Review of potential collision between tidal stream devices and marine animals

Page 1 of 69

Review of potential collision between

tidal stream devices and marine

animals

Report No: 444

Author Name: Frost, N., Goldsmith, N., San Martin, E., West, V.

Author Affiliation: ABPMer

June 2020

Page 2 of 69

About Natural Resources Wales

Natural Resources Wales’ purpose is to pursue sustainable management of natural

resources. This means looking after air, land, water, wildlife, plants and soil to improve

Wales’ well-being, and provide a better future for everyone.

Evidence at Natural Resources Wales

Natural Resources Wales is an evidence-based organisation. We seek to ensure that our

strategy, decisions, operations and advice to Welsh Government and others are

underpinned by sound and quality-assured evidence. We recognise that it is critically

important to have a good understanding of our changing environment.

We will realise this vision by:

• Maintaining and developing the technical specialist skills of our staff;

• Securing our data and information;

• Having a well resourced proactive programme of evidence work;

• Continuing to review and add to our evidence to ensure it is fit for the challenges facing

us; and

• Communicating our evidence in an open and transparent way.

This Evidence Report series serves as a record of work carried out or commissioned by

Natural Resources Wales. It also helps us to share and promote use of our evidence by

others and develop future collaborations. However, the views and recommendations

presented in this report are not necessarily those of NRW and should, therefore, not be

attributed to NRW.

Report series: NRW Evidence Reports

Report number: 444

Publication date: June 2020

Contractor: ABPMer

Contract Manager: L.G. Murray

Title: Review of potential collision between tidal stream devices and marine animals

Author(s): Frost, N., Goldsmith, N., San Martin, E., West, V.

Restrictions: None

ABPmer Document reference: Report No. R.3322

ABPmer Project no.: R/4780/1

ABPmer preparation by N J Goldsmith; Approved by N J Frost; Authorised by S C Hull.

Page 3 of 69

ABPmer Notice:

ABP Marine Environmental Research Ltd ("ABPmer") has prepared this document in

accordance with the client’s instructions, for the client’s sole purpose and use. No third

party may rely upon this document without the prior and express written agreement of

ABPmer. ABPmer does not accept liability to any person other than the client. If the client

discloses this document to a third party, it shall make them aware that ABPmer shall not

be liable to them in relation to this document. The client shall indemnify ABPmer in the

event that ABPmer suffers any loss or damage as a result of the client’s failure to comply

with this requirement.

Sections of this document may rely on information supplied by or drawn from third party

sources. Unless otherwise expressly stated in this document, ABPmer has not

independently checked or verified such information. ABPmer does not accept liability for

any loss or damage suffered by any person, including the client, as a result of any error or

inaccuracy in any third party information or for any conclusions drawn by ABPmer which

are based on such information.

ABPmer: Quayside Suite, Medina Chambers, Town Quay, Southampton, Hampshire

SO14 2AQ. T: +44 (0) 2380 711844 W: http://www.abpmer.co.uk/

Page 4 of 69

Distribution List (core)

NRW Library, Bangor 2

National Library of Wales 1

British Library 1

Welsh Government Library 1

Scottish Natural Heritage Library 1

Natural England Library (Electronic Only) 1

Recommended citation for this volume:

ABPmer 2020. Review of Potential Collision Between Tidal Stream Devices and Marine

Animals. NRW Evidence Report No: 444, 65 pp, NRW, Bangor.

Page 5 of 69

Contents

About Natural Resources Wales .......................................................................................... 2

Evidence at Natural Resources Wales ................................................................................ 2

Distribution List (core) .......................................................................................................... 4

Recommended citation for this volume: ............................................................................... 4

Contents .............................................................................................................................. 5

List of Tables ....................................................................................................................... 7

1. Crynodeb Gweithredol ..................................................................................................... 8

2. Executive summary........................................................................................................ 10

3. Introduction .................................................................................................................... 12

4. Approach ....................................................................................................................... 15

4.1. Task 1 – Identified series of projects/devices ....................................................... 15

4.2. Task 2 – Set up a standardised evidence template .............................................. 16

4.3 Task 3 – Evidence gathering and gap analysis ........................................................ 18

5. Evidence review ............................................................................................................. 20

5.1. Marine mammals ..................................................................................................... 20

5.1.1. Monitoring approaches ..................................................................................... 20

5.1.2. Monitoring studies and results .......................................................................... 23

5.1.3. Wider evidence and assessment of collision .................................................... 25

5.1.4. Summary of current knowledge ........................................................................ 27

5.2. Seabirds ............................................................................................................... 28

5.2.1. Monitoring approaches ..................................................................................... 28

5.2.2. Monitoring studies and results .......................................................................... 29

5.2.3. Wider evidence and assessment of collision .................................................... 30

5.2.4. Summary of current knowledge ........................................................................ 31

5.3. Fish ....................................................................................................................... 32

5.3.1. Monitoring approaches ..................................................................................... 32

5.3.2. Monitoring studies and results .......................................................................... 34

5.3.3. Wider evidence and assessment of collision .................................................... 37

5.3.4. Summary of current knowledge ........................................................................ 40

6. Gap Analysis .................................................................................................................. 46

Page 6 of 69

7. Recommendations for Addressing Key Gaps ................................................................ 50

8. Conclusions ................................................................................................................... 52

9. References .................................................................................................................... 54

10. Appendices .................................................................................................................. 59

Appendix A - Agreed List of Tidal Devices ..................................................................... 59

Appendix B - Full list of Contacted organisations ........................................................... 68

Data Archive Appendix ...................................................................................................... 69

Page 7 of 69

List of Tables

• Table 1: Overview of devices and developers: Evidence spreadsheet. ....................... 17

•

Table 2: Parameters assessed for each tidal development: Evidence spreadsheet. ... 17

•

Table 3: References spreadsheet. ............................................................................... 18

•

Table 4: Summary table of the collision risk monitoring techniques used to date. ....... 42

•

Table 5: Gap analysis of evidence available for undertaking robust collision risk

assessments. ............................................................................................................... 47

•

Table 6: Initial long list of tidal stream energy devices/developers agreed with NRW

and the status of the monitoring and/or reporting. ........................................................ 59

•

Table 7: Organisations contacted (same order as Appendix A) that have/had devices in

situ or pre-consent stage. ............................................................................................. 68

•

Table 8: Organisations contacted (same order as Appendix A) that are involved in tidal

energy. ......................................................................................................................... 68

Page 8 of 69

1. Crynodeb Gweithredol

Mae gwrthdrawiadau posibl rhwng anifeiliaid y môr a dyfeisiau llif llanw yn bryder o ran

trwyddedu defnydd o'r fath oherwydd ansicrwydd presennol. Comisiynodd Cyfoeth

Naturiol Cymru ABPmer i gasglu ac adolygu unrhyw ddata sydd ar gael a gasglwyd oddi

wrth ddyfeisiau yn y fan a'r lle, a'r llenyddiaeth ehangach, a ymchwiliodd i wrthdrawiadau

rhwng mamaliaid y môr, adar môr a physgod a dyfeisiau llif llanw. Ar sail canlyniadau'r

adolygiad hwn, darperir argymhellion ar y bylchau allweddol sy'n parhau yn y dystiolaeth

ac sydd angen eu llenwi er mwyn cefnogi'r gwaith o gydsynio ac asesu datblygiadau o

fewn y sector hwn yng Nghymru ymhellach.

Roedd y dull o fynd ati i gynnal yr adolygiad tystiolaeth hwn yn cynnwys tair tasg allweddol.

Roedd Tasg 1 yn gofyn am nodi'r holl brosiectau llif llanw a gynlluniwyd ac a weithredwyd

yn fyd-eang a dod i benderfyniad ynglŷn â'u prosesau o ran asesu a monitro

gwrthdrawiadau posibl. O hyn, dewiswyd is-set o 21 o ddyfeisiau/datblygwyr/prosiectau llif

llanw a oedd wedi monitro neu asesu'r risg gwrthdrawiad er mwyn llywio'r adolygiad

tystiolaeth hwn.

Roedd Tasg 2 yn gofyn am sefydlu templed safonol i gasglu manylion monitro ac asesu

gwrthdrawiadau ar gyfer pob un o'r prosiectau llif llanw a ddewiswyd. Nod hyn oedd

sicrhau bod y dystiolaeth wedi'i chasglu mewn modd cyson, a bod yr holl ganlyniadau

wedi'u darparu ar daenlen ar wahân. Darparwyd taenlen hefyd yn cynnwys amrediad

eang o lenyddiaeth yn ymwneud â gwrthdrawiadau posibl rhwng anifeiliaid morol a

dyfeisiau llif llanw, dulliau monitro, datganiadau amgylcheddol, modelu risg

gwrthdrawiadau, ac unrhyw lenyddiaeth berthnasol arall.

Roedd Tasg 3 yn gofyn am gasglu data monitro a gwybodaeth o'r datblygiadau/dyfeisiau

llif llanw a ddewiswyd, o nifer o ffynonellau, gan gynnwys trafodaethau â datblygwyr ac

academyddion llif llanw, yn ogystal ag adolygu asesiadau amgylcheddol, adroddiadau

monitro, a’r llenyddiaeth ehangach o ymchwil a adolygwyd gan gymheiriaid ac adolygiadau

strategol. Wedyn, adolygwyd y wybodaeth hon er mwyn penderfynu: dulliau presennol o

fonitro gwrthdrawiadau posibl rhwng dyfeisiau llif llanw a phob un o'r tri grŵp derbynnydd

ar gyfer anifeiliaid morol (mamaliaid morol, adar môr a physgod); gwerth ac effeithiolrwydd

y gwaith monitro hwn; dulliau o ddeall effeithiau posibl gwrthdrawiadau; y bylchau

allweddol mewn data ac argymhellion ar gyfer ymchwil i'r dyfodol; ac, yn olaf, sut y gellid

trosglwyddo'r holl wybodaeth hon, o bosib, i ddatblygiadau llif llanw yn nyfroedd Cymru.

Darganfu'r adolygiad hwn fod technegau monitro yn y maes a ddefnyddiwyd i benderfynu

patrymau dosbarthiad gofodol-amserol anifeiliaid morol (yn bennaf mamaliaid morol ac

adar môr) wedi darparu gwybodaeth werthfawr ar gyfer disgrifio presenoldeb, dosbarthiad

a bregusrwydd tebygol rhywogaethau o ran dyfeisiau llif llanw. Cafodd yr astudiaethau

arsylwi gweledol cychwynnol hyn eu cwblhau cyn i ddyfais gael ei gosod (monitro

gwaelodlin) ac ar ôl ei gosod (monitro effaith) fel ei bod yn bosib monitro symudiadau

mewn dosbarthiad (e.e. osgoi o bell). Maent hefyd yn darparu amcangyfrifon dwysedd sy'n

baramedr mewnbwn angenrheidiol ar gyfer modelu risg gwrthdrawiadau. Nid yw'r dulliau

hyn yn darparu tystiolaeth uniongyrchol o wrthdrawiadau ond maent yn galluogi'r gwaith o

asesu a monitro rhai o'r effeithiau a achosir o ganlyniad i osod y dyfeisiau.

Hyd yn hyn, nid yw'r un o'r astudiaethau monitro ar famaliaid morol ac adar môr wedi

cofnodi gwrthdrawiad uniongyrchol yn erbyn dyfais lanwol. Fodd bynnag, bu problemau

Page 9 of 69

methodolegol (e.e. cymal diffodd, diffyg dadansoddi'r holl ddata oedd ar gael a/neu ddiffyg

monitro gwirioneddol o wrthdrawiad uniongyrchol) sy’n awgrymu na fyddai gwrthdrawiad

wedi cael ei ganfod pe byddai wedi digwydd. Cofnodwyd gwrthdrawiadau yn erbyn

tyrbinau llanwol mewn un o'r astudiaethau monitro ar bysgod, yn arbennig pysgod ifanc

sy'n heigio. Ceir prinder o ddata monitro gan mai nifer fach yn unig o ddyfeisiau llanwol

sydd wedi'u gosod a’u monitro hyd yn hyn. Er hyn, mae'r data sydd wedi'i gasglu hyd yn

hyn yn darparu tystiolaeth werthfawr ar ymddygiad (e.e. osgoi o bell) a gorgyffyrddiad

tebygol gwahanol rywogaethau morol o gwmpas dyfeisiau. Mae data symudiadau tri

dimensiwn ar raddfa fân, trwy ddyfeisiau telemetreg a hydroacwstig, wedi darparu ychydig

o dystiolaeth gychwynnol ar achosion o osgoi agos. Fodd bynnag, mae'r dulliau hyn yn

gymharol gostus ac yn cynhyrchu maint sylweddol o ddata sydd angen llawer o amser ac

adnoddau i'w brosesu a dadansoddi.

Modelu sy'n parhau i fod y dull mwyaf cyffredin a ddefnyddir er mwyn asesu risg

gwrthdrawiad o ran anifeiliaid y môr. Mae amrediad o offerynnau modelu risg

gwrthdrawiadau ar gael, a phob un ohonynt â gofynion paramedr mewnbwn a

rhagdybiaethau gwahanol, sydd yn aml yn geidwadol. Mae'n ymddengys mai dilysiad

cyfyngedig o'r modelau hyn a gafwyd o ran y canlyniadau monitro yn ystod y cyfnod

gweithredol. Felly, lefel isel o hyder sydd yng nghanlyniadau'r offerynnau modelu hyn,

ond, hyd yn hyn, dyma yw'r ffordd orau o asesu'r risg bosibl o ran gwrthdrawiadau.

Mae'r bylchau allweddol mewn tystiolaeth ar gyfer pob anifail morol yn ymwneud â

chyfraddau osgoi neu daro, a hefyd cadarnhau os yw gwrthdrawiad gwirioneddol wedi

digwydd a beth fyddai effeithiau gwrthdrawiad. Yn ogystal, mae'r data monitro cyfyngedig

sydd ar gael ar hyn o bryd yn benodol i rywogaethau, lleoliadau a dyfeisiau, ac felly efallai

na fydd modd ei drosglwyddo, neu na fydd modd ei gymhwyso, i waith asesu prosiectau llif

llanw eraill. Mae bylchau allweddol eraill yn cynnwys goblygiadau posibl marwolaeth o

ganlyniad i wrthdrawiad ar lefel y boblogaeth a'r effeithiau cronnus yn sgil gosod dyfeisiau

ac araeau llanwol lluosog yn yr amgylchedd morol.

Un o'r prif argymhellion ar gyfer ymdrin â'r bylchau allweddol hyn yw casglu tystiolaeth

bellach ar ymddygiad o dan y dŵr (gan gynnwys osgoi agos) fel ei bod yn bosib creu

cyfraddau osgoi cadarn. Gellid archwilio technolegau eraill, fel synwyryddion gwasgedd a

osodir ar lafnau neu ddelweddau hydroacwstig (sy'n dechnoleg sy'n datblygu'n gyflym), a'u

datblygu ymhellach er mwyn cadarnhau a ydynt yn effeithiol wrth benderfynu a oes

gwrthdrawiad wedi digwydd. Mae'n ofynnol hefyd casglu rhagor o wybodaeth o ran

goblygiadau ffisegol gwrthdrawiad (â'r llafn neu’r differyn yn y gwasgedd) er mwyn deall y

posibilrwydd o farwolaeth neu anaf yn llawn.

Efallai y bydd rhywfaint o dystiolaeth berthnasol neu wersi y gellir eu dysgu o fathau tebyg

eraill o ddatblygiad (e.e. lagwnau llanwol neu brosiectau amrediad llanw) sydd â'r potensial

i arwain at wrthdrawiad, ond canolbwyntiodd yr adolygiad hwn yn gyfan gwbl ar ddyfeisiau

llanw llif. O ddiddordeb arbennig yw effaith unrhyw ddifferyn mewn gwasgedd a achoswyd

gan y llafnau wrth iddynt droi. Cynhaliwyd ymchwil i hyn mewn dyfeisiau ynni cefnforol

eraill ond prin yw'r dystiolaeth sydd ar gael o ddyfeisiau llif llanw ar hyn o bryd.

Page 10 of 69

2. Executive summary

The potential for collision between marine animals and tidal stream devices is a concern in

relation to consenting such deployments due to current uncertainties. NRW commissioned

ABPmer to collate and review any available data collected from in situ devices and the

wider literature that investigated collision between marine mammals, seabirds and fish and

tidal stream devices. Based on the outcomes of this review, recommendations are

provided on the key outstanding evidence gaps that need to be resolved to further support

consenting and assessment of developments within this sector in Wales.

The approach to this evidence review comprised three key tasks. Task 1 involved

identifying all planned and implemented tidal stream projects globally and determining their

assessment and monitoring of potential collision. From this, a subset of 21 tidal stream

devices/developers/projects that had monitored or assessed the risk of collision was

selected to inform this evidence review.

Task 2 involved setting up a standardised template to collate details of the collision

monitoring and assessment for each of the selected tidal stream projects. This was

designed to ensure the evidence was captured in a consistent manner, with all results

provided in a separate spreadsheet. An additional spreadsheet was also provided

containing a wide range of literature relating to potential collision between marine animals

and tidal stream devices; monitoring methods; environmental statements; collision risk

modelling and any other relevant literature

Task 3 involved gathering monitoring data and information from the selected tidal stream

developments/devices from a number of sources, including discussions with tidal stream

developers and academics as well as reviewing environmental assessments, monitoring

reports and wider literature from peer-reviewed research and strategic reviews. A review

of this information was then undertaken to determine: the current methods of monitoring

potential collisions between tidal stream devices and each of the three marine animal

receptor groups (marine mammals, seabirds and fish); the value and effectiveness of this

monitoring; approaches to understanding potential impacts of collision; the key data gaps

and recommendations for future research; and finally, how all this knowledge could

potentially be transferred to tidal stream developments within Welsh Waters.

This review found that field monitoring techniques used to determine the spatial-temporal

distribution patterns of marine animals (mainly marine mammals and seabirds), provided

valuable information for describing the presence, distribution and likely vulnerability of

species to tidal stream devices. These initial visual observation studies were undertaken

both before the installation of a device (baseline monitoring) and after its deployment

(impact monitoring) enabling distribution shifts (e.g. far field avoidance) to be monitored.

They also provide density estimates that are a necessary input parameter for collision risk

modelling. These methods do not provide direct evidence of collision but enable some of

the consequences of installation of the devices to be assessed and monitored.

To date, none of the monitoring studies on marine mammals and seabirds have been able

to record a direct collision with a tidal device. This may reflect an absence of collisions or

because of methodological limitations (e.g. shut down clause, no analysis of all available

data and/or no actual monitoring of direct collision) that may have prevented detection of a

collision even if it had occurred. One of the monitoring studies undertaken on fish have

Page 11 of 69

recorded collisions with tidal turbines, particularly shoaling juvenile fish. There is a paucity

of monitoring data because there have only been a small number of tidal devices deployed

and monitored thus far. Despite this, the data that has been collected to date provides

valuable evidence on the behaviour (e.g. far-field avoidance) and likely overlap of different

marine species around devices. Fine-scale 3D movement data, through telemetry and

hydroacoustic devices, has provided some initial evidence for near-field evasions.

However, these methods are relatively costly and generate a considerable amount of data

which require a large amount of time and resource to process and analyse.

Modelling continues to be the most commonly used approach to assess the risk of collision

of marine animals. There are a range of collision risk modelling tools available, each with

different input parameter requirements and assumptions which are often conservative.

There appears to have been limited validation of these models with the results of

monitoring during operation. The level of confidence in the outputs of these modelling

tools is therefore low, but to date, they are still the best way to assess the potential risk of

collision.

The key evidence gaps for all marine animals relate to avoidance or encounter rates, as

well as confirming if an actual collision has occurred and what the effects of a collision

would be. In addition, the limited monitoring data that is currently available is species,

location and device specific and may therefore, not be transferable or applicable to the

assessment of other tidal stream projects. Other key gaps are the potential implication of

collision mortality at the population level and the cumulative effects of deploying multiple

tidal devices and arrays in the marine environment.

One of the main recommendations for addressing these key gaps is to collect further

evidence on underwater behaviour (including near field evasion) to be able to generate

robust avoidance rates. Other technologies, such as blade mounted pressure sensors or

rapidly improving hydroacoustic imagery, could be explored and developed further in order

to confirm if they are effective in determining a collision event. More information on the

physical consequences of a collision (with the blade or pressure differential) is also

required to fully understand the potential for death or injury.

There may be some relevant evidence or lessons that can be learned from other similar

types of development (e.g. tidal lagoons or tidal range projects) that have the potential to

result in a collision, but this review focussed wholly on tidal stream devices. Of particular

interest is the impact and effect of any pressure differential, caused by the rotating blades,

this has been studied in other ocean energy devices but available evidence from tidal

stream devices is currently lacking.

Page 12 of 69

3. Introduction

The marine renewable energy industry is expanding globally in response to concern

around the impacts of climate change and increased energy demands. Within the UK,

Wales has the potential for the development of diverse marine renewable technologies

(Roche et al., 2016). In line with Welsh Government’s road to decarbonisation there are

aspirations to increase the contribution of marine renewable energy to Wales’ electricity

generation, and the recent introduction of demonstration zones for tidal energy aims to

facilitate developers in device deployment (Roche et al., 2016; Welsh Government, 2019).

Potential collisions with marine animals are a concern in relation to consenting tidal stream

energy deployments and a high level of uncertainty surrounds the likelihood of collisions

and the population consequences (e.g. mortality). NRW commissioned ABPmer to review

available evidence about the interaction and collision risk of seabirds, fish and mammals

with tidal stream energy devices.

This study has involved collating and evaluating existing evidence from tidal stream

deployments in the UK and worldwide. Recommendations are provided on how the

findings of the study can be used by NRW to advise on the development of the evidence to

support the growth of tidal stream energy sector in Wales while ensuring the sustainable

management of natural resources.

The key objectives of this study were to:

• Determine the overall status of evidence relating to collision;

• Evaluate the value and effectiveness of monitoring and other approaches to

understanding potential collision risk;

• Undertake a gap analysis to understand the further data and information requirements

and associated recommendations for future research; and

• Consider how the evidence can be applied to potential tidal stream energy

developments in Wales in the knowledge of the levels of confidence and outstanding

uncertainties.

In this study ‘collision’ refers to the situation in which a transit through the swept area of a

tidal turbine would be predicted to result in either a direct physical contact between the

individual animal and the turbine blade or an indirect impact as a result of the pressure

differential associated with the turbine blade. These interactions have the potential to

result in injury or death. Where the text refers to ‘collision risk’, this is the probability of

collision for an animal when making a single transit through the swept area of a turbine.

Once account is taken of the likely number of such transits, ‘collision rate’ is the overall

number of collisions estimated within a given period (usually one year). These estimates

can then be used to determine the potential consequences for the population.

The main approaches that are used to assess the potential risk of collision between tidal

energy devices and marine animals are modelling tools, monitoring in the field and

laboratory studies. These different approaches are described in general terms below.

Further detail is provided in the subsequent evidence review with illustrated examples.

Models allow for an estimate of the number of individuals of different species that might

collide with a turbine device to be predicted. Simple models assume equal distribution of

Page 13 of 69

animals through the water column and at different times of tide, day and season. More

complex models incorporate depth distribution information and may be refined for

particular species or for a specific device design. The three main types of model available

to determine the potential collision rate in marine mammals and seabirds (and which could

also be used modified for fish) are the Encounter Rate Model (ERM), the Collision Risk

Model (CRM) and the Exposure Time Population Model (ETPM). Existing fish collision risk

models include kinematic models and agent-based models. Available models tend to

assume there is no avoidance action taken by animals. Avoidance rate is considered

separately as part of the assessment process and is usually based on judgement. Models

also tend to assume that all collisions are fatal, irrespective of the blade speed, which

varies with tidal speed, blade length and distance along blade (increasing blade speed with

distance from hub).

There are three main basic approaches to collision monitoring in the field. One involves

recording the spatial-temporal distribution of animals to estimate their density and

determine the probability of encounters with a device. The second is directly recording the

near or “far field” behaviour (e.g. avoidance or evasion, respectively) and collision of

animals with operating turbines. The third is the opportunity to examine the physical

consequences of potential collisions through post mortem examination and/or necropsy of

stranding individuals. Laboratory studies have also been used to assess the collision risk

of various turbine blade designs and survival rates of animals following a strike.

The information collated during field monitoring can be used to provide the required input

parameter data into models (e.g. density estimates) and/or to refine model predictions of

collision risk and collision/avoidance rate. Each of the monitoring approaches has different

strengths and limitations which can affect the level, resolution and/or quality of the data

that can be collected. This limits the extent of model validation, which in turn affects the

confidence in the modelled outputs, as well as what the results might mean for predictions

of population level effects.

Estimation of population level effects can be undertaken using population models which

take account of population dynamics (fecundity, lifespan etc). Such population models

have been applied to seabirds in relation to onshore and offshore wind farm collision

mortality, fish in relation to commercial fishing and power station cooling water entrapment

and for marine mammals in relation to by-catch and tidal turbine collision mortality.

A more detailed review of the main approaches outlined above available to determine

collision risk of different marine animals (namely marine mammals, seabirds and fish) with

tidal turbine devices has been undertaken and is provided in the following section. This

review examined the available evidence that has been gathered by several planned and

implemented tidal stream projects in the UK, Europe and North America.

The report has been structured as follows:

• Introduction – provides background context as to what is included within the study and

sets out the key objectives to the study;

• Approach – presents an outline of the method applied to inform the project objectives;

• Evidence review – presents the main findings of the review structured according to key

receptor groups (marine mammals, seabirds and fish); and

Page 14 of 69

• Discussion – provides a summary of the key project findings, data gaps and

recommendations.

Page 15 of 69

4. Approach

The Evidence Review captured information from tidal stream projects that are planned

and/or have been implemented as well as wider literature sources. This included

discussions with tidal stream developers and academics whose research is focussed on

collision risk as well as the review of environmental assessments, monitoring reports and

wider published materials.

The method followed for this Evidence Review was split into three key inter-linked tasks as

outlined below:

• Task 1 – Identify past, present and future tidal stream deployments for which evidence

on collision might be available;

• Task 2 – Create a standardised template to capture details of the collision evidence for

each of the planned/implemented tidal stream projects; and

• Task 3 – Evidence gathering and reporting.

Each of these tasks is explained in greater detail below.

4.1. Task 1 – Identified series of projects/devices

The most comprehensive list of planned and implemented tidal stream projects, both in the

UK and overseas, is hosted on the Tethys website (tethys.pnnl.gov). This combined with

both NRW and ABPmer knowledge was used to develop an over-arching list of tidal

stream projects (see Appendix A). Key characteristics of each of the identified projects

were collated to determine their potential relevance to understanding collision risk. This

included consideration of the following, and allowed prioritisation of efforts as part of the

evidence review process:

• Developer;

• Device

• Project;

• Location;

• Status; and

• Monitoring & Reporting.

From this list of tidal stream projects, a subset were selected to inform the evidence review

process. This resulted in a total of 21 projects being captured within the evidence review

including planned deployments, for which the likelihood of collision has been evaluated at

a pre-consented stage:

• Minesto – Deep Green Tidal Kite

• Tidal Energy Ltd – DeltaStream turbine

• Mentor Mon (Morlais Energy)

• EMEC – Multiple: Demonstration area

• Nova Innovation – NOVA M100 Turbine

• SIMEC Atlantis Energy – MeyGen and SeaGen

• SmartBay – Multiple: Demonstration area

• Atlantis Marine Energy Test Site– Multiple: Demonstration area

• Atlantis Operations – Atlantis Resources AR1500 turbine

Page 16 of 69

• Cape Sharp Tidal – OpenHydro turbine

• FORCE – Multiple: Demonstration area

• Sustainable Marine Energy: PLAT-1 and PLAT-0

• Clean Current – Clean Current Turbine

• Ocean Renewable Power – RivGen and TideGen

• Verdant Power – Gen4 Free Flow

• Atlantis Resources – AK-1000

• Perpetuus – Multiple: Demonstration area

• Sabella – D10 turbine

• SEENEOH – Multiple: Riverine test area

The list of devices/deployments to be considered within the evidence review was agreed

with NRW at the project inception phase. Several projects were not progressed within the

review either due to a lack of data availability or lack of progress with the development

(see Appendix A for full list of devices, information on location, technologies, and reasons

behind screening out/in).

4.2. Task 2 – Set up a standardised evidence template

A standardised template was set up to capture details of the collision risk evidence base

for each of the planned/implemented tidal stream projects. This was designed to ensure

the evidence was captured in a consistent manner for each of the projects reviewed.

An excel spreadsheet was set up in which to capture the evidence. The first tab of the

evidence spreadsheet presents key information for each of the reviewed projects (see

Table 1). The subsequent tabs within the spreadsheet contain more detailed information

for each of the identified projects (Table 2). A separate references spreadsheet has also

been produced to provide a comprehensive source of current tidal stream collision risk

literature (Table 3).

Page 17 of 69

Table 1: Overview of devices and developers: Evidence spreadsheet.

Table 2: Parameters assessed for each tidal development: Evidence spreadsheet.

Parameter

Description

Developer

Name of developer

Device

Details of the type of device that has been/will be deployed

Project

Project name

Location

Project location

Years Deployed/

Operational

Timescales of device deployment

Development stage

Project status, either - Pre-consent/in development,

Consented, Deployed or Decommissioned

Assessment of

Environmental Effects

A signpost as to how the potential collision risk has been

evaluated for each receptor type. This is further broken

down in to:

• Theoretical – Modelling

• Monitoring Pre-construction

• Monitoring Post-construction

Data availability

An indication as to whether the respective data is publicly

available (and where it is held)

Key references

Key references of relevance to that project

Parameter

Description

Developer

Name of developer

Device

Details of the type of device that has been/will be deployed

Project

Project name

Location

Project location

Years Deployed

Timescales of device deployment

Current Status

Project status, either - Pre-consent/in development,

Consented, Deployed, Decommissioned or Not operational

Assessment and

Monitoring Undertaken

Summary of the project specific evidence collected for the

development through:

• Modelling

• Monitoring Pre-construction

• Monitoring Post-construction

This has been further broken down for each receptor type

Results from monitoring

Results of any in situ monitoring undertaken for both pre-

construction (baseline) and post-construction (operational)

monitoring. This has been broken down for each receptor

type

Limitations of monitoring

Constraints associated with pre- and post-construction

monitoring results

Limitations of modelling

Constraints associated with modelling results

References

Key reference of relevance to a particular project

Page 18 of 69

Table 3: References spreadsheet.

4.3 Task 3 – Evidence gathering and gap analysis

Evidence gathering was focused solely on projects related to tidal stream energy

developments. It is acknowledged that wider evidence (including assessment of number of

collision events through modelling) and monitoring techniques are available for other forms

of green energy devices however these fell outside the scope of this review. Collision risk

evidence was gathered from a number of sources including:

• Project specific details in the form of environmental assessments, application and

post consent monitoring documentation;

• Interviews with developers and academics working in the field of tidal stream

collision risk; and

• Literature review.

Evidence was sourced from tidal stream developments in the UK and worldwide. This was

designed to capture all available data and evidence to inform collision risk assessment and

to identify any knowledge which could potentially be transferred to tidal stream

developments within Welsh Waters.

The types of project specific documentation reviewed included Environmental Statements,

Habitats Regulations Assessments, as well as wider assessment and application details.

These were extracted from publicly available sources such as the Tethys website, The

Crown Estate’s Marine Data Exchange, the Marine Management Organisation’s (MMO)

Parameter

Description

Reference Number

Internal reference provided to each report. This links to the

reference numbers provided in the evidence spreadsheet

Year

Year report was published

First author

First Author or Company Name

Additional authors

Other named authors or contributors

Title

Report title

Journal/ Report Number

The publishing journal for the report and report number or

reference

Type

The form of the report. This was categorised as:

• Research Paper (Original research)

• Research Paper (Review article)

• Book chapter

• Conference Presentation

• Scoping Report (EIA)

• Environmental Statement (EIA)

• Environmental Appraisal

• HRA

• SEA

• Technical Report

• Website page

Link

Website link to the report for online access where available

Page 19 of 69

marine licence public register (Marine Case Management System (MCMS)), local

authorities’ planning portals, planning inspectorate’s website and/or individual developers’

websites.

Contact was made with over 36 individuals at 30 organisations (including developers,

academics and industry bodies) to try to obtain evidence in relation to collisions which is

not necessarily in the public domain. A full list of organisations contacted can be found in

Appendix B. A total of 18 responses were received in answer to this information request.

Interviews were held with five developers/demonstration areas to obtain further evidence,

or clarification, with respect to their individual projects and any wider experiences with

respect to collision risk (European Marine Energy Centre (EMEC) (Scotland), Morlais

(Wales), NOVA Innovation (Scotland), SEENEOH (France) and Sustainable Marine

Energy (Canada)). The main point of the telephone interviews was to understand the

amount and type of monitoring that had been undertaken for the tidal device or at the test

centre. Additionally, the success and limitations of the different monitoring techniques

undertaken were discussed to understand which types of monitoring were not practical or

feasible. It was also requested that where available, any unpublished monitoring data were

shared with the project team to further develop the evidence base. Where information has

been provided from these telephone interviews it is referenced within the separate

Evidence Spreadsheet.

Similarly, discussions were also held with academics from two universities (Swansea, and

Bangor) to understand the nature of ongoing research studies and how these contribute to

the tidal stream collision risk evidence base.

A wider literature review was undertaken to determine the current state of knowledge on

collision risk with tidal stream devices globally. This included both published research

papers as well as strategic reviews (e.g. Ocean Energy Systems Technology Collaboration

Programme (OES) and The Offshore Renewables Joint Industry Programme ORJIP).

Search tools such as ScienceDirect and Google Scholar were used to identify key

literature. A systematic approach was used to identify literature through the use of pre-

defined search terms.

The tidal stream collision risk evidence derived from each of these sources was

synthesised for each of the main receptor types (marine mammals, seabirds and fish). The

evidence is detailed within the supporting standardised evidence templates and

summarised in the subsequent sections of this report. An assessment of the effectiveness

and limitations of the evidence for each receptor has also been discussed below.

Additionally, in all instances, where it was not possible to get hold of the actual data or

evidence this has been signposted within the evidence spreadsheet.

Once the evidence had been collated gaps within the current knowledge were identified.

Page 20 of 69

5. Evidence review

A synthesis of the results from the evidence review and an assessment of the predicted

collision between marine mammals, seabirds and fish and tidal stream devices is provided

below and within the separate evidence spreadsheet. Each section aims to review:

• The overall status of the evidence relating to collision by summarising in situ

monitoring approaches that have been used to date and looking at wider evidence

(including modelling techniques that have been employed to predict number of

collisions, wider academic literature and non field-based studies);

• The potential value and/or limitation of monitoring approaches to understanding

potential collision between tidal stream energy devices and marine animals; and

• Any gaps in data/information requirements to collision assessment.

Consideration has been given to how the evidence can be applied to potential tidal stream

energy developments in Wales throughout the review accounting for the levels of

confidence and outstanding uncertainties with the current monitoring approaches

5.1. Marine mammals

Marine mammals are regularly recorded foraging within high energy environments

indicating a potential risk of collision with tidal stream devices (Benjamins et al., 2015;

Copping et al., 2016). This spatial overlap means that the potential risk of collision with

such devices requires full consideration when evaluating the impacts that could arise from

a project. Marine mammals are offered high levels of international protection with any

detrimental effect needed to be fully assessed and mitigated where necessary. Several

species of marine mammal have small population sizes and therefore even a small

number of injuries/mortalities could have potential population level impacts and therefore

risk of collision is an especially challenging consenting risk for the industry.

5.1.1. Monitoring approaches

Monitoring to understand the potential for collision between marine mammals and tidal

stream devices, and the effect of these collisions, has been undertaken through three main

approaches:

• The first approach focuses on observing the spatial-temporal overlap between

marine mammals and the tidal stream device, and therefore the probability of

encounters. This approach also monitors far field avoidance;

• The second approach focuses directly on detecting collision and monitoring near

field evasion with tidal stream devices; and

• The third approach looks at the aftermath of a potential collision through post

mortem examination and/or necropsy.

Spatial and temporal overlap

The first monitoring approach aims to map the distribution and density of species within the

vicinity of a proposed and/or operational tidal stream deployment. This can be achieved by

Page 21 of 69

using visual observations via vantage point, boat or aerial surveys (for cetaceans and

seals), by placing data loggers onto the animals (for seals only) or by acoustic surveys

(cetaceans only). Each of these survey types have different methodologies depending on

the geography of the tidal stream area.

Vantage point, boat and aerial surveys are the most common example of monitoring that is

undertaken to provide baseline information in the vicinity of a proposed device and is often

continued post-deployment to understand the behavioural response of marine mammals.

The vast majority of projects included within this Evidence Review have undertaken some

form of visual observation. Vantage point surveys are undertaken from a set location,

scanning the near-field environment at regular intervals. The scanned area is usually

divided into sections to help with spatial analysis. These surveys are limited to daylight

hours and the difficulty of accurately locating sightings over large distances is well known

(JNCC, 2005). Boat and aerial surveys are undertaken along pre-defined transects often

zig-zagging over the device footprint with one or two trained observers recording all

species of marine mammal observed and the location of each sighting. Aerial surveys,

including digital aerial surveys, can be undertaken over a large area but analysis to

species level can sometimes be difficult (Hammond et al., 2017).

Data loggers can be glued onto the heads/back of seals to provide data on fine-scale

movement patterns (Hastie et al.¸ 2014). Within each datalogger a wide variety of sensors

can be housed including time-depth recorders, satellite positioning (e.g. global positioning

system (GPS)), accelerometers, and magnetometers. Each sensor provides a different

parameter which potentially can then be interpreted together to produce fine-scale 3D

movements and evidence of behaviour (e.g. rapid swim away from a turbine). However,

they are attached to only a small sample of the population. These loggers can be limited

by battery life, storage availability and can only transmit data on surfacing and connecting

with satellites or when animals are recaptured. The more sensors within the device the

greater the battery drain, and more data storage capacity needed. Another limitation of

applying data loggers is that there is also potential that once deployed the seal may not

return to the same haul out site again, this can be overcome by using certain technologies

that transmit data remotely.

Cetaceans are vocal organisms and the vocalisations can be recorded using acoustic

devices to understand distribution, behaviour, relative abundance and other aspects of

ecology (Zimmer, 2011). Acoustic receivers (such as hydrophones, C-PODS, T-PODS, or

SoundTraps) may be placed close to the tidal stream device, in the surrounding area, or

on the devices themselves. These passive acoustic monitoring (PAM) devices are placed

at either the surface or seabed and detect acoustic signals at a set frequency. Single

acoustic receivers can show the presence/absence of marine mammals, but multiple

devices can be placed within an array or cluster to locate and detect movement of animals

by investigating the detection rate/intensity at the varying devices (Hastie et al., 2014;

Williamson et al., 2015; Malinka et al., 2018). PAM can be undertaken 24 hours a day in

turbid environments, making it appropriate for these high energy environments with several

multi-year studies undertaken to date (Zimmer, 2011; Tollit et al., 2019). There are,

however, some limitations; for example, cetaceans are not constantly vocalising so there is

the potential not to detect some individuals. The range of PAM is limited, and cetacean

sound is also directional, meaning animals swimming away from the device may not be

detected.

Page 22 of 69

Direct collision monitoring

Direct monitoring approaches to detect collision use either technologies which can “see”

the device, through hydroacoustic monitoring (e.g. sonar or echosounders) or underwater

video cameras or technologies which monitor impacts with tidal stream devices (e.g.

accelerometers or strain gauges) but does not distinguish between objects (debris or

animals).

Modern hydroacoustic imaging devices (e.g. sonar, echosounders, split- and multi-beam

devices), operating at high frequencies, can acquire detailed acoustic images of the

underwater environment. Variables such as occurrence, size class and behaviour of a

variety of aquatic species of fish, seabirds, and mammals that occur in high energy marine

environments can be monitored using imaging sonar systems. The specific approach to

the deployment of such acoustic devices has varied between studies and sites. To date

hydroacoustic devices have been incorporated into the tidal stream device or positioned on

specifically designed platforms set away from the tidal stream device (Hastie et al., 2014;

Williamson et al., 2015). Each device differs in the technology used, but in general

hydroacoustic devices pulse acoustic energy from the echosounder via a transducer which

is reflected off the animal/object. The high pulse frequency enables movement to be

detected and tracked through the observational window.

In order to ascertain the movement patterns of each marine mammal species, computer

algorithms have been developed to automatically assign a detection to a certain species.

This has been achieved by hydroacoustic devices that were mounted on a vessel to

provide worked examples of each species’ movement pattern (Hastie, 2013; Hastie et al.,

2019). This automation of detections is critical in the future deployment of hydroacoustic

devices in order to increase processing speed by reducing human involvement which

currently is a major limitation for this type of monitoring.

The placement, observational window and frequency of recording varies between studies

but will result in a trade-off between data resolution and field of data capture. Only

echosounder instruments with a sufficient range provide the practical means to investigate

the behaviour of marine mammals throughout the entire water column of a typical tidal

channel. The functionality of echosounders in energetic environments has so far been

limited due to the operational difficulties of data collection in such conditions and the

intense interference caused by backscatter related to turbulence (Fraser et al., 2017; Fox

et al., 2018). Another limitation of hydroacoustic monitoring is the large amount of data

produced, which is then time consuming to analyse and restricts the ability to have long-

term datasets (Hastie et al., 2019).

Underwater video techniques involve the placement of a camera system on the turbine

structure to record the presence of marine mammals in the vicinity of the turbine. Where

possible the passage of species through the turbine device to capture any collision and/or

injury is overserved/recoded. One key limitation of video systems is that due to light

requirements they are only able to capture data during daylight hours, at certain depths

and turbidity levels. Additionally, the camera view can be obscured at times by the turbine

blades or the field of view may not cover the entire rotor area, meaning not all encounters

may be captured. Cameras are also prone to biofouling. Footage collected can provide

clear evidence of collision or injury caused to marine mammals, however data analysis is

extremely time consuming.

Page 23 of 69

Accelerometers and strain gauges have been placed on tidal stream devices with the

primary purpose of monitoring the physiological stress on the blades as the device is

operational. These monitoring devices could also provide an indication, through adverse

parameters, as to when an object collides with the blade. Marine mammals have the

largest mass and it is likely that if an animal was to collide with the blade any vibration or

reduction in velocity of the blade should be recorded. The main limitation is that the

amount of deviation from a normal reading that would indicate a collision is unknown. Even

if an abnormal reading is recorded (indicating a collision) there is no way of knowing what

object it was.

Individual consequences of collision

The third approach to monitoring collisions in marine mammals is to understand the

consequences of collision to the individual. This can involve searching and then

investigating stranded (either alive or dead) marine mammals and looking for any signs of

impact. If a collision occurs, it is unknown whether a lethal or sublethal injury could occur.

Investigation of any carcasses within the vicinity of a tidal device and wider region could

provide further information on the potential impact injury. However, after finding an injured

carcass it would be hard to ascertain how and when this injury occurred, so there are

major caveats to this monitoring. If there was an overall increase in strandings with similar

injuries within the vicinity of a new device/operation then this method could allude to a

direct impact; however, there would be high levels of uncertainty of causation.

The potential impact of a blade can also be assessed theoretically based on a series of

physiological and biological assumptions. This type of prediction has been supported by

studies using dead carcasses and subjecting them to strikes in water from an object similar

to a blade or investigating tissue tensile strength (Carlson et al., 2014; Copping et al.,

2017; Onoufriou et al., 2019).

5.1.2. Monitoring studies and results

Since the first deployment of a tidal turbine several developers have undertaken in situ

monitoring to record marine mammal species in the vicinity of tidal devices. Multiple

approaches, as discussed above, have been used for monitoring potential collision risk,

with varying degrees of success. This section reviews some of the monitoring undertaken

across a range of developments, summarises the findings of the monitoring with regards to

collision risk and discusses any limitations of the findings.

Marine Current Turbines’(MCT) SeaGen device (now owned by SIMEC Atlantis) provides

the best example of a long-term project that has undertaken multiple examples of

monitoring (both pre-deployment and during operation). The project was located in

Strangford Lough, Northern Ireland and involved implementing a range of monitoring

approaches between 2005 (pre-deployment) until decommissioning and removal in

2018/19.

Over the lifetime of the project several of the monitoring approaches (described above)

have been undertaken within the vicinity of the device. These comprised visual

observations via boat, aerial, shore and device-based surveys, harbour seal telemetry,

passive acoustic monitoring (via Timed POrpoise Detectors (T-PODs)), carcass post

Page 24 of 69

mortems and active SONAR (Royal Haskoning, 2011; Hastie et al.¸ 2014; Savidge et al.,

2014).

Following the initial commissioning of the device, three years of monitoring was required in

accordance with the Environment Monitoring Plan (2008 to 2011). This monitoring

concluded that there were “no major impacts on marine mammals” (Royal Haskoning,

2011). The monitoring identified a slight change in the distribution of species, but no barrier

effect or reduction in the overall population size (through seal haul-out sites aerial

monitoring). It should be noted, however, that there was a licence condition to stop the

operation of the device if a marine mammal was spotted within its vicinity. This meant

near-scale evasion was not able to be ascertained from the monitoring, and direct

collisions would be prevented. Initially the shutdown clause was for a 250 m buffer zone,

but when the active sonar device was active, the shutdown clause was reduced to a 30 m

buffer zone (Hastie et al., 2014; Savidge et al., 2014).

As direct collision monitoring was not able to be undertaken within Strangford Lough, novel

techniques were used to see if a population level change could be detected. Biannual

aerial surveys (from a helicopter and thermal imagery) of the known haul out sites within

the Lough and wider region were undertaken between 2006 and 2014. There was a

decline in number of seals at monitored haul out sites, but the rate of decline was continual

since 2002, and not expected to have been impacted by the SeaGen device (Savidge et

al., 2014).

More recent analysis of the harbour seal telemetry data collected between 2006 and 2010

looked in greater detail at any barrier effect and the number of transitions through the

Narrows (via the device) (Joy et al., 2018; Sparling et al., 2018). Comparisons were made

between data collected in 2006 (pre-deployment), 2008 (installation period) and 2010

(operational). The main conclusion drawn from this analysis was that the 32 tagged seals

exhibited avoidance behaviour away from the operational device suggesting there was a

reduction in potential collision risk. Sparling et al. (2018) demonstrated that the

operational device reduced the number of transitions through the Narrows, by 20 %

(overall) and 57 % (during daylight hours), however when transitions did occur, they were

approximately 250 m either side of the device.

Using the same seal telemetry data, Joy et al. (2018) looked at how this monitoring could

feed into collision risk models. By modelling the depth at which the seals transited past the

device, the number of seals that pass through the “impact zone” was reduced by 90 %

compared to modelling using only the 2D locational data. Within the study, only 10 % of

the transit lines occurred at depths at which the device was located. Therefore, in addition

to any avoidance/evasion, the period for which a seal is within the “impact zone” is greatly

reduced.

Similar to MCT’s SeaGen device, Minesto’s quarter scale device was trialled in Strangford

Lough between February 2013 and June 2014. The marine mammal observer present

recorded a 95% reduction in the presence of seals within 50 m of the device following its

deployment (Minesto, 2016). This supports the previous study that found the number of

animals in the vicinity reduced during the operational phase.

Both devices and the respective monitoring within Strangford Lough were unable to

monitor direct collision due to the shutdown licence condition described above. This means

while there was empirical evidence relating to potential collision via near field evasion and

Page 25 of 69

far field avoidance, which can inform avoidance rates in CRM/ERM, there was no direct

understanding of collision from either SeaGen or Minesto.

Different devices deployed within other locations have also monitored the spatial-temporal

overlap between the device and marine mammals. EMEC’s wildlife observations data

indicates there was a noticeable change in the distribution of marine mammals around the

test devices (Long, 2017). Low densities of cetaceans made it difficult to draw conclusions

on how they might have been affected. For seals, however, there was an initial decrease in

abundance around the turbines on and immediately after installation. Numbers appeared

to recover gradually thereafter, including when turbines were operational, although not to

baseline levels. It is thought the initial decrease related to the associated increase in

shipping activity rather than the presence or operation of the turbines.

Another method of monitoring spatial temporal overlap is PAM. PAM was undertaken

around the DeltaStream device in Wales and at the Fundy Ocean Research Centre for

Energy (FORCE) in Canada. Both the resulting analysis of the data indicated that harbour

porpoise click detection reduced close to the device when the devices were operational

(Joy et al., 2018; Malinka et al.¸ 2018; Tollit et al., 2019). Each study found that overall

detection did not change but the number of detections decreased at the PAM devices

closest to the device potentially suggesting avoidance behaviour was taken by harbour

porpoise. Results found at the DeltaStream site indicate a preference for foraging at night

with 71 % of harbour porpoise detections occurring during darkness. This limits the

effectiveness of monitoring approaches involving visual observations which are commonly

employed (Malinka et al., 2018). PAM at DeltraStream included arrays of hydrophones

which allowed animals to be localised, resulted in three detections of porpoise close to the

device when it was operational (Malinka et al., 2018).

The monitoring techniques used at MCT’s SeaGen to monitor direct collision

(hydroacoustic monitoring) have also been replicated at other locations including EMEC

and Ramsey Sound. However, the hydroacoustic device at EMEC (FLOWBEC) did not

detect any marine mammals during six 14-day deployments and data at Ramsey Sound

has not been publicly released (three months of deployment) (Williamson et al., 2017).

5.1.3. Wider evidence and assessment of collision

The wider evidence with respect to the risk of collision is also available through scientific

literature as well as from modelling studies undertaken to inform impact assessments.

The most applicable models used in the assessment of collision between marine mammals

and tidal devices are largely derived from offshore wind farm collision risk models (CRMs)

(e.g. Band, 2000, Grant et al., 2014). Some of the models have been refined further for

particular species (e.g. harbour seals, Band et al., 2016), or for a different device design

(e.g. tidal kite, Schmitt et al., 2017). Often after a model has been run, the results are

inserted into a Population Viability Analysis (PVA) to understand what the results mean for

the affected population. Recent examples of CRM/ERM and then PVA analysis was

undertaken for Morlais, due to unconfirmed phasing and number of devices, the range of

outcomes were large, with several of the estimations (with no avoidance) indicating that

the number of animals killed annually would be larger than the population size. This

highlights the inaccuracy that can exist within some of these models, especially when no

avoidance behaviour is factored in.

Page 26 of 69

In summary there are three main types of model - ERMs, CRMs and Exposure Time

Population Models (ETPM). Scottish Natural Heritage (SNH) undertook a detailed review

of modelling methods for tidal stream devices. The review included a detailed outline of

each model, the methods used to gather input parameters and guidance on how to

undertake and interpret the results (SNH, 2016). The SNH review still provides the most in-

depth overview of the knowledge on collision risk modelling and marine mammals. There

has been more recent, device specific modelling which is described below.

The approaches of the ERM and CRM are broadly similar in that they both use a physical

model of the rotor and the body size and swimming activity of the animal to estimate the

potential collision rate. The ERM model focuses on the volume per unit time swept by each

blade, while the CRM focuses on the number of animal transits through a rotating rotor and

the collision risk during each transit. In both models, the shape of the rotor blades and

animal are highly simplified, and single mean values are used for tidal current, animal and

rotor speeds (Band et al., 2016; SNH, 2016).

The ETPM uses population modelling to assess the critical additional mortality due to

collisions which would cause an adverse effect to an animal population. The model

translates that into the collision rate for each animal within the volume swept by the rotors

which would be sufficient to cause such an effect.

Schmitt et al.’s (2017) model was refined for non-static devices within the water column

and used real underwater movement data from Minesto’s tidal kite to define the

parameters. The assumptions within the previously described models is that an animal

swims at a constant speed perpendicular to the device and would therefore encounter the

device at a certain rate. However, as the device moves in this case the rate of encounter

would be lower.

All models have the limitation that they are only as good as the input data, and there is still

an overall lack of understanding around near field evasion and far field avoidance of tidal

stream devices by marine mammals, with an estimated avoidance rate between 0 and 100

% used in the SNH recommended CRM/ERM. In addition, the models presume that all

collisions would result in mortality; this is unlikely for larger animals like marine mammals

(Carlson et al., 2014; Copping et al., 2017). Work undertaken by Band et al. (2016)

addressed some of the uncertainties by including refined input parameters by including

telemetry derived movement data (Thompson et al., 2016) and reducing the likelihood of

mortality if a collision event occurs.

The density of marine mammals used as an input parameter for the models suggested

within SNH’s review are largely provided from visual observations, which as mentioned

previously are susceptible to error. The density estimates will relate to number of

mammals observed in two dimensions at the surface of the water, this does not account

for use of the water column (i.e. three dimensions) described by dive time, dive depth or

swimming profile, which will all impact on the potential collision risk. It is also possible that

if turbines act as fish aggregation devices then this might alter mammal densities in the

vicinity. Furthermore, current models are highly precautionary due to the large number of

assumptions.

In addition to monitoring and modelling studies, wider investigations have also been

undertaken to help understand collision risk, for example the impact of a collision for

harbour and grey seals in Scotland, south resident killer whales in Canada and harbour

Page 27 of 69

seals in the USA has been examined (Onoufriou et al., 2019; Carlson et al., 2014; and

Copping et al., 2017, respectively). Through direct field observations and lab studies

Carson et al. (2014) and Copping et al. (2017) investigated dead stranded animals, taking

tissue samples to understand the tensile strength of the samples and determining collision

at varying points along the body at differing speeds. These studies concluded that

collisions between the device and killer whales would likely lead to “some subcutaneous

damage…while laceration of the skin is thought unlikely”. This did not represent a lethal

impact if a collision where to occur (Carlson et al., 2014). To understand the impacts on a

smaller species, the same testing was done on harbour seals in the USA and the same

conclusions were drawn. The chance of a serious, fatal injury occurring was estimated to

be minimal (0.005 % chance). For this to a occur a unique set of circumstances had to

happen whereby a marine mammal would need to hit the tip of the blade while the blade

was at full rotational speed, with no avoidance behaviour shown. The likelihood of a

sublethal effect occurring was concluded to be more likely (Copping et al., 2017).

Onoufriou et al. (2019) simulated the potential impacts of collision using dead seal

carcasses (18 grey seals, and one harbour seal) and a boat, the keel of which had been

modified at the bow to replicate a turbine blade. Pre- and post-impact x-rays were taken to

assess the impact of the collision on the animal. The speed of the boat (blade), was the

key factor in determining the level of injury and whether it could potentially be lethal, with

speed greater than 5.1 m/s indicating lethal skeletal damage. During the study, 48 % of the

collisions produced sufficient skeletal trauma to be considered likely to have been fatal.

The study was heavily caveated, as there were multiple assumptions and acceptance of

the oversimplified impact. There were also some caveats to the study, in terms of

representing a worst-case example by using the thinnest part of the blade.

5.1.4. Summary of current knowledge

This section provides an overview of the empirical evidence from tidal stream devices and

summarises the results for the monitoring approaches described above for marine

mammals (see Table 4 for a summary of all monitoring techniques to date). Currently there

is little actual monitoring data or limited direct evidence relating to collision risk between

marine mammals and turbine devices and as such there is a large gap in current

understanding of actual encounter rates as well as direct and indirect mortality rates in the

event of a collision.

Several projects have recorded a distribution shift of marine mammals pre- and post-

instalment, and then another shift, once operation has started. Avoidance behaviour has

been exhibited at Strangford Lough, EMEC and DeltaStream. This reduces the spatial

temporal overlap between marine mammals and tidal stream devices and therefore a

reduction in potential collision that might have otherwise occurred. This far-field avoidance

is understood to some extent, but each individual population of marine mammals is likely

to exhibit a different response, and therefore the changes recorded elsewhere should not

be presumed to apply to all populations. The spatial temporal patterns for some species

derived from visual observations have shown a clear preference to certain tidal periods,

applying a tidal restriction to operation could reduce potential risk.

Overall there is no evidence of an observed collision between a marine animal and a tidal

stream device. However, this may be due to limitations of the monitoring methods (e.g.

shut down clause, partial coverage of swept area, biofouling) indicating that if a collision

Page 28 of 69

had occurred it would have not been detected, and the small amount of operational time

that has been monitored. The evidence gathered to date, indicated marine mammals show

distribution shifts away from a device, suggesting some degree of far-field avoidance.

However, there has not been enough information on near-field evasion to provide an

overall conclusion about any near-field responses. The monitoring undertaken for these

projects/studies has provided useful insights into animal movements within the vicinity of

tidal turbines and have proven the potential applicability of several technologies.

If a collision were to occur, there is not enough information on what the impact on the

individual would be. Current thinking suggests that all collisions may not be lethal, and

would depend on how fast the rotors were turning and where on the bladed the collision

occurred, with the sublethal effects hard to test and quantify (Copping et al., 2017;

Onoufriou et al., 2019).

5.2. Seabirds

Seabirds are attracted to high tidal energy areas due to increased prey resources

associated with the high-energy environment (Benjamins et al., 2015; Waggit et al., 2016).

Most species of seabird are attracted to these high energy areas, but it is diving species

that are likely to be affected the most, but all species might potentially be impacted by

surface placed devices. Both surface diving species (e.g. auks and cormorants) and

plunge diving species (e.g. gannets) can dive to depths at which a bottom mounted tidal

device could be positioned (Furness et al., 2012). Surface foraging species (e.g. gulls and

terns) could also be impacted from a surface device, with these species able to “dive” to

one or two meters. This direct overlap between the swept area of the blade/device and a

foraging seabird means that a seabird could be struck by a blade (Langton et al., 2011;

Furness et al., 2012; McCluskie et al., 2012; Benjamins et al., 2015).

5.2.1. Monitoring approaches

Monitoring to understand the potential risk of collision between seabirds and tidal stream

devices has been undertaken through two main approaches:

• The first approach focuses on understanding the spatial-temporal overlap

between seabirds and the tidal stream device, and therefore the probability of

encounters; and

• The second approach focuses directly on monitoring collision with tidal stream

devices.

Spatial and temporal overlap

Seabirds are recorded during the same visual observation surveys as marine mammals,

with experienced personal able to record both seabirds and marine mammals concurrently.

Due to the size of some seabird species, there are increased challenges of correctly

identifying and placing individual seabirds into the correct section to provide fine-scale

distribution data (Waggit et al., 2014). The range of successful species identification is

reduced compared to marine mammals. New technologies that incorporate laser range

finders into binoculars have been used during visual surveys to increase the accuracy of

the sighting locations (Cole et al.¸ 2019).

Page 29 of 69

Like marine mammal monitoring, boat and aerial surveys are undertaken along pre-defined

transects often zig-zagging over the device footprint with one or two trained observers

recording all species of seabird observed and the location of each sighting. Aerial surveys,

including digital aerial surveys, are used to survey large areas quickly and particular areas

where certain species of seabird are known to be easily flushed (fly away or dive) e.g.

divers and seaducks by the presence of a boat. If seabirds are flushed outwith the

detection area (common for sensitive species) the estimate calculated after the survey

may underestimate the true number of seabirds present. However, some species are

harder to observe from the aerial surveys due to their size and colouration (e.g. species

with dark feathering on the upperparts can be overlooked, due to the feathering blending

into the sea when viewed dorsally).

Data loggers have been glued onto the back of seabirds or attached via a harness

providing fine-scale movement patterns. Within each datalogger a wide variety of sensors

can be housed including time-depth recorders, satellite positioning (e.g. global positioning

system (GPS)) and accelerometers. Each sensor provides a different parameter which can

then be interpreted together to produce fine-scale 3D movements and evidence of

behaviour (Collins et al., 2015). These loggers can be limited by battery life, storage

availability and accepted size of device in order to avoid impacting the seabird. The more

sensors within the device, the greater the battery drain, and more data storage capacity

needed and therefore the larger size. There is also potential that once deployed the

seabird may not be recaptured. This is considered unlikely during the breeding season

when seabirds need to return to the nest, but wintering patterns are very hard to ascertain.

The most common deployment is for a short timeframe due to the size of the device

(Owen, 2015; Shoji et al., 2015; Johnston et al., 2018).

Direct collision monitoring

Direct measures to determine collision use technologies which can “see” the device, either

through hydroacoustic monitoring (e.g. sonar or echosounders) or underwater video

cameras. In theory, these technologies could detect when an object, whether it be debris,

or seabird directly collides with the device, but to date this has not happened. The

methods that are used to detect seabirds are similar to the methods used to detect marine

mammals, with all technologies applicable to both receptors. Please refer to Section 6.1.3

for full explanation of the monitoring approaches used.

A seabird specific limitation of the device, especially hydroacoustic, is that due to the small

size of seabirds it is impossible to identify the species of seabird recorded (Williamson et

al., 2017). Currently, to understand species specific direct collision monitoring, the use of

video cameras is required. However, for future projects hydroacoustic monitoring could be

used concurrently with either data loggers or vantage point surveys to identify which

species are detected on the hydroacoustic monitoring device. In reality this would be time

consuming and during mixed-species feeding would be virtually impossible.

5.2.2. Monitoring studies and results

Since the first in situ placing of a tidal turbine several developers have implemented

monitoring to record seabird species in the vicinity of tidal stream devices. Multiple

approaches, as discussed above, have been used for monitoring potential collision risk,

Page 30 of 69

with varying degrees of success. This section reviews some of the monitoring undertaken

across a range of developments, summarises the findings of the monitoring with regard to

collision between seabirds and tidal stream devices and discusses key limitations of the

findings.

EMEC provides the best example in Europe of a long-term tidal stream energy project that

has undertaken monitoring of seabirds (between 2005 and 2015) and published

interpreted results (both pre-deployment and during operation). FORCE in Canada has

also undertaken multiple years (2009 to 2012 for the baseline and then 2016 to present for

operational) of monitoring and analysis of seabird distribution. Monitoring via visual

observation through dedicated vantage point surveys and the use of hydroacoustic devices

have been deployed in both locations to monitor seabird distribution and potential collision.

Other projects/devices which have published data on seabirds include, SeaGen in

Northern Ireland, Verdant Power in USA and DeltaStream in Wales.

During vantage point surveys at EMEC, FORCE and SeaGen changes in the distribution of

seabird species between pre-deployment, instalment and operational phases were

observed (Robbins, 2012; Savidge et al., 2014; Long, 2017; Envirosphere, 2018). Several

seabird species increased in abundance in the vicinity of the device, for example

cormorants were more abundant after the device was installed (but not operational), likely

due to the devices acting as fish aggregation devices (FAD) (Long, 2017). Once the

devices become operational a distribution shift was observed and avoidance occurred for

some species (including cormorants and divers) (Long, 2017; Envirosphere, 2018). In

contrast, Verdant Power’s Roosevelt Island Tidal Energy (RITE) Project observed no

measurable change in the number of diving species following the installation of the devices

(Double-breasted Cormorant specifically). However, it should be noted that the location of

RITE restricted the number of seabirds present due to the urbanised riverine location close

to New York (Verdant Power, 2010).

Acoustic monitoring of seabirds has often been a secondary aim of the hydroacoustic

devices deployed at tidal stream projects. This is due to the difficulty in identifying species

owing to the relatively small size of seabirds compared to marine mammals. However,

several hydroacoustic devices have successfully tracked seabirds by tracking dives on

acoustic imagery (Savidge et al., 2014). The FLOWBEC platform deployed at EMEC was

specifically designed to monitor seabirds, by using novel algorithms that aid detection of

seabirds within high energy areas (Williamson et al., 2019). Six trials lasting 14 days each

detected a single seabird at both the control location and the location with a turbine

present. During the short trial period the technology was proven, and the algorithms

refined to ensure the technology would be able to record a collision (if one were to

happen). Similarly, at the DeltaStream device in Ramsey Sound, the hydroacoustic device

detected seabirds on multiple occasions, with no collision observed.

5.2.3. Wider evidence and assessment of collision

The wider evidence with respect to the risk of collision is also available through scientific

literature as well as from modelling studies undertaken to inform impact assessments.

Page 31 of 69

Similar to marine mammals, SNH undertook a detailed review of different models that

predict collision between seabirds and tidal stream devices (see Section 5.1.3). The SNH

review still provides the most comprehensive knowledge on modelling for seabirds.

Several recent examples of the models are described below.

Recent assessments in support of applications for tidal stream projects have used both the

CRM and ERM models to predict collision for seabirds. As there is no consensus on which

is the most appropriate for use underwater both of the methods are often used (nrp, 2014:

Minesto, 2016; Morlais, 2019). Similar input parameters are required for all of the models

including, body length, time at the swept area depth, swim speed and density, but none

are able to provide an avoidance rate, so each model includes avoidance estimates

between 50 and 99 %. This wide variety of avoidance estimates means that worse case

scenarios are often high (over 1 % of a population could be “struck” annually).

Studies evaluating potential for interaction have also considered seabird behaviour and

environmental factors within highly energetic tidal areas. Some species of seabird are

often recorded in largest numbers during periods of lowest tidal movement whereas other

species have been observed in the largest numbers when the tidal increases in speed.

The tidal cycle has been observed to influence foraging rates at multiple tidal steam

locations, both positively and negatively (Wade, 2015; Waggit et al., 2016; Goldsmith,

2017; Lieber et al., 2019). These results have also been observed during surveys of tidal

stream devices with significant relationships between abundance estimates and tidal

strength observed (Robbins, 2012; Savidge et al., 2014; Long, 2017; Envirosphere, 2018).

Furness et al. (2012) used a vulnerability index similar to ones used in offshore windfarms

(Garthe & Hüppop, 2004) to investigate which species are most vulnerable to collision in

Scottish waters. By assessing the species conservation status via four parameters (status

in relation to the Birds Directive, percentage of the biogeographic population that occurs in

Scotland, adult survival rate, and UK threat status) and seven biological parameters

(drowning risk, mean and maximum diving depth, benthic foraging, use of tidal races for

foraging, feeding range, disturbance by ship traffic, and habitat specialisation) a metric of

impact was determined. In conclusion, the paper identified Black Guillemot, Razorbill,

European Shag, Common Guillemot, Great Cormorant, divers and Atlantic Puffin as the

species most vulnerable to the adverse effects from tidal turbines in Scottish waters. The

method used within this study could be applied to other areas to provide site specific

vulnerability estimates once initial surveys have ascertained which species are present.

5.2.4. Summary of current knowledge

This section provides an overview of the empirical evidence from tidal stream devices and

summarises the results for the monitoring approaches described above for seabirds (see

Table 4 for a summary of all monitoring techniques to date).

Currently there is little actual monitoring data or limited direct evidence relating to collision

risk between birds and turbine devices and as such there is a large gap in current

understanding of actual encounter rates as well as direct and indirect mortality rates in the

event of a collision. From visual surveys to date, there is some evidence that seabird

species (particularly cormorants and divers) do not habitually forage during periods of the

fastest currents, with a clear preference for areas/times of a lower tidal flow. Some smaller

species like auks have been observed to increase at tidal flow increase, at different

Page 32 of 69

locations. This highlights the importance of site-specific baseline surveys to address this. If

the tidal patterns show a decrease in usage with an increase in current speed it would

restrict the number of seabirds that are active when the devices would be operational.

The evidence gathered to date, indicating birds show distribution shifts away from a

device, suggested far-field avoidance. However, there has not been enough information on

near-field evasion to provide an overall conclusion. The monitoring undertaken for these

projects/studies has provided useful insights into animal movements within the vicinity of

tidal turbines and have proven the potential applicability of several technologies. Telemetry

is widely used in ornithological studies, but there do not appear to be any published

studies where data telemetry devices (either GPS or underwater accelerometers) have

been deployed on seabirds close to tidal stream devices. Such studies would provide

important empirical data.

Overall there is no evidence of an observed collision between a seabird and a tidal stream

device. However, this may be due to the limitations associated with all monitoring methods

and the limited amount of monitoring that has been undertaken. As direct collision is very

hard to monitor for small species like seabirds, data are few and far between. The use of

hydroacoustic devices provides the clearest “picture” to date. Hydroacoustic devices have

tracked diving seabirds in the vicinity of a device (Williamson et al., 2015). Neither of the

locations in which this monitoring has been successfully deployed have recorded a

collision event, with very few encounters recorded.

The largest gap within our current understanding is driven by the lack of empirical

avoidance data. Distribution shifts of several seabird species after the deployment of a

device have been observed, but due to the challenges of monitoring direct collision, there

is still unknown potential for collision, especially when considering arrays of devices.

5.3. Fish

Several studies have been undertaken evaluating the impact of tidal turbine devices on

fish behaviour and fish collision risk with rotating turbine blades. This has included in situ

monitoring, in laboratory settings and through collision risk modelling. A synthesis of

current understanding of fish collision risk with tidal stream devices is provided below and

within the evidence spreadsheet.

Within the UK, migratory fish have been highlighted as the main concern in regards to fish

interactions with tidal stream devices. However, various fish species also contribute to the

diet of diving seabirds and marine mammals.

The review has therefore considered all fish species (for which evidence exists) including

commercially important species and those that are protected through environmental

designations. Physical injuries to fish caused by mechanical strike, shear and cavitation

are the principle risks identified.

5.3.1. Monitoring approaches

Monitoring to understand the potential collision risk of fish and elasmobranchs with tidal

stream devices is in its infancy and there is limited data available to inform collision risk

assessments. Direct sampling of fish (typically undertaken by nets and trawls) is

Page 33 of 69

impractical in the energetic conditions in which most tidal stream developments are placed

and is therefore not an acceptable form of monitoring (Fraser et al., 2018). Currently

monitoring has been undertaken through two main approaches:

• The first approach uses underwater video monitoring to assess fish distributions

and monitor the device to ascertain potential rate of collision or the consequences

of collision; and

• The second approach uses a hydroacoustic device to determine the spatial-

temporal overlap between fish and the tidal stream device, and therefore the

likelihood of encounters.

These methods can be used to enable an estimation of fish density. In addition, fish

behaviour such as shoaling, avoidance and collision with turbines can be recorded.

Underwater video monitoring can also allow species identification.

Underwater video techniques involve the placement of a camera system on the turbine

structure. Video cameras can be used to record the presence of particular fish species in

the vicinity of the turbine, information on tidally-induced behaviour and, to some degree,

the passage of species through the turbine device to capture any collision or injury. One

key limitation of video systems is that due to light requirements they are only able to

capture data during daylight hours. Additionally, the field of view may not cover the entire

rotor and the camera view can be obscured at times by the turbine blades meaning that

not all encounters may be captured. Cameras may also become obscured by biofouling.

Footage collected can provide clear evidence of collision or injury caused to fish, however,

data analysis is extremely time consuming. The use of baited video cameras may provide

additional information on species present but also risks biasing recording towards

predatory species.

Echosounders, split-beam acoustic transducers (SBT), high definition sonar or other

similar hydroacoustic devices can also be used to monitor fish presence in the vicinity of

turbine devices. The deployment of such acoustic devices has varied between studies.

Echosounders can be mounted on a vessel to conduct transects across the development

area. Specific over-the-turbine transects are necessary to generate a representative strike

risk model but transects can also be conducted across the wider area to assess fish

populations and diurnal movement patterns. Data collection through this mechanism is

limited by vessel operational periods and weather down-time which might prevent data

collection.

Hydroacoustic devices can also be placed directly onto the turbine device and left in situ

for a set period of time or can be seabed-mounted and deployed within the wake of the

tidal stream turbine. This fixed-location method includes the use of Dual Frequency

IDentification SONar (DIDSON) acoustic cameras (Sound Metrics Corp., Seattle, WA),

which can provide acoustic imagery to monitor movements of fish within the vicinity of tidal

turbines. As devices can be continuously recording they can collect data across a 24-hour

window and are not light sensitive so are able to monitor during the dark and in highly

turbid environments. However, data analysis is resource intensive.

The placement, observational window and frequency of recording vary between studies

but will result in a trade-off between data resolution and field of data capture. Only

Page 34 of 69

echosounder instruments with a sufficient range provide the practical means to investigate

the behaviour of fish throughout the entire water column of a typical tidal channel. The

functionality of echosounders in energetic environments has so far been limited due to the

operational difficulties of data collection in such conditions and the intense interference

caused by backscatter related to turbulence (Fraser et al., 2018).

Two additional techniques which have been trialled to assess fish collision risk are injury

assessment (direct sampling of fish injuries) and fish tagging (tracking studies of fish

movements), however, their accuracy and reliability to assess risk are inconclusive.

Direct sampling can be undertaken to conduct an injury assessment. During site visits, fish

can be collected and any injured fish identified. The type of injury, such as bruising,

laceration or descaling, can be identified and the likely cause assessed. In addition,

discussions with fishermen can also provide data to inform injury assessment if they have

noted any injured fish during catches. However, to date many assessments have been

unable to conclude the sources of injury or determine if all injuries have come from a single

source.

Attaching tracking tags to fish, is another method which has been used to analyse fish

movements in the vicinity of tidal turbines. The use of tags can, for example, allow the

tracking of seasonal and diurnal movements of fish species. It can also provide

information on swimming velocity and direction and has been highlighted as particularly

useful in areas with migratory fish species. The use of this method is relatively new and in

many cases data is not yet publicly available. One of the main challenges faced in

detecting acoustically tagged fish is poor receiver efficiency due to excessive noise

interference when current speeds exceed 2 m s-1 (Redden et al., 2014). To date, such

methods have had limited success in informing collision risk.

Use of electro-mechanical (strain/ accelerometers) devices can be used to measure force

on a turbine blade from object collision and assess where on a turbine the collision has

occurred. However, due to the nature of tidal stream environments accelerometers are

always under strain and there is uncertainty over the sensitivity of the devices and the

force of impact required to register a collision event. Currently these devices have not

been utilised for fish collision risk monitoring and there is no clear evidence as to their

effectiveness, however there is potential applicability for their use in monitoring basking

shark collisions.

Despite the availability of the above monitoring techniques, it should be noted that many

developments have not undertaken project specific monitoring and instead rely on desk-

based reviews and historic data to inform the potential distribution of fish species in the

vicinity of a site. Although the literature can identify potential species present in the vicinity

of a tidal device and provide background on fish behaviour, the specific impacts through

collision risk are difficult to predict. This is a significant limitation with respect to the current

understanding of fish collision risk assessments.

5.3.2. Monitoring studies and results

Since the first in situ placing of a tidal turbine several developers have implemented

monitoring to record fish species in the vicinity of tidal devices. Multiple approaches, as

discussed above, have been used for monitoring potential collision risk, with varying

Page 35 of 69

degrees of success. This section reviews some of the monitoring undertaken across a

range of developments, summarises the findings of the monitoring with regards to collision

risk and discusses any limitations of the findings.

Ocean Renewable Power Company (ORPC), as part of their RivGen project, similarly used

underwater imagery to assess fish distributions in the vicinity of their turbines in the Igiugig

River, Alaska. Underwater imagery was collected 24 hours per day from 19 to 25 July and

19 to 27 August 2015 (Priest and Nemeth, 2015). Due to the large time and resource

implications of analysing the video footage initial analysis focused on only the first 10

minutes of the footage each hour.

The initial results identified 1,020 fish from six species observed across the monitoring

period. Several instances of fish moving through the turbine were recorded but there was

no direct evidence of physical injury or collision. However, the footage did record some

evidence of disorientation by juvenile salmon moving downstream (Priest and Nemeth,

2015). Following later analysis of the remaining footage results revealed a total 2,538 fish

within the vicinity of the turbine. Across the period a total of 20 collisions were recorded

(0.8 %), the majority of which involved shoals of juvenile fish (Matzner et al., 2017). The

differences in results shows the limitation in only part assessing video footage and indicate

the time constraints required to undertake accurate analysis. Additionally, the method of

post processing and analysis of footage can be key in determining collision risk.

As part of the TidGen Project, ORPC undertook a down-looking hydroacoustic survey to

assess the impact of their TidGen horizontal axis turbines. The survey was undertaken

between August 2012 and September 2013, monitoring fish presence, abundance and

vertical distribution. Key species recorded in the vicinity of the turbines at all tidal

conditions were Atlantic herring, Atlantic mackerel, winter flounder, silver hake, haddock

and white hake. The results of the study showed a significant decline in fish density closer

to the turbine, starting from approximately 140 m from the device. Fish were more likely to

be recorded at the same depth as the turbine during the night compared to day time, the

tidal stage did not appear to have an impact. On the basis of this level of avoidance from

the turbine it was concluded that the probability of a fish encountering the turbine’s blade

would be less than 2.9%, based on the density of fish in the study area (Shen et al., 2015;

FORCE, 2018). However, it should be noted that the use of hydroacoustics limited the

ability to isolate individual fish species within mixed shoals and limited the area/ range in

which behaviour could be observed.

The Cape Sharp Tidal project in Minas Passage, Canada, also utilised a downward facing

hydroacoustic echosounder mounted onto a vessel as part of the baseline fish monitoring.

The OpenHydro open centred turbine was deployed between 2009 and 2010 and 2016

and 2018. As part of defining the environmental baseline a fish-monitoring programme was

implemented. Three 24-hour surveys in May, August and October 2016 were undertaken

to assess fish abundance and behaviour. Following the start of operations four additional

24 hour surveys were undertaken in November 2016, January, March and May 2017.

Preliminary findings suggested no significant effect of the turbine on the density of fish in

the mid-field i.e. less than 1 km from the turbine or on fish vertical distributions or at

different tidal states. However, monitoring did record highly variable fish densities

seasonally with highest densities observed in November and January. Specific over-the-

turbine transects were necessary to generate a representative strike risk model.

Page 36 of 69

A further project example in which fish interaction with tidal stream devices have been

examined through the use of hydroacoustic devices is the Ocean Renewables turbine

testing platform in Cobscook Bay, Maine. In this example, two DIDSON acoustic cameras

were deployed for a period of 22 hours in September 2010 (Viehman and Zydlewski,

2015). This time span included approximately 11 hours of daylight and 11 hours of

darkness, and nearly two tidal cycles. The two DIDSON units were mounted upstream and

downstream of the device and were operated in high-frequency mode (1.8 MHz), which

provides better resolution at short ranges, however this limits the viewing window to 10 m.

Behaviours of individual fish and schools were classified (e.g. entering, avoiding, passing,

or remaining in the wake of the turbine) and the effects of turbine motion (rotating or not

rotating), diel condition (day or night), and fish size (small, ≤10 cm; large, >10 cm) were

analysed.

Turbine motion significantly affected the probability of fish entering, avoiding, and passing

by the turbine. The turbine began rotating (and generating power) when current speeds

exceeded 1 m s−1. 11,377 fish were detected while the turbine was not rotating (24% of

the time), and 17,611 were detected while it was rotating (76% of the time). When the

turbine was rotating, the probability of fish entering the turbine decreased by over 35%

from when it was not. The probability of fish entering the turbine was also greater at night.

Overall, no direct collisions were detected, but 19% of fish were recorded entering the

turbine while the turbine was operational and therefore at risk of collision. The largest

shortcoming of the DIDSON technology in this study was the resolution, although DIDSON

image resolution is among the best available, the results could not provide information on

direct blade strike of fish or the condition of fish exiting the turbine for the same reason.

Additionally, the rotation of the turbine caused a slight blurring around the blade edges, so

everything within approximately 5 cm of the blades was not discernible (Viehman and

Zydlewski, 2015).

Verdant at Roosevelt Island Tidal Energy (RITE) project used acoustic sampling via mobile

split-beam acoustic transducers (SBT) to monitor fish. SBTs were attached and mounted

downward on a vessel. Four baseline surveys were conducted between September 2005

and November 2005, six further surveys were undertaken during operation between

October and December 2008. Additionally, 24 stationary SBTs were deployed to monitor

passing fish. These were monitored once a month for the first six months of deployment,

January to June 2007. From the stationary SBTs 38 schools and 82 individual fish were

observed within the 112 minutes of video footage collected when turbines were rotating

and operational. Thorough assessment of the footage revealed five occurrences (4%) of

what appeared to be direct encounters with the rotor blade. However, a key limitation of

the analysis was the limited field of view of the SBT which was blocked by turbine blades,

as such some collision incidences may have been missed (Bevelhimer et al., 2016).

FORCE in the upper Bay of Fundy is a tidal energy test facility. A multi-year fish tracking

study (2010-2013) has been undertaken at this site to address questions related to the

potential risks of turbine operation to migratory species. VEMCO animal tracking

technology was used to detect near year-round animal movements (path, velocity and

depth) and behaviour of 386 tagged Atlantic salmon, Atlantic sturgeon, American eel and

striped bass.

Hydo-acoustic receivers were placed in lines at 300 to 400 m intervals across both the

Minas Passage (5 km wide) and the FORCE test site (1 km wide) to detect the presence of

Page 37 of 69

transmitters surgically implanted in fish as they moved within the Minas Passage during

migrations into and out of the Minas Basin. Results show that the FORCE test area

represents an important migratory corridor for all fish species examined and provided

evidence of frequent use of the passage rather than use for just in- and out-migration.

Compared to the other species monitored, striped bass were within the detection range of

acoustic receivers for surprisingly long periods of time (up to 10 months) and were

considered at significant risk of interaction with tidal devices.

Despite these findings, the results of the tagging study could not indicate direct interactions

with tidal devices and therefore provides no assessment of collision risk. Additionally, the

study found poor receiver detection efficiency during periods of high current velocity

(greater than 2 m s-1). The tag transmission dataset was therefore predicted to represent

40% or less of the actual transmissions within the general range of the receivers which

was a key limitation to the assessment (Redden et al., 2014).

As an additional part of this project specific fish injury monitoring was also undertaken.

Incidence of injured fish were monitored through visits and discussions with fishers of the

southern Minas Basin. The survey was conducted from May 2017 until approximately two

weeks after the tidal turbine was removed. However, the cause of the injuries could not be

determined, nor could it be concluded that all injuries were from a single source and

therefore the assessment could not provide specific results on fish collision rates (FORCE,

2018).

5.3.3. Wider evidence and assessment of collision

Potential impacts of tidal turbines on fish species have been predicted through collision

risk modelling. Various types of models have been used to predict the risk of fish colliding

with tidal turbine devices as there is currently no single recommended model type for this

purpose. As such studies have adopted different approaches to collision risk modelling.

The different collision risk models have some similarities but differ in scope (coverage of

the collision risk pathway) and in consideration of animal behaviour, such as natural

seasonal or diurnal movement patterns, movement in different tidal flows and avoidance

capacity.

Two of the key models which are used to directly assess collision risk include kinematic

models which are mathematical models that describe the motion of objects without

consideration of forces, this includes fish movement and tidal turbine operation; and Agent-

Based Models (ABM) which simulate animal-structure interactions to predict collision risk.

A Kinetic Hydropower System (KHPS)-Fish Interaction Model was developed and applied

by Verdant Power to support assessment of its Gen4 KHPS device at the RITE project.

Using the results from acoustic fish monitoring surveys undertaken between 2007 and

2009 collision risk modelling was undertaken using Echoview analysis. Data on the

location of 34,708 fish was used including the location, heading and velocity of each fish

that passed through the multibeam field, alongside data on turbine operation and velocity,

current velocity and tidal state (ebb/flow). The model predicted the probability of a blade

strike on fish passing the turbine to be below 0.50 % for all arrays up to 30 turbines

(Bevelhimer et al., 2016). However, the rotational speed of the turbine blades was also

considered as a known constant at 35 ft s-1, the model did therefore not include

parameters to assess the varying rotational speed at different points along the blade and

Page 38 of 69

therefore did not account for where on the blade the strike would occur. The blade was

considered to be rotating above 1 m s-1 and all collision with a rotating blade were

considered to be lethal.

To further assess the impacts of its TidGen device, Ocean Renewable Power conducted

an encounter probability model to assess the likely number of fish collisions. The model

was based on previous monitoring work at both the RivGen and TidGen test sites. The

model took account of fish abundance, vertical distribution and avoidance behaviour (Shen

et al., 2015). From the baseline monitoring it was assessed that the probability of fish

being at the depth of the blade and therefore at risk of collision was 0.793 %. However,

when accounting for avoidance behaviour and for time when the turbine was actually

operational (minimum tidal current of > 1 m s-1) the probability of a collision fell to 0.083%.

One caveat to this finding is that no interpretation has been made as to the impact on

wider fish population dynamics and no specification is made as to what probability means,

i.e. per day, per year, or based of fish density.

The probability of collision risk calculated in the above model differs from that determined

as part of a separate operational monitoring study undertaken at the TidGen site. Based

on the outputs of the monitoring study the probability of a fish encountering the turbine’s

blade was calculated to be less than 2.9 %, a 34-fold difference in collision probability

(Viehman and Zydlewski, 2015; FORCE, 2018). This indicates the potential limitations of

modelling (also noting all the potential inaccuracies associated with monitoring) and shows

that care needs to be taken when interpreting results. However, the probability of collision

from this model is comparable to the results from modelling by Bevelhimer et al. (2016),

which predicted a 0.5 % risk for fish passing through the swept area of a 30-turbine array.

No further assessment of the exposure time of fish to turbines was undertaken to analyse

wider population impacts.

An ABM was developed to predict the likely collision risk of migrating silver eels passing a

tidal turbine in Strangford Lough. The ABM aimed to simulate interactions between fish

and consider the natural population cycle and behaviours. The dimensions of the device,

the number of blades, current speed and the size of the fish were all characterised within

the model. An additional parameter was also included where combined collision speeds

greater than 5 m s-1 were assumed to be fatal.

Results predicted low rates of collisions, with just 1.1 % of eels passing through Strangford

Lough predicted to collide with the turbines. The model also predicted that more collisions

would occur for fish swimming upstream (against the flow and increased with longer body

lengths). However, risk decreased the faster the fish swam. As with other modelling

studies this project did not include an assessment of avoidance behaviour and therefore

further work could include modifying swimming behaviours to include active turbine

avoidance (Rossington and Benson, 2019).

The current outputs of collision risk models are primarily derived from density data

indicating utilisation of fish at a set location. However, a recognised limitation is that these

models do not generally account for avoidance behaviour of fish and thus it is difficult to

assess the level of exposure to collision risk pressure. Modelling approaches are also

limited by the degree of species, site and tidal device specific data for any given location.

In addition, there is relatively little validation data of actual collisions to verify predictions

that are made. Care therefore needs to be taken when interpreting the results from

Page 39 of 69

modelling studies as specific input parameters may mean results are not comparable to

real world situations, as shown by Shen et al., (2015). However, modelling can provide an

indication of likely collision risk when monitoring data is not available.

Alongside gathering of empirical data from locations in which tidal stream devices have

been deployed, theoretical studies have also been undertaken to help understand collision

risk. These studies look at fish activity to determine the likelihood of encounter based on

behaviour alone. Interaction with turbine devices is only considered conceptually and not

through direct interaction with tidal stream devices.

A study by Hammar et al. (2015), for example, assessed collision risk based upon video

data of fish movements in strong tidal currents. Only fish species known to fully or partly

utilise the mid and upper parts of the water column were included in the study as they were

considered most important in the context of collision risk. Fish movements (directions,

depth, speed), fish length, avoidance behaviour and tidal current were included in the

analysis.

The study generated three important findings regarding the probability of co-occurrence

between fish and tidal turbines. Firstly, results showed that as current speed increased fish

movement decreased as fish seek shelter. Fish were very rare in currents as strong as 1 m

s−1, therefore many species will have a very low probability of co-occurring with operating

tidal turbines. Secondly, the findings indicate that in strong tidal currents fish are most

likely to swim in the direction of the current, increasing the probability of entering turbines

in the direction of the flow. The authors also noted that large fish have a higher probability

of collision compared to small fish, including high probabilities of blade incident and

damage.

Hammar et al. (2015) further assessed the implications of turbine design on likely collision

impact and noted that turbine design has a large influence on potential mortality rates. The

theoretical assessment found that small turbines were easier to avoid than large ones, and

slow rotational speeds reduce the probability of turbine injury. Turbines with rotors moving

fast and of a large diameter were more likely to cause severe injury. Among the many

developing turbine designs, the Minesto Deep Green design is distinguished by the fact

that it moves very fast (>10 m s-1) has a very large diameter and can operate in relatively

slow currents (1 m s−1) where fish activity is higher than in stronger currents. The study

suggested that the installation of such turbines in areas frequented by large fish of

vulnerable populations should therefore be carefully assessed with regards to ecological

risks before installation (Hammar et al., 2015).

Another factor which is not generally well understood is the survivability of fish following a

collision. Combining strike probability with strike mortality provides a measure of turbine

passage survival. As such a multi-year study was initiated by EPRI to evaluate the

importance of turbine design, including leading-edge blade thickness, shape, and impact

velocity, on fish survival (EPRI, 2011).

The study used modelling and laboratory testing to develop a blade design criterion.

Initially the modelling indicated that a semi-circular shaped blade created the highest

differential forces (leading edge pressures) and therefore had the greatest potential to

deflect a fish prior to impact. Laboratory testing was undertaken using rainbow trout, white

sturgeon and American eel, of various lengths to test the model findings. Turbines were

Page 40 of 69

installed in a large, recirculating flume and fish were exposed to blades of differing

thicknesses (9.5, 25.4, 50.8, 101.6, and 152.4 mm) traveling at speeds up to 30 ft s-1.

The ratio of fish length to blade thickness (L/t) was used to standardize the results. Strike

survival rates greater than 90% were observed when the L/t ratio was 1 or less (i.e., fish

length was equivalent to or greater than the leading-edge blade thickness).

A similar study by Amaral et al. (2015) considered the survivability of fish following a

collision by monitoring delayed mortality to fish injured during a collision. Underwater video

cameras were used to record fish movements in the flume, and to test survivability of

species for up to 48 hours following collision.

The results found that survivability was variable between species but ranged from 1 to 0.96

one-hour post collision. However, for some species survivability dropped 48 hour post

testing, ranging from 1 to 0.91. The predominant form of injury observed was bruising

(seen on 23% of all fish). This assessment therefore highlights the potential indirect/

delayed mortality related to collisions something which is not captured in current turbine

monitoring studies (Amaral et al., 2015).

In both studies, across the species monitored the observed survival rates were generally

greater than 95%. However, it should be noted that survivability decreases with increasing

blade diameter (as blades are moving across a larger volume of water) and with increasing

strike speed (Hammar et al., 2015). Survivability also varied between species. It should

also be noted that due to the scale of the projects, full size turbines could not be used. In

the study by Amaral et al., (2015) only a turbine with 1.5 m blade diameter could be used,

which may have resulted in fewer injuries to fish.

5.3.4. Summary of current knowledge

Monitoring techniques utilised to monitor fish populations predominantly include either

underwater video assessment or use of hydroacoustic devices (including Echosounders,

SBTs, DIDSON). However, there are multiple limitations to the monitoring methods

currently undertaken (see Table 4 for a summary of all monitoring techniques to date).

For hydroacoustic devices these limitations include a narrow sampling/ viewing window

limiting the area of analysis, a limited resolution which makes estimating fish species and

fish size particularly difficult and being unable to provide information of direct blade strike

on fish or the condition of fish exiting the turbine. In contrast, video equipment allows for

the identification of individual fish species and for the effects of turbines on swimming

behaviour to be analysed. However, video cameras are not able to sample at night without

artificial lighting, and often footage is obstructed by turbine blades which reduces the

accuracy of the data collected.

Overall, Viehman and Zydlewski, (2015) concluded that if blade strike is the focus of a

study, video may be a more useful tool, but that DIDSON is a useful tool for monitoring fish

interaction with tidal turbine devices and is especially well suited to sampling at night or in

turbid conditions.

Page 41 of 69

Currently there is little actual monitoring data or limited direct evidence relating to collision

risk between fish and turbine devices and as such there is a large gap in current

understanding of actual encounter rates as well as direct and indirect mortality rates in the

event of a collision. Fish species composition and abundance vary spatially between

different tidal stream project sites, and temporally over seasonal or diurnal cycles, which

means site specific studies over an appropriate timescale are, necessary to assess

potential device impact. The potential interactions between fish and tidal turbines have

been identified as a research gap for tidal stream power generation in the UK as a whole,

and Wales in particular (Roche et al., 2016).

Limitations to current collision risk assessment include:

1. Uncertainly around all input parameters and models which do not include fish

avoidance behaviour;

2. Limited analysis of individual species - behaviours of species, e.g. demersal vs

pelagic or fish size, will change likely impacts and risk of collision;

3. Results are species, location and device specific and may not be appropriate for

the assessment of other devices;

4. Delayed mortality from injury caused during collision has not been assessed.

Survivability predicted from current studies are likely an underestimate; and

5. No interpretation as to what impact the probability of collision risk has on wider

fish population dynamics.

Additionally, there has been little verification of model outputs and the associated

predictions given the lack of available monitoring data and the limitations around this.

These limitations lead to uncertainty concerning the reliability of results and limit the

potential extrapolation or use of results to inform other tidal developments. Therefore, care

should be taken when interpreting results from previous studies.

Page 42 of 69

Table 4: Summary table of the collision risk monitoring techniques used to date.

Monitoring

Technique

Category of

Monitoring

Receptor

Monitored

Main Pros Main Cons In situ example(s)

Visual

observations -

vantage point

Spatial-

temporal

overlap

Marine

mammals and

seabirds

• Relatively cheap

compared to the

other observations

allowing a longer-

term evidence base

to be collected;

• Provides information

on behaviour and

occupancy patterns

(surface only)

• Does not provide

information on sub-

surface collision risk;

• Restricted to daylight

monitoring;

• Long-term datasets

may be needed to

undertake robust

analysis;

• Can be hard to

provide accurate

spatial information.

• All sites/projects to

date have

undertaken visual

observations.

Visual

observations -

boat and aerial

surveys

Spatial-

temporal

overlap

Marine

mammals and

seabirds

• Able to cover a large

area quickly (aerial);

• Only methods able

to provide density

estimates for

modelling (true

density provided

over the whole site).

• Does not provide

information on

collision;

• Restricted to daylight

monitoring;

• Can be hard to

identify all animals to

species (some stay

at species group

level).

• All sites/projects to

date have

undertaken visual

observations.

Page 43 of 69

Monitoring

Technique

Category of

Monitoring

Receptor

Monitored

Main Pros Main Cons In situ example(s)

Device

mounted video

camera(s)

Direct collision

and spatial-

temporal

overlap

Marine

mammals,

seabirds and

fish

• Allows direct visual

observation of any

collisions;

• Provides data on

near-field presence

and behaviours

around turbines

• Able to have multiple

cameras on tidal

stream device;

• Technology is

cheap.

• Quantity of data

generated,

associated

processing and

storage issues;

• Specific

environmental

conditions required

to allow visual

observations to be

made (i.e. during

daytime, low turbid

conditions);

• Not usually viewed

live, not able to stop

any collision.

• Nova Innovation’s

devices in Bluemull

Sound;

• Sustainable Marine

Energy’s PLAT-I

devices.

Passive

Acoustic

Monitoring

(PAM)

Spatial-

temporal

overlap

Marine

mammals

• Provide 24/7

monitoring;

• Relatively easy to

deploy and retrieve

(if not attached to

device).

• Only monitors

cetaceans;

• May not provide

directional

information;

• Does not provide

information on actual

collision.

• Minesto (at the

Holyhead Deep Site

and Strangford

Lough)

• FORCE

Page 44 of 69

Monitoring

Technique

Category of

Monitoring

Receptor

Monitored

Main Pros Main Cons In situ example(s)

Active SONAR

Direct collision

and spatial-

temporal

overlap

Marine

mammals,

seabirds and

fish

• Provide real time

feedback which

could stop a collision

from occurring;

• Tracks 3D

movement of

animals.

• Large initial cost;

• Not able to

determine species

for fish, seabirds or

some mammal

species;

• Produces vast

amounts of data

which can be

challenging to

process/store.

• DeltaStream device

in Ramsey Sound;

• SeaGen in

Strangford Lock.

Blade mounted

pressure

gauges/

accelerometers

Direct collision

Marine

mammals,

seabirds and

fish

• Can provide

operational

performance data at

same time as

environmental.

• Not able to ascertain

what has hit (maybe

debris);

• Inbuilt to turbines so

hard to repair if

broken;

• Often hypersensitive

and recording water

flows.

• Several devices at

EMEC.

Page 45 of 69

Monitoring

Technique

Category of

Monitoring

Receptor

Monitored

Main Pros Main Cons In situ example(s)

Animal-

attached

technology

(GPS/accelero

meters/

magnetometry)

Direct collision

and spatial-

temporal

overlap.

Marine

mammals and

seabirds (and

fish)

• Provides a large

amount of data

within a small tag,

which can last a long

time.

• May not be able to

retrieve device (may

not be needed by

some type of

device);

• Provides an

individual level of

detail, may not be

applicable to the

population and

therefore need to tag

a lot to understand

patterns fully.

• May be limited

information on near-

field behaviour

• Seal population of

Strangford Loch.

• Seals within Ramsey

Sound.

• Seabirds off north

coast of Anglesey.

Page 46 of 69

6. Gap Analysis

Given the small number of tidal stream deployments to date and therefore the limited

evidence base, a gap analysis of the evidence has been undertaken to determine the

information that is required to undertake robust assessments of collision risk and also the

best approach for future consenting.

The results of the gap analysis are presented in Table 5. In summary, the evidence that

can be collected from baseline surveys to determine the spatial and temporal distribution

and density estimates of marine mammals and seabirds is considered to be largely

adequate, although it should be noted that not all tidal stream developments have

undertaken baseline monitoring. Gathering this same spatial and temporal information for

fish can be more challenging as most monitoring methods are based only on passage data

and are dependent on the swimming behaviour or life history of particular fish species.

The key evidence gaps for all receptor groups during operational monitoring are in relation

to determining avoidance or encounter rates of different marine species, as well as

confirming if an actual collision has occurred and what the effects of a collision are. For

example, there is evidence that some fish may avoid high tidal flows and thus not be

exposed to collision risk and evidence that marine mammals may also avoid tidal turbines

to some extent. Evidence on near-field evasion is also very limited thus creating

challenges in estimating avoidance rates. For example, there is no evidence of what

happens when fish approach tidal devices as a result of the pressure differential

associated with turbine blades. Operational monitoring of fish is additionally challenging

given the limitations of the particular methods that are available and that approaches are

not species specific.

In addition to these gaps, the limited monitoring data that is currently available is species,

location and device specific and may therefore not be transferable or applicable to the

assessment of other tidal stream projects. In particular, species composition and

abundance can vary spatially between different tidal stream project sites, and temporally

over seasonal or diurnal cycles, which means site specific studies over an appropriate

timescale are necessary to be able to assess the potential impact of a device.

Another key gap is the potential implication of collision mortality at the population level.

Whilst it might be possible to estimate the collision risk for an individual, understanding

what the consequence might be for the population is challenging. Methods do exist to

assess population level effects of tidal stream devices, and these have been applied to

marine mammals and seabirds but there is no evidence that these have been applied for

fish.

The cumulative effects of deploying multiple tidal devices and arrays in the marine

environment is a further key uncertainty. This is particularly the case for marine species

that travel large distances and that have the potential to overlap with more than one project

site.

Page 47 of 69

Table 5: Gap analysis of evidence available for undertaking robust collision risk assessments.

Receptor Baseline Evidence Operational Monitoring Modelling

Marine Mammals

• Survey methods available for

collecting general baseline

(density) data but limited

information available on finer

scale behaviour within tidal

stream areas (non-device and

device areas);

• Density estimates between

acoustic and visual surveys can

be significantly different, a

multi-method approach to

baseline collection would be

beneficial.

• Uncertainty around avoidance

rates and actual strikes;

• No evidence of the

effectiveness of accelerometers

and these are generally not

considered sufficiently sensitive

to accurately register collision

events (although the larger the

animal the more effective this

method is likely to be);

• Real-time assessment of

collision requires improved

algorithms for identifying marine

mammals approaching turbine

blades.

• Currently based on hypothetical

avoidance rates (no avoidance

behaviour); avoidance rates

need to be well defined in order

for models to provide accurate

collision estimates

• Assumes that all collisions are

fatal; better information required

on the consequences of

collision.

Page 48 of 69

Receptor Baseline Evidence Operational Monitoring Modelling

Seabirds

• Survey methods available for

collecting general baseline

(density) data but fine-scale

distribution is hard to gather

over large spatial scales using

traditional methods (vantage

point).

• Unknown near-field avoidance

rates and subsequent

consequences if an actual strike

was to occur;

• Currently near-field observation

methods cannot identify seabird

species, therefore unknown

impact on the population;

• Telemetry is widely used in

ornithology, but there has been

no published work on bird

movement within tidal stream

environments using this

technology.

• Currently based on hypothetical

avoidance rates (no avoidance

behaviour); avoidance rates

need to be well defined in order

for models to provide accurate

collision estimates

• Currently assume that all

collisions are fatal, better

information required on

consequences of collision.

Page 49 of 69

Receptor Baseline Evidence Operational Monitoring Modelling

Fish

• Baseline data collection is

lacking, with few methods used

to understand which species

are present. Unknown best

monitoring approach;

• The period that fish spend

within the vicinity of the tidal

stream device, is not well

understood, seasonally,

temporally, or tidally driven

variation;

• The behaviour of basking

sharks in high energy

environments; is there a similar

attraction observed within

seabirds and marine mammals?

• The extent to which devices,

moorings and inter-array areas

may act as fish aggregation

devices;

• Better understanding needed

on the use of tidal stream areas

by fish, including:

- Migratory species pathways

and behaviour;

- Fish swimming behaviour –

Swimming depth preference

and avoidance capability.

• Probability of collision not

related to wider population

impacts;

• Real-time assessment of

collision not currently

undertaken;

• Additional information on

specific species impact.

Fish are currently classed

assessed as a

homogenous entity, but

likely to be differences in

exposure between

demersal and pelagic

species;

• No assessment of impacts

to fish from pressure

differential across the

blade, which may cause

injury and/or mortality;

• There is little information

currently on the sublethal

effects of collision.

• Currently based on

hypothetical avoidance

rates (no avoidance

behaviour); avoidance rates

need to be well defined in

order for models to provide

accurate collision estimates

• No agreed approach for

Collision Risk Modelling for

fish species;

• Modelling parameters do

not currently account for

pressure differential across

the blade;

• Rotational speeds at

different points along the

blade not assessed within

models and all collisions

are considered fatal.

Page 50 of 69

7. Recommendations for Addressing Key Gaps

Recommendations on addressing the key gaps in data and information have been

identified and are discussed below.

Further evidence on realistic animal densities and near field evasion is likely to be needed

to generate robust avoidance rates. This is particularly important for marine mammals and

fish which may swim through tidal arrays on each tide and thus regularly be exposed to

collision risk.

While acoustic techniques for tracking fine scale behaviour of marine animals close to

turbines have matured recently, they still require further development. Other relevant

technologies, such as blade mounted pressure sensors for instance, also need to be

explored and developed further in order to confirm if they are effective in determining a

collision event. More information on the behaviour of animals in the presence of a turbine

and on the physical consequences of a collision (with the blade or pressure differential) is

also required to fully understand the potential for death or injury.

Development of and improvement to fish monitoring techniques is a key recommendation

and research priority to improve the knowledge of fish behaviour within tidal stream areas.

Further research is also needed to accurately determine fish behaviour around tidal turbine

devices, as well as to detect and record collision events to quantify the occurrence and

frequency of collisions.

There may be some relevant evidence or lessons that can be learned from other similar

types of development that have the potential to result in a collision risk, notwithstanding

that the design of these developments are different to tidal energy devices and therefore

the actual collision risk would not be the same. For example, there is growing experience

of applying collision risk modelling to seabirds in relation to offshore wind farms. There is

also a good understanding of the impact of hydropower turbines on fish. Established

projects with longer term monitoring programmes would provide some further insight into

the interaction of different marine animals with moving structures in the marine

environment and the likelihood of evasion. They would also provide evidence on the

potential impact of pressure differentials which is currently lacking from the evidence

available from tidal stream development.

Population Viability Analysis (PVA) methods are available to determine population level

effects and these are often applied to marine mammals and seabirds and can be applied

to fish too but this is not often done. Assessing the population level effects on all fish

receptors is considered to be more challenging. Monitoring techniques are not species

specific and stock assessment data against which to compare impacts is quite limited.

Equivalent Adult Values (EAVs) are often used to assess population level fish impacts but

there are limitations to this approach. For example, published EAV’s are highly variable

and monitoring is not species specific. The requirement to improve our confidence in

monitoring and assessing fish impacts is therefore key to a reliable assessment of the

population level effects becoming possible.

Page 51 of 69

Page 52 of 69

8. Conclusions

A primary concern with tidal turbines is the risk of marine animal collision, however, there

is a lack of clear evidence to illustrate how animals interact with the turbines and to what

degree this represents a real risk. This is largely due to the uncertainties relating to the

likelihood and potential effects of collision and the individual and population consequences

of injury/mortality. The main approaches that are used to help determine the possible risk

of collision are modelling tools, monitoring in the field and laboratory studies. This

Evidence Review has undertaken an in-depth review of these approaches, focussing in

particular on information from field monitoring, and the empirical evidence available from a

number of planned and implemented tidal stream projects in the UK, Europe and North

America.

The range of available monitoring techniques, and their strengths and weaknesses, are

summarised in Table 4. Monitoring requirements will vary from site to site, but could

include:

• Animal behaviour around turbine structures (e.g. can they detect turbines, do

they avoid them, can they escape once detected etc.);

• Quantification of number of collisions and near misses (primarily dependent

on accuracy of assumed or modelled impact);

• Outcome of animal collisions (e.g. injury/damage to animal, noting that where

evidence is not available, collision is currently assumed to be fatal on a

precautionary basis); and

• Identification of object/species types (to inform behaviour and impact

studies).

The review found that field monitoring techniques used to determine the spatial and

temporal overlap between a tidal stream device and marine animal, primarily to

characterise the baseline environment (baseline monitoring) but also during the

operational phase (impact monitoring), are valuable in determining the presence,

distribution and likely vulnerability of species to tidal stream devices. They also provide

density estimates that are a key input parameter for collision risk models. However, these

types of surveys need to be carefully designed to ensure that the data collected is of

sufficient spatial resolution (e.g. include depth distribution) and accounts for temporal

variability (e.g. tidal cycle and seasonality).

To date, none of the monitoring studies on marine mammals and seabirds have recorded a

direct collision with a tidal device; however, there are limitations with all of these studies

(e.g. shut down clause, no analysis of all available data and/or no actual monitoring of

direct collision) such that if a collision had occurred it may not have been detected. One of

the monitoring studies undertaken on fish have recorded collisions with tidal turbines,

particularly shoaling juvenile fish. There has been an overall paucity of monitoring data

given that there have only been a small number of tidal devices deployed and monitored

thus far. Despite this, the data that has been collected to date, provides valuable evidence

on the behaviour (e.g. far-field avoidance) and likely overlap of different marine species

around devices.

Page 53 of 69

Fine-scale 3D movement data, through telemetry and hydroacoustic devices, have

provided some initial evidence for near-field evasions. However, these methods are

relatively costly and generate a considerable amount of data which require a large amount

of time and resource to process and analyse. In addition to the challenges of monitoring in

a tidal environment, not all these methods are able to provide conclusive evidence that an

actual strike has occurred, and each method has different limitations that need to be taken

account of in terms of the resolution and quality of the data that can be collected (e.g.

battery capacity of telemetry tags, narrow sampling/viewing window of hydroacoustic

devices, video footage not possible to collect at night or in turbid conditions etc.).

Modelling continues to be the most commonly used approach to assessing the risk of

collision. There are a range of modelling tools available, each with different input

parameter requirements (e.g. the physical characteristics of turbines, physical and

behavioural characteristics of animals and local density estimates). Model assumptions

are often conservative, for example, they may assume there is no avoidance behaviour

and that all collisions are fatal. The three main types of model available to determine the

potential collision rate in marine mammals and seabirds (and which could also be used

modified for fish) are ERMs, CRMs and ETPMs. Existing fish collision risk models include

kinematic models and agent-based models.

From a review of the evidence currently available, there has been limited validation of

collision risk models with the results of monitoring during operation. The level of

confidence in the outputs of these modelling tools is therefore quite limited.

Over time, as additional tidal stream energy devices are deployed and monitored, the

evidence base that can then be used in the consenting of future projects will increase. The

monitoring techniques and ultimately the predictions of collision and its potential effects on

the population will therefore also improve.

Page 54 of 69

9. References

Amaral SV, Bevelhimer MS, Cada GF, Giza DJ, Jacobson PT, McMahon BJ. & Pracheil

BM. 2015. Evaluation of Behavior and Survival of Fish Exposed to an Axial-Flow

Hydrokinetic Turbine. North American Journal of Fisheries Management 35 (1), 97-113.

Band B. 2000. Windfarms and seabirds: calculating a theoretical collision risk assuming no

avoiding action. Scottish Natural Heritage Guidance Note.

Band B, Sparling C, Thompson D, Onoufriou J, San Martin E. & West N. 2016. Refining

estimates of collision risk for harbour seals and tidal turbines. Scottish Marine and

Freshwater Science, 7(17), pp.1786-1.

Benjamins S, Dale AC, Hastie G, Waggitt JJ, Lea MA, Scott B. & Wilson, B. 2015.

Confusion reigns? A review of marine megafauna interactions with tidal-stream

environments. Oceanography and Marine Biology: An Annual Review, 53(53), 1-54.

Bevelhimer M, Colby J, Adonizo MA, Tomichek, C, & Scherelis C. 2016. Informing a Tidal

Turbine Strike Probability Model through Characterization of Fish Behavioural Response

using Multibeam Sonar Output. Oak Ridge National Laboratory, Report No. ORNL/TM-

2016-219. pp.75.

Carlson TJ, Elster JL, Jones ME, Watson BE, Copping AE, Watkins ML, Jepsen RA. &

Metzinger K. 2014. Assessment of strike of adult killer whales by an OpenHydro tidal

turbine blade (No. PNNL-21177). Pacific Northwest National Lab.(PNNL), Richland, WA

(United States).

Cole EL, Waggitt JJ, Hedenstrom A, Piano M, Holton MD, Börger L. & Shepard EL. 2019.

The Ornithodolite as a tool to quantify animal space use and habitat selection: a case

study with seabirds diving in tidal waters. Integrative Zoology, 14(1), pp.4-16.

Collins PM, Green JA, Warwick‐Evans V, Dodd S, Shaw PJ, Arnould JP. & Halsey LG,

2015. Interpreting behaviors from accelerometry: a method combining simplicity and

objectivity. Ecology and Evolution, 5(20), pp.4642-4654.

Copping A, Grear M, Jepsen R, Chartrand C & Gorton A., 2017. Understanding the

potential risk to marine mammals from collision with tidal turbines. International Journal of

Marine Energy, 19, pp. 110-123.

Copping A, Sather N, Hanna L, Whiting J, Zydlewski G, Staines G, Gill A, Hutchison I,

O’Hagan A, Simas T, Bald J, Sparling C, Wood J. & Masden E. 2016. Annex IV 2016 State

of the Science Report: Environmental Effects of Marine Renewable Energy Development

Around the World.

Envirosphere, 2018. Marine Seabirds Monitoring Program. Tidal Energy Demonstration

Site – Minas Passage, Year 2: 2017–2018.

EPRI. 2011. Fish Passage Through Turbines: Application of Conventional Hydropower

Data to Hydrokinetic Technologies. Technical Report No. 1024638. October 2011. pp 56.

Page 55 of 69

FORCE. 2018. Environmental Effects Monitoring programme, Annual Report 2017. Fundy

Ocean Research Center for Energy.

Fraser S, Nikora V, Williamson BJ. & Scott BE, 2017. Automatic active acoustic target

detection in turbulent aquatic environments. Limnology and Oceanography: Methods,

15(2), pp.184-199.

Fraser, S., Williamson, B.J., Nikora, V. and Scott, B.E., 2018. Fish distributions in a tidal

channel indicate the behavioural impact of a marine renewable energy installation. Energy

Reports, 4, pp.65-69.

Furness, R.W., Wade, H.M., Robbins, A.M. and Masden, E.A., 2012. Assessing the

sensitivity of seabird populations to adverse effects from tidal stream turbines and wave

energy devices. ICES Journal of Marine Science, 69(8), pp.1466-1479.

Garthe S, Hüppop O. 2004. Scaling possible adverse effects of marine wind farms on

seabirds: developing and applying a vulnerability index, Journal of Applied Ecology , 2004,

vol. 41 (pg. 724-734).

Goldsmith N., 2017. Quantifying seabird habitat use and occurrence across tidal-stream

environments in Anglesey, UK. MSc Thesis, Bangor University.

Grant M, Trinder M. and Harding N, 2014. A diving seabird collision risk assessment

framework for tidal turbines. Scottish Natural Heritage Commissioned Report No 773.

Hammar L, Eggertsen L, Andersson S, Ehnberg J, Arvidsson R, Gullstrom M, & Molander

S. 2015. A Probabilistic Model for Hydrokinetic Turbine Collision Risks: Exploring Impacts

on Fish. PLoS One. 10(3): e0117756.

Hammond, P.S., Lacey, C., Gilles, A., Viquerat, S., Boerjesson, P., Herr, H., Macleod, K.,

Ridoux, V., Santos, M., Scheidat, M. and Teilmann, J., 2017. Estimates of cetacean

abundance in European Atlantic waters in summer 2016 from the SCANS-III aerial and

shipboard surveys. Wageningen Marine Research.

Hastie G. 2013. Tracking Marine Mammals Around Marine Renewable Energy Devices

Using Active Sonar. Report by SMRU Consulting. pp 99.

Hastie GD, Gillespie DM, Gordon JC, Macaulay JD, McConnell BJ. & Sparling CE. 2014.

Tracking technologies for quantifying marine mammal interactions with tidal turbines:

pitfalls and possibilities. In Marine Renewable Energy Technology and Environmental

Interactions (pp. 127-139). Springer, Dordrecht.

Hastie GD, Wu GM, Moss S, Jepp P, MacAulay J, Lee A, Sparling CE, Evers C. &

Gillespie D. 2019. Automated detection and tracking of marine mammals: A novel sonar

tool for monitoring effects of marine industry. Aquatic Conservation: Marine and

Freshwater Ecosystems, 29, pp.119-130.

JNCC 2005. UK Common Standards Monitoring Guidance for Marine Mammals.

Johnston, D.T., Furness, R.W., Robbins, A.M., Tyler, G., Taggart, M.A. and Masden, E.A.,

2018. Black guillemot ecology in relation to tidal stream energy generation: An evaluation

Page 56 of 69

of current knowledge and information gaps. Marine environmental research, 134, pp.121-

129.

Joy R, Wood JD, Sparling CE, Tollit DJ, Copping AE. & McConnell B.J. 2018. Empirical

measures of harbor seal behavior and avoidance of an operational tidal turbine. Marine

Pollution Bulletin, 136, pp.92-106.

Langton R, Davies IM, & Scott BE. 2011. Seabird conservation and tidal stream and wave

power generation: Information needs for predicting and managing potential impacts.

Marine Policy, 35(5), 623-630.

Lieber L, Nimmo-Smith WAM, Waggitt JJ. & Kregting L. 2019. Localised anthropogenic

wake generates a predictable foraging hotspot for top predators. Communications Biology,

2(1), p.123.

Long C. 2017. Analysis of the possible displacement of seabird and marine mammal

species related to the installation and operation of marine energy conversion systems.

Scottish Natural Heritage Commissioned Report No. 947.

Malinka CE, Gillespie DM, Macaulay JD, Joy R. & Sparling CE. 2018. First in situ passive

acoustic monitoring for marine mammals during operation of a tidal turbine in Ramsey

Sound, Wales. Marine Ecology Progress Series, 590, pp.247-266.

Matzner S, Trostle C, Staines G, Hull R, Avila A. & Harker-Klimes GEL. 2017. Triton:

Igiugig fish video analysis.

McCluskie AE, Langston RHW. & Wilkinson NI. 2012. Seabirds and wave & tidal stream

energy: an ecological review. RSPB Research Report No. 42. Bedfordshire, UK: Royal

Society for the Protection of Seabirds (RSPB).

Minesto, 2016. Depp Green Holyhead Deep Project – Phase 1 (0.5 MW). Environmental

Statement.

Morlais, 2019. Morlais Project Environmental Statement Volume 1. Chapter 11: Marine

Ornithology.

nrp, 2014. Collision Risk to Diving Seabirds. Appendix 10B of Perpetuus Tidal Energy

Centre’s Environmental Statement.

Onoufriou J, Brownlow A, Moss S, Hastie G. & Thompson D. 2019. Empirical

determination of severe trauma in seals from collisions with tidal turbine blades. Journal of

Applied Ecology, 56(7), pp.1712-1724.

Owen E. 2015. Black guillemot (Cepphus grylle) tracking in Orkney, 2013 and 2014.

Scottish Natural Heritage Commissioned Report No. 903.

Priest J. & Nemeth M. 2015. Data Analysis for Monitoring of the RivGen in the Kvichak

River, 2015.

Redden AM, Broome J, Keyser F, Stokesbury M, Bradfors R, Gibson J. & Halfyard E.

2014. Use of Animal Tracking Technology to Assess Potential Risks of Tidal Turbine

Interactions with Fish. Proceedings of the 2nd International Conference on Environmental

Page 57 of 69

Interactions of Marine Renewable Energy Technologies (EIMR2014), 28 April – 02 May

2014, Stornoway, Isle of Lewis, Outer Hebrides, Scotland.

Robbins A. 2012. Analysis of Seabird and Marine Mammal Data for the Fall of Warness

Tidal Test Site, Orkney. Scottish Natural Heritage Commissioned Report No. 614.

Roche RC, Walker-Springett K, Robins PE, Jones J, Veneruso G, Whitton TA, Piano M,

Ward SL, Duce CE, Waggitt JJ, Walker-Springett GR, Neill SP, Lewis MJ, & King JW.

2016. Research priorities for assessing potential impacts of emerging marine renewable

energy technologies: Insights from developments in Wales (UK). Renewable Energy, 99,

1327-1341.

Rossington K, & Benson T. 2019. An agent-based model to predict fish collisions with tidal

stream turbines. Renewable Energy. [Online] Available at:

http://eprints.hrwallingford.co.uk/1740/1/TP-060_INSITE-KRO-A0-2019-R1.pdf

Royal Haskoning 2011. SeaGen Environmental Monitoring Programme – Final Report.

Savidge G, Ainsworth D, Bearhop S, Christen N, Elsaesser B, Fortune F, Inger R,

Kennedy R, McRobert A, Plummer KE, Pritchard DW, Sparling CE. & Whittaker TJT. 2014.

Strangford Lough and the SeaGen tidal turbine. In Marine renewable energy technology

and environmental interactions (pp. 153-172). Springer, Dordrecht.

Schmitt P, Culloch R, Lieber L, Molander S, Hammar L. & Kregting L, 2017. A tool for

simulating collision probabilities of animals with marine renewable energy devices. PloS

One, 12(11), p.e0188780.

Scottish Natural Heritage 2016 Assessing collision risk between underwater turbines and

marine wildlife. Scottish Natural Heritage Guidance Note.

Shen H, Zydlewski G, Viehman H. & Staines G. 2015. Estimating the Probability of Fish

Encountering a Marine Hydrokinetic Device. Renewable Energy. 97, 746-756.

Shoji A, Elliott KH, Greenwood JG, McClean L, Leonard K, Perrins CM, Fayet A. & Guilford

T. 2015. Diving behaviour of benthic feeding Black Guillemots. Seabird Study, 62(2),

pp.217-222.

Sparling C, Lonergan M. & McConnell B., 2018. Harbour seals (Phoca vitulina) around an

operational tidal turbine in Strangford Narrows: No barrier effect but small changes in

transit behaviour. Aquatic Conservation: Marine and Freshwater Ecosystems, 28(1),

pp.194-204.

Thompson D, Onoufriou J, Brownlow A. & Morris C. 2016. Data based estimates of

collision risk: an example based on harbour seal tracking data around a proposed tidal

turbine array in the Pentland Firth. Scottish Natural Heritage Commissioned Report No.

900.

Tollit D, Joy R, Wood J, Redden AM, Booth C, Boucher T, Porskamp P. & Oldreive M.,

2019. Baseline presence of and effects of tidal turbine installation and operations on

harbour porpoise in Minas Passage, Bay of Fundy, Canada. Journal of Ocean Technology,

14.

Page 58 of 69

Verdant Power, 2010. Pilot License Application Roosevelt Island Tidal Energy Project

FERC NO. 12611. Exhibit E – Environmental Report.

Viehman, H.A. and Zydlewski, G.B., 2015. Fish interactions with a commercial-scale tidal

energy device in the natural environment. Estuaries and Coasts, 38(1), pp.241-252.

Wade HM. 2015. Investigating the potential effects of marine renewable energy

developments on seabirds. PhD Thesis, University of Aberdeen

Waggitt, J.J., Bell, P.S. and Scott, B.E., 2014. An evaluation of the use of shore-based

surveys for estimating spatial overlap between deep-diving seabirds and tidal stream

turbines. International Journal of Marine Energy, 8, pp.36-49.

Waggitt JJ, Cazenave PW, Torres R, Williamson BJ. & Scott B.E. 2016. Predictable

hydrodynamic conditions explain temporal variations in the density of benthic foraging

seabirds in a tidal stream environment. ICES Journal of Marine Science, 73(10), pp.2677-

2686.

Welsh Government, 2019. Prosperity for All: A Low Carbon Wales [online]. Available at:

https://gov.wales/sites/default/files/publications/2019-06/low-carbon-delivery-plan_1.pdf

(accessed on 09/12/19)

Williamson BJ, Blondel P, Armstrong E, Bell PS, Hall C, Waggitt JJ. & Scott BE. 2015. A

self-contained subsea platform for acoustic monitoring of the environment around Marine

Renewable Energy Devices–Field deployments at wave and tidal energy sites in Orkney,

Scotland. IEEE Journal of Oceanic Engineering, 41(1), pp.67 81.

Williamson BJ, Fraser S, Blondel P, Bell PS, Waggitt JJ. & Scott BE. 2017. Multisensor

Acoustic Tracking of fish and seabird behavior around tidal turbine structures in Scotland.

IEEE Journal of Oceanic Engineering, 42(4), pp.948-965.

Williamson B, Fraser S, Nikora V, Blondel P, Couto A, Chapman J & Scott, B 2019.

Interaction of Marine Renewable Energy and Marine Organisms: Active Acoustic

Assessment. International Conference on Fisheries Engineering, Nagasaki, Japan,

21/09/19 - 24/09/19.

Zimmer WM. 2011. Passive acoustic monitoring of cetaceans. Cambridge University

Press.

Page 59 of 69

10. Appendices

Appendix A - Agreed List of Tidal Devices

Table 6: Initial long list of tidal stream energy devices/developers agreed with NRW and the status of the monitoring and/or reporting.

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Wales

Minesto

Deep Green

Tidal Kite DG500

One 0.5 MW tidal

kite, no moving

parts.

Holyhead

Deep

Off Holyhead,

Anglesey, Wales

Testing in

Strangford

Lough 2017

and then

deployment in

Holyhead in

2018.

EIA undertaken.

Currently undertaking

operational monitoring.

Monitoring a condition of

Marine Licence (ORML

1618).

Yes

Wales

Tidal Energy

Ltd.

DeltaStream

400 kW device of

3-bladed turbines

N/A

Ramsey Sound,

Pembrokeshire,

Wales

Operational

March –

December

2016.

Operational monitoring

occurred and scientific

papers of the results

published. Company no

longer exists so no

contact was made.

Yes

Page 60 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Wales

Morlais

Multiple

Consent

application is

being prepared

for the

demonstration

zone

West

Anglesey

Demonstrat

ion Zone

Off Holyhead,

Anglesey Wales

At the pre-

consent

phase.

EIA undertaken with

baseline data collected,

but no device has been

put into operation and

therefore no operational

monitoring.

Yes

Wales

SIMEC

Atlantis

Energy

Up to nine

Atlantis

Resource

SeaGen devices

Skerries

Tidal

Stream

Array

Between Carmel

Head and the

Skerries, off

Anglesey, Wales.

Consented in

2015. Project

put on hold in

2016.

EIA undertaken, baseline

data collected but no

operational monitoring,

as project has stalled.

Scotland

Multiple

Multiple:

Testing site in

which multiple

devices have

been placed. Will

review the

monitoring at

EMEC as well as

individual

projects if

specific

monitoring has

occurred.

EMEC

Falls of Warness,

Orkney,

Scotland.

The DA

opened in

2006, with the

first device

placed in the

water in

September

2007.

EMEC undertakes

wildlife observations

throughout the year

alongside various

projects. Multiple

researchers have

published papers in the

area. Annual reports up

to 2014 online.

Yes

Scotland

Nova

Innovation

Nova M100

turbine

Twin 4.5 m

blades 100 kw

Shetland

Tidal Array

Bluemull Sound,

Shetland,

Scotland

Two devices

operational

since 2016

with an

additional

device

deployed in

2017.

Non-statutory

environmental

assessment undertaken,

including collision risk

modelling. Monitoring

started in 2010 and is

continuing. Undertook

phone interview for

additional information.

Yes

Page 61 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Scotland

SIMEC

Atlantis

Energy

Atlantis

Resources

AR1500 and

Andritz Hydro

Hammerfest

(AHH) AH1000

MK1.

Four No.1.5 MW

three 18 m blade

turbines, one

Atlantis and

three AHH

MeyGen

Between

Scotland’s

northernmost

coast and

Stroma.

Currently in

Phase 1 with

first turbine

installed in

October 2016,

up to 4 in situ

(by April

2018).

Additional

phases are

planned with

more turbines

consented.

EIA undertaken.

Currently undertaking

operational monitoring.

No information was

found about this project.

Atlantis did not partake in

the evidence review.

Yes

Scotland

Nautricity

CoRMaT tidal

stream turbine

Contra-rotating

turbine with two

blades moving in

opposite

directions.

Argyll Tidal

Demonstrat

or Project

Mull of Kintyre,

Scotland

Approved in

2013 but never

installed.

Device

installed at

EMEC in 2017

(see EMEC

above).

Environmental Appraisal

for Mull of Kintyre site,

using baseline data. No

operational phase

monitoring undertaken at

this location.

See EMEC, for

monitoring at that

location.

Scotland

DP Energy

Multiple

West Islay

Tidal

Project

West of Islay,

Scotland.

Consented in

2017 by

Crown Estate

and Marine

Scotland but

no device has

been placed in

the water.

EIA undertaken, baseline

data collected but no

operational monitoring as

project has halted.

Page 62 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Ireland

SIMEC

Atlantis

Energy

SeaGen

1.2 MW device

with 2

No.600 kW

powertrains

Strangford

Narrows

Strangford Lough

Narrows,

Northern Ireland.

Operational

from 2008-

2016.

Decommission

ed in 2019.

Large amounts of

monitoring data were

collected and reported as

part of this project.

Yes

Ireland

SmartBay

Multiple

Ireland’s national

marine and

energy test and

demonstration

site.

SmartBay

Galway Bay,

west of Ireland

Operational

since 2006.

Operational monitoring is

being undertaken as part

of the licence conditions.

Yes

Ireland

Sustainable

Energy

Authority of

Ireland

(SEAI)

Multiple

Demonstration

area.

Atlantic

Marine

Energy

Test Site

Annagh Head,

west of

Belmullet, Ireland

Fully

consented in

2015 but yet to

have devices

put in place.

EIA undertaken with

baseline data collected,

but no device has been

put into operation and

therefore no monitoring.

Will follow up with

company as they put out

a tender for monitoring

services – may still be

collecting data.

Yes

Canada

Atlantis

Operations

Canada Ltd.

(a joint

venture of

Atlantis

Resources

Ltd. and Rio

Fundo Ltd. (a

DP Energy

affiliate))

Atlantis

Resources

AR1500

Three

No.1.5 MW three

18 m blade

turbines

N/A

Minas Passage,

Bay of Fundy,

Canada (Fundy

Ocean Research

Center for

Energy

(FORCE))

Unclear when

the device was

put into the

water.

FORCE undertakes

monitoring reporting

annually and reports on

their environmental

effects.

Scientific publications

have been undertaken at

this location.

Yes

Page 63 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Canada

Cape Sharp

Tidal

(OpenHydro

and Emera)

OpenHydro

Open centred

2 MW turbine

(16 m diameter)

N/A

FORCE

First device

2009 – 2010.

Second device

2016 – 2018

until

OpenHydro

went into

liquidation.

FORCE undertakes

monitoring reporting

annually and reports on

their environmental

effects.

Scientific publications

have been undertaken at

this location.

Yes

Canada

DP Energy

Six Andrtiz

Hammerfest

Hydro (AHH)

MK1 (up to 9

MW)

Uisce Tapa

FORCE

Post-consent.

Will use Berth

E and C at

FORCE. Not

yet in water.

FORCE undertakes

monitoring reporting

annually and reports on

their environmental

effects.

Scientific publications

have been undertaken at

this location.

Canada

Sustainable

Marine

Energy Ltd

and

SCHOTTEL

Hydro

PLAT-1

Floating platform

with four

SCHOTTEL

SIT250 70 kW

turbines

N/A

Grand Passage

(between Long

Island and Brier

Island, in Digby

County, Nova

Scotia)

Operational

Sep 2018 and

June 2019 for

phase 1

testing.

Operational monitoring

occurred via video

cameras.

Yes

Page 64 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

Canada

Clean

Current

Clean Current

turbine

65 kW Horizontal

axis bi-directional

ducted turbine.

N/A

Race Rocks, Off

Vancouver

Island, Canada

Operational

between 2005-

2011

Environmental

Monitoring Report

produced after 1 year of

operation. No information

was found about this

project. Archipelago

(consultancy who

undertook the

monitoring) did not

partake in the evidence

review.

Yes

USA

Ocean

Renewable

Power

Company

RivGen Tidal

Turbines

Horizontal Axis

Turbines

Igiugig

River

Energy

Project

Igiugig River,

Alaska, USA.

Operational

2014-2015.

Monitoring reports on

fish available using

EyeSea, 43 hours

detected 20 fish

interactions with no

injury.

Yes

USA

Ocean

Renewable

Power

Company

TidGen Tidal

Turbines

Horizontal Axis

Turbines (750

kW)

Cobscook

Bay Tidal

Energy

Project

Cobscook Bay,

part of the bigger

Bay of Fundy, off

the Maine coast,

USA.

Operational

2012-2017.

Monitoring reports 2012-

2016 are available

online.

Yes

Page 65 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

USA

Verdant

Power

Gen4 Free Flow

System

Grid-connected

demonstration

array of six

Kinetic

Hydropower

System (KHPS)

3-bladed

turbines. Gen5 is

being licensed at

the moment – for

deployment in

2020.

Roosevelt

Island Tidal

Energy

(RITE)

Project

Demonstrat

ion

East Channel of

East River - New

York, NY, USA.

Operational

2006-2009

(9,000 hours

of operation).

Operational monitoring

occurred, specifically for

fish.

Yes

Australia

Atlantis

Resources

100 kW

Aquanator™

device, a 150 kW

AN-150™

(Nereus™ I)

device, and a

400 kW AN-

400™ (Nereus™

II) device where

used over the

projects lifespan.

San Remo

Test Site

Newhaven

Wharf, near San

Remo, Victoria,

Australia

Operational

between 2006

and 2015

Tethys mentions “zero

environmental impact”

after two years of

independent testing. No

information was found

about this project.

Atlantis did not partake in

the evidence review.

Yes

Australia

Tanax

Energy

N/A

Clarence

Strait Tidal

Energy

Project

Clarence Strait,

Northern

Territory,

Australia

Still in the pre-

consent

planning

stage.

Impact assessment

undertaken with baseline

data collection but no

operational monitoring.

Page 66 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

England

Sustainable

Marine

Energy Ltd

PLAT-0

Submerged

platform with two

turbines.

N/A

Off Yarmouth,

IOW. Then

EMEC.

Blank cell

Operational monitoring

occurred at IOW and

also continued at EMEC.

See EMEC, for

monitoring at that

location.

Yes

England

Perpetuss

and Isle of

Wight

Council

Multiple

Demonstration

area.

Perpetuus

Tidal

Energy

Centre

(PTEC)

Off St.

Catherine’s

Point, IOW,

England

Fully

consented in

2016 by MMO

however put

on hold in

2017 due to

financial

concerns.

EIA undertaken with

baseline data collected,

but no device has been

put into operation and

therefore no operational

monitoring.

Netherlands

Multiple

Dutch Marine

Energy Centre

(DMEC)

DMEC

Marsdiep

between Den

Helder and the

Wadden island of

Texel, Holland

Operates one

test site, but

only mentions

one user.

Little information online

with no clear monitoring

plan. Only device

mentioned also placed in

EMEC (see EMEC

above).

France

Sabella

D10-1000

Multiple 10 m

blades (1 MW)

N/A

Between Brest

and Ushant

Island, France

Operational

2015-2016

and then again

in 2018-

present

Email communication

confirmed video cameras

where placed on device.

Yes

France

SEENEOH

Multiple – small

scale devices

can be tested in

a riverine

environment

N/A

Gironde Estuary,

Bordeaux

Operational

since 2016

No monitoring to date but

plan for fish mortality

studies. Undertook

telephone interview to

find out about project.

Yes

Page 67 of 69

Country Developer Device Project Location Status

Monitoring and

Reporting

Included

within this

review?

South Korea

South

Korean

Government

Cross-flow

Helical Turbine

(1 MW)

Uldolmok

Tidal Power

Station

Uldolmok Strait

in the Yellow

Sea, at Jindo

Island, South

Jeolla, South

Korea

Operational

since 2009

and currently

in use

No environmental

assessment or

monitoring has taken

place.

Multiple

countries

OpenHydro

(a Naval

Energies

company)

OpenHyrdo

Open centred

N/A

Off Brittany,

France;

Seattle,

Washington,

USA;

EMEC; and

FORCE

Operational

between 2007

until 2018 (not

continuous)

when

company went

into liquidation.

OpenHydro undertook

several EIAs due to the

different locations that

the device has been

placed. Operational

monitoring also occurred

at several of the sites.

Included fish, marine

mammals and seabirds.

See EMEC and FORCE,

for monitoring at that

location.

Company no longer

exists so no contact was

made.

Yes

Page 68 of 69

Appendix B - Full list of Contacted organisations

Table 7: Organisations contacted (same order as Appendix A) that have/had devices in situ or pre-

consent stage.

Table 8: Organisations contacted (same order as Appendix A) that are involved in tidal energy.

Organisation Contacted

Response

Minesto

No response received.

SIMEC Atlantis Energy

No response received.

Morlais

Response received and interview undertaken.

EMEC

Response received and interview undertaken.

Nova Innovation

Response received and interview undertaken.

DP Energy

No response received.

SmartBay

No response received.

Atlantic Marine Energy Test Site

No response received.

Sustainable Marine Energy

Response received and interview undertaken.

Archipelago (undertook monitoring

of Clean Current at Race Rocks)

No response received.

Ocean Renewable Power Company

No response received.

Verdant Power

No response received.

Sabella

Response received.

SEENEOH

Response received and interview undertaken.

Organisation Contacted

Response

AZTI

No response received.

Bangor University

Response received.

Edinburgh University

Response received.

FORCE

No response received.

Juno Energy

Response received and interview undertaken.

Marine Energy Wales

Response received.

Marine Power Solutions

No response received.

ORJIP

Response received.

Plymouth University

No response received.

ScotMER

No response received.

SEACAMS

Response received and interviews undertaken.

SMRU (and SMRU Consulting)

No response received.

Page 69 of 69

Data Archive Appendix

The data archive contains:

[A] The final report in Microsoft Word and Adobe PDF formats.

Metadata for this project is publicly accessible through Natural Resources Wales’ Library

Catalogue https://libcat.naturalresources.wales (English Version) and

https://catllyfr.cyfoethnaturiol.cymru (Welsh Version) by searching ‘Dataset Titles’.

Resources Wales.

Further copies of this report are available from library@cyfoethnaturiolcymru.gov.uk