34
A Review of Biological Reference Points in the Context of the
Precautionary Approach
Wendy L. Gabriel and Pamela M. Mace
NMFS, Northeast Fisheries Science Center, 166 Water Street, Woods Hole, MA 02543-1097.
E-mail address: [email protected]
Abstract.- Draft guidelines for National Standards under the Magnuson-Stevens Act state that Councils should adopt a precau-
tionary approach to specification of Optimum Yield (OY), and list three features which characterize this approach: 1.) target refer-
ence points such as OY should be set safely below limit reference points (such as the Maximum Sustainable Yield, MSY); 2.) stocks
at sizes below the level that produces MSY should be harvested at lower rates than stocks at sizes above the level that produces MSY;
3.) as uncertainty about stock status or productive capacity increases, target catch levels should be more cautious. The guidelines
indicate limit reference points which include a maximum fishing mortality rate which produces MSY and minimum stock size
thresholds from which a stock could be rebuilt to MSY within ten years. Reference points can be direct estimates or proxies for
direct estimates, depending on adequacy of available data. In this paper, we review desirable properties of directly-estimated and
potential proxy biological reference points, in the contexts of the National Standards, guidelines, approaches adopted by interna-
tional management bodies, and other more generic contexts of the precautionary approach. We compare alternative candidate
reference points in terms of their utility and potential performance as limit or target reference points in risk-averse management
frameworks.
Introduction
The objective of this paper is to review model-based
approaches to the estimation of biological reference
points, review precautionary reference points as a com-
ponent of the precautionary approach, and describe the
relevant subset of biological reference points which are
consistent with the MSY-related focus of the UN Agree-
ment on Straddling Fish Stocks and Highly Migratory
Fish Stocks (1995) and the revised Magnuson-Stevens
Fisheries Conservation and Management Act
(MSFCMA).
Biological Reference Points: A Brief Review
A biological reference point (BRP) in its most ge-
neric form is a metric of stock status from a biological
perspective. The biological reference point often re-
flects the combination of several components of stock
dynamics (growth, recruitment and mortality, usually
including fishing mortality) into a single index. The
index is usually expressed as an associated fishing mor-
tality rate or a biomass level. The procedure for esti-
mating the reference point and the underlying model is
agreed within the scientific community.
The three most common models that underlie bio-
logical reference points have been summarized by
Sissenwine and Shepherd (1987): (1) spawner-recruit,
(2) dynamic pool and (3) production models. The choice
of model is predicated on life history and availability of
catch, relative abundance, stock-recruitment, and age-
specific mortality, growth, and maturity data (Table 1).
Spawner-Recruit Reference Points (Semelparous Popu-
lations)
Ricker (1975) describes multiple features of the
spawner-recruit relationship which may serve as bio-
logical reference points for semelparous populations
such as Pacific salmon. In these models, spawners and
recruits are both represented in terms of numbers. Age-
structure is not incorporated, because spawners are as-
sumed to spawn once and then die; and because recruits
produced by spawners are all assumed to return to spawn
at the same time. The underlying dynamic mechanism
is density-dependent compensation in the stock-recruit-
ment relationship, which results in an increased produc-
tion of recruits per capita at lower spawner abundances
and reduced per capita production at high stock sizes.
This may arise when the survival of eggs and/or larvae
is affected by density-dependent competition for food
or space, compensatory predation, or cannibalism of
young by adults (Ricker, 1975). Those reference points
are derived from continuous spawner-recruit functions
and include spawners needed for maximum recruitment
(P
m
in Ricker’s notation), the replacement spawner abun-
dance at which recruitment equals parent stock (P
r
);
spawners needed for maximum sustainable yield
(MSY)(P
s
) and rate of exploitation at MSY (u
s
).
Dynamic Pool (Per-Recruit) Reference Points
Dynamic pool models were initially described by
Thompson and Bell (1934) and Beverton and Holt
(1957). These models serve as the basis for biological
The views expressed herein are those of the authors, not necessarily NMFS’
35
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
reference points on a cohort or year class basis, stan-
dardized to the number recruited to the cohort, and so
are also referred to as yield-per-recruit, egg-per-recruit
and spawning-stock-biomass-per-recruit models. Age
structure is incorporated in terms of age-specific sched-
ules of mortality, growth, and sexual maturity. Age-
specific fishing mortality rates reflect the effects of a
fishery selection (or exploitation) pattern, in which the
vulnerability of a cohort changes as it ages. This could
reflect changing patterns in availability to the fishery or
vulnerability to the gear. “Knife-edge” exploitation pat-
terns are approximations that assume that below the age
at first capture, fishing mortality = 0, but at or above the
age at first capture, the cohort is fully vulnerable to the
same rate of fishing mortality. Age-specific schedules
of weights in the spawning stock or weights in the
catches (as landings or discards) are specified, as are
age-specific maturity rates. The models enable an evalu-
ation of the effects of alternative exploitation patterns
and fully-recruited fishing mortality rates on the amount
of yield or spawning stock biomass per recruit, over the
lifetime of the cohort, independent of the initial size of
the cohort at recruitment. The models do not usually
incorporate density-dependent compensation: the same
age-specific mortality, maturity, and growth schedules
are assumed to apply regardless of year class size ini-
tially or at subsequent ages. The models do not incorpo-
rate density-independent effects: the age-specific rates
are assumed to apply regardless of any changing envi-
ronmental conditions, fishery behavior, predation lev-
els, or prey availability over the life of the cohort. The
schedules must be obtained over the entire lifespan of
the cohort in order for the calculated yield per recruit or
spawning stock biomass per recruit to be realized. In
the case of spawning-stock-biomass-per-recruit analy-
ses, all kilograms of spawning stock biomass are as-
sumed to be equally productive in terms of recruitment,
i.e., the production of viable eggs per kilogram of spawn-
ing stock biomass is assumed to be equal regardless of
age composition, size composition, and number of pre-
vious spawning seasons of spawners contributing to the
spawning stock biomass. For this reason, metrics other
than spawning biomass are sometimes used (e.g., egg
production). Length-based estimates of yield per re-
cruit and spawning stock biomass per recruit are pos-
sible when growth, maturity at length and length-com-
position data are available (e.g., Gallucci et al., 1996).
Reference points derived from yield-per-recruit
analyses include F
max
, the (fully-recruited) fishing mor-
tality rate which produces the maximum yield per re-
cruit; and F
0.1
, the fishing mortality rate corresponding
to 10% of the slope of the yield-per-recruit curve at the
origin (Gulland and Boerema, 1973). The F
0.1
refer-
ence point was conceptualized as a biologically precau-
tionary target relative to F
max
: at F
0.1
, catch per unit ef-
fort is not reduced substantially, but the fishing mortal-
ity rate is lower than F
max
. Because the yield-per-re-
cruit analyses only reflect schedules of mortality and
weight at age in the catch, both F
max
and F
0.1
are refer-
ence points in the context of growth overfishing, not
recruitment overfishing.
A wide variety of reference points have been de-
rived from spawning-stock-biomass-per-recruit models.
In isolation, spawning-stock-biomass-per-recruit analy-
ses reflect schedules of mortality, maturity, and spawn-
ing weight at age for a cohort. Under conditions of no
fishing mortality, 100% of a stock’s spawning potential
is obtained. As fishing mortality rates increase, spawn-
ing stock biomass per recruit decreases, as more spawn-
ing opportunities are lost over the lifetime of the cohort.
The reduction in spawning stock biomass per recruit
relative to the unfished level can be reflected as a per-
centage of the maximum spawning potential (MSP), e.g.,
a fishing mortality rate denoted F
35%MSP
would allow a
Age structure in
population
S-R data
required
S-R
function
required
Model type Example citation Reference points Comments
Unknown No No Surplus production Schaefer, 1957
Prager, 1993
F
M
SY
, B
M
SY
Very risk-prone without auxiliary
data on recent relative recruitment
No (semelparous) Yes Yes Spawner-recruit Ricker, 1975
P
s
, u
s
Yes (iteroparous) No No Dynamic pool,
Y/R
Thompson and
Bell, 1934
F
max
, F
0.1
No information about reproductive
dynamics
By
analogy
By
analogy
Dynamic pool,
SSB/R
F
20%SPR
,
F
35%SPR
No stock-recruitment relationship,
except by analogy
Yes No Dynamic pool,
SSB/R
Shepherd, 1982
F
med
Yes Yes Dynamic pool,
SSB/R
Mace, 1994
F
τ
Yes Yes Production Sissenwine and
Shepherd, 1987
F
M
SY,
B
M
SY
Table 1. Summary of principal models that underlie biological reference points, and associated specification of age-
structured and stock-recruitment data.
36
stock to attain only 35% of the maximum spawning
potential which would have been obtained under condi-
tions of no fishing mortality. It is thus possible to cal-
culate spawning stock biomass per recruit as a function
of fishing mortality rate, in terms of either kilograms of
spawning stock biomass per number of recruits or in
terms of percentage of the maximum spawning poten-
tial (the ratio of kilograms of spawning stock biomass
per recruit under a specific F compared to kilograms of
spawning stock biomass per recruit under no F). These
give rise to reference points of the form of e.g., F
20%SPR
or F
35%SPR
, where SPR stands for spawning (products)
per recruit, and “products” are biomass, egg production,
or related metrics, and x% SPR has exactly the same
meaning as x% MSP.
Results of spawning-stock-biomass-per-recruit
analyses can be combined with stock-recruitment data
to provide reference points in the context of recruitment
overfishing. If a stock-recruitment model can be fitted,
then the fishing mortality rate which corresponds to the
slope of the function at the origin can be estimated, F
τ
(Mace and Sissenwine, 1993). This is possible because
the slope of the stock-recruitment function has units of
R/SSB and if this value is inverted to units of SSB/R, a
corresponding fishing mortality rate can be found from
the relationship between SSB/R and F as described
above.
It may not be possible to fit a stock-recruitment re-
lationship because the range of observed stock sizes is
narrow, data are dominated by environmental variabil-
ity, or stock or recruitment estimates are imprecise or
inaccurate, for example. In that case, it still may be
possible to define fishing mortality reference points
based on the distribution of observed R/SSB, from a ra-
tio of observed SSB and subsequent recruitment. Such
reference points include those introduced by Shepherd
(1982) and ICES (Anon., 1984), F
low
, F
med
, and F
high
,
corresponding to the lower 10-percentile, 50-percentile,
and upper 90-percentile of the observed R/SSB ratios,
respectively. These reference points represent fishing
mortality rates which can be supported by observed sur-
vival rates from spawning to recruitment in 90%, 50%,
and 10% of the years, respectively. The same short-
comings in the data which would prevent fitting a stock-
recruitment relationship make other reference points
based on different forms of the same data less reliable,
however. Depending on which part of the stock size
range is observed, F
med
may be close to F at the slope at
the origin, F
msy
, or close to zero. F
med
may also be un-
sustainable depending on the age structure of the stock
and degree of temporal correlation in survival ratios:
although high recruitment rates may balance low recruit-
ment rates over the long term, if age structure in a stock
is severely truncated, (e.g., to four age classes) there is a
higher probability the stock may collapse under exploi-
tation at F
med
(e.g., if four consecutive years of poor R/
SSB were obtained). G
loss
(Cook, 1998) is a more elabo-
rately formulated reference point which includes uncer-
tainty in the estimation of the stock-recruitment data and
the R/SSB calculations using simulation procedures, and
a smoothed trend rather than a fitted stock-recruitment
relationship: the distribution of R/SSB at the lowest ob-
served stock size is simulated, and compared with the
distribution of R/SSB at the current fishing mortality
rates.
Surplus Production Reference Points I
The surplus production model is among the sim-
plest of the models used for stock assessment: it does
not reflect any age structure in a population, and the
dynamics of natural mortality, growth, and recruitment
are aggregated into a single intrinsic rate of population
biomass increase, modified by fishing mortality. Model
dynamics are also affected by the size of the population
with respect to its carrying capacity. Data requirements
are modest: the model can be fitted based on an abun-
dance (catch per unit effort) index and catch. Models
have been formulated which do not require an equilib-
rium assumption (e.g., Prager, 1994). However, both
observation and process errors occur (random variation
in the observed abundance index and catch of the stock;
and in the population dynamics of the stock, respec-
tively). Although observation error estimators have been
fairly well developed, if process error is large, then pa-
rameter estimation may be poor (Prager, 1994; Chen
and Andrew, 1998). Thus, if there is a trend or cycle to
natural or fishing mortality, growth, or recruitment, for
example, this type of model will perform poorly or sepa-
rate fits would be required for each period in the stock’s
history. Although surplus production models produce
relatively precise estimates of MSY and f
MSY
, absolute
values of F
MSY
and B
MSY
are usually not precise and re-
quire good estimates of q (the parameter that scales abun-
dance indices into biomass estimates)(Prager, 1994).
Surplus Production Reference Points II
The production model described in Sissenwine and
Shepherd (1987), in contrast, is one of the more data-
intensive and complex models. It requires a functional
stock-recruitment relationship, a spawning-stock-biom-
ass-per-recruit analysis, and a yield-per-recruit analy-
sis. For any specified rate of fishing mortality, an asso-
ciated value of SSB/R is defined, incorporating the as-
sumptions detailed in the previous section on dynamic
pool models. When this value of SSB/R is inverted and
superimposed on the stock-recruitment function as a
slope (R/SSB), the intersection of this slope with the
stock-recruitment function defines an equilibrium level
of recruitment. When this value of recruitment is mul-
tiplied by the yield per recruit calculated for the same
37
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
fishing mortality rate, the equilibrium yield associated
with the fishing mortality rate emerges. F
MSY
, the fish-
ing mortality rate which maximizes the yield from the
system (conditional on selection pattern, schedules of
growth and maturity, accuracy of stock-recruitment
function, etc. as detailed in the preceding section on
dynamic pool models) can be found; and B
MSY
, the as-
sociated stock biomass which produces that yield can
also be found.
Biological Reference Points and Fishery Management
Reference Points
In a management context, a biological reference
point can serve as a performance standard or a land-
mark for a fishery management regime. Other types of
performance standards are also available in the domain
of economics (e.g., the fishing mortality rate which pro-
duces maximum economic yield). Some performance
standards are not quantitative but only directional (e.g.,
some social anthropological elements such as the social
stability of local fishing villages). Others are not articu-
lated (e.g., minimum sustainable whinge, sensu Pope)
The biological reference point itself is not equiva-
lent to a management regime or the management objec-
tives. If the management objective were to maximize
economic efficiency, for example, the effect of any pro-
posed measures would presumably be evaluated in terms
of economic impact. Those proposed measures would
also be evaluated with respect to the impact on stock
status in terms of growth overfishing, recruitment over-
fishing, or the sustainability of yields from the stock,
however. Appropriate biological reference points would
provide standards by which to judge the performance of
that management regime in a biological context, even
though the ultimate aim of the management regime might
be to achieve some specific economic end. As noted
above, F
max
or F
0.1
are biological reference points com-
monly used to index growth overfishing; F
35%SPR
, F
med
,
or F
τ
have been used to index recruitment overfishing;
and F
MSY
and B
MSY
index stock conditions which pro-
duce surplus production as maximum sustainable yield
(MSY). In the context of this paper, MSY or OY are con-
sidered emergent properties of other reference points or
harvest control policies.
Precautionary Reference Points
Two types of precautionary reference points, limits
and targets, and their management contexts are described
in Annex II of the UN Straddling Stocks Agreement
(1995): Limit reference points set boundaries which are
intended to constrain harvesting within safe biological
limits within which the stocks can produce maximum
sustainable yield.... Fishery management strategies shall
ensure that the risk of exceeding limit reference points
is very low. If a stock falls below a limit reference point
or is at risk of falling below such a reference point, con-
servation and management action should be initiated to
facilitate stock recovery... The fishing mortality rate
which generates maximum sustainable yield should be
regarded as a minimum standard for limit reference
points. For stocks which are not overfished, fishery
management strategies shall ensure that fishing mor-
tality does not exceed that which corresponds to maxi-
mum sustainable yield, and that the biomass does not
fall below a predefined threshold.
In Annex II, Target reference points are intended
to meet management objectives...Fishery management
strategies shall ensure that target reference points are
not exceeded on average. For overfished stocks, the
biomass which would produce maximum sustainable
yield can serve as a rebuilding target.
Thus, the UN Straddling Stocks Agreement defines
two fishery management reference points to achieve pre-
cautionary objectives: limit reference points and target
reference points. These reference points are cast en-
tirely in terms of biological reference points related to
maximum sustainable yield, B
MSY
and F
MSY
.
The FAO guidelines on the precautionary approach
(1995a) discuss operational targets and constraints, and
treat biological reference points as measurable terms to
express those targets and constraints. The guidelines
recognize that what is measurable will vary, depending
on species and fishery. Operational targets are associ-
ated with desirable outcomes to be attained, such as par-
ticular abundance levels or fishing mortality rates. Op-
erational constraints are associated with undesirable
outcomes to be avoided, such as risk of declining re-
cruitment. The constraint is directly comparable to the
limit: “it is highly desirable... to maintain acceptable low
levels of probability that the constraints are violated.”
Under the precautionary approach, operational targets
may require adjustment to be consistent with constraints,
e.g., so that target fishing mortality rates are lower than
F
MSY
. Constraints have precedence over targets: if B
MSY
(target) were lower than the biomass where there is a
high probability of reduced recruitment (constraint), then
the probability of violating the constraint while meeting
the target would be too large. If targets can be ap-
proached rapidly, then there may be a possibility of over-
shooting the target and violating the constraints, which
should be avoided.
The critical reference point within the precaution-
ary context is the limit reference point. Within Annex
II, paragraph 7 states that: The fishing mortality rate
which generates maximum sustainable yield should be
regarded as a minimum standard for limit reference
points. Within the revised MSFCMA, Section 3(29)
38
states that : The terms ‘overfishing’ and ‘overfished’
mean a rate or level of fishing mortality that jeopar-
dizes the capacity of a fishery to produce the maximum
sustainable yield on a continuing basis. There thus is
indirect correspondence between limit reference points
recommended under the UN Straddling Stocks Agree-
ment and overfishing definitions under the revised
MSFCMA: in both cases, F
MSY
represents an upper bound
to fishing mortality rates. Similarly, there is correspon-
dence between target reference points intended to meet
management objectives under the Straddling Stocks
Agreement and OY under the revised MSFCMA.
Garcia (1995) distinguishes among limit, target, and
threshold reference points in the precautionary context:
limit points should never be reached, and if they were to
be reached, severe and corrective management actions
should be implemented. He indicates that limits should
be minimum rebuilding targets to be reached before any
rebuilding measures are relaxed. The threshold refer-
ence point is defined as an “early warning” reference
point, to reduce the probability that a target or limit point
would be exceeded due to estimation or observation
uncertainty or due to slow management reaction.
Thresholds are advisable when there is an especially
high probability of a negative outcome when the limit is
crossed, e.g., in a highly variable environment, when
species are at the edge of their geographic range or are
relatively unresilient; or other circumstances when the
cost of exceeding the limit is high (Garcia, 1995).
A relatively wide variety of precautionary reference
points has been proposed by various different national
and international working groups and fishery manage-
grebnesoR
late 6991,.
lanoitaN
WAS
,GSN/CS
8991
desiveR
AMCFM
SECIOFANOCSANTACCI
:mreT
timiL
etulosbA
dlohserht
dlohserhTtimiLB:timiL
mil
F,
mil
B:timiL
mil
F,
mil
noitavresnoC
timil
gnihsifrevO
sPRB
:mreT
dlohserhT
yranoituacerP
dlohserht
yranoituacerP
tegrat
yranoituacerP
B:eulav
ap
F,
ap
B:reffuB
,fub
F
fub
:mreT
tegraT
tegraTtegraTtegraT
B:tegraT
tegrat
,
F
tegrat
B:tegraT
rt
F,
rt
tnemeganaM
tegrat
:PRB F
timiL
erehwF
5.0=)R(E
R(E
xam
)
=F
F(niM
YSM
ro,
Fseixorp
RPS%
,
F
1.0
=)M=F,
TMFM
F
YSM
F
mil
F=
hsarc
F,
ssol
Fro
dem
tfel(
)bmil
F
mil
F=
YSM
,
F
xam
F,
dem
ro
F
RPS%03
tnemepacsE
gnicudorp
B
YSM
F=F
YSM
F,
1.0
Fro
xam
:PRB F
dlohserhT
57.0=F
TMFM
F
ap
F=
mil
e
s2-
Fro
gpl
Fro
dem
F
fub
F=
mil
e
s2-
,
Fro,M
YSM
2/
:PRB F
tegraT
F
YO
F<
YSM
F
YO
F<
YSM
BSS:PRB
timiL
erehwBSS
5.0=)R(E
R(E
xam
)
=B
B(xaM
YSM
,2/
BotB
YSM
ni
=)sraey01
TSSM
B
YSM
rof(
)gnidliuber
B
mil
B=
ssol
ro
LABM
B
mil
B=
ssol
,
ro,LABM
B*2.0
xam
)yevrus(
tnemepacsE
gnicudorp
B
YSM
B=B
YSM
BSS:PRB
dlohserhT
B
ap
B=
ssol
ro
B
mil
e
s2-
B
fub
B=
mil
e
s2-
,
B3/2
YSM
ro,
B*5.0
xam
)yevrus(
BSS:PRB
tegraT
B
YO
B
YO
B
YSM
Table 2. Summary of limit, threshold and target reference points as defined by U.S. advisory documents and legislation, and
international management institutions. Initial definition of limit, threshold and target reference points are by Garcia (1995).
39
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
ment organizations. While biological reference points
are based on scientifically agreed models, precaution-
ary reference points reflect individual organizations’ in-
terpretations and implementations of precautionary
management. Using Garcia’s distinctions between limit,
threshold, and target reference points as a basis for or-
ganization, we summarize a range of precautionary ref-
erence points currently or recently under consideration
by various working and management groups (Table 2).
Additional information on different organizations’ ap-
plications of the precautionary approach is summarized
in Mace and Gabriel, this volume. It is important to
note that in the U. S. National Standard Guidelines, Tech-
nical Guidance on the Use of Precautionary Approaches
to Implementing National Standard 1 of the Magnuson-
Stevens Fishery Conservation and Management Act
(Restrepo et al., 1998), and Scientific Review of Defi-
nitions of Overfishing in U.S. Fishery Management Plans
(Rosenberg et al., 1994), the use of the term “threshold”
corresponds to the “limit” reference point as defined in
FAO guidelines rather than to the “early warning” ref-
erence point indicated by Garcia (1995).
Precautionary Management: Harvest Control Rules,
Uncertainty and Precautionary Reference Points
The FAO Code of Conduct for Responsible Fisher-
ies (FAO 1995b) summarizes the relationship between
precautionary reference points and harvest control rules:
When precautionary or limit reference points are ap-
proached, measures should be taken to ensure that they
will not be exceeded. These measures should where
possible be pre-negotiated. If such reference points are
exceeded, recovery plans should be implemented imme-
diately to restore the stocks. The biological reference
points which serve as limits, thresholds, or targets are
triggers for management actions or are parameters in
harvest control rules. The harvest control rule is a pre-
agreed course of management action as a function of
stock status and other economic or environmental con-
ditions. A recovery plan may be considered a special-
ized control rule which applies when the stock is out-
side safe biological limits. Harvest control rules (in-
cluding their component biological reference points)
should be developed in the management planning stage
with the involvement of all stakeholders, and then evalu-
ated for robustness to uncertainties in statistical estimates
of stock status, environmental conditions, harvester be-
havior, and managers’ ability to change harvest levels
(FAO, 1995b). If harvest control rules are based on large
amounts of uncertainty in terms of model, observation,
process, or implementation errors (including estimation
of reference points), then the formulation of the control
rule should be more precautionary. If, on the other hand,
inputs to harvest control rules are based on little uncer-
tainty and/or if resulting controls more stringent, then a
less precautionary formulation of the control rule should
be successful.
In a different approach to the development of har-
vest control rules, the management community could
specify performance criteria for harvest rules (includ-
ing robustness) at the outset, and then alternative har-
vest control rules would be developed which meet those
performance criteria. This different approach is imple-
mented in the International Whaling Commission’s re-
vised management procedure, and focusses pre-agree-
ment on the performance criteria rather than on any par-
ticular control rule or component reference points.
The need for simultaneous consideration of refer-
ence points and actions to be taken if they are exceeded
is made in both the FAO Code of Conduct for Respon-
sible Fisheries (1995b) and Article 6 of the United Na-
tions agreement relating to the conservation and man-
agement of straddling fish stocks and highly migratory
fish stocks (1995). In the FAO Code of Conduct,
7.5.2 In implementing the precautionary approach,
States should take into account, inter alia, uncer-
tainties relating to the size and productivity of the
stocks, reference points, stock condition in rela-
tion to such reference points, levels and distribu-
tion of fishing mortality and the impact of fishing
activities, including discards, on non-target and
associated or dependent species as well as envi-
ronmental and socio-economic conditions.
7.5.3 States and subregional or regional fisheries man-
agement organizations and arrangements should,
on the basis of the best scientific evidence avail-
able, inter alia, determine:
a. stock specific target reference points, and,
at the same time, the action to be taken if
they are exceeded; and
b. stock specific limit reference points, and, at
the same time, the action to be taken if they
are exceeded; when a limit reference point
is approached, measures should be taken
to ensure that it will not be exceeded.
In Article 6 of the Straddling Stocks Agreement (1995):
3. In implementing the precautionary approach,
States shall:
(a) improve decision-making for fishery re-
source conservation and management by
obtaining and sharing the best scientific in-
formation available and implementing tech-
niques for dealing with risk and uncertainty;
(b) apply the guidelines set out in Annex II and
determine, on the basis of the best scientific
40
information available, stock-specific refer-
ence points and the action to be taken if they
are exceeded;
(c) take into account, inter alia, uncertainties
relating to the size and productivity of the
stocks, reference points, stock condition in
relation to such reference points, levels and
distribution of fishing mortality and the im-
pact of fishing activities on non-target and
associated and socio-economic conditions...
4. States shall take measures to ensure that, when
reference points are approached, they will not be
exceeded. In the event that they are exceeded,
States shall, without delay, take the action deter-
mined under paragraph (3) to restore the stocks.
Implementation of the precautionary approach re-
quires consideration of uncertainty in stock size and pro-
ductivity. Unless stock sizes are known with perfect
certainty, the estimation of uncertainty associated with
a reference point is only part of the precautionary pro-
cess, and the uncertainty associated with the current es-
timate of stock size or stock status is a critical part of the
evaluation. The probability that the currently observed
fishing mortality rate, for example, exceeds the limit
reference point then would become conditional on the
estimate of the limit reference point (e.g., Conser and
Gabriel, 1992).
The harvest control rule has two components: the
specification of the reference points (and other relevent
parameters), and a functional form relating current stock
status and reference points to management reaction (e.g.,
catch). The two components act together to determine
the degree of precaution afforded by the rule. Rosenberg
and Restrepo (1996) discuss the interaction among the
acceptable probability of overfishing, the consequences
of exceeding limit reference points, and the action to be
taken when the stock is overfished. For example, an
acceptable probability of overfishing could be higher if
the action to be taken when the limit is exceeded is im-
mediate and drastic. An acceptable probability of over-
fishing could be higher if stock conditions were excep-
tionally favorable, or if the result is simply that the prob-
ability of poor recruitment increases slightly in only one
year, rather than resulting in a significant increase in the
probability of repeated recruitment failure.
Parameterizing Limit Control Rules under National
Standard Guidelines
As noted previously, within the precautionary con-
text, the limit reference point is the critical reference
point. Both the UN Straddling Stocks Agreement and
the revised MSFCMA focus on MSY-related reference
points as limits. This constrains the range of relevent
biological reference points to a subset of those described
earlier.
A default limit control rule is outlined in the Tech-
nical Guidance on the Use of Precautionary Approaches
to Implementing National Standard 1 of the Magnuson-
Stevens Fishery Conservation and Management Act,
which defines limits to fishing mortality rate as a func-
tion of stock biomass (Restrepo et al., 1998). The rule
is based on three parameters, F
MSY
, B
MSY
and c, a factor
which reflects the expectation that a stock fished at F
MSY
would naturally fluctuate around B
MSY
:
MSY
MSY
Bc
BF
BF
=)(
for all B ≤ c B
MSY
F(B) = F
MSY
for all B > c B
MSY.
The extent of that fluctuation is likely related to the natu-
ral mortality rate, and so c is defined as the maximum of
(1-M, 1/2). The fishing mortality rate cannot exceed
F
MSY
, regardless of stock size, and must be reduced be-
low F
MSY
to zero as biomass declines below cB
MSY
to zero.
A minimum stock size threshold (MSST) is also speci-
fied: in no case should MSST be less than half the level
which produces MSY (i.e., MSST >1/2 B
MSY
), and MSST
may be approximated as cB
MSY
. The rule provides an
approximate estimate of the maximum fishing mortal-
ity rate (MFMT).
The NMFS National Standard Guidelines for Stan-
dard 1 define MSY as “the largest long-term average
catch or yield that can be taken from a stock or stock
complex under prevailing ecological and environmen-
tal conditions” with MSY stock size defined as “the long-
term average size of the stock or stock complex, mea-
sured in terms of spawning biomass or other appropri-
ate units, that would be achieved under an MSY control
rule in which fishing mortality rate is constant.” The
MSY control rule is defined as “a harvest strategy which,
if implemented, would be expected to result in a long-
term average catch approximating MSY.” In this con-
text, the MSY stock size would be reflected by the bio-
logical reference point B
MSY
, and the MSY fishing mor-
tality rate would correspond to the biological reference
point F
MSY
.
Situations Requiring the Use of Proxies for F
MSY
and
B
MSY
The MSFCMA allows for the use of proxies in situ-
ations where there is insufficient knowledge to imple-
ment approaches outlined above. In general, proxies
would be needed when MSY-related parameters cannot
be estimated at all from available data, or when their
estimated values are deemed to be unreliable for vari-
ous reasons (e.g., extremely low precision, insufficient
41
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
contrast in the data, or inadequate models). We refer to
these situations as “data-poor” and “data-moderate”, re-
spectively. However, it should also be noted that there
may also be circumstances under which proxies would
also be useful in “data-rich” situations (e.g., when they
are believed to be more robust or reliable than the esti-
mates of MSY-related parameters). Thus, our use of the
term “data-moderate” can be more generally interpreted
as meaning “information-moderate”.
In this report, proxies are substitutes for key bio-
logical reference points, which are used in place of those
key reference points because they are easier to calcu-
late, or require fewer data, or are more robust. MSY-
based reference points are often difficult to estimate,
particularly when the calculations involve estimation of
the parameters of a stock-recruitment relationship. How-
ever, MSY has been the central focus of management
objectives for several decades in many national and in-
ternational agreements, and many proxies have been
developed and applied. In addition, empirical studies
and computer models have suggested which proxies can
generally be considered reasonable for use as “default”
substitutes (point estimates or ranges corresponding to
life history strategies) for MSY-related parameters.
The list of proxies presented in the following sec-
tions is not all-inclusive and fisheries scientists are en-
couraged to develop and examine alternatives.
Data-Moderate Situations
In general, reference points from yield-per-recruit
(YPR) and spawning-stock-biomass-per-recruit (SPR)
analyses are easy to calculate because relatively few data
are required; in particular, it is not necessary to obtain
stock-recruitment data. For this reason, YPR and SPR
reference points are often used as proxies for other ref-
erence points that do require stock and recruitment data.
Proxies for F
MSY
F
max
was one of the earliest measures used as a proxy
for F
MSY
. However, it was often believed to be an over-
estimate of F
MSY
, because it does not account for the fact
that recruitment must decline at low spawning stock
sizes. Computer models have also demonstrated that
F
max
invariably overestimates F
MSY
if a Beverton-Holt
(1957) stock-recruitment relationship applies, although
F
MSY
can sometimes exceed F
max
with a Ricker (1958)
curve. For this reason, and taking into account economic
considerations, F
0.1
was developed and promoted as a
more prudent alternative (Gulland and Boerema, 1973).
Although F
0.1
is commonly interpreted as a conserva-
tive or cautious estimate of F
MSY
, this is not always the
case (Mace, 1994; Mace and Sissenwine, 1993). And
even when F
0.1
does underestimate F
MSY
, the equilibrium
yields associated with the two reference points may be
relatively very close (based on the argument that the
difference between the equilibrium yields associated
with F
max
and F
0.1
are usually small, and F
MSY
is usually
less than F
max
).
Another class of reference points that has gained
prominence as proxies or independent measures of tar-
gets and limits are those based on F
%SPR
. In particular,
values in the range F
20%
to F
30%
have frequently been
used to characterize recruitment overfishing thresholds
(Rosenberg et al., 1994), while values in the range F
30%
to F
40%
have been used as proxies for F
MSY
. These de-
faults are supported by Mace and Sissenwine (1993) who
advocated F
20%
as a recruitment overfishing threshold
for well-known stocks with at least average resilience
and F
30%
as a recruitment overfishing threshold for less
well-known stocks or those believed to have low resil-
ience, by Clark (1991, 1993) who advocated F
35%
as a
robust estimator of F
MSY
applicable over a wide range of
life histories, or F
40%
if there is strong serial correlation
in recruitment, and by Goodyear (1993) who advocated
at least 20% SPR unless there were evidence of excep-
tionally strong density dependence.
Finally, in the uncommon situation where stocks
have been maintained near B
MSY
, F
med
may be consid-
ered a reasonable proxy for F
MSY
.
Proxies for B
MSY
The equilibrium biomass corresponding to the
above-mentioned fishing mortality reference points can
be used as proxies for B
MSY
. In addition, B
MSY
has been
approximated by various percentages of the unfished
biomass, usually in the range 30-60% B
0
(higher per-
centages being used for less resilient species, and lower
percentages for more resilient species). B
MSY
can also
be approximated by the mean recruitment (R
mean
) multi-
plied by either (a) the level of spawning per recruit at
F
MSY
; namely SPR(F
MSY
), or some proxy thereof, or (b)
30-60% SPR
F=0
. Note that if F
MSY
is overestimated, then
SPR(F
MSY
) and B
MSY
will both be underestimated, thus
compounding the riskiness of control rules that use esti-
mates of F
MSY
and B
MSY
in combination.
If catch and CPUE data are available, production
models may provide useful proxies such as CPUE
MSY
which can be used as a relative index of B
MSY
(in addi-
tion, the nominal effort (e.g., in boat-months) corre-
sponding to F
MSY
can be used as a relative index of F
MSY
).
The risks of using CPUE as an index of or proxy
for stock size with associated assumptions of constant
catchability over all stock sizes and time may be size-
able and have been well-described (e.g., Hilborn and
Walters, 1992).
42
Proxies for B
0
Where B
0
is unknown, it can be approximated by
the product of average recruitment and SPR
F=0
(Myers
et al., 1994); however, this approximation assumes that
there have been no density-dependent changes in growth,
survival, or age at maturity during the “fishing down”
period.
Proxies for MSY
The equilibrium yield corresponding to the above-
mentioned F and/or B reference points can be used as a
proxy for MSY, although of course such estimates of
MSY must be considered long-term averages, and not
treated as constant annual catches. For a fishery where
annual quotas remain constant over a prolonged period
(perhaps because there are insufficient data to update
stock assessments), such quotas should be set at a level
of 60-90% of the equilibrium or static estimate of MSY,
with the high end of the scale applying to species with
low natural variability or low M, and the low end apply-
ing to species with high natural variability or high M
(Mace and Sissenwine, 1989).
Constraints on Acceptable Proxies
In addition, there are a number of estimators of, or
approximations to, the limit reference F
t
points based
on the slope at the origin of stock-recruitment relation-
ships (variously called F
extinction
, F
ext
, F
τ
(Mace, 1994),
F
crash
(ICES 1997a)). These estimators include F
med
(if
calculated from data collected during a period when the
stock was at low biomass), F
high
(the fishing mortality
corresponding to the 90th percentile of survival ratios),
F
20%
, F
loss
(the fishing mortality corresponding to the
lowest observed spawning stock and associated recruit-
ment — Cook, 1998), and F
COMFIE
(the minimum of F
MSY
,
F
med
and F
crash
). Suggested biomass limits that have been
considered dangerously close to the origin include
MBAL (the minimum biologically acceptable level of
spawning biomass; Serchuk and Grainger, 1992), B
50%R
(the spawning biomass corresponding to 50% of the
maximum recruitment in a stock recruitment relation-
ship; Mace, 1994; Myers et al.; 1994), B
90%R,90%R/S
(the
biomass corresponding to the intersection of the 90th
percentile of observed recruitment and the 90th percen-
tile of survival; Serebryakov, 1991; Shepherd, 1991),
and B
loss
(the biomass corresponding to the lowest ob-
served spawning stock; ICES, 1997a). Any proxies used
for F
MSY
or B
MSY
should be more conservative than these
extremes.
Recommended Data-Moderate Defaults
The recommended data-moderate default limit con-
trol rule is the limit control rule described above as in
Restrepo et al., 1998, using proxies for F
MSY
and B
MSY
as
described below.
We recommend that fishing mortality rates in the
range F
30%SPR
to F
40%SPR
be used as general default prox-
ies for F
MSY
, in cases where the latter cannot be reliably
estimated. In the absence of data and analyses that can
be used to justify alternative approaches, it is recom-
mended that F
30%SPR
be used for stocks believed to have
relatively high resilience, F
40%SPR
for stocks believed to
have low to moderate resilience, and F
35%SPR
for stocks
with “average” resilience. Less-preferred alternatives
(in order of decreasing preference) are to use F
0.1
, M,
F
max
, or F
med
(when F
med
is calculated from data collected
when the stock was believed to be fluctuating around
B
MSY
) as the proxies for F
MSY
. The equilibrium or aver-
age biomass levels corresponding to these fishing mor-
tality rates should then be used as proxies for B
MSY
, in
the same order of preference. The default limit control
rule would then be defined with fishing mortality set to
this default level when biomass exceeds (1-M)*B
MSY
or
1/2 B
MSY
, whichever is greater, and would decline lin-
early to zero for biomass levels below this threshold..
The recommended default MSST corresponds to 1/2 B
MSY
(the absolute lowest limit triggering the need for a re-
building plan) for species with M ≥ 0.5; but occurs at a
larger biomass for species with smaller M.
Data-Poor Situations
If data are insufficient data to conduct YPR and SPR
analyses, or if estimates of F and B cannot be obtained
for comparison with YPR and SPR reference points,
there are far fewer options for defining meaningful tar-
gets and limits. Priority should be given to bringing the
knowledge base at least up to “data-moderate” standards.
Proxies for F
MSY
The natural mortality rate M has often been consid-
ered to be a conservative estimate of F
MSY
; however, it is
becoming more and more frequently advocated as a tar-
get or limit for fisheries with a modest amount of infor-
mation. In fact, in several fisheries, F=0.8*M and
F=0.75*M have been suggested as default targets for
data-poor cases (Thompson, 1993; NMFS, 1996). In
data-poor situations, M may not be reliably estimated
either, however.
Proxies for B
MSY
The equilibrium biomass corresponding to F=M or
F=0.8*M can be used as a proxy for B
MSY
. However, in
most data-poor situations, it will not be possible to cal-
culate this quantity.
43
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
Proxies for B
0
If there are no data on recruitment, some function
of CPUE might conceivably be used as a relative index
of initial biomass. If information (perhaps anecdotal)
exists on resource conditions prior to or shortly after the
onset of fishing, some inferences of initial biomass (B
0
)
may be possible. Because the geographic area occu-
pied by a stock may contract with declines in abundance,
the contrast between present and early geographic dis-
tributions of the resource may be used to obtain a rough
approximation of pre-fishery abundance. Early sport
fishing records may provide useful information on re-
source conditions prior to intense exploitation (MacCall
1996). Estimates of early CPUE may relate to B
0
, but
care must be taken to correct for the general tendency
for CPUE to underestimate declines in resource abun-
dance. For example, this may require geographic strati-
fication, correction for temporal changes in fleet com-
position (e.g., loss of less efficient vessels as catch rate
declines) and a variety of behavioral and biological in-
teractions. Nonequilibrium production modeling
(Hilborn and Walters, 1992; Prager, 1994) also may pro-
vide an inference of initial CPUE for the fishery.
Proxies for MSY
If there is absolutely no information available to
estimate fishing mortality or biomass reference points,
it may be reasonable to use the historical average catch
as a proxy for MSY, taking care to select a period when
there is no evidence that abundance was declining. In
recognition of the danger of continually setting annual
quotas at a constant level equal to the historical average
catch (a common situation in data-poor fisheries), it
might be best to scale down the historical average catch
by multiplying by a factor in the range 0.6-0.9 where
smaller multipliers would be used for highly variable
stocks and larger numbers for less variable stocks (Mace
and Sissenwine 1989).
Recommended Data-Poor Defaults
In the absence of data and analyses that can be used
to justify alternative approaches, it is recommended that
the default limit control rule be implemented by multi-
plying the average catch from a time period when there
is no quantitative or qualitative evidence of declining
abundance (“Recent Catch”) by a factor depending on a
qualitative estimate of relative stock size:
Above B
MSY
:
Limit catch = 1.0*(Recent catch)
Above MSST but below B
MSY
:
Limit catch = 0.67*(Recent catch)
Below MSST (i.e., overfished):
Limit catch = 0.33*(Recent catch).
The multipliers 1.0, 0.67 and 0.33 were derived by
dividing the default precautionary target multipliers in
Section 3.3.2 of Restrepo et al. (1998) by 0.75, in order
to maintain the 0.75 ratio recommended as the default
distance between the limit and target reference points
for stocks above (1-M)*B
MSY
. Since it probably will not
be possible to determine stock status relative to B
MSY
ana-
lytically, an approach based on “informed judgement”
(e.g., a Delphi approach) may be necessary.
Concluding Observations
The MSY-related reference points in the MSCFMA
National Standard Guidelines and the FAO guidelines
appear stringent in the context of current fishing mor-
tality rates observed in open-access fisheries. Yet, his-
torically, the risk associated with MSY-related reference
points has been well detailed from a qualitative perspec-
tive: Larkin’s (1977) famous summary cites the relative
instability of stocks harvested at MSY which may arise
from a relatively high proportion of young and first-
time spawners (which may reduce the viability of eggs
which are deposited) and the reduction in the number
of spawning age classes; the risk to local subpopula-
tions or substocks with lower productivities than the
stock as a whole; the risk to less productive co-occur-
ring species when highly productive species are fished
at MSY levels; and the risk to productivity of competi-
tors and predators when all stocks cannot be fished si-
multaneously at their respective MSY-related levels.
Typical single-species reference points still treat all units
of spawning stock biomass as equivalent, regardless of
age structure or spawning history; and rarely include
diversity of age structure as a component of the refer-
ence point. Maintenance of genetic diversity and prob-
lems of technological and biological interactions must
be dealt with through compromises, and so there con-
tinue to be elements in the fishery system which are at
risk under this approach.
For many of the model-based estimates of biologi-
cal reference points, uncertainty has been evaluated us-
ing Monte Carlo or bootstrap procedures (e.g., surplus
production models: Prager, 1994; Polacheck et al., 1993;
yield-per-recruit models: Restrepo and Fox, 1988;
Pelletier and Gros, 1991; spawning-stock-biomass-per
recruit models: Cook, 1998). As more and more infor-
mation about uncertainty becomes available or is in-
cluded in the estimation process, the estimate of uncer-
tainty related to the reference point increases. This may
occur as more sources of observation error are included,
or if process error is also included. The method used to
fit a model may also affect the the estimate of uncer-
tainty, as a function of the types of errors included in
the model (e.g., Chen and Andrews, 1997). Conse-
quently, although procedures for estimating the refer-
ence point and the underlying model may be agreed
44
within the scientific community, the procedures for es-
timating uncertainty in those reference points are less
standardized. In the most information-poor situations,
quantification of uncertainty associated with the refer-
ence point may not even be possible. Thus, paradoxi-
cally, statistical uncertainty may appear to increase while
the quality of information is increasing. The evalua-
tion of alternative reference points as proxies for MSY-
related reference points becomes problematic, because
each reference point may be estimated with a different
degree of certainty, some of which may be due to statis-
tical artifact but which may affect its performance as a
proxy in a precautionary context.
There is a fast-burgeoning body of literature which
reviews various biological reference points as candidates
for limit and target reference points in the precaution-
ary contexts of various management institutions. Those
papers almost universally endorse the evaluation of the
performance of those reference points and associated
harvest control rules using simulation modelling. We
propose that for many management systems, the results
of these simulations may show that the effect of the
choice of a particular estimate of F
MSY
or B
MSY
(or its
respective proxy) as a limit reference point may be tan-
gential to the success of a precautionary managment
regime when compared to the effects of the form of the
associated harvest control rules, and historically ob-
served implementation errors (the difference between
the intended effect of management action and the real-
ized result; e.g., the difference between total allowable
catch and actual catch in a year). There have been cases
where biological reference points and stock status have
been defined with reasonable quality data and with rea-
sonable certainty, but associated management regimes
have led to significant and undesirable stock declines.
Although the specification of MSY-related limit refer-
ence points based on poor-quality data may be daunt-
ing, for many fishery systems it is likely to be the easi-
est component of the precautionary process to imple-
ment.
Literature Cited
Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of
exploited fish populations. Fish. Invest. Ser. 2, vol. 19.
U.K. Ministry of Agriculture, Food and Fisheries, Lon-
don.
Chen, Y. and N. Andrew. 1998. Parameter estimation in
modelling the dynamics of fish stock biomass: are currently
used observation-error estimators reliable? Can. J. Fish.
Aquat. Sci. 55: 749 - 760.
Clark, W. G. 1991. Groundfish exploitation rates based on
life history parameters. Can. J. Fish. Aquat. Sci. 48:734-
750.
Clark, W. G. 1993. The effect of recruitment variability on
the choice of a target level of spawning biomass per re-
cruit. Pp. 233-246 In G. Kruse, D. M. Eggers, R. J. Marasco,
C. Pautzke and T. Quinn II (eds), Proceedings of the Inter-
national Symposium on Management Strategies for Ex-
ploited Fish Populations. Alaska Sea Grant College Pro-
gram, P. O. Box 755040, Fairbanks AK.
Cook, R. M. 1998. A sustainability criterion for the exploita-
tion of North Sea cod. ICES J. Mar. Sci. 55: 1061-1070.
Conser, R., and W. Gabriel. 1992. Probabilities associated
with biological reference points in relation to current stock
status [Abstract]. Second NMFS National Stock Assess-
ment Workshop. Southwest Fisheries Science Center, La
Jolla, CA., April, 1992.
Food and Agriculture Organization (FAO). 1995a. Precau-
tionary approach to fisheries. Part I: Guidelines on the
precautionary approach to capture fisheries and species
introductions. Elaborated by the Technical Consultation
on the Precautionary Approach to Capture Fisheries (In-
cluding Species Introductions). Lysekil, Sweden, 6-13 June
1995. FAO Technical Paper No. 350, Part 1. 52 p.
Food and Agriculture Organization (FAO). 1995b. Code of
conduct for responsible fisheries. FAO Mimeo. Rome,
FAO. 41 p.
Garcia, S. M. 1995. The precautionary approach to fisheries
and its implications for fisher research, technology and
management: an updated review. In Food and Agricultural
Organization (FAO)1995a. Precautionary approach to fish-
eries. Part II: Scientific papers. Prepared for the Technical
Consultation on the Precautionary Approach to Capture
Fisheries (Including Species Introductions). Lysekil, Swe-
den, 6-13 June 1995. FAO Technical Paper No. 350, Part
2. 210 p.
Gallucci, V. F., B. Amjoun, J. Hedgepeth, and H. L. Lai. 1996.
Size-based methods of stock assessment of small-scale fish-
eries In Gallucci, V. F., S. B. Saila, D. J. Gustafson and B.
J. Rothschild [eds.]. Stock assessment: quantitative meth-
ods and applications for small-scale fisheries. CRC Press.
Boca Raton, FL. 515 pp.
Goodyear, C.P. 1993. Spawning stock biomass per recruit in
fisheries management: foundation and current use. pp. 67-
81 In S.J. Smith, J.J. Hunt and D. Rivard [eds.]. Risk evalu-
ation and biological reference points for fisheries manage-
ment. Can. Spec. Publ. Fish. Aquat. Sci. 120.
Gulland, J.A. and L.K. Borema. 1973. Scientific advice on
catch levels. Fishery Bulletin 71(2):325-335.
Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries
stock assessment: choice, dynamics anduncertainty.
Chapman and Hall, N.Y. 570 p.
International Council for the Exploration of the Sea (ICES).
1984. Report of the working group on methods of fish
stock assessment. ICES CM 1984/Assess:19: 56 p.
International Council for the Exploration of the Sea (ICES)
1997a. Report of the study group on the precautionary
approach to fisheries management. ICES CM 1997/As-
sess:7. 37pp.
International Council for the Exploration of the Sea (ICES).
1997b. Report of the
Comprehensive Fishery Evaluation Working Group. ICES
45
Proceedings, 5
th
NMFS NSAW. 1999. NOAA Tech. Memo. NMFS-F/SPO-40.
CM 1997/Assess:15. 140 p.
Larkin, P. A. 1977. An epitaph for the concept of maximum
sustained yield. Trans. Am. Fish. Soc. 106: 1- 11.
MacCall, A.D. 1996. Patterns of low-frequency variability in
fish populations of the California Current. CalCOFI Reps.
37:100-110.
Mace, P.M. 1994. Relationships between common biological
reference points used as thresholds and targets of fisher-
ies management strategies. Canadian Journal of Fisheries
and Aquatic Sciences 51:110-122.
Mace, P.M. and W. Gabriel. 1999. Evolution, scope and cur-
rent applications of the precautionary approach in fisher-
ies. Proceedings of the 5th National Stock Assessment
Workshop, V. Restrepo (editor). National Marine Fisher-
ies Service, Office of Science and Technology, 1315 East-
West Highway, Silver Spring, MD 20910.
Mace, P.M. and M.P. Sissenwine. 1989. Biological reference
points for New Zealand fisheries assessments. New Zealand
Fisheries Assessment Research Document 89/11. 11pp.
Mace, P.M. and M.P. Sissenwine. 1993. How much spawn-
ing per recruit is enough? In S.J. Smith, J.J. Hunt and D.
Rivard [eds.] Risk evaluation and biological reference
points for fisheries management. Canadian Special Publi-
cations in Fisheries and Aquatic Sciences 120:101-118.
Myers, R.A., A.A. Rosenberg, P.M. Mace, N. Barrowman,
and V.R. Restrepo. 1994. In search of thresholds for re-
cruitment overfishing. ICES Journal of Marine Science
51:191-205.
National Marine Fisheries Service (NMFS). 1996. Environ-
mental Assessment/Regulatory Impact Review for Amend-
ment 44 to the Fishery Management Plan for the Ground-
fish Fishery of the Bering Sea and Aleutian Islands Area
and Amendment 44 to the Fishery Management Plan for
the Groundfish Fishery of the Gulf of Alaska to Redefine
Acceptable Biological Catch and Overfishing, Appendix
B. Alaska Fisheries Science Center, National Marine Fish-
eries Service, 7600 Sand Point Way NE., Seattle, WA
98115-0070.
Pelletier, D. and P. Gros. 1991. Assessing the impact of sam-
pling error on model-based management advice: compari-
son of equilibrium yield per recruit variance estimators.
Can. J. Fish.Aquat. Sci. 48: 2129-2139.
Polacheck, T., R. Hilborn and A. Punt. 1993. Fitting surplus
production models: comparing methods and measuring
uncertainty. Can. J. Fish. Aquat. Sci. 50: 2597-2607.
Prager, M. H. 1994. A suite of extensions to a nonequilibrium
surplus-production model. Fish. Bull. 92: 374-389.
Restrepo, V. R. and W. W. Fox. 1988. Parameter uncertainty
and simple yield-per-recruit analysis. Trans. Am. Fish.
Soc. 117: 282-289.
Restrepo, V. R, G. G. Thompson, P. M. Mace, W. L. Gabriel,
L. L. Low, A. D. MacCall, R. D. Methot, J. E. Powers, B.
L. Taylor, P. R. Wade, and J. F. Witzig. 1998. Technical
Guidance on the Use of Precautionary Approaches to Imple-
menting National Standard 1 of the Magnuson-Stevens
Fishery Conservation and Management Act. NOAA Tech-
nical Memorandum NMFS-F/SPO-31. 54 p.
Ricker, W. E. 1958. Handbook of computations for biologi-
cal statistics of fish populations. Fish. Res. Bd. Canada,
Bull. No. 119.
Ricker, W. E. 1975. Computation and interpretation of bio-
logical statistics of fish populations. Fish. Res. Bd.
Canada, Bull. No. 191.
Rosenberg, A., P. Mace, G. Thompson, G. Darcy, W. Clark,
J. Collie, W. Gabriel, A. MacCall, R. Methot, J. Powers,
V. Restrepo, T. Wainwright, L. Botsford, J. Hoenig, and
K. Stokes. 1994. Scientific review of definitions of over-
fishing in U.S. Fishery Management Plans. NOAA Tech.
Memo. NMFS-F/SPO-17. 205 pp.
Rosenberg, A., and V. Restrepo. 1995. Precautionary man-
agement reference points and management strategies. In
Food and Agricultural Organization (FAO)1995a. Precau-
tionary approach to fisheries. Part II: Scientific papers.
Prepared for the Technical Consultationon the Precaution-
ary Approach to Capture Fisheries (Including Species In-
troductions). Lysekil, Sweden, 6-13 June 1995. FAO
Technical Paper No. 350, Part 2. 210 p.
Serchuk, F.M. and R.J.R. Grainger. 1992. Development of
the basis and form of ICES fisheries management advice:
historic background (1976-1990) and the new form of
ACFM advice. ICES C.M. 1992/Assess:20. 13pp.
Serchuk, F., D. Rivard, J. Casey and R. Mayo. 1997. Report
of the ad hoc working group of the NAFO Scientific Coun-
cil on the precautionary approach. NAFO SCS Document
97/12.
Serebryakov, V.P. 1991. Predicting year-class strength under
uncertainties related to survival in the early life history of
some North Atlantic commercial fish. NAFO Science
Council Studies 16:49-56.
Shepherd, J.G. 1991. Report of special session. NAFO Sci-
ence Council Studies 16:7-12.
Shepherd, J. G. 1982. A versatile new stock-recruitment rela-
tionship of fisheries and construction of sustainable yield
curves. Cons. Perm. Int. Explor. Mer. 40: 67-75.
Sissenwine, M. P. and J. G. Shepherd. 1987. An alternative
perspective on recruitment overfishing and biological ref-
erence points. Can. J. Fish. Aquat.Sci. 44: 913 - 918.
Thompson, G. G. 1993. A proposal for a threshold stock size
and maximum fishing mortality rate. In S. J. Smith, J. J.
Hunt, and D. Rivard (editors), Risk evaluation and bio-
logical reference points for fisheries management, p.
303-320. Can. Spec. Publ. Fish. Aquat. Sci. 120.
Thompson, W. F. and F. H. Bell. 1934. Biological statistics
of the Pacific halibut fishery. 2. Effects of changes in in-
tensity upon total yield and yield per unit of gear. Rep.
Int. Pac. Halibut Comm. 8: 49 p.
United Nations (UN). 1995. Agreement for the implementa-
tion of the provisions of the United Nations convention
on the law of the sea of 10 December 1982 relating to the
conservation and management of straddling fish stocks and
highly migratory fish stocks. United Nations Conference
on Straddling Fish Stocks and Highly Migratory Fish
Stocks. Sixth session. New York, 24 July - 4 August
1995. A/CONF. 164/37 8 September 1995.
