ASSESSING THE IMPACTS OF SHRIMP FISHING ON … · Sabellaria reefs has long been accepted (Marine...

ASSESSING THE IMPACTS OF SHRIMP FISHING ON

SABELLARIA SPINULOSA REEF AND ASSOCIATED

BIODIVERSITY IN THE WASH AND NORTH NORFOLK

SAC, INNER DOWSING RACE BANK NORTH RIDGE

SAC AND SURROUNDING AREAS

PILOT STUDY

Kim Last1, Vicki Hendrick

1, Ian Sotheran

2, Bob Foster-Smith

2,

Dan Foster-Smith2, and Zoë Hutchison

1

Final Report for Natural England

May 2012

1 SAMS Research Services Ltd.

Scottish Association of Marine Science

Scottish Marine Institute

Oban, Argyll, PA37 1QA

2 Envision

Stephenson House

Horsley Business Centre

Horsley, Newcastle upon Tyne, NE15 0NY

Impacts of shrimp fishing on S. spinulosa reef

1

TABLE OF CONTENTS

Background ................................................................................................................................... 4

Aims .............................................................................................. Error! Bookmark not defined.

Introduction ................................................................................................................................... 4

Strategy ......................................................................................................................................... 8

Sampling Design ........................................................................................................................... 9

Equipment ................................................................................................................................... 10

Multibeam (Interferometric Bathymetry System) ..................................................................... 10

Towed Video Sled ................................................................................................................... 11

Video Systems for Remote Observation ................................................................................... 13

Laser line Profiler .................................................................................................................... 13

Lander ..................................................................................................................................... 14

Scanning Sonar ........................................................................................................................ 15

Positioning .................................................................................................................................. 17

Trial Surveys ............................................................................................................................... 17

Equipment Trial – Dunstaffnage Bay 4th October 2011 ............................................................. 17

Survey Trial – Loch Etive 21st October 2011 ............................................................................ 18

Experimental Trawl ................................................................................................................. 23

Pilot Study – Southern North Sea 15th – 17

th January 2012 ....................................................... 23

Critique of Techniques and Equipment ........................................................................................ 34

Moorings ................................................................................................................................. 34

Multibeam ............................................................................................................................... 34

Towed Video Sled ................................................................................................................... 34

Video System .......................................................................................................................... 35

Laser Profiler ........................................................................................................................... 35

Lander ..................................................................................................................................... 36

Scanning Sonar ........................................................................................................................ 36

Diver Survey ........................................................................................................................... 38

Experimental Trawl ................................................................................................................. 38

Decision – Support Tool .............................................................................................................. 39

Recommendations for a Full Scale Survey ................................................................................... 41

Literature Cited ........................................................................................................................... 42

Acknowledgements ..................................................................................................................... 44

Appendix 1 : Technical Specifications of the GeoSwath System. ............................................. 45

Shallow Water Multibeam and Side Scan System ..................................................................... 45

Appendix 2 : Abbreviations ..................................................................................................... 47


2

LIST OF FIGURES

Figure 1 : A) The twin beam trawler the “Lynn Princess”. B) Starboard net showing the

beam, bobbins and shoes. C) The “heal” on the shoe. D) The cod end of the net

being hauled aboard showing protective net “chaffers” (Photographs © Kim Last). ......... 6

Figure 2 : Hypothetical example of sampling design. ...................................................................... 8

Figure 3 : The surface data acquisition PC and Sonar head of the GeoSwath Plus system.

(Photographs © Kongsberg GeoAcoustics Ltd). ............................................................ 11

Figure 4 : The towed video sled with umbilicals on which several cameras and a vertical laser scanner were mounted (Photograph © Ian Sotheran). .................................................... 12

Figure 5 : An anterior view of the video sled deployed in Loch Etive (Image © Hugh Brown,

NFSD). ......................................................................................................................... 12

Figure 6 : The laser line generators used, left hand image shows the Savante offshore system mounted on the sled with right hand image showing the simplified battery powered

system which was mounted in a the same manner (Photographs © Ian Sotheran)........... 14

Figure 7 : Light-weight lander on which a scanning sonar unit and a high definition video camera were mounted (Photograph © Ian Sotheran). ..................................................... 14

Figure 8 : Second lander configuration trialled in the southern North Sea (Image © Zoë

Hutchison). ................................................................................................................... 15

Figure 9 : Image of an ensonified object obtained from a digital scanning sonar system

(Image ©Tritech International Ltd). .............................................................................. 16

Figure 10 : Scanning sonar deployed on the lander system successful in sheltered conditions

(Image © Ian Sotheran)................................................................................................. 17

Figure 11 : Hand towed system being trailed in Dunstaffnage Bay in shallow waters. (Image ©

Ian Sotheran). ............................................................................................................... 18

Figure 12 : Location of trial survey site in Loch Etive (56.46108602°N 5.34433104°W). .............. 19

Figure 13 : Mooring for survey in Loch Etive (Image © Kim Last). ................................................ 19

Figure 14 : Plots of the sled tows conducted in the Loch Etive survey (Image © Envision).............. 21

Figure 15 : The screenshot on the left (inverted) shows the 10cm Sabellaria „hummock‟ whilst the screenshot on the right shows the 5cm „hummock‟. The corresponding laser

profile of each hummock is plotted below the relevant frame grab image. ..................... 22

Figure 16 : The scanning sonar image of the artificial S. spinulosa „hummock‟ (Image ©

Envision). ..................................................................................................................... 23

Figure 17 : Location of areas of search for suitable study site (Map © Envision). ............................ 24

Figure 18 : Survey site location (Map © Envision). ......................................................................... 25

Figure 19 : Bathymetric grid and contours around the survey area (Map © Envision). ..................... 25

Figure 20 : Cross sections showing bathymetric profiles of the survey site, top image is a

north-south profile and the lower image a west to east profile (Figures © Envision). ..... 26

Figure 21 : Typical catch after a 10 minute haul showing crustaceans, echinoderms, fish and

macroalgae. Note abundance of Flustra sp. also seen as most dominant faunal constituent from seabed video tows (Photograph © Kim Last). ..................................... 27

Figure 22 : Comparable paths of the trawler on the initial and reciprocal trawl with direction of

tow indicated (Image © Envision). ................................................................................ 28


3

Figure 23 : The video sled survey tracks and path of the trawler across the transect with

direction of travel indicated (Map © Envision). ............................................................. 29

Figure 24 : Organism assemblage composition at study site with five pre-trawl replicate video

sled tows....................................................................................................................... 30

Figure 25 : Comparative assemblage composition of pre- and post- shrimp trawl video tows. ......... 30

Figure 26 : Scanning sonar image from the pilot site survey (Image © Envision). ........................... 31

Figure 27 : Scanning sonar image from the pilot site survey (Image © Envision). ........................... 32

Figure 28 : Example of image, digitised laser line (Photograph © Envision). .................................. 33

Figure 29 : Scanning sonar image of an object, specifically the walls of a test tank, from a controlled test in Dunstaffnage Marine Laboratory (Image © Envision). ....................... 37

Figure 30 : Example of scanning sonar image of Sabellaria reef with scanning head positioned

1.5m above the sea floor. The accompanying image is a frame grab from a bullet camera attached to the lander (Images © Envision). ...................................................... 38

Figure 31 : Assessing Sabellaria reef and management options. ...................................................... 39

Figure 32 : A hypothetical example of a Sabellaria reef scored for the physical characteristics

of reef extent, patchiness, elevation and consolidation (screen-grab). ............................ 40

Figure 33 : Hypothetical pre- and post trawl assessment of a Sabellaria reef following shrimp

trawling (screen-grab). .................................................................................................. 41

LIST OF TABLES

Table 1 : Examples of sample sizes needed to detect a reduction in height (one-tailed test) between trawled versus untrawled areas for two levels of significance (0.05, 0.1)

and two effect sizes (20% and 40%). ............................................................................. 10

Table 2 : Mean rugosity values for each transect line and from each camera. ............................... 33

Table 3: Variability of the rugosity values for each transect line. ................................................ 33


4

BACKGROUND

The „ross worm‟, Sabellaria spinulosa Leuckart 1849 is a sedentary, epifaunal polychaete that builds

rigid tubes from sand or shell fragments. It is a suspension feeder that is generally found individually

or in small aggregations although it can be gregarious in favourable conditions. Large colonies

consisting of fused sand-tubes may form thin crusts or extensive reefs (Hayward & Ryland, 1990, Schäfer, 1972).

The reefs, commonly known as „Ross‟, can be many metres across and raised above the sea bed by up

to 40 cm (Foster-Smith & White, 2001, Vorberg, 2000), providing a biogenic habitat that allows many other associated species, including epibenthos and crevice fauna, to become established

(UK Biodiversity Group, 1999). The fauna is distinct from other biotopes and the structure allows

species to become established in predominantly sedimentary areas where they would not otherwise be

found (Foster-Smith et al., 1997a). In this guise therefore, S. spinulosa reef has been identified as a priority habitat under the UK Biodiversity Action Plan (BAP). „Reefs‟ are also listed under Annex I of

the EC Habitats Directive (Council Directive EEC/92/43 on the Conservation of Natural Habitats and

of Wild Fauna and Flora) as a marine habitat to be protected by the designation of Special Areas of Conservation (SACs).

S. spinulosa reef is a designated feature of both the Wash and North Norfolk (WNN) SAC and the

Inner Dowsing Race Bank North Ridge (IDRBNR) SAC, and there is concern that a predicted increase in the intensity of the pink shrimp fishery in these areas may have an adverse effect on the

reef structures. This pilot study was therefore undertaken in order to assess the best techniques and

survey methodology with which to gather robust evidence on the effects and/or impacts of this

activity on S. spinulosa reef; evidence on which Natural England can base their statutory advice to the managing authorities.

INTRODUCTION

Association between pink shrimp and Sabellaria spinulosa

An association between the commercially valuable pink shrimp, Pandalus montagui Leach, 1814 and

Sabellaria reefs has long been accepted (Marine Ecological Surveys Ltd. et al., 2011, Mistakidis,

1957, Warren, 1973, Warren & Sheldon, 1967). Support for this arises from the belief that S. spinulosa, when available, is a major item in the diet of P. montagui, and laboratory observations

which show that the shrimps probe the Sabellaria dwelling tubes to extract fragments of the worms

with their claws (Mistakidis, 1957). However, in his study of the pink shrimp fishery in the Wash,

Warren (1973) deduced that the abundance of Sabellaria was insufficient of itself to feed the whole of the shrimp population in the main area of the fishery. He consequently concluded that various

alternative food sources must also be taken and that the preference of pink shrimp for Sabellaria may

simply be due to the fact that in a large Sabellaria colony suitable food is available in high concentrations. This observation is supported by further work in which trawls with large numbers of

shrimps/prawns were not found to be associated with Sabellaria (Foster-Smith, et al., 1997a, Foster-

Smith et al., 1997b), although few samples were taken during the study. A close link between pink

shrimp and Sabellaria may therefore be somewhat speculative, and any association may be as much a function of the often prolific nature of the benthic food supply associated with Sabellaria colonies, as

of the presence of the worms themselves (Rees & Dare, 1993, Warren, 1973). Nevertheless, the

perceived association between Pandalus and Sabellaria reef has led to reports that fishermen pursuing Pandalus have used small trawls to search for lumps of S. spinulosa which they regard as an

indication of good fishing grounds (Warren & Sheldon, 1967).

Such reports give rise to concern because physical disturbance, and fisheries activities in particular, are considered to be one of the greatest threats to S. spinulosa reefs (Holt et al., 1997, Holt et al.,

1998, Rees & Dare, 1993). Trawling for shrimp or finfish, dredging for oysters and mussels, net

fishing and potting are all believed to cause physical damage to erect S. spinulosa reef communities

(Jones, 1999). The impact of mobile fishing gear is thought to break the reefs down into small chunks,


5

thus changing the habitat for the rich infauna and epifauna associated with this biotope. If the

individual worms themselves escape direct injury, they can be left vulnerable to predation if their dwelling tubes are damaged, particularly if they are dislodged from them.

Pink shrimp fishery of the Wash

A discussion with the skipper of the Lynn Princess1 revealed that the fishery in the Wash and general

environs is currently limited. Moreover, he sees an expansion of the fishery as unlikely particularly since the cable laying to the Skegness windfarm development appears, in his view, to have

dramatically reduced the pink shrimp catches. This assessment contrasts with the alternative notion

that the fishery is likely to see a significant increase in intensity, mainly due to technological advances

in the peeling of the shrimp. If the latter view proves correct, the fishery is likely to consist predominantly of twin beam trawlers of up to 8 m, though there may also be a limited number of

vessels in the area utilising small otter trawls.

Fishing success for shrimp appears very much dictated by the behaviour of the animals which are believed to bury themselves deeper into the sediment during spring tides, inclement weather or when

the sea temperatures drop during the winter months. This reduces catch success and is the driving

factor in using heavier gear during these times, presumably to increase disturbance to the seabed to improve catches. There also appears to be more favourable shrimping in deeper offshore water during

the winter months when the Lynn Princess specifically targets the “Silver Pit”.

The beam trawler – Lynn Princess

The Lynn Princess has twin, 6 m beam trawls (Figure 1A, B), which consist of a solid steel beam, with “shoes” at either end between which is strung a line of bobbins. Each shoe has a “heel” (function

unknown) which projects to a depth of ~50 mm and which we believe would act as a small plough

when in contact with the seabed (Figure 1C). Together the beam, shoes and bobbins weigh ~1.5 tons.

The net which has ground chafers trails behind the beam (Figure 1D) and it is the action of the bobbins “running” along the seabed which causes the shrimp to dart vertically off and out of the sand

with tail flips, ending in the catch.

1 As the only known shrimper in the area, the Lynn Princess was commissioned to do trial trawls for the

purposes of this study.


6

A

B

C

D

Figure 1 : A) The twin beam trawler the “Lynn Princess”. B) Starboard net showing the

beam, bobbins and shoes. C) The “heal” on the shoe. D) The cod end of the net

being hauled aboard showing protective net “chaffers” (Photographs © Kim Last).

The skipper of the vessel felt that the impact of the nets dragging over the seabed would be small

compared with the impact of the shoes and bobbins since the nets, even when full, supposedly do not

have much contact with the seabed when towed. Each haul takes approximately two hours and fishing

speed is about 2-3 knots. Consequently, a hypothetical beam haul of the Lynn Princess covers ~11 hectares of seabed, and in a typical 12 hour working day ~132 hectares.

The pink shrimp fishery of the Wash and Sabellaria

The occurrence of Sabellaria in the nets of the Lynn Princess is infrequent but when it does occur,

Sabellaria chunks can be of various sizes. Of particular interest is that its occurrence is highly consistent with site. The locations of Sabellaria reef are probably well known since they are

associated with high quality shrimp (referred to as ghost shrimp), but even at these locations

Sabellaria is not consistently picked up as by-catch. Anecdotal evidence further suggests that Sabellaria is more commonly caught in nets on neaps than on spring tides. Semi-lunar sediment burial

may well expose/inundate sections of reef and hence “protect” the reef to the impacts of shrimp


7

trawling. Considering Sabellaria‟s tolerance to sediment burial, at least in laboratory studies (Last et

al., 2011), this may well be a common characteristic of the habitat in which it exists.

Evidence for impact of shrimp trawling on Sabellaria reef

Shrimp fisheries of both the pink shrimp and brown shrimp, Crangon crangon (Linnaeus, 1758), have

been repeatedly implicated in the decline of S. spinulosa reefs. The loss of large S. spinulosa reefs

between 1924 and the 1980s from the subtidal shallows and channels of the northern Wadden Sea, for instance, is thought to have been a consequence of the long-term effects of shrimp-fishing trawls

(Reise, 1982, Reise & Schubert, 1987, Riesen & Reise, 1982). Local fishermen were reported to have

deliberately ground the reefs with heavy gear because the reefs ripped apart the nets when fishing for

shrimp (Riesen & Reise, 1982), though there does not appear to be any other specific evidence of causation. The fishing effort of the brown shrimp beam-trawl fishery increased considerably in the

1980s, despite a decrease in the number of vessels (Berghahn & Vorberg, 1997), simultaneous with

the changes in benthos of the Wadden Sea. This further reinforced the view that the fisheries were responsible for the demise of S. spinulosa reefs which have effectively been replaced by beds of

mussel, Mytilus edulis Linnaeus, 1758, and sand-dwelling amphipods, Bathyporei spp (Reise &

Schubert, 1987). It is possible, however, that the change may be attributed, at least in part, to an increase in coastal eutrophication favouring Mytilus, or to changes in currents affecting larval supply.

Pink shrimp fisheries have been similarly implicated in the loss of S. spinulosa reefs in the approach

channels to Morecambe Bay (Mistakidis, 1956, Taylor & Parker, 1993). Here, subsequent surveys

suggest recovery of S. spinulosa did not occur in the Bay despite the cessation of pink shrimp fishing (Sankey, 1987), and this seems most likely to be due either to lack of larval supply, or to permanent or

ongoing alterations to the habitat (Holt, et al., 1998). In this regard it is of note that the brown shrimp

continued to be fished commercially in the general area (Sankey, 1987). In the Thames Estuary and the Wash, it was reported that “the accepted practice among commercial fishermen [was] to search

with a small hand dredge for the polychaete worm S. spinulosa and then trawl for [pink] shrimp in

areas where this was found” (Warren & Sheldon, 1967). As a consequence, S. spinulosa in these

areas have also been considered vulnerable to the fishery of the pink shrimp.

Despite the widespread belief that shrimp trawling can and has impacted on Sabellaria reef,

supporting evidence is relatively limited. One of the few specific studies in this regard was

undertaken by Vorberg (2000) who used underwater video techniques to make direct observations of the fishing gear of the Crangon fishery in action on the sea bottom. The images revealed the bobbins

of the trawl regularly jumped off the reef surface and only had bottom contact for 39% of the overall

trawling time (20 min duration). The rollers stirred up the top sediment layer when touching the reef surface producing clouds of fine-grain material, such that damage to the reef construction itself could

not be observed, nor could the trawl shoes be observed directly. The trawl shoes do, however, appear

to have permanent bottom contact (Vorberg, 1997), and were found to have left clear impressions in

the surface of a reef constructed by the sister species Sabellaria alveolata (Linnaeus, 1767) reef following a single-pass experimental beam trawl undertaken without a net (due to local authority

restrictions) (Vorberg, 2000). Nevertheless, in-situ measurements of the periodically exposed reef

indicated that all traces caused by the fishing gear had disappeared four to five days later due to the building activities of the S. alveolata worms. Furthermore, empirical calculations of the load of the

fishing gear and the compressive strength of S. spinulosa reef sections led Vorberg to conclude that

the relatively light trawls used in Crangon fisheries can not cause serious damage to sabellariid reef constructions (Vorberg, 2000).

It should be noted, that the findings of Vorberg (2000) relate exclusively to short-term effects

following once-only disturbance. They also appear to relate to large, intact reefs which the trawls pass

over the top of, rather than the patchy clumps of Sabellaria more commonly encountered in the Wash and general environs which the trawls are more likely to strike from the side. The possibility of

impairment by shrimping in the medium to long-term can not be ruled out in the event of more

intensive fishing, despite the relatively light gear, and more evidence is needed before firm conclusions can be reached on the impact of such activity on S. spinulosa reef.


8

STRATEGY

We believe that assessing and measuring the physical expression of Sabellaria and changes (damage)

that are inflicted by trawling is the most important purpose of the study. Once the physical nature of

the impact (if any) across the trawl path is understood, an appropriate sampling strategy can be

determined subsequently to address the ecological effects of pink shrimp fishing on S. spinulosa reef and associated biota. Thus for this pilot study, efforts have been concentrated on means of observing

and measuring reef structure in a replicable and systematic way as a basis for testing for statistically

significant changes that may occur during trawling.

The overall survey strategy encompassed four broad approaches to assess the impact of trawling on

the reef structure at progressively finer levels of detail. In combination it was expected that the

different approaches would enable measurement of reef parameters such as tube height, reef area,

patchiness, orientation of dwelling tubes etc.

Remote survey of the reef

Firstly, it was proposed that the complete reef would be surveyed via remote sensing (multibeam) in

order to allow the extent and outline (plan view) to be mapped, and for the path of a single-pass trawl

to be planned (see Figure 2 for hypothetical example of trawl path). An appropriate transect line could then be laid perpendicular to the proposed trawl path for more detailed assessment, defined

either end with buoys. The acoustic survey would also be repeated within the trawled area following

the trawl.

Figure 2 : Hypothetical example of sampling design.

Remote sampling: Video sled and laser

It was suggested that finer detail of the reef structure could be gleaned from the use of small, high

quality or high definition video systems. Mounted on a light weight, hand-towed sled, this could then

be towed across the sampling area which it was suggested should be approximately 3-4 times the width of the trawl path. Sampling further away from the proposed trawl track would be avoided as

there may be spatial trends across the reef that would obscure the effect of trawling.

Concurrent with the video camera, it was proposed that a laser line generator also be mounted on the sled. It was hoped that the moving platform would then allow the reef profile to be built up through

sequential images by the laser, in turn allowing measurement of reef height. Repeated tows of the sled

would be undertaken across the trawl path post-trawl for comparison.

Extent of reef

Trawl path

Moorings

Lander drops

Transect

Sled tows

Extent of reef

Trawl path

Moorings

Lander drops

Transect

Sled tows


9

Point sampling: Lander with video and scanning sonar

An alternative point sampling technique was also proposed consisting of a small, light-weight lander

equipped with video and scanning sonar that would enable the heights of features to be assessed from

the sonar records. This methodology was utilised in a MALSF MAL008 project with some success (Foster-Smith et al., 2010), and it was felt that further refinement (e.g., by using a smaller frame with

a lower position of the scanning sonar head) would provide a very useful tool capable of deployment

in a wide range of environmental conditions.

Pre-trawl sampling would sample the requisite number of locations (see Sampling Design) randomly

chosen from an area that would encompass the proposed trawl line. This sampling would then be

repeated post-trawl along the trawl path. No attempt would be made to precisely relocate the samples

within the trawled area and all the measurements would be treated as independent samples. Although this would require increased sampling compared to paired samples, it is felt that precise sampling

could not be guaranteed without greatly increasing survey complexity and costs. Independent

sampling is considered to offer greater flexibility within the recommended survey protocols that would be one of the main outputs of this first stage of the trials.

Diver surveys

If spatial heterogeneity of the seabed is high, then to gain any statistically meaningful data sampling

effort has to also be correspondingly high. To avoid an extensive sampling campaign and associated high cost it was anticipated that a very targeted diver survey trial would be employed if conditions

permitted across the tow path of the beam trawl. The divers would video and survey a line transect

pre- and post-trawl using hand-held cameras and by undertaking point sample measurements along the length of the transect at regular intervals. In this way spatial heterogeneity is of little consequence

since it is possible that exactly the same piece of seabed will be surveyed before and after the trawl.

Although it was considered that remote viewing would provide the main sampling technique, it was

hoped that the use of divers under favourable conditions would serve as the ultimate yardstick for assessing the ability of other techniques to measure reef structure, and that they may allow precise

identification of the path of the trawl on the sea bed and assessment of any cross-sectional variation

across the path of the trawl path ie between the shoes and the cod-end of the net.

SAMPLING DESIGN

A further purpose of the first stage trial was to investigate statistical power of the sampling protocol and what standard is feasible to adopt for the full trial, given the costs and sampling, measurement

error and effect size. Table 1 illustrates the power of a hypothetical sampling program (sample size

per treatment – trawled and untrawled) to detect a change in a measurement (effect size) for two significance levels, 95% and 90%. Although 95% significance is the standard for scientific research

for the avoidance of a Type 1 error (the chance of finding a positive effect when there was none), this

can be relaxed for ecological work which is prone to large variation in measurement, especially when

the avoidance of a Type 2 error is arguably more important (the chance of missing an effect when in fact there was a change between treatments).

Given that there may also be quite large measurement errors as well as stochastic variation in any reef

metric, it is suggested that consideration be given to the adoption of a large effect size (e.g., 40%) is combined with a 90% confidence level.


10

Table 1 : Examples of sample sizes needed to detect a reduction in height (one-tailed test)

between trawled versus untrawled areas for two levels of significance (0.05, 0.1)

and two effect sizes (20% and 40%).

% Change Significance Sample size per

treatment

20 0.05 80 0.1 45

40 0.05 40

0.1 25

The exact plan for the sampling program would depend upon the size of the reef chosen (and available) for the trial survey. The pre-trawl and post trawl sampling would enable comparisons to be

made between an area previously untrawled that is then trawled as well as between an area that has

been trawled versus one that has not been trawled. The former would show if a change could be detected and the latter would tell if there is a difference between a trawled and untrawled area.

EQUIPMENT

More detailed descriptions of the equipment used for the pilot survey are given in the sections below.

Multibeam (Interferometric Bathymetry System)

In order to assess the site in context with the surrounding environs, the pilot survey area was mapped

using a GeoAcoustics GeoSwath Plus Interferometric system (Figure 3), which provides simultaneous true sidescan sonar and bathymetric data which are geographically coincident and correct for tide,

vessel movement and position. The data are also corrected for distortion caused by variations in the

speed of sound through water: sound velocity profiles are taken at intervals through the survey using an Odum SVP transducer lowered to the seafloor. The system gives sidescan sonar images across a

swath 10-12 times water depth as well as swath bathymetry. The swath data can reveal topographic

features (often quite large) which may be undetected by other acoustic methods. The system was

deployed at a tracking speed of ~10 km.hr-1

and survey lines were run 500 m either side of the centre of the survey site. Technical specifications for the system are provided in Appendix 1.


11

Figure 3 : The surface data acquisition PC and Sonar head of the GeoSwath Plus system.

(Photographs © Kongsberg GeoAcoustics Ltd).

Towed Video Sled

A relatively light weight sled made of aluminium and PVC was constructed for deployment during

this project (Figure 4). The system was designed for use in low visibility conditions and with the task of detecting topographic changes in the seabed as a main focus. To this effect, cameras were mounted

close the seabed at an oblique angle, with two cameras being used to increase the field of view visible

from the sled. A lighting system was used to provide vertical illumination close to the seafloor (Figure 5). The design of the system also enabled a laser line device to be attached (see section below).


12

Figure 4 : The towed video sled with umbilical and two cameras in the main frame. The front

gantry supports the lights and vertical laser scanner (Photograph © Ian Sotheran).

Figure 5 : Anterior view of the video sled deployed in Loch Etive (Photograph © Hugh

Brown, NFSD).

Laser

Camera

Cameras

Light

Laser

Camera

Cameras

Light

Laser

Lights

Cameras

Laser

Lights

Cameras


13

The system was tested successfully under controlled conditions in Dunstaffnage Bay prior to deployment in Loch Etive with divers and then subsequently used at the test site off the Norfolk coast.

Video Systems for Remote Observation

Envision have pioneered the use of small drop-down/towed camera systems suitable for deployment

in a range of environmental conditions. For the purposes of this study, small high-resolution Sony cameras enclosed in underwater housings were utilised, linked to the surface via an umbilical which

provides power to the cameras and video signal to the surface. The cameras were mounted on both

the sled and the landers close to the sea floor (~ 15 cm height), and angled forwards to allow

assessment of the height of reef structures. The video signal from one camera was overlain with vessel position using overlay module and both signals were recorded onto MiniDV tape and could be viewed

in real time. Lighting was angled downwards onto the focal objects and was provided by FAMI

underwater LED video lighting, powered by a submersible battery pack. This system was considered preferable to water lens systems where the camera is higher and vertically mounted.

Laser Line Profiler

Laser line systems were utilised for the project with the aim of providing a technique which would

enable quantitative measurements of seafloor rugosity to be obtained. The laser line generator was mounted vertically to the camera system so that if a perfectly flat surface were to be viewed, then a

straight horizontal line would be observed on the video footage. Any changes in the height of the

surface would cause the horizontal line to be deflected upwards such that changes in height would be observed and recorded on the video footage.

Initially a commercial system – a Lumeneye Laser Profile System - was hired from Savante Offshore

Services Ltd. which provided the laser line generator, a video recorder and post processing system

which produced x-y plots from specific frames of video footage. This system was connected to the surface via umbilical, and power and video signals were transmitted via this. This system was tested

in Dunstaffnage Bay and Loch Etive to determine the effectiveness of the technique and how suitable

this would be for assessing the vertical complexity (rugosity) of Sabellaria reefs.

Once the technique had proved worthy of further trials, a more cost effective solution was developed

for the main pilot survey. A simple battery powered underwater laser line generator was constructed

(Figure 6). This was then mounted to the camera sled and used to project a green (532 nm 40-50 mW)

laser line onto the seabed in front of the cameras mounted on the video sled.


14

Figure 6 : The laser line generators used, left hand image shows the Savante offshore system

mounted on the sled with right hand image showing the simplified battery powered

system which was mounted in a the same manner (Photographs © Ian Sotheran).

Lander

The primary consideration of the lander design for this project was of a light-weight construction to

facilitate quick and easy hand-deployment from the survey vessel together with stability in strong currents. To this end a low, tripod construction was initially adopted with a central, adjustable spindle

on which the scanning sonar was mounted. Video cameras were mounted facing inwards on the legs

(Figure 7). The design proved satisfactory in the relatively sheltered conditions of Dunstaffnage Bay and Loch Etive but was not sufficiently stable in the stronger currents encountered at the pilot site.

Figure 7 : Light-weight lander on which a scanning sonar unit and a high definition video

camera were mounted (Photograph © Ian Sotheran).

Camera

Scanning

sonar

Camera

Scanning

sonar


15

Under the higher current speeds in the southern North Sea, an alternative lander design was adopted

with a more compact design and lower centre of gravity (Figure 8). The scanning sonar was again mounted on a central axis with the camera mounted at approximately 15 cm from the seabed.

Figure 8 : Second lander configuration trialled in the southern North Sea (Photograph © Zoë Hutchison).

Positioning of the lander relied on the differential GPS on the vessel (accuracy within 1 m), offset for

deployment point on boat. It is expected that the video would reach the sea floor with minimal divergence if the vessel was fully stopped and times of the strongest tides were avoided.

Scanning Sonar

Scanning sonar is an established technology based on sidescan sonar theory used for ROV navigation

and collision avoidance. However, instead of the sonar being static within the towed „fish‟, the scanning sonar has a rotating transducer operating like a radar. The system emits a series of sonar

pings as the sonar head rotates through 360˚ (or other operator configured angle), and the reflected

return signal is measured by the sonar head allowing an image of the reflective surroundings to be built up as the sonar rotates (for example see Figure 9). The images are directly comparable to

sidescan sonar and at a known scale allowing the altitude of features to be determined using

trigonometry from shadow length and known height of the sonar head.

Camera

Scanning

sonar

Camera

Scanning

sonar


16

Figure 9 : Image of an ensonified object obtained from a digital scanning sonar system

(Image ©Tritech International Ltd).

For the purposes of this survey, the scanning sonar was attached to the lander system in a variety of

configurations (Figure 8 and Figure 10) in order to produce an image of reflective objects with a

certain radius of the sonar head. The different combinations were used to compensate for the effects

of water movement, tidal currents, on the systems. The systems were lowered to the seafloor and left in at a point location whilst the sonar performed a scan of the area. The sonar was set to perform the

fastest scan possible (~1 min per 360 degree sweep) and a short range (5 m in order to minimise the

time the system had to remain in fixed location.


17

Figure 10 : Scanning sonar deployed on the lander system successful in sheltered conditions

(Image © Ian Sotheran).

POSITIONING

The surveys utilised a differentially corrected GPS (dGPS) system for recording position of the vessel which has a published accuracy of 1 metre or better accuracy with dGPS, 3 metres or better accuracy

with WAAS (Wide Area Augmentation System).

TRIAL SURVEYS

Equipment Trial – Dunstaffnage Bay 4th

October 2011

The proposed equipment was successfully trialled in Dunstaffnage Bay on the 4th October 2011.

Specifically, the hand-towed sled mounted with video cameras and the hired laser projector was hand-

towed along the shore (Figure 11), and a light-weight lander mounted with scanning sonar and video camera (Figure 7) was deployed from the local pontoon.

High quality images of the seabed were obtained from the cameras on both sled and lander, with

images from the sled showing the beam of the laser across the seabed. Good quality images were also

obtained from the scanning sonar of objects on the seabed and in a laboratory tank (see Figure 29).

Following the trial, modifications were made to the system to increase its stability and the lighting for

the cameras was altered. The trial also showed that ambient light in shallow waters made a low power

laser difficult to detect and thus a brighter (30-50 mW) laser was subsequently used to compensate.

Camera

Scanning

sonar

Camera

Scanning

sonar


18

Figure 11 : Hand towed system being trailed in Dunstaffnage Bay in shallow waters

(Photograph © Ian Sotheran).

Survey Trial – Loch Etive 21st October 2011

The proposed survey techniques and methodology were trialled successfully in Loch Etive, Argyll following cancellation of the planned survey off the north Norfolk coast due to strong winds. The trial

included deployment of the tripod lander (with scanning sonar and video), sled (with hired laser

projector and video) and divers in a sheltered area of Loch Etive from RV Seol Mara. The trial also included an experimental 2 m beam trawl across a transect line.

Site selection and survey strategy

A drop down video camera was used to assess site suitability for assessing the impacts of trawling on

the seabed during the trial. The south shore of the Loch Etive was chosen for this (56.46108602°N 5.34433104°W) since the habitat was consistent with what would be expected in the environs of

Sabellaria reef off the Norfolk coast: sand and pebbles with algal turf, and it was relatively sheltered

from the strong winds. Two moorings were then laid providing markers to aid positioning of a) the

video tows across the site, parallel to the detailed transect line; b) the experimental beam trawl at right angles to the transect; and c) to allow divers to descend “shot” lines to assess the transect line.


19

Figure 12 : Location of trial survey site in Loch Etive, Argyll, Scotland (56.46108602°N

5.34433104°W).

Transect mooring

A mooring was laid which consisted of two anchors (20 kg chain) connected by a weighted transect

ground line (25 m) with polypropeline riser lines and pellet buoys at either end (Figure 13).

Figure 13 : Mooring for survey in Loch Etive (Image © Kim Last).


20

Deployment of the mooring involved dropping one anchor, steaming at half a knot until all the ground

line had been paid out and then dropping the second anchor so that the ground line was taught over the seabed.

Paired divers were deployed who laid sequentially numbered tiles (20 x 20 cm) along the ground line

at 1 m intervals. After the tiles had been laid the ground line was removed in order to avoid the trawl

potentially snagging the line. The purpose of the tiles was:

To mark the position of the transect line;

To serve as discrete, scaled points for still and video photography;

To allow distances along transect line to be easily determined;

To show possible disturbance by the tow path.

Matt, beige tiles were used as markers since they contrasted sufficiently well with the background

seabed to allow easy relocation, and they didn‟t result in glare from the camera flash or upset the contrast of the photographic image.

Diver survey

A pre-trawl diver survey was carried out along the transect line once the mooring had been positioned.

This survey was repeated post experimental trawl.

For each survey two divers were employed, one for video, the other for still photography. The video

transect was taken at an oblique angle ~1 m above the seabed in an attempt to allow semi-qualitative

observations of seabed height whilst the still photographs were taken directly from above each tile. These still photographs encompassed an area of seabed equating to ~9 tiles (0.36 m

2). It was

anticipated that direct height measurements would also be made in a systematic way of any Sabellaria

encountered in close proximity to each tile laid during the main pilot survey, but none was

encountered in Loch Etive. At all times divers swam up current and avoided any seabed contact to minimise sediment disturbance and damage to the habitat.

Sled tows

Sled tows were conducted pre- and post-trawl and on each occasion an attempt was made to run

directly parallel to the trawl line (Figure 14). As can be seen from the plot there was considerable leeway in all tracks. This is believed to have been primarily due to currents having a considerable

impact on the slow towing vessel.


21

Figure 14 : Plots of the sled tows conducted in the Loch Etive survey (Image © Envision).

Laser profiles

To assess the potential of using the video sled and laser to determine changes in seabed height i.e.

rugosity, laser scans were made of “artificially” created dead S. spinulosa “hummocks” deployed by

divers. The hummocks were approximately 5 cm and 10 cm high and were placed on fairly uniform flat seabed. The sled, with laser mounted vertically on the front gantry, was then manually pushed

over the hummocks by a diver allowing the laser line to scan over them. The result is a line on the

seabed which, when it passes over a raised obstacle, becomes visibly deformed. The position of the line in space can be calibrated against a video screenshot from a camera mounted on the sled which

allows calculation of the distance from the laser emitter and the seabed as show in Figure 15.


22

Figure 15 : The screenshot on the left (inverted) shows the 10 cm Sabellaria „hummock‟ whilst

the screenshot on the right shows the 5 cm „hummock‟. The corresponding laser

profile of each hummock is plotted below the relevant frame grab image (Images ©

Savante). Both axes are scaled in mm.

Scanning sonar

A similar procedure was employed using the scanning sonar lander which was placed in close

proximity to the artificial reef such that the Sabellaria hummock was positioned between two of the

lander legs in order that it could be easily located and verified (Figure 16). The sonar was set to

perform a full 360° sweep and the sonar images recorded.

Figure 16 demonstrates that the „hummock‟ is visible but other targets are also identifiable which give

some indication of the heterogeneity of the seabed with several raised features being detected within a

5 m radius.


23

Figure 16 : The scanning sonar image of the artificial S. spinulosa „hummock‟ (Image ©

Envision).

Experimental Trawl

As part of the Loch Etive trial a 2 m beam trawl was towed across the experimental transect line. The net had been filled prior to the experimental trawl and on recovery it was found that it was filled with

benthic invertebrates, brown fucoids, and a boulder estimated to weight at least 100 kg. The

subsequent diver survey found no evidence that the transect line tiles had been displaced as a consequence of the trawl. This finding was somewhat surprising and suggests that even though there

may have been considerable removal of the seabed (it is not known when the boulder was caught)

there was no immediate evidence of this impact using our assessment technique.

Pilot Study – Southern North Sea 15th

– 17th

January 2012

The pilot survey off the north Norfolk coast was undertaken at short notice in January taking

advantage of a short but advantageous weather window, together with availability of surveyors and

boats. Unfortunately previous cancellations due to weather combined with the necessary timescale for

completion of the pilot survey meant that it was felt that this opportunity needed to be taken. Although this timing meant that it coincided with neap tides the tidal range was still high (~4.0 m) and

hence was less than ideal from the perspective of currents.


24

Identification of study site

In order to minimise survey costs, the identification of study site was facilitated through liaison with

Natural England, the Inshore Fisheries and Conservation Authority (IFCA) and others rather than on

the basis of a specific survey to locate S. spinulosa reef. Two potential sites were identified on the basis of existing knowledge out with designated protected areas, both of which were visited and

checked using a small drop-down video camera. Unfortunately no S. spinulosa was identified at either

site (see green search areas Figure 17), so following consultation with Natural England the decision was taken to trial the survey methodology on cobble habitat near Wells next the Sea.

Figure 17 : Location of areas of search for suitable study site (Map © Envision).

Survey

The selected site (6 miles NE off Wells, see Figure 18) was surveyed using GeoSwath. Five pre- and four post-trawl sled tows were completed, but the lander deployment was of limited success in the

strong tidal currents. Unfortunately the divers were unavailable at short notice for this survey although

the limited weather window and strong currents is likely to have precluded their deployment regardless.


25

Figure 18 : Survey site location NE off Wells-on-Sea (Map © Envision).

Bathymetry

The multibeam data require post processing to enable compensations for tide, vessel movement,

vessel installation and communications between systems to be incorporated, and this was done to

produce a bathymetric model at 1m resolution of the site and its surrounding environs (Figure 19).

Figure 19 : Bathymetric grid and contours around the survey site (Map © Envision).


26

The bathymetric data shows the site and adjacent area to be flat with very few distinct features

present. Although there are small variations (<10 cm) which are visible as „waves/ripples‟ to the south of the site, these are artefacts of the processing system which show up when vertically

exaggerated but consist of 20 cm variations over 100 m. The site chosen for survey typically varies

between 16.5 m and 16.3 m with a maximum depth of 17 m and a minimum of 14 m over the whole

area. This shows the site to have no raised features which are likely to be impacted when trawled.

Figure 19 shows the bathymetry of the site and the locations where two cross-sections of the

bathymetric data have been take to provide profiles of the site in a west to east and a north to south

direction as show in Figure 20.

Figure 20 : Cross sections showing bathymetric profiles of the survey site, top image is a north-

south profile and the lower image a west to east profile (Figures © Envision).

Figure 20 illustrates that the site is very flat with only a 20 cm variation in both directions over 150 m.

Experimental trawls

Several trawls were carried out by the Lynn Princess which included a trawl across the experimental transect line, a number of short ten minute trawls to look for evidence of Sabellaria in the vicinity and

demonstrate gear to KSL, and finally, a reciprocal trawl to determine precision of repeat trawls.

Prior to the trawl across the transect, a two hour trawl was carried out outside the search area defined in Figure 17 just north of Blakeney Overfalls in order to fill the nets with the aim of maximising

impact of the experimental trawl. The catch was so poor however with almost empty nets (and no

evidence of Sabellaria) that it was decided to use empty nets for the experimental trawl. In this

regard, it is of note that both the skipper of the trawler and Vorberg (2000) consider that the impact of the net would be minimal, even if full.

Unfortunately due to strong currents the marker pellets on the moorings were dragged beneath the

surface of the water just prior to carrying out the experimental trawl, hence their location was difficult to determine and it was decided that the Lynn Princess should simply follow the survey vessel across

the transect.

A number of ten minute trawls to search (unsuccessfully) for Sabellaria were carried out post the

experimental trawl and the typical catches can be seen in Figure 21.


27

Figure 21 : Typical catch after a 10 minute haul showing crustaceans, echinoderms, fish and

macroalgae. Note abundance of Flustra sp. also seen as most dominant faunal

constituent from seabed video tows (Photograph © Kim Last).

It is suggested that a full scale trial of the effect of shrimp trawling would benefit from multiple trawls over the same area. To that end, attempts were made to determine the practicalities of organising a

commercial trawler to repeat trawls precisely. Unfortunately the plotter of the shrimp trawler was not

working at the time of the trial so rather than being given a pre-set track to follow as planned, the

trawler was asked to simply re-run the route of the previous trawl. The comparable paths are illustrated in Figure 22.


28

Figure 22 : Comparable paths of the trawler on the initial and reciprocal trawl with direction

of tow indicated (Image © Envision).

An attempt had been made at overlap with the original trawl after the double right angled change in

direction (brown line Figure 22) which represents recovery and deployment of the nets under strong

NE currents. It can be seen that the reciprocal trawl direction (SE) did not even cross the path of the

original trawl path (WNW).

Video images

Video assessment of the habitat using sled tows was intended to determine degree of spatial

heterogeneity as well as the effects of pink shrimp trawling on a range of epibenthic organisms. Five

video tows were undertaken prior to the experimental trawl and three tows following it (Figure 23).


29

Figure 23 : The video sled survey tracks and path of the trawler across the transect with

direction of travel indicated (Map © Envision).

Videos from the sled tows (see supplementary video material) were visually assessed using playback in real-time to determine faunal/floral composition to family level. Organisms identified on the videos

were limited to the most dominant families and absolute numbers were calculated per unit area and

corrected for tow length (Figure 24 and Figure 25). Identification to species level was not practical and considered inappropriate for the purposes of this study.

Five pre-trawl video sled tows were carried out at the study site to determine physical/biological

heterogeneity (Figure 24), and three post-trawl tows were undertaken to determine the potential

impact of shrimp trawling. However, pre- and post-trawl comparative analysis was confined to only those video sled tows which covered the shrimp trawl area as seen in Figure 23 hashed line at bottom.

Thus a comparison of family composition was carried out on the ends of pre-trawl 2 and 5 and the

start of post-trawls 1 and 3.

Family assemblage composition showed no significant differences between videos assessed pre-trawl

(ANOVA, fcrit 2.71, p = 0.58, Figure 24). However, there was a significant drop in assemblage

composition when videos were compared pre and post shrimp trawling (ANOVA, fcrit 3.07, p = 0.1,

Figure 25). The most dominant families observed in all videos were the bryozoan, Flustra and anemones.


30

Figure 24 : Organism assemblage composition at study site with five pre-trawl replicate video

sled tows.

Figure 25 : Comparative assemblage composition of pre- and post- shrimp trawl video tows.

Although no significant difference is seen between pre-trawl sled tows, a trend is observed of

decreasing number of organisms with sled tow number. Since sled tows were taken sequentially over

time, and the first tows coincided with slack water, we hypothesise that with increasing sled tow speed (due to increasing tidal flow), decreasing numbers or organisms are identified on the video. We


31

therefore propose that the significant difference in assemblage composition recorded between pre- and

post-trawl videos shown in Figure 25 is due to increasing underestimation of assemblages post-trawl, and not due to the effects of the shrimp trawl. This suggestion is supported by the fact that post-trawl

videos from 1 and 3 showed frequent loss of sled contact with the sea bed which was not witnessed in

pre-trawl 2 and 5.

We conclude that the site is characteristically uniform physically and in its organism assemblage. However significant sampling bias has probably occurred which has resulted in an underestimation of

organisms post shrimp trawling due to increased tow speed. This may be corrected for in future

studies by increasing tow number to cover all states of the tide, or to limit sled tows to the same tidal state which would be more challenging. Either approach would improve statistical power and

confidence in the data.

Sonar

The second lander configuration with the lower centre of gravity (Figure 8) was successful in enabling the system to remain at fixed location at the pilot study site whilst a scan was performed (see Figure

26 and Figure 27). The system was deliberately deployed during slack water to reduce the effect of

tidal currents.

Figure 26 : Scanning sonar image from the pilot site survey (Image © Envision).


32

Figure 27 : Scanning sonar image from the pilot site survey (Image © Envision).

Nevertheless, the low topographic complexity of the main survey site surveyed, corroborated by the

accompanying video mounted on the tripod lander, meant there were very few objects which could be identified and measured at this location, limiting the overall success of the sonar system for this pilot

study.

Laser

Video footage with laser lines was collected over the eight tow lines from the selected survey area (Figure 23), five pre- and three post-trawl. The footage from each transect was extracted and static

frames were exported every two seconds. These frames were then reviewed and a frame selected

which best represented, in terms of quality, 20 seconds of footage. The representative frame was then reviewed in image processing software and the laser line digitised (Figure 28). From this digitised

line it was then possible to determine a rugosity (or habitat complexity) value, R, using the ratio of

actual length of the line to the distance between start and end points using the formula:

R = Sr/Sg

Where:

Sr = the real surface distance between two points

Sg = the straight line geometric distance between the two points.


33

Figure 28 : Example of image, digitised laser line (Photograph © Envision).

Rugosity values were determined along each transect and mean values and variances determined for

each transect.

Table 2 : Mean rugosity values for each transect line and from each camera.

Transect

T1 T2 T3 T4 T5

POST_LEFT 1.180926494 1.226575458 1.202278113

POST_RIGHT 1.184065703 1.206394656 1.171648735

PRE_LEFT 1.382360929 1.478658447 1.413521074 1.225056238 1.228021926

PRE_RIGHT 1.242742774 1.298787161 1.196901898 1.484891765 1.232531406

Table 3: Variability of the rugosity values for each transect line.

Transect

T1 T2 T3 T4 T5

POST_LEFT 0.021824739 0.021458958 0.025944652

POST_RIGHT 0.001670487 0.002330323 0.005671576

PRE_LEFT 0.039864737 0.035841495 0.027831221 0.023709599 0.004969388

PRE_RIGHT 0.00376548 0.01245716 0.004813417 0.097090448 0.004719264

The rugosity values from the pre- and post- trawl transect lines show a slight decrease in rugosity post

trawl data, but it is suspected this is within the natural variation found within the survey area and the


34

very slight change is minimal. If large changes were observed, and or the seabed type was inheritably

more variable with district raised features which could be impacted, then it is expected that more definitive conclusions could be drawn.

Full statistical analysis of these data is possibly inappropriate due to the inaccuracies introduced

during field operations, lack of repeatability and replication. However measuring rugosity of the

seabed is feasible using a laser line profiler but the levels at which change can be measured require further investigations and a more robust dataset.

CRITIQUE OF TECHNIQUES AND EQUIPMENT

Moorings

The mooring deployment involving the ground line and tiles was very successful in Loch Etive. The

surface pellet buoys allowed for easy location of the transect line and the riser lines allowed the divers

to descend down “shot” lines. No ground line was laid at the pilot study site in the southern North Sea since this was unnecessary in the absence of divers. The effectiveness of the moorings laid at this site

was, however, markedly different.

The North Sea mooring surface marker buoys were intended to have been separated by ~40 m but

since there was no ground line it was difficult to pilot the vessel accurately over this distance. As a consequence the transect line became far longer than intended, probably ~100 m. Strong currents also

meant that the pellets were submerged for part of the tidal cycle severely hampering relocation efforts.

This issue was not experienced in Loch Etive suggesting that under benign current regimes (< 0.5 knots) the methodology is appropriate.

Future recommendation when not using a fixed ground line would be to use a slip line. This entails

looping a line around a slip ring on the first anchor deployed. This loop would be paid out from the steaming vessel until the desired transect length had been achieved and then the second mooring

anchor dropped. The marker line could then simply be slipped and retrieved. In this way the distance

of the transect is fixed. Although our survey was targeted to approaching neaps, currents were still

reasonably strong at peak flow (> 1.5 knots), which resulted in the marker pellets being pulled underwater. For this we suggest the use of very buoyant large (1 m diameter) fenders for future

deployments.

Multibeam

The GeoSwath plus multibeam system provided an accurate bathymetric model of the survey site and the surrounding environs. Whilst this is useful to enable the survey site to be placed in context, the

very low relief encountered over the site meant there were no distinct features to detect and measure.

If a site with more vertical relief was to be surveyed, it may be possible to determine areas and heights of distinct features such as reefs, and it may also be possible to detect any disturbance from fishing

gear.

Towed Video Sled

An attempt was made to run video sled tows parallel to the transect line but, as can be clearly seen from Figure 23, this proved a challenge in the prevailing currents. Two factors are probably of

greatest significance: poor boat handing from a survey perspective and strong currents. The latter

highlights the importance of conducting the full-scale survey under improved tidal conditions.

On a similar note the trawl path was intended to cross the middle of the transect, as in the Loch Etive trial (see Figure 14) but quite clearly did not in the North Sea trial (see Figure 23). This is possibly

due to the submergence of the transect marker buoys at the critical moment combined with lack of a

working GPS plotter, though a hand-held GPS was used aboard by KSL. A recommendation would be


35

to carry out several repeat tows and trawls (with gear up) during slack water. Once satisfied that a

good tow path is achievable, the survey should commence.

A critique of using any form of towed sled is the potential for disturbance by the sled on the seabed

from direct crushing or from side impact during the tow. We therefore qualitatively assessed the

video sled tow path crosses (total 5, see Figure 23) for signs of damage from sled i.e. sediment disturbance, scared hard surfaces, loss of Flustra sp. etc. Determining the exact location of tow path

crosses without knowing exact times and locations of the sled on the seabed was not possible but an

assessment of video during the approximate expected times of crosses showed no obvious disturbance. This is perhaps not surprising since the sled used in this survey is constructed of

aluminium sheet and plastic runners (Figure 4) and weighs ~15 kg in air which equates to ~5.5 kg in

water. The footprint of the sled is ~0.3 m2 so the overall weight of the sled in water when stationary is

~17 g/cm2. Sled weight on the ground would also be much lower during the tow. Comparatively the

brown shrimp trawl (beam and shows) used by Vorberg (2000) was calculated to exert a downward

pressure of 230 g/cm2 when stationary. Since there were also no observable signs of impact from the

shrimp trawl itself in this study, on either the benthic organism composition or on rugosity we can conclude that the impact of the sled tow be minimal.

Video System

The camera systems used for this pilot survey worked well. Visibility was reasonably good and only

one sled tow had to be abandoned after the sled got tangled in the mooring pellet riser. However the quality of data extracted from video was strongly influenced by tow speed (which needed to be

increased as current speed increased) leading to significant underestimation of organism numbers post

trawl. When the sled was slow moving, identification was much better than when tow speed increased, to the extreme when excessive tow speed resulted in the sled lifting off the ground

altogether. Abundance counts were also open to positive bias to organisms which are either more

abundant and/or more visible (larger or less camouflaged). The analysis technique employed here

attempted to minimize this bias in focusing on only the most abundant and visible organisms. Animals seen more often when the video sled was stationary or only moving very slowly i.e. hydroids, brittle

stars and hermit crabs were removed from subsequent analysis as counts where considered under

representative.

A serious short fall of the current investigation is the limited video of seabed which actually

corresponds to the trawled area. As can be seen in Figure 23, only two tows crossed the trawl path

before and two crossed after which equates to a very small percentage of the seabed assessed. Confidence in the belief that there is no impact is therefore limited.

Laser Profiler

The laser profiler was tested in the Loch Etive trials and the controlled conditions of Dunstaffnage

Bay, following which modifications were made. The main drawback of the original, hired system was the large amount of surface side equipment that was required for data processing, the additional

deep-sea specified umbilical for power and an additional camera signal. This made the system

cumbersome to deploy in very shallow water with a significant chance that the sled would be inverted

and the system damaged. Nevertheless, under controlled or benign conditions the laser profiler produce good results and enabled small pieces of Sabellaria tubes to be profiled and the relative

measurements to the surrounding seabed made.

The simplified, battery-powered laser profiler that was subsequently developed by Envision Ltd. was significantly easier to deploy, especially in the more exposed conditions and dynamic environment of

the southern North Sea. The battery system did suffer the drawback of loss of power during survey

operations but this was easily detected by the live video feed and subsequently remedied. Two benthic habitat video cameras were deployed on the sled and the laser line spread equally across both images.


36

Unfortunately this did not cover either field of view fully, and it is recommended that in future the

laser be concentrated on one camera to give a good/full coverage of the whole frame.

The original system under hire used an automated process to detect the laser line profile whilst the

Envision laser required some manual inputs in determining rugosity which may introduce operator

bias into the results. For future laser line rugosity calculation, it is recommended that an automated

process is developed, with a consistent and repeatable line detection method not amenable to observer bias. Alternatively data needs to be processed “blind”.

As with the video system the speed of tow is critical to the quality of data and if tides or vessel

movement cause the system to be towed too rapidly then the quality of the data is compromised.

The use of a laser line profiler to detect seabed rugosity is possible, but assessing if the system can be

used to detect change is difficult especially if the natural variations in rugosity are either low or high.

In this study the sea bed surveyed was flat and relatively featureless and any changes in could not be identified. This lack of detection may be due to the natural variations in rugosity which are at a level

at which change introduced experimentally cannot be detected and discerned against background

measurements. Further work would be required on a more appropriate seabed type to ascertain if a

laser line profiler can detect change in seabed rugosity due to shrimp trawling.

Lander

The tripod lander (Figure 7) was successfully deployed in controlled conditions of Dunstaffnage Bay

and Loch Etive with the frame resting on the sea floor for sufficient time to extract frame grabs from

the video and to build an image with the scanning sonar. Being light-weight, the lander was also capable of very rapid deployment, and the frame could be lifted briefly from the sea floor and

repositioned to obtain multiple views from a restricted sampling area. However, the lander system

was liable to the effect of tidal currents and vessel movements whilst the system was being deployed and had a tendency to be „blown‟ or pulled over if the conditions were not ideal.

The modified lander design with the lower centre of gravity (Figure 8) initially proved stable under

the higher current speeds in the North Sea survey, providing a stable platform for the sonar which

required being on the seabed for at least one revolution of the sonar head (~1 minute). However due to the time required for the lander to complete a scan, rapid vessel drift in strong currents required large

amounts of umbilical to be rapidly paid out. When drift distance during the time required to scan

exceeded the length of umbilical, the lander would become knocked down. Hence it is recommended that any lander equipment utilised during a full scale survey should stipulate that the system requires

deployment during a slack water window and in calm conditions, with an extended umbilical and/or

using a vessel with dynamic positioning to maintain position of the vessel relative to the seabed and ensure suitable quality data is collected.

Scanning Sonar

The scanning sonar and lander were deployed successfully in the controlled conditions of

Dunstaffnage Bay and Loch Etive allowing objects on the seafloor to be identified and measured (Figure 29).


37

Figure 29 : Scanning sonar image of an object, specifically the walls of a test tank, from a

controlled test at SAMS (Image © Envision).

The lander/sonar system was also deployed during operations at the experimental survey site off the Norfolk coast but tidal currents made it unfeasible for the tripod configuration (Figure 7) to remain in

a fixed location or upright and the lander was „blown‟ over. The second configuration (Figure 8) was

relatively more successful in the strong currents and enabled the system to remain on at fixed location whilst a scan was performed (see Figure 26 and Figure 27).

The low topographic complexity of the main survey site surveyed meant there were very few objects

which could be identified and measured at this location. If a more complex topography were to be

surveyed, then it is possible that digital scanning sonar could enable objects to be detected and the distribution of objects around the sonar measured which could give an indication of the spatial

heterogeneity of objects raised from the surrounding surface. This was demonstrated in the Loch Etive

survey (Figure 16 and Figure 30), which illustrates the potential of scanning sonar trialled on a S. spinulosa reef in the Wash trialled during a previous survey.


38

Figure 30 : Example of scanning sonar image of Sabellaria reef with scanning head positioned

1.5 m above the sea floor. The accompanying image is a frame grab from a bullet

camera attached to the lander (Images © Envision).

Diver Survey

Discussions with the NFSD dive team revealed that although currents at Loch Etive site were weak

they had an impact on the survey strategy. The intention had been to conduct the video transects up current (avoiding disturbing the seabed sediment) but this was abandoned in favour of a survey with

the current. The consequence of this was that the divers were travelling very fast and although the

resultant footage was of general interest, it was difficult to determine height of seabed structures due to the speed at which footage had been collected.

An important consideration when using divers is that bottom time is restricted by depth. For the Loch

Etive trial at 12 m depth, bottom time was only half an hour. This will obviously reduce still further if current speeds are higher or if up current surveying is attempted. Since the timing is limited, so too

are the number of tasks. A dive pair could lay a transect and video at the same time, especially if the

seabed was not too disturbed, but still images of tiles together with direct Sabellaria measurement

would probably then require another diver pair survey.

The transect tiles were very visible and provided good markers of distance along the transect as well a

scale reference for determining point samples. The tiles however were also intended to provide a

reference for seabed disturbance and so it was a little surprising that the trawl (with 100 kg boulder) had showed no visible signs of tile displacement. A recommendation would be to attach short lengths

of threaded rod bolted vertically to the middle of the tile thereby providing a “hook” to promote tile

displacement by a passing trawl, and hence direct evidence, of the trawl having crossed the transect.

Experimental Trawl

The combination of a damaged plotter on the Lynn Princess and the (unintentional) sub-surface

marker buoys made accurate headings challenging, both for the transect line crossing and for the

reciprocal trawl. Recommendations have already been made to improve marker buoy visibility and

future surveys should consider some training of skippers (trawler and survey vessels) in maintaining course to waypoints or trawl/survey paths


39

DECISION – SUPPORT TOOL

Prior to any management decision being made in regard to Sabellaria spinulosa reef, specifically in

regard to shrimp trawling operations, the following have to be known:-

How „reefy‟ a reef is, and how important is it in the general scheme of reef habitat locally and

globally?

How sensitive reefs of different „reefiness‟ are to fishing operations?

How damaging are the different intensities of fishing to the reefs?

How likely is damage to take place? (From real or potential fishing intensity)?

What are the theoretical options for management?

The management decision then needs to be made by balancing the first three points and deciding what management option can be justified.

A management tool in its entirety needs to address all the points outlined above. The idea for a

framework was to build up a scoring system for each component („reefiness‟/reef status, sensitivity, likelihood of damage) that could be applied (with discussion and negotiation on exactly what the

scores translate into) to real situations (Figure 31).

Figure 31 : Assessing Sabellaria reef and management options.

As the formative stage of this management framework, a decision-support tool has been developed to

the stage of providing a numerical scoring and visual representation of the “reefiness” of a test site following assessment through a variety of survey techniques. The tool is based on the theoretical

scoring system for evaluating „reefiness‟ in the context of the Habitats Directive previously developed

by the authors (Hendrick & Foster-Smith, 2006). Figure 32 illustrates a hypothetical example of

Sabellaria reef scored for four physical characteristics: reef extent, patchiness, elevation and consolidation, together with default scores for feature importance and estimates of confidence as

described in Hendrick & Foster-Smith (2006). It is unfortunate that the absence of Sabellaria during

the course of this pilot study has meant that the tool could not be tested on a reef habitat.


40

Figure 32 : A hypothetical example of a Sabellaria reef scored for the physical characteristics

of reef extent, patchiness, elevation and consolidation (screen-grab).

As a further stage in the framework development, the tool has been progressed to allow its use in

assessing how sensitive reefs of different „reefiness‟ are to fishing operations (bullet 2 above). In the

context of this pilot study, for instance, it was expected that the features of the target reef could be scored both pre- and post- experimental trawl with any changes in overall score giving an indication

of impact from the trawling event. Figure 33 further illustrates the hypothetical example suggesting a

decline in the „reefiness‟ of the hypothetical Sabellaria aggregation following the trawl with a

numerical comparison of the overall reefiness scores both pre- and post-trawl together with an indication of confidence in the comparison.


41

Figure 33 : Hypothetical pre- and post trawl assessment of a Sabellaria reef following shrimp

trawling (screen-grab).

Had a Sabellaria aggregation been encountered during this pilot survey, we had hoped to demonstrate

what is possible through local application of the tool, with specific scoring guidance for the impact of shrimp trawling on S. spinulosa reefs. It is unfortunate that this did not prove possible. However, we

see added potential from the generic nature of the tool in that the approach (the way scores are

compiled and compared) could be applied to any feature, not just Sabellaria reefs but any biogenic, or indeed artificial, structure. We expect further development of the full tool to occur through use, and

only following extensive consultations with relevant stakeholders.

RECOMMENDATIONS FOR A FULL SCALE SURVEY

This report was envisaged to encompass a wide range of equipment and methods designed and used to

test physical impacts of shrimp trawling on Sabellaria reef. However, since Sabellaria reef was not found we have focussed on providing a critique of equipment and methods capable of such a

hypothetical assessment which will hopefully inform future survey improvements.

In general, techniques worked adequately and were able to address different habitat assessments.

However, the importance of deployment of divers or equipment in minimal currents cannot be overstated and a priority should be to target slack water and neap tides at times of the year when the

tidal range is lowest at this locality i.e. <3 m. The implication of this targeting is that there are actually

very few opportunities in the year to conduct such a survey, particularly when survey times need to be


42

coincident with a suitable weather window. A solution to this maybe to select a reef site which is very

shallow or even interidal where the drag on remote sensing lines/umbilicals will be reduced and diver bottom time will be increased.

We cannot at this stage suggest alternative techniques that are likely to perform any better even with

significant additional costs. The equipment used during this trial is relatively inexpensive to purchase

or hire (other than the GeoSwath system) and proved effective in the Wash habitat. Furthermore the sampling gear is relatively (compared to the trawl itself) light-weight and our pilot trial showed no

evidence of disturbing the seabed.

A full scale trial would necessitate repeated trawling of the same ground to limit the amount of sampling effort. The spatial scale of the survey also needs careful consideration, as even with slack

water and calm conditions location and relocation of survey lines will probably be difficult. Since our

efforts at repeated trawling were very poor we recommend backup/independent GPSs and a certain level of vessel master training prior to a full scale survey so as not to impact on valuable

current/weather window time.

As part of this pilot study, a day was dedicated to finding Sabellaria reef outwith any protected

habitat. This proved unsuccessful. Therefore any future survey will require considerable effort and cost to locate the experimental reef of sufficient size so that at least one discrete part could be

experimentally trawled.

And finally it is perhaps also important to consider what level of change detected in trawled Sabellaria reef is of interest and/or importance to the stakeholders? This question will be central to

planning any future impact study and can be used to inform and build the decision support tool

described as part of this trial.

LITERATURE CITED

Berghahn R. & Vorberg R. (1997). Shrimp Fisheries and Nature Conservation in the National Park Wadden Sea of Schleswig-Holstein. UBA-Texte 82/97. 197 pp.

Foster-Smith R.L., Egerton J., Foster-Smith D., Sotheran I. & Meadows W. (2010). Developing new

ground truthing techniques for seabed mapping. Report for the Marine Aggregate Sustainability Levy Fund (MASLF). MEPF/08/64. 21 pp.

Foster-Smith R.L., Sotheran I.S. & Walton R. (1997a). Broadscale mapping of habitats and biota of

the sublittoral seabed of the Wash, North Norfolk and Lincolnshire coasts. Draft interim report of the

1997 Broadscale Mapping Project (BMP) Survey. Report for English Nature. 9 pp.

Foster-Smith R.L., Sotheran I.S. & Walton R. (1997b). Broadscale mapping of habitats and biota of

the sublittoral seabed of the Wash. Final report of the 1996 Broadscale Mapping Project (BMP)

Survey. Report for English Nature. 25 pp.

Foster-Smith R.L. & White W. (2001). Sabellaria spinulosa in the Wash and north Norfolk cSAC and

its approaches: Part I: Mapping techniques and ecological assessment. A report for the Eastern Sea

Fisheries Joint Committee and English Nature. 43 pp.

Hayward P. & Ryland J. (1990). The Marine Fauna of the British Isles and Western Europe. Oxford University Press, New York 627 pp.

Hendrick V.J. & Foster-Smith R.L. (2006). Sabellaria spinulosa reef : a scoring system for evaluating

'reefiness' in the context of the Habitats Directive. Journal of the Marine Biology Association, UK. 86: 665-677.

Holt T.J., Hartnoll R.G. & Hawkins S.J. (1997). Sensitivity and vulnerability to man-induced change

of selected communities: intertidal brown algal shrubs, Zostera beds and Sabellaria spinulosa reefs. English Nature Research Reports. No. 234. 97 pp.


43

Holt T.J., Rees E.I., Hawkins S.J. & Seed R. (1998). Biogenic Reefs: an overview of dynamic and

sensitivity characteristics for conservation management of marine SACs. Natura 2000 report prepared for Scottish Association of Marine Science (SAMS). Oban. (UK Marine SACs Project, Vol. IX). 170

pp.

Jones L. (1999). Habitat Action Plan: Sabellaria spinulosa reefs. English Nature. 4 pp.

Last K.S., Hendrick V.J., Beveridge C.M. & Davies A.J. (2011). Measuring the effects of suspended particulate matter and burial on the behaviour, growth and survival of key species found in areas

associated with aggregate dredging. Report by SAMS for the marine Aggregate Levy Sustainability

Fund (mALSF). Project 08/P76. 70 pp.

Marine Ecological Surveys Ltd., MarLIN & DeepBlueSky (2011). Pandalus. The online Marine

Macrofauna Genus Trait Handbook Update. Funded by Defra through the Marine Aggregate Levy

Sustainability Fund (MALSF). Project Ref: MEPF 10/P142. Accessed: March 2012. Available from: http://www.genustraithandbook.org.uk/.

Mistakidis M. (1956). Survey of the pink shrimp fishery in Morecambe Bay. Lancashire and Western

Sea Fisheries Joint Committee. 14 pp.

Mistakidis M. (1957). The biology of Pandalus montagui Leach. Fisheries Investigations, London Series 2. 21: 52pp.

Rees H. & Dare P. (1993). Sources of mortality and associated life-cycle traits of selected benthic

species: a review. MAFF Directorate of Fisheries Research. Lowestoft. 37 pp.

Reise K. (1982). Long term changes in the macrobenthic invertebrate fauna of the Wadden Sea: are

polychaetes about to take over? Netherlands Journal of Sea Research. 16: 29-36.

Reise K. & Schubert A. (1987). Macrobenthic turnover in the subtidal Wadden Sea: the Norderaue revisited after 60 years. Helgoländer Meeresunters. 41: 69-82.

Riesen W. & Reise K. (1982). Macrobenthos of the subtidal Wadden Sea: revisited after 55 years.

Helgoländer Meeresunters. 35: 409-423.

Sankey S. (1987). The shrimp fishery and its bycatch in the North Western and North Wales Sea Fisheries Committee district. North Western and North Wales Sea Fisheries Committee. Original

reference not seen. Cited by Holt et al. (1997). Sensitivity and vulnerability to man-induced change of

selected communities: intertidal brown algal shrubs, Zostera beds and Sabellaria spinulosa reefs, English Nature Research Reports No. 234.

Schäfer W. (1972). Ecology and Palaeoecology of Marine Environments. Translation of Aktuo-

paläontologie nach Studien in der Nordsee. University of Chicago Press, Chicago 568 pp.

Taylor P.M. & Parker J.G. (1993). An Environmental Appraisal: The Coast of North Wales and North West England. Hamilton Oil Company Ltd. 80 pp.

UK Biodiversity Group (1999). Tranche 2 Action Plans - Volume V: Maritime species and habitats.

English Nature.

Vorberg R. (1997). Auswirkungen der Garnelenfischerei auf den Meeresboden und die Bodenfauna

des Wattenmeeres. Verlag Kovac, Hamburg. Original reference not seen. Cited by Vorberg R.

(2000). Effects of shrimp fisheries on reefs of Sabellaria spinulosa (Polychaeta). ICES Journal of Marine Science. 57: 1416-1420.

Vorberg R. (2000). Effects of shrimp fisheries on reefs of Sabellaria spinulosa (Polychaeta). ICES

Journal of Marine Science. 57: 1416-1420.

Warren P. (1973). The fishery for the pink shrimp Pandalus montagui of the Wash. Laboratory Leaflet (New Series) No. 28. Ministry of Agriculture, Fisheries and Food. Lowestoft. 46 pp.

Warren P. & Sheldon R. (1967). Feeding and migration patterns of the pink shrimp Pandalus

montagui in the estuary of the River Crouch, England. Journal of Fisheries Research Board of Canada. 24: 569-580.

http://www.genustraithandbook.org.uk/


44

ACKNOWLEDGEMENTS

We are grateful to the respective skippers and crew of the three survey boats used during the course of

this study: RV Seol Mara (SAMS); Kirkspray and the Vanguard (both of Safety Boat Services), and to

the skipper, Dave Mott, and crew of the shrimp trawler Lynn Princess (Lynn Shellfish Ltd.). We are

also indebted to the divers from the National Facility for Scientific Diving: Elaine Azzopardi, Hugh Brown and Simon Thurston for their help with the diving survey. We are also very grateful to Judy

Foster-Smith and Alison Benson (both of Envision) for their help with data processing. Finally, we

are grateful to Charlotte Johnson (Natural England) and Michael Coyle (Marine Management Organisation) for their advice and support throughout this study.


45

APPENDIX 1 : TECHNICAL SPECIFICATIONS OF THE GEOSWATH SYSTEM.

Shallow Water Multibeam and Side Scan System

Description

GeoSwath Plus offers very efficient simultaneous swath bathymetry and side scan seabed mapping with accuracies that have been shown to exceed the IHO Standards for Hydrographic Surveys. The

applied phase measuring bathymetric sonar technology provides data coverage of up to 12 times the

water depth, giving unsurpassed survey efficiency in shallow water environments.

System Components

The GeoSwath Plus turn-key solution comprises a dual transducer head with versatile mounting

options as well as a deck unit containing the complete sonar electronics together with a high spec PC,

running the GeoSwath Plus software. The software provides full acquisition, calibration and data processing capabilities for producing the final bathymetry map and side scan mosaic data products.

All customary ancillary sensors can be directly interfaced

Dual Transducer wet end

The rugged port and starboard transducers, available in three frequency options (125, 250, 500 kHz),

can be attached to a supplied sonar head assembly for boat over-the side or bow-mount options, which also can accommodate a range of ancillary sensors. Alternatively they can be deployed on bespoke

boat hull mount, as well as bespoke ROV and AUV assemblies.

Deck Unit

The compact workstation contains the complete system electronics as well as a high spec PC. All peripheral sensors (position, motion, heading, transducer face sound velocity, sound velocity profiler

and tide) are interfaced directly.

Software

Operating on Windows XP, the GeoSwath Plus software package provides a complete project based solution, including acquisition, storing and editing of sonar and ancillary data, grid-based patch test

calibration, data processing with audit trail, advanced bathymetry data gridding and side scan

mosaicing, data visualisation including 3D fly-through capability

Features

Ultra high resolution swath bathymetry

IHO SP-44, special order

Co-registered geo-referenced side scan

Frequency versions: 125, 250, 500 kHz

Up to 12 times water depth coverage

240º view angle

Dual transducer set-up with versatile mounting options

Full software solution included: data acquisition, processing, presentation


46

Interfaces to all customary peripheral sensors

Interfaces to all customary software packages

Technical Specifications

GeoSwath Plus 125 kHz 250 kHz 500 kHz

max Water Depth Below Transducers

200 m 100 m 50 m

max Swath Width 780 m 390 m 190 m

max Coverage up to 12 x depths

Depth Resolution 6 mm 3 mm 1.5 mm

Two Way Beam Width

(Horizontal)

0.85º 0.75º 0.5º

Transmit Pulse Length 128 μS to 896 μS 64 μs to 448 μS 32 μs to 224 μs

max Swath Update Rate 30 per second (range dependant)

Transducer Dimensions 540 x 260 x 80 mm 375 x 170 x 60 mm 255 x 110 x 60 mm

Transducer Weight 11.6 kg (in air)

3.3 kg (in water)

3.8 kg (in air)

1.8 kg (in water)

1.5 kg (in air)

0.5 kg (in water)


47

APPENDIX 2 : ABBREVIATIONS

BAP Biodiversity Action Plan

dGPS Differential Global Positioning System

GPS Global Positioning System

IDRBNR SAC Inner Dowsing Race Bank North Ridge Special Area of Conservation

IFCA Inshore Fisheries and Conservation Authority

MALSF Marine Aggregate Levy Sustainability Fund

MPA Marine Protected Area

NFSD National Facility of Scientific Diving

PVC Polyvinyl chloride

SAC Special Area of Conservation

SAMS Scottish Association of Marine Science

WAAS Wide Area Augmentation System

WNN SAC Wash and North Norfolk Special Area of Conservation

ASSESSING THE IMPACTS OF SHRIMP FISHING ON … · Sabellaria reefs has long been accepted (Marine...

Documents

Transcript of ASSESSING THE IMPACTS OF SHRIMP FISHING ON … · Sabellaria reefs has long been accepted (Marine...