panels to

extend and then immediately retract was about 1/2 second. The ground

spoilers did not move.

• Autospoiler extension with roll inputs: Normal autospoiler extensions were

conducted with full left and right roll inputs.

• Engine response: The throttles were quickly advanced from idle to between

1 3/8 and 1 3/4 inches and then immediately retracted. When the engines were

at ground idle, no movement of any engine parameters could be seen in the

cockpit. When the engines were at flight idle, a momentary movement of the

EPR and N1 indicators was seen in the cockpit.

The CVR was running continuously throughout the tests, and the cockpit door

was closed during portions of the tests to better replicate in-flight sounds. Recordings

were made of the autospoiler actuator operating with the spoiler handle in the armed

position, the left throttle at 1 3/4 inches above idle and at idle, and the spoiler handle in

the unarmed position. The CVR also recorded the sounds associated with arming and

unarming of the spoiler handle.

1.16.2 Main Landing Gear Tire Examination

All four main landing gear tires from the Little Rock accident airplane were

inspected at the Michelin Aircraft Tire Corporation, Greenville, South Carolina, on

October 10, 2000. The entire tread surface (internal and external) of each tire was

inspected, and all four tires exhibited typical wear characteristics for bias-ply tires and

were estimated to be about 50 percent worn. None of the tires showed any evidence of

external reverted rubber or internal ply separation. Superficial scrub marks were found

laterally on the tread surface. The tires’ inner liners showed no evidence of underinflation

or excessive load.

To determine whether the tires had been heated because of reverted rubber

hydroplaning,

Shore hardness tests were performed using a device, known as a

durometer, to measure the hardness of the rubber in the tires. (Reverted rubber

hydroplaning tends to soften the rubber on the tires.) Durometer measurements were

taken every 90° on the tread surface, and all readings were within the tires’ expected

operating range. Michelin engineers believed that storage length and conditions would

An article, titled “Landing on Slippery Runways,” in Boeing’s October to December 1992

Airliner

magazine, explained reverted rubber hydroplaning as follows: “when a tire locks up on a

smooth, wet, or icy surface, the friction heat generates steam. The steam pressure then lifts the

tire off the runway, and the steam heat reverts the rubber to a black gummy deposit.”

Factual Information 58 Aircraft Accident Report

have had only minor effects on the hardness readings and would not have affected

reverted rubber if it were present.

At the request of investigators, Michelin conducted a more extensive

measurement of tire tread hardness by taking measurements every 10° on all four tire ribs.

These measurements confirmed that the hardness of each tire was within operational

standards.

Further tests (holographic imaging and sectional, rubber, and microscopic

analyses) were considered but not performed because they would not provide any

additional data on whether reverted rubber hydroplaning occurred during the accident

sequence.

Michelin engineers presented information on reverted rubber hydroplaning and

the conditions that cause internal reverted rubber. The engineers stated that reverted

rubber hydroplaning, which occurs on the molecular level on the surface of the tire tread,

happens very quickly and does not cause internal heating of the rubber because of its very

low thermal conductivity. The engineers indicated that evidence of reverted rubber

hydroplaning could be worn away fairly quickly. The engineers further indicated that, for

internal heating of the tires to occur, the tires must have a significant load and be rolling

for a substantial distance and that damage from internal heating could accumulate over

time.

1.16.3 Cockpit Voice Recorder Sound Spectrum Study

To determine whether the autospoiler actuator operated at touchdown, the Safety

Board conducted a CVR sound spectrum study at its headquarters in Washington, D.C.

The sound spectrums from the CVRs aboard the Little Rock accident airplane, Palm

Springs incident airplane, flight test airplanes, and ground test airplane were examined on

a spectrum analyzer, which gives a visual presentation of the frequency contents of

signals, and a computer signal analyzer, which presents the specific frequency content of

the signals and detailed timing and waveform information. Charts were prepared for all of

the airplanes to document the sounds heard on the cockpit area microphone.

The sound spectrum study indicated that most of the noise made by the

autospoiler actuator was at a frequency centered around 1200 hertz (Hz). For the Little

Rock accident airplane, the sound associated with the autospoiler actuator lasted about

0.08 second. A “clunk” sound was heard about 0.32 second after the initial actuator

sound. For the Palm Springs incident airplane, the sound associated with the autospoiler

actuator lasted 0.06 second, and the “clunk” sound was heard 0.30 second after the initial

actuator sound.

The sound spectrums were compared with those from the flight and ground tests.

As stated in section 1.11.1, the flight tests were conducted in August 1999 aboard two

The tires had been in storage for more than 1 year in a hangar that was not environmentally

controlled.

Two of the tires had localized soft spots, which were associated with tread contamination

or rib damage as a result of the impact sequence.

Factual Information 59 Aircraft Accident Report

American Airlines flights. The flight crews armed the autospoiler handle during approach

and conducted a normal touchdown, upon which the wheel spin-up sensors triggered the

motor to automatically deploy the spoiler handle. For the first flight test airplane, the

sound associated with the autospoiler actuator lasted 0.18 second, and the “clunk” sound

was heard 0.30 second after the initial actuator sound. For the second flight test airplane,

the sound associated with the autospoiler actuator lasted 0.19 second, and the “clunk”

sound was heard 0.32 second after the initial actuator sound.

As stated in section 1.16.1.1, the ground tests were conducted in August and

October 2000 aboard the Palm Springs incident airplane. The CVR was run continuously

during the tests, and several autospoiler actuator sequences were analyzed. For the test

that triggered the autospoiler system with the spoiler handle in the unarmed position, the

sound associated with the autospoiler actuator lasted about 0.06 second, and the “clunk”

sound was heard 0.18 second after the initial actuator sound. For the test that involved the

autospoiler system with an armed spoiler handle, two actuator sequences were analyzed.

For both sequences, the sounds associated with the autospoiler actuator lasted

0.16 second, and the “clunk” sounds were heard 0.30 second after the initial actuator

sounds.

1.16.4 Airplane Performance Study

An Airplane Performance Study was conducted to determine the motion of

American Airlines flight 1420 and the physical forces that produced that motion. The

study considered data from the following sources: wreckage location, runway scars and

markings, radar data, FDR and CVR data, weather information, and ground deceleration

computer program results. The radar, CVR, and FDR data times were synchronized to a

single reference time.

The airplane performance parameters that were of primary interest in this accident

were those that defined the motion of the airplane on the runway, including the airplane’s

position; ground speed; heading, track, and drift angles; and deceleration. The approach

parameters of primary interest included the airplane’s position relative to the ILS

localizer and glideslope beams; airspeed; heading, bank, pitch, drift, and flightpath

angles; and the rate of climb or descent. The control inputs, power settings, and winds

that affected the airplane were also important considerations during the approach and

landing segments.

The FDR and radar data were used together, along with weather information,

runway tire markings, and wreckage location, to calculate the airplane’s flightpath. The

position of the airplane was determined in two different segments: the “in air” trajectory,

which incorporated data from 440 to 20 feet afl, and the “on ground” trajectory, which

incorporated data from 20 feet afl to about 600 feet beyond the end of runway 4R (at

which point the FDR ceased to record data). The calculated ground trajectory is discussed

further in section 1.16.4.1.

Factual Information 60 Aircraft Accident Report

The altimeter setting reported by air traffic controllers during flight 1420’s

approach to the airport was 29.86 inches of Hg. The Airplane Performance Study

indicated that, because the atmospheric pressure was rising rapidly during the descent

and landing, an altimeter setting of 29.92 inches of Hg would have been more accurate

for the final approach segment of the flight.

An altimeter setting of 29.86 inches of Hg

would have resulted in an indicated altitude that was about 55 feet lower than the altitude

with an altimeter setting of 29.92 inches of Hg. (After the accident, the altimeter setting

for the airport was recorded as 29.98 inches of Hg.)

1.16.4.1 Calculated Ground Trajectory

On final approach, the accident airplane’s airspeed averaged 156 knots, which

was about 25 knots faster than the V

ref

, and jumped erratically within a band of ±5 knots,

which was consistent with the gusty and turbulent winds on approach.

The wind was

blowing mostly along the airplane’s lateral axis, from the left to the right sides of the

airplane. The airplane touched down 2,000 feet from the runway threshold at a ground

speed of 160 knots, drifting about 5º to the right,

and a tailwind component of about

5 knots was present. A ground speed of 160 knots was about 20 knots faster than the

zero-wind touchdown ground speed that would result from an approach at the reference

airspeed plus 10 knots.

Flight 1420 continued to drift while on the runway by as much as 16º both to the

right and the left of the direction of travel. Just before the FDR data ended, the airplane’s

heading was 20º to the right of the direction of travel. The airplane was returning to the

extended centerline of the runway when it impacted the runway 22L approach lighting

system support structure.

From 3,000 to 5,800 feet beyond the runway 4R approach end threshold, the

airplane’s rudder was consistently in the trailing edge right direction (nose right), but the

airplane’s heading was continuously decreasing (nose left) at 1º to 3º per second. The

heading stopped decreasing between 4,000 and 4,600 feet, coinciding with FDR data

indicating a brief stowing of the thrust reversers. About 5,200 feet beyond the runway

threshold, as full right rudder was being applied, the heading decreased about 1.5º per

second until 5,800 feet, when FDR data indicated that the thrust reversers were stowed

again. At this point, the yaw rate reversed, and the heading started to increase up to 7º per

second. About 6,600 feet beyond the runway threshold, the left reverser was deployed,

but the right reverser remained stowed; with the engine EPRs at an idle power level, the

airplane continued to yaw nose right about 4º per second. The ground trajectory and FDR

data showed that, between 2,800 and 5,000 feet beyond the threshold, both the right and

left elevator surfaces were deflected full nose down (15º).

The FDR records pressure altitude, that is, altitude data based on an altimeter setting of 29.92

inches of Hg (the standard pressure at sea level).

Gusty winds involve a fluctuation in wind speed of 10 knots between lulls and peaks. Turbulent

winds imply vertical motion and rolling.

The drift angle is the difference between the airplane’s heading and the direction of the velocity

vector of the center of gravity.

Factual Information 61 Aircraft Accident Report

The left engine EPR was greater than 1.3 almost continuously between 3,200 and

5,800 feet beyond the runway 4R approach end threshold while the thrust reverser was

deployed. The right engine EPR also reached levels above 1.3 several times while the

thrust reverser was deployed. The left brake pedal was relaxed briefly about 5,500 feet

beyond the runway threshold, coinciding with a stowing of the thrust reversers and a loss

of deceleration. The calculated ground trajectory indicated that the flight 1420 airplane

departed runway 4R at about 97 knots and impacted the runway 22L approach lighting

system support structure at about 83 knots.

1.16.4.2 Ground Deceleration Study

The effects of various deceleration devices on flight 1420’s stopping distance was

evaluated using The Boeing Company’s MD-80 Operational Landing Program. The

conditions tested in the program were as follows:

• airplane weight, 127,000 pounds;

• center of gravity position, 16.7 percent of mean aerodynamic chord;

• temperature, 25º C (77º F);

• rolling friction coefficient, 0.02;

• wet runway braking friction coefficient, ranging from 0.21 to 0.28 at

flight 1420’s speeds from touchdown to the end of the runway;

• touchdown speed, 152 and 162 knots;

• reverse thrust, none and constant symmetric reverse at 1.3 EPR;

• braking, none, flight 1420’s braking profile (initial braking at 5 seconds after

touchdown and full braking 6 seconds later), and a braking profile in which

initial braking occurred 1/4 second after touchdown and full braking

1 1/4 seconds later; and

• spoilers, deployed and not deployed.

Computer runs were made to test airplane deceleration with different

combinations of spoiler deployment, reverse thrust, and braking scenarios.

Selected computer runs showing ground speed versus distance from the runway

4R threshold were compared with flight 1420’s calculated ground speed profile. All of

the selected computer runs were based on an initial ground speed of 162 knots, which

was close to the 160-knot touchdown ground speed determined from the ground

trajectory calculation, and an initial position on the runway of 2,000 feet from the

threshold, which was about the point where flight 1420’s tire markings began. Constant,

symmetrical reverse thrust at 1.3 EPR was maintained until the ground speed was zero,

and all of the braking was based on flight 1420’s braking profile.

With no braking, spoilers, or reverse thrust, the airplane would depart the runway

with a ground speed of about 142 knots, which is 45 knots faster than the flight 1420

airplane. With no spoilers or brakes but with a constant, symmetrical reverse thrust at

Factual Information 62 Aircraft Accident Report

1.3 EPR, the airplane would leave the runway at 117 knots, which is 20 knots faster than

the flight 1420 airplane. With no spoilers but with 1.3 EPR reverse thrust and the flight

1420 braking profile, the airplane would have departed the runway at 95 knots. This

computer run matched the flight 1420 data best and indicated that the accident airplane

experienced a braking coefficient of at least 0.21 to 0.23 at speeds between 160 and

140 knots, respectively.

Figure 16 shows the results of the computer runs compared with

the flight 1420 profile.

Figure 16. Effects of Spoilers, Brakes, and Reverse Thrust on Stopping Distance

Two additional computer runs demonstrated that proper spoiler deployment is

critical to deceleration and stopping distance. Without reverse thrust and with the spoilers

deployed and the flight 1420 braking profile, the airplane would depart the runway at

20 knots, which is 75 knots less than the airplane with no spoilers but with reverse thrust

and braking. With constant reverse thrust, deployed spoilers, and the flight 1420 braking

profile, the airplane could have stopped about 700 feet before the end of the runway.

Computer runs were also performed to determine the effect of the touchdown

ground speed on stopping distance. One computer run indicated that an airplane with no

Typical braking coefficients for a hydroplaning airplane range from 0.02 to 0.04.

0 1000 2000 3000 4000 5000 6000 7000 8000

100

120

140

160

180

Notes:

* For reverse thrust runs, reverse thrust at 1.3 EPR

is maintained until ground speed is zero.

* Braked runs are based on AA flight 1420 braking profile.

Line Spoilers Brakes Reverse Thrust

A *** AA flight 1420 data ***

BNoNoNo

CNo Yes Yes

DNoNoYes

EYes Yes No

F Yes Yes Yes

Runway

end

Touchdown

point

Ground speed, kts.

Distance from runway 4R threshold, ft.

Factual Information 63 Aircraft Accident Report

spoiler deployment but with reverse thrust and a 10-knot reduction in the touchdown

ground speed would have departed the runway at 60 knots and would have impacted the

runway 22L approach lighting system support structure at a speed of 20 knots, which is a

94-percent reduction in the kinetic energy compared with the calculated 83-knot impact

speed in the flight 1420 profile. Another run that allowed for a loss in braking

performance after the airplane departed the hard surface of the runway indicated that the

impact speed would have been 45 knots, which is a 72-percent reduction in the kinetic

energy compared with the calculated 83-knot impact speed. In addition, without reverse

thrust but with spoilers deployed and a 10-knot reduction in the ground speed, an airplane

would have been able to stop about 400 feet before the end of the runway.

Finally, computer runs were conducted to determine the effect on stopping

distance of a braking profile in which initial braking occurred 1/4 second after touchdown

and full braking 1 1/4 seconds later. (Because of a limitation in Boeing’s Operational

Landing Program, this comparison could only be done using a touchdown ground speed

of 152 knots.) The computer runs indicated that, without spoiler deployment but with

reverse thrust, the normal braking profile would stop an airplane 200 feet sooner that the

flight 1420 braking profile. With spoiler deployment and no reverse thrust, the normal

braking profile would stop an airplane 800 feet sooner that the flight 1420 braking

profile.

1.16.5 Engineered Materials Arresting System Computer Model

The Safety Board requested that Engineered Arresting Systems Corporation of

Lester, Pennsylvania, survey the accident area and provide information on any safety

benefit that a soft-ground aircraft arresting system would have provided for flight 1420 if

such a system had been installed at the departure end of runway 4R.

At the Board’s

public hearing on this accident, a consultant for Engineered Arresting Systems discussed

the company’s Engineered Materials Arresting System (EMAS). The consultant

described the EMAS as a passive system that decelerates an airplane when its wheels roll

through a soft foamy material.

He added that the system was designed to be compatible

with all of the airplanes with which it would come in contact.

The computer program that was used for the analysis assumed the following

conditions based on data from the accident sequence:

• runway departure speed, 98 knots;

• gross weight, 128,000 pounds;

On April 14, 1999, the FAA, the Little Rock Municipal Airport Commission, and the airport’s

tenants met at a joint planning conference to identify airport development needs for the next 5 years.

According to the conference report, an arresting system for runway 4R/22L was identified as a

recommended project for 2000. In the fall of 2000, an EMAS measuring 304 feet long and 200

feet wide was installed at the departure end of runway 4R.

The consultant stated that the material has an average strength in compression of 80 psi and

that the material would retain its properties over a wide temperature range.

Factual Information 64 Aircraft Accident Report

• center of gravity, 16.7 percent mean aerodynamic chord;

• yaw angle, 20º;

• yaw rate, 0.069 radians per second;

• lateral velocity, 8 knots;

• roll angle, 2º right wing down; and

• distance offset from runway, 7 feet.

The analysis also included an EMAS that was designed to reflect the conditions

that existed at the departure end of runway 4R. The total length of this EMAS was

402 feet.

In a December 7, 1999, report to the Safety Board, Engineered Arresting Systems

concluded that the benefit of an EMAS in this accident would have been limited by the

airplane traveling partly outside the runway edges; thus, the airplane would not have been

able to use the full length of the EMAS. The report also concluded that an EMAS would

have reduced the speed of the airplane by 15 knots but would not have enabled the

airplane to stop within runway 4R’s runway safety area. In addition, the report concluded

that, in this accident, an EMAS could have led to two landing gear failures because the

loads exceeded the ultimate values determined by the airplane manufacturer.

1.17 Organizational and Management Information

American Airways was incorporated in 1930, and its name changed to American

Airlines, Inc., in 1934. American is owned by the AMR Corporation and is headquartered

in Dallas, Texas. American provides passenger and cargo service throughout North

America, the Caribbean, Latin America, Europe, and the Pacific. AMR Corporation also

owns and operates American Eagle, a regional airline that provides service at American’s

hubs and other cities throughout the United States, Canada, the Bahamas, and the

Caribbean.

100

In February 1999, American acquired Reno Air, which was fully integrated

The Safety Board notes that, although the Engineered Arresting Systems report indicated that

an EMAS would not have stopped the flight 1420 accident airplane, the system’s safety benefit

was demonstrated in the May 8, 1999, accident involving American Eagle flight 4925, which overran

the approach end of runway 4R at John F. Kennedy International Airport, Jamaica, New York, during

landing. The airplane crossed the runway threshold at a speed of about 180 knots and touched down

7,000 feet beyond the end of the runway at a speed of about 157 knots. The flight crew applied

reverse thrust and maximum braking, but the airplane departed the 8,400-foot runway at a speed

of about 75 knots. Approximately 300 feet of skid marks were observed before the end of the

runway. The airplane then traveled over a 6-inch deflector and approximately 248 feet across a

400-foot long EMAS. The landing gear sank approximately 30 inches into the EMAS, and the airplane

came to a stop. Of the 30 people aboard the airplane, 1 person was seriously injured, and the rest

were not injured. The airplane received substantial damage. See section 1.18.6 for additional details

on this accident.

100

At the public hearing, the AMR Corporation’s Vice Chairman stated that, even though American

Eagle operates under its own certificate, the airline is considered to be a “sister company” to American

Airlines, and both airlines are increasingly sharing human and other resources.

Factual Information 65 Aircraft Accident Report

into American’s operations at the end of August 1999. In April 2001, American acquired

Trans World Airlines.

According to the AMR Vice Chairman, American Airlines experienced a period

of substantial growth during the 1980s, both in the number of aircraft and the number of

employees. American’s Web site indicated that, as of March 2000, the airline had

649 transport-category airplanes in its fleet with an average age of 8 years. The fleet

consisted of Airbus A300; Boeing 727, 737, 757, 767, and 777; Fokker F.100; and

McDonnell Douglas DC-10, MD-11, and MD-80 airplanes.

101

At the time of the accident, American Airlines employed 9,661 pilots, 2,812 of

whom were qualified on the MD-80 (1,440 captains and 1,372 first officers). The

company had 279 MD-80 series airplanes in its fleet and 10 MD-80 bases throughout the

United States. The company had a total of 498 check airmen, 108 of whom were MD-80

check airmen.

American Airlines underwent an executive reorganization on January 5, 2000.

According to the AMR Corporation’s Vice Chairman, changes to the company’s

organization were proposed before the Little Rock accident but had not been

implemented because of a lack of consensus that “any particular change was appropriate

or meaningful.” The AMR Vice Chairman indicated that the new organizational setup

was a “highly integrated but extensive attempt to provide more safety emphasis at the

company.”

At the time of the accident, the Managing Director of Flight Safety reported to the

Vice President of Flight/Chief Pilot, who was responsible for hiring pilots, training them,

and managing flight operations. The Vice President of Flight/Chief Pilot reported to the

Executive Vice President of Operations,

102

who reported to the Chairman of American

Airlines and AMR Corporation.

Under the reorganization, the Executive Vice President of Operations was

elevated to the Office of the Chairman as the AMR Corporation Vice Chairman, and he

retained the primary responsibility for operations. Eight organizational units report to the

Office of the Chairman, including a new unit, Safety, Security, and Environmental.

103

These three functions existed separately under the former organization but are now

101

The AMR Vice Chairman indicated that American was planning to retire its older fleets—

the 727 and DC-10.

102

The Executive Vice President of Operations (now the AMR Corporation Vice Chairman) was

responsible for eight other organizational units: Maintenance and Engineering, Operations Planning

and Performance, Cargo, Reno Air Integration, Corporate Real Estate, Purchasing, Security, and Safety.

In public hearing testimony, the AMR Vice Chairman indicated that the areas that received most

of his attention as Executive Vice President of Operations were maintenance and flight.

103

The seven other organizational units that report to the Office of the Chairman are Human

Resources, General Counsel, Operations, Customer Services, Marketing and Planning, Finance, and

Government Affairs.

Factual Information 66 Aircraft Accident Report

placed under a single vice president-level leadership.

104

The Managing Director of Safety

reports to the Vice President of Safety, Security, and Environmental and coordinates with

the Vice President of Flight/Chief Pilot, who is responsible for the Flight Department.

105

Four main organizational units exist within the Flight Department, one of which is

Flight Training and Standards. Ground school and simulator training are separate

organizational units within Flight Training and Standards rather than a combined unit

within the training organization. The AMR Vice Chairman indicated that the training and

standards (that is, checking) functions were separated to provide improved objectivity

and standardization.

1.17.1 Aviation Safety Action Program

The American Airlines Safety Action Program (ASAP) was implemented in 1994

after 2 years of development and coordination with the FAA and the Allied Pilots

Association.

106

ASAP is a voluntary, confidential pilot reporting program designed to

collect and disseminate information on safety issues and incidents to prevent their

recurrence. The program receives about 3,600 reports each year.

107

No disciplinary actions

are taken against pilots as a result of an ASAP report submission. The Managing Director

of Flight Safety administered the program at the time of the accident. At the Safety

Board’s public hearing, the Vice Chairman of AMR Corporation testified the following:

[ASAP] is a true accident prevention program. It operates on the basis of the

following principle, and that is that the best safety information is the information

that we don’t know, and, so, any way that we can create to bring information that

otherwise would not be known to those who can effect change, to bring that

information to the surface is very, very valuable, and that’s what ASAP’s

principle concept is all about.

ASAP requires that pilots submit a report of an event within 24 hours of its

occurrence (or within 24 hours of the time at which the pilot became aware that an event

occurred). ASAP reports are sent to the ASAP Event Review Team, which consists of a

representative from American, the FAA’s Certificate Management Office for American,

and the Allied Pilots Association.

108

The team meets weekly to review submitted reports,

104

Two new functions were also added to the Safety, Security, and Environmental organizational

unit: Operational Audits and Compliance. The AMR Vice Chairman testified that, “in these areas,

we intend to substantially pick up...our auditing processes by third parties who are not responsible

for...particular operational functions throughout the company.”

105

The Vice President of Flight/Chief Pilot reports to the Managing Director of Operations, who

reports to the Managing Director of Safety.

106

The program is consistent with the guidance in AC 120-66, “Aviation Safety Action Programs.”

107

ASAP reports do not include deliberate or criminal acts or events that are already known

within the FAA.

108

Confidential reports are also sent to the manager of the ASAP program, the appropriate manager

within American’s Flight Safety Department, and NASA for inclusion in its Aviation Safety Reporting

System. In addition, the FAA has immediate access to ASAP reports.

Factual Information 67 Aircraft Accident Report

decide which ones represent a significant safety concern or deviation from procedure, and

determine who should investigate the events and recommend corrective actions.

According to American’s pilot magazine, Flight Safety,

109

ASAP reports identify,

in order of frequency, altitude deviations; heading deviations; other flight irregularities,

communication problems with ATC, or clearance deviations during flight; deviations

from regulations or operational procedures, including the MEL; general irregularities or

communication problems with ATC on the ground; runway or taxiway incursions; and

other categories, including aircraft damage, turbulence encounters, and mechanical

problems. The Managing Director of Safety stated that ASAP does not currently collect

information about the relationship between an event and the flight crew’s flight and duty

time.

110

Selected information from deidentified ASAP reports is distributed to company

pilots via bulletins every 6 weeks. Information from ASAP reports can also be

communicated quarterly through the company’s pilot magazine. In addition, ASAP

reports are entered into an American Airlines database, which provides the company with

trend information on particular subjects and areas of focus for training and surveillance.

1.17.2 Flight Crew Training

American’s flight crew training academy is located in Dallas/Fort Worth. All

first-time pilots at American Airlines attend a basic indoctrination course, where they are

taught general information on the way the company operates. According to the MD-80

Fleet Manager at the time of the accident,

111

basic indoctrination training includes an

alertness strategies course, which focuses on fatigue countermeasures more than fatigue

recognition.

Pilots then attend initial and/or transition ground school and simulator flight

training. According to the American Airlines DC-9 Initial and Transition Training

Syllabus (dated December 15, 1998), the typical initial and transition training consists of

10 days of ground school, 10 days of simulator flight training, and 25 hours of initial

operating experience (IOE).

Ground school training is presented using self-paced computer-based training

with graphics and audio, videotapes, and a performance workbook with printed practice

problems. Days 1 through 5 of simulator training are conducted by an American Airlines

simulator instructor, and days 6 through 10 are conducted by an American Airlines check

airman. For their IOE, new and upgrade pilots fly with a check airman, who references a

worksheet that details the airplane maneuvers, procedures, or functions that are required

109

Chidester, T. “ASAP Turns Five and a Half.”

Flight Safety

. American Airlines, Vol. 1, No. 2,

First Quarter (2000): pp. 11-16.

110

Although the reports contain time-of-day information, the Managing Director of Flight Safety

stated that the information is used only to correlate reports with ATC records.

111

American Airlines’ MD-80 Fleet Manager at the time of the accident became the company’s

Managing Director of Flight Crew Relations in December 1999.

Factual Information 68 Aircraft Accident Report

to be covered. In addition, all pilots are required to attend recurrent training (a 2-hour

line-oriented flight training [LOFT] session)

and perform a 2-hour proficiency check ride

every year.

112

The recurrent training and the proficiency check ride are conducted by

company check airmen.

The former MD-80 Fleet Manager indicated that, in all phases of training and

IOE, American attempts to put the pilot in a situation in which he or she is required to

make the best decision for the particular circumstances. The former fleet manager also

indicated that first officers are trained to provide feedback to captains; specifically, first

officer trainees are presented with various situations and are critiqued on their actions and

the consequences of those actions.

1.17.2.1 Simulator Flight Training

Day 1 of the 10-day initial and transition simulator flight training course includes

a briefing on American’s checklist philosophy for flying and nonflying pilots. The

simulator profile during that day emphasizes how each checklist is accomplished and

which items require a response.

Day 6 of the course is dedicated to takeoff and landing exercises. According to the

former MD-80 Fleet Manager, the day begins with a 2-hour briefing that covers the key

points regarding the airplane’s handling characteristics and the windshear escape and

recovery maneuvers. The former Fleet Manager indicated that the simulator session starts

out with little or no crosswind and a dry runway and that the crosswind component is

gradually increased until the MD-80’s demonstrated maximum crosswind component is

attained. Afterward, the runway surface friction component is reduced. This process is

repeated until the pilots have experienced the control difficulties that they need to learn to

correct.

113

The former MD-80 Fleet Manager stated that the simulator can replicate rudder

blanking (that is, the decrease in rudder effectiveness resulting from an increase in reverse

power) and that this feature allows the company to train pilots to reduce the amount of

reverse thrust until they have regained directional control.

The training syllabus indicates that windshear profiles, slippery runways,

crosswind landings,

114

and nighttime landings are discussed during day 6. The syllabus did

112

American’s recurrent training syllabus changes on February 1 of each year. According to the

former MD-80 Fleet Manager, the recurrent training course beginning on February 1, 2000, was expected

to include a briefing on manual spoiler operation and the spoilers’ effect on braking and stopping.

Each pilot would then have a chance to manually operate the spoilers in the simulator. The recurrent

training was also expected to include a review of the stabilized approach concept, wet and slippery

runway reversing, and automatic braking.

113

American’s Instructor/Check Airman Guide, MD-80 Supplement, did not contain any guidance

for instructing landings with crosswinds or on wet runways. According to American, its instructor

guides refer the instructor to the appropriate flight and operating manuals for this information.

114

The former MD-80 Fleet Manager stated that crosswind training is also taught in ground

school during a day in which the contents of the performance manual are discussed. The manual

contains a chart depicting the headwind, tailwind, and crosswind components in relation to wind

speed and the angle between the wind and the runway direction.

Factual Information 69 Aircraft Accident Report

not contain any description of the specific scenarios taught during those sessions but did

contain a reference to the company’s DC-9 Operating Manual.

Day 8 of the training includes the captain’s rating ride preparation and the first

officer’s check ride. The captain’s rating ride occurs on Day 9 of the training and is

administered entirely in the DC-9 flight simulator. The rating ride consists of about

15 basic profiles, and the training manual indicates that the rating ride is conducted “in

real time and in as realistic an ATC environment as possible.”

LOFT is conducted on Day 10 of flight training. The LOFT session consists of

two legs of a real-time flight in the simulator. According to the syllabus, the first leg is

usually routine and flown by the first officer, and the second leg is an abnormal situation

that is flown by the captain. The manual stated that the check airman provides realistic

flight plans, a weather briefing, and takeoff performance system data for each leg of the

flight and that communications with ATC and the company are provided.

1.17.2.2 Observations of Simulator Sessions

In July 1999, two members of the Operations Group observed separate sessions of

day 6 (takeoffs and landings) of American’s MD-80 simulator training. The sessions

were conducted by two different instructors.

One of the simulator sessions (referred to in this report as session A) did not

include any “failed spoiler” events during the landing portion of the training. The other

simulator session (referred to in this report as session B) included seven failed spoiler

events. The students in session B noticed two of the seven events; in both instances, the

first officer student manually extended the spoilers. However, American’s procedures

stated that the captain was to manually extend the spoilers if they did not automatically

extend during landing (see section 1.17.4.2). The effects of the spoilers not extending

during landing were discussed.

115

In session A, the instructor recommended that heavy manual braking be used on

contaminated runways and stated that he did not like the autobrakes for landing. In

session B, the instructor stated that medium autobrakes should be used when landing on

slippery runways. However, American’s procedures at the time of the accident stated that

aggressive manual brakes or maximum autobrakes should be used with wet runway

conditions (see section 1.17.4.3). No discussion occurred in either session regarding what

constitutes a slippery runway and what distinguishes it from a wet runway.

During session A, the training focused on using the 1.6 EPR reverse thrust setting

(the normal setting used by American for landing on dry runways). There was no

discussion or training on company procedures to limit reverse thrust to 1.3 EPR during

115

The two Operations Group members passed along their spoiler training observations to the

former MD-80 Fleet Manager who, in turn, met with the simulator instructors to inform them of

the observations and implement actions to improve the training. The former fleet manager indicated,

during the public hearing, that training on recognizing no-spoiler extensions and performing the

appropriate response had been added to day 6.

Factual Information 70 Aircraft Accident Report

landing on a slippery runway (see section 1.17.4.4). Rudder blanking was discussed.

During session B, the simulator instructor taught that 1.6 EPR was acceptable for landing

on a slippery runway unless a crosswind was present,

116

and the students applied 1.6 EPR

reverse thrust during 10 to 12 landings on slippery runways. The instructor subsequently

realized that 1.3 EPR should have been the maximum reverse thrust used and informed

the students of this information. Rudder blanking was discussed.

Session A did not include a discussion of company procedures for crosswind

limits for reduced visibility operations and contaminated runways but emphasized the

importance of avoiding thunderstorms. Session B included a discussion of crosswinds

and their effects on takeoffs and landings.

1.17.2.3 Human Factors and Safety Training

In 1990, American Airlines established a manager position dedicated to human

factors and safety training.

117

This manager is supported by 10 staff facilitators who are

company line captains and first officers on detail to the Human Factors and Safety

Training program.

118

At the time of the accident, five of the facilitators were qualified on

the MD-80.

The human factors and safety training program courses emphasize four

fundamental principles—situational awareness, communication, teamwork, and technical

proficiency—through lectures, slide and videotape presentations, and group discussions.

The videotapes present event scenarios that were recreated in a simulator. Many of the

recreated events were derived from ASAP reports. The facilitators guide the group

discussion to ensure that every element of the event is addressed, including factors that

may have prevented the event from occurring.

American provides separate human factors and safety training courses to its

pilots, flight attendants, and dispatchers. The courses for pilots include a 3-hour basic

indoctrination course, a 4-hour first officer upgrade course, a 1-day captain upgrade

course, a 2-hour recurrent course, and a 2-day check airman course. The first officer

upgrade course, attended by new first officers, emphasizes their roles and responsibilities

in the cockpit, including the need to speak up when an airplane is being handled

improperly or placed in jeopardy. The captain upgrade course, attended by new captains,

emphasizes strategies for being an efficient manager of events. The recurrent course,

attended jointly each year by captains, first officers, and flight engineers, emphasizes

areas that parallel those taught during recurrent simulator training.

119

116

A member of the Operations Group informed the instructor of the discrepancy between this

statement and the operating manual requirement to limit reverse thrust to 1.3 EPR when landing

on slippery runways.

117

Before 1996, this program was known as the crew resource management program.

118

While on detail to this program, captains serve half of the time as check airmen for an

airplane fleet, and first officers fly as line pilots two times per month and then every third month.

119

The 1999 recurrent training course emphasized decision-making and managing the environment

when faced with conflicting demands and information.

Factual Information 71 Aircraft Accident Report

1.17.3 Approach Procedures

1.17.3.1 Approach Briefing

American Airlines’ Flight Manual, Part I, Section 10, “Approach and Landing,”

page 12 (dated April 7, 1999), states that the captain will ensure that the first officer and

the flight engineer (if applicable) are briefed before every approach.

120

American’s DC-9

Operating Manual, Volume 1, Normals, page 87 (dated April 26, 1999), indicates that,

for all instrument approaches, the approach briefing is to include the following:

121

• landing runway and reported visibility or RVR;

• type of approach to be conducted;

• if a nonprecision or Category I ILS approach, who will be the flying pilot;

• approach chart to be used and the applicable minimum visibility or RVR;

• approach facility and frequency;

• final approach course;

• airport elevation;

• outer marker crossing altitude or minimum crossing altitude at final

approach fix;

• minimum descent altitude, decision altitude, and decision height, as

applicable;

• missed approach procedure;

• minimum safe altitude;

• initial approach altitude; and

• terrain awareness.

1.17.3.2 Before Landing Checklist

The Before Landing checklist is accomplished using a mechanical checklist in the

cockpit. According to American’s DC-9 Operating Manual, Normals, page 7 (dated

December 21, 1998), the mechanical checklist shows 10 items: hydraulic pumps,

altimeters, flight instruments and bugs (that is, moveable markers for key airspeed

values), seat belt/no smoking signs, tail deice, gear, spoiler lever, autobrakes, flaps and

slats, and annunciator lights.

122

American’s Instructor/Check Airman Guide, MD-80 Pilot

Supplement, Section 3.01, page 9 (dated July 15, 1996), states that the nonflying pilot

120

At the public hearing, the former MD-80 Fleet Manager stated that either pilot could conduct

an approach briefing but that the captain was responsible for ensuring that a briefing was completed.

121

The first officer stated in a postaccident interview that the captain conducted a formal briefing

for runway 22L.

122

The Before Landing checklist was designed so that the first five items are completed early

in the approach and that the last five items are completed late in the approach.

Factual Information 72 Aircraft Accident Report

was to accomplish the Before Landing checklist and was to “discuss those items

requiring responses and emphasize proper challenges and responses.” American’s DC-9

Operating Manual, Normals, page 71 (dated December 21, 1998), states the following

regarding the Before Landing checklist:

123

After each item has been accomplished, the pilot-not-flying will call out that item

on the checklist, call out the appropriate response and then move the

corresponding switch on the Mechanical Checklist. Any item that cannot be

verified by the pilot-not-flying as accomplished will require a challenge and

response. ALTIMETERS and FLT INSTR & BUGS will be challenged by the

pilot-not-flying and responded to by both pilots. When all items have been

accomplished, the pilot-not-flying will advise, “Before Landing checklist

complete.”

American’s DC-9 Before Landing checklist is found in the DC-9 Operating

Manual, Normals, pages 71 through 74 (dated April 26, 1999). Page 72 indicates that the

nonflying pilot is responsible for announcing that the spoiler lever has been armed and

that the spoilers should not be armed if the AUTO SPOILER DO NOT USE light is

illuminated, but no reference could be found to indicate which pilot was responsible for

physically arming the spoiler lever. According to postaccident interviews with American

Airlines pilots, instructors, and check airmen, pilots were instructed during simulator

training that the nonflying pilot was to arm the spoilers. However, company line pilots

said that, during actual flights, either pilot could arm the spoilers but that the captain

usually did because the spoiler handle was on that side of the center pedestal.

The Before Landing checklist also indicates that the flying pilot is responsible for

commanding the landing gear down and that the nonflying pilot is to respond “down

three green.” In addition, the Before Landing checklist indicates that the nonflying pilot

is to arm the autobrakes as required (see section 1.17.4.3); verbally verify the flap/slat

handle position, the flaps position indicator, and SLAT LAND light illumination, and

advise the flying pilot of the status of the annunciator panel.

For information on American’s postaccident changes to its Before Landing

checklist guidance, see section 1.17.5.1.

1.17.3.2.1 Manufacturer’s Information

Boeing’s Flight Crew Operating Manual (FCOM),

124

Volume II, section 2,

contains the following guidance on MD-80 spoiler and autobrake arming under the

heading “Before Landing Expanded Procedures”:

Lift SPOILER lever, observe lever remains up when released and red arm placard

is visible at the base of the lever.

123

The FAA reviewed human factors principles of checklist design and incorporated them into

its January 1995 publication, Human Performance Considerations in the Use and Design of Aircraft

Checklists

, which is available to airlines and FAA inspectors.

Factual Information 73 Aircraft Accident Report

Rotate AUTO BRAKE selector to desired deceleration rate (MIN, MED, or

MAX). Move ARM/DISARM switch to ARM.

1.17.3.3 Crew Coordination Procedures

American’s DC-9 Operating Manual, Volume 1, Normals, pages 101 and 102

(dated December 21, 1998), states the following regarding crew coordination procedures

for a Category I approach:

• Use of autopilot (if operative) is recommended with less than 4000 RVR

• Either pilot callout—“Radio Altimeter Alive”

• Pilot-Flying callouts:

–“Track—Track”

–“Outer Marker” and MSL crossing altitude

–“Auto Go—Auto Land” (if applicable)

The manual also states the following:

Approach

• Pilot-Flying: Fly approach.

• Pilot-Not-Flying: Monitor approach.

• Pilot-Not-Flying: Callouts:

–“1000” AFL on barometric altimeter—Verbally verify when

Flaps/Slats at landing setting

–“500” AFL on barometric altimeter—Airspeed ± Approach Speed and

Descent Rate

At 100 feet above DA [decision altitude] (on Baro[metric] Altimeter)

Pilot-Flying: When advised that visual references are in sight, confirm

requirements to descend below DA are satisfied, callout—“Landing” and

complete approach and landing.

124

The Chief Pilot for Flight Operations at Boeing’s Long Beach Division (who, at the time

of the flight 1420 accident, was Boeing’s Chief Test Pilot for Flight Operations at Boeing’s Long

Beach Division) indicated that the company’s FCOM is the primary tool used by the company to

conduct its training program and establish training programs for operators. Boeing’s FCOM provides

guidelines for operating procedures, but an operator, along with its principal operations inspector

(POI), may change the procedures. The Boeing Chief Pilot stated that the company sends the operators

updated information of proposed changes to the manual and that the operator is responsible for

coordinating the proposed changes with its POI.

Factual Information 74 Aircraft Accident Report

Pilot-Not-Flying: Callouts:

•“100 above”

• visual references when in sight

Pilot-Not-Flying: Direct primary attention to monitoring instruments.

At DA (on the Baro[metric] Altimeter)

Pilot-Flying: Execute a missed approach if not completing the landing.

Pilot-Not-Flying: Call out—“Decision Altitude”

Pilot-Not-Flying: Callouts:

•“100” AGL [above ground level] on Radio Altimeter

•“50, 40, 30, 20, 10” AGL on Radio Altimeter (if automated voice callouts are

inoperative)

1.17.3.4 Stabilized Approach Concept

American’s DC-9 Operating Manual, Volume 1, Techniques,

125

page 19 (dated

November 15, 1995), indicates the following under the heading “General”:

The stabilized approach concept requires that, before descending below the

specified minimum stabilized approach altitude, the airplane should be –

• in the final landing configuration (gear down and final flaps),

• on Approach Speed,

• on the proper flight path and at the proper sink rate,

• and at stabilized thrust.

These conditions should then be maintained throughout the rest of the approach.

125

Regarding American’s use of the word “techniques,” its DC-9 Operating Manual, Volume 1,

Conditionals, page 1 (dated November 15, 1995), states the following: “Techniques are not procedures,

but are suggested ways of accomplishing a task. These suggestions are based on experience and

recognized practices. Generally, they offer the best method to complete the task in most cases. However,

another technique may be just as appropriate, or even better, when considering the particular

circumstances.” The former MD-80 Fleet Manager indicated that American was gradually editing

out the Techniques section of all of its manuals. The information contained in this section will

be integrated into the Normals section so that all of the information for a particular phase of flight

will appear in one section.

Factual Information 75 Aircraft Accident Report

The minimum recommended stabilized approach altitudes are:

• VFR [visual flight rules] – 500 feet AFL

• IFR – 1000 feet AFL

In all cases, select landing flaps by 1000 feet AFL.

For information on American’s postaccident changes to its stabilized approach

procedures, see section 1.17.5.2.

American’s DC-9 Operating Manual, Volume 1, Normals, page 89 (dated

December 21, 1998), states that “on final, a callout will be made anytime any

crewmember observes LOC [localizer] displacement greater than 1/3 dot and/or G/S

[glideslope] displacement greater than 1/2 dot. The other pilot will acknowledge this

deviation.”

126

Page 89 also states that the pilot-not-landing is to make the following

deviation callouts: with landing flaps, any time airspeed varies more than ±5 knots from

approach speed and, inside the final approach fix, when rate of descent exceeds

1,000 feet per minute (fpm).

1.17.3.4.1 Federal Aviation Administration Guidance

FAA Order 8400.10, “Air Transportation Operations Inspector’s Handbook,”

volume 4, chapter 2, section 3, paragraph 511, “Stabilized Approach Concept,” states that

“maintaining a stable speed, descent rate, vertical flightpaths, and configuration is a

procedure commonly referred to as the stabilized approach concept” and that

“operational experience has shown that the stabilized approach concept is essential for

safe operations with turbojet aircraft.” The guidance also includes the following

information:

A stabilized approach for turbojet aircraft means that the aircraft must be in an

approved landing configuration…, must maintain the proper approach speed with

the engines spooled up, and must be established on the proper flightpath before

descending below the minimum “stabilized approach height” specified for the

type of operation being conducted. These conditions must be maintained

throughout the rest of the approach for it to be considered a stabilized approach.

Operators of turbojet aircraft must establish and use procedures which result in

stabilized approaches.

In addition, the guidance indicates that a stabilized approach needs to be

established before descending below 500 feet above the airport elevation during VFR

conditions or visual approaches and 1,000 feet above the airport or touchdown zone

elevation during any straight-in instrument approach in IFR conditions.

126

In postaccident interviews, the MD-80 Fleet Manager and several MD-80 check airmen explained

that captains used their discretion to determine the maximum acceptable deviation from the glideslope

or the localizer.

Factual Information 76 Aircraft Accident Report

1.17.3.5 Thunderstorm and Windshear Avoidance

American’s Flight Manual, Part I, Section 12, “Weather,” page 12 (dated

November 30, 1998), states the following regarding thunderstorm avoidance:

Do not enter or depart terminal areas when such areas are blanketed[

127

] by

thunderstorms except where known thunderstorm-free routes exist and are

followed. Airborne radar and all available weather reports will be used to make

this determination.

American’s DC-9 Operating Manual, Volume 1, Environmental, page 13 (dated

August 22, 1997), states the following regarding windshear avoidance:

Avoid areas of known severe windshear.[

128

] PIREPS [pilot reports] of windshear

in excess of 20 knots or 500 fpm climb or descent below 1000 feet AFL are a

good indication of such areas. Consider the time elapsed since the report and the

change in reported or observed (radar or visual) weather. Microbursts in

particular can create severe windshear conditions, but these conditions develop,

change, and dissipate rapidly.

The most dangerous form of windshear is a convective microburst. Some have

been documented with wind changes in excess of 200 knots. Because

microbursts intensify for several minutes after they first impact the ground, the

severity may be up to twice which is initially reported.

Search for clues which may indicate the presence of severe windshear. Severe

windshear has been encountered under the following conditions:[

129

]

• Thunderstorm and convective clouds

• Rain and snow showers

• Frontal systems

• Low altitude jet streams

• Strong or gusty surface winds

Page 14 of the operating manual states that, when positive indications of severe

windshear exist, avoid the areas by diverting around them, initiating a go-around

maneuver on approach, or holding on approach until conditions improve. Page 14 also

indicates that LLWAS can detect microbursts within 2 1/2 miles of the airport.

127

The former MD-80 Fleet Manager testified that “blanketed” meant “a significant amount of

coverage of the area.”

128

A precise definition of “severe windshear” was not found in American’s DC-9 operating or

flight manuals.

129

In public hearing testimony, the former MD-80 Fleet Manager stated that he believed that

at least four of the five criteria were probably present during the approach to Little Rock. However,

this official also stated that conditions conducive to windshear would necessitate a heightened level

of awareness but would not require an approach to be abandoned.

Factual Information 77 Aircraft Accident Report

American’s Flight Manual, Part I, Section 12, page 13 (dated December 19,

1997), indicates that LLWAS wind reports are “advisory only.” The manual also states

that “the reported surface winds, as presently obtained from the centerfield

instrumentation, are controlling for our Flight Operations.”

The former MD-80 Fleet

Manager stated at the public hearing that he would expect a pilot to evaluate an LLWAS

alert by considering the magnitude of the shear and its direction and then decide, with the

other pilot, whether to continue or abort the approach. If the crew’s decision was to

continue the approach, the former fleet manager stated that he would expect the pilot to

increase speed and be ready to immediately perform the escape maneuver if an

uncommanded change in pitch, roll, or rate of descent occurred because of the windshear.

The former MD-80 Fleet Manager also testified that, in November 1999,

American strengthened its guidance for handling windshear warnings. Specifically, the

guidance indicates that, if pilots receive a “microburst alert,” they are required to execute

a go-around or escape maneuver.

1.17.3.6 Continuation of an Approach Below the Decision Height

American’s Flight Manual, Part I, Section 10, page 13 (dated November 30,

1998), states the following regarding weather deterioration after the final approach

segment has started, in accordance with 14 CFR 121.651(c):

130

After the aircraft is established on the final approach segment, if the weather is

reported to be below published minima, the approach may be continued to the

appropriate DH [decision height] or MDA [minimum descent altitude], and

landing may be accomplished in accordance with the conditions for the type

approach being conducted.

The flight manual further states, “the final approach segment for an ILS approach

begins on the glide slope at the glide slope intercept altitude as shown in the [approach

chart] profile.”

The DC-9 Operating Manual, Volume 1, Normals, page 102, requires the

following to continue an approach below the MDA or DH, in accordance with

14 CFR 121.651(c)(1), (2), and (3):

• Airplane must be continuously in a position from which a descent to a

landing on the intended runway can be made at a normal rate of descent using

normal maneuvers, and where that descent rate will allow touchdown to

occur within the touchdown zone of the runway of intended landing.

• Flight visibility must not be less than the visibility prescribed in the standard

instrument approach being used.

130

Title 14 CFR 121.651(c) states the following: “If a pilot has begun the final approach segment

of an instrument approach procedure…and after that receives a later weather report indicating

below-minimum conditions, the pilot may continue the approach to the DH or MDA.”

Factual Information 78 Aircraft Accident Report

• Except for Category II or Category III approaches where any necessary

visual reference requirements are specified by authorization of the [FAA]

Administrator, at least one of the following visual references for the intended

runway is distinctly visible and identifiable to the pilot:

– Approach light system, except that the pilot may not descend below

100 feet above the touchdown zone elevation using the approach lights

as reference unless the red terminating bars or the red side row bars are

also distinctly visible and identifiable.

–Threshold

– Threshold markings

– Threshold lights

– Runway End Identifier Lights (REIL)

– Visual Approach Slope Indicator (VASI)

– Touchdown zone or touchdown zone markings

– Touchdown zone lights

– Runway or runway markings

– Runway lights

1.17.3.7 Missed Approach Policy

American’s Flight Manual, Part I, section 10, page 21 (dated April 15, 1998),

includes the following procedure for performing a missed approach:

When landing cannot be accomplished and, upon reaching the MAP [missed

approach point] defined on the approach chart, the pilot must comply with the

missed approach procedure or with an alternate missed approach procedure

specified by ATC.

The missed approach procedures were revised on August 15, 1999, to add an

introductory paragraph that is titled “General” and states the following:

American Airlines has a no-fault go-around policy, recognizing that a successful

approach can end in a missed approach. Captains are required to execute/order a

missed approach if the aircraft is not stabilized by 1000’ AFL (IFR) or 500’ AFL

(VFR), or if in the pilot’s judgement a safe landing cannot be accomplished

within the touchdown zone, or the aircraft cannot be stopped within the confines

of the runway.

131

The former MD-80 Fleet Manager indicated that American always had a policy of not holding

a pilot accountable for performing a go-around and that this revision would ensure that all pilots

were aware of this policy.

Factual Information 79 Aircraft Accident Report

1.17.4 Landing Procedures

1.17.4.1 Wind Landing Limits

American’s Flight Manual, Part 1, section 10, page 20 (dated April 7, 1999),

indicates that “pilots shall secure the latest surface wind direction and velocity prior to

making a landing at an airport.” Section 12 of the flight manual, page 3 (dated April 4,

1997), indicates that “the latest surface wind for use for crosswind and/or headwind and

tailwind limitations will be that in the latest weather observation unless a more recent oral

report is available from an operating control tower.”

Section 10 of the flight manual, page 20, also states that, for approaches conducted

with an RVR of 4,000 feet or a visibility of 3/4 mile or greater, the dry runway

132

maximum demonstrated landing crosswind component for a DC-9 is 30 knots. The

manual further states that, for approaches conducted with an RVR of less than 4,000 feet

or 3/4-mile visibility, the maximum landing crosswind component is 15 knots and that, for

approaches conducted with an RVR of less than 1,800 feet or 1/2-mile visibility, the

maximum landing crosswind component for the DC-9 is 10 knots. According to the

manual, if the captain believes that “environmental conditions or braking reports indicate

that the runway is wet or slippery, the maximum acceptable crosswind should be reduced

to 20 knots.”

American’s DC-9 Operating Manual, Limitations, page 3 (dated April 26, 1999),

states that the maximum tailwind limit for takeoffs and landings is 10 knots. The manual

also states that the tailwind limit might be further reduced by performance requirements.

1.17.4.1.1 Manufacturer’s Information

According to Boeing, the MD-80 is required to safely land with a crosswind of at

least 20 knots.

133

An aerodynamics engineer from Boeing testified at the public hearing

that, during crosswind certification testing, the airplane was found to be capable of

handling 30 knots of crosswind. The Procedures section of Boeing’s MD-80 Airplane

Flight Manual states the following:

The limiting crosswind value has not been determined; however, the maximum

demonstrated crosswind component for takeoff and landing is 30 knots reported

wind at the fifty foot height. This value was demonstrated on a dry runway….

The accepted industry practice is that the operator use this demonstrated

capability and their operational experience to construct their crosswind landing

guidance material which is approved by the FAA.

132

Section 10 of the flight manual, page 25 (dated April 7, 1999), states that an airport runway

is considered to be dry when no snow, slush, ice, or water have been reported and “no more than

the following conditions” have been reported: scattered showers in the airport area; intermittent drizzle

with an intensity no greater than moderate, intermittent light rain with surface temperatures above

freezing; and light snow with surface temperatures below 28º F. Precise definitions for a “wet”

or “slippery” runway were not found in American’s DC-9 Operating Manual or Flight Manual.

133

The FAA’s crosswind certification requirement is that an airplane must be capable of safely

landing with a 90º crosswind component of at least 20 percent of the stall speed for dry runway conditions.

Factual Information 80 Aircraft Accident Report

Boeing’s MD-80 FCOM states, under the heading “Operational Limitations,” that

the limiting tailwind component is 10 knots.

1.17.4.2 Spoiler Deployment Procedures

American Airlines’ DC-9 Operating Manual, Volume 1, Techniques, page 21

(dated November 15, 1995), states that the automatic deployment of the spoilers after

touchdown should be monitored and that, if the spoilers do not deploy automatically, the

captain should manually deploy them. The manual also indicated, in its Normals section,

page 75 (dated December 21, 1998), that “if Spoiler Lever does not move back to full aft

(EXT) [extend] position, the Captain, regardless of which pilot is making the landing,

will manually deploy spoilers.” Page 75 also indicated that the flying pilot and the

nonflying pilot are to check that the spoiler lever is full aft after the airplane has touched

down. For information on American’s postaccident changes to its spoiler deployment

procedures, see section 1.17.5.1.

1.17.4.2.1 Manufacturer’s Information

Boeing’s FCOM, Volume II, section 2, contains the following guidance on

MD-80 spoiler deployment under the heading “Landing Roll Expanded Procedures”:

For automatic deployment of inboard (ground) spoilers, throttles must be at idle.

If throttles are above idle at touchdown, outboard and inboard spoilers may

deploy and retract and ABS will disarm.

If SPOILER lever does not move aft or does not remain at EXT [extend] position,

PNF call, “No Spoilers,” PF move lever aft to full extend position and up to

latched position.[

134

]

1.17.4.3 Braking Procedures

American’s DC-9 Operating Manual, Volume 1, Techniques, page 21 (dated

November 15, 1995), states that “the prudent use of manual braking has been shown to

reduce brake wear” and that “unless circumstances dictate otherwise, manual braking is

generally recommended for landing.” Volume 1, Normals, page 74 (dated April 26,

1999), indicates that pilots should use aggressive manual braking or maximum

autobrakes on short or slippery runways. Volume 1, Environmental, page 3 (dated

November 15, 1995), states under the section “Manual Brake Stopping” that “for short or

slippery runways, immediately after nose gear touchdown, use full brake pedal.”

Volume 1, Environmental, page 27 (dated April 26, 1999), states that, for landing on a

134

The Boeing MD-80 Chief Pilot stated his belief that this procedure had been a part of the

company’s manual since 1991. The Chief Pilot indicated that most, if not all, of the domestic operators

coordinate with the manufacturer, normally by requesting a letter of no technical objection, before

making any changes to a manual. The Chief Pilot pointed out that there is no legal requirement

for the operator to request a letter of no technical objection. In its April 20, 2000, letter to the

Safety Board (see section 1.6.2.2), Boeing indicated that it could not find any record of issuing

a letter of no technical objection to American regarding alteration of the manufacturer’s recommended

spoiler or autobrake procedures.

Factual Information 81 Aircraft Accident Report

slippery runway, “use aggressive manual braking or maximum auto brakes and auto

spoilers.”

American’s DC-9 Operating Manual, Volume 1, Environmental, page 4 (dated

January 31, 1997), states the following regarding the anti-skid system:

When brakes are applied on a slippery runway, several skid cycles will occur

before the anti-skid system establishes the right amount of brake pressure for the

most effective braking. If the pilot modulates the brake pedals, the anti-skid

system is forced to readjust the brake pressure to re-establish optimum braking.

During this readjustment time, braking efficiency is lost.

On extremely slippery runways at high speeds, the pilot is confronted with a

rather gradual deceleration and may interpret the lack of an abrupt sensation of

deceleration as a total anti-skid failure. The natural response might be to pump

the brakes or turn off the anti-skid. Either action will degrade braking

effectiveness.

For information on American’s postaccident changes to its braking procedures,

see section 1.17.5.1.

1.17.4.3.1 Manufacturer’s Information

Boeing’s FCOM, Volume II, section 2, contains the following guidance on

MD-80 braking procedures under the heading “Landing Roll Expanded Procedures”:

When nose gear is firmly on runway, apply sufficient down elevator after nose

gear contact to increase weight on the nosewheel for improved steering

effectiveness. (An excessive amount of down elevator will unload the main gear

and reduce braking efficiency.)

Volume II, section 3, of the manual contains the following information on

operating on wet or slippery runways:

On contaminated surfaces, full braking should be used to realize optimum

antiskid operation. Autobrakes if available, should be used in the maximum

setting. The normal braking technique on slippery runways is that immediately

after nose gear touchdown, apply brake pressure smoothly and symmetrically

with maximum pedal pressure and hold until a safe stop is assured.

If a landing is planned on a runway contaminated with snow, slush, standing

water, or during heavy rain, the following factors must be considered: available

runway length; visibility of runway markers and lights; wind direction and

velocity; crosswind effect on directional control; braking action; and the

probability of hydroplaning and its effect on stopping distances.

If a skid develops, especially in crosswind conditions, reverse thrust will increase

the sideward movement of the airplane, in this case release brake pressure and

reduce reverse thrust to reverse idle, and if necessary, forward idle. Apply rudder

Factual Information 82 Aircraft Accident Report

as necessary to realign the airplane with the runway and reapply braking and

reversing to complete the landing roll.

1.17.4.4 Use of Reverse Thrust

American Airlines’ DC-9 Operating Manual, Volume 1, Techniques, page 21,

(dated November 15, 1995), indicates the following:

The application of reverse thrust tends to blank out the rudder. The effectiveness

of the rudder starts decreasing with the application of reverse thrust and at

90 knots, at 1.6 EPR (in reverse) it is almost completely ineffective.

If the airplane starts drifting across the runway while reversing, immediately

return the reverse thrust levers to idle reverse to assist in regaining directional

control and to restore rudder effectiveness.

American’s DC-9 Operating Manual, Volume 1, Environmental, page 7 (dated

April 26, 1999), states the following under the heading “Slippery Runway – Crosswind”:

One of the worst situations occurs when there is a crosswind and sufficient water

and speed to produce total tire hydroplaning. Reverse thrust tends to disrupt

airflow across the rudder and increase the tendency of the airplane to drift

downwind, especially if a crab or yaw is present.

As reverse thrust increases above 1.3 EPR, rudder effectiveness decreases until it

provides no control at about 1.6 EPR. Do not exceed 1.3 EPR reverse thrust on

the slippery portions of the runway, except in an emergency.

Page 27 includes the following guidance for landing on a slippery runway:

• Apply reverse thrust as soon as possible after nosewheel touchdown. Do not

exceed 1.3 EPR reverse thrust on the slippery portions of the runway, except

in an emergency.

• When reversing, be alert for yaw from asymmetric thrust. If directional

control is lost, bring engines out of reverse until control is regained.

• Do not come out of reverse at a high RPM. Sudden transition of reversers

before engines spool down will cause a forward acceleration.

For information on American’s postaccident changes to its guidance on the use of

reverse thrust, see section 1.17.5.1.

1.17.4.4.1 Manufacturer’s Information

At the Safety Board’s public hearing on this accident, the Chief Pilot for Flight

Operations at the Boeing Commercial Airplane Group’s Long Beach, California,

Division presented information on MD-80 rudder effectiveness during reverse thrust

operation. The MD-80 Chief Pilot described thrust reversers as “clam shells” that open up

Factual Information 83 Aircraft Accident Report

and redirect engine air flow to help slow the airplane. He indicated that, during reverse

thrust operation, the overall effectiveness of the rudder to steer the airplane could be

reduced because of the disrupted air flow over the rudder and the vertical stabilizer.

According to the Chief Pilot, the thrust reversers disrupt the air that would normally be

flowing in a streamline across the rudder and vertical stabilizer, creating a field of air that

would be turbulent.

The Chief Pilot stated that an MD-80 traveling at 90 to 100 knots with a 1.6 EPR

(dry runway) setting would lose directional control from the rudder. He also stated that,

because of the reduced rudder effectiveness, Boeing recommended coming out of reverse

thrust until the airplane was going in the intended direction and that, once directional

control from the rudder is restored, reverse thrust could be reapplied.

Boeing’s FCOM, Volume II, section 2, contains the following guidance on

MD-80 reverse thrust operations under the heading “Landing Roll Expanded

Procedures”:

On a dry runway, reverse thrust of no more than 1.6 EPR should be used, except

in an emergency.

On wet or contaminated runways, reverse thrust of no more than 1.3 EPR should

be used, except in an emergency.

If difficulty in maintaining directional control is experienced during reverse

thrust operations, reduce thrust as required and select forward idle, if necessary,

to maintain or regain directional control. Do not attempt to maintain directional

control by using asymmetric reverse thrust.

1.17.4.4.2 MD-80 All Operators Letter

On February 15, 1996, McDonnell Douglas issued an all operators letter to all

MD-80 operators regarding handling characteristics when landing on wet or slippery

runways and a change to reverse thrust management techniques. The letter stated that,

because of the reverser efflux pattern resulting from the MD-80’s canted reverse thrust

buckets, “the aerodynamic forces acting on the vertical stabilizer and rudder are disrupted

by an increase in reverse thrust above approximately 1.3 EPR, thus reducing the ability of

the rudder and vertical stabilizer to provide optimum directional control.” The letter

further stated that, “as reverse thrust increases above approximately 1.3 EPR, rudder and

vertical stabilizer effectiveness continue to decrease until at reverse thrust greater than

approximately 1.6 EPR the rudder and vertical stabilizer provide little or no directional

control.”

The letter emphasized that rudder effectiveness is extremely important when

surface friction is low and crosswind or tailwind conditions are present. The letter

indicated that, under these conditions, directional control may only be available from the

rudder. The letter also indicated that McDonnell Douglas was revising its recommended

FCOM procedures to limit reverse thrust to 1.3 EPR when landing on wet or slippery

Factual Information 84 Aircraft Accident Report

runways. (Section 1.17.4.4.1 shows the revised recommended procedures, which were in

effect at the time of the Little Rock accident. The procedures that were in effect at the

time the all operators letter was issued recommended reverse thrust settings up to

1.6 EPR.)

The letter also stated that, on wet or slippery runways, sufficient nose down

elevator should be applied after nose gear contact for improved steering effectiveness.

The letter cautioned that an “excessive amount” of down elevator should not be applied

because it would unload the main gear and reduce braking efficiency.

1.17.5 Postaccident Actions

After the Little Rock accident, American conducted a three-phase operational and

safety evaluation from August to November 1999.

135

The first part of the evaluation was

an independent audit that focused on flight training and the organization and operations

of the Flight Department; American sought the input of professional pilots outside the

organization who had managed airline operations. According to the AMR Vice

Chairman, American “wanted the benefit of having outside eyes looking at how we did

business.” The second part of the evaluation was a review by the Allied Pilots

Association of the same subjects as in the first evaluation.

The third part of the evaluation was conducted by the “System Analysis Team,”

which consisted of representatives from American, the Allied Pilots Association, and the

FAA. The team performed a 120-day review of six accidents and two incidents that

occurred during the previous 5 years at American to determine why those events occurred

and what approaches the airline could take to prevent such events from recurring.

The three phases of the evaluation produced 85 recommendations.

136

American

and the Allied Pilots Association prioritized the recommendations according to

importance or safety impact and created a timeline in which to accomplish the

recommendations.

137

At the public hearing, the AMR Vice Chairman indicated that the list

of recommendations had been reviewed by the FAA’s Certificate Management Office for

AMR Corporation and presented to the Flight Standards Group at FAA headquarters.

Figure 17 shows a table, presented by the AMR Vice Chairman, that highlights

recommendation categories, their expected completion dates, and their safety impact.

135

The former MD-80 Fleet Manager cited two more company actions that were also the result

of the Little Rock accident. First, American presented an in-depth review of material related to landing

with adverse runway conditions to all MD-80 check airmen at their third quarter 1999 standardization

meeting. Second, American retained a radar expert to conduct a 1-day seminar for its check airmen

on the use of airborne weather radar. These seminars, which were held between October and

December 1999, reviewed each airplane’s individual weather radar systems, the types of echoes that

the systems might portray, and the conditions that would be associated with those echoes.

136

Because the audits were conducted independently, many of the recommendations addressed

common issues. The audits made recommendations in the following general issue areas: structure

and effectiveness of the Flight Safety Department, flight training and flight standards, and flight operations.

137

The FAA’s POI for American Airlines stated that 17 recommendations were deemed “high

priority” with the goal for immediate implementation.

Factual Information 85 Aircraft Accident Report

Figure 17. Categories of the Postaccident Audit Recommendations According to

Expected Completion Date and Safety Value

On July 20, 2001, American Airlines provided the Safety Board with an update on

actions to address the issues that had been identified during the three postaccident safety

and operational audits. In addition to the actions involving the reorganization of the

company that were previously discussed at the beginning of section 1.17 and the changes

to the go-around policy previously stated in section 1.17.3.7, American made changes to

its checklist philosophy to require challenges by the nonflying pilot and responses by the

flying pilot for all mechanical checklist items (section 1.17.5.1 for the changes to the

Before Landing checklist items) and expanded its stabilized approach guidance (see

section 1.17.5.2).

American’s Flight Safety Department undertook actions to increase its ability to

identify and track operational trends and enhanced its internal evaluation program.

American is also developing a formal risk management program and a Flight Operations

Quality Assurance program that will be integrated with ASAP. In addition, Systems

Analysis Team projects were completed, including one associated with pilot

decision-making during the approach and landing phases of flight and one associated

with stabilized approaches.

American has also acted to improve its severe weather operational guidance. For

example, the company revised and expanded its thunderstorm avoidance policies,

clarified its wind landing limits to include gusts, expanded the quick reference crosswind

table, strengthened guidance for landing on slippery runways during crosswinds,

provided all pilots with the manufacturer’s weather radar booklet, and entered into a

Highlights by Category

• Ensure all crewmembers attend

the fatigue countermeasures

program

• Develop computer-based

training and digital video training

for special airports

• Improve Chief Pilot

communication skills

• Independent Flight Standards

program to monitor flight training

• Implement Special Procedures

Operational Training

• Flight Operations/System

Operations Control Department

reorganization

• Reexamine Advanced

Qualification Program and

Single Visit Exemption

•Publish Flight Safety magazine

• Improve human factors focus on

pilot decision-making

• Implement Flight Operations

Quality Assurance

• Enhance safety audits

• Develop safety performance

metrics

• Upgrade safety tracking system

• Revise thunderstorm avoidance

policy

• Revise approach and landing

policy

• Safety Department reorganization

• Vice President of Safety

• Departmental Safety Directors

• Restore relationship with Allied

Pilots Association Safety

Committee

Jan-Mar 00 April-June 00 July 00-Jan 02

Factual Information 86 Aircraft Accident Report

contract to retrofit airborne predictive windshear weather radar on all company airplanes.

In addition, American added pilot decision-making guidance during the approach and

landing phases of flight to its Airplane Flight Manual Part I and distributed a technical

bulletin to MD-80 pilots that addressed rollout deceleration factors and considerations

specific to that airplane.

In addition, training and evaluation standards were revised. For example,

American provided its check airmen with additional weather radar training, updated its

Human Factors and Safety Training to include additional events that involve flight crew

decision-making, and modified the frequency and duration of training cycles under the

Single Visit Exemption of the Advanced Qualification Program. Also, “Hot Briefing

Items” on thunderstorm avoidance, weather radar usage, landing on contaminated

runways, spoiler and deceleration device logic, and autobrake usage have been presented

by all of American’s fleets. Further, American is developing an intranet-based continuing

education training program on radar use and is considering the use of Special Procedures

Operational Training as a replacement to LOFT.

American also indicated that it established and implemented fatigue and reserve

rest policies for its flight crews. In addition, the company published a “Commitment to

Safety” statement to all employees and work groups.

1.17.5.1 DC-9 Operating Manual Changes

American made several changes to its DC-9 Operating Manual after the Palm

Springs incident, which occurred 8 months after the Little Rock accident. On

February 23, 2000, 1 week after the Palm Springs incident, American issued

Revision No. 31 to its DC-9 Operating Manual. The revision indicated that the Before

Landing checklist would be accomplished as follows:

• The pilot-not-flying will call out each item on the mechanical checklist.

• The pilot-flying will visually verify that the item has been accomplished, and

make the appropriate responses aloud.

• The pilot-not-flying will also visually verify each item, and after obtaining a

response, will move the corresponding switch on the mechanical checklist.

• Both pilots will respond to the ALTIMETERS and FLT INSTR & BUGS

items.

The revision also stated that the “captain will always arm the spoiler for landing.”

On June 22, 2000, American issued DC-9 Operating Manual Bulletin DC-9-42,

which further revised the procedures for accomplishing the Before Landing checklist.

The bulletin indicated that pilots were to “accomplish the Before Landing checklist by

challenge and response as follows”:

Factual Information 87 Aircraft Accident Report

Pilot-Not-Flying

• Call out each item on the mechanical checklist.

• Verify that item has been accomplished.

• Call out response and move corresponding switch on mechanical checklist.

Pilot-Flying

• Verify and respond to the following items:

– ALTIMETERS

– FLT INSTR & BUGS

–GEAR

– SPOILER LEVER

– FLAPS & SLATS

In conjunction with the 1000 foot AFL callout, both pilots will visually confirm

that all tabs on mechanical checklist are closed-out, and the pilot-not-flying will

call out – “Before Landing checklist complete.”

According to American’s February 23, 2000, revision to its DC-9 Operating

Manual, both pilots are to check for spoiler deployment. The revision indicates that the

captain is still to manually deploy the spoilers if the spoiler lever does not move back to

the full aft (extend) position. The revision further indicates that, when the spoilers

deploy, the nonflying pilot is to call out “deployed”; if the spoilers do not deploy or fail

to remain deployed, the nonflying pilot is to call out “no spoilers,” and the captain is to

manually deploy the spoilers.

On June 27, 2000, American issued DC-9 Operating Manual Bulletin DC-9-43,

which revised the policy on the use of autobrakes for landing. The revision stated the

following:

Autobrakes

If operative, must be armed prior to landing when any of the following conditions

exist:

• Runway length less than 7000 feet

• RVR less than 4000 or visibility less than 3/4 mile

• Runway contaminated with standing water, snow, slush, or ice

• Braking conditions reported less than good

In addition, the use of autobrakes is recommended when landing with gusty

winds or crosswinds.

Factual Information 88 Aircraft Accident Report

Autobrake settings should be appropriate to the conditions: MAX must be used

when minimum stopping distance is required (MAX autobrake deceleration rate

is slightly less than that produced by full manual braking).

After landing, intervene with manual braking as necessary to slow the airplane at

the desired rate.

American’s February 23, 2000, revision to its DC-9 Operating Manual indicated

that the following reverse thrust procedures were effective immediately for all MD-80

landings:

When nosewheel steering is on the ground, the PF will select idle reverse and

apply slight forward pressure on the yoke.

With spoilers deployed and directional control assured, reverse thrust may be left

at idle or increased to a target of approximately 1.3 EPR.

Reverse thrust of more than 1.3 EPR should not be used unless stopping distance

is in doubt.

Do not remain in reverse thrust if directional control cannot be maintained and do

not use asymmetrical reverse thrust to regain directional control.

1.17.5.2 Flight Manual Changes

American’s Flight Manual, Part I, Section 10, page 3 (dated July 21, 2000),

revised the company’s stabilized approach criteria to state the following:

138

Significant speed and configuration changes during an approach can complicate

aircraft control, increase the difficulty of evaluating an approach as it progresses,

and complicate the decision at the decision point; i.e., DA, DH, MDA. A pilot

must assess the probable success of an approach before reaching the decision

point. This requires the pilot to determine that requirements for a stabilized

approach have been met and maintained.

To limit configuration changes at low altitude, the airplane must be in a landing

configuration by 1,000 feet AFL (gear down and landing flaps).

A stabilized approach must be established before descending below the following

minimum stabilized approach heights:

• IMC [instrument meteorological conditions] – 1000 feet AFL

• VMC [visual meteorological conditions] – 500 feet AFL

138

This information also appears in American’s DC-9 Operating Manual in a section titled

“Approach—Landing—Go Around” (dated April 27, 2001).

Factual Information 89 Aircraft Accident Report

A stabilized approach means that the airplane must be:

• At Approach Speed (V

REF

+ additives)

• On the proper flight path at the proper sink rate

• At stabilized thrust

These requirements must be maintained throughout the rest of the approach for it

to be considered a stabilized approach.

If the stabilized approach requirements cannot be satisfied by the minimum

stabilized approach heights or maintained throughout the rest of the approach, a

go-around is required.

Section 10, pages 1 and 2 (dated July 21, 2000), provide expanded approach and

landing decision-making guidance. For example, under the heading “Thunderstorms and

Microburst,” the guidance explicitly states that “takeoffs and landings are not permitted

when thunderstorms are near the airport unless the runway and flight path are clear of

thunderstorm hazards.” In addition, under the heading “Decision Factors,” the guidance

provides a list of factors that can influence the decision to begin and continue an

approach to either a landing or missed approach, including convective activity,

turbulence, visibility/RVR, crosswind, precipitation, and weather trends. Further, the

guidance indicates that “pilots should be especially aware of the cumulative factors and

trends. For example…thunderstorms approaching an airport suggests holding until

passage.”

1.17.6 Federal Aviation Administration Oversight

The FAA Certificate Management Office for AMR Corporation is located in

Dallas, Texas. The principal operations inspector (POI) for American Airlines stated that

the Certificate Management Team was staffed with 17 air safety inspectors and 1 cabin

safety inspector at the time of the flight 1420 accident. Of the 17 air safety investigators,

10 were operations inspectors, including the MD-80 Aircrew Program Manager (APM)

and the Assistant MD-80 APM, and 7 were geographic inspectors who worked on the

American certificate and were assigned to the Air Transportation Oversight System

(ATOS) program office.

139

The ATOS program was implemented at American Airlines in February 1999. As part of

the program, each carrier was to be assigned an FAA analyst to determine trends in reported data.

At the time of the accident, the analyst for American had not been assigned. In fact, at the time

of the Safety Board’s public hearing on this accident, only 1 of the 10 carriers had been assigned

an analyst. American’s analyst has now been hired and was expected to begin work on April 24,

2001. The ATOS program will be discussed in detail in the Board’s final report on the January 31,

2000, Alaska Airlines flight 261 accident.

Factual Information 90 Aircraft Accident Report

1.17.6.1 Principal Operations Inspector

The POI for American Airlines authorizes the airline’s operations through the

issuance of operations specifications. The POI is also responsible for approving manuals

and their revisions and reviewing surveillance records of training and line operations.

The POI indicated during public hearing testimony that, when a manufacturer has

recommended revisions to its operating manual, he routes the revisions to the appropriate

APM, who reviews the information with the carrier. The POI stated that he relies very

heavily on the opinion of the appropriate APM in deciding whether to approve a change

to the operating manual because the APM has the technical expertise for the specific

airplane and works closely with the FAA’s Aircraft Evaluation Group, the manufacturer,

and the carrier. The POI also indicated that a carrier might choose not to make a

manufacturer’s suggested change because of the way that carrier has configured the

particular airplane.

According to his testimony, the POI indicated that he ensures that pilot training is

being performed satisfactorily through the APMs. The POI also stated that he is notified

when any training event is not in compliance with company procedures and when line

operations do not comply with company procedures.

The POI testified that a hiring freeze had severely impacted his office’s ability to

conduct surveillance. According to the POI, his emphasis areas at the time of the

flight 1420 accident included the surveillance of American’s two new airplane fleets (the

737 and the 777) and the company’s South American operations. The POI indicated that,

at the time of the public hearing, he had 16 air safety inspectors (6 of which were

geographic inspectors) but wanted a total of 30 air safety inspectors (10 of which would

be geographic inspectors). In addition, the POI indicated that he did not have a say in

where the geographic inspectors were physically located. For example, at the time of the

public hearing, the POI’s staff included a geographic inspector in Las Vegas but not one

in New York or San Francisco, where American has a large volume of operations.

140

1.17.6.2 Aircrew Program Manager

The FAA’s APM for American’s MD-80 program at the time of the accident had

worked in this position since 1984. The APM, along with the Assistant APM for

American’s MD-80 program, were responsible for all operational aspects of American’s

MD-80 fleet, including the flight training program. The APM and Assistant APM were

current and qualified on the MD-80.

In a postaccident interview, the APM stated that workload and personnel

constraints did not allow him or the Assistant APM to monitor most of American’s

proficiency check rides and type rating certifications of pilots. The APM indicated during

140

In August 2000, the POI for American indicated that he was “finally” assigned a geographic

inspector who is located in New York. The POI indicated that the inspector would be ready to

conduct inspections at the beginning of 2001—after he completes ATOS training and receives

carrier-specific briefings.

Factual Information 91 Aircraft Accident Report

that interview that, under his supervision, approximately 17 check airmen from American

(appointed by the FAA as aircrew program designees) did most of the airman

certification activities.

141

He further indicated that the only certification work he

performed was examining a new aircrew program designee or rechecking a pilot who had

failed two consecutive check rides. The APM stated that he tried to observe the aircrew

program designees each year before they were renewed, in accordance with volume 5 of

FAA Order 8400.10, “Air Transportation Operations Inspector’s Handbook.”

The APM indicated that workload constraints prevented him from observing all of

American’s check airmen on a regular basis. The APM stated that he observed all check

airmen at least every 2 years, in accordance with FAA Order 8400.10, volume 3.

The APM explained that he and the Assistant APM were the only ones

responsible for observing American’s MD-80 training. He further stated that they were

“spread thin” and that it was “extremely difficult to cover all of the things that are going

on at American Airlines, from an observation standpoint of training.” The APM stated

that he tried to review the training manuals once a year and that the only simulator

training he had time to observe was when a check airman was performing a proficiency

check.

The APM testified that he relied heavily on American Airlines to ensure

standardization in its simulator training program. According to his testimony, American’s

senior designated examiners, referred to as coordinators, observe every check airman and

every simulator instructor at least once a year. The APM stated that these coordinators

were “highly qualified.” In addition, the APM indicated that he tried to further efforts to

standardize simulator training through discussions with the MD-80 check airmen during

American’s quarterly check airmen meetings.

The APM stated that the MD-80 was among the airplane models in American’s

junior fleet and that about 75 percent of American’s new upgrade captains and 50 percent

of the company’s new hire pilots were assigned to the MD-80 fleet. He also indicated that

American’s MD-80 fleet averages about 28 new captains per month and that the number

of check airmen tripled and the number of designated examiners doubled between the

second half of 1998 and the end of 1999. The APM stated that, because of the training

workload and personnel constraints, geographic inspectors (including those from offices

in Chicago, New York, and Nashville) provided some assistance in observing IOE and

line checks.

The FAA’s PTRS showed that, for American’s MD-80 fleet, surveillance of

check airmen was performed 27 times in 1998; surveillance of check airmen

administering a line check, 25 times; and surveillance of check airmen during IOE,

60 times. From January to June 1, 1999, surveillance of check airmen had been

performed 9 times for American’s MD-80 fleet; surveillance of check airmen

141

The APM testified that, 10 to 15 years ago, 5 or 6 inspectors from his program did all

of the certification work but that the FAA did not have that “luxury” any longer because of a

reduction in personnel.

Factual Information 92 Aircraft Accident Report

administering a line check, 18 times; and surveillance of check airmen during IOE,

49 times.

According to the APM’s testimony, he and the Assistant APM conducted en route

inspections to observe whether American’s procedures were being correctly performed

during line operations. The APM indicated that he critiques flight crews immediately if

differences are observed. He further indicated that he has never seen a major problem

during en route inspections but that flight crewmembers are on “good behavior” when

inspectors are present. The FAA’s PTRS showed that, for American’s MD-80 fleet, en

route surveillance was performed 795 times in 1998 and 537 times from January to

June 1, 1999.

1.18 Additional Information

1.18.1 Runway Friction Information

The senior research engineer from NASA’s Langley Research Center presented

information at the public hearing about the three types of friction loss that can occur on a

wet runway surface.

142

They are viscous hydroplaning, dynamic hydroplaning, and

reverted rubber skidding (or locked tires). His description of each follows:

The contributing factors for viscous hydroplaning are a damp or wet pavement,

medium to high speed, poor pavement texture, and worn tire tread. If a runway

has good microtexture

143

and grooving and the aircraft tires have a good tread

design, viscous hydroplaning could be alleviated.

144

The contributing factors for dynamic hydroplaning are a flooded pavement, high

speed, low tire pressure, and worn tire tread. If a runway has good macrotexture

145

and grooving and the aircraft tires have high pressure and good tread design,

dynamic hydroplaning could be alleviated.

146

142

The NASA Langley engineer presented the following runway wetness classifications: damp—

moisture present on the surface to a depth of less than 0.01 inch, wet—standing water on the surface

to a depth between 0.01 and 0.1 inch, and flooded—standing water on the surface to a depth that

exceeds 0.1 inch. The NASA Langley engineer testified that the left side of runway 4R was flooded

but that the right side was not. He also indicated that the crosswind conditions were not severe

enough for the flooding from the left side to have exceeded the runway’s centerline.

143

Microtexture is a small sandpaper-type of texture that can only be felt.

144

According to the article, “Landing on Slippery Runways,” in Boeing’s October to December 1992

Airliner magazine, “viscous hydroplaning occurs on all wet runways and is a technical term used

to describe the normal slipperiness or lubricating action of the water.” The article also stated that

viscous hydroplaning reduces friction but not to the level that would prevent an airplane’s wheel

from spinning up shortly after touchdown.

145

Macrotexture is the large roughness in the surface that is visible to the eye.

146

According to the article, “Landing on Slippery Runways,” in Boeing’s October to December 1992

Airliner magazine, dynamic hydroplaning “lifts the tire completely off the runway and causes such

a substantial loss of tire friction that wheel spin up may not occur.”

Factual Information 93 Aircraft Accident Report

The contributing factors for reverted rubber skidding

147

are a wet or flooded

pavement, high speed, poor pavement texture, and a deficient brake system. To

alleviate reverted rubber skidding, a good pavement structure and grooving and

improved antiskid control devices are necessary.

The NASA Langley engineer indicated that a pavement’s capability to alleviate

slipperiness improves as its microtexture and macrotexture increase. The potential to

alleviate slipperiness on smooth pavement surface that is damp or flooded is poor,

whereas the potential to alleviate slipperiness on a porous pavement surface that is damp

or flooded is excellent. For a transverse grooved runway, such as runway 4R at Little

Rock, the potential to alleviate slipperiness during damp conditions is excellent, and the

potential to alleviate slipperiness during flooded conditions is good to excellent. The

engineer stated that runway 4R’s microtexture was above average, macrotexture was

excellent, and grooving was satisfactory. In addition, he stated that the runway’s ability

to prevent hydroplaning and other braking problems was excellent.

148

The engineer stated that dynamic hydroplaning was not a factor in the accident

because of the scrub marks on the runway.

149

He further stated that the water depth that

would have produced a dynamic hydroplaning effect on runway 4R was approximately

0.28 inches and that, at flight 1420’s touchdown point, there would have been less than

0.10 inch of water in that area of the runway located 15 feet from the centerline because

of the crosswinds that were moving from left to right.

150

In addition, the engineer stated

that he found no evidence of reverted rubber skidding because the runway had good

pavement texture and grooving and there was no tread reversion on the four main landing

gear tires and no reverted rubber on the runway surface.

The NASA research engineer also provided testimony about antiskid operations.

The engineer indicated that normal antiskid operation would be expected with high wheel

spin-up accelerations on a high-to-medium runway traction surface and early spoiler

deployment at touchdown. He also stated that abnormal antiskid operation would be

expected with low- to no-wheel spin-up accelerations on a low- to no-traction runway

surface and delayed spoiler deployment at touchdown. According to the NASA engineer,

a dilemma with antiskid control systems during crosswind operations is that, to preserve

147

The terms “reverted rubber skidding” and “reverted rubber hydroplaning” can be used

interchangeably.

148

The Boeing aerodynamics engineer stated during his testimony that runway 4R is “very good”

because it has been grooved and has good surface texture.

149

The NASA Langley engineer indicated that the scrub marks (white-appearing tracks on the

pavement that were lighter in color than the adjacent areas, as described in section 1.10.3.1) resulted

from the high pressure between the tire print and the wet pavement. He indicated that the marks

were caused by the braking action between the tire and the pavement and the steering forces being

developed between the tire and the pavement. The NASA Langley engineer stated that the scrub

marks led up to tire marks going through the grass and to the accident site. He also stated that

no black rubber tire marks were evident on the runway because the pavement surface was wet.

150

The NASA Langley engineer explained that, under the accident conditions and with the high

vertical sink rate at which the airplane touched down, he had “no problem or no question about

the wheels spinning up on touchdown.”

Factual Information 94 Aircraft Accident Report

braking, the antiskid must operate at increasing slip ratios as the airplane yaws but that, to

preserve cornering, the antiskid must operate at low slip ratios.

151

The engineer stated his

belief that flight 1420 could not develop any cornering friction with the wet runway

conditions because of the yaw angles the airplane experienced during the accident

sequence.

1.18.2 Study on Thunderstorm Penetration in the Terminal Area

Research staff at the Massachusetts Institute of Technology’s Lincoln Laboratory

conducted a study, sponsored by NASA, of the variables that are correlated with arriving

pilots’ convective weather penetrations or deviations in the Dallas/Fort Worth terminal

airspace.

152

(This airspace encompasses Dallas/Fort Worth International Airport and

Dallas/Love Field.) The study analyzed 63 hours of convective weather and flight data

for arriving aircraft. The data were collected over nine stormy days between late April

and early July 1997. During that time, 1,279 aircraft encountered storm cells 1,952 times;

about one-third of the encounters (642) resulted in deviations, and about two-thirds

(1,310) resulted in penetrations. The study did not include information from the flight

crews regarding the reasons for the actions documented by the flight data.

The data included very few encounters with NWS level 1 (very light)

precipitation because of the threshold values used to identify penetrations and deviations.

The data indicated that the number of penetrations for level 2 (light to moderate)

precipitation was about six times greater than the number of deviations. Further, the data

indicated that pilots were more likely to penetrate than deviate around level 3 (strong)

thunderstorms, but pilots were more likely to deviate around rather than penetrate level 4

(very strong) and 5 (intense) thunderstorms. Finally, the data showed very few

encounters with level 6 (extreme) thunderstorms, but all of the pilots encountering such

thunderstorms deviated around them.

In addition, the data indicated that most of the encounters within 20 to

30 kilometers (10 to 16 nm) of the destination airport resulted in penetrations. For

example, there were 918 encounters with level 3, 4, and 5 thunderstorms during the data

collection period. Of the 297 such encounters within 25 kilometers of the destination

airport, 266 (90 percent) resulted in penetrations. However, of the 611 such encounters

farther than 25 kilometers from the destination airport, 157 (26 percent) resulted in

penetrations. The study also found that, farther from the airport, pilots nearly always

deviated around intense storms and penetrated weaker storms and that, closer to the

airport, pilots mostly penetrated the storms regardless of their intensity.

Several flight-related variables were tested to determine whether they were

correlated to a pilot’s decision to penetrate thunderstorms. The tests determined that

151

Slip ratios measure the amount of rolling and skidding in tire motion. For a rolling tire,

the ratio is zero; for a skidding tire (or locked wheel), the ratio is one.

152

Rhoda, D.A. and Pawlak, M.L. 1999.

An Assessment of Thunderstorm Penetrations and Deviations

by Commercial Aircraft in the Terminal Area. Massachusetts Institute of Technology, Lincoln Laboratory,

Project Report NASA/A-2.

Factual Information 95 Aircraft Accident Report

pilots were more likely to penetrate convective weather when they were following

another aircraft, behind schedule by more than 15 minutes, or flying after dark.

According to public hearing testimony by the primary author of this study, there were no

discernible differences among air carriers regarding the propensity to penetrate or deviate

around thunderstorms or among jet versus turboprop airplanes.

1.18.3 Studies on Flight Crew Decision-Making

1.18.3.1 Safety Board Study of Flight Crew Involvement in Major Accidents

The Safety Board issued a 1994 safety study that examined the operating

environments and errors made by flight crewmembers in 37 major accidents between

1978 and 1990.

153

The Board identified 302 flight crew errors in the 37 accidents, 232 of

which were considered primary errors and 70 of which were considered secondary errors.

The 232 primary errors were grouped according to type of error. The primary

error categories and the number of errors in each category were as follows: procedural

(for example, not conducting or completing required checklists or not following

prescribed checklist procedures), 73 errors; tactical decision (for example, improper

decision-making, failing to change a course of action in response to a signal to do so, and

failing to heed warnings or alerts that suggest a change in course of actions), 51 errors;

aircraft handling, 46 errors; situational awareness (for example, controlling the airplane

at an incorrect target altitude), 19 errors; communication, 13 errors; systems operation,

13 errors; resource management, 11 errors; and navigational, 6 errors. Secondary errors

resulted from the failure of a crewmember to monitor or challenge a primary error made

by another crewmember.

The study also examined the effect of the length of time since awakening (TSA)

on the errors made by flight crewmembers. The performances of flight crews in which the

captain and the first officer had been awake an average of 13.8 and 13.4 hours,

respectively (referred to as high TSA crews), were compared with the performances of

flight crews in which the captain and the first officer had been awake an average of 5.3

and 5.2 hours, respectively (referred to as low TSA crews).

154

The Safety Board found that both the number and type of errors made by the

flight crews varied significantly according to the TSA length. Specifically, high TSA

crews made an average of 40 percent more errors than low TSA crews. Also, high TSA

crews made significantly more procedural and tactical decision errors than low TSA

crews. According to the study report, these results suggested that the degraded

performance by high TSA crews tended to involve ineffective decision-making and

procedural slips rather than a deterioration of aircraft handling skill.

153

National Transportation Safety Board. 1994.

A Review of Flightcrew-Involved Major Accidents

of U.S. Air Carriers, 1978 Through 1990. Safety Study NTSB/SS-94/01. Washington, DC.

154

The safety study’s findings related to fatigue are based on only 12 of the 37 accidents examined

in the study. There were six high TSA crews and six low TSA crews.

Factual Information 96 Aircraft Accident Report

1.18.3.2 National Aeronautics and Space Administration Study on Flight

Crew Decision Errors

Researchers at NASA’s Ames Research Center in Moffett Field, California,

conducted a study that examined the Safety Board’s findings in its 1994 safety study.

155

The purpose of the NASA study was to analyze the accident data on the most common

flight crew decision errors to determine any themes or patterns within which the errors

occurred.

The NASA researchers found that the most common decision errors occurred

when the flight crew decided to “continue with the original plan of action in the face of

cues that suggested changing the course of action.” The study stated that cues that signal

a problem are not always clear and that a decision-maker’s situation assessment may not

keep pace with conditions that deteriorate gradually. The study also stated that

individuals have a natural tendency to maintain their originally selected course of action

until there is clear and overwhelming evidence that the course of action should be

changed. Further, the study stated the following:

[A] recurring problem is that pilots are not likely to question their interpretation

of a situation even if it is in error. Ambiguous cues may permit multiple

interpretations. If this ambiguity is not recognized, the crew may be confident

that they have correctly interpreted the problem. Even if the ambiguity is

recognized, a substantial weight of evidence may be needed to change the plan

being executed.

In addition, the study noted that pilots under stress might not evaluate the

consequences of various options.

1.18.4 Technologies to Detect and Locate Downed Airplanes

Several technologies could help ATC facilities and ARFF units in detecting,

locating, and expediting emergency response efforts after aircraft accidents. These

technologies include emergency locator transmitters (ELT) and the DEVS. An ELT is a

device aboard an airplane that is intended to transmit a signal after an accident to aid

rescue personnel in locating the airplane. Airplanes operating under 14 CFR Part 121 are

not required to be equipped with an ELT, but it is a recommended practice, under

Annex 6 to the Convention on International Civil Aviation, for all airplanes to carry an

automatically activated ELT. According to AC 150/5210-19, “Driver’s Enhanced Vision

System (DEVS),” dated December 23, 1996, the purpose of the DEVS program is to

reduce ARFF response times in poor visibility conditions. The system is “aimed at the

difficult aspects of poor visibility response: locating the accident, navigating to the

accident site, and avoiding obstacles and locating people on the way to the accident site.”

155

Orasanu, J.; Martin, L.; and Davison, J.

Errors in Aviation Decision Making: Bad Decisions

or Bad Luck? NASA Ames Research Center, Moffett Field, California. Presented to the Fourth Conference

on Naturalistic Decision Making, Warrenton, Virginia, May 29-31, 1998.

Factual Information 97 Aircraft Accident Report

DEVS consists of three components: a forward-looking infrared device, a global

positioning system, and a tracking system.

The FAA’s Manager of the Airport Safety and Certification Branch, Office of

Airport Safety and Standards, stated that the most important part of DEVS is the

forward-looking infrared device because it allows a vehicle’s driver to see in almost zero

or zero visibility. The range that the system can look ahead depends mostly on the

weather conditions. According to the Branch Manager, the device is looking for heat

sources, so an airplane on fire in the rain may be harder for the device to find because the

rain would cool the plume of smoke from the fire. He added that the device would be

effective in snow but that the device’s range would be less. The Branch Manager

indicated that the FAA requires a forward-looking infrared device to be installed on all of

its new fire trucks that carry 1,500 or more gallons. There is no requirement, however, for

ARFF personnel aboard these trucks to use this device.

1.18.5 Previous Weather-Related Accidents

USAir Flight 1016, Charlotte, North Carolina, July 2, 1994

On July 2, 1994, USAir flight 1016, a Douglas DC-9-31, N954VJ, collided with

trees and a private residence after executing a missed approach to runway 18R at

Charlotte/Douglas International Airport, Charlotte, North Carolina.

156

At the time, adverse

weather conditions existed over the airport and along the flightpath. Of the 57 people

aboard the airplane, 37 passengers were killed; 2 flight attendants and 14 passengers

received serious injuries; and the 2 flight crewmembers, 1 flight attendant, and 1

passenger received minor injuries. The airplane was destroyed by impact forces and fire.

The Safety Board determined that the probable causes of this accident were

(1) the flight crew’s decision to continue an approach into severe convective activity that

was conducive to a microburst; (2) the flight crew’s failure to recognize a windshear

situation in a timely manner; (3) the flight crew’s failure to establish and maintain the

proper airplane attitude and thrust setting necessary to escape the windshear, and (4) the

lack of real-time adverse weather and windshear hazard information dissemination from

ATC, all of which led to an encounter with and failure to escape from a

microburst-induced windshear that was produced by a rapidly developing thunderstorm

located at the approach end of runway 18R.

Contributing to this accident were (1) the lack of ATC procedures that would have

required the controller to display and issue ASR-9 weather information to the pilots of

flight 1016; (2) the Charlotte tower supervisor’s failure to properly advise and ensure that

all controllers were aware of and reporting the reduction in visibility and the RVR value

information; (3) the inadequate remedial actions by USAir to ensure adherence to

156

National Transportation Safety Board. 1995.

Flight Into Terrain During Missed Approach, USAir

Flight 1016, DC-9-31, N954VJ, Charlotte/Douglas International Airport, Charlotte, North Carolina,

July 2, 1994. Aircraft Accident Report NTSB/AAR-95/03. Washington, DC. Section 1.18.7.2 discusses

the ARFF emergency response to this accident.

Factual Information 98 Aircraft Accident Report

standard operating procedures; and (4) the inadequate software logic in the airplane’s

windshear warning system that did not provide an alert upon entry to the windshear.

Delta Air Lines Flight 191, Dallas, Texas, August 2, 1985

On August 2, 1985, Delta Air Lines flight 191, a Lockheed L-1011-385-1,

N726DA, crashed while approaching to land on runway 17L at Dallas/Fort Worth

International Airport.

157

The airplane entered a microburst, which the pilot was unable to

traverse successfully. The airplane struck the ground about 6,300 feet short of the

runway, hit a car on the highway north of the runway, struck two water tanks at the

airport, and broke apart. The 3 flight crewmembers, 5 flight attendants, and 126

passengers were killed; 1 flight attendant and 14 passengers received serious injuries;

2 flight attendants and 10 passengers received minor injuries; and 2 passengers were not

injured. In addition, the driver of the car was killed, and 1 rescuer received minor injuries.

The airplane was destroyed by impact forces and a postcrash fire.

The Safety Board determined that the probable causes of this accident were the

flight crew’s decision to initiate and continue the approach into a cumulonimbus cloud,

which the crewmembers observed to contain visible lightning; the lack of specific

guidelines, procedures, and training for avoiding and escaping from low altitude

windshear; and the lack of definitive, real-time windshear hazard information. This

resulted in the airplane’s encounter at low altitude with a microburst-induced, severe

windshear from a rapidly developing thunderstorm located on the final approach course.

1.18.5.1 Previous Weather-Related Safety Recommendations

Since 1974, the Safety Board has issued at least 90 weather-related safety

recommendations to the FAA, NWS, National Oceanic and Atmospheric Administration,

and others. Three of these recommendations have relevance to the current investigation

and are detailed below.

A-95-48 and -52

On April 4, 1995, as part of its final report on the USAir flight 1016 accident, the

Safety Board issued Safety Recommendations A-95-48 and -52 to the FAA and NWS,

respectively. These recommendations asked each agency to cooperate with the other to

Reevaluate the Central Weather Service Unit program and develop procedures to

enable meteorologists to disseminate information about rapidly developing

hazardous weather conditions, such as thunderstorms and low altitude windshear,

to FAA TRACONs and tower facilities immediately upon detection.

On August 14, 1995, the FAA stated that it had reevaluated its existing

procedures to disseminate information about rapidly developing hazardous weather

157

National Transportation Safety Board. 1986.

Delta Air Lines, Inc., Lockheed L-1011-385-1,

N726DA, Dallas/Fort Worth International Airport, Texas, August 2, 1985. Aircraft Accident Report

NTSB/AAR-86/05. Washington, DC.

Factual Information 99 Aircraft Accident Report

conditions to ATC facilities and believed that the procedures were appropriate. On

July 26, 1996, the Safety Board stated that additional procedures must be developed to

efficiently use and disseminate hazardous weather information that is available from

current and planned meteorological systems for the center weather service units

(CWSU),

158

such as WSR-88D. The Board also indicated that the FAA, along with the

NWS, must make a timely, detailed, and comprehensive effort to reevaluate the CWSU

program.

On January 6, 1997, the FAA stated that it had been working with the NWS

regarding air traffic aviation weather needs associated with the performance of the

CWSUs and that it was developing a program to explore the concept of using enhanced

weather-trained FAA ATC personnel to perform all CWSU duties. On June 9, 1997, the

Safety Board acknowledged the FAA’s efforts to date but expressed concerns about its

initiative to replace NWS meteorologists at the CWSUs with enhanced weather-trained

FAA ATC personnel. On February 28, 1998, the FAA stated that NWS meteorologists

would continue to staff the CWSUs. The FAA also stated that it would work with the

NWS to standardize CWSU operations and rewrite the national CWSU order based on air

traffic requirements. On May 29, 1998, the Board recognized these efforts.

On June 5, 2000, the FAA stated that it and the NWS held a meeting in

February 2000 with most of the CWSU meteorologists-in-charge and representatives

from both agencies to review current CWSU products and procedures. The FAA stated

that the participants agreed on the need to review and update all documentation on

CWSU duties and standardize procedures and products. The FAA also stated that the

participants determined that equipment should be installed to enable the FAA to relay all

weather information, including hazardous weather conditions, to controllers without the

need for manual intervention. As a result, the FAA indicated that it was installing an

ASOS Controller Equipment–Information Display System (designed to consolidate all

weather and aeronautical information into one controller display) in eight TRACON or

associated facilities. The FAA anticipated that all eight systems would be installed by

September 2001 if funding were available.

On July 14, 2000, the Safety Board indicated that, although the FAA was taking

the actions specified in this recommendation, its work was not scheduled to be completed

until 6 1/2 years after the recommendation was issued. On August 21, 2001, the FAA

indicated that two of the eight ASOS Controller Equipment–Information Display

Systems were operational and that the installation of the remaining systems would be

completed by the fall of 2001, as long as funding was available for this project. The FAA

also indicated that it and the NWS would revise, in early 2002, FAA Order 7210.38A,

Center Weather Service Unit, to standardize CWSU operations nationwide. On

October 22, 2001, the Board reemphasized its concern about the length of time that had

passed since the recommendation was issued but recognized that the FAA appeared to be

in the final stages of completing its planned actions. Pending completion of the FAA’s

158

There are 21 CWSUs throughout the United States.

Factual Information 100 Aircraft Accident Report

planned actions, Safety Recommendation A-95-48 was classified “Open—Acceptable

Response.”

On November 17, 1997, the NWS indicated that it had several programs

underway that were relevant to the intent of Safety Recommendation A-95-52, including

programs at NASA and the FAA. On April 15, 1998, the Safety Board stated that it

wanted to review the NWS’ actions to identify future goals and responsibilities of the

CWSUs, identify any deficiencies in the CWSU program, and institute any necessary

improvements. On August 28, 2000, the NWS indicated that the Integrated Terminal

Weather System, which would produce fully automated and integrated terminal weather

information, had been developed by the FAA and would be fully implemented in 2003.

On January 5, 2001, the Board expressed its concern that the NWS appeared to have

made little progress in addressing this recommendation and requested that the NWS

provide, within 3 months, a schedule of all activities taken in response to this

recommendation. On April 5, 2001, the NWS explained that it was beyond the scope of

present CWSU operations for weather forecasters to provide continuous weather

information services to support terminal operations. The NWS further stated that any

increase in the scope of responsibilities for CWSU forecasters would require direction

from the FAA, which, according to the NWS, has “reserved the right to make the final

decision regarding any changes to the CWSU concept of operations.”

On August 7, 2001, the Safety Board acknowledged that the FAA funds and

makes final decisions regarding the CWSU program, but the Board did not believe that

the concept of operations regarding center weather support was driven solely by the FAA.

The Board indicated that the NWS and FAA should discuss the issues in Safety

Recommendation A-95-48 and -52 as part of routine discussions regarding CWSU

operations and make plans to address the issues. The Board further noted that, in the

6 years since this recommendation was issued, the NWS has provided few updates on the

status of its actions, including the schedule of all recommendation-related activities

requested in the Board’s January 2000 letter. Pending receipt of this information and

completion of the recommended action, Safety Recommendation A-95-52 was classified

“Open—Unacceptable Response.”

A-90-84

On June 18, 1990, the Safety Board issued a safety recommendation letter to the

FAA about its windshear program. The letter indicated that the program had generally

been well managed and productive and that elements of the program had been well

coordinated within the Government and industry. However, the Board determined that

additional actions were required by the FAA to ensure maximum protection against

low-altitude windshear. As a result, the Board issued Safety Recommendation A-90-84,

which asked the FAA to

Develop the modular windshear detection enhancement to the Airport

Surveillance Radar (ASR-9) and implement the enhanced ASR-9 at air carrier

airports not scheduled for a Terminal Doppler Weather Radar (TDWR) system or

as an interim measure at airports scheduled for late TDWR installations.

Factual Information 101 Aircraft Accident Report

On February 21, 1995, the FAA stated that it was proceeding with the

development of an ASR Weather Systems Processor (WSP). The FAA explained that the

WSP is a “cost-effective, add-on, passive radar receiver and weather processor on a host

ASR” and that the WSP adds “full TDWR-like performance to existing ASR-9 search

radars” by providing windshear and microburst detection, storm cell motion, and

prediction of gust fronts for use by air traffic controllers and control tower operators. The

FAA further explained that the WSP would consolidate other weather data from the

LLWAS or ASOS into a single display. The FAA indicated that WSP component testing

would conclude at the end of 1995 and that the system would be integrated by

September 2002 at 33 airports that have medium air traffic density, are subject to

windshear, and are not qualified to receive TDWR. On June 20, 1995, the Safety Board

indicated that it was pleased with the FAA’s plan to develop the WSP but urged the FAA

to expedite its efforts to provide windshear protection at airports that are not served by

TDWR.

The FAA’s letters to the Safety Board between August 1996 and December 1999

detailed the FAA’s progress in developing and procuring the WSP systems. The Board’s

letters to the FAA between December 1996 and March 2000 acknowledged the FAA’s

efforts to date and restated the Board’s concern about the hazard of low-altitude

windshear at airports before the WSP systems are delivered.

On July 24, 2000, the FAA stated that it was continuing its efforts to complete the

development of the production version of the WSP. The FAA indicated that two WSP

prototypes were operating continuously at Albuquerque, New Mexico, and Austin,

Texas. The FAA also stated that three limited-production WSP systems were installed at

the FAA’s Academy, the FAA’s William J. Hughes Technical Center, and Albuquerque

in January, May, and June 2000, respectively; a fourth limited-production system was in

the process of being installed at Austin; a fifth limited-production system would remain

with the contractor to serve as a depot repair system; and a sixth limited-production

system was scheduled to be delivered to Norfolk, Virginia, in September 2000. In

addition, the FAA stated that operational tests and evaluations were to be completed in

December 2000 and that full production system deliveries would occur between March

2001 and July 2002 at a rate of two systems per month. On September 14, 2000, the

Safety Board expressed its concern about the WSP project’s slow pace and pointed out

that the project’s anticipated completion date of mid-2002 would be 12 years after this

recommendation was first issued.

On March 16, 2001, the FAA indicated that, between July and December 2000, it

completed initial operational testing and evaluation of the WSP and that all six

limited-production systems were continuously operating. The FAA also indicated that, in

November 2000, it procured an additional system to be installed in at the FAA’s program

support facility in Oklahoma City, Oklahoma. In addition, the FAA stated that it had

conducted the first operators’ training course in November 2000 and the first

maintenance training course in December 2000. Further, the FAA stated that follow-on

operational testing and evaluation was scheduled to be completed in April 2001 and that

Factual Information 102 Aircraft Accident Report

initial full-production deliveries would begin then. On May 16, 2001, the Safety Board

stated that it would appreciate receiving updates on the delivery schedule.

On July 2, 2001, the FAA stated that, in April 2001, it held a dedication ceremony

to commemorate the installation of the first three key site operational WSP systems in

Austin, Albuquerque, and Norfolk; the delivery of the first of 32 full-production units to

the FAA’s program support facility in Oklahoma City; and the initiation of WSP

production. The FAA indicated that it had updated WSP software to improve system

maintainability and that WSP deliveries would begin by August 2001. On August 27,

2001, the Safety Board noted its concern about the amount of time that has passed since

this recommendation was first issued but acknowledged that the FAA was working to

have the WSP fully deployed soon. Pending full deployment and operation of all WSP

systems by July 2002, Safety Recommendation A-90-84 was classified “Open—

Acceptable Response.”

1.18.6 Previous Fatigue-Related Accidents

American Eagle Flight 4925, Jamaica, New York, May 8, 1999

On May 8, 1999, American Eagle flight 4925, a Saab 340B, N232AE, overran the

end of runway 4R during landing at John F. Kennedy International Airport, Jamaica, New

York.

159

The captain flew the approach with excessive altitude, airspeed, and rate of

descent and remained above the glideslope, and the first officer failed to make required

callouts, including a missed approach callout. As previously discussed in section 1.16.5,

an EMAS installed at the end of the runway stopped the airplane. Of the 30 persons on

board, 1 was seriously injured, and 29 were not injured. The airplane received substantial

damage.

During postaccident interviews, both pilots stated that they were fatigued. They

had been working a continuous-duty overnight schedule (that is, a trip sequence that is

flown during the late night hours and the early morning hours of the next day in which

the break between flights is not sufficient to qualify as a duty rest period, even if the

flight crew is provided with a hotel room for rest). On May 7, 1999, the flight

crewmembers awoke during the morning hours, did not sleep during the day, and

reported for duty about 2200 for a flight scheduled about 2246. The flight was delayed

but arrived at the destination airport—Baltimore-Washington International Airport—

about 0025 on May 8. The pilots went to sleep about 0130 and awoke about 0445 for the

accident flight, which was scheduled to depart about 0610.

The Safety Board determined that the probable cause of this accident was the

captain’s failure to perform a missed approach, as required by company procedures.

Contributing to the cause of this accident were the captain’s improper in-flight decisions

159

The description for this accident, NYC99FA110, can be found on the Safety Board’s Web

site at <http://www.ntsb.gov>.

Factual Information 103 Aircraft Accident Report

and failure to comply with FAA regulations and company procedures, inadequate crew

coordination, and fatigue.

Korean Air Flight 801, Nimitz Hill, Guam, August 6, 1997

On August 6, 1997, Korean Air flight 801, a Boeing 747-300, Korean registration

HL7468, crashed at Nimitz Hill, Guam.

160

The airplane had been cleared to land on

runway 6L at Guam International Airport, Agana, Guam, and crashed into high terrain

about 3 miles southwest of the airport. Of the 254 persons on board, 228 were killed

(including the three flight crewmembers), and 23 passengers and 3 flight attendants

survived the accident with serious injuries. The airplane was destroyed by impact forces

and a postcrash fire.

The accident happened after midnight in the flight crew’s home time zone (0142

Guam local time). According to the final report on this accident, research has indicated

that this time of day is often associated with degraded alertness and performance and a

higher probability of errors and accidents. Also, the arrival time for flight 801 was several

hours after the captain’s (the flying pilot) normal bedtime. At the time of the accident, the

captain had been awake for 11 hours,

161

and the CVR recorded unsolicited comments

made by the captain related to fatigue. The Safety Board concluded that the captain was

fatigued, which degraded his performance and contributed to his failure to properly

execute the approach.

The Safety Board determined that the probable cause of this accident was the

captain’s failure to adequately brief and execute the nonprecision approach and the flight

crew’s failure to effectively monitor and cross-check the captain’s execution of the

approach. Contributing to these failures were the captain’s fatigue and Korean Air’s

inadequate flight crew training. Contributing to the accident was the FAA’s intentional

inhibition of the minimum safe altitude warning system at Guam and the agency’s failure

to adequately manage the system.

1.18.6.1 Previous Fatigue-Related Safety Recommendation

Since 1989, the Safety Board has issued at least 70 safety recommendations

related to fatigue for all modes of transportation. In addition, human fatigue in transport

operations has been included in the Board’s annual list of Most Wanted Transportation

Safety Improvements since the list’s inception in September 1990. One recent aviation

fatigue-related recommendation is detailed below.

160

National Transportation Safety Board. 2000.

Controlled Flight Into Terrain. Korean Air Flight 801,

Boeing 747-300, HL7468, Nimitz Hill, Guam, August 6, 1997.

Aircraft Accident Report

NTSB/AAR-2000/01. Washington, DC.

161

The captain’s wife indicated that, on August 5, 1997, he woke up about 0600 and took a

nap between 1100 and 1340 (Seoul local time).

Factual Information 104 Aircraft Accident Report

A-99-45

In May 1999, the Safety Board adopted a safety report that evaluated the

U.S. Department of Transportation’s (DOT) efforts in the 1990s to address operator

fatigue.

162

In its report, the Board noted that in 1989 it issued three recommendations to

the DOT addressing needed research, education, and revisions to hours-of-service

regulations.

163

The Board stated that, even though the DOT and modal administrations had

responded positively to the recommendations addressing research and education, little

action had occurred with respect to revising the hours-of-service regulations.

The safety report discussed the activities and efforts by the DOT and the modal

administrations to address operator fatigue and the resulting progress that has been made

over the past 10 years to implement the actions called for in the Safety Board’s

fatigue-related recommendations. The report also provided background information on

current hours-of-service regulations, fatigue, and the effects of fatigue on transportation

safety. As a result of its findings, the Safety Board issued Safety Recommendation

A-99-45, which asked the FAA to

Establish within 2 years scientifically based hours-of-service regulations that set

limits on hours of service, provide predictable work and rest schedules, and

consider circadian rhythms and human sleep and rest requirements.

On July 15, 1999, the FAA stated that, on December 11, 1995, it issued Notice of

Proposed Rulemaking (NPRM) 95-18, “Flight Crewmember Duty Period Limitations,

Flight Time Limitations and Rest Requirements.” The NPRM proposed amending

existing regulations to establish one set of duty period limitations, flight time limitations,

and rest requirements for flight crewmembers involved in air transportation. The FAA

stated that the NPRM considered scientific data from studies conducted by NASA

relating to flight crewmember duty periods, flight times, and rest and that Safety

Recommendation A-99-45 would be included in this rulemaking project. The FAA

indicated that its Aviation Rulemaking Advisory Committee was tasked to review reserve

162

National Transportation Safety Board. 1999.

Evaluation of U.S. Department of Transportation

Efforts in the 1990s to Address Operator Fatigue. Safety Report NTSB/SR-99/01. Washington, DC.

163

Safety Recommendation I-89-1 asked the DOT to expedite a coordinated research program

on the effects of fatigue, sleepiness, sleep disorders, and circadian factors on transportation system

safety. Safety Recommendation I-89-2 asked the DOT to develop and disseminate educational material

for transportation industry personnel and management regarding shift work; work and rest schedules;

and proper regimens of health, diet, and rest. Safety Recommendation I-89-3 asked the DOT to

review and upgrade regulations governing hours of service for all transportation modes to ensure

that they are consistent and that they incorporate the results of the latest research on fatigue and

sleep issues. Safety Recommendations I-89-1 and -2 were classified “Closed—Acceptable Action”

on July 19, 1996, and May 25, 2001, respectively. Safety Recommendation I-89-3 was classified

“Closed—Unacceptable Action/Superceded” with the issuance of Safety Recommendation I-99-1, which

asked the DOT to “require the modal administrations to modify the appropriate Codes of Federal

Regulations to establish scientifically based hours-of-service regulations that set limits on hours of

service, provide predictable work and rest schedules, and consider circadian rhythms and human

sleep and rest requirements. Seek Congressional authority, if necessary, for the modal administrations

to establish these regulations.” The Safety Board classified Safety Recommendation I-99-1 “Open—

Acceptable Response” on December 4, 2000.

Factual Information 105 Aircraft Accident Report

issues related to the NPRM but was unable to agree on a recommendation. The FAA

further indicated that it was conducting a risk assessment to determine the probability of

preventing future incidents related to fatigue and did not know when a supplemental

NPRM would be issued. In addition, the FAA stated that, on June 15, 1999, it published a

Notice of Intent in the Federal Register to indicate the agency’s intent to enforce

regulations concerning flight time limitations and rest requirements.

On January 3, 2000, the Safety Board stated that it was pleased that the FAA

would include this safety recommendation in any future action regarding NPRM 95-18

and that the FAA was committed to enforcing its existing regulations. However, the

Board indicated that it was disappointed that, even though the NPRM was issued over

4 years earlier, the existing regulations had not been upgraded. Also, the Board indicated

that it wanted to understand the FAA’s rationale for a supplemental NPRM. In addition,

at an October 7, 1999, meeting to discuss aviation issues on the Board’s list of Most

Wanted Transportation Safety Improvements, FAA representatives said that the FAA

would not be able to meet the specified 2-year timeframe for a new rule.

On December 5, 2000, the FAA stated that it planned to issue a supplemental

NPRM that would address the technical and operational concerns that were raised during

the NPRM comment period. According to the FAA, the supplemental NPRM would

prescribe “a maximum duty period linked to a maximum flight time restriction that is

associated with a minimum rest period based on the number of pilots.” Also, the FAA

indicated that the supplemental NPRM, which was expected to be issued by spring 2001,

would address the issue of fatigue “concretely” and give the airlines the flexibility they

need to operate. The FAA further indicated that it would have to issue a supplemental

NPRM rather than a final rule because of the numerous comments that were received as a

result of NPRM 95-18.

On April 26, 2001, the Safety Board indicated that it was frustrated by the FAA’s

lack of progress concerning this safety issue. The Board stated that, in the 5 years since

the issuance of NPRM 95-18 and the 1 1/2 years since the need for a supplemental

NPRM was first communicated, the FAA has not taken action. The Board urged the FAA

to expeditiously complete action to develop and implement flight and duty regulations.

Pending the issuance of the supplemental NPRM and expeditious action to implement a

final rule, Safety Recommendation A-99-45 was classified “Open—Unacceptable

Response.”

1.18.6.2 Additional Fatigue Information

On April 28, 2000, a fatigue researcher from the University of Pennsylvania

School of Medicine was interviewed as part of the investigation of the flight 1420

accident.

164

The researcher stated that the term “fatigue” generally refers to an individual’s

164

The fatigue researcher is a Professor of Psychology in the Department of Psychiatry and the

Chief of the Division of Sleep and Chronobiology at University of Pennsylvania School of Medicine.

He also directs the Unit for Experimental Psychology, a laboratory that conducts research studies

related to alertness and performance capabilities in individuals who work long or unusual hours.

Factual Information 106 Aircraft Accident Report

difficulty maintaining a certain level of performance as a function of time. The researcher

described how fatigue could affect human performance. For example, individuals who are

fatigued tend to have decreased vigilance in tasks that require monitoring or detecting

signals and increased short-term and working memory errors. Also, individuals who are

fatigued tend to keep trying one specific solution to a situation even if it no longer works,

take more risks (that is, cut corners and accept lower standards), and pay less attention or

not be aware of peripheral events.

The fatigue researcher reviewed factual documents related to the flight 1420

accident. He stated that, because the accident happened a couple of hours after the

captain’s usual bedtime, his circadian system was already in a downward phase. The

professor indicated that the captain could have been feeling “pretty good” and “alert”

when he left Dallas but that he could have experienced “an increased physiologic fatigue

level” as he began to approach Little Rock. The researcher further indicated that the

captain’s total time awake (over 16 hours) might have made him more vulnerable to

making errors. In addition, he stated that the captain’s prolonged wakefulness and the

time of the accident (compared with his usual bedtime) made it “highly likely” that the

captain was fatigued at the time of the crash.

1.18.7 Other Previous Related Accidents

1.18.7.1 Accidents Involving Rescue Efforts

The Safety Board has investigated or learned of accidents in which the survival of

airplane occupants was aided by or was dependent on the ability of rescue personnel to

enter an airplane to perform interior firefighting or remove a person to safety.

165

A brief

discussion of three such accidents follows.

British Airtours Charter Flight, Manchester, England, August 22, 1985

On August 22, 1985, a British Airtours charter Boeing 737-226, G-BGJL, was

taking off from Manchester International Airport, England, when an uncontained failure

in the left engine occurred.

166

The wing fuel tank was punctured, which led to a fuel-fed

fire. The flight crew abandoned the takeoff immediately and ordered the evacuation of

the passengers through the exits on the right side of the airplane. By the time the airplane

had come to a stop, fire had penetrated the hull, and smoke had entered the cabin. Of the

6 crewmembers and 131 passengers aboard the airplane, 55 were killed, 15 were

165

The Safety Board had previously commented on the need for additional ARFF resources.

Specifically, in an August 1996 letter to the FAA, the Board commented on its review of the FAA’s

draft report,

Aircraft Rescue and Firefighting Services—Mission Response Study

. The Board stated

that the mission set forth in 14 CFR Part 139 for ARFF personnel to “provide an escape path

from a burning airplane” was no longer sufficient and that it supported a broader mission in which

adequate ARFF resources would be available “to rapidly extinguish interior fires and extricate aircraft

occupants from such interior fires.”

166

For more information, see

Air Accidents Investigation Branch. Report on the Accident to Boeing

737-236 Series 1, G-BGJL, at Manchester International Airport on 22 August 1985, Aircraft Accident

Report 8/88. Farnborough, England.

Factual Information 107 Aircraft Accident Report

seriously injured, and 67 received minor injuries or were not injured. The airplane was

destroyed. About 5 1/2 minutes after the airplane had stopped, a 14-year-old boy was

pulled out of the right overwing exit by a firefighter. He was the last evacuee to survive

the accident.

Horizon Air Flight 2638, Seattle, Washington, April 15, 1988

On April 15, 1988, Horizon Air flight 2638, a DeHavilland DHC-8, N819PH, had

departed from Seattle-Tacoma International Airport when the right engine lost power.

167

The flight crew decided to return to the airport for a precautionary landing. As the

landing gear was lowered, a fire broke out in the right engine nacelle. The engine was

then shut down. After the airplane landed, almost all of the directional control and

braking capability had been lost. The airplane departed the paved surface of the runway

and struck several objects on a paved ramp area. Of the 40 airplane occupants, 4 received

serious injuries, and 36 received minor or no injuries. The airplane was destroyed by the

fire and impact. Firefighters entered the cabin and extricated two passengers who were

trapped by wreckage—one of whom had a lacerated aorta—and assisted other airplane

occupants, including the captain and the first officer, off the airplane. The Safety Board

determined that the firefighters’ rescue of the passenger with a lacerated aorta (which

occurred before the fire was extinguished) was “instrumental” in saving his life.

USAir Flight 1493/Skywest Flight 5569, Los Angeles, California,

February 1, 1991

On February 1, 1991, USAir flight 1493, a Boeing 737-300, N388US, and

Skywest Flight 5569, a Fairchild Metroliner SA-227-AC, N683AV, collided on

runway 24L at Los Angeles International Airport, California.

168

The USAir airplane was

landing, and the Skywest airplane was positioned on the same runway awaiting clearance

for takeoff. There was an explosion and fire upon impact, and the two airplanes slid into

an unoccupied fire station. Of the 95 airplane occupants aboard the USAir airplane, 22

were killed, 13 were seriously injured, 15 received minor injuries, and 37 were uninjured.

All of the 12 occupants aboard the Skywest airplane were killed. Both airplanes were

destroyed by impact forces and postcrash fire.

Firefighters assisted some passengers in evacuating the 737. One of the

firefighters removed the first officer from the wreckage and, with the help of another

firefighter, moved him to a safe area. Also, firefighters brought hand lines into the

airplane, including one firefighter who brought a hand line to the cockpit to protect the

captain from fire (he was pinned in the wreckage and appeared lifeless).

169

The firefighters

167

For more information, see National Transportation Safety Board. Horizon Air, Inc., DeHavilland

DHC-8, Seattle-Tacoma International Airport, Seattle, Washington, April 15, 1988. Aircraft Accident

Report NTSB/AAR/89-02. Washington, DC.

168

For more information, see National Transportation Safety Board.

Runway Collision of USAir

Flight 1493, Boeing 737, and Skywest Flight 5569, Fairchild Metroliner, Los Angeles International

Airport, Los Angeles, California, February 1, 1991. Aircraft Accident Report NTSB/AAR/91-08.

Washington, DC.

169

The first officer survived the accident. The captain died as a result of traumatic injury to the head.

Factual Information 108 Aircraft Accident Report

remained in the cabin until the interior fire was extinguished. The Safety Board’s final

report stated that “the rapid availability of adequate numbers of ARFF-trained

firefighters…allowed ARFF personnel to implement an interior fire attack immediately.”

1.18.7.2 Accidents Involving Delayed Emergency Response

The Safety Board has investigated accidents in which the emergency response

was delayed because the crash was not detected or the wreckage was not located in a

timely manner. A brief discussion of three such accidents follows.

Northwest Airlines Flights 1482 and 299, Detroit, Michigan, December 3, 1990

On December 3, 1990, Northwest Airlines flight 1482, a Douglas DC-9-14,

N3313L, and Northwest flight 299, a 727, N278US, collided near the intersection of two

runways in dense fog at Detroit Metropolitan/Wayne County Airport, Romulus,

Michigan.

170

Of the 40 passengers and 4 crewmembers aboard the DC-9, 8 were killed, 10

received serious injuries, and 26 received minor or no injuries. None of the 146 passengers

and 8 crewmembers aboard the 727 were injured. The DC-9 was destroyed, and the 727

was substantially damaged. In its final report on the accident, the Safety Board stated that

the low visibility and the lack of immediate, accurate location information available in the

control tower resulted in the fire department being unable to locate the DC-9 until about 5

minutes after the collision. In addition, by the time the first ARFF vehicle reached the

DC-9, its cabin was fully involved with fire. That ARFF vehicle, however, had no heavy

fire-fighting equipment, so the driver had to radio the other units (which were responding

to the 727) and request that they respond to the DC-9.

USAir Flight 1016, Charlotte, North Carolina, July 2, 1994

In the July 2, 1994, USAir flight 1016 accident in Charlotte, North Carolina,

(previously discussed in section 1.18.5), the control tower’s initial notification to the

ARFF station was made about 2 minutes after the accident occurred. The initial

notification, however, did not identify any specific location of the downed airplane

because of the restricted visibility conditions. The fire trucks began to traverse the

airport, in heavy rain, searching for the airplane. The controller then saw a large area of

smoke and described the location to the ARFF units. The first ARFF unit arrived at the

accident scene about 8 minutes after the accident.

170

For more information, see National Transportation Safety Board. 1991.

Runway Incursion and

Collision, Northwest Airlines, Inc., Flights 1482 and 299, Detroit Metropolitan/Wayne County Airport,

Romulus, Michigan, December 3, 1990.

Aircraft Accident Report NTSB/AAR/91-05. Washington, DC.

Factual Information 109 Aircraft Accident Report

American Airlines Flight 1340, Chicago, Illinois, February 9, 1998

On February 9, 1998, American Airlines flight 1340, a Boeing 727-223,

N845AA, crashed 180 feet short of the runway 14R threshold at Chicago O’Hare

International Airport.

171

During the ground impact sequence, the main landing gear and aft

air stairs separated from the airplane, and the nose gear folded back into the avionics

compartment. The airplane came to rest on its fuselage 2,245 feet from the initial impact

point. Of the 3 flight crewmembers, 3 flight attendants, and 116 passengers aboard the

airplane, 1 flight attendant and 22 passengers received minor injuries. The airplane

received substantial damage. A City of Chicago Department of Aviation vehicle radioed

the control tower about 5 minutes after the accident had occurred, informing the local

controller that an airplane was down and that debris was on the runway. Until that point,

the controller was not aware that flight 1340 had crashed.

172

Fire department personnel

arrived at the accident scene 2 minutes after receiving notification, which was about

7 minutes after the accident.

1.18.8 Other Previous Related Safety Recommendations

1.18.8.1 Aircraft Operations in the Airport Environment

A-94-211

On April 27, 1994, Action Air Charters flight 990, a Piper PA-31-350 Navajo

Chieftain, N990RA, crashed into a blast fence at the end of runway 6 at Sikorsky

Memorial Airport, Stratford, Connecticut.

173

Nine airplane occupants were killed, and one

was seriously injured.

174

The airplane was destroyed by impact forces and a postcrash fire.

As a result of this accident, the Safety Board issued Safety Recommendation A-94-211 on

December 13, 1994, which asked the FAA to

Inspect all 14 CFR Part 139 certificated airports for adequate runway safety areas

and nonfrangible objects, such as blast fences, and require that substandard

171

The description for this accident, DCA98MA023, can be found on the Safety Board’s Web

site at <http://www.ntsb.gov>.

172

Before he became aware of the crash, the controller had been clearing arriving aircraft for

landing on the same runway. One airplane had completed a landing with the debris on the runway,

and a second airplane had touched down momentarily on the runway while executing a go-around

maneuver.

173

National Transportation Safety Board. 1994.

Impact With Blast Fence Upon Landing Rollout,

Action Air Charters Flight 990, Piper PA-31-350, N990RA, Stratford, Connecticut, April 27, 1994.

Aircraft Accident Report NTSB/AAR-94/08. Washington, DC.

174

The blast fence was located 342 feet northeast of the runway 24 displaced threshold. The

departure end of runway 6 (the approach end of runway 24) had no runway safety area. This runway

was exempt from the standards in the FAA’s June 5, 1991, version of AC 150/5300-13 (see section 1.10.1);

otherwise, a runway safety area of 800 feet would have been required. In its final report on this

accident, the Safety Board concluded that the destruction of the airplane and the resulting occupant

injuries were a direct result of the airplane’s collision with the blast fence but that the crash forces

resulting from the collision were survivable (the nine fatalities resulted from smoke inhalation and/or

thermal injuries).

Factual Information 110 Aircraft Accident Report

runway safety areas be upgraded to AC 150/5300-13 minimum standards

wherever possible.

On October 15, 1997, the FAA indicated that it had conducted a study and found

that 58 percent of the runways at 14 CFR Part 139 certificated airports have runway

safety areas that meet this standard, 25 percent have safety areas that do not meet the

standard but could with feasible improvements, and 17 percent have safety areas that

could not be feasibly improved to meet the standard. The FAA stated that runway safety

area improvement projects would be scheduled as part of overall runway improvement

projects because of the associated cost and infrequency of aircraft overruns and

undershoots. On February 10, 1999, the Safety Board classified Safety Recommendation

A-94-211 “Closed—Unacceptable Action” because runway safety area upgrades might

be delayed by many years, allowing nonstandard conditions to continue.

A-84-36

The Safety Board conducted a special investigation in response to three accidents

in 1982 that involved long-standing concerns regarding the safety of aircraft operations

in the airport environment.

175

On April 16, 1984, the Board issued Safety

Recommendation A-84-36, which asked the FAA to

Initiate research and development activities to establish the feasibility of

submerged low-impact resistance support structures for airport facilities and

promulgate a design standard if such structures are found to be practical.

On October 30, 1996, the FAA stated that it and the National Institute of

Standards and Technology studied the feasibility of submerged low-impact structures and

concluded that, with the technology currently available, any structure deemed frangible

would most likely be destroyed by wave motion produced by small storms. On the basis

of the FAA’s actions, the Safety Board classified Safety Recommendation A-84-36

“Closed—Acceptable Action” on December 20, 1996.

1.18.8.2 Stabilized Approach Guidance

A-97-85

On October 19, 1996, Delta Air Lines flight 554, an MD-88, N914DL, struck the

approach lighting structure at the end of the runway deck during an approach to runway

13 at LaGuardia Airport, Flushing New York.

176

The 2 flight crewmembers, the 3 flight

attendants, and 55 passengers were not injured; 3 passengers received minor injuries. The

airplane was substantially damaged. In its final report on this accident, the Safety Board

175

These accidents were Air Florida flight 90 near Washington National Airport, January 13,

1982; World Airways flight 30H at Boston Logan International Airport, January 23, 1982; and Pan

American World Airways flight 759 near New Orleans International Airport, July 9, 1982.

176

National Transportation Safety Board. 1997.

Descent Below Visual Glidepath and Collision

With Terrain, Delta Air Lines Flight 554, McDonnell Douglas MD-88, N914DL, LaGuardia Airport,

New York, October 19, 1996

. Aircraft Accident Report NTSB/AAR-97/03. Washington, DC.

Factual Information 111 Aircraft Accident Report

stated that Delta’s manuals did not specify operational criteria for a stabilized approach

or contain procedural guidance for pilots to follow if an approach became unstabilized.

The Board concluded that the lack of guidance on the stabilized approach concept did not

contribute to the accident. However, the Board was concerned that, if Delta’s manuals at

the time of the accident contained inadequate information on stabilized approaches, other

air carriers’ guidance might also be inadequate. As a result, the Board issued Safety

Recommendation A-97-85, which asked the FAA to

Require all 14 CFR Part 121 and 135 operators to review and revise their

company operations manuals to more clearly delineate flight crewmember (pilot

flying/pilot not flying) duties and responsibilities for various phases of flight and

more clearly define terms that are critical for safety-of-flight decision-making,

such as “stabilized approach.”

On June 26, 1998, the FAA stated that it issued Flight Standards Handbook

Bulletin for Air Transportation 98-22, “Stabilized Approaches,” on May 26, 1998. The

bulletin directed 14 CFR Part 121 and 135 POIs to review operators’ training and

operations manuals to ensure that the stabilized approach concept was addressed. The

POIs were also to ensure that the manuals contained the minimum requirements for a

stabilized approach, the immediate actions that needed to be taken if the stabilized

approach conditions were not met, and the flying and nonflying pilots’ responsibilities

during the approach phase of flight. On the basis of the FAA’s actions, the Safety Board

classified Safety Recommendation A-97-85 “Closed—Acceptable Action” on

November 20, 1998.

1.18.8.3 Manufacturer’s Operating Information

A-98-102

On January 9, 1997, Comair flight 3272, an Embraer EMB-120RT, N265CA,

crashed during a rapid descent after an uncommanded roll excursion near Monroe,

Michigan.

177

The 2 flight crewmembers, 1 flight attendant, and 26 passengers were killed,

and the airplane was destroyed by impact forces and a postcrash fire. In its investigation of

this accident, the Safety Board concluded that the FAA’s policy of allowing air carriers to

elect not to adopt airplane flight manual operational procedures without clear written

justification could result in air carriers using procedures that may not reflect the safest

operating practices. As a result, the Board issued Safety Recommendation A-98-102,

which asked the FAA to

Require air carriers to adopt the operating procedures contained in the

manufacturer’s airplane flight manual and subsequent approved revisions or

provide written justification that an equivalent safety level results from an

alternate procedure.

177

National Transportation Safety Board. 1998.

In-flight Icing Encounter and Uncontrolled Collision

With Terrain, Comair Flight 3272, Embraer EMB-120RT, N265CA, Monroe, Michigan, January 9,

1997.

Aircraft Accident Report NTSB/AAR-98/04. Washington, DC.

Factual Information 112 Aircraft Accident Report

On September 16, 1999, the FAA stated that, on May 28, 1999, it issued Joint

Flight Standards Handbook Bulletin for Air Transportation, Airworthiness, and General

Aviation, Flight Standards Policy—Company Operating Manuals and Company Training

Program Revisions for Compliance. According to the FAA, the bulletin directed that

POIs encourage their operators to have a reliable delivery system in place for flight

manual revisions, which ensures that the operators receive the revisions within

30 calendar days of approval, and develop an action plan to notify, in writing, respective

POIs of new flight manual revisions within 15 days after receipt. The FAA also stated

that it was considering a regulatory change that would require the bulletin’s safety policy

to be mandatory.

On April 11, 2000, the Safety Board noted that the bulletin, although not

regulatory in nature, was a positive step toward meeting the goals of the

recommendation. On July 7, 2000, the FAA stated that it had initiated an NPRM

proposing to revise 14 CFR Part 121, Subparts N and O, and that the policy included in

the handbook bulletin would be reflected in the NPRM. On January 12, 2001, the Board

acknowledged the FAA’s actions and stated that, pending the issuance of the NPRM and

the implementation of the proposed regulation, Safety Recommendation A-98-102 was

classified “Open—Acceptable Response.” On August 2, 2001, the FAA stated that it was

continuing to develop the NPRM.

113 Aircraft Accident Report

2. Analysis

2.1 General

The captain and the first officer of American Airlines flight 1420 were properly

certificated and qualified under Federal and company requirements. No evidence

indicated any preexisting medical or behavioral conditions that might have adversely

affected the flight crew’s performance during the accident flight.

The accident airplane was properly certified, equipped, and maintained in

accordance with Federal regulations and approved company procedures. No evidence

indicated preexisting engine, system, or structural failures.

This analysis focuses primarily on the flight crew’s performance and the

airplane’s spoiler system. The flight crew’s performance on approach to the airport is

examined during three segments of the approach and in the context of the weather

information and cues that were available. The flight crew’s and the airplane’s

performance during the landing and overrun sequences are also examined. The analysis

also addresses the roles of situational stress and fatigue in the accident sequence;

meteorological support, including ATC services; emergency response efforts; airport

issues; and American and FAA oversight.

2.2 Accident Scenario

2.2.1 The Approach

The flight crew had multiple sources of information indicating that thunderstorms

might become a factor during the approach. The preflight weather package (discussed in

section 1.7.3) included two NWS weather advisories for severe thunderstorms and a

company SIGMEC [significant meteorological condition] for a widely scattered area of

thunderstorms along the planned route. Also, the dispatcher’s 2254 message via the

aircraft communication addressing and reporting system informed the crew of the lines of

thunderstorms to the left and right of the planned route and suggested that the crew

expedite the arrival to the airport “to beat the thunderstorms.” In addition, NWS

Convective SIGMET [significant meteorological information] 15C, received about 2304,

warned of a line of severe thunderstorms moving southeast through Arkansas with hail up

to 2 inches in diameter and the possibility of wind gusts to 70 knots. Finally, automatic

terminal information service (ATIS) information Romeo, which was current beginning

about 2326, indicated that a thunderstorm with frequent lightning was located west

through northwest of the airport and moving northeast.

Analysis 114 Aircraft Accident Report

The flight crewmembers also had weather information from the airborne weather

radar and their view outside the cockpit. Statements from the CVR indicated that the

flight crew had discussed the weather and the need to expedite the approach. For

example, at 2325:47, the captain said, “we got to get over there quick.” The first officer

then stated, “I don’t like that...that’s lightning,” to which the captain replied, “sure is.” At

2328:30, the captain repeated, “we gotta get there quick.” At 2329:55, the first officer

stated, “I say we get down as soon as we can.”

The CVR also indicated that the crew had the city of Little Rock and the airport

area in sight by at 2326:59. Further, the CVR recorded the captain’s announcement to the

passengers, beginning at 2327:31, that “quite a light show” was to the left of their course

and that they would be passing the lightning on the way to Little Rock.

2.2.1.1 Descent Into the Terminal Area

At 2334:11, the local controller told the flight crew, upon initial contact, that a

thunderstorm located northwest of the airport was “moving through the area now” and

that the winds were from 280º at 28 knots gusting to 44 knots. Even though the flight

crewmembers had previously discussed the need to expedite the approach because of the

weather, the CVR indicated that they had not discussed the possibility that the

thunderstorm might reach the airport before the flight landed. In a postaccident interview,

the first officer stated that, during the descent, the weather appeared to be about 15 miles

away from the airport and that he and the captain thought that there was “some time” to

make the approach. At 2339:12, the first officer told the controller, “that storm is moving

this way like your radar says it is but a little bit farther off than you thought.” At that

point, flight 1420 was about 11 miles south of the airport. Figure 18 shows a map of the

weather conditions at 2339:12 and flight 1420’s location. After receiving the controller’s

next wind report (330º at 11 knots), the first officer indicated that the winds were “a little

bit better” than they had been earlier.

Weather data obtained after the accident depicted the weather conditions in the

area shortly before the time of the accident. For example, airport weather observation

data showed that heavy rain—defined by the NWS as 0.03 inch of rain within

6 minutes—had begun falling at the airport by about 2338, 12 minutes before the

accident. In addition, the National Lightning Detection Network detected

903 cloud-to-ground lightning strikes within 20 miles of the airport in the 15 minutes

before the accident and 46 strikes within 5 miles of the airport in the 5 minutes before the

accident; most of the strikes were located northeast, west, and southwest of the airport

and along a line that was parallel to (and to the west of) flight 1420’s final approach path.

American’s policy did not prohibit flight crews from continuing an approach with

thunderstorms in the terminal area as long as the crews ensured that their intended route

was clear of the thunderstorms. During the descent into the terminal area, there was no

evidence to indicate that the route was not clear. In public hearing testimony, the first

officer stated that, during the descent, the weather was to the left and moving off to the

right and that the airport looked clear. Thus, the Safety Board concludes that, during the

Analysis 115 Aircraft Accident Report

descent into the terminal area, the flight crewmembers could have reasonably believed

that they could reach the airport before the thunderstorm.

Figure 18. Weather and Flight Information for 2339:12

2.2.1.2 Maneuvering to the Airport for Final Approach

The flight crewmembers had been previously told by the controller to expect an

instrument landing system (ILS) approach to runway 22L. At 2339:45, the local

controller broadcast the first of two windshear alerts, reporting that the centerfield wind

was 340º at 10 knots.

178

At this point, the airplane was about 8 miles from the airport.

Afterward, the flight crew requested a change in runways from 22L to 4R because the

winds had shifted to the northwest. The use of runway 22L could have resulted in a

tailwind at the time of landing, so runway 4R was a more appropriate choice because of

the expectation of a headwind. (NWS radar data indicated that the leading edge of a line

of thunderstorms was over the airport at this time with the heaviest activity located

northwest through northeast 5 miles from the runway.)

179

Figure 19 shows a map of the

weather conditions at 2339:45 and flight 1420’s location.

178

The windshear alert also indicated that the north boundary wind was 330º at 25 knots and

that the northwest boundary wind was 010º at 15 knots.

179

The leading edge is one of the most hazardous areas in thunderstorms because of the

updraft-downdraft interaction.

Analysis 116 Aircraft Accident Report

Figure 19. Weather and Flight Information for 2339:45

The controller instructed the flight crew to turn right to a heading of 250º for

vectors to the runway 4R approach course. The right turn, which routed the airplane away

from the airport, was necessary because the airplane was only about 4 miles southeast of

runway 4R/22L. However, the turn, and the subsequent maneuvering south of the airport,

meant that the flight crew had temporarily lost its ability to use the airborne weather radar

(because of its forward-looking design) to monitor the storm’s intensity and location

relative to the airport. The flight crew would regain the use of the airborne weather radar

to evaluate the storm and its location relative to the airport when the airplane turned back

toward the airport to intercept the final approach course for runway 4R. However, until

the airplane completed its base-to-final turn and leveled off, the radar data would have

likely been difficult to interpret, especially for detecting features in the direction of the

turn. (Thus, for about 7 minutes, the crew did not have precise information from the

airborne weather radar about the location of the storm relative to the airport.)

After the runway change from 22L to 4R, the CVR recorded the first officer

performing an abbreviated briefing for the newly assigned runway.

180

This briefing

consisted of the localizer frequency and course, the minimum safe altitude, the missed

approach procedure, and the decision altitude. Regarding the missed approach procedure,

180

Although American’s flight manual states that the captain is to ensure that a briefing is conducted

before every approach, it does not contain any procedure for performing an abbreviated briefing for

an instrument approach to a changed runway. The first officer stated in a postaccident interview that

the captain conducted a formal briefing for runway 22L. The CVR recording did not include an

approach briefing for runway 22L, but it is possible that the briefing occurred before the beginning

of the recording at 2319:44.

Analysis 117 Aircraft Accident Report

the CVR recorded the first officer’s statement, “missed approach right turn to four

thousand,” followed immediately by an unintelligible comment. The Jeppesen approach

plate for the airport indicated that the right turn at 4,000 feet was to be accomplished via

a 110° heading. The Safety Board could not conclusively determine whether the first

officer briefed that part of the missed approach procedure. However, the CVR contained

no discussion between the flight crewmembers about the missed approach procedure in

relation to the location of the storm, specifically, that the right turn would maneuver the

airplane clear of the weather.

The flight crew continued to receive, but did not discuss, wind reports from the

controller. The flight crew’s actions and conversations during this phase of the approach

largely consisted of steps to expedite the approach to the airport. For example, at

2342:40, the first officer asked the captain if he wanted to conduct a “short approach” and

“keep it in tight.” The captain accepted the suggestion but with the qualification “if you

see the runway, ‘cause I don’t quite see it.” (The airport was off the right, or first

officer’s, side of the airplane.) The first officer began to radio the controller at 2342:54 to

suggest a visual approach but interrupted this transmission 1 second later when he stated

to the captain, “it’s going right over the…field.” In a postaccident interview, the first

officer indicated that he was referring to a temporary obstruction of his view of the

airport because of clouds. The first officer’s next transmission, 4 seconds later, informed

the controller that “we got the airport” and that “we’re going between clouds.” Figure 20

shows a map of the weather conditions at 2342:55 and flight 1420’s location.

Figure 20. Weather and Flight Information for 2342:55

Analysis 118 Aircraft Accident Report

The controller offered the flight crew a visual approach to runway 4R, which the

crew accepted. The CVR indicated that, during the attempted visual approach, the captain

could not see the airport and was relying on the first officer to guide him toward the

airport.

181

The airplane’s position (that is, its orientation, proximity, and altitude) relative

to the airport likely made visual contact with the airport environment difficult for the

captain.

At 2344:19, the captain stated, “see we’re losing it. I don’t think we can maintain

visual.” The first officer notified the controller at 2344:30 that visual contact with the

airport had been lost because of a cloud between the airplane and the airport. The

controller then cleared the airplane to fly a heading of 220º for radar vectors for the ILS

approach to runway 4R. (ATC radar indicated that, at this time, flight 1420 was about

6 miles south of runway 4R, heading westbound and away from the runway. Weather

radar data indicated that heavy rain was occurring at the airport at the time that the

airplane was flying on the 220° heading.) Figure 21 shows a map of the weather

conditions at 2344:30 and flight 1420’s location.

Figure 21. Weather and Flight Information for 2344:30

181

At 2343:26, the first officer said, “…there’s the airport right there. Okay?” Five seconds later,

the captain asked, “where?” At 2343:31, the first officer said, “okay, you’re set up on a base for

it. Okay?” to which the captain questioned, “I’m on a base now?” At 2343:35, the first officer said,

“well, you’re on a dog leg. You’re comin’ in. There’s the airport.” Three seconds later, the captain

said, “I lost it,” and the first officer said, “right there, you’re you’re downwind. See it’s right there.”

The captain then said, “I still don’t see it...well just vector me. I don’t know.”

Analysis 119 Aircraft Accident Report

The weather south of the airport became the next focus of the flight crew’s

attention. NWS radar data indicated that, in the 6 minutes before the accident, a line of

thunderstorms, with several large areas of intense and extreme activity, was

encompassing the Little Rock airport area and runway 4R approach path. At 2345:47, the

first officer told the controller “we’re getting pretty close to this storm. we’ll keep it tight

if we have to.” The controller indicated to the flight crew that, “when you join the final,

you’re going to be right at just a little bit outside the marker if that’s gonna be okay for

ya.” The captain stated, “that’s great,” and the first officer told the controller, “that’s

great with us.” Figure 22 shows a map of the weather conditions at 2345:47 and

flight 1420’s location.

Figure 22. Weather and Flight Information for 2345:47

The airplane was once again turned westbound, and the controller provided the

flight crew with heading instructions to position the airplane to intercept the runway 4R

final approach course. At 2346:39, the controller informed the flight crew that the

airplane was 3 miles from the outer marker, instructed the crew to maintain 2,300 feet

until the airplane was established on the localizer, and cleared the crew for the runway 4R

ILS approach. In a postaccident interview, the first officer stated that, at this point, there

was urgency to land because the weather was “up against” the airport. However, the first

officer also stated that, after being vectored to the runway 4R ILS approach course, he

had visual contact with the runway throughout the rest of the approach.

At 2346:52, the controller advised the flight crew that heavy rain was falling at

the airport, visibility was less than 1 mile, ATIS information Romeo was no longer

current, and the runway visual range (RVR) was 3,000 feet. Simultaneously, the captain

Analysis 120 Aircraft Accident Report

stated, “aw, we’re goin’ right into this.” Figure 23 shows a map of the weather conditions

at 2346:52 and flight 1420’s location. About this point in the approach, the flight crew

should have regained the use of the airborne weather radar to depict precipitation levels

in front of the airplane. However, the Safety Board could not determine whether the radar

was displaying this information or whether the flight crewmembers, given their

workload, were able to perceive and interpret the information that the radar was

providing.

Figure 23. Weather and Flight Information for 2346:52

At 2347:08, the controller cleared flight 1420 to land and stated that the wind was

350º at 30 knots gusting to 45 knots. The CVR indicated that flight crew did not discuss

the wind information, the heavy rain that was already falling at the airport, or the

depiction of the weather on the airborne weather radar. Figure 24 shows a map of the

weather conditions at 2347:08 and flight 1420’s location.

Statements from the CVR and the first officer’s postaccident interviews indicated

the flight crew’s concern about the location of the thunderstorm in relation to the airport

and the airplane. However, the Safety Board concludes that, because the first officer was

able to maintain visual contact with the runway as the airplane was vectored for the final

approach course, both flight crewmembers might still have believed that flight 1420

could arrive at the airport before the thunderstorm. The Board understands that some

other flight crews might continue an approach to a runway under the same circumstances.

On the other hand, the Board also recognizes that the approaching storm and the reports

of heavy rain, dropping visibility, and increasing crosswinds (from 10 to 30 knots with

Analysis 121 Aircraft Accident Report

gusts to 45 knots) would be sufficient for some flight crews to hold until the storm passed

or proceed to an alternate airport.

Figure 24. Weather and Flight Information for 2347:08

2.2.1.3 Final Approach Segment

The final approach segment began when the airplane intercepted the localizer, at

2347:16. Simultaneously, the first officer erroneously read back the controller’s previous

wind report of 350º at 30 knots gusting to 45 knots as “zero three zero at four five.” The

flight crew then discussed the effect of the previously reported RVR on the approach.

The captain stated, “three thousand RVR. we can’t land on that.” The first officer then

correctly told the captain that the approach required a 2,400-foot RVR, and the captain

said, “ok, fine.”

The CVR contained no discussion between the flight crew about the fact that the

winds reported by the controller exceeded the company’s maximum crosswind

component for wet runways and the prevailing RVR. According to American’s flight

manual, the maximum crosswind component for a DC-9 landing on a runway with an

RVR less than 4,000 feet was 15 knots. The latest winds reported by the controller

resulted in a crosswind component of 23 knots for the steady-state wind and 34 knots for

the gusting wind. Thus, the landing for runway 4R could no longer be conducted in

accordance with company procedures. However, the flight crew likely did not recognize

that an operational criterion for conducting the landing had been exceeded because the

incorrect wind information read back by the first officer (030º at 45 knots) would have

resulted in a crosswind component under 10 knots.

Analysis 122 Aircraft Accident Report

The CVR recorded the captain commanding the landing gear down at 2347:44.

According to American’s procedures, this callout indicated that the second half of the

Before Landing checklist—landing gear, spoiler lever, autobrakes, flaps and slats, and

annunciator lights—was to be accomplished using a mechanical checklist in the cockpit

(see section 1.17.3.2). The CVR recorded the sound of the landing gear being operated

2 seconds after the callout and the captain’s statement “and lights, please” 5 seconds after

the callout. The CVR indicated that, at this point in the approach, none of the remaining

checklist items had been performed.

At 2347:53, the controller provided the flight crew with the second of the two

windshear alerts, stating that the centerfield wind was 350º at 32 knots with gusts to

45 knots.

182

The captain then indicated that he wanted 20 knots added to the approach

speed.

183

There was no further discussion of the winds, which still exceeded American’s

maximum crosswind limitations for landing.

184

(NWS and ATC radar data indicated that,

about this time, flight 1420 was on the leading edge of a line of thunderstorms.) The

Safety Board concludes that, when the second windshear alert was received, the flight

crew should have recognized that the approach to runway 4R should not continue

because the maximum crosswind component for conducting the landing had been

exceeded. Additionally, the flight crew should have recognized an apparent wind shift of

40° (a change from the mistaken 030º read back to the current 350º wind direction). This

apparent wind shift should have given the flight crew the impression of an acutely

changing weather situation, yet no related discussion was recorded on the CVR.

At 2348:12, the controller informed the flight crew that the RVR was 1,600 feet.

This RVR reading further reduced the company’s maximum crosswind component for

landing to 10 knots, which was required when an RVR was less than 1,800 feet.

However, the CVR still contained no discussion between the flight crew about the

exceeded maximum crosswind component. During public hearing testimony, the first

officer stated that he was still able to see the airport and that the RVR information

provided by the controller “did not concur” with what he and the captain were seeing.

Because the maximum crosswind component had been exceeded, the landing on

runway 4R would no longer be conducted in accordance with company procedures.

However, the Federal Aviation Regulations did not prohibit the approach to the runway

from continuing, even with weather that was below published minimums. Specifically,

14 CFR 121.651 (and American’s flight manual) allowed the crew to continue to the

decision height (200 feet afl, or 460 feet msl) because the airplane was already

established on its final approach segment. After hearing that the runway 4R RVR had

decreased to 1,600 feet, the captain stated, “well we’re established on the final,” and the

182

The windshear alert also indicated that the north boundary wind was 310º at 29 knots and

that the northeast boundary wind was 320º at 32 knots.

183

The former MD-80 Fleet Manager stated at the public hearing that, if a flight crew decided

to continue an approach after receiving an LLWAS alert, he would expect the pilots to increase speed.

184

In a postaccident interview, the first officer indicated that neither he nor the captain had

checked the crosswind limitation in the flight manual. The first officer also indicated that he had

started to look at his flight manual but that the captain had signaled him to put the manual away.

Analysis 123 Aircraft Accident Report

first officer informed the controller that the airplane was “established inbound.” The

controller then cleared the flight crew to land and provided updated winds.

At this point in the approach, there were cues (for example, the heavy rain that

was previously reported by the controller, the rapidly decreasing RVR, and the shifting

and gusting winds) that the weather at and around the airport had deteriorated

substantially. However, the CVR contained no discussion between the flight

crewmembers about whether the approach should continue. In the Safety Board’s

judgment, the deteriorating weather conditions would have prompted some flight crews

to abort the approach.

The CVR recorded the first officer’s required announcement of the 1,000-foot

altitude at 2348:50. At 2349:02, the first officer recognized that the final landing flap

configuration had not been established and asked the captain whether he wanted landing

flaps. The captain indicated that he thought he had already called for the landing flaps,

185

after which the first officer stated “forty now” and “forty forty land” to verify the flap/slat

handle position, the flaps position indicator, and the illumination of the SLAT/LAND

light. American’s DC-9 Operating Manual stated that landing flaps are to be selected by

1,000 feet afl. The first officer indicated in a postaccident interview that he thought the

airplane was about 900 feet afl when he selected the 40º final flap setting for landing,

which was consistent with calculations based on FDR data. The CVR did not indicate any

other Before Landing checklist callouts.

At 2349:10, the controller indicated that the winds were 330º at 28 knots. About

2 seconds later, the captain stated, “…this is a can of worms.” FDR data indicated that the

airplane was about 1,140 feet msl (880 feet afl) at this time and that the captain was

making active control inputs to keep the airplane on the localizer and glideslope. Also,

the FDR data showing fluctuating airspeed (±5 knots, with one excursion to +8 knots and

one to –10 knots) and vertical acceleration (±0.2 G with occasional excursions to

±0.25 G)

186

were consistent with the presence of gusty and turbulent winds. Figure 25

shows a map of the weather conditions at 2349:12 and flight 1420’s location.

The first officer called the runway in sight at 2349:24, and the captain identified

the runway about 7 seconds later. FDR data indicated that, at the time that the first officer

called the runway in sight, the airplane’s altitude was about 990 feet msl (730 feet afl),

airspeed was 160 knots, glideslope deviation was 0.3 dot high, and localizer deviation

was less than 0.1 dot to the left. FDR data also indicated that, at the time that the captain

identified the runway, the airplane’s altitude was 940 feet msl (680 feet afl), airspeed was

158 knots, glideslope deviation was 0.6 dot high, and localizer deviation was 0.1 dot to

the left. At 2349:41, the CVR recorded a sound consistent with windshield wiper motion;

the airplane’s altitude at this time was about 820 feet msl (560 feet afl).

185

The captain also did not call for flaps 28 earlier in the approach. The CVR indicated that,

at 2348:03, the first officer had asked the captain, “flaps twenty eight?”.

186

G is a unit of measurement equivalent to the acceleration caused by the earth’s gravity

(32.174 feet/sec

Analysis 124 Aircraft Accident Report

Figure 25. Weather and Flight Information for 2349:12

The first officer made his required 500-foot altitude callout at 2349:46. About

10 seconds later, an unidentified voice in the cockpit stated, “…we’re off course.” One

second later, the CVR recorded an unintelligible statement made by an unidentified voice

in the cockpit. In a postaccident interview, the first officer indicated that, about this time,

he said quietly “go around.” At 2350:00, the first officer stated, “we’re way off.” In a

postaccident interview, the first officer indicated that he made this statement because the

airplane was off the localizer and the right-side runway edge lights were drifting to his

left. Calculations based on FDR data indicated that the airplane was about 477 feet msl

(217 feet afl) and that the localizer deviation was between 0.9 and 1.0 dot to the right. As

the airplane descended through the decision height (460 feet msl), the glideslope

deviation was between 1.0 and 1.5 dots high and increasing.

American’s DC-9 Operating Manual states that, on final approach, a callout will

be made whenever either crewmember observes localizer displacement greater than

1/3 dot and/or glideslope displacement greater than 1/2 dot and that the other

crewmember will acknowledge this deviation. FDR data indicated that the localizer and

glideslope were both displaced in excess of these values, but the CVR contained no

callout from either the captain or the first officer regarding the fact that the airplane was

high on the glideslope and was to the right on the localizer. The first officer’s statement,

“we’re way off,” at 2350:00 (apparently in reference to the localizer displacement) was

not a proper callout according to company procedures because it did not specifically

inform the captain which ILS component was “off” and by what amount.

Analysis 125 Aircraft Accident Report

Although American’s DC-9 Operating Manual instructed pilots to call out if the

glideslope or localizer was displaced, the guidance did not indicate what displacement

amount would result in an unstabilized approach and require a missed approach. In

postaccident interviews, the MD-80 Fleet Manager and several check airmen explained

that captains had the discretion to determine the maximum acceptable deviation from the

glideslope or the localizer. (The adequacy of American’s stabilized approach guidance is

discussed in section 2.6.1.) In addition, FAA guidance on the stabilized approach concept

in the Air Transportation Operations Inspector’s Handbook does not provide any specific

criteria for a stabilized approach; the guidance instead states that “operators of turbojet

aircraft must establish and use procedures which result in stabilized approaches” (see

section 1.17.3.4).

187

Thus, the Safety Board was not able to gauge whether the flight 1420

approach was considered unstabilized according to company procedures or FAA

guidance.

At 2350:02, the captain stated, “I can’t see it.” Calculations based on FDR data

indicated that the airplane was between 5 and 20 feet above the decision height at that

time. The Safety Board was not able to determine whether the captain’s comment

referred to his loss of visual contact with the runway or with its environment. Figure 26

shows a map of the weather conditions at 2350:02 and flight 1420’s location.

Figure 26. Weather and Flight Information for 2350:02

187

The Safety Board notes that the FAA’s guidance does provide general parameters and that

American based its stabilized approach guidance on these parameters rather than specific ones.

Analysis 126 Aircraft Accident Report

Title 14 CFR 121.651 (c)(3) and American’s DC-9 Operating Manual allow a

pilot to continue an approach below the decision height if a visual reference for the

intended runway is “distinctly visible and identifiable.” Such references include the

approach lighting system; the runway threshold or its markings; and the runway, its

markings, or its lights. At 2350:05, the captain stated, “I got it”; calculations based on

FDR data indicated that the airplane was between 10 and 30 feet below the decision

height at the time. However, the Safety Board could not determine whether the captain’s

comment referred to the runway or to its environment and the altitude at which the

captain regained visual contact. Further, the Board could not determine what altitude

would have been displayed on the altimeters as the airplane descended below the decision

height.

188

At 2350:11, the first officer made his required 100-foot callout. The CVR then

recorded the airplane’s GPWS announcement of the first of two sink rate alerts about

2350:13. FDR data indicated that the airplane was about 330 feet msl (70 feet afl) at the

time. The second sink rate alert was recorded about 1 second later, and FDR data

indicated that the airplane was about 310 feet msl (50 feet afl) at the time. These alerts

indicated that the airplane’s air data computer system was computing rates of descent that

exceeded predetermined thresholds. The first officer made his required 50-, 40-, 30-, 20-,

and 10-foot callouts between about 2350:13 and 2350:18, and the captain continued the

approach to a landing.

2.2.1.4 Summary of the Flight Crew’s Performance During the Approach

As the flight crew was maneuvering the airplane for landing, there were events

that, individually, might not necessitate aborting an approach: a runway change because

of a shifting wind, a failed visual approach to the newly assigned runway, the temporary

inability of the airborne weather radar to show the weather conditions at the airport

because of the airplane’s direction of travel, the controller’s report of the second part of

the thunderstorm moving through the airport area, and the acceptance of a short approach

near the outer marker because of the airplane’s location in relation to the storm. However,

these events, collectively, should have heightened the crewmembers’ awareness that they

might not be able to safely continue the approach. Thus, it would have been appropriate

for the flight crew to have discussed specific options (holding, diverting to one of the two

alternate airports, or performing a missed approach after the airplane was established on

the final approach segment) in the event that the weather would necessitate aborting the

approach later.

189

As the airplane intercepted the ILS final approach course for runway 4R, the

flight crew entered an event-dependent, high workload phase of flight. Under normal

188

Flight 1420 received an altimeter setting of 29.86 inches of Hg. An altimeter with this setting

and zero instrument error would display an altitude that was about 55 feet lower than the airplane’s

actual altitude during the final approach segment. However, because the Federal Aviation Regulations

(14 CFR Part 43, Appendix E) allow instrument error of up to ± 75 feet, the altimeters could

have been reading up to 20 feet higher than the airplane’s actual altitude.

189

The Safety Board notes that the airplane would have had an adequate fuel supply for each

of these options.

Analysis 127 Aircraft Accident Report

conditions, tasks during this phase of flight include controlling and maneuvering the

airplane, configuring the airplane to land, performing final landing checks, and

evaluating the airplane’s performance relative to the landing criteria. In this case, the

flight crew’s earlier decision to accept a short approach increased the crew’s already high

workload by compressing the amount of time that was available to accomplish required

tasks. In fact, the first officer highlighted this issue at the public hearing when he stated,

“I remember around the time of making that base-to-final turn, how fast and compressed

everything seemed to happen.”

During the final approach, the flight crew had a significant amount of weather

information that had to be simultaneously evaluated. This information included the

controller’s previous report of heavy rain at the airport with visibility less than 1 mile, the

second windshear alert,

190

a rapidly decreasing RVR, and several wind reports. Under

these circumstances, some flight crews would have decided to abandon the approach.

The flight 1420 crew then poorly performed, and did not complete, the second

half of the Before Landing checklist. Although the sound of the landing gear being

operated was recorded by the CVR, there is no CVR evidence to indicate that the first

officer verbalized this checklist item as “down, three green” as required, which would

have indicated that all three landing gear systems were in the down and locked position.

Also, the captain commanded the 40° final landing flap configuration only after being

queried by the first officer. As previously stated, the captain did not realize that he had

not yet called for the flaps; as a result, this checklist item was performed late. In addition,

there is no CVR evidence to indicate that the first officer called out that the spoiler lever

was armed, checked the annunciator lights, and completed the Before Landing checklist

or that either pilot had armed the spoiler lever and considered the use of automatic rather

than manual braking in light of the deteriorating weather conditions.

191

As previously discussed, the flight crew should have initiated a go-around during

the final approach segment when a specific operational criterion was not met, that is,

when the company’s maximum crosswind component for conducting the landing was

exceeded. The flight crewmembers’ failure to establish the final landing flap

configuration before reaching 1,000 feet afl and their failure to maintain a normal rate of

descent, under different circumstances, might not necessitate a go-around. However, the

Safety Board concludes that, because of the flight crew’s failure to adequately prepare for

the approach and the rapidly deteriorating weather conditions, the likelihood of safely

completing the approach was decreasing, and the need to take a different course of action

190

Even though American’s flight manual indicated that windshear alerts were advisory, its DC-9

Operating Manual instructed pilots to avoid areas of known severe windshear and search for clues

to indicate the presence of severe windshear. The operating manual indicated that such clues included

thunderstorm and convective clouds, rain showers, and strong or gusty surface winds. All of these

conditions were present at some point and to some degree during flight 1420’s approach to the airport.

191

The Safety Board notes that several events were competing for the first officer’s attention

at the time that he was performing the Before Landing checklist, as detailed in section 2.2.1.3.

The Board further notes that, even though the first officer did not verbalize that the Before Landing

checklist had been completed, as required, the CVR indicated that he did verbalize (early in the

approach) that the Descent checklist had been completed.

Analysis 128 Aircraft Accident Report

was progressively increasing; as a result, the flight crew should have abandoned the

approach. Factors that contributed to the flight crew’s performance during the accident

flight are discussed in section 2.2.3.

Finally, it is important to note that a microburst, with a peak wind gust of 76 knots

and rainfall rates of 9 inches per hour, impacted the airport shortly after the flight 1420

accident and thus was not a factor. NWS radar data, however, detected that the

microburst was over runway 4R at 2353:00. Thus, if flight 1420’s takeoff, en route flight,

or approach to landing had been delayed by less than 2 minutes, the flight could have

encountered the microburst on final approach. Microbursts can result in vertical and

horizontal windshear that can be extremely hazardous to aircraft, especially at low

altitudes, as demonstrated by the 1994 USAir flight 1016 accident in Charlotte, North

Carolina, and the 1985 Delta Air Lines flight 191 accident in Dallas, Texas (see

section 1.18.5). As a result, the Safety Board is concerned that the flight crew was

operating in an environment that was conducive to microburst conditions.

192

Section 2.2.3.1.1 presents an industry-wide recommendation to develop operational

strategies and guidance to promote better flight crew decision-making regarding the

penetration of severe convective activity.

2.2.2 The Landing

Flight 1420 touched down on runway 4R at 2350:20 at a speed of 160 knots. The

airplane touched down about 2,000 feet down the 7,200-foot runway, slightly to the right

of the centerline and sliding to the right. According to calculations based on FDR data,

the airplane was subjected to a 5-knot tailwind component upon touchdown and a 20- to

25-knot left-to-right crosswind component during the landing.

The NWS’ Automated Surface Observing System (ASOS) weather data indicated

that surface winds from 290° at 16 knots gusting to 22 knots were present about the time

that flight 1420 touched down, but this information was not available to the flight crew or

the controller because the system’s 2-minute wind data are not directly reported to the

control tower. The controller’s final wind report to the flight crew (320º at 23 knots,

which was transmitted 27 seconds before touchdown) would not have indicated the

possibility of a tailwind component at touchdown. Although the 5-knot tailwind

component was within the 10-knot limitation required by American’s flight manual and

advised by Boeing in its MD-80 Flight Crew Operating Manual (FCOM), the Safety

Board notes that the flight crew’s purpose in changing runways from 22L to 4R was to

avoid a tailwind component. The 20- to 25-knot crosswind component, however,

exceeded the 10-knot limitation required by American’s flight manual for a runway with

an RVR of less than 1,800 feet.

192

The Safety Board notes that, during flight 1420’s final approach segment, the microburst was

located northwest of the airport and that the missed approach procedure would have taken the airplane

east of the airport.

Analysis 129 Aircraft Accident Report

Flight 1420 departed the end of runway 4R sliding to the left, with its nose gear

on the left edge of the runway and the main gear off the left edge of the runway, at a

calculated speed of 97 knots. The airplane sustained no damage before its departure from

the runway. The airplane likely collided with the runway 22L approach lighting system

support structure at a calculated speed of about 83 knots.

With the use of Boeing’s MD-80 Operational Landing Program, the Safety Board

predicted that the accident airplane experienced a wet runway braking coefficient of at

least 0.23 at 140 knots and 0.25 at 160 knots. Typical braking coefficients to indicate

dynamic hydroplaning range from 0.02 to 0.04. Thus, flight 1420 experienced a

maximum braking coefficient that was over six times greater than the maximum typical

hydroplaning braking coefficient.

Examination of the accident airplane’s hydraulic brake system and the lack of

hydraulic fluid contamination on the runway indicated that the hydraulic brake system

was most likely capable of functioning within operational parameters. Examination and

testing of the airplane’s antiskid system and the lack of flat spots or reverted rubber on

any main landing gear tire indicated that the antiskid system was most likely capable of

functioning within operational parameters.

The Safety Board’s examination of runway 4R determined that it was in good

condition with normal or better-than-normal surface conditions for friction levels and

counter-hydroplaning effectiveness. A senior research engineer from NASA’s Langley

Research Center stated at the public hearing that runway 4R’s grooving was satisfactory

but that its microtexture was above average and macrotexture was excellent. More

importantly, the engineer stated that the runway’s ability to prevent hydroplaning and

other braking problems was excellent. The Safety Board concludes that dynamic or

reverted rubber hydroplaning did not occur during the accident airplane’s landing rollout.

Flight 1420’s substantial drift angle while on the runway (as much as 16º both to

the right and the left of the direction of travel) was evidence of a directional control

problem. The Safety Board examined the lack of spoiler deployment upon touchdown,

the use of reverse thrust greater than 1.3 engine pressure ratio (EPR), and the use of

manual rather than automatic brakes and evaluated the role that each played in the flight

crew’s inability to maintain directional control of the airplane on the runway and stop the

airplane within the remaining available runway length (5,200 feet).

2.2.2.1 Lack of Spoiler Deployment

The spoiler system aboard the accident airplane was reported to be operating

properly by the previous flight crew. FDR data from the airplane’s previous landing

showed that the left outboard and right inboard flight spoilers (the only two spoiler

parameters recorded) deployed upon touchdown and remained fully deflected for about

33 seconds. Also, FDR data from the accident flight indicated that the flight spoilers

extended symmetrically (in response to a pilot input using the spoiler handle) for about

55 seconds during the descent into Little Rock.

Analysis 130 Aircraft Accident Report

After the accident, the ground and flight spoilers were found in the retracted

position. FDR data showed that, other than a momentary deflection by the left outboard

flight spoiler concurrent with a left aileron deflection, the spoilers did not deploy upon

touchdown. FDR data also showed a momentary full deflection of the right inboard flight

spoiler concurrent with a full right aileron roll input during the landing rollout. The

spoiler movement recorded on the FDR indicated that the spoiler position sensors and the

spoiler control and hydraulic systems were working properly before and during the

landing rollout.

2.2.2.1.1 Autospoiler Arming

As stated in section 2.2.1.3, autospoiler arming was one of the 10 items on the

Before Landing mechanical checklist. According to American’s DC-9 Operating Manual

procedures that were in effect at the time of the accident, the nonflying pilot was

responsible for ensuring that each checklist item had been accomplished. For autospoiler

arming, the nonflying pilot was to state “spoiler lever,” announce that the spoilers had

been armed, and move the spoiler lever switch on the mechanical checklist. American’s

DC-9 Operating Manual procedures, however, did not specify which pilot was

responsible for physically arming the spoilers. The company’s June 25, 2001, party

submission indicated that, for flight 1420, the first officer (the nonflying pilot) was

responsible for arming the spoilers.

According to postaccident interviews with company line pilots, instructors, and

check airmen, American’s MD-80 pilots were instructed during simulator training that

the nonflying pilot was to arm the spoilers. In contrast, the line pilots indicated that it was

accepted practice for captains to arm the spoilers, regardless of whether they were the

flying or nonflying pilot, because the spoiler handle is located on the forward left, or

captain’s, side of the center pedestal.

193

The Safety Board notes that, on February 23,

2000, American revised its procedures to state that the captain is always responsible for

arming the spoilers.

As the chief pilot at American’s Chicago base of operations, the captain was

required to fly less frequently than line pilots. However, the captain had become a chief

pilot only months before the accident and had flown 54 hours in the 90 days before the

accident flight. Thus, it is likely that he was aware of the common practice during line

operations for the captain, rather than the nonflying pilot, to arm the spoilers. However,

as a check airman, the captain would have likely been ensuring during check rides that

the nonflying pilot was arming the spoilers.

193

To arm the spoilers on the MD-80 series airplane, the handle must be grasped and squeezed

while it is lifted up toward its full travel limit. The pilot in the left seat of the airplane can reach

the handle without obstruction, but the pilot in the right seat of the airplane must reach in front

of, or behind, the throttle levers (and possibly around the captain’s arm if his or her hand were

on the throttle at the time) to manipulate the spoiler handle. A search of NASA’s Aviation Safety

and Reporting System database using the terms “spoiler handle” and “spoiler arming” did not reveal

any reports consistent with flight crews having difficulty arming the spoiler handle.

Analysis 131 Aircraft Accident Report

The first officer for flight 1420 stated, in a postaccident interview, that he did not

arm the spoilers because he thought that the captain had armed them. The first officer

indicated that he had moved the switch on the mechanical checklist to indicate that the

spoilers had been armed. However, the CVR did not indicate any mention of the first

officer’s required “spoiler lever” callout or any verbal verification that the spoilers were

armed.

194

In addition, there were no sounds or other indications on the CVR that were

consistent with the spoiler handle being armed.

During postaccident examination, the spoiler handle appeared to be in the full aft

position with the red ARM indicator stripe partially visible; however, the cockpit center

pedestal showed deformation and displacement in the left downward direction as a result

of the damage to the cockpit floor structure during the impact sequence. Teardown of the

center pedestal revealed no significant marks on the spoiler handle or handle slot to

indicate the handle’s position at impact. In addition, the teardown revealed that the

autospoiler crank arm was found in its fully extended position and was positioned above

the spoiler handle’s roller—the contact point against which the crank arm pushes to

extend the handle during normal autospoiler extension.

Teardown of the autospoiler actuator revealed that the actuator was in the fully

extended (deployed) position, which was consistent with proper autospoiler operation

with the autospoiler handle in the unarmed position. Thus, the spoiler handle was actually

in the stowed position at impact but appeared to be in the fully extended position only

because of impact damage. Further, testing of the primary components of the autospoiler

system—the autospoiler actuator, the autospoiler switching unit, the ground spoiler

control box, and the two ground nose oleo switches—revealed that all were capable of

functioning properly.

The CVR transcript identified the sound of two “thuds” similar to the sound of an

airplane touching down on a runway along with a “squeak” sound at 2350:20.2,

2 seconds before the first officer stated, “we’re down.” Analysis of the CVR sound

spectrum for the accident airplane indicated that the first “thud” was main landing gear

touchdown and that the second “thud” was nose gear touchdown. Safety Board

investigators determined that the squeak sound, which occurred in between the two

“thud” sounds, was the autospoiler actuator operating. Because of the placement of the

squeak sound, Board investigators further determined that the actuator was triggered at

main landing gear touchdown by the wheel spin-up transducers.

194

There are straightforward visual cues to indicate the position of the spoiler handle to the

flight crew (see section 1.6.2), and a visual assessment of the handle’s status can be made from

either seat in the MD-80 cockpit. The pilot in the left seat has an unobstructed view of the entire

length of the spoiler handle. As a result, no change in head or body position is required to visually

assess the status of the handle. However, the pilot in the right seat of the airplane normally has

a partially obstructed view of the spoiler handle. Obstructions in the normal field of view include

the throttle handles, the thrust reverser levers, and the right hand of the pilot in the left seat. If

the pilot in the right seat were to adjust his or her head position, an unobstructed view of the

handle could be achieved to assess the status of the spoiler handle. A search of the Aviation Safety

and Reporting System database using the terms “spoiler handle” and “spoiler arming” did not reveal

any reports consistent with flight crew difficulty confirming the spoiler handle’s status on MD-80

series airplanes.

Analysis 132 Aircraft Accident Report

The Safety Board compared the CVR sound spectrum for the accident airplane

with those for other CVR recordings obtained during the Palm Springs incident and the

flight and ground tests that were conducted as part of the investigations of this accident

and the Palm Springs incident.

195

According to the CVR sound spectrum study, the sound

durations associated with the autospoiler actuator extension for the Little Rock and Palm

Springs flights (0.08 and 0.06 second, respectively) were consistent with the ground test

in which the spoiler handle was unarmed (0.06 second). The study also indicated that, for

the two flight tests and two ground tests involving normal autospoiler operation (that is,

when the spoiler handle was armed before touchdown and the spoilers fully deployed

after touchdown), the sounds associated with the activation of the autospoiler actuator

lasted at least two times longer for the flight and ground test airplane recordings (between

0.16 and 0.19 second) than the sounds for the Little Rock airplane recording.

The ground tests conducted on the Palm Springs airplane also indicated that,

when the left throttle was at least 1 3/8 inches above idle and the autospoiler actuator

operated, the spoiler handle would fully extend (by means of the autospoiler actuator), be

knocked down, and retract about 0.5 second later to the stowed position. Because the two

spoiler positions—the left outboard and right inboard—are recorded alternately every

0.5 second with about 0.25 second in between recordings, it is likely that an autoretract

event would be captured by an FDR. In fact, the ground test FDR captured all of the

autoretract events, and the two spoiler autoretract events reported by American (see

section 1.6.2.2) were captured on FDR data. However, no indication of an autoretract

event was found on the flight 1420 FDR data.

The CVR sound spectrum study, physical evidence, and testing and teardown

results indicate that the autospoiler actuator on the flight 1420 airplane operated properly

at touchdown and that the spoiler handle was in the unarmed position when the

autospoiler actuator extended. Accordingly, the Safety Board concludes that the

autospoiler system operated properly and that the spoilers did not automatically deploy

because the spoiler handle was not armed by either pilot before landing.

2.2.2.1.2 Checklist Design Regarding Spoiler Arming

At the time of the accident, American did not require spoiler arming to be a dual

crewmember challenge and response checklist item. Dual callouts protect against the

failure that one pilot would identify an item as complete before it has been accomplished

or omitting an item entirely during a high workload phase of flight.

196

The human factors principles of checklist design established by the FAA and

incorporated in its publications are available to airlines and FAA inspectors. These

195

As discussed in section 1.16, the Little Rock accident and the February 16, 2000, Palm Springs

incident were similar in that they both involved American Airlines MD-80 series airplanes that did

not experience autospoiler extension at touchdown.

196

See U.S. Department of Transportation, Federal Aviation Administration. 1995.

Human

Performance Considerations in the Use and Design of Aircraft Checklists

. Also see Degani, A., and

Wiener, E.L. 1990. Human Factors of Flight-Deck Checklists: The Normal Checklist. National Aeronautics

and Space Administration Contractor Report 177549.

Analysis 133 Aircraft Accident Report

principles prescribe that checklist items affecting safety of flight should ensure that the

proper levels of operational redundancy have been achieved. For example, many

checklists require both flight crewmembers to positively confirm the status of flap

settings before takeoff and gear position before touchdown. The Safety Board recognizes

that MD-80 series airplanes can be dispatched with the autospoiler system inoperative;

thus, the system is not considered to be an item that is critical to safety of flight.

However, spoiler deployment after touchdown is crucial to optimal landing performance,

and the autospoiler system, if installed and operative, is the most reliable and efficient

way to engage the flight and ground spoilers during landing.

The Safety Board notes that, on June 22, 2000, American revised its Before

Landing checklist procedures, requiring both the flying and nonflying pilots to verify and

respond that the spoilers have been armed. The Board acknowledges that American’s

procedures now require dual crewmember confirmation of the spoiler arming checklist

item but is concerned that other airplane operators may not include this requirement in

their operating procedures. The Safety Board concludes that a high level of operational

redundancy should exist to ensure that spoiler arming has been completed before landing.

Therefore, the Safety Board believes that the FAA should require, for all 14 CFR

Part 121 and 135 operators of airplanes equipped with automatic spoiler systems, dual

crewmember confirmation before landing that the spoilers have been armed and that the

FAA should verify that these operators include this procedure in their flight manuals,

checklists, and training programs.

2.2.2.1.3 Manual Spoiler Deployment

At the time of the accident, American’s DC-9 Operating Manual required both

pilots to monitor the automatic deployment of the spoilers after touchdown. Pilots can

verify that the spoilers have automatically deployed by the extensive movement of the

handle to the extend position and the distinct “clanking” sound associated with the

handle’s travel. No written procedure in any of American’s manuals required either pilot

to announce the failure of the spoilers to automatically deploy. However, American’s

DC-9 Operating Manual indicated that, if the spoilers failed to deploy automatically, the

captain was responsible for manually extending them regardless of which pilot was

making the landing.

Even though American’s manuals did not require pilots to announce if the

spoilers failed to automatically extend, the MD-80 Fleet Manager and several MD-80

check airmen stated, during postaccident interviews, that pilots were trained to make this

announcement. (The adequacy of American’s spoiler system training at the time of the

accident is discussed in section 2.6.2.) However, the CVR did not record any

announcement by either pilot that the spoilers had failed to deploy automatically or any

sounds that could be associated with an attempt to manually extend the spoilers.

The Safety Board concludes that the flight crew failed to verify that the spoilers

had automatically deployed after landing and that the captain failed to manually extend

the spoilers when they did not deploy. As a result, the wings continued to support most of

the airplane’s weight, and very little weight was transferred to the landing gear,

Analysis 134 Aircraft Accident Report

preventing the main landing gear tires from developing the braking and cornering forces

required to achieve expected deceleration and directional control performance.

In the Little Rock accident, the pilots’ workload associated with attempting to

maintain directional control of the airplane on the runway may have prevented them from

detecting that the spoilers had not automatically deployed. The flight crew’s failure to

detect that the spoilers had not deployed might have been avoided if a procedural

requirement similar to the one in Boeing’s MD-80 FCOM had been in place at the time.

According to the Boeing manual, the nonflying pilot should call out “no spoilers” if the

spoiler lever does not move aft after touchdown, and the flying pilot should then move

the lever aft to the full extend position and up to the latched position. American’s

February 23, 2000, revision to its DC-9 Operating Manual incorporated a similar

requirement. The Safety Board acknowledges American’s revision but is disappointed

that this change and those related to spoiler arming (that is, the captain will always arm

the spoilers and both crewmembers will confirm that the spoilers have been armed) were

not put into effect until after the Palm Springs incident, which occurred more than 8

months after the Little Rock accident.

The Safety Board is concerned that other air carriers may not require callouts

regarding the status of the spoilers after touchdown. The Safety Board concludes that,

because spoiler deployment is critical for optimal landing performance, procedures to

ensure that the spoilers have deployed after touchdown should be a required part of all air

carriers’ landing operations. Therefore, the Safety Board believes that the FAA should

require, for all 14 CFR Part 121 and 135 operators, a callout if the spoilers do not

automatically or manually deploy during landing and a callout when the spoilers have

deployed and that the FAA should verify that these operators include these procedures in

their flight manuals, checklists, and training programs. The procedures should clearly

identify which pilot is responsible for making these callouts and which pilot is

responsible for deploying the spoilers if they do not automatically or manually deploy.

The Safety Board’s Airplane Performance Study demonstrated that spoiler

deployment is critical to an airplane’s braking force. For example, when the spoilers on

an MD-80 series airplane weighing 127,000 pounds and traveling at 140 knots are

extended, about 65 percent of the airplane’s weight is supported by the main landing

gear; when the spoilers are not extended, only about 15 percent of the airplane’s weight is

supported by the main landing gear.

197

In this accident, the light loading of the landing

gear substantially reduced the effectiveness of the brakes and the ability of the gear to

develop cornering loads to counter the aerodynamic side loads produced by the

197

These findings were consistent with information presented by the Boeing aerodynamics engineer

at the public hearing for this accident regarding the spoilers’ effect on the weight applied to the

wheels immediately after touchdown for an MD-80 series airplane with a similar landing weight

as the accident airplane.

Analysis 135 Aircraft Accident Report

crosswind.

198

The Safety Board concludes that the lack of spoiler deployment led directly

to the flight crew’s problems in stopping the airplane within the remaining available

runway length and maintaining directional control of the airplane on the runway.

Regarding the inability to stop an airplane within an available runway length, the lack of

spoiler deployment is far more critical than the use of reverse thrust (discussed in

section 2.2.2.2).

2.2.2.2 Use of Reverse Thrust Above 1.3 Engine Pressure Ratio

American’s DC-9 Operating Manual indicated that, for landings on slippery

runways, pilots were not to exceed 1.3 EPR on the “slippery portions of the runway”

except in an emergency situation. Likewise, Boeing’s MD-80 FCOM indicated that

reverse thrust of no more than 1.3 EPR should be used on wet or contaminated runways,

except in an emergency. However, FDR evidence indicated that reverse thrust exceeded

1.3 EPR several times during flight 1420’s landing sequence.

199

Further, American’s and

Boeing’s maximum reverse thrust setting for landings on dry runways was 1.6 EPR, and

FDR data showed that even this setting was exceeded many times during the landing.

The CVR recorded the first officer’s statement “we’re sliding ” about 4 seconds

after the airplane touched down on the runway. FDR data indicated that, about

1 1/2 seconds after this statement, 1.89 and 1.67 reverse EPR were being applied to the

left and right engines, respectively. FDR data indicated that the thrust reversers were

returned to the unlocked status as the airplane was continuing to slide.

200

However, the

thrust reversers were deployed again; the left engine reached a maximum setting of

1.98 reverse EPR, and the right engine reached a setting of 1.64 reverse EPR. The

reversers were returned again to the unlocked status; as the right thrust reverser was

moving to the unlocked status, the right engine reached a maximum setting of

1.74 reverse EPR. By this point, the captain was likely applying excessive reverse thrust

because he perceived that the landing had become an emergency situation.

198

The effects of the light loading on the main landing gear may have been exacerbated by

the position of the elevator surfaces during the landing roll. McDonnell Douglas’ 1996 all operators

letter indicated that, when operating on wet or slippery runways, pilots should not apply an excessive

amount of down elevator because it will unload the main gear and reduce braking efficiency. However,

the left and right elevator surfaces were deflected full nose down between 2,800 and 5,000 feet

beyond the runway 4R threshold.

199

The thrust reverser deployment recorded on the FDR when the airplane powered back from

the gate during its departure from Dallas/Fort Worth to Little Rock indicated that the thrust reverser

position sensors were working properly.

200

American’s DC-9 Operating Manual stated that, if an airplane begins to drift across the runway

while reversing, pilots should immediately come out of reverse thrust to help regain directional control

and restore rudder effectiveness.

Analysis 136 Aircraft Accident Report

On the MD-80 series airplane, excessive reverse thrust can reduce or eliminate

rudder and vertical stabilizer effectiveness.

201

In fact, FDR data for the flight 1420

airplane were consistent with deteriorated rudder and vertical stabilizer performance

because of the use of excessive reverse thrust.

202

Specifically, at the time that the heading

of the airplane was moving nose left from 1º to 3º per second despite substantial nose

right rudder inputs, the thrust reversers were deployed. The left engine EPR reached

values of 1.3 or greater almost continuously; the right engine EPR reached values of 1.3

or greater several times. The heading stopped moving left, and rudder effectiveness was

restored, when the thrust reversers were briefly stowed.

When reverse thrust was applied again, the airplane started to yaw left once more

(the heading was moving left about 1.5º per second) despite full right rudder inputs. The

airplane then reacted dramatically to the rudder inputs when the thrust reversers were

stowed for the second time; the yaw rate reversed, and the heading started to move right

up to 7º per second. The airplane was continuing to yaw to the right (the heading was

increasing 4º per second) when it departed the left side of the runway; at that time, the left

engine thrust reverser was deployed, but the engine’s EPR was at an idle power level.

The Safety Board concludes that the use of reverse thrust at levels greater than 1.3 EPR

significantly reduced the effectiveness of the airplane’s rudder and vertical stabilizer and

resulted in further directional control problems on the runway.

Postaccident observations of American’s MD-80 simulator training sessions

revealed that the training focused on applying 1.6 EPR reverse thrust when landing.

There was no discussion of the company’s procedures to limit reverse thrust to 1.3 EPR

when landing on slippery runways, and pilots were observed exceeding 1.3 EPR when

slippery runway conditions were presented. In fact, one of the simulator instructors

initially taught that 1.6 EPR was acceptable for landing on a slippery runway unless a

crosswind was present (see section 1.17.2.2). On February 23, 2000, American revised its

DC-9 Operating Manual, indicating that, for all MD-80 landings, reverse thrust should

not exceed 1.3 EPR unless stopping distance is in doubt.

The Safety Board concludes that the maximum reverse thrust for MD-80 landings

on wet or slippery runways should be 1.3 EPR, except when directional control can be

sacrificed for a marginal increase in deceleration. Therefore, the Safety Board believes

that the FAA should issue a flight standards information bulletin that requires the use of

1.3 EPR as the maximum reverse thrust power for MD-80 series airplanes under wet or

slippery runway conditions, except in an emergency in which directional control can be

201

American’s DC-9 Operating Manual stated that the rudder is almost “completely ineffective”

at 1.6 EPR and a speed of 90 knots. McDonnell Douglas, in its 1996 all operators letter, warned

that, as reverse thrust increases above approximately 1.3 EPR, rudder effectiveness continues to decrease

and that, at a reverse thrust greater than approximately 1.6 EPR, the rudder provides little or no

directional control. During the time that the accident airplane’s reverse thrust settings were 1.6 EPR

and greater, the airplane was traveling down the runway at speeds greater than 90 to 100 knots.

202

Loss of directional control associated with high reverse thrust EPRs has been observed in

other accidents involving DC-9 airplanes landing on wet or contaminated runways, including the

March 5, 1997, American Airlines flight 320 accident in Cleveland, Ohio. The description for this

accident, IAD97FA052, can be found on the Safety Board’s Web site at <http://www.ntsb.gov>.

Analysis 137 Aircraft Accident Report

sacrificed for decreased stopping distance. The Safety Board also believes that the FAA

should require principal operations inspectors (POI) of all operators of MD-80 series

airplanes to review and determine that these operators’ flight manuals and training

programs contain information on the decrease in rudder effectiveness when reverse thrust

power in excess of 1.3 EPR is applied. The Safety Board further believes that the FAA

should require all operators of MD-80 series airplanes to require a callout if reverse thrust

power exceeds the operators’ specific EPR settings.

2.2.2.3 Use of Manual Braking

The preflight weather package (which contained weather information issued

about 2205) indicated that the runways at Little Rock airport were wet with no

measurable rain and no braking action reports. According to the CVR, the captain had

decided, at 2331:24 (before establishing contact with the Little Rock tower), to use

manual brakes. At that point in the approach, the flightpath was still free of convective

activity. However, the CVR did not record any discussion between the pilots regarding

whether the use of manual brakes for the landing was still prudent in light of the changing

weather conditions on approach to the airport.

The flight crewmembers had received a report from the controller that heavy rain

was falling at the airport, so they should have been concerned that the runway would be

slippery. The crewmembers had also received a report from the controller about a strong

crosswind, so they should have been aware that significant rudder inputs could be

required to maintain directional control of the airplane on the runway and that these

inputs could take away from their ability to maximize manual braking. In addition, the

flight crew recognized that runway 4R was considered a short runway for a DC-9

landing,

203

so maximum efficiency of the brakes was critical. These conditions justified

the use of autobrakes for the landing on runway 4R. In fact, Boeing’s procedures indicated

that maximum autobrakes should be used for landing on a wet or slippery runway.

204

According to FDR data, the initial application of manual brakes began about

5 seconds after touchdown, and full application of the manual brakes was not recorded

until about 11 seconds after touchdown.

205

These time intervals are not indicative of

aggressive manual braking.

The use of autobrakes requires either automatic or manual spoiler deployment at

touchdown, which did not occur in this situation; therefore, autobrakes would not have

helped decelerate the accident airplane. However, if the spoilers had deployed and the

203

In a postaccident interview, the first officer stated that the captain briefed 40° flaps for the

landing because of the short runway.

204

As stated in section 1.17.4.3, American’s DC-9 Operating Manual stated that maximum autobrakes

could be used on short or slippery runways but did not specify the type of braking that should

be used on wet runways or during crosswind conditions.

205

As stated in section 1.17.4.3, American’s DC-9 Operating Manual stated that aggressive manual

braking could be used on short or slippery runways. The manual further stated that, for manual

brake stopping on short or slippery runways, the full brake pedal should be used immediately after

nose gear touchdown.

Analysis 138 Aircraft Accident Report

flight crew had selected maximum autobrakes for the landing, initial brake application

could have occurred about 4 seconds sooner.

On June 27, 2000, American issued a revision to its DC-9 Operating Manual,

requiring the use of autobrakes for four specific circumstances: when a runway is less

than 7,000 feet long; an RVR is less than 4,000 feet or visibility is less than 3/4 mile; a

runway is contaminated with standing water, snow, slush, or ice; or braking conditions

are reported to be “less than good.” Also, the revision recommended, but did not require,

the use of autobrakes when landing with gusty winds or crosswinds. The Safety Board

acknowledges that American has implemented specific criteria requiring the use of

autobrakes but is concerned that the company still does not require autobrakes during

landings with crosswinds.

In the aviation industry, it is generally understood that landing during high

crosswinds (that is, in excess of 10 knots) requires the pilot to make significant rudder

pedal inputs to maintain directional control of the airplane. Because manual braking is

accomplished by applying pressure to the upper portion of the rudder pedals, the Safety

Board is concerned that a pilot who uses manual brakes during high crosswind conditions

might not be able to immediately apply and maintain aggressive manual braking. In this

accident, the captain likely did not apply full manual braking for 11 seconds because he

was making significant rudder pedal inputs to keep the airplane on the runway. In fact,

the first officer stated that he had to help the captain with his braking efforts as the

airplane was nearing the end of the runway.

FDR data showed that the left brake pedal was relaxed momentarily after full

braking was achieved. However, this brake relaxation occurred while the airplane was

drifting to the right (that is, its nose was pointed to the left of the direction of travel) and

coincided with the application of full right rudder. Thus, the brake pedal relaxation may

have been the result of the captain’s attempt to apply differential brakes to correct the

airplane’s heading or his inability to maintain full braking while applying full right

rudder.

The use of automatic brakes is also important for airplanes landing on wet or

slippery runways. Brakes decrease stopping distance most effectively when they are

applied at a high speed. Any delay in brake application after touchdown results in a

considerable increase in the required stopping distance because the highest speed during

the ground roll (the most distance traveled per unit of time) occurs immediately after

touchdown. An interruption in brake application at lower speeds is less critical because

the airplane does not travel as far (that is, it does not consume as much runway) in the

same time at low speed as it does at high speed. Because wet or slippery runway

conditions degrade an airplane’s landing performance, fast brake application in these

circumstances is critical.

The Safety Board recognizes that airplane operators may not choose to require

automatic braking because that type of braking, compared with manual braking, will

wear out brakes faster and thus require brake replacement more often. The Board also

recognizes that, during optimal landing situations, pilots can apply manual brakes more

Analysis 139 Aircraft Accident Report

quickly than automatic brakes.

206

However, during high workload landing situations that

may require active or aggressive use of the rudder pedals, the use of automatic brakes

provides pilots with a faster, more consistent means for stopping an airplane within the

available runway length. The Safety Board concludes that automatic brake systems

reduce pilot workload during landings in wet, slippery, or high crosswind conditions.

Therefore, the Safety Board believes that the FAA should require, for all 14 CFR

Part 121 and 135 operators, the use of automatic brakes, if available and operative, for

landings during wet, slippery, or high crosswind conditions and that the FAA should

verify that these operators include this procedure in their flight manuals, checklists, and

training programs.

2.2.2.4 Summary of the Landing

The airplane’s performance during the landing roll indicates that flight 1420

experienced many of the difficulties discussed in McDonnell Douglas’ February 1996

MD-80 all operators letter regarding landing operations on wet or slippery runways:

greatly reduced reaction forces on the gear (because of the spoiler position), unloading of

the main gear because of large nose-down elevator inputs, strong crosswinds, loss of

vertical stabilizer and rudder effectiveness because of reverse thrust greater than 1.3 EPR,

and a slight tailwind (see section 1.17.4.4.2). In addition, the airplane touched down

about 2,000 feet down the 7,200-foot runway going 29 knots faster than the zero-wind

touchdown ground speed that would result from an approach at the reference airspeed.

The resulting ground trajectory of the airplane is consistent with the expected airplane

performance, as determined from Boeing’s Operational Landing Program, and the

operational experience outlined in the all operators letter.

The Safety Board’s Airplane Performance Study indicated that the accident

airplane could have stopped about 700 feet before the end of the runway if the spoilers

had deployed, a constant symmetrical reverse thrust at 1.3 EPR had been maintained, and

the flight 1420 manual braking profile had been applied. In contrast, with the spoilers not

extended, the airplane could not have stopped within the remaining runway length even if

maximum manual braking had been applied immediately after touchdown and

symmetrical reverse thrust at 1.3 EPR had been maintained throughout the landing roll.

Thus, the Safety Board concludes that the lack of spoiler deployment was the single most

important factor in the flight crew’s inability to stop the accident airplane within the

available runway length.

206

Boeing’s MD-80 FAA-approved Airplane Flight Manual, Appendix 5, “Automatic Brake System,”

states that, because of the delay in automatic brake application and the conservative testing conditions

that were used to construct the automatic brake landing distance data, stopping performance in the

MAX setting does not achieve the same level of performance compared with manual braking. In

addition, the manual states, “stopping distances are provided for guidance information only to assist

in the selection of the most desirable setting.”

Analysis 140 Aircraft Accident Report

2.2.3 Human Factors

During the accident flight, both crewmembers made basic errors in flight

management and the completion of routine tasks, including required callouts.

207

addition, the flight crew did not appear to be effectively evaluating the weather cues that

were available or considering their cumulative effect, specifically, that the thunderstorm

had likely already arrived at the airport. The Safety Board recognizes that the flight crew

was provided with only general, advisory information on severe weather avoidance rather

than specific operational decision-making criteria regarding the penetration of convective

activity (see section 2.3.1.1). However, the Safety Board concludes that the flight

crewmembers’ performance during the accident flight was degraded, as evidenced by

their operational errors and impaired decision-making.

The flight crew’s degraded performance was inconsistent with the level of

performance that would have been expected from both pilots, considering that the captain

was a chief pilot and check airman and that the first officer, as a new hire, had been

recently trained in American’s standards and procedures. Also, the flight crew’s

performance deviated significantly from the positive statements that other pilots made

about both pilots’ skills, abilities, and cockpit style. The captain was described in

postaccident interviews as a conservative pilot who used common sense, demonstrated

wisdom and experience, and was professional. The first officer was described in

postaccident interviews as an above-average new hire who was very competent and

knowledgeable and an experienced pilot with good cockpit discipline, and his

probationary file contained favorable comments about his performance. In addition, the

captain was appointed to the chief pilot position because he possessed good technical

skills and leadership abilities. Factors that contributed to the flight crew’s degraded

performance—situational stress and fatigue—are discussed in section 2.2.3.1 and 2.2.3.2,

respectively.

The pairing of the first officer, a new hire who was 5 months into his probationary

year, with the captain, a chief pilot and check airman with over 10,000 flight hours (more

than half of which were as an MD-80 captain), was evaluated and determined not to be a

factor in the accident. Although it is possible that a probationary first officer might find

speaking and challenging a captain who is a chief pilot to be difficult, CVR evidence

indicated that this first officer was assertive during most of the flight for which CVR

information was available. For example, the first officer initiated an abbreviated

approach briefing after the change in runways and queried the captain when he failed to

207

In addition to the callout errors discussed in section 2.2.1.3, the captain did not call out “Track—

Track” when he had the initial movement of the localizer needle on his horizontal situation indicator,

nor did he call out “Outer Marker” and the msl crossing altitude as the airplane crossed the outer

marker beacon on the ILS approach course. Finally, he did not call out “Landing” before descending

below the decision altitude, which would have confirmed that he had adequate visual reference with

the runway. The first officer did not challenge the captain on these required Category I ILS callouts.

Also, neither pilot made the required “radio altimeter alive” callout, and the first officer did not

call out, as required, “decision altitude,” speed deviations of ±5 knots, and rates of descent exceeding

1,000 feet per minute.

Analysis 141 Aircraft Accident Report

command the final landing flap configuration. Further, the first officer was instrumental

in directing the captain to the airport during the attempted visual approach.

2.2.3.1 The Role of Situational Stress

The presence of weather as a potential threat to the safety of flight and efforts to

expedite the landing were stresses to the flight crew. Research has demonstrated that

decision-making can be degraded when individuals are under stress because they

selectively focus on only a subset of cues in the environment. As a result, any situation

assessment may be incomplete, and the resulting decision, even when made by an expert,

may be degraded. Stress can also impede an individual’s ability to evaluate an alternative

course of action, resulting in a tendency to proceed with an original plan even though it

may no longer be optimal. Research on decision-making has demonstrated a natural

tendency for individuals to maintain their originally selected course of action until there

is clear and overwhelming evidence that the course of action should be changed (see

section 1.18.3.2).

The CVR contained no evidence to indicate that the flight crew had reevaluated

its original plan to expedite the landing because of the approaching weather. The actions

taken by the flight crewmembers throughout the approach were consistent with their

original plan. Despite several cues that indicated that the weather at the airport had

deteriorated, neither crewmember discussed a need to initiate a go-around, enter a

holding pattern, or divert to an alternate airport. The Safety Board also notes that any

delay in landing would have further extended the pilots’ duty day, but there is no

evidence to indicate that this factor affected the flight crew’s decision to continue the

approach.

The flight crewmembers’ intention to expedite the landing despite the weather

diverted their attention away from other activities during the final minutes of the flight

and, as a result, affected the crew’s ability to properly assess the situation and make

effective decisions. Therefore, the Safety Board concludes that the flight crewmembers’

focus on expediting the landing because of the impending weather contributed to their

degraded performance.

2.2.3.1.1 Industry Standards and Practices

The Safety Board evaluated the flight crew’s decision to conduct an approach to

an airport environment surrounded by severe convective activity in relation to

contemporary industry standards and practices. Most airlines and flight training programs

instruct pilots to avoid thunderstorms during routine operations. However, data from

accidents and incidents demonstrate that pilots penetrate thunderstorms—in some cases

with catastrophic results, as shown by the USAir flight 1016 and the Delta flight 191

accidents. In fact, in its final report on the Delta flight 191 accident, the Safety Board

stated that “there is an apparent lack of appreciation on the part of some, and perhaps

many, flight crews of the need to avoid thunderstorms and to appraise the position and

severity of the storms pessimistically and cautiously.”

Analysis 142 Aircraft Accident Report

A June 1999 report sponsored by NASA and conducted by research staff at the

Massachusetts Institute of Technology’s Lincoln Laboratory (see section 1.18.2) used

weather radar and ATC radar data sources to document flight crew behavior during

60 hours of observations in the Dallas/Fort Worth terminal area during convective

activity. This research documented that pilots routinely penetrated thunderstorms with

NWS precipitation intensity levels of 3 (strong), 4 (very strong), and 5 (extreme) rather

than deviated around them, especially when approaching an airport to land. Of the 1,952

encounters with thunderstorm cells recorded in these data, pilots penetrated the

thunderstorms 1,310 times (67 percent). However, the study did not include information

from the pilots regarding the reasons for the actions documented by the flight data. The

study concluded that pilots were more likely to penetrate a thunderstorm when they were

flying after dark, flying within 10 to 16 miles of the airport, following another aircraft, or

running behind schedule by more than 15 minutes. All but one of these factors (following

another aircraft) applied to the accident flight.

In its final report on the Delta flight 191 accident, the Safety Board stated its

concern that “the present training within the industry for windshear encounters on the final

approach seems to advocate the philosophy that the retrieval of the approach profile is the

desired end result and not escape from the environment.” The results of the NASA study,

which was completed 13 years after the Delta flight 191 accident report was issued,

demonstrate that this industry philosophy can also apply to the penetration of severe

thunderstorms.

Some air carriers, including American, provide their flight crews with only

general, advisory information on severe weather avoidance. As a result, the individual

flight crews are responsible for making decisions on whether an approach near

convective activity should continue, and such decisions are typically based on the pilots’

subjective assessment of the severity of the situation and their experience. The Safety

Board is aware that some other air carriers provide their pilots with specific operational

guidance, including decision aids and flow charts in quick reference checklists, from

which flight crews can make “go” or “no go” decisions concerning operations near

hazardous weather. Such information includes a detailed list of specific cues and

operational criteria from which pilots can easily assess weather conditions and

objectively determine whether they can safely continue or need to take a different course

of action. As a result, the pilots do not have to rely on an open-ended decision-making

process regarding whether and at what point to deviate around weather. Further, these

explicit, formalized cue recognition and decision aids minimize the potential for

thunderstorm penetrations resulting from impaired judgment and decision-making

because of situational stress or fatigue.

In addition, as demonstrated in this accident, airborne weather radar does not

always facilitate a flight crew’s assessment of a thunderstorm regarding the storm’s

location and movement relative to the airport and its severity, including the potential for

microburst conditions. The Safety Board is aware that recent technologies, such as

moving airport map displays integrated with airborne weather radar displays and real-time

wind readouts, are available in new-generation airplanes with “glass cockpits.”

208

Also,

Analysis 143 Aircraft Accident Report

NASA, the FAA, and avionics manufacturers are testing whether ground-based advanced

weather graphics, such as regional radar mosaics and single Doppler radar images, can be

up-linked to airplanes. These graphics can show enhanced detail of a thunderstorm

(including its intensity, movement, and tops) and other weather information; therefore,

they have the potential for providing flight crews with in-flight information to improve

their situational awareness and decision-making regarding hazardous weather.

Because the NASA study showed no discernible differences among operators and

airplane types regarding the propensity to penetrate thunderstorms, the Safety Board

concludes that aircraft penetration of thunderstorms occurs industry-wide. Therefore, the

Safety Board believes that the FAA should establish a joint Government-industry

working group to address, understand, and develop effective operational strategies and

guidance to reduce thunderstorm penetrations and should verify that these strategies and

guidance materials are incorporated into air carrier flight manuals and training programs

as the strategies become available. The working group should focus its efforts on all

facets of the airspace system, including ground- and cockpit-based solutions. The

near-term goal of the working group should be to establish clear and objective criteria to

facilitate recognition of cues associated with severe convective activity and guidance to

improve flight crew decision-making.

2.2.3.2 The Role of Fatigue

The CVR contained no statements to indicate that either pilot was tired, but the

CVR did record a yawn at 2324:13 from one of the pilots. The first officer stated, after the

accident, that it had been a long day and that he was getting tired but that he felt fine when

flying to Little Rock. In addition, the first officer stated that he did not remember talking

with the captain about whether he was tired, but the first officer was not concerned about

the captain being fatigued. However, research indicates that self-assessment of fatigue

impairment and detection of fatigue in others are inaccurate. Thus, the Safety Board

examined whether cumulative sleep loss, continuous hours of wakefulness, and the time

of the accident relative to the flight crew’s normal schedule were consistent with the

development of fatigue.

209

First, the captain and the first officer reportedly received a normal amount of

sleep the night before the accident; both went to sleep about 2200 and awoke about 0730.

208

The term “glass cockpit” refers to cockpits with cathode ray tubes or flat plate screens that

integrate multiple sources of flight information formerly displayed on analog dials and gyro instruments.

An example of a glass cockpit system is the Enhanced Flight Information System.

209

In its final report on the American International Airways flight 808 accident in Guantanamo

Bay, Cuba, the Safety Board explained that these three background factors are commonly examined

for evidence related to fatigue. For more information, see National Transportation Safety Board. 1994.

Uncontrolled Collision With Terrain, American International Airways Flight 808, Douglas DC-8-61,

N814CK, U.S. Naval Air Station, Guantanamo Bay, Cuba, August 18, 1993. Aircraft Accident Report

NTSB/AAR-94/04. Washington, DC. For additional information on fatigue-related factors, see Federal

Aviation Administration. 1998.

An Overview of the Scientific Literature Concerning Fatigue, Sleep,

and the Circadian Cycle. Prepared for the FAA’s Office of the Chief Scientific and Technical Advisor

for Human Factors.

Analysis 144 Aircraft Accident Report

Also, there was no evidence that either pilot had experienced cumulative sleep loss in the

days before the accident.

Second, at the time of the accident (2350:44), the captain and the first officer had

been continuously awake for at least 16 hours.

210

Research indicates that the normal

waking day is between 14 and 16 hours and that lapses in vigilance increase and become

longer if the normal waking day is extended.

211

In addition, the Safety Board’s 1994 study

of flight crew-related major aviation accidents (see section 1.18.3.1) found that flight

crews that had been awake for an average of about 13 hours made significantly more

errors, especially procedural and tactical decision errors, than crews that had been awake

for an average of about 5 hours. Thus, the flight crew’s extended continuous hours of

wakefulness was consistent with the development of fatigue.

Third, the 2350:44 accident time was nearly 2 hours after the time that both pilots

went to bed the night before the accident and the captain’s routine bedtime (between

2130 and 2200). According to a recognized expert in fatigue research who reviewed the

flight and duty time and CVR data associated with this accident, because the flight

crewmembers were conducting an approach at a time of night when they would have

normally been asleep, their circadian systems were not actively promoting alertness in

the last hours of their duty period. Thus, the time at which the accident occurred was

consistent with the development of fatigue.

212

Research indicates that the ability to consider options decreases as people who are

fatigued become fixated on a course of action or a desired outcome (which is also the case

with situational stress, as discussed in section 2.2.3.1.1) and that it can be more difficult

for a fatigued person to remember whether tasks have been accomplished.

213

In this

accident, the flight crew did not consider delaying or diverting the landing, the first officer

did not ensure that the autospoilers had been armed for landing, and the captain did not

realize that he had not called for flaps 40. Also, automatic processes (such as radio calls

and routine behavior) are affected less by fatigue than controlled processes (such as more

210

The flight crew had accumulated 7 hours 49 minutes of flight time during this period. Because

of differences in check-in times, the captain had accumulated 5 hours 24 minutes of ground time,

and the first officer had accumulated 5 hours 44 minutes of ground time. Therefore, when the accident

occurred, the captain’s duty day was 13 hours 13 minutes, and the first officer’s duty day was

13 hours 33 minutes. Both pilots’ total continuous time awake at the time of the accident was

at least 16 hours 21 minutes.

211

Kruger, G.P. 1989. “Sustained Work, Fatigue, Sleep Loss, and Performance: A Review of

the Issues.”

Work and Stress

. Vol. 3, pp. 129-141.

212

Continuous hours of wakefulness and accident time were also factors in the August 1997

Korean Air flight 801 accident in Guam (see section 1.18.6). The captain had been awake for 11 hours

at the time of the crash, which occurred after midnight in the flight crew’s home time zone (0142

Guam local time). The time of the crash was also several hours after the captain’s (the flying pilot)

normal bedtime.

213

Caldwell, J.A. 1997. “Fatigue in the Aviation Environment: An Overview of the Causes and

Effect as Well as Recommended Countermeasures.” Aviation, Space, and Environmental Medicine.

Vol. 68, pp. 932-938.

Analysis 145 Aircraft Accident Report

complex behavior, responses to new situations, and decision-making);

214

in this accident,

however, both automatic and controlled processes were affected during the flight. Further,

fatigue deteriorates performance on work-paced tasks that are characterized by time

pressure and task-dependent sequences of activities, as demonstrated by the flight crew’s

failure to properly perform routine tasks during the final approach phase of flight.

Therefore, the Safety Board concludes that the flight crew’s degraded performance was

consistent with known effects of fatigue.

Fatigue in transportation operations has been on the Safety Board’s list of Most

Wanted Safety Improvements since the list’s initiation in September 1990. In May 1999,

the Board issued a safety report that evaluated the DOT’s efforts to address operator

fatigue among the transportation modes, including aviation. (The Board first asked the

DOT to upgrade flight and duty times and hours-of-service regulations for all modes in

1989.) The Board’s report concluded that, despite acknowledgement by the DOT that

fatigue is a significant factor in transportation accidents, little progress has been made to

revise the hours-of-service regulations to incorporate the results of the latest research on

fatigue and sleep issues. As a result, the Board issued Safety Recommendation A-99-45,

asking the FAA to “establish within 2 years scientifically based hours-of-service

regulations that set limits on hours of service, provide predictable work and rest schedules,

and consider circadian rhythms and human sleep and rest requirements.”

On July 15, 1999, the FAA stated that, on December 11, 1995, it issued Notice of

Proposed Rulemaking (NPRM) 95-18, “Flight Crewmember Duty Period Limitations,

Flight Time Limitations and Rest Requirements.” The NPRM proposed amending

existing regulations to establish one set of duty period limitations, flight time limitations,

and rest requirements for flight crewmembers involved in air transportation. At an

October 7, 1999, meeting with the Safety Board, FAA representatives stated that the FAA

would not be able to meet the recommendation’s time requirement for a new rule. On

January 3, 2000, the Board indicated that, even though the NPRM was issued over 4 years

earlier, the existing regulations concerning flight time regulations and rest requirements

had not been upgraded. On December 5, 2000, the FAA stated that it planned to issue, in

spring 2001, a supplementary NPRM that would address the issue of fatigue “concretely”

and give the airlines the flexibility they need to operate. On April 26, 2001, the Board

indicated that, in the 5 years since the issuance of NPRM 95-18 and the 1 1/2 years since

the need for a supplemental NPRM was first communicated, the FAA has not taken action.

As a result, Safety Recommendation A-99-45 was classified “Open—Unacceptable

Response.”

In a May 14, 2001, press release, the FAA stated that it “is confident that, overall,

the airline industry complies with current FAA rules on pilot time limitations and rest

requirements.” The press release also stated the following:

On Nov. 20, 2000, the FAA responded to a letter from the Allied Pilots

Association that set forth specific scenarios that could affect a very small number

214

Humphrey, D.G.; Kramer, A.F.; and Stanny, R.R. 1994. “Influence of Extended Wakefulness

on Automatic and Nonautomatic Processing.”

Human Factors.

Vol. 36, pp. 652-669.

Analysis 146 Aircraft Accident Report

of all commercial pilots. The FAA’s response was consistent with the agency’s

long-standing interpretation of the current rules. In summary, the FAA reiterated

that each flight crew member must have a minimum of eight hours of rest in any

24-hour period that includes flight time. If a pilot’s actual rest was less than nine

hours in the 24-hour period, the next rest period must be lengthened to provide

for the appropriate compensatory rest. Ensuring that all pilots, especially those on

reserve duty, receive adequate rest is key to maintaining a safe aviation system.

On May 17, 2001, the FAA published a notice in the Federal Register that

reiterated its interpretation of pilot flight time and rest rules. The notice stated that the

FAA intended to enforce its rules in accordance with the interpretation and that, 6 months

after the issuance of the notice, the FAA would review airline flight scheduling practices

and deal stringently with any violations.

215

The Safety Board is encouraged by the FAA’s increased efforts to enforce the

current pilot flight time and rest rules. However, the Little Rock accident and the

May 1999 American Eagle flight 4925 accident in New York (see section 1.18.6) highlight

the need to expedite efforts to comprehensively address the issue of fatigue in aviation.

Therefore, the Safety Board reiterates Safety Recommendation A-99-45.

216

2.3 Meteorological Support

2.3.1 Weather Information Provided by the Local Controller

The local controller was working all of the tower cab positions on the night of the

accident and was not handling any other in-flight traffic, so flight 1420 had virtually his

full attention. The controller responded promptly to all of the flight crew’s requests. The

CVR indicated that, between 2339:59 and 2340:12, the first officer and the controller

discussed a change from runway 22L to 4R (after the wind shift to the northwest) and that

the controller responded at 2340:20 with a heading change for vectors to the runway 4R

ILS approach course. At 2344:30, the first officer informed the controller that the visual

approach to runway 4R could not continue because of a cloud between the airplane and

the airport, and the controller provided vectors for the ILS approach at 2344:39. The first

officer told the controller, at 2345:47, that the airplane was getting close to the storm, and

the controller returned with a new heading at 2345:52.

215

In June 2001, the Air Transport Association and the Regional Airline Association asked the

U.S. Court of Appeals for the District of Columbia to stay the pending enforcement of the FAA’s

interpretation of pilot flight time and rest rules, citing that enforcement of these rules constituted

illegal rulemaking. On September 5, 2001, the U.S. Court of Appeals granted the Air Transport

Association’s and the Regional Airline Association’s request for a stay of enforcement. As of October 29,

2001, the court was expected to hear arguments on this case in January 2002.

216

The Safety Board notes that this safety recommendation asked the FAA to take the recommended

action “within 2 years.” By reiterating this safety recommendation, the Board is not suggesting that

the recommended action should occur within 2 years from the data of this reiteration. Rather, because

the original 2-year period has already expired, the Board urges the FAA to expedite its efforts to

accomplish the action specified in this recommendation.

Analysis 147 Aircraft Accident Report

The controller also provided the flight crew with ongoing information about the

wind direction and speed, including the two windshear alerts, while the airplane was

approaching the airport and updated the wind information four times while the airplane

was on final approach. The controller also kept the crew apprised of the progress of the

thunderstorm. When the crew made initial contact with the tower, the controller indicated

that a thunderstorm located northwest of the airport was moving through the area. He also

informed the crew when the second part of the storm was moving through the area and

when heavy rain was falling at the airport. Further, the controller alerted the crew when

the RVR for runway 4R had decreased first to 3,000 feet and then to 1,600 feet and when

ATIS information Romeo was no longer valid. The Safety Board could not find any

instance in which the controller did not provide the flight crew with aviation weather

information that was available to him in the tower or any delay in relaying this

information, which is especially noteworthy considering that the weather conditions were

rapidly changing during the last several minutes of flight 1420’s approach to the airport.

The Safety Board concludes that the local controller provided appropriate,

pertinent, and timely weather information to the flight crew regarding the conditions on

approach to and at the airport. The controller’s actions after the crash occurred are

discussed in section 2.4.

2.3.1.1 Weather Information Depiction on Air Traffic Control Radar Systems

Although the controller accurately reported the weather information available in

the tower, he appeared to lack confidence in the tower’s radar weather depiction. For

example, the controller asked the flight crew how the final approach to runway 22L

looked on the airplane’s radar presentation because the airplane’s radar was “a lot better”

than what he had available in the tower. In addition, the ATC transcript indicated that a

controller from the Memphis ARTCC called the local controller to determine whether

flights headed toward Little Rock would be able to land. As part of his response to this

query, the local controller said, “my radar is not that good by the weather you know.” The

center controller’s response to this comment was “better than ours.”

The Safety Board notes that the radar used in ATC facilities was designed to

depict air traffic; it was not designed to show weather. If near-real-time color weather

radar had been available at ATC facilities, the Little Rock local controller would likely

have been able to relay to the flight 1420 crew that a thunderstorm with extreme

reflectivities had moved over the airport. In this case, the Board cannot determine whether

such a report would have changed the flight crew’s course of action because of the

workload at the time that the report would have been received, as well as the flight

crewmembers’ impaired performance. Nevertheless, ATC near-real-time color weather

radar information would enable controllers to provide flight crews with a better source of

weather information than is currently available in the tower.

The Safety Board concludes that, if near-real-time color weather radar showing

precipitation intensity were available, it would provide air traffic controllers with

improved representation of weather conditions in their areas of responsibility. Therefore,

the Safety Board believes that the FAA should incorporate, at all ATC facilities, a

Analysis 148 Aircraft Accident Report

near-real-time color weather radar display that shows detailed precipitation intensities.

This display could be incorporated by configuring existing and planned Terminal

Doppler Weather Radar (TDWR) or Weather Systems Processor (WSP) systems with this

capability or by procuring, within 1 year, a commercial computer weather program

currently available through the Internet or existing stand-alone computer hardware that

displays the closest single-site Weather Surveillance Radar 1988 Doppler (WSR-88D)

data or regional mosaic images.

2.3.2 Additional En Route Weather Information

2.3.2.1 Dispatch Office Weather Radar

After flight 1420 was underway (about 2240), the flight dispatcher transitioned

from a flight-releasing to a flight-following role, which required him to provide the

pilot-in-command with any safety-of-flight information that was pertinent to the flight’s

operation. However, the FAA does not generally provide Part 121 flight dispatch offices

with access to TDWR real-time weather radar information. Although American’s

dispatchers receive high-resolution weather radar mosaic updates every 15 minutes on

their workstations, the mosaics are delayed several minutes so that a clutter-free image

can be presented. Even though the 15-minute updates indicate the organization and

intensity of weather activity, the inherent delay in displaying the information (so that

images can be compared with other weather observations and corrected for beam height

and distance errors) prevents it from being depicted to the dispatcher in a timely manner.

In this accident, the thunderstorm activity was moving rapidly, and the 15-minute radar

updates could not adequately portray to the dispatcher the real-time conditions that

flight 1420 could encounter.

Dallas/Fort Worth is 1 of the 41 airports at which the FAA has installed TDWR;

Little Rock airport does not have the system, and the FAA does not plan to install the

system there. The availability of TDWR data to the flight dispatcher would not have

affected the outcome of the accident because TDWR presents only site-specific data;

thus, the TDWR at Dallas/Fort Worth would not have provided the dispatcher with

information about the weather conditions in the Little Rock airport area. However, for

those airports equipped or planned to be equipped with TDWR, information from that

radar system relayed by dispatchers would allow flight crews to have more detailed

current weather information en route than their airborne weather radar systems are able to

depict. This information would also help the dispatchers in planning, releasing, and

following flights. WSP systems, when they become available (which the FAA expects to

be in mid-2002), could provide the same benefits as TDWR for dispatchers located at

airports without TDWR. (Little Rock is not among the airports that will be receiving the

WSP system.)

The Safety Board concludes that the ability of flight dispatchers to provide timely

and accurate weather support would be enhanced if they had access to TDWR

information at airports where it is available and WSP information when the system

becomes available. Therefore, the Safety Board believes that the FAA should provide

Analysis 149 Aircraft Accident Report

U.S. air carriers operating under 14 CFR Part 121 access to TDWR, at airports where the

system is available, and access to the WSP system, when it becomes available, so that

their flight dispatch offices can use this information in planning, releasing, and following

flights during periods in which hazardous weather might impact safety of flight.

2.3.2.2 Center Weather Service Unit Staffing

Flight 1420 was handled by the Memphis ARTCC before the flight entered Little

Rock airspace. The controllers at this center did not have access to real-time weather

radar data, and no internal meteorological support was available to them because the

center weather service unit (CWSU) had closed. The CWSU at the Memphis center was

not staffed for 24-hour operation and had closed on the night of the accident about 2130,

even though severe weather was predicted to affect the center’s airspace. The CWSU

meteorologists have access to WSR-88D weather products and thus could have provided

the center controller with better information regarding the line of thunderstorms moving

into the area. However, the availability of this information likely would not have affected

the outcome of the accident because of the flight crew’s impaired performance.

In its final report on the USAir flight 1016 accident, the Safety Board issued

Safety Recommendations A-95-48 and -52, which asked the FAA and NWS, in

cooperation with each other, to reevaluate the CWSU program and develop procedures to

enable meteorologists to immediately disseminate information about rapidly developing

hazardous weather conditions to Terminal Radar Approach Control and tower facilities.

On October 22, 2001, and August 7, 2001, the Board acknowledged that the FAA and

NWS, respectively, were working to address the actions specified in the

recommendations but expressed concern that the work was not scheduled to be

completed in a timely manner. Pending completion of the FAA’s and NWS’ planned

actions, Safety Recommendations A-95-48 and -52 were classified “Open—Acceptable

Response” and “Open—Unacceptable Response,” respectively.

Even after the FAA and NWS have completed actions to address these

recommendations, their intent cannot be fully achieved unless the CWSUs are adequately

staffed at all times when rapidly developing hazardous weather conditions are possible.

(In letters to the Safety Board regarding the progress in implementing these safety

recommendations, neither agency has described such staffing for CWSUs.) The Safety

Board concludes that CWSUs should be staffed at all times when any significant weather

is predicted to affect their areas of operation, even if the weather is predicted to occur

before or after normal operating hours. Therefore, the Safety Board believes that the FAA

and the NWS, in cooperation with the other, should ensure that CWSUs are adequately

staffed at all times when any significant weather is forecast.

2.3.2.3 Automated Surface Observing System Lockout Period

The lockout feature on the ASOS (that is, the time period between 47:20 and

53:20 after each hour when METARs [meteorological aerodrome reports] are prepared,

edited, and transmitted) prevented the system from issuing pertinent weather information

for the flight crew. If the lockout had not been in place, the system would have issued a

Analysis 150 Aircraft Accident Report

special observation when the reduced visibility, heavy rain, and strong gusting winds

associated with the thunderstorm were detected. An additional special observation would

have been issued when the visibility was further reduced. (The ASOS edit log indicated

that a special observation at 2347:22 was canceled, and a 2350:31 special observation

was recorded but not disseminated.)

The canceled observation would have likely indicated that the thunderstorm was

at the airport and provided the flight crew with critical situational awareness information

about the intensity of the storm. Because the accident airplane did not touch down until

2350:20, the information in the canceled 2347:22 special observation would have

provided the flight crew with another indication that it was unsafe to land. The NWS

advised the Safety Board that the next ASOS software implementation would eliminate

the lockout period but that a target date for implementation has not been established

because of problems with the software. The Safety Board concludes that the ASOS

lockout period can prevent the relay of critical weather information to flight crews.

Therefore, the Safety Board believes that the NWS should eliminate the ASOS lockout

feature as soon as possible.

2.3.3 Airport Weather Equipment

2.3.3.1 Runway Visual Range System

Although the new-generation RVR system at Little Rock was designed to provide

a more accurate reading than that provided by the previous RVR system, the Safety

Board has two concerns about the new system. First, the RVR data were not directly

transmitted to the ASOS. Second, the Board did not have access to 1-minute RVR data

for this accident because an event log was not started.

The 1-minute RVR data were not included in the FAA’s initial specifications for

ASOS recorded data. As a result, certified weather observers are required to contact

tower controllers to obtain the 10-minute average RVR readings, and the weather

observers use this information in preparing METARs and SPECIs [special weather

observations]. However, it is a recommended practice, under Annex 3 to the Convention

on International Civil Aviation, for RVR data to be included in automated weather

observation systems because of the data’s importance to takeoff and landing operations.

This accident demonstrates how an RVR reading can decrease drastically in a

short timeframe; the RVR reading of 3,000 feet at 2346:52 had decreased to 1,600 feet less

than 1 1/2 minutes later. Because a change in visibility is one of the conditions that

generates a special weather observation, the Safety Board concludes that RVR data

should be directly reported to automated weather systems. Therefore, the Safety Board

believes that the FAA should modify automated weather systems to accept RVR data

directly from RVR sensors.

In addition, an RVR event log presents a total of 12 hours of the system’s data.

When an event log is started, the previous 2 hours of recorded RVR data are saved, and

Analysis 151 Aircraft Accident Report

the next 10 hours of RVR data are recorded and saved. Thus, after this accident, an event

log needed to be started no later than 0150 on June 2, 1999, to have preserved 1-minute

RVR data before and at the time of the accident. Because an event log was not started,

these data were overwritten by newer data. Airways facility personnel are responsible for

starting event logs; however, these personnel are not always present in the 2 hours after

an event occurs. This small timeframe during which personnel are required to start event

logs does not account for the possibility that RVR data will need to be retrieved.

The Safety Board concludes that the current 2-hour RVR archiving capability is

inadequate to ensure that data can be preserved for future use. Therefore, the Safety

Board believes that FAA should maintain at least a 48-hour archive of 1-minute RVR

data. Such an archive could be accomplished either by modifying RVR systems or by

interfacing RVR systems with local automated weather systems.

2.3.3.2 Low Level Windshear Alert System

The Low Level Windshear Alert System (LLWAS) alerts about 2339 and 2347

correctly detected actual windshear conditions associated with the gust front and other

wind surges. No windshear alerts were current or were being issued at the time that

flight 1420 touched down.

217

LLWAS centerfield wind sensors are typically mounted between 70 and 100 feet

above field elevation; at Little Rock, the sensor is mounted at 70 feet. ASOS wind

sensors are mounted at a standard height of 32 feet; thus, ASOS wind data may be more

representative of the surface winds that will be present when an airplane is landing. At

the time that flight 1420 landed on runway 4R, the ASOS was measuring the wind from

290° at 16 knots with gusts to 22 knots, resulting in a 5-knot tailwind component upon

touchdown that increased the airplane’s speed on the runway and affected the airplane’s

directional control and braking performance. As stated in section 2.2.2, the controller’s

last wind report of 320° at 23 knots (at 2349:53) would not have indicated to the flight

crew the possibility of a tailwind at touchdown. Thus, the LLWAS centerfield wind

information does not always reflect surface wind conditions, and the difference in height

between LLWAS and ASOS sensors, in some cases, may be critical.

The FAA’s Aeronautical Information Manual (AIM) Section 1, “Meteorology,”

Part 7, “Safety of Flight,” includes only general information on the LLWAS. The

information does not indicate that, in some circumstances, LLWAS centerfield wind

information alone may not accurately represent the winds that are present at the runway

surface. The information also does not caution that the LLWAS alerts at some airports

(including Little Rock) currently do not distinguish between windshear and microburst

events. (A future software change to LLWAS will allow all system models to

differentiate between microburst and windshear alerts, but the only LLWAS systems that

currently make this differentiation are those that are integrated with TDWR systems.

217

The LLWAS first detected winds associated with the microburst at 2351:30 and issued alerts

from 2352:10 to 0005:10.

Analysis 152 Aircraft Accident Report

Airports with such LLWAS systems include Dallas/Fort Worth, Chicago O’Hare, Denver

International Airport, and Atlanta Hartsfield.)

In addition, at the public hearing on this accident, the expert on LLWAS from the

Massachusetts Institute of Technology’s Lincoln Laboratory indicated his concern that

pilots may be disregarding LLWAS alerts and continuing to operate into the terminal area

because they perceive that the alerts are false and that no windshear threat exists. This

situation may be occurring because pilots may not realize that the LLWAS sensors in use

today are not the same as those used in the late 1970s through the late 1980s, which

alerted when normal gusting winds were present. The latest LLWAS sensors include

technologies to reduce such false alerts, yet this information also does not appear in the

AIM.

The Safety Board concludes that, if detailed information on the LLWAS were

contained in the FAA’s AIM, pilots could have a better understanding of the system.

Therefore, the Safety Board believes that the FAA should provide additional information

on the LLWAS in the AIM, including that an LLWAS alert is a valid indicator of

windshear or a microburst.

2.4 Emergency Response

The local controller reported that he called the Aircraft Rescue and Fire Fighting

(ARFF) units on the crash phone about 2352 after several attempts to contact the flight

crew after the airplane landed. The controller indicated the possibility of an accident at

the end of runway 4R but did not specify which end of the runway. The ARFF units

proceeded to the approach end of runway 4R, but the airplane was off the departure end

of the runway. As a result, the ARFF units had to travel back to the taxiway at which they

entered the runway and then proceed to the other end of the runway. The ARFF units

located the airplane about 0003, 11 minutes after the initial call from the local controller.

However, they did not arrive on scene until 5 minutes later, about 0008 (16 minutes after

the initial notification), because they had to travel in the opposite direction to an access

road, turn onto a perimeter road back in the direction of the accident site, stop to

manually unlock a perimeter security gate, and then continue on the perimeter road to the

accident site.

If the ARFF units had known the approximate location of the airplane when they

left the fire station, the time spent traveling from the taxiway to the approach end of the

runway and back would have been saved. The ARFF units reported that they were

initially traveling very slowly because of the limited visibility toward the approach end of

the runway and the unknown location of the airplane. The Safety Board recognizes that

the controller could have provided a more precise description of the accident location to

the ARFF units, especially since he knew the direction in which the airplane was landing

and had seen the airplane travel past midfield. However, the Board also recognizes that

the ARFF personnel could have queried the controller to see if he knew any additional

information about the airplane’s location.

Analysis 153 Aircraft Accident Report

The Safety Board concludes that part of the delay in locating the flight 1420

wreckage was preventable and that several minutes in the emergency response time

might have been saved if the ARFF units had proceeded directly to the departure end of

runway 4R. Because the delay can be partly attributed to the incomplete location

information provided to the ARFF units by the local controller, the Safety Board believes

that the FAA should issue a mandatory briefing item to tower controllers that describes

the circumstances of this accident, including the interactions between the controller and

ARFF crews. This briefing item should emphasize that location information provided to

ARFF crews should be as complete and specific as possible to minimize opportunities for

confusion. The Safety Board also believes that the FAA should amend FAA

Order 7110.65, “Air Traffic Control,” to require controllers to monitor the progress of

ARFF crews responding to emergencies to ensure that the response is consistent with

known location information. In addition, the Safety Board believes that the FAA should

amend FAA Order 7210.3R, “Facility Operation and Administration,” to direct tower

managers to establish mutual annual briefings between ATC and ARFF personnel to

ensure that these personnel have a common understanding of the local airport emergency

plan and sections of the FAA’s AC 150/5210-7C, “Aircraft Rescue and Firefighting

Communications,” that are applicable to local ATC/ARFF emergency response

procedures.

The accident was not survivable for those who were seated on the forward left

side of the airplane in the area of the collisions with the runway 22L approach lighting

structure (the captain and the passengers in seats 3A and 8A) and those who were

immediately exposed to lethal impact forces (seats 17B and 18A and B) or fire (seats

19A, B, and C) in the area where the fuselage separated.

218

The accident, however, was

potentially survivable for the passenger fatalities in seats 27E and 28D.

219

Because the accident was potentially survivable for the passengers in seats 27E

and 28D, the Safety Board considered whether a shorter ARFF response time could have

prevented the fatalities but determined that the passengers’ lives would not have been

218

In a 2001 safety report, the Safety Board defined a survivable accident as follows: “For the

accident to be deemed survivable, the forces transmitted to occupants through their seat and restraint

system cannot exceed the limits of human tolerance to abrupt accelerations, and the structure in the

occupants’ immediate environment must remain substantially intact to the extent that a livable volume

is provided for the occupants throughout the crash.” See National Transportation Safety Board. 2001.

Survivability of Accidents Involving Part 121 U.S. Air Carrier Operations, 1983 Through 2000

. Safety

Report NTSB/SR/01-01. Washington, DC.

219

The passenger in seat 27E died of smoke and soot inhalation. After stating “that’s everyone—

that’s all” in response to another passenger’s query about whether anyone was still inside the cabin,

the passenger in seat 27E continued farther aft in the cabin (for undetermined reasons) and was

overcome by smoke; his body was found in the extreme aft part of the cabin on the right side.

The passenger would have most likely survived if he had evacuated the airplane when he was near

the right aft overwing exit. The passenger in seat 28D received serious burns when she evacuated

the airplane through the left aft overwing exit and died 15 days later of complications from the

burn injuries. This passenger would have likely survived if she had exited the airplane through the

right aft overwing exit or the tailcone exit. The Safety Board recognizes that three other passengers

who used the left aft overwing exit survived the accident, one of whom was burned severely. The

passenger in seat 28D may have experienced a more intense fire than the other passengers because

of the variable winds that were present or the progression of the fire.

Analysis 154 Aircraft Accident Report

saved if emergency responders had arrived on scene earlier. Even with the shortest

possible response time, the passenger in seat 28D would have already received the

second- and third-degree burns to over half of her body and the severe inhalation injury

from which she later died. The passenger in seat 27E remained on the airplane and

therefore needed to be rescued from the wreckage. However, the four ARFF personnel

that responded to the accident were not available to enter the airplane because they were

involved in positioning the fire trucks and operating fire suppression equipment.

220

Thus,

an interior search of the airplane could not be conducted until off-airport firefighters

arrived on scene about 0022.

2.4.1 Aircraft Rescue and Fire Fighting Staffing Levels

The Safety Board could not determine whether the passenger in seat 27E would

have survived if sufficient ARFF personnel had been available to perform a rescue.

However, previous accidents in which the occupants’ survival was aided by or depended

on the abilities of rescue personnel to enter an airplane (see section 1.18.7.1) provide

lessons learned that highlight the need for an adequate number of ARFF personnel to

perform rescue operations.

The FAA’s January 1997 final report, Aircraft Rescue and Firefighting Services—

Mission Response Study, indicated that evacuation of an aircraft was a primary

responsibility of the air carriers and that the carriers have crew complements trained for

that function. This finding concerns the Safety Board because, in the event that

crewmembers are incapacitated or the conditions aboard the airplane deteriorate to the

point that the crew is forced to leave, the remaining airplane occupants must rely on

ARFF personnel to assist in the evacuation. In fact, the first officer and two of the four

flight attendants in the Little Rock accident sustained serious injuries and were unable to

assist with the evacuation.

Title 14 CFR 139.319(j) requires that “sufficient rescue and firefighting personnel

are available during all air carrier operations to operate the vehicles, meet the response

times, and meet the minimum agent discharge rates required by this part.” However, the

regulation does not contain any specific staffing requirement for ARFF units. Thus, the

regulation does not ensure that ARFF units will be staffed at a level that would allow

timely entry into an airplane for rescue and firefighting activities.

Insufficient ARFF staffing levels were demonstrated in two recent events. First,

on October 10, 2000, a Canadair Challenger Model 604, C-FTBZ, owned by Bombardier

Inc., and being operated as a test flight, crashed into terrain and collided with an airport

perimeter fence during a failed takeoff from runway 19 at the Wichita Mid-Continent

Airport, Kansas.

221

A fuel-fed fire erupted after the collision. Two ARFF fire trucks and

220

The Safety Board recognizes that Little Rock airport is now staffed with six ARFF personnel

at all times.

221

The description for this accident, CHI01MA006, can be found on the Safety Board’s Web

site at <http://www.ntsb.gov>.

Analysis 155 Aircraft Accident Report

three ARFF personnel responded within about 90 seconds and applied a mass application

of firefighting agent to extinguish the exterior fire. The firefighters stated that they could

hear screams for help coming from the cockpit. One of the ARFF trucks carried a

“penetrating nozzle”;

222

however, the nozzle could not be used because two trained

firefighters were required to operate it, and only one was available. (Two of the three

personnel were occupied in their vehicles with firefighting activities.) The pilot and flight

test engineer were killed, the copilot received serious injuries and died more than

1 month later, and the airplane was destroyed.

Second, on August 8, 2000, Air Tran flight 913, a DC-9-32, N838AT, made an

emergency landing in Greensboro, North Carolina, because of dense smoke in the

cockpit.

223

The airplane landed successfully, and an emergency evacuation was conducted.

All occupants were able to evacuate the airplane. Four crewmembers received minor

injuries from smoke inhalation in flight, 1 passenger received a minor injury during the

evacuation, and 1 crewmember and 57 passengers were uninjured. As with the flight 1420

emergency response, three ARFF vehicles and four ARFF personnel responded to the Air

Tran event. If the occupants aboard the Air Tran flight had not been able to evacuate, there

would not have been adequate ARFF resources to enter the airplane and rescue

individuals. In fact, no ARFF personnel entered the airplane until after off-airport

emergency responders arrived, despite the fire progressing through the airplane.

The Safety Board concludes that ARFF units may not be staffed at a level that

enables ARFF personnel, upon arrival at an accident scene, to conduct exterior

firefighting activities, an interior fire suppression attack, and a rescue mission. Therefore,

the Safety Board believes that the FAA should amend 14 CFR 139.319(j) to require a

minimum ARFF staffing level that would allow exterior firefighting and rapid entry into

an airplane to perform interior firefighting and rescue of passengers and crewmembers.

2.4.2 Crash Detection and Location Technology

The accident airplane was not equipped with a technology, such as an emergency

locator transmitter (ELT), that might have assisted the controller in directing the ARFF

units to the airplane’s location after it crashed.

224

Also, the ARFF vehicles were not

equipped with the Driver’s Enhanced Vision System (DEVS), which was designed to help

reduce emergency response times in poor visibility conditions such as those experienced

after the flight 1420 crash. The DEVS includes a forward-looking infrared device, which

searches for heat sources. Even though the heavy rain at the airport was cooling the plume

222

A penetrating nozzle is a tool that is used to puncture the skin of a burning airplane and

apply extinguishing agent to the interior of the airplane.

223

The description for this accident, DCA00MA079, can be found on the Safety Board’s Web

site at <http://www.ntsb.gov>.

224

Tower facilities monitor ELT frequencies at all times. The Little Rock ARFF units did not

have an ELT receiver.

Analysis 156 Aircraft Accident Report

of smoke from the postcrash fire, it is likely that the device would have detected the smoke

plume sooner than the ARFF units were able to see it.

225

The Safety Board has investigated other accidents in which the use of crash

detection and location equipment would have significantly helped with the emergency

response effort (see section 1.18.7.2). It is extremely important that ATC facilities

receive immediate information about a downed aircraft and that ARFF units and other

emergency responders arrive at the accident scene in the shortest possible time. ELTs,

DEVS, and other current technologies that can be used to help detect, locate, or respond

to downed aircraft offer the potential for improving emergency response times. The

Safety Board concludes that a crash detection and location technology would help

expedite the arrival of emergency responders at an accident scene, thus maximizing the

possibility for saving lives and reducing the severity of injuries. Therefore, the Safety

Board believes that the FAA should evaluate crash detection and location technologies,

select the most promising candidate(s) for ensuring that emergency responders could

expeditiously arrive at an accident scene, and implement a requirement to install and use

the equipment.

2.4.3 Interagency Emergency Response Critique

Little Rock National Airport did not conduct a postaccident interagency

emergency response critique shortly after the flight 1420 accident. Nine months after the

accident, the airport completed individual critiques with all of the agencies involved with

the emergency response and a group critique with some of these agencies. All of the

agencies involved with the emergency response were invited to attend the group critique;

however, the Little Rock Fire and Police Departments, the Little Rock Office of

Emergency Services, and Metropolitan Emergency Medical Services did not attend. In

addition, although the agenda for the group critique included many areas of discussion

affecting all facets of an emergency response, the only documented information resulting

from these discussions that was provided to the Safety Board was a summary of the

hospitals’ recommendation, observations, and concerns.

The FAA does not currently require airport operators to perform postaccident

emergency response critiques. The Federal Railroad Administration (FRA), however,

requires rail carriers to conduct postaccident emergency response critiques. Specifically,

49 CFR 239.105 requires that “each railroad operating passenger train service shall

conduct a debriefing and critique session after each passenger train emergency situation”

to determine the effectiveness of the railroad’s emergency preparedness plan and amend

or improve the plan according to the information gleaned. The FRA requires the critique

to assess, among other items, how much time elapsed between the emergency situation

and the notification to the emergency responders, whether the emergency responders

arrived quickly on scene after receiving notification, and whether the emergency

225

As stated in section 1.18.4, the FAA requires a forward-looking infrared device to be installed

on all of its new fire trucks that carry 1,500 or more gallons. DEVS is currently in use at Logan

International Airport in Boston, Massachusetts.

Analysis 157 Aircraft Accident Report

response was effective. The FRA further requires that the debriefing and critique session

be conducted within 60 days after the passenger train emergency situation and that the

railroad maintain records of the session (including the names of all the participants) and

make these records available to FRA representatives.

Although a formal postaccident interagency emergency response critique was not

required by 14 CFR Part 139, such a critique, if performed in a timely manner after an

aviation accident, would enable participants to take immediate, appropriate actions to

rectify any identified emergency response deficiencies.

The Safety Board investigated two aviation accidents, 3 years apart, at Dallas/Fort

Worth International Airport that demonstrate how corrective actions implemented after

one accident response prevented a recurrence of the problem during a subsequent

accident response. In the August 2, 1985, Delta flight 191 accident investigation (see

section 1.18.5), the Board determined that, although the on-airport emergency response

was timely and effective and contributed significantly to saving a number of lives, the

amount of time required to complete all of the emergency notifications was excessive

(45 minutes). The Board recommended that the Dallas/Fort Worth Airport Board improve

its Airport Emergency Plan to provide for more efficient and timely notification of the

mutual aid agencies and area hospitals.

On August 31, 1988, Delta flight 1141, a Boeing 727-232, N473DA, crashed

during its takeoff roll.

226

Of the 108 airplane occupants, 14 were killed, 26 were seriously

injured, 50 received minor injuries, and 18 were uninjured. The airplane was destroyed

by impact forces and postcrash fire. In its final report on this accident, the Board found

that the time to complete emergency notifications, including those to the mutual aid

agencies and hospitals, had been significantly reduced (21 minutes). The decreased

notification time was partly attributed to the installation and use of the Automated Voice

Notification System in the airport’s Emergency Operations Center. The Board concluded

that the corrective actions taken by the Dallas/Fort Worth Airport Board after the

flight 191 accident “greatly improved” the communications and coordination of the

ARFF personnel and medical units responding to the flight 1141 accident.

The Safety Board concludes that a timely postaccident interagency emergency

response critique that identifies deficiencies that need corrective action and successes that

should be repeated in similar circumstances would be beneficial for all parties involved in

an aviation accident response. Therefore, the Safety Board believes that the FAA should

develop specific criteria, using the FRA’s requirements as guidance, to be evaluated

during a postaccident interagency emergency response critique and amend 14 CFR

Part 139 to require airport operators to conduct this critique within 60 days after any air

carrier accident and provide the results of the critique to the FAA.

226

For more information, see National Transportation Safety Board. 1989.

Delta Air Lines, Inc.,

Boeing 727-232, N473DA, Dallas/Fort Worth International Airport, Texas, August 31, 1988. Aircraft

Accident Report NTSB/AAR/89-04. Washington, DC.

Analysis 158 Aircraft Accident Report

2.5 Airport Factors

2.5.1 Runway Safety Areas

Runway 4R/22L, which was opened in September 1991, has runway safety areas

of 1,000 feet at the departure end of 22L and 450 feet at the departure end of 4R.

Although the FAA’s June 5, 1991, version of AC 150/5300-13 stated that the standard

runway safety area was 1,000 feet, runway 4R/22L was exempt from this standard under

the provisions of 14 CFR 139.309(a)(1), which allowed runways that had a safety area on

December 31, 1987, to be maintained, as long as no reconstruction or significant

expansion of the runway had begun after January 1, 1988.

227

The Safety Board notes that,

in this accident, an extra 550 feet at the departure end of runway 4R would not have

prevented the airplane from departing the end of the runway or impacting the approach

lighting system; however, the airplane’s speed would have further decreased with an

extra 550 feet at the end of the runway, resulting in a lower impact speed. Because safety

areas of at least 1,000 feet would provide an extra margin of safety under most

circumstances, the Board is concerned about runway safety areas that are less than the

current FAA standard.

Another recent accident involved an overrun beyond the threshold of a runway

with a nonstandard safety area. Specifically, on March 5, 2000, Southwest Airlines

flight 1455, a Boeing 737-300, N668SW, departed the end of runway 8 during landing at

Burbank-Glendale-Pasadena Airport, Burbank, California. The airplane traveled through

a nonfrangible metal blast fence beyond the departure end of the runway and came to rest

on a highway outside the airport perimeter. Of the 142 airplane occupants, 2 received

serious injuries, 42 received minor injuries, and 98 were uninjured.

228

The runway safety

area at the approach (west) end of runway 8/26 is 200 feet; no safety area exists at the

departure (east) end of the runway. As with runway 4R/22L in Little Rock, runway 8/26 in

Burbank was exempt, under 14 CFR 139.309(a)(1), from the 1,000-foot runway safety

area standard in AC 150/5300-13.

On December 13, 1994, the Safety Board issued Safety Recommendation

A-94-211, which asked the FAA, among other things, to require that substandard runway

safety areas be upgraded to AC 150/5300-13 minimum standards wherever possible.

229

its October 15, 1997, response, the FAA indicated that 25 percent of the runways at

14 CFR Part 139 certificated airports have safety areas that do not meet AC 150/5300-13

minimum standards but could with feasible improvements and that 17 percent have safety

areas that could not be feasibly improved to meet the standard. However, the FAA stated

that runway safety area improvement projects would be scheduled only as part of overall

227

Construction of runway 4R/22L began as early as September 1982. Work on the runway

continued after January 1, 1988, but none of the efforts involved reconstruction or significant expansion.

228

The description for this accident, DCA00MA030, can be found on the Safety Board’s Web

site at <http://www.ntsb.gov>.

229

This recommendation was issued as a result of the April 27, 1994, Action Air Charters flight 990

accident in Stratford, Connecticut (see section 1.18.8.1).

Analysis 159 Aircraft Accident Report

runway improvement projects because of the associated cost and infrequency of aircraft

overruns and undershoots. On February 10, 1999, the Board expressed its concern that the

delay in runway safety area upgrades would allow nonstandard conditions to continue and

classified Safety Recommendation A-94-211 “Closed—Unacceptable Action.”

The Safety Board recognizes that the design of some airport runways makes it

difficult for runway safety areas to be upgraded to the standards of AC 150/5300-13 and

that those airport runways that can be upgraded may not be improved for some time

based on the FAA’s current plans. However, those runways should provide equivalent

runway protection. One way to achieve this goal is to install a type of soft-ground aircraft

arresting system, such as the Engineering Materials Arresting System (EMAS). The

safety benefit of EMAS was demonstrated by the American Eagle flight 4925 accident

when the airplane departed an 8,400-foot runway but was stopped approximately 248 feet

into a 400-foot EMAS (see section 1.16.5).

According to a report by Engineered Arresting Systems Corporation, which

developed EMAS, the flight 1420 airplane would not have been significantly slowed by a

standard EMAS installed at the approach end of runway 22L because the airplane’s track

was outside the extended runway edges. Thus, the flight 1420 airplane would not have

been able to use the full length of a standard EMAS. However, according to FAA

AC 150/5220-22, “Engineered Materials Arresting System for Aircraft Overruns,” most

airplane runway overruns “come to rest within 1,000 feet of the runway end and between

the extended edges of the runway.” Therefore, the Safety Board continues to support the

installation of EMAS, especially for those runways in which the safety area is less than

the minimum standards established in AC 150/5300-13. The Board notes that an EMAS

was installed at the departure end of runway 4R at Little Rock in the fall of 2000. The

Board further notes that Little Rock airport is working with Federal and local government

agencies to extend the runway safety area at the departure end of runway 4R to 1,000 feet

by July 2002.

2.5.2 Nonfrangible Structures

The FAA determined that the runway 22L approach lighting system at Little

Rock, which is located in a flood plain area of the Arkansas River, could not be retrofitted

to a frangible design because of the possibility that moving water, ice, and floating debris

would affect the structural integrity of the system. The Safety Board recognizes the

current design limitations of this approach lighting system and acknowledges that, if the

approach lighting system had been frangible, it is possible that the accident airplane

would not have been stopped on the ground and would have gone into the Arkansas

River. However, the Board also recognizes that frangible structures, because of their

ability to break, distort, or yield on impact with aircraft, generally present less risk than

nonfrangible ones. In this accident, the airplane’s collision with the nonfrangible

approach lighting system was the direct cause of the fatal blunt force trauma injuries

sustained by the captain and the passengers in seats 3A and 8A and the destruction to the

airplane on the left side of the fuselage.

Analysis 160 Aircraft Accident Report

In 1984, the Safety Board issued Safety Recommendation A-84-36, which asked

the FAA to “initiate research and development activities to establish the feasibility of

submerged low-impact resistance support structures for airport facilities and promulgate

a design standard if such structures are found to be practical.” The FAA conducted

research in this area with the National Institute of Standards and Technology. In

October 1996, the FAA concluded that, with the current technology, any submerged

low-impact frangible structure would most likely be destroyed by wave motion from

small storms. Because of the FAA’s research activities, the Board classified the

recommendation “Closed—Acceptable Action.”

The FAA has had an effort underway for some time to replace selected

nonfrangible structures with frangible ones. However, technological advances since the

time of the FAA/National Institute of Standards and Technology research activities,

especially those involving the use of new materials, might allow some additional

nonfrangible structures to be replaced by frangible ones. The Safety Board concludes that

the development of recent technologies to convert nonfrangible structures to frangible

ones would provide a safety benefit to airport facilities. Therefore, the Safety Board

believes that the FAA should conduct research activities to determine if recent

technological advances would enable submerged low-impact structures and other

nonfrangible structures at airports to be converted to frangible ones.

2.6 American Airlines

2.6.1 Stabilized Approach Criteria

At the time of the accident, American’s only written guidance for MD-80 pilots

regarding the stabilized approach concept was in the “Techniques” section of the DC-9

Operating Manual. The guidance indicated that the minimum recommended stabilized

approach altitudes for instrument flight rules (IFR) and visual flight rules (VFR)

conditions were 1,000 and 500 feet, respectively, and that landing flaps were to be selected

by 1,000 feet afl. The guidance also stated that, before descending below the specified

minimum stabilized approach altitude, the airplane was to be in the final landing

configuration (gear down and final flaps), on approach speed, on the proper flightpath, at

the proper sink rate, and at stabilized thrust; these conditions were expected to be

maintained throughout the rest of the approach. However, the guidance did not define

what was meant by “on” approach speed, “on” the proper flightpath, and “at” the proper

sink rate. In addition, the guidance did not describe the necessary flight crew actions if the

stabilized approach criteria were not met. Further, information presented in the

“Techniques” section was not considered by American to be required procedures but

rather suggested ways of accomplishing a task.

230

The Safety Board notes that American revised its stabilized approach criteria after

the accident and included this information in both its Airplane Flight Manual and DC-9

Operating Manual. The revised company procedures state that the airplane must be “at

approach speed (V

ref

plus additives)” rather than “on approach speed.” The procedures

Analysis 161 Aircraft Accident Report

also state that the minimum stabilized approach height is 1,000 feet afl in instrument

meteorological conditions and 500 feet afl in visual meteorological conditions rather than

present minimum recommended stabilized approach altitudes for IFR and VFR

conditions. In addition, the procedures explicitly state that a go-around is required if

stabilized approach requirements cannot be maintained until landing.

At the public hearing on this accident, the first officer discussed the training and

guidance that he had received at American concerning a decision to execute a go-around.

The first officer indicated that this decision was based on the stabilized approach theory.

He stated that, if the sink rate was excessive or if the airplane was deviated to the left or

right of course (among other criteria), then the approach would not meet the stabilized

approach definition and a go-around should be performed.

American’s flight manual contains the only written guidance regarding the

performance of a missed approach. At the time of the accident, the manual stated that,

when a landing cannot be accomplished, the pilot must comply with the missed approach

procedure, or an alternate missed approach procedure specified by ATC, upon reaching

the missed approach point defined on the approach chart. The missed approach

procedures were revised on August 15, 1999, to state that American Airlines has a

“no-fault go-around policy” and recognize that a successful approach can end in a missed

approach. The revised procedures require captains to execute or order a missed approach

if the aircraft is not stabilized by 1,000 feet afl (in IFR conditions) or 500 feet afl (in VFR

conditions) or if the captain believes that a safe landing cannot be accomplished within

the touchdown zone or that the airplane cannot be stopped within the available runway

length.

The Safety Board acknowledges that American revised its missed approach

procedures to ensure that its pilots are aware that they will not be faulted if they perform

a missed approach and that the revised procedures specify the altitude at which a

go-around is required if an approach is not stabilized. However, the new stabilized

approach requirements still do not explicitly state what is meant by “proper” flightpath

and “proper” sink rate. Further, the requirements do not provide pilots with specific

amounts of glideslope and localizer displacement for determining whether an approach is

stabilized. The new requirements also do not contain specific, corrective actions for pilots

to take if an approach becomes unstabilized before the required minimum stabilized

approach heights.

231

Thus, the Safety Board concludes that American Airlines has

230

The Safety Board recognizes that American plans to integrate the “Techniques” section with

the “Normals” section of the DC-9 Operating Manual, which contains required, rather than suggested,

procedures for accomplishing tasks. However, American did not provide any timetable for combining

the two sections, stating at the public hearing that it was “gradually” editing out the “Techniques”

section. Thus, the minimal stabilized approach guidance that does exist may remain only as a suggested

procedure for some time. Even after the information is placed in the “Normals” section, it will still

lack the necessary specificity to assist pilots in recognizing an unstabilized approach.

231

In the February 6, 1997, American Airlines flight 699 accident in St. John’s, Antigua, the

Safety Board determined that a contributing factor in the accident was American’s inadequate procedures

to address corrective actions for approaches that become unstabilized. The description for this accident,

DCA97LA027, can be found on the Safety Board’s Web site at <http://www.ntsb.gov>.

Analysis 162 Aircraft Accident Report

insufficient guidance to assist its pilots in performing a stabilized approach and

recognizing when an approach has become unstabilized.

On August 29, 1997, the Safety Board issued Safety Recommendation A-97-85,

which asked the FAA to require all 14 CFR Part 121 and 135 operators to review and

revise their company operations manuals to more clearly define terms that are critical for

safety-of-flight decision-making, such as “stabilized approach.”

232

On May 26, 1998, the

FAA issued Flight Standards Handbook Bulletin for Air Transportation (HBAT) 98-22,

“Stabilized Approaches,” which directed 14 CFR Part 121 and 135 POIs to review

operators’ training and operations manuals to ensure that they addressed, among other

things, the minimum requirements for a stabilized approach and the immediate actions

that needed to be taken if the stabilized approach conditions were not met. On the basis of

the FAA’s actions, the Safety Board classified Safety Recommendation A-97-85

“Closed—Acceptable Action” on November 20, 1998.

The Safety Board is concerned that, even with the requirement for POIs to ensure

that their carrier’s stabilized approach guidance is in accordance with Flight Standards

HBAT 98-22, some carriers may still have stabilized approach guidance that lacks

specificity (as demonstrated by American’s revised MD-80 stabilized approach

guidance.) In addition, guidance to air carriers on stabilized approach criteria in the

FAA’s Air Transportation Operations Inspector’s Handbook is not sufficiently detailed to

ensure that carriers provide their pilots with defined guidelines for determining a

stabilized approach and deciding when a missed approach is necessary. The Safety Board

concludes that, because a stabilized approach is a critical part of safe flight operations, it

is imperative that air carriers have specific stabilized approach criteria. Therefore, the

Safety Board believes that the FAA should define detailed parameters for a stabilized

approach, develop detailed criteria indicating when a missed approach should be

performed, and ensure that all 14 CFR Part 121 and 135 carriers include this information

in their flight manuals and training programs.

2.6.2 Spoiler System Training

During observations of day 6 (takeoffs and landings) of American Airlines’

MD-80 simulator training sessions, spoiler system training deficiencies were noted. One

session presented no landings in which the spoilers had failed to deploy. The other

session presented seven landings in which the spoilers had failed to deploy, but none of

the trainees had called out that the spoilers had not extended. As indicated in section

2.2.2.1.3, American’s MD-80 Fleet Manager at the time of the accident and several

company MD-80 check airmen indicated that pilots were trained to announce if the

spoilers failed to automatically extend, yet the instructor did not point out to the students

that they had failed to make this announcement.

232

This recommendation was issued as a result of the October 19, 1996, Delta Air Lines flight

554 accident in New York (see section 1.18.8.2).

Analysis 163 Aircraft Accident Report

The spoilers were manually deployed in only two of the failed spoiler landings,

but a first officer trainee was the pilot who extended the spoilers. This action was

contrary to American’s policy, which indicated that the captain was to extend the spoilers

if they failed to deploy, regardless of which pilot was making the landing.

Further evidence of problems with American’s spoiler system training came to

light after the Palm Springs incident. During the 8 months in between the Little Rock

accident and the Palm Springs incident, company managers had been aware of the spoiler

deployment problems experienced by the Little Rock pilots. In fact, the former MD-80

Fleet Manager testified at the public hearing that spoiler system training now included

recognizing the lack of spoiler extensions and performing the appropriate response.

The Palm Springs captain had completed MD-80 captain upgrade training in

November 1999, during which time he would have attended day 6 of the simulator

training and should have been instructed on the revised simulator spoiler training

procedures. The Palm Springs first officer had completed MD-80 recurrent training in

August 1999, during which time he should have also been instructed on the revised

simulator spoiler training procedures. However, both flight crewmembers in the Palm

Springs incident failed to verify that the spoilers had automatically deployed after

landing, and the captain failed to manually extend the spoilers when they did not deploy.

The Safety Board recognizes that, on February 23, 2000 (1 week after the Palm Springs

incident), American revised its DC-9 Operating Manual to prevent a situation similar to

the Little Rock and Palm Springs events from recurring. The revisions specified that the

nonflying pilot was responsible for making a “deployed” or “no spoilers” callout at

touchdown and that the captain was always responsible for deploying the spoilers if they

did not extend after touchdown.

2.6.3 Spoiler and Braking Systems Procedures

At the time of the accident, American had not adopted Boeing’s MD-80 spoiler

deployment and autobrake procedures. Boeing’s procedures recommended a “no

spoilers” callout by the nonflying pilot if the spoiler handle did not move aft after

touchdown and the use of maximum autobrakes for landings on wet or slippery runways.

As stated in section 2.2.2.1.3, the flight crew’s failure to detect that the spoilers had not

deployed at touchdown might have been avoided if a procedure similar to Boeing’s had

been in place at American at the time. As stated in section 2.2.2.3, if the spoilers had

deployed and the flight crew had selected maximum autobrakes for the landing, initial

brake application could have occurred about 4 seconds sooner.

At the public hearing on this accident, Boeing’s MD-80 Chief Pilot for Flight

Operations indicated that operators are not required to adopt the operating procedures in

Boeing’s FCOM and that operators, along with their POIs, can change the procedures.

The Chief Pilot also indicated that most, if not all, of the domestic operators coordinate

with the manufacturer—normally by requesting a letter of no technical objection—before

making any changes to their flight manual, even though there is no legal requirement to

do so. In an April 20, 2000, letter to the Safety Board, Boeing indicated that it could not

Analysis 164 Aircraft Accident Report

find any record of issuing a letter of no technical objection to American regarding

alteration of the manufacturer’s recommended spoiler deployment or autobrake

procedures.

On November 30, 1998, Safety Recommendation A-98-102 was issued because

air carriers had the prerogative not to adopt certain manufacturer procedures without

clear written justification.

233

Safety Recommendation A-98-102 asked the FAA to “require

air carriers to adopt the operating procedures contained in the manufacturer’s airplane

flight manual and subsequent approved revisions or provide written justification that an

equivalent safety level results from an alternate procedure.” In response to this

recommendation, the FAA issued, in May 1999, the Joint Flight Standards HBAT,

Airworthiness, and General Aviation, Flight Standards Policy—Company Operating

Manuals and Company Training Program Revisions for Compliance. The handbook

bulletin directed that POIs encourage their operators to (1) have a reliable delivery system

in place for flight manual revisions, which ensures that the operators receive the revisions

within 30 calendar days of approval, and (2) develop an action plan to notify, in writing,

respective POIs of new flight manual revisions within 15 days after receipt.

In addition, on July 7, 2000, the FAA stated that it had initiated an NPRM

proposing to revise 14 CFR Part 121, Subparts N and O, to reflect the policy included in

the May 1999 Joint Flight Standards HBAT, Airworthiness, and General Aviation, Flight

Standards Policy—Company Operating Manuals and Company Training Program

Revisions for Compliance. On January 12, 2001, the Safety Board acknowledged the

FAA’s actions and stated that, pending the issuance of the NPRM and implementation of

the proposed regulation, Safety Recommendation A-98-102 was classified “Open—

Acceptable Response.” On August 2, 2001, the FAA stated that it was continuing to

develop the NPRM.

At the public hearing on this accident, the POI for American indicated that a

carrier might choose not to make a manufacturer’s suggested change because of the way

that the carrier has configured the particular airplane. However, it is critical that the

carrier provide written justification to the FAA regarding the reasons for not making a

change or for implementing an alternative procedure in case the manufacturer’s

performance data do not support the carrier’s justification. It is also critical that the

carrier make its POI and respective aircrew program manager (APM) aware of any

manufacturer’s recommended procedure that is not being adopted or is being altered.

The Safety Board recognizes that American, since the time of the accident, has

revised its DC-9 Operating Manual to include spoiler deployment and autobrake

procedures similar to Boeing’s. The Board further recognizes that the FAA has taken

positive steps toward implementing the intent of this recommendation. However, this

accident highlights the need for timely action to ensure that pilots are operating airplanes

according to procedures that reflect the manufacturer’s safest operating practices.

Therefore, the Safety Board reiterates Safety Recommendation A-98-102.

233

This recommendation was issued as a result of the January 9, 1997, Comair flight 3272 accident

near Monroe, Michigan (see section 1.18.8.3).

Analysis 165 Aircraft Accident Report

2.7 Federal Aviation Administration Oversight

Within the FAA’s Certificate Management Office for the AMR Corporation, the

POI for American Airlines is responsible for the overall oversight of American’s training

and line operations and for approving training and flight manuals and their revisions. At

the public hearing, the POI stated that he needed more inspectors to conduct surveillance

but that a hiring freeze was in effect. The POI also indicated that he needed almost double

the number of air safety inspectors he had in his office at the time and that his inability to

hire more inspectors had severely impacted his office’s surveillance activities. Further,

the POI expressed concern that his office did not have geographic air safety inspectors in

some locations where American has a large volume of operations.

An APM and an Assistant APM were responsible for providing oversight of

American’s MD-80 fleet. This oversight responsibility included reviewing and approving

the contents of the flight manuals, monitoring the training program, reviewing

recommended changes to the manuals and training program, and monitoring daily line

operations. The APM indicated that his ability to oversee American’s MD-80 fleet was

affected by personnel constraints. For example, the APM stated that, because he and the

Assistant APM were the only ones responsible for observing American’s MD-80

training, they were “spread thin” and found it “extremely difficult” to observe all facets

of the training program. In addition, the APM stated that budget constraints had limited

the amount of oversight that could be performed and the locations where it could be

performed.

The APM said that a thorough job of oversight required many personnel, which

he did not have. According to the APM, selected check airmen from American Airlines,

named aircrew program designees, performed most of the airman certification activities

under the APM’s supervision. Also, senior designated examiners at American were

responsible for observing every company check airman and simulator instructor at least

once a year. Further, the APM relied heavily on American Airlines to ensure

standardization in its simulator training program.

It is clear that the FAA’s oversight responsibilities for American’s MD-80

training were highly dependent on the work of American’s check airmen, aircrew

program designees, and senior designated examiners. Although the Safety Board

acknowledges that American and the FAA’s Certificate Management Office for the AMR

Corporation have developed a cooperative working relationship with each other, the

Board is concerned about the lack of direct FAA oversight of American’s MD-80 fleet.

The problems found during observations of American’s simulator training sessions (for

example, the students’ use of reverse thrust above 1.3 EPR on wet runways and failure to

notice the lack of spoiler extension) might have been detected earlier if the FAA had been

directly monitoring the training.

In its final report on the USAir flight 1016 accident, the Safety Board expressed

its concern about the relationship between USAir and the FAA POI for USAir.

Specifically, the Board faulted the POI for relying entirely on USAir to rectify a situation

Analysis 166 Aircraft Accident Report

in which many pilots were not in compliance with a standard operating procedure. The

Board stated that “overreliance on the air carrier to carry out its [the FAA’s]

responsibility could limit the POI’s ability to maintain an adequate oversight program

and monitor the operation for noncompliance.” The flight 1420 accident has brought

attention to another circumstance of FAA overreliance on a carrier to perform oversight

activities. Independent oversight for American’s MD-80 fleet is necessary, especially

because three-quarters of the new upgrade captains and one-half of the new hire pilots are

assigned to the MD-80 fleet, according to the APM.

The Safety Board concludes that effective FAA oversight of American Airlines’

MD-80 flight training and flight operations has not occurred. In light of the comments

made by the POI for American at the public hearing, the Board is concerned that similar

oversight problems might be occurring in the company’s other fleets. Therefore, the

Safety Board believes that the FAA should provide additional personnel to accomplish

direct oversight of American Airlines’ flight training and flight operations and include

the POI for American in decisions regarding where these personnel are to be placed. In

addition, the Board encourages the FAA to review oversight staffing levels at all Part 121

and 135 carriers and make appropriate changes to ensure that effective oversight of flight

training and flight operations is occurring.

167 Aircraft Accident Report

3. Conclusions

3.1 Findings

1. The captain and the first officer of American Airlines flight 1420 were properly

certificated and qualified under Federal and company requirements. No evidence

indicated any preexisting medical or behavioral conditions that might have adversely

affected the flight crew’s performance during the accident flight.

2. The accident airplane was properly certified, equipped, and maintained in accordance

with Federal regulations and approved company procedures. No evidence indicated

preexisting engine, system, or structural failures.

3. During the descent into the terminal area, the flight crewmembers could have

reasonably believed that they could reach the airport before the thunderstorm.

4. Because the first officer was able to maintain visual contact with the runway as the

airplane was vectored for the final approach course, both flight crewmembers might

still have believed that flight 1420 could arrive at the airport before the thunderstorm.

5. When the second windshear alert was received, the flight crew should have

recognized that the approach to runway 4R should not continue because the

maximum crosswind component for conducting the landing had been exceeded.

6. Because of the flight crew’s failure to adequately prepare for the approach and the

rapidly deteriorating weather conditions, the likelihood of safely completing the

approach was decreasing, and the need to take a different course of action was

progressively increasing; as a result, the flight crew should have abandoned the

approach.

7. Dynamic or reverted rubber hydroplaning did not occur during the accident airplane’s

landing rollout.

8. The autospoiler system operated properly, and the spoilers did not automatically

deploy because the spoiler handle was not armed by either pilot before landing.

9. A high level of operational redundancy should exist to ensure that spoiler arming has

been completed before landing.

10. The flight crew failed to verify that the spoilers had automatically deployed after

landing, and the captain failed to manually extend the spoilers when they did not

deploy.

Conclusions 168 Aircraft Accident Report

11. Because spoiler deployment is critical for optimal landing performance, procedures to

ensure that the spoilers have deployed after touchdown should be a required part of all

air carriers’ landing operations.

12. The lack of spoiler deployment led directly to the flight crew’s problems in stopping

the airplane within the remaining available runway length and maintaining directional

control of the airplane on the runway.

13. The use of reverse thrust at levels greater than 1.3 engine pressure ratio significantly

reduced the effectiveness of the airplane’s rudder and vertical stabilizer and resulted

in further directional control problems on the runway.

14. The maximum reverse thrust for MD-80 landings on wet or slippery runways should

be 1.3 engine pressure ratio, except when directional control can be sacrificed for a

marginal increase in deceleration.

15. Automatic brake systems reduce pilot workload during landings in wet, slippery, or

high crosswind conditions.

16. The lack of spoiler deployment was the single most important factor in the flight

crew’s inability to stop the accident airplane within the available runway length.

17. The flight crewmembers’ performance during the accident flight was degraded, as

evidenced by their operational errors and impaired decision-making.

18. The flight crewmembers’ focus on expediting the landing because of the impending

weather contributed to their degraded performance.

19. Aircraft penetration of thunderstorms occurs industry-wide.

20. The flight crew’s degraded performance was consistent with known effects of fatigue.

21. The local controller provided appropriate, pertinent, and timely weather information

to the flight crew regarding the conditions on approach to and at the airport.

22. If near-real-time color weather radar showing precipitation intensity were available, it

would provide air traffic controllers with improved representation of weather

conditions in their areas of responsibility.

23. The ability of flight dispatchers to provide timely and accurate weather support would

be enhanced if they had access to Terminal Doppler Weather Radar information at

airports where it is available and Weather Systems Processor information when the

system becomes available.

24. Center Weather Service Units should be staffed at all times when any significant

weather is predicted to affect their areas of operation, even if the weather is predicted

to occur before or after normal operating hours.

Conclusions 169 Aircraft Accident Report

25. The Automated Surface Observing System lockout period can prevent the relay of

critical weather information to flight crews.

26. Runway visual range data should be directly reported to automated weather systems.

27. The current 2-hour runway visual range archiving capability is inadequate to ensure

that data can be preserved for future use.

28. If detailed information on the Low Level Windshear Alert System were contained in

the Federal Aviation Administration’s Aeronautical Information Manual, pilots

could have a better understanding of the system.

29. Part of the delay in locating the flight 1420 wreckage was preventable, and several

minutes in the emergency response time might have been saved if the Aircraft

Rescue and Fire Fighting units had proceeded directly to the departure end of runway

4R.

30. Aircraft Rescue and Fire Fighting (ARFF) units may not be staffed at a level that

enables ARFF personnel, upon arrival at an accident scene, to conduct exterior

firefighting activities, an interior fire suppression attack, and a rescue mission.

31. A crash detection and location technology would help expedite the arrival of

emergency responders at an accident scene, thus maximizing the possibility for

saving lives and reducing the severity of injuries.

32. A timely postaccident interagency emergency response critique that identifies

deficiencies that need corrective action and successes that should be repeated in

similar circumstances would be beneficial for all parties involved in an aviation

accident response.

33. The development of recent technologies to convert nonfrangible structures to

frangible ones would provide a safety benefit to airport facilities.

34. American Airlines has insufficient guidance to assist its pilots in performing a

stabilized approach and recognizing when an approach has become unstabilized.

35. Because a stabilized approach is a critical part of safe flight operations, it is

imperative that air carriers have specific stabilized approach criteria.

36. Effective Federal Aviation Administration oversight of American Airlines’ MD-80

training and line operations has not occurred.

3.2 Probable Cause

The National Transportation Safety Board determines that the probable causes of

this accident were the flight crew’s failure to discontinue the approach when severe

Conclusions 170 Aircraft Accident Report

thunderstorms and their associated hazards to flight operations had moved into the airport

area and the crew’s failure to ensure that the spoilers had extended after touchdown.

Contributing to the accident were the flight crew’s (1) impaired performance

resulting from fatigue and the situational stress associated with the intent to land under the

circumstances, (2) continuation of the approach to a landing when the company’s

maximum crosswind component was exceeded, and (3) use of reverse thrust greater than

1.3 engine pressure ratio after landing.

171 Aircraft Accident Report

4. Recommendations

As a result of the investigation of this accident, the National Transportation Safety

Board makes the following recommendations:

—To the Federal Aviation Administration

For all 14 Code of Federal Regulations Part 121 and 135 operators of

airplanes equipped with automatic spoiler systems, require dual

crewmember confirmation before landing that the spoilers have been

armed, and verify that these operators include this procedure in their flight

manuals, checklists, and training programs. (A-01-49)

For all 14 Code of Federal Regulations Part 121 and 135 operators, require

a callout if the spoilers do not automatically or manually deploy during

landing and a callout when the spoilers have deployed, and verify that these

operators include these procedures in their flight manuals, checklists, and

training programs. The procedures should clearly identify which pilot is

responsible for making these callouts and which pilot is responsible for

deploying the spoilers if they do not automatically or manually deploy.

(A-01-50)

Issue a flight standards information bulletin that requires the use of

1.3 engine pressure ratio as the maximum reverse thrust power for MD-80

series airplanes under wet or slippery runway conditions, except in an

emergency in which directional control can be sacrificed for decreased

stopping distance. (A-01-51)

Require principal operations inspectors of all operators of MD-80 series

airplanes to review and determine that these operators’ flight manuals and

training programs contain information on the decrease in rudder

effectiveness when reverse thrust power in excess of 1.3 engine pressure

ratio is applied. (A-01-52)

Require all operators of MD-80 series airplanes to require a callout if

reverse thrust power exceeds the operators’ specific engine pressure ratio

settings. (A-01-53)

For all 14 Code of Federal Regulations Part 121 and 135 operators, require

the use of automatic brakes, if available and operative, for landings during

wet, slippery, or high crosswind conditions, and verify that these operators

include this procedure in their flight manuals, checklists, and training

programs. (A-01-54)

Recommendations 172 Aircraft Accident Report

Establish a joint Government-industry working group to address,

understand, and develop effective operational strategies and guidance to

reduce thunderstorm penetrations, and verify that these strategies and

guidance materials are incorporated into air carrier flight manuals and

training programs as the strategies become available. The working group

should focus its efforts on all facets of the airspace system, including

ground- and cockpit-based solutions. The near-term goal of the working

group should be to establish clear and objective criteria to facilitate

recognition of cues associated with severe convective activity and

guidance to improve flight crew decision-making. (A-01-55)

Incorporate, at all air traffic control facilities, a near-real-time color

weather radar display that shows detailed precipitation intensities. This

display could be incorporated by configuring existing and planned

Terminal Doppler Weather Radar or Weather Systems Processor systems

with this capability or by procuring, within 1 year, a commercial computer

weather program currently available through the Internet or existing

stand-alone computer hardware that displays the closest single-site

Weather Surveillance Radar 1988 Doppler data or regional mosaic images.

(A-01-56)

Provide U.S. air carriers operating under 14 Code of Federal Regulations

Part 121 access to Terminal Doppler Weather Radar, at airports where the

system is available, and access to the Weather Systems Processor, when it

becomes available, so that their flight dispatch offices can use this

information in planning, releasing, and following flights during periods in

which hazardous weather might impact safety of flight. (A-01-57)

In cooperation with the National Weather Service, ensure that Center

Weather Service Units are adequately staffed at all times when any

significant weather is forecast. (A-01-58)

Modify automated weather systems to accept runway visual range (RVR)

data directly from RVR sensors. (A-01-59)

Maintain at least a 48-hour archive of 1-minute runway visual range data.

(A-01-60)

Provide additional information on the Low Level Windshear Alert System

(LLWAS) in the Aeronautical Information Manual, including that an

LLWAS alert is a valid indicator of windshear or a microburst. (A-01-61)

Recommendations 173 Aircraft Accident Report

Issue a mandatory briefing item to tower controllers that describes the

circumstances of this accident, including the interactions between the

controller and Aircraft Rescue and Fire Fighting (ARFF) crews. This

briefing item should emphasize that location information provided to

ARFF crews should be as complete and specific as possible to minimize

opportunities for confusion. (A-01-62)

Amend Federal Aviation Administration Order 7110.65, “Air Traffic

Control,” to require controllers to monitor the progress of Aircraft Rescue

and Fire Fighting crews responding to emergencies to ensure that the

response is consistent with known location information. (A-01-63)

Amend Federal Aviation Administration (FAA) Order 7210.3R, “Facility

Operation and Administration,” to direct tower managers to establish

mutual annual briefings between air traffic control (ATC) and Aircraft

Rescue and Fire Fighting (ARFF) personnel to ensure that these personnel

have a common understanding of the local airport emergency plan and

sections of the FAA’s Advisory Circular 150/5210-7C, “Aircraft Rescue

and Firefighting Communications,” that are applicable to local

ATC/ARFF emergency response procedures. (A-01-64)

Amend 14 Code of Federal Regulations 139.319(j) to require a minimum

Aircraft Rescue and Fire Fighting staffing level that would allow exterior

firefighting and rapid entry into an airplane to perform interior firefighting

and rescue of passengers and crewmembers. (A-01-65)

Evaluate crash detection and location technologies, select the most

promising candidate(s) for ensuring that emergency responders could

expeditiously arrive at an accident scene, and implement a requirement to

install and use the equipment. (A-01-66)

Develop specific criteria, using the Federal Railroad Administration’s

requirements as guidance, to be evaluated during a postaccident

interagency emergency response critique, and amend 14 Code of Federal

Regulations Part 139 to require airport operators to conduct this critique

within 60 days after any air carrier accident and provide the results of the

critique to the Federal Aviation Administration. (A-01-67)

Conduct research activities to determine if recent technological advances

would enable submerged low-impact structures and other nonfrangible

structures at airports to be converted to frangible ones. (A-01-68)

Define detailed parameters for a stabilized approach, develop detailed

criteria indicating when a missed approach should be performed, and

ensure that all 14 Code of Federal Regulations Part 121 and 135 carriers

include this information in their flight manuals and training programs.

(A-01-69)

Recommendations 174 Aircraft Accident Report

Provide additional personnel to accomplish direct oversight of American

Airlines’ flight training and flight operations, and include the principal

operations inspector for American in decisions regarding where these

personnel are to be placed. (A-01-70)

—To the National Weather Service

In cooperation with the Federal Aviation Administration, ensure that

Center Weather Service Units are adequately staffed at all times when any

significant weather is forecast. (A-01-71)

Eliminate the Automated Surface Observing System lockout feature as

soon as possible. (A-01-72)

In addition, the Safety Board reiterates the following recommendations to the

Federal Aviation Administration:

Establish within 2 years

234

scientifically based hours-of-service regulations

that set limits on hours of service, provide predictable work and rest

schedules, and consider circadian rhythms and human sleep and rest

requirements (A-99-45)

Require air carriers to adopt the operating procedures contained in the

manufacturer’s airplane flight manual and subsequent approved revisions

or provide written justification that an equivalent safety level results from

an alternate procedure. (A-98-102)

234

As previously stated, because the 2-year timeframe specified in this recommendation has already

expired, the Safety Board urges the FAA to expedite its efforts to accomplish this recommendation.

BY THE NATIONAL TRANSPORTATION SAFETY BOARD

MARION C. BLAKEY

Chairman

JOHN A. HAMMERSCHMIDT

Member

CAROL J. CARMODY

Vice Chairman

JOHN J. GOGLIA

Member

GEORGE W. BLACK, JR.

Member

Adopted: October 23, 2001

175 Aircraft Accident Report

5. Appendixes

Appendix A

Investigation and Hearing

Investigation

The National Transportation Safety Board was initially notified of this accident

on June 2, 1999, about 0115 eastern daylight time. A full go-team was assembled and

departed about 0430 eastern daylight time from Ronald Reagan National Airport in

Washington, D.C., for Little Rock. The team and arrived on scene about 0700 eastern

daylight time (0600 central daylight time). Accompanying the team to Little Rock was

Board Member George Black.

The following investigative teams were formed: Aircraft Operations, Human

Performance, Aircraft Structures, Aircraft Systems, Powerplants, Maintenance Records,

Air Traffic Control, Meteorology, Aircraft Performance, Survival Factors, and

Airport/Search/Fire/Rescue. Specialists were also assigned to conduct the readout of the

FDR and transcribe the CVR in the Safety Board’s laboratory in Washington, D.C.

Parties to the investigation were the Federal Aviation Administration (FAA),

American Airlines, the Boeing Commercial Airplane Group, Pratt & Whitney Engines,

the Allied Pilots Association, the Association of Professional Flight Attendants, the

National Air Traffic Controllers Association, the National Weather Service, Little Rock

National Airport, and Little Rock Fire Department.

Public Hearing

The Safety Board held a public hearing on this accident on January 26 through 28,

2000, in Little Rock. Former Chairman Jim Hall presided over the public hearing. The

issues presented at the hearing were flight crew decision-making, availability and

dissemination of weather data, aircraft performance, passenger safety and emergency

response, runway overrun protection, American Airlines’ operational practices and

procedures, American’s internal oversight, and the FAA’s oversight of American.

Parties to the public hearing were the FAA, American, Boeing, the Allied Pilots

Association, the Association of Professional Flight Attendants, the National Weather

Service, Little Rock National Airport, and Little Rock Fire Department.

176 Aircraft Accident Report

Appendix B

Cockpit Voice Recorder Transcript

The following is the transcript of the Fairchild A-100A cockpit voice recorder

serial number 53282, installed on American Airlines flight 1420, a McDonnell Douglas

MD-82, N215AA, that overran the end of the runway at Little Rock National Airport, on

June 1, 1999.

Note 1: Times are expressed in central daylight time (CDT).

Note 2: Generally only radio transmissions to and from the accident aircraft were transcribed.

Note 3: Words shown with excess vowels, letters, or drawn out syllables are a phonetic representation of the words as

spoken.

LEGEND

RDO

Radio transmission from accident aircraft

CAM

Cockpit area microphone voice or sound source

PA voice transmitted over aircraft public address system

INT

Voice transmitted over aircraft interphone system

CTR

Radio transmission from Little Rock center controller

APR

Radio transmission from the Little Rock approach/tower controller

-1

Voice identified as Pilot-in-Command (PIC)

-2

Voice identified as Co-Pilot (SIC)

-3

Voice identified as 1

female flight attendant

-4

Voice identified as 2

female flight attendant

-5

Voice identified as aircraft mechanical voice

Voice unidentified

Unintelligible word

Non-pertinent word

Expletive

- - - Break in continuity

( )

Questionable insertion

[ ]

Editorial insertion

.....

Pause

Appendix B 177 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1119:44

START of RECORDING

START of TRANSCRIPT

1119:53

CAM-2 I warmed it up pretty good.

1119:55

CAM-1 aah.

1119:56

CAM-2 like it. ** .…

1119:56

CAM-1 actually, it's getting pretty hot.

1120:07

CAM-2 did they complain about the temperature?

1120:09

CAM-1 naw, it's getting warm up here.

1120:11

CAM-2 ah, okay.

1120:15

CTR American fourteen twenty, are you gonna want still want

lower?

1120:18

CAM-1 ah, so far it's okay.

1120:20

RDO-2 so far so good ma'am. fourteen twenty we'll let you know.

1120:24

CTR right.

Appendix B 178 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1120:52

CAM-2 twenty five for twenty four. set and armed.

1121:01

CAM-2 this stuff is working out pretty well. * get ahead of that stuff.

1121:45

CAM-1 **, we're almost down to max landing weight.

1121:55

CAM-1 we'll be there.

1121:57

CAM-2 yeah.

1122:17

CAM-2 you want to use one thirty, right?

1122:19

CAM-1 yeah, well. I don't know. we've got a hundred miles to go.

yeah, I guess so.

1122:32

CAM-1 and we'll use flaps forty since **.

1122:35

CAM-2 sure.

1122:47

CAM-? **.

1122:50

CAM-? **.

1123:02

CAM-1 we're right on the edge of this **.

Appendix B 179 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1123:24

CAM-? **.

1123:58

CAM-2 this is the ground over here on the right.

1124:00

CAM-1 yeah I see an occasional ground **.

1124:13

CAM-? [sound of yawn]

1124:24

CAM-1 boy, this is too much (return).

1124:44

CAM [sound of "ding dong" similar to flight attendant call chime]

1124:47

CAM-2 there's a moon out there. or a space ship.

1124:53

CAM-1 yeah. the mother ship.

1124:56

CAM-2 [sound of chuckle] got your Nike's on?

1125:00

CAM-1 yeah, right.

1125:01

CAM-? [sound of chuckle]

1125:03

CAM-1 what was that guy's name?

1125:04

CAM-2 @, @ or.

Appendix B 180 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1125:06

CAM-1 yeah @.

1125:10

CAM-2 center pumps comin' off.

1125:11

CAM-1 all right.

1125:12

CAM [sound of two clicks]

1125:17

CAM-2 there's your big wadiddily.

1125:19

CAM-1 yeah.

1125:23

CAM-2 thirteen miles?

1125:25

CAM-? ***.

1125:30

CAM [sound similar to ice bag being struck in galley]

1125:47

CAM-1 we got to get over there quick.

1125:52

CAM-2 I don't like that…. that's lightning.

1126:00

CAM-1 sure is.

Appendix B 181 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1126:24

CAM-2 oh.

1126:40

CAM-1 that's about as far as we can go.

1126:41

CAM-2 yeah, I would say right about. maybe a little bit more and that's

about it. we could start down here pretty soon.

1126:49

CAM-1 I'm gonna ask her to come **….

1126:52

CAM-1 this is the bowling alley right here.

1126:54

CAM-2 yeah, I know.

1126:59

CAM-1 in fact those are the city lights straight out there.

1127:01

CAM-2 that's it.

1127:07

CAM-2 want to go down?

1127:09

CAM-1 uuh, not just yet…. but, pretty soon.

1127:14

CAM-1 (seventy two), yeah.

1127:15

CAM-? **.

Appendix B 182 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1127:15

CTR American fourteen twenty descend and maintain one zero

thousand. the, Little Rock altimeter, is two niner eight six.

1127:24

RDO-2 ten thousand, two niner eight six. American fourteen twenty,

thanks.

1127:27

CAM-1 ten set and armed.

1127:28

CAM-2 thanks.

1127:31

PA-1 uh, we're now just uh, eighty miles from the airport and we

have started our descent uh, toward it. quite a light show off

the left hand side of the aircraft. we'll be passing that on our

way toward Little Rock…. and we should be landing here in

about uh, probably about twenty minutes. I'm gonna have to

slightly over-fly the airport, in or.… order to turn back around to

land. it's been a pleasure having you on board for this short

flight and I'd like to take this opportunity to thank you for flying

American Airlines.

1128:06

CAM-2 descent checks are complete.

1128:08

CAM-1 okay.

1128:23

CTR American fourteen twenty contact Memphis center one three

five point eight. good day.

1128:27

RDO-2 thirty five eight, American fourteen twenty, good night.

Appendix B 183 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1128:30

CAM-1 we gotta get there quick.

1128:31

CAM-2 yep.

1128:44

RDO-2 American fourteen twenty leaving two two zero for one zero

thousand.

1128:51

CTR Amer…. fourteen twenty, Memphis rog....

1129:02

CAM-2 sit'em down early?

1129:03

CAM [sound of "ding dong, ding dong" similar to flight attendant call

chime]

1129:06

INT-3 this is Nancy.

1129:07

INT-1 yeah, how you guys uh, doing back there?

1129:08

INT-4 this is Jennifer.

1129:09

INT-1 yeah, how you guys doing back there?

1129:10

INT-4 um, pretty okay.

1129:11

INT-3 they're still out in the in the aisle with the cart doing the service.

Appendix B 184 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1129:14

INT-4 yeah.

1129:14

INT-1 really uh.…

1129:15

INT-3 yeah.

1129:15

INT-1 it's uh, I think it's gonna get a little bumpy here again and if you

don't mind uh….

1129:18

INT-4 do we need to sit down?

1129:19

INT-1 yeah, how far through are you?

1129:21

INT-4 we're almost done but not quite, so….

1129:23

INT-1 okay, well, finish it real quick.

1129:24

INT-4 okay.

1129:25

INT-1 all right.

1129:25

INT-4 'bye.

1129:26

INT-1 'bye.

Appendix B 185 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1129:35

CTR American fourteen twenty, roger. Little Rock altimeter's two

niner eight six.

1129:40

RDO-2 two niner eight six, American fourteen twenty.

1129:47

CAM-2 yeah, that alley's getting' big…. closing to the west.

1129:51

CAM-1 yeah it is.

1129:52

CAM-2 * be okay.

1129:55

CAM-2 I say we get down as soon as we can.

1129:59

CAM-1 two nine eight six?

1130:00

CAM-2 * nine eight six. altimeters are set and cross checked.

1130:09

CAM-2 aw #, no right side **.

1130:52

CAM-2 okay, hydraulic pumps are on, high, and on.

1130:55

CAM-1 okay.

1130:55

CAM-2 altimeters? two nine eight six.

Appendix B 186 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1130:59

CAM-1 reset, two nine eight six.

1131:00

CAM-2 flight instruments and bugs?

1131:02

CAM-1 uuh, I got a hundred, and thirty.

1131:06

CAM-2 yeah.

1131:08

CAM-1 with the flaps forty, a hundred and thirty thousand pounds.

four hundred and sixty feet, two hundred feet ***….

1131:16

CAM-2 set and cross checked.

1131:18

CAM-2 tail de-ice? uh, not required?

1131:21

CAM-1 uh, not required.

1131:22

CAM-2 manual brakes?

1131:24

CAM-1 uuh, manual's fine.

1131:32

CAM-1 I have to go a little to the right here.

1131:33

CAM-2 yeah.

Appendix B 187 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1131:34

CAM-? (don't turn left)

1131:38

CAM-2 actually there's the city right there.

1131:39

CAM-1 yeah.

1131:42

CAM-2 breaking out of this (crud). good…. doing good.

1131:55

CAM-2 whoa. looks like it's movin' this way though.

1131:57

CAM-1 yeah *.

1131:58

CAM-2 ***.

1132:08

CAM-1 * just some lightning straight ahead.

1132:14

CAM-2 *** think we're gonna be okay. right there.

1132:18

CAM-? *.

1132:31

CAM-1 down the bowling alley.

1132:47

CAM-2 as my friends would say, California cool.

1132:51

CAM-1 cool.

Appendix B 188 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1132:52

CAM-2 [sound of chuckle]

1132:54

CAM-1 peachy.

1132:55

CAM-2 exactly.

1133:48

CAM-1 that's forty miles.

1133:49

CTR American fourteen twenty, contact Little Rock approach one

three five point four.

1133:50

CAM-2 yeah.

1133:55

RDO-2 thirty five four, American fourteen twenty. you have a good

night.

1133:57

CTR good night.

1134:05

RDO-2 American uh, fourteen twenty at uh, eleven three for ten thou-

sand.

1134:11

APR American fourteen twenty, Little Rock approach roger. ah we

have a thunderstorm just northwest of the airport moving uh,

through the area now. wind is two eight zero at two eight,

gusts four four and uh, I'll have new weather for you in just a

moment I'm sure.

Appendix B 189 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1134:23

RDO-2 yeah we can see the uh, lightning and uh, you wanta repeat

those winds again.

1134:28

APR right now the wind current wind is two niner zero at two eight,

gusts four four.

1134:34

CAM-1 all right two eight zero at four four.

1134:36

CAM-2 gusts to forty four *.

1134:38

CAM-1 right near the limit.

1134:39

CAM-2 yeah, it's uh, forty degrees off. what's our cross(wind) *.

1134:43

APR American fourteen twenty expect an ILS runway two two left.

1134:46

CAM-1 thirty.

1134:47

RDO-2 two two left, we've got that, fourteen twenty.

1134:50

CAM-2 no that's that's *, you're, not out of the limits because of the

angle *, but it's pretty close.

1134:56

CAM-1 yeah.

1135:21

CAM-2 two two left is the right one…. so uh….

Appendix B 190 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1135:29

CAM-2 I uh, I didn't realize that.

1135:32

CAM-? eerraaw.

1135:37

APR American fourteen twenty, descend at pilot's discretion. main-

tain four thousand.

1135:40

RDO-2 * down to four thousand, American uh, fourteen twenty.

1135:46

CAM-1 four thousand set.

1135:50

CAM-2 okay, ten thousand foot, seatbelt sign no smoking.

1135:52

CAM [sound of "ding dong" similar to flight attendant call chime]

1135:53

CAM-1 yeah I'll get down in a second *.

1135:55

CAM-2 okay.

1136:02

CAM-2 yeah it's ten knots uh.…

1136:04

CAM-1 thirty knots is the crosswind limitation but….

1136:06

CAM-1 thirty knots is the…. wet, well.

Appendix B 191 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1136:08

CAM-2 that's the dry.

1136:09

CAM-1 yeah, dry.

1136:10

CAM-2 what about wet?

1136:11

CAM-1 wet.

1136:12

CAM-2 yeah.

1136:12

CAM-1 is twenty.

1136:13

CAM-2 ah, it's twenty five. aw, what the #.

1136:30

PA-1 flight attendants prepare for landing please.

1136:40

CAM-2 you got the NOTAMS, with ya?

1137:17

CAM-2 see the airport?

1137:18

CAM-1 see it blinking out there.

1137:20

CAM-2 ** to the north,

1137:20

CAM-1 straight ahead.

Appendix B 192 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1137:21

CAM-2 well there's a couple runways here so, the problem is we're

sixteen miles south of the VOR and the airport's another five

miles past that.

1137:29

CAM-1 all right. (doesn't) matter.

1137:32

CAM-2 so we've still got a little ways to go.… bad part..… I'll tell you

what. I'm gonna stay on the run.… the VOR till we get a little

closer.

1138:22

CAM-1 oh I think I see, I see where it is.

1138:25

CAM-2 yeah it's on **.

1138:26

CAM-1 it's straight up there, yeah….

1138:27

CAM-2 * (blinking) *.

1138:28

CAM-1 it looks like there's stratus a layer, right over there.

1138:36

CAM-2 *** I definitely got **. (I'll show you this later).

1138:54

CAM-1 he said there was a storm just northwest of the field?

1138:56

CAM-2 he said northwest.

Appendix B 193 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1138:57

CAM-1 yeah.

1138:58

CAM-2 lightning strike he said storm, uh.

1139:00

APR American fourteen twenty, descend and maintain, three thou-

sand.

1139:03

RDO-2 out of four for three, American uh, fourteen twenty.

1139:06

APR American fourteen twenty uh, you're equipment's a lot better

than uh, what I have. how 's the final for two two left lookin'?

1139:12

CAM-1 what's that?

1139:12

RDO-2 okay, we can uh, see the airport from here. we can barely

make it out but uh, we should be able to make two two. uh,

that storm is moving this way like your, radar says it is but a lit-

tle bit farther off than you thought.

1139:23

APR American fourteen twenty roger, would you just want to shoot

a visual approach?

1139:27

CAM-1 naw.

1139:28

RDO-2 uh, at this point we can't really make it out. we're gonna have

to stay with you as long as possible.

Appendix B 194 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1139:32

APR American fourteen twenty roger. and uh, the winds kinda

kicked around a little bit right now. it's three three zero, at uh,

one one.

1139:38

CAM-1 whoa.

1139:39

RDO-2 okay, well that's a little bit, better than it was.

1139:42

CAM-1 * thirty is a, tailwind though.

1139:45

APR *.

1139:45

APR and uh, right now I have a uh, windshear alert. the center field

wind is three four zero at one zero north boundary wind is

three three zero at two five. northwest boundary wind is zero

one zero at one five.

1139:53

CAM-? *.

1139:56

CAM-1 ** be landing on four?

1139:59

RDO-2 is there a possibility to get runway four?

1140:01

APR American fourteen twenty yes sir. we can do runway four if *

you'd prefer that.

1140:05

CAM-1 it'd be a headwind.

Appendix B 195 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1140:06

CAM-2 yeah.

1140:06

CAM-2 I think we're gonna need.…

1140:08

RDO-2 …we would rather do the headwinds sir.

1140:09

APR I'm sorry, say again American fourteen twenty.

1140:12

RDO-2 yeah, we're gonna want the headwind of course…. runway

four.

1140:19

CAM-1 we're going to three, right?

1140:20

APR American uh, fourteen twenty uh, turn right heading of uh, two

five zero vectors for the ILS runway four right final approach

course.

1140:22

CAM-2 yeah, three thousand.

1140:26

RDO-2 okay, a right turn to two five zero uh, the long way around?

1140:29

APR uh, yes sir, you're a little close to the airport.

1140:31

CAM-1 yeah right.

Appendix B 196 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1140:32

RDO-2 two five zero, that'll work.

1140:36

CAM-2 *, runway four.

1140:46

CAM-2 four right. one one one point three…. zero four two. I think

we were, I think that was the airport right below us.

1141:02

CAM-1 yeah it was. okay, one eleven three.

1141:07

CAM-2 one eleven three. zero four two. four sixty on decision alti-

tude.

1141:14

CAM-2 four thousand for three thousand, is armed.

1141:16

CAM-1 okay.

1141:19

CAM-2 uh, MSA is thirty three hundred feet all the way around.

1141:22

APR American fourteen twenty uh, maintain three thousand three

hundred for now please.

1141:25

RDO-2 three thousand three hundred. we just saw it, thanks.

1141:28

CAM-1 yeah, the uh *.

Appendix B 197 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1141:31

CAM-2 okay. and two two seventeen glide slope intercept all the way

down missed approach right turn to four thousand…. ***.

1141:57

CAM-2 let's see, you got the airport? tell you what. *.

1142:00

CAM-1 yeah. ** I don't have the airport.

1142:03

CAM-2 **, I'm saying you got the ILS.

1142:04

CAM-1 yeah, I got the ILS

1142:07

CAM-1 it's uh….

1142:13

CAM-2 yeah, there it is. I got the airport.

1142:16

CAM-1 okay, and decision height is four sixty.

1142:17

CAM-2 yeah.

1142:19

CAM-1 do you have the airport?

1142:20

CAM-2 *.

1142:20

CAM-1 is that it right there?

Appendix B 198 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1142:21

CAM-? okay.

1142:23

CAM-2 * see, I can't.

1142:24

CAM-1 I don't see a runway.

1142:26

CAM-2 go out this way.

1142:27

APR American fourteen twenty, it appears we have uh, second part

of this storm moving through. the winds now, three four zero

at one six, gusts three four.

1142:34

CAM-1 okay.

1142:35

RDO-2 roger that.

1142:40

CAM-2 you wanna accept a short approach? want to keep it in tight?

1142:42

CAM-1 yeah, if you see the runway. 'cause I don't quite see it.

1142:45

CAM-2 yeah, it's right here, see it?

1142:48

CAM-1 [sound of grunt] you just point me in the right direction and I'll

start slowing down here. give me flaps eleven.

Appendix B 199 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1142:54

RDO-2 and uh….

1142:55

CAM-2 #, it's going right over the.… f-field.

1142:55

CAM-1 *.

1142:56

APR American fourteen twenty, did you call me?

1142:59

RDO-2 well we got the airport. we're going between clouds. I think

it's right off my uh, three o'clock low, about four miles.

1143:05

APR American fourteen twenty, that's it. do you wanna shoot the

visual approach or you wanna go out for the ILS?

1143:09

RDO-2 I can, we'll, we'll (start) the visual. if we we can do it.

1143:11

APR American fourteen twenty's cleared visual approach runway

four right. if you lose it, need some help. let me know please.

1143:15

RDO-2 I'll stay with you as long as possible, OK?

1143:18

APR that's fine, I'm working everything, American fourteen twenty.

Appendix B 200 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1143:20

RDO-2 that works for me.

1143:21

APR all right.

1143:23

CAM-1 well you keep me straight.

1143:23

CAM-2 keep it right here, keep it right here, ** right here.

1143:25

CAM-1 what?

1143:26

CAM-2 okay, did you notice something? there's the airport right there.

okay?

1143:31

CAM-1 where?

1143:31

CAM-2 okay, you're set up on a base for it. okay?

1143:33

CAM-1 I'm on a base now?

1143:35

CAM-2 well, you're on a dogleg. you're comin' in. there's the airport.

1143:38

CAM-1 uh, I lost it.

1143:39

CAM-2 right there, you're you're downwind. see it's right there.

Appendix B 201 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1143:44

CAM-1 I still don't see it. [sound of chuckle] well just vector me. I

don't know.

1143:47

CAM-2 okay, well just go * right here.

1143:49

CAM-1 okay.

1143:59

APR American fourteen twenty, you can monitor one one eight

point seven, runway four right, cleared to land. the wind right

now three three zero at two one.

1144:05

RDO-2 eighteen seven, we'll monitor, American fourteen twenty,

thanks. cleared to land runway four.

1144:10

CAM-1 ******.

1144:13

CAM-2 if you look at ….

1144:14

CAM-1 those red lights out there. where, where's that in relation to….

1144:18

CAM-2 there's another, there's two runways here. there's three run-

ways.

1144:19

CAM-1 yeah I know. see we're losing it. I don't think we can maintain

visual.

1144:22

CAM-2 ** yeah.

Appendix B 202 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1144:23

RDO-2 hold on and uh…..

1144:26

CAM-1 oh, you're on tower.

1144:27

CAM-2 oh, I'm sorry.

1144:28

RDO-2 and approach American fourteen twenty.

1144:29

APR American fourteen twenty, yes sir.

1144:30

RDO-2 and there's a cloud between us and the airport. we just lost

the field and I'm uh, on this vector here, I have the uh, basi-

cally last vector you gave us, we're on kind of a dog leg it

looks like.

1144:39

APR American fourteen twenty, can you fly heading two two zero?

I'll take you out for the ILS.

1144:42

CAM-1 **.

1144:43

RDO-2 yeah two two zero's fine.

1144:45

APR and it will be just one probably one turn on from uh, downwind

to final, for the ILS.

Appendix B 203 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1144:49

RDO-2 'K that's how it's gonna have to be, thanks.

1144:51

CAM-2 yeah, I had it but I lost it with the clouds and that's what I was

saying.

1144:54

CAM-1 okay.

1144:54

APR American fourteen twenty, descend and maintain two thou-

sand three hundred.

1144:56

RDO-2 two thousand three hundred, American fourteen twenty.

1144:59

CAM-2 two thousand three hundred.

1145:00

CAM-1 set and armed. uh, now it is.

1145:07

CAM-2 #, * we had it.

1145:09

CAM-1 yeah. I just, I never saw the runway.

1145:11

CAM-2 no no, it's okay. I **.

1145:12

CAM [sound similar to stabilizer-in-motion horn]

1145:13

CAM-5 stabilizer motion

Appendix B 204 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1145:15

CAM-1 I hate droning around visual at night in weather without, having

some clue where I am.

1145:23

CAM-2 yeah but, the longer we go out here the.…

1145:24

CAM-1 yeah, I know.

1145:25

CAM [sound similar to stabilizer-in-motion horn]

1145:26

CAM-5 stabilizer motion.

1145:29

CAM-2 see how we're going right into this crap.

1145:31

CAM-1 right.

1145:47

RDO-2 and approach American fourteen twenty, I know you're doing

your best sir. we're getting pretty close to this storm. we'll

keep it tight if we have to.

1145:52

APR * American fourteen twenty uh, turn right heading of uh, two

seven zero.

1145:56

CAM [sound similar to stabilizer-in-motion horn]

1145:57

RDO-2 two seven zero, American fourteen twenty.

Appendix B 205 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1145:59

APR and uh, when you join the final, you're going to be right at just

a little bit outside the marker if that's gonna be okay for ya.

1146:04

CAM-1 that's great.

1146:05

RDO-2 that's great with us.

1146:06

APR American fourteen twenty, roger.

1146:11

CAM [sound similar to stabilizer-in-motion horn]

1146:11

CAM-2 see we're right on the base of these clouds so.…

1146:13

CAM-1 yeah.

1146:14

CAM-2 … it's not worth it.

1146:15

CAM [sound similar to stabilizer-in-motion horn]

1146:20

CAM-2 two seven zero, two thousand three hundred?

1146:23

CAM-1 yes sir. * where I am.

1146:25

APR American fourteen twenty, turn right heading three, zero zero.

Appendix B 206 Aircraft Accident Report

30 o 38

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1146:29

RDO-2 right turn three zero zero American fourteen twenty.

1146:39

APR American fourteen twenty is uh, three miles from the

marker. turn right heading zero two zero. maintain two thou-

sand three hundred 'til established on the localizer. cleared

ILS runway four right approach.

1146:43

CAM [brief sound of Morse Code identifier]

1146:47

RDO-2 zero two zero 'til established, American fourteen twenty,

cleared four left approach.

1146:52

CAM-1 aw, we're goin' right into this.

1146:52

APR American fourteen twenty, right now we have uh, heavy rain

on the airport. the uh, current weather on the ATIS is not cor-

rect. I don't have new weather for ya, but the uh, visibility is

uh, less than a mile. runway four right RVR is three thousand.

1146:53

CAM [sound similar to stabilizer-in-motion horn]

1147:04

CAM-1 three thousand.

1147:04

RDO-2 roger that, three thousand, American uh, fourteen twenty. this

is four right, correct?

1147:07

CAM [sound similar to stabilizer-in-motion horn]

Appendix B 207 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1147:08

APR American fourteen twenty, that's correct sir. and runway four

right, cleared to land. the wind three five zero at three zero,

gusts four five.

1147:10

CAM-1 can we land?

1147:16

RDO-2 zero three zero at four five, American fourteen twenty.

1147:19

CAM-2 ** zero forecast right down the runway.

1147:22

CAM-1 three thousand RVR. we can't land on that.

1147:24

CAM-2 three thousand if you look at uh….

1147:26

CAM [sound similar to stabilizer-in-motion horn]

1147:27

CAM-1 what do we need?

1147:28

CAM-2 no it's twenty four hundred RVR.

1147:29

CAM-1 okay, fine.

1147:30

CAM-2 yeah, we're doing fine.

1147:31

CAM-1 all right.

Appendix B 208 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1147:34

CAM-1 uh, fifteen.

1147:36

CAM [sound of clicks similar to flap handle movement]

1147:40

CAM [sound similar to stabilizer-in-motion horn]

1147:44

CAM-1 lllanding gear down.

1147:46

CAM [sound similar to landing gear being operated]

1147:47

CAM [sound similar to stabilizer-in-motion horn]

1147:49

CAM-1 and lights ** please.

1147:51

CAM [sound similar to stabilizer-in-motion horn]

1147:52

CAM-5 stabilizer motion

1147:53

APR windshear alert, center field wind, three five zero at three two,

gusts four five. north boundary wind three one zero at two

niner. northeast boundary wind three two zero at three two.

1148:01

CAM [sound similar to stabilizer-in-motion horn]

1148:02

CAM-5 stabilizer motion.

Appendix B 209 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1148:03

CAM-2 flaps twenty eight?

1148:10

CAM-1 add twenty.

1148:12

CAM-2 right.

1148:12

CAM-1 add twenty knots.

1148:12

APR American fourteen twenty, the runway four right RVR now is

one thousand six hundred.

1148:14

CAM-2 okay.

1148:17

CAM-2 aw #.

1148:18

CAM-1 well we're established on the final.

1148:20

CAM-2 we're established we're inbound, right.

1148:24

RDO-2 okay, American fourteen twenty, we're established inbound.

1148:26

APR American fourteen twenty roger, runway four right, cleared to

land, and the wind, three four zero at three one. north wind,

north uh, boundary wind is three zero zero at two six, north-

east boundary wind three two zero at two five, and the four

right RVR is one thousand six hundred.

Appendix B 210 Aircraft Accident Report

3o38

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1148:36

CAM [sound similar to stabilizer-in-motion horn]

1148:41

RDO-2 American uh, fourteen twenty, thanks.

1148:43

CAM-2 that's a good point.

1148:45

CAM [unidentified intermittent tone]

1148:47

CAM-2 keep the speed.

1148:50

CAM-2 thousand feet.

1148:54

CAM-1 I don't see anything. lookin' for four sixty.

1148:58

CAM [sound similar to stabilizer-in-motion horn]

1149:00

CAM-2 it's there.

1149:02

CAM-2 want forty flaps?

1149:04

CAM-1 oh yeah, thought I called it.

1149:05

CAM-2 forty now. thousand feet. twenty, forty forty land.

1149:10

CAM [unidentified tone similar to sound at time 1148:45]

Appendix B 211 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1149:10

APR wind is three three zero at two eight.

1149:12

CAM-1 this is, this is a can of worms.

1149:17

CAM [sound similar to stabilizer-in-motion horn]

1149:22

CAM [sound similar to stabilizer-in-motion horn]

1149:24

CAM-1 (I'm gonna stay above it a little)

1149:24

CAM-2 there's the runway off to your right, got it?

1149:26

CAM-1 no.

1149:27

CAM-2 I got the right runway in sight.

1149:30

CAM-2 you're right on course. stay where you're at.

1149:31

CAM-1 I got it, I got it.

1149:32

APR wind three three zero at two five.

1149:37.7

CAM-? wipers.

Appendix B 212 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1149:41.4

CAM [sound similar to windshield wiper motion]

1149:46.4

CAM-2 five hundred feet.

1149:50.1

CAM-? *.

1149:53.1

APR wind three two zero, at two three.

1149:53.7

CAM-1 plus twenty.

1149:56.6

CAM-? aw #, we're off course.

1149:57.6

CAM-? **.

1150:00.4

CAM-2 we're way off.

1150:01.5

CAM-1 I can't see it.

1150:04.4

CAM-2 got it?

1150:05.1

CAM-1 yeah I got it.

1150:07.9

CAM-2 hundred feet.

1150:09.4

CAM-? above.

Appendix B 213 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1150:11.1

CAM-2 hundred.

1150:12.8

CAM-5 sink rate.

1150:13.7

CAM-2 fifty.

1150:14.2

CAM-5 sink rate.

1150:14.5

CAM-2 forty.

1150:15.8

CAM-2 thirty.

1150:17.6

CAM-2 twenty.

1150:18.3

CAM-2 ten.

1150:20.2

CAM [sound of two thuds similar to aircraft touching down on run-

way concurrent with unidentified squeak sound]

1150:22.2

CAM-2 we're down.

1150:24.4

CAM-2 we're sliding.

1150:26.1

CAM-1 #.… #.

Appendix B 214 Aircraft Accident Report

INTRA-COCKPIT COMMUNICATION AIR-GROUND COMMUNICATION

TIME & TIME &

SOURCE CONTENT SOURCE CONTENT

1150:31.8

CAM-? on the brakes.

1150:33.1

CAM-? oh sh ….

1150:33.5

CAM [sound similar to increase in engine RPM]

1150:35.1

CAM-? other one, other one, other one.

1150:40.9

CAM-? aw #.

1150:41.6

CAM-? ##.

1150:43.8

CAM [sound of impact]

1150:44.3

CAM-? ##.

1150:46.9

CAM [sound of several impacts]

1150:48.1

END of RECORDING

END of TRANSCRIPT

215 Aircraft Accident Report

Appendix C

Automated Surface Observing System Data

Time Wind information Precipitation information

2 minute 5-second gust 1 minute 15 minute

2330 203° at 08 knots 244° at 09 knots 0.00 inch Trace amount

2331 241° at 11 knots 274° at 20 knots 0.00 inch

2332 273° at 19 knots 278° at 29 knots 0.00 inch

2333 291° at 25 knots 295° at 35 knots 0.00 inch

2334 298° at 27 knots 290° at 33 knots Trace amount

2335 298° at 26 knots 289° at 27 knots Trace amount

2336 289° at 21 knots 281° at 25 knots 0.01 inch

2337 286° at 15 knots 285° at 18 knots 0.00 inch

2338 304° at 17 knots 318° at 18 knots 0.01 inch

2339 320° at 17 knots 336° at 20 knots 0.02 inch

2340 320° at 17 knots 336° at 20 knots 0.01 inch

2341 338° at 17 knots 333° at 22 knots 0.01 inch

2342 351° at 18 knots 356° at 27 knots 0.01 inch

2343 359° at 21 knots 360° at 28 knots 0.02 inch

2344 352° at 20 knots 357° at 20 knots 0.02 inch

2345 320° at 17 knots 336° at 20 knots 0.03 inch 0.14 inch

2346 322° at 14 knots 333° at 16 knots 0.03 inch

2347 328° at 15 knots 329° at 26 knots 0.06 inch

2348 314° at 14 knots 283° at 11 knots 0.07 inch

2349 296° at 12 knots 291° at 20 knots 0.03 inch

2350 285° at 16 knots 302° at 22 knots 0.04 inch

2351 281° at 18 knots 291° at 21 knots 0.04 inch

Appendix C 216 Aircraft Accident Report

Note: In the body of this report, wind directions were rounded to the nearest tenth.

2352 284° at 19 knots 287° at 24 knots 0.06 inch

2353 281° at 18 knots 287° at 21 knots 0.08 inch

2354 277° at 14 knots 264° at 14 knots 0.13 inch

2355 287° at 13 knots 296° at 18 knots 0.08 inch

2356 299° at 23 knots 317° at 76 knots 0.15 inch

2357 308° at 20 knots 281° at 16 knots 0.14 inch

2358 291° at 10 knots 241° at 21 knots 0.11 inch

2359 297° at 09 knots 012° at 08 knots 0.05 inch

0000 358° at 09 knots 023° at 14 knots 0.02 inch 1.09 inches

Time Wind information Precipitation information

2 minute 5-second gust 1 minute 15 minute