Scottish Marine and Freshwater Science Volume 5 Number 14: Electrofishing for Razor Clams (Ensis siliqua and E. arquatus): Effects on Survival and Recovery of Target and Non-Target Species

Materials and Methods

Boat Trials

Vessels and Electrofishing Equipment

The study was carried out on three commercial inshore fishing vessels (Table 1). Each vessel was equipped with an AC generator and a rig of either two ( FV Lucky Lucy) or three pairs of brass electrodes ( FV Ensis and FV Nicola Jane). The electric rig on these vessels was connected by copper cables to the generator via a transformer to reduce the output voltage to ~ 25 v and the current to ~ 80 A (Table 2). These voltages were the maximum output possible for the equipment used on these vessels. Previous reports on electrofishing have looked at DC systems (Woolmer et al. 2011) where the electrodes have been shown to corrode and release chlorine gas when the generator is switched on. This has not been observed using an AC system either in tanks (see below) or whilst fishing (Video A1). Furthermore, razor clam divers have reported metal components of the divers' equipment, such as the inflator hose connecting the drysuit and the exposed metal components of regulators and cylinders also corrode when a DC system is used (R. Grieve pers. com.). As the use of electric currents is prohibited under Article 31 of EU regulations 850/98 "unconventional fishing methods" a derogation was required to legally use the equipment in this study. Derogations for each vessel were issued by Marine Scotland Licencing on the basis that a research scientist was on board at all times the equipment was in use.

Table 1: Vessel data (from Marine Management Organisation ( MMO) lists of registered vessels, UK Government 2014)

Vessel	FV Ensis	FV Nicola Jane	FV Lucky Lucy
Length	10.65	11.32	11.25
Port letters and number	OB1004	OB1043	SR48
Administrative port	Oban	Oban	Ayr
Home port	Fort William	Oban	Ayr
Launch port	Pittenweem	Mallaig	Largs
Registry of Shipping	C18291	A16424	B10363
Licence number	41516	50107	41644
Skipper	David Simister	Robbie Grieve	Michael Crowe

Study Areas

The study was carried out in three sites around Scotland: East Fife, Loch Nevis and the Clyde sea area around the Isle of Cumbrae (Figure 3). The sites are in shallow inshore waters (< 10 m) with a sandy seabed; and are currently viable fishing grounds used by the vessels. Measurements of seawater salinity (using a hand held optical refractometer) and temperature at the seabed were taken at each location (Table 2). Sediment cores (5 cm diameter, 10 cm high, 196 cm ³) were collected (6 in Fife and Loch Nevis, 4 in the Clyde) and particle size analysis was conducted ( Table A1, Figure A1).

Table 2: Location, water temperature, salinity, target species and electrical output for each vessel.

Vessel	FV Ensis	FV Nicola Jane	FV Lucky Lucy
Date	24 Sept 2013	30 Sept 2013	9 Dec 2013
Location	East Fife	Loch Nevis	Isle of Cumbrae
Target species	Ensis siliqua	Ensis arcuatus	Ensis siliqua
Water temperature (°C)	12	12	9
Salinity	34	31	33
Voltage	24	24	Missing value*
Current (A, mean ± 1 SD)	83 ± 2.07	82 ± 1.73	80

* The transformer on this vessel did not have a voltage meter, however, current was standardised across vessels so the voltage is believed to be comparable.

Figure 3: Map of trial locations by vessel. Projection: WGS1984.

Experimental Design

Electrical Stimulus

An electrical stimulus was applied to the seabed, designed to mimic commercial electrofishing activity as closely as possible. The vessel anchored in shallow water and let out a 100 m line. Pairs of brass electrodes connected to an onboard generator were lowered to the seabed to tow behind the vessel (a more detailed description of the electrical equipment is provided below in the Tank Trials section). Once the line was let out, the diver entered the water and took up position behind the electrodes. The diver instructed the skipper when to start the generator powering the electrodes and to start the winch, either through pulling a rope attached to a clanger on the vessel or via a divers voice communication unit installed in the mask. The on-board winch was then used to drag the vessel towards the anchor, towing the brass electrodes across the seabed, a technique known as fly-dragging (Figure 4). The speed at which the vessel is fly-dragged was determined by the diver based on visibility and the population density of razor clams (mean speed of 3 m min ^-1). When the electrodes exited a patch of razor clams, the diver signalled the skipper to turn off the electrodes. Replicate short tows were conducted: seven on the FV Ensis, six on the FV Nicola Jane and four on the FV Lucky Lucy.

Figure 4: Schematic diagram of electrofishing for Ensis spp. (Figure 1.7, Breen et al. 2011)

In situ Recovery

Once the electrodes were switched off, the diver placed quadrat frames (1 m ² base, n = 2) where the electrodes had been active within minutes of switching off the electricity (time between electrodes turned off and quadrats placed: mean ± sd = 2 min 13 s ± 54 s). A high definition camera (GoPro Hero3 Black Edition, Woodman Labs, USA) was mounted to the top of each frame and used to record high-definition, wide-angle video footage (resolution: 1920 x 1440 pixels, 48 frames s ^-1) to a 32 GB memory card accessed after recovery of the cameras. Video footage was downloaded to a laptop computer between replicate runs. The footage recovered was used to observe the recovery of Ensis spp. and other species, and to identify the non-target species encountered during electrofishing activity. The quadrats were left in situ until all the animals within the quadrat had recovered or for a maximum of 30 minutes. The quadrat frames were rotated between electrode pairs between runs. Four replicate videos were taken for each pair of electrodes (12 videos for the FV Ensis and FV Nicola Jane, 8 videos for the FV Lucky Lucy).

The recovery start times were recorded as the first sign of movement from each individual animal in the videos after the electrodes were turned off. The quadrat placement time was recorded as the recovery time for non-target species that were displaying normal movement patterns when the quadrat was placed. This gave a conservative estimate of recovery times. For Ensis spp. missing values for the recovery start times due to recovery starting before the quadrat was placed were omitted from recovery analyses, but these individuals were accounted for in calculating the population densities. A recovery end point was also noted for Ensis spp. when the clams had less than 1 cm of shell left showing above the sediment. The shell lengths of the razor clams were estimated from freeze frame stills taken from the videos. In order to minimise error due to visual distortion in the videos, shells were measured relative to the side of the quadrat they were closest and most parallel to. Lengths were recorded to the nearest cm. Videos from the FV Lucky Lucy were excluded from the analysis owing to logistical problems including: time constraints forcing FV Lucky Lucy trials to be conducted in December; use of a sheltered bay as the weather was poor, which restricted the space available for fly-dragging; and low light levels resulting in a lower quality of video footage in which it was more difficult to distinguish organisms from the seabed.

Table 3: Location, depth and speed data for tows conducted and analysed.

Date	Vessel	Trawl	Electrodes on			Electrodes off			Length of trawl
			Time (min)	Location	Depth (m)	Time	Location	Depth (m)	Time (min)	Distance (m)	Direction	Speed (m min -1)	Organism exposure time to electric field (decimal min)
24 Sept 2013	Ensis	1	0948	056°12.133 N 002°53.915W	4.1	0952	056°12.133 N 002°53.923 W	4.1	4	4	270°	1	3
24 Sept 2013	Ensis	2	1040	056°12.132 N 002°53.925 W	4.2	1044	056°12.132 N 002°53.934 W	4.2	4	9	270°	2.25	1.33
24 Sept 2013	Ensis	3	1105	056°12.138 N 002°53.947 W	4.1	1107	056°12.140 N 002°53.948 W	4.1	2	3	344°	1.5	2
24 Sept 2013	Ensis	4	1251	056°12.231 N 002°54.464 W	5.8	1253	056°12.232 N 002°54.462 W	5.8	2	2	048°	1	3
24 Sept 2013	Ensis	5	1438	056°12.112 N 002°54.097 W	6.8	1446	056°12.108 N 002°54.085 W	6.8	8	14	120°	1.8	1.67
24 Sept 2013	Ensis	6	1517	056°12.100 N 002°54.068 W	7.2	1519	056°12.101 N 002°54.064 W	7.2	2	4	065°	2	1.5
24 Sept 2013	Ensis	7	1623	056°12.104 N 002°54.042 W	7.8	1626	056°12.104 N 002°54.039 W	7.8	3	3	089°	1	3
30 Sept 2013	Nicola Jane	1	1215	057°01.699 N 005°44.782 W	8.0	1217	057°01.695 N 005°44.766 W	7.9	2	17	114°	8.5	0.35
30 Sept 2013	Nicola Jane	2	1256	057°01.688 N 005°44.763 W	8.4	1258	057°01.690 N 005°44.753 W	8.2	2	10	069°	5	0.6
30 Sept 2013	Nicola Jane	3	1332	057°01.704 N 005°44.743 W	7.2	1334	057°01.703 N 005°44.732 W	7.9	2	11	099°	5.5	0.55
30 Sept 2013	Nicola Jane	4	1418	057°01.702 N 005°44.730 W	8.1	1419	057°01.698 N 005°44.723 W	8.2	1	10	136°	10	0.3
30 Sept 2013	Nicola Jane	5	1503	057°01.689 N 005°44.736 W	8.8	1505	057°01.690 N 005°44.728 W	8.7	2	8	077°	4	0.75
30 Sept 2013	Nicola Jane	6	1618	057°01.686 N 005°44.729 W	8.6	1620	057°01.687 N 005°44.717 W	8.6	2	12	081°	6	0.5

Sandeel Recovery

A diver survey was conducted in Ayrshire to focus on potential impacts on sandeels ( Ammodytes marinus) caused by electrofishing. Sandeels are fish which bury into the seabed, and as such are potentially at risk from electrofishing activity. Sandeel species, notably A. marinus, are ecologically important as prey for breeding seabirds such as puffins, guillemots and kittiwakes (Wanless et al. 2005). Two short transects (055° 18.780 N, 004° 50.488 W - 055° 18.743 N, 004° 50.594 W, and 055° 18.743 N, 004° 50.594 W - 055° 18.709 N, 004° 50.698 W) were surveyed by divers from the FV Ensis launching out of Girvan (14 May 2014). Electric rods were fly-dragged across the seabed as described above. Divers followed the rods and collected any stunned sandeels in a hand held net. The survey was recorded by Go-Pro cameras mounted to the diver and the number of sandeels collected was counted from the video footage. Stunned fish were placed in a bucket of seawater on deck within 10 minutes of collection from the seabed and their recovery, taken as restoration of normal swimming behaviour, was monitored. The number of transects was limited by mechanical failure of the winch the vessel was using to fly-drag.

Physical Impact of Fishing Gear on the Seabed

An additional diver survey was conducted in Loch Nevis to make observations on the impact of the electric fishing equipment on the seabed. A second diver entered the water to video normal fishing activity conducted by a diver fisherman with a mounted GoPro camera.

Statistical Analysis

Linear models were constructed to test for the effects of species identity and location in non-target species recovery and for the effects of species identity, location, population density and shell length on the recovery start and end times of Ensis spp. Criteria for normality and homogeneity of variance were met (following Quinn and Keough 2002) for the recovery times of non-target species and the recovery end times for Ensis spp.. However, during the initial model building phase of the Ensis recovery start time analysis, diagnostic residual plots indicated the presence of heterogeneity of variance due to differences in density variances, not accounted for by the fixed effects. A linear model with a generalised least squares( GLS) estimation was used, which allows variance functions to be explicitly included to model the variance structure, avoiding the need for data transformation. Therefore, a density-specific variance-covariate was included to model the variance (Pinheiro & Bates 2000). Following the inclusion of density-specific variance covariates, diagnostic residual plots indicated homogeneity of variance. Parameters in the final models were estimated using restricted maximum likelihood ( REML, following West et al. 2007). REML was used in preference to maximum likelihood ( ML) as it accounts for the loss of degrees of freedom in estimating the fixed effects, thus producing unbiased estimates of the covariance parameters (West et al. 2007). All statistical analyses were carried out in the R statistical and programming environment (version 2.15.0, R Development Core Team). GLS analysis was conducted using the nlme package (v3.1-101; Pinheiro et al. 2011) in R.

Tank Trials

Two large fibreglass stimulus tanks (Tanks A and B, 3.76 m x 1.86 m x 1 m deep) and five smaller fibreglass holding tanks (Tanks C and D: 1 m diameter x 0.80 m deep; Tanks E and F: 1.07 m x 0.83 m x 0.6 m deep; Tank G: 0.7 m x 2.1 m x 0.82 m deep) were assembled at Loch Leven Shellfish Ltd, Onich, Scotland. Sand (30 cm deep in tanks A and B, 5 cm deep in tanks C-G, Table A1, Figure A1) was overlain with seawater (20 cm depth in tanks A and B, mean depth of 36 cm in tanks C-G) which was pumped continuously to all tanks from Loch Leven by a submersible pump at a depth of 15 m on a flow through system, and returned to the loch. Seawater temperature, salinity and dissolved oxygen were measured daily on all tanks for the duration of all experiments ( Table A2). The sand in tank A was replaced between the initial survival trials and the later stimulus trials. No other tanks were used in the initial survival trials.

Electrical Tests

The electrical stimulus applied and the rig geometry used during vessel and tank trials represent one of a number of systems thought to be in use commercially. The illegal and unregulated nature of the fishery means that little information is available as to type of electrical stimulus and the construction and design of the electrode system as there is no defined industry or regulatory standard. The parameters used were those presented to the research team as being a "commonly used" commercial system. (J and R. Grieve pers. comm.). Observations, measurements and conclusions reported relate only to these specific gear parameters.

Tank tests were designed to look at the electrical properties of a fishing rig being used for harvesting razor clams in the Scottish fishery. All electrical tests were conducted in stimulus tank A and survival tests were conducted in both stimulus tanks A and B. The system used, supplied by Lochleven Shellfish Ltd., was an alternating current ( AC) single phase version of the three phase system normally used on commercial vessels and the same type used on the FV Nicola Jane. One pair of brass electrodes (12.75 mm diameter, 2.3 m length) spaced 1 m apart was placed in each stimulus tank (A and B, Figure 5). They were supported by pieces of slate to prevent them sinking into the sand and connected to an auto-transformer connected to a generator. The input power was estimated as 10 kW. The transformed output voltages could be varied using rotary selector switches and varied from 16.6 - 25.89 v on no load. The auto-transformer was fitted with an RCD (Residual Current Device) and a remote kill switch to control the output. A Dinse type cable connector (positive locking) was used to connect the transformer to the deck cables. These were connected to the underwater electrode pair using a male to female Dinse plug set. For the survival experiments the 220 - 240 v, 50 Hz input was reduced to 25 v. The current passing between the electrodes (mean = 42 A) was lower than those recorded on the boats (mean = 82 A) as a lower current was required to deliver 25 v to one pair of electrodes rather than two or three pairs.

Figure 5: Set up of stimulus tanks (A and B). Electric rig is submerged and lying on the sand. It is connected to a transformer (inset) and controlled by a switch. An electric field is generated between the electrodes that stimulate razor clams buried in the sand.

A rig was constructed with test points at fixed distances to establish the effective field strength or voltage gradient around the electrodes (Figure 6). One electrode was used as the reference and voltages were measured at the measuring points using an AC voltmeter. The output waveform was recorded on a number of occasions (for example see Figure A2). Three different configurations were tested in order to establish voltages around the electrodes. The selector switches on the generation system matched the settings as reportedly used in the fishery, with the settings selected giving the largest power output.

Figure 6: Schematic diagram and photographs of the test rig used to measure the properties of the electric field produced by the electrodes.

Configuration 1

The layout of electrodes in Configuration 1 was identical to that used by the commercial vessels involved in the project. The electrodes were placed on top of the sediment in the tank at 1 m separation. The reference electrode passed between vertical columns 7 and 8 of the measurement rig. In order to prevent the electrodes sinking into the sediment, small pieces of roofing slate were used below them to support the load. Additional wires were attached directly to the electrodes along their length in order to measure possible voltage drops that might occur due to the resistance of the electrode material. To ensure that no harmful contaminants might be released into adjoining tanks, and to evaluate if chlorine gas was being evolved at the electrodes, the water circulation system was disabled during the trials. Weighted upturned plastic boxes (200 cm ³ volume) were placed over a section of the electrodes to collect any gas released. Water in the collectors was tested at the end of the trial for chlorides using a Palin test (detects 0.1 - 3 mg/L Cl ^-, DPD01 tablets used).

Configuration 2

The distance between the electrodes was reduced to 70 cm with the reference electrode remaining in the same position between columns 7 and 8. This allowed a fuller voltage plot to be made to investigate symmetry. This was necessary due to the limited dimensions of the measurement rig used that could not span to the voltage mid-point in the commercial electrode configuration.

Configuration 3

It had been reported by divers in the fishery that razor clams also emerge from the sediment on the outside of the electrodes in some instances, i.e. not between electrodes. In order to investigate this phenomenon the reference electrode was moved to between columns 4 & 5 of the measurement rig. Electrode spacing was retained at approximately 70 cm. This configuration allowed the field shape to be more closely investigated outside the electrodes.

Ensis Survival and Behaviour

All razor clams used in both survival and stimulus trials were collected by divers who hand pulled the clams in Loch Nevis and the Clyde Sea area near the Isle of Cumbrae. A preliminary survival trial was conducted in Tank A in September 2013. 44 individuals of Ensis siliqua were labelled with plastic identifying tags adhered to the shell surface with superglue, weighed, measured and allowed to burrow into the sand. Initial burrowing times were recorded and survival was monitored over a three week period.

Stimulus trials were conducted using E. siliqua in February 2014. Six individuals were introduced to both tanks A and B. Initial burrowing behaviour was recorded using high definition GoPro cameras mounted above each tank. Any individuals that did not burrow within 12 hours were replaced. Following successful burial, the razor clams were left for 24 hours to acclimate to the tanks. In order to account for possible tank effects the experiment was repeated four times over four consecutive weeks. Stimulus and control tanks were alternated weekly. Razor clams in the stimulus tank (Tank A on weeks 1 and 3, Tank B on weeks 2 and 4) were exposed to an electric field for 2 minutes, based on the observed exposure of subjects in the commercial fishery (v = 25, A = mean 42). An average exposure time of 1 min was calculated from fishing activity conducted on the FV Nicola Jane, 2 min exceeded the maximum recorded exposure time, and was therefore considered appropriate to give a cautious estimate of exposure effects. Current (LEM current clamp, Model LH1040) and voltage (Tektronix multimeter, Model TDS 3034) outputs to the electrodes were recorded ( Table A3). Emergent and "kicking" behaviours during exposure were recorded. The razor clams were videoed for the duration of the stimulus and for one hour following the stimulus to record reburying behaviour.

Non-Target Species Survival and Behaviour

Individuals (n = 5) of the Atlantic surf clam Spisula solida, the sea star Asterias rubens, and the hermit crab Pagurus bernhardus were collected using non-electrical methods to examine the effects of electrofishing on non-target species. A. rubens and P. bernhardus were selected as these were the most frequently observed non-target species from the boat trial videos. S. solida were selected because they are known to emerge from the sediment on exposure to electricity, and there is the potential for fishing vessels to exploit electrofishing to harvest them (J. Grieve pers. comm.) A. rubens and P. bernhardus used in the trials were collected using Nephrops creels in Loch Leven. S. solida were dredged by the FV Ensis. Five stimulus and five control groups of five randomly assigned individuals of each species were used in six consecutive experimental runs to account for potential tank and run effects. Each group was individually transferred to the stimulus tank A. Stimulus groups were exposed to an electric field for two minutes and then transferred to holding tanks C-G to monitor survival over a five day period. Control groups were also transferred to the stimulus tank for two minutes and then transferred to holding tanks to eliminate handling effects. Current and voltage outputs to the electrodes were recorded for each run ( Table A3).

Statistical Analysis

ANOVAs (Analysis of Variance) were used to test the relationships between the length of the shells and emergence time, the time of the first movement following the end of the electrical stimulus and the reburial times of the razor clams. A linear model was used to test the effects of species identity and electrical treatment on survival in the non-target species. All statistical analysis was conducted in R.