Spatial Ecology and Fisheries Interactions of Rajidae in the Uk

Home , Arctic skate, Blonde ray, Common skate, Leucoraja naevus, Plymouth Breakwater, Smalleyed ray

UNIVERSITY OF SOUTHAMPTON

FACULTY OF NATURAL AND ENVIRONMENTAL SCIENCES

Ocean and Earth Sciences

SPATIAL ECOLOGY AND FISHERIES INTERACTIONS OF RAJIDAE IN THE UK

Samantha Jane Simpson

Thesis for the degree of DOCTOR OF PHILOSOPHY APRIL 2018

UNIVERSITY OF SOUTHAMPTON

ABSTRACT

FACULTY OF NATURAL AND ENVIRONMENTAL SCIENCES

Ocean and Earth Sciences

Doctor of Philosophy

FINE-SCALE SPATIAL ECOLOGY AND FISHERIES INTERACTIONS OF RAJIDAE IN UK WATERS

by Samantha Jane Simpson

The spatial occurrence of a species is a fundamental part of its ecology, playing a role in shaping the evolution of its life history, driving population level processes and species interactions. Within this spatial occurrence, species may show a tendency to occupy areas with particular abiotic or biotic factors, known as a habitat association. In addition some species have the capacity to select preferred habitat at a particular time and, when species are sympatric, resource partitioning can allow their coexistence and reduce competition among them.

The Rajidae (skate) are cryptic benthic mesopredators, which bury in the sediment for extended periods of time with some species inhabiting turbid coastal waters in higher latitudes. Consequently, identifying skate fine-scale spatial ecology is challenging and has lacked detailed study, despite them being commercially important species in the UK, as well as being at risk of population decline due to overfishing. This research aimed to examine the fine-scale spatial occurrence, habitat selection and resource partitioning among the four skates across a coastal area off Plymouth, UK, in the western English Channel. In addition, I investigated the interaction of Rajidae with commercial fisheries to determine if interactions between species were different and whether existing management measures are effective. First using a combination of research surveys, conventional and electronic tagging I investigate the fine-scale spatial ecology of four sympatric skates. Second I use stable isotope analysis of Rajidae eye lenses to provide an insight into juvenile feeding and spatial ecology. Finally this research used commercial landings data and conventional tagging to investigate fisheries interactions and current management efficacy.

Results show that Rajidae were not randomly distributed at fine-scales within the coastal zone but instead associated with particular locations and depths. In addition, only one of the four species inhabited both marine and brackish habitats. I present evidence demonstrating inter- and intra- species groups partitioned by both habitat and diet along with evidence for active habitat selection. Habitat partitioning between these four species influenced their interaction with commercial fisheries and the degree of protection offered by existing Marine Protected Areas (MPAs). I also demonstrate that legislation specifically designed to protect skates may not be effectively enforced and indicate where further investigation will be required to ensure that the conservation of skates is realised.

List of Contents Chapter 1. Fine-scale spatial ecology and fisheries interactions of Rajidae in UK waters ………………………………………………………………………………………………………………….19 1.1 Concepts in spatial ecology ...... 20 1.1.1 Spatial occurrence ...... 20 1.1.2 Habitat association ...... 21 1.1.3 Habitat selection and habitat preference...... 22 1.1.4 Habitat partitioning ...... 24 1.1.5 Dietary partitioning...... 24 1.1.6 Temporal resource partitioning ...... 25 1.1.7 Sexual segregation ...... 26 1.1.8 Resource partitioning among age groups ...... 27 1.2 Study species ...... 28 1.2.1 Rajidae resource use ...... 28 1.2.2 Fishery history and management of Rajidae ...... 32 1.3 Aims and objectives ...... 33 Chapter 2. Spatial and temporal occurrence of four sympatric Rajidae species in coastal waters of the Western English Channel: are there fine-scale habitat differences? ...... 35 2.1 Introduction ...... 35 2.2 Methods ...... 39 2.2.1 Sampling ...... 39 2.2.2 Comparing occurrence of species within and among sites ...... 42 2.2.3 Depth of capture ...... 42 2.2.4 Seasonality analysis ...... 43 2.2.5 Sex ratio analysis...... 43 2.3 Results ...... 43 2.3.1 Comparing occurrence of species within and among sites ...... 46 2.3.2 Depth of capture ...... 50 2.3.3 Seasonality analysis ...... 50 2.3.4 Sex ratio analysis...... 54 2.4 Discussion ...... 55 2.4.1 Rajidae habitat associations ...... 55 2.4.2 Temporal habitat association ...... 57 2.4.3 Ontogenetic shifts in habitat associations ...... 58

2.4.4 Sexual segregation in Rajidae ...... 58 2.4.5 Management implications ...... 59 2.4.6 Conclusion ...... 59 Chapter 3. The spatial ecology of Rajidae from mark-recapture tagging and its implications for assessing fishery interactions and efficacy of Marine Protected Areas .... 61 3.1 Introduction ...... 61 3.1.1 The UK Rajidae fishery ...... 62 3.1.2 Monitoring and management ...... 62 3.1.3 Rajidae ecology ...... 63 3.2 Method ...... 65 3.2.1 Ethics statement ...... 65 3.2.2 Analysis ...... 65 3.3 Results ...... 66 3.3.1 Rajidae spatial ecology ...... 68 3.3.2 Seasonality of recaptures ...... 74 3.3.3 Method of recapture...... 75 3.3.4 Recaptures in MPAs ...... 76 3.4 Discussion ...... 77 3.4.1 Rajidae spatial ecology ...... 78 3.4.2 Seasonality of recapture ...... 80 3.4.3 Method of recapture...... 80 3.4.4 Efficacy of existing MPAs for Rajidae ...... 81 3.4.5 Conclusion ...... 82 Chapter 4. Investigating fine-scale habitat selection, spatial partitioning and sexual segregation in Rajidae using passive acoustic telemetry ...... 83 4.1 Introduction ...... 83 4.2 Method ...... 85 4.2.1 Acoustic networks ...... 89 4.2.2 Spatial overlap analysis ...... 90 4.3 Results ...... 90 4.3.1 Species detections ...... 91 4.3.2 Spatial overlap...... 94 4.3.3 Differences in the detection of sexes ...... 98 4.4 Discussion ...... 103

4.4.1 Habitat selection in Rajidae ...... 103 4.4.2 Resource partitioning in Rajidae ...... 104 4.4.3 Differences in movement among species ...... 106 4.4.4 Sexual segregation in Rajidae ...... 106 4.4.5 Conclusion ...... 108 Chapter 5. Diel vertical migration and central place foraging in benthic predators .... 109 5.1 Introduction ...... 109 5.2 Methods ...... 111 5.2.1 Tags and tagging ...... 111 5.2.2 Data analysis ...... 112 5.3 Results ...... 115 5.3.1 Summary ...... 115 5.3.2 Movement analysis ...... 117 5.3.3 Temperature as a possible driver ...... 119 5.3.4 Depth ranges...... 119 5.3.5 Timing of events ...... 120 5.3.6 Seasonality of events ...... 121 5.4 Discussion ...... 123 5.4.1 Benthic rather than pelagic movement ...... 123 5.4.2 Nektobenthic DVM as possible central place foraging ...... 124 5.4.3 Potential drivers of nektobenthic DVM ...... 125 5.4.4 Conclusions ...... 128 Chapter 6. Resource partitioning and ontogenetic shifts in the trophic geography of sympatric skates (Rajidae) inferred from eye lens isotopic signatures ...... 131 6.1 Introduction ...... 131 6.2 Methods ...... 135 6.2.1 Statistical analysis ...... 137 6.3 Results ...... 137 6.3.1 Allometry and the use of Rajidae eye lens tissues in stable isotope analysis ...... 137 6.3.2 Isotopic differences among species at throughout ontogeny ...... 138 6.3.3 Standard ellipse area (SEAc) analysis ...... 140 6.3.4 δ15N and δ13C and eye lens diameter ...... 143 6.3.5 Magnitude of δ15N difference in eye lens tissue ...... 148

6.4 Discussion ...... 149 6.4.1 Comparative trophic geography and resource partitioning ...... 149 6.4.2 Maternal egg provisioning ...... 150 6.4.3 Trophic ecology across ontogeny...... 151 6.4.4 Spatial ecology across ontogeny ...... 151 6.4.5 Conclusions and implications ...... 152 Chapter 7. Are critically endangered fish back on the menu? Analysis of U.K. fisheries data suggest post-ban landings of prohibited skates in European waters ...... 153 7.1 Introduction ...... 154 7.2 Method ...... 156 7.3 Results ...... 157 7.3.1 Prohibited species ...... 157 7.3.2 Other species ...... 161 7.3.3 Skates and rays...... 161 7.4 Discussion ...... 163 7.4.1 Are prohibited, endangered skates being landed into UK ports from fishing areas where bans are in place? ...... 163 7.4.2 Are fishers misidentifying or misreporting prohibited skate species? ...... 165 7.4.3 Are systematic errors in data input being made? ...... 167 7.4.4 Concluding Remarks ...... 167 Chapter 8. General discussion ...... 171 8.1 The spatial ecology of Rajidae ...... 171 8.1.1 Spatial occurrence ...... 171 8.1.2 Habitat associations ...... 171 8.1.3 Habitat selection and preference ...... 174 8.1.4 Resource partitioning ...... 175 8.1.5 Temporal resource partitioning ...... 175 8.1.6 Interaction with the commercial fishery ...... 177 8.2 Wider implications ...... 178 8.3 Overall limitations ...... 180 8.4 Future directions ...... 181 8.4.1 Acoustic array...... 181 8.4.2 Accelerometer / magnetometer ...... 182 8.4.3 Maternal offspring biochemical transfer and eye lens allometry...... 183

8.4.4 Investigating landings ...... 184 8.4.5 Fisheries ...... 184 Additional information ...... 185 References 221

List of Figures and Tables

Figure 1: The spatial occurrence of Bettongia tropica (Burbidge and Woinarski, 2016a) ...... 21 Figure 2: The spatial occurrence of the blue whale Balaenoptera musculus (in red) (Reilly et al., 2008)...... 21 Figure 3: Rajidae spatial occurrence A= Raja brachyura B=Raja clavata C=Raja microocellata D=Raja montagui (Ellis, 2016, Ellis, 2006, Ellis et al., 2009, Ellis et al., 2007)...... 30 Figure 4: EUNIS substrate types off the coast of Plymouth, UK (European Environment Agency, 2018)...... 40 Figure 5: Sampling sites off Plymouth Sound...... 40 Figure 6: Capture locations for four species 2000-2016. Hollow circles represent empty hauls, colours indicate presence. A=R. brachyura, B=R. clavata, C=R. microocellata, D=R. montagui ...... 45 Figure 7: CPUE at each of the five sites during the study period 2000-2016. Note the magnitude of difference in CPUE between Bigbury Outer and Bigbury Inner which are adjacent sites (their centre points are only 4.831km apart)...... 46 Figure 8: Within site difference in species counts. A) All sites combined B) Bigbury Inner) C) Standard haul D) Whitsand E) River F) Bigbury Outer ...... 47 Figure 9: Between site differences in counts. a) R. brachyura b) R. clavata c) R. microocellata d) R. montagui ...... 49 Figure 10: Difference in depth of capture between four species and age classes. The box represents 25th and 75th percentiles, whiskers, 5th and 95th percentiles, horizontal line is the median ...... 50 Figure 11: % presence/absence across the year. A) R. brachyura at Bigbury Outer B) R. clavata at Standard haul C) R. microocellata at Whitsand D) R. montagui at Bigbury Outer. Colour represent seasons, blue= winter, yellow= spring, orange= summer, red= autumn...... 51 Figure 12: Seasonal difference in captures by site and species. a) R. brachyura in Bigbury Outer b) R. montagui at Bigbury Outer c) R. clavata in the River d) R. microocellata at Whitsand e) R. clavata at Whitsand ...... 53 Figure 13: Sex ratio in the catches for four Rajidae species. There is a significant bias towards females in R. brachyura, R. clavata and R. microocellata...... 54 Figure 14: Marine protected areas within the study area. 1 The Lizard, 2 Plymouth Sound and Estuaries, 3 Fal and Helford, 4 South Devon Shore Dock, 5 Blackstone Point, 6 Polruan to Polperro, 7 Lyme Bay and Torbay, 8 Lizard Point, 9 Start Point to Plymouth Sound and Eddystone. Red area is the designated Whitsand and Looe Bay MCZ...... 63 Figure 15: Left- total number recaptured during the study period. Expected numbers are based on the total number released. There was no significant difference between species in the number released. Right- the rate of recapture, where R. brachyura are recaptured more quickly than the other species. Note the increase in recaptures of R. montagui after 281 days...... 67

Figure 16: Kernel density estimation of R. brachyura, Top= Release, Bottom=Recapture. Hotspots of recapture were in the outer section Bigbury Bay...... 68 Figure 17: Kernel density estimation of R. clavata, Top= Release, Bottom=Recapture. Hotspots of recapture were in the estuary, in the Cawsands Bay area...... 69 Figure 18: Kernel density estimation of R. microocellata, Top= Release, Bottom=Recapture. Note the wider dispersal of recapture hotspots in this species. Recaptures were primarily in the Whitsand Bay area...... 69 Figure 19: Kernel density estimation of R. montagui, Top= Release, Bottom=Recapture. Hotspots of recapture were in the Bigbury Bay area...... 70 Figure 20: Differences in depth (m) of capture between species. Box represents 25th and 75th percentile, whiskers 5th 9th percentiles and middle horizontal line is the median. R. clavata and R. microocellata were caught at significantly shallower depths than R. brachyura and R. montagui (P<0.001)...... 71 Figure 21: Proportion of recaptures at <10km, 10-30km and >30km from release. R. clavata were predominantly caught within 10km of where they were released, while in R. microocellata a higher proportion were caught greater displacements...... 72 Figure 22: Displacement as the crow flies. Green=release, red=recapture. A- R. brachyura, B- R. montagui, C- R. clavata, D- R. microocellata. Note the difference between C and D, where R. clavata (C) dispersed less than R. microocellata (D)...... 73 Figure 23: Seasons of recaptures. A. R. brachyura B. R. clavata C. R. microocellata D. R. montagui ...... 75 Figure 24: Proportion of species captured by fishing gear type. Note the high proportion of R. clavata caught in nets relative to R. brachyura which were more frequently caught by trawl...... 75 Figure 25: Rajidae recapture locations. MPA boundaries marked in blue...... 76 Figure 26: Potential new MPA (yellow) containing 28% of all recaptured Rajidae species...... 77 Figure 27: Relative fishing effort by trawls (sightings per unit effort). Data and method found at Vanstaen and Breen 2014...... 78 Figure 28: The line is the animals track, the red dots are known locations. How much of an individual’s movement can we sample using three different techniques? Fishing survey data can only provide a single known location and we cannot infer active selection, while a passive acoustic array can detect movement at a finer scale and therefore active movement towards particular areas...... 83 Figure 29: One of 12 receiver deployments on large tripod landers off the coast of Plymouth. The data-logging acoustic receiver is arrowed...... 86 Figure 30: Location of the Whitsand and Plymouth arrays off the south west coast, Plymouth, UK. Bottom left highlights the Whitsand array in more detail...... 87 Figure 31: Range testing in Whitsand array. Red circles are the receiver positions and arrows show the distance (m) of detection ranges...... 87 Figure 32: Tag detections over time from 2010 to 2017...... 90 Figure 33: Observed networks left, Example randomised networks on the right A= R. brachyura B= R. clavata C= R. microocellata D= R. montagui. Bottom left box highlights the Whitsand array in more detail. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections...... 92 Figure 34: Proportion of detections or connections by species at the Plymouth and Whitsand array and connections between the 2 arrays...... 93

Figure 35: Observed and average random detections. Proportions of detections for each species at the 12 different receivers. Brighter colours are the Whitsand array, darker colours indicate the Plymouth array...... 94 Figure 36: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. brachyura Right= R. montagui vs R. clavata...... 96 Figure 37: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. clavata Right= R. clavata vs R. brachyura...... 96 Figure 38: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. montagui Right= R. brachyura vs R. montagui...... 97 Figure 39: Acoustic networks comparing sexes. A= R. brachyura B= R. clavata. Left = female, Right = male. Bottom left box highlights the Whitsand array in more detail where required. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections...... 99 Figure 40: Acoustic networks comparing sexes. A= R. microocellata B= R. montagui. Left = female, Right = male. Bottom left box highlights the Whitsand array in more detail where required. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections...... 100 Figure 41: Proportion of detections by sex for each species, normalised by total number of days with at least one ping to account for tagging bias...... 101 Figure 42: Spatial overlap between sexes over time. Red lines indicate the timing and extent of overlap between species. Left= R. brachyura, Right= R. clavata...... 102 Figure 43: Spatial overlap between sexes over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata Right= R. montagui...... 102 Figure 44: Example DVM plots ...... 113 Figure 45: Frequency of events per individual ...... 116 Figure 46: Event movement analysis ...... 118 Figure 47: Plateau vs deep temperatures and depths ...... 119 Figure 48: Frequency of events by start depth for species and sex...... 120 Figure 49: Timing of events in relation to sunrise and sunset times ...... 121 Figure 50: Observed – Expected event counts by species, sex and season ...... 122 Figure 51: Study location off Plymouth, south-west UK. Filled circles represent the results from a tagging study (green= release, red=recapture) defining the spatial extent of the study area...... 136 Figure 52: The linear relationship between the total length (mm) of an individual on recapture and their maximum eye lens diameter (mm). Standard error is represented by the grey areas. Each data point plotted denotes an individual skate...... 138 Figure 53: Differences in δ15N values among species at 3 ontogenetic stages. The box represents the 25th and 75th percentile, the whiskers are the 5thand 95th percentiles and the horizontal line is the median. No difference was found in the adult phase, but differences occurred in both the juvenile and pre-hatch phase...... 139 Figure 54: Difference in δ13C among species at 3 ontogenetic stages. The box represents the 25th and 75th percentile, the whiskers are the 5th and 95th percentiles and the horizontal line is the median. No significant differences were found in the pre-hatch phase, but differences were found in the adult and juvenile phases...... 139

Figure 55: Standard ellipse areas and convex hulls by species. A= Adults/muscle tissue, B=Juvenile/eye lens tissue, C=Pre-hatching/eye lens core ...... 141 Figure 56: δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases. Note the similar pattern across all species, where δ15N reaches a low point shortly after hatching...... 144 Figure 57: δ13C values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases...... 145 Figure 58: Individual level δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases...... 146 Figure 59: Individual level δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases...... 147 15 15 Figure 60: The magnitude of change in δ N in the eye lens tissues (i.e.  Nmax-min). The box represents the 25th and 75th percentile, the whiskers are the 5th and 95th percentiles and the horizontal line is the median. Note the higher magnitude of difference in R. microocellata compared to R. montagui...... 148 Figure 61 Reported landings (tonnes) of common skate by ICES area, 2011-2014, in UK MMO data...... 158 Figure 62: Quantity (tonnes) of common skate reported landed by port, 2011-2014. Only the highest landing ports (>0.085t) are named...... 159 Figure 63: The relative quantity (tonnes) of common skate landed by ICES area 2014...... 159 Figure 64: Reported landings (tonnes) of white skate by ICES area 2011-2014. Hatched areas are not prohibited from landing white skate...... 160 Figure 65: Reported landings (tonnes) of white skate landed by port, 2011-2014. Hatched areas are not prohibited from landing white skate. Only the highest landing ports are named (>0.233)...... 160 Figure 66: Skates and rays (tonnes) landed in the UK ICES areas 2011-2014. Landings are from UK and foreign vessels landing into the UK and UK vessels landing abroad...... 162 Figure 67: Quantity (tonnes) of 'skate and ray' landed by port 2011-2014. Only ports with landings >5t are displayed and only the highest landing ports (>10t) are named...... 162 Figure 68: Multiple R. brachyura spatial datasets. The blue area represents the average depth range (plus standard error) (Humphries et al., 2016c), red circles represent locations of recapture (dispersed in ARC GIS to visualise multiple point in the same location) (Chapter 3), triangles represent research survey data (Chapter 2), pentagons and lines represent acoustic networks (Chapter 4). Note the high overlap of all data in the Bigbury Bay area...... 173 Figure 69: The Bigbury Bay area. Red lines are boundaries of MPAs, triangles represent R. brachyura in Chapter 2. Black lines represent 2 meter bathymetry contour lines. DEFRA bathymetry highlights the complex bathymetry inside the MPAs (designated to protect reef features) and the relatively simple bathymetry between the two MPA boundaries which may indicate a sandy habitat...... 174 Figure 70: Multiple R. clavata spatial datasets. The blue area represents the average depth range (plus standard error) (Humphries et al., 2016c), red circles represent locations of recapture (dispersed in ARC GIS to visualise multiple point in the same location) (Chapter 3), triangles represent research survey data (Chapter 2), pentagons and lines represent acoustic networks (Chapter 4). Note the wider distribution of data from the different data sets relative to R. brachyura...... 176

Table 1: Rajidae depth ranges and occurrence (Wheeler, 1969b, Whitehead et al., 1984, Ellis et al., 2005a) ...... 37 Table 2: Details of the sampling sites. Average depth(m) of the site, the maximum depth of the site(m), the number of hauls conducted and the time spent trawling (hours)...... 41 Table 3: Rajidae length at maturity ...... 43 Table 4: Within site Chi Square Goodness of Fit test results ...... 48 Table 5: Between site differences Chi Square Goodness of Fit test results...... 48 Table 6: Overall % presence of all Rajidae species in hauls/sites ...... 51 Table 7: Seasonal differences in catches, Chi Squared Goodness of Fit results ...... 52 Table 8: Sex ratio in catches, Chi Squared Goodness of Fit results ...... 54 Table 9: Quantifying the sampling effort and return rates ...... 67 Table 10: Displacement metrics for each species ...... 74 Table 11: Chi-Squared results of differences in season of recapture ...... 74 Table 12: Proportion of a species recaptured within the boundary of a designated MPA ...... 76 Table 13: Of those recaptured in an MPA, the proportion of different gears types ...... 76 Table 14: Receiver array depths and names ...... 86 Table 15: Number of acoustic tags released by species and sex ...... 88 Table 16: Observed versus randomised overlap coefficient. ↓= Overlap coefficient was lower than expected by chance, → not significantly different than expected by chance ↑ higher than expected by chance...... 95 Table 17: Frequency of DVM events per day tracked by species ...... 116 Table 18: Χ 2 results from the frequency by sex analysis ...... 116 Table 19: Summarised statistical results ...... 117 Table 20: Χ 2 results from the seasonal analysis, significant results highlighted ...... 122 Table 21: Summary of sample sizes and median capture lengths for each species ...... 135 Table 22: Statistical differences in δ15N and δ13C values among species at 3 ontogenetic stages 140 Table 23: The difference and the overlap in standard ellipse area (SEAc) for each stage of development...... 142 15 Table 24: Average  Nmax-min for each species ...... 148

List of accompanying material

Appendix 1: Chapter 2 supplementary information

S Table 1: Gear types and trawl detail for each site

S Table 2: Research vessel details

Appendix 2: Chapter 4 supplementary information

S Table 3: Statistical results from Wilcoxon signed ranked pairs test, comparing observed to random node metrics.

S Table 4: Statistical results from Wilcoxon signed ranked pairs test, comparing observed to random edge metrics.

Appendix 3: Chapter 5 publication

Figure S1: Observed event deltas, ascent deltas and event durations for all events initially extracted. The red lines indicate the cut off for events during filtering.

Figure S2: No significant correlation between individual length and the number of events per day tracked.

Humphries NE, Simpson SJ, Sims DW. Diel vertical migration and central place foraging in benthic predators. Marine Ecology Progress Series. 2017;582:163-80.

Appendix 4: Chapter 6 supplementary information

Overall difference in δ15N and δ13C in eye lens tissue between species

S Table 5: Statistical results-differences in overall eye lens δ15N values between species

S Table 6: Statistical results-differences in overall eye lens δ13C values between species

Appendix 5: Chapter 7 publications

Sims DW, Simpson SJ. Fisheries: Better policing for fishy catch data. Nature. 015;520(7549):623.

Simpson SJ, Sims DW. Are critically endangered fish back on the menu? Analysis of UK fisheries data suggest post-ban landings of prohibited skates in European waters. Marine Policy. 2016;69:42-51.

Academic Thesis: Declaration Of Authorship

I, SAMANTHA JANE SIMPSON declare that this thesis and the work presented in it are my own and has been generated by me as the result of my own original research.

I confirm that:

1. This work was done wholly or mainly while in candidature for a research degree at this University;

2. Where any part of this thesis has previously been submitted for a degree or any other qualification at this University or any other institution, this has been clearly stated;

3. Where I have consulted the published work of others, this is always clearly attributed;

4. Where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work;

5. I have acknowledged all main sources of help;

6. Where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself (overleaf);

7. Either none of this work has been published before submission, or parts of this work have been published as: [please list references below]:

Sims DW, Simpson SJ. Fisheries: Better policing for fishy catch data. Nature. 2015;520(7549):623.

Simpson SJ, Sims DW. Are critically endangered fish back on the menu? Analysis of UK fisheries data suggest post-ban landings of prohibited skates in European waters. Marine Policy. 2016;69:42-51.

Humphries NE, Simpson SJ, Sims DW. Diel vertical migration and central place foraging in benthic predators. Marine Ecology Progress Series. 2017;582:163-80.

Humphries NE, Simpson SJ, Wearmouth VJ, Sims DW. Two’s company, three’s a crowd: fine-scale habitat partitioning by depth among sympatric species of marine mesopredator. Marine Ecology Progress Series. 2016;561:173-87.

Signed: S.Simpson………………………………………………………………………

Date: 30/03/2018…………………………………………………………………………

Declaration of authorship continued: outline of work conducted jointly with or by others presented in this thesis (information pertaining to point 6 on previous page). Chapter 2

The long term time series data were collected by the MBA. S. Simpson assisted on research surveys throughout the PhD project.

Chapter 3 and 4

Tagging was started by the MBA prior to the beginning and continued through the PhD project, a summary of which follows. These numbers are an overall extract from the tagging databases and include tags that may have failed and were unused in the analysis.

Tag Tagged by Tagged Total type S. Simpson by the MBA Acoustic 174 27 201 DST 77 179 256 Total 251 206 457

All acoustic receivers were deployed by the MBA. S. Simpson aided in the deployment of six receivers.

Chapter 5

All tags mentioned in this chapter were deployed by the MBA. S. Simpson and N. Humphries contributed equally to the work. S. Simpson and N. Humphries designed the study, performed the analysis and wrote the paper. This chapter was published as:

Humphries NE, Simpson SJ, Sims DW. Diel vertical migration and central place foraging in benthic predators. Marine Ecology Progress Series. 2017;582:163-80.

Chapter 7

Data were provided by the MMO statistics department. This chapter was published as:

Sims DW, Simpson SJ. Fisheries: Better policing for fishy catch data. Nature. 015;520(7549):623.

Simpson SJ, Sims DW. Are critically endangered fish back on the menu? Analysis of UK fisheries data suggest post-ban landings of prohibited skates in European waters. Marine Policy. 2016;69:42-51.

Acknowledgments

First I would like to thank NERC and the Marine Biological Association for funding this PhD and giving me the opportunity to pursue this research topic. I greatly appreciate being part of the SPITFIRE DTP and value the training opportunities provided for me. I also received a research grant from the Fisheries society of the British Isles which enabled me to complete the stable isotope work and a travel grant from PLYMSEF to present my research at a FSBI conference.

I would like to thank my panel chair Professor Stephen Hawkins, my supervisors Dr. Clive Trueman and Professor David Sims who have provided advice and feedback, without which I could not have completed my thesis. Clive’s expertise provided an entirely different perspective to the thesis and has broadened my knowledge greatly. I thank him for guiding me through this new method. I thank David Sims for sharing his expertise and experience with me. I would especially like to thank him for allowing me to independently explore the research area, whilst guiding and advising me in the right direction.

I cannot thank Nick Humphries enough for his assistance throughout the PhD. The fantastic software that he provided has allowed me to explore my data efficiently and I thank him for the training, advice and many discussions that have led to publications. The feedback Nick provided has been incredibly helpful and he has been a joy to work with.

I thank both Dr. Victoria Wearmouth for all the tagging that was done before I arrived and Dr. Stephen Cotterell for the training and the deployment of the receiver array. I also thank the crew of the MBA Sepia, Alix Harvey, Bastian Hambach, Megan Spencer and fishers who returned tags, without whom this would not have been possible.

I feel privileged to have worked at the MBA over the past few years and I am very proud to have been part of this outstanding institute. While working at the MBA I have been incredibly lucky to have the support and friendship of my fellow PhD students, Charlotte Walker, Harry Teagle and Harriet Dale, as well as Costanza Zanghi who I have known since our masters degree. My greatest thanks for listening to me constantly chatting with excitement about my PhD, as well as at the low times.

My family have always been my rock and whatever situation I am in. I can always channel their wisdom. I am so incredibly lucky to have such a strong foundation, how could I fail to sagasiate my PhD! Thank you for letting me talk on and on about fish, statistics and boring you to tears, I know that you understand my passion!

Finally I would like to thank Adam, who has borne the brunt of this PhD ride! You have been so incredibly patient and understanding throughout this experience and your faith in me has been unwavering.

Abbreviations

MMO Marine Management Organisation IFCA Inshore Fisheries Conservation Agency EUNIS European Nature Information System WW1 World War 1 WW2 World War 2 IUCN International Union for Conservation of Nature MPA Marine Protected Area ICES International Council for the Exploration of the Sea IBTS International Bottom Trawl Survey CPUE Catch per Unit Effort DEFRA Department for Environment, Food and Rural Affairs ANOVA Analysis of variance ANCOVA Analysis of covariance MCZ Marine Conservation Zone WEC Western English Channel AWERB Animal Welfare and Ethical Review Body FAO Food and Agriculture Organization DST Data Storage Tag MBA Marine Biological Association SAC Special Area of Conservation SI Straightness Index DVM Diel Vertical Migration rDVM reverse Diel Vertical Migration nDVM normal Diel Vertical Migration DBM Diel Bank Migration GAMM General Additive Mixed Model wAIC weighted Akaike Information Criterion SEAc Standard Ellipse Area MAFF Ministry of Agriculture, Fisheries and Food TAC Total Allowable Catch

Chapter 1. Fine-scale spatial ecology and fisheries interactions of Rajidae in UK waters

Movement is defined as a change in the spatial location over time by a whole organism that is driven by processes of different spatial and temporal scales (Nathan et al., 2008). The ability of an organism to move allows an individual to change the characteristics of the surroundings they are experiencing, for example, resource availability including mates, food or predation risk or environmental factors such as temperature (Moorter et al., 2013). These different factors can drive different patterns of movement at different times (Humphries et al., 2010). For instance, movement rates could decrease or become more tortuous when animals forage within areas with plentiful resources, while they may make longer meandering movements when searching for food (Pyke et al., 1977, Klaassen et al., 2006). The study of movement therefore requires the observation of an organism through space and time. Herein lies the problem, how to record complex dynamic movement patterns over the life-time of an animal, yet at a temporal and spatial resolution that will capture fine-scale movement that is important to the ecology of the animal. This is particularly challenging in the marine environment where species are often not directly observable, are well camouflaged and rarely breach the surface (Hussey et al., 2015, Cooke et al., 2013a). Therefore, to understand movement, the behaviour of animals can be sampled in different spatial and temporal scales to explore their spatial ecology and infer movement patterns.

Understanding the extent of a species movement patterns can provide information relevant to the protection and management of vulnerable species. For instance, it is difficult to manage effectively and plan spatial protection for species when their spatial extent or which habitats critical to their survival are unknown. Furthermore, identifying and designating boundaries of spatial protection is complicated by demands from other stakeholders that use the space, such as fishers in the marine environment or land developers in terrestrial systems (Mangi and Austen, 2008, Hein et al., 2006). Without accurate evidence of the efficacy of the spatial protection, spatial plans become difficult to implement or defend (Ward et al., 1999). It seems likely that any management that recognises species differences in spatial ecology may have a greater chance of conserving vulnerable species. This is especially important for species that broadly occupy the same area, but differing spatial ecology may result in different habitats occupied and, therefore, varying management strategies for each species in the assemblage may be needed (Speed et al., 2010).

1.1 Concepts in spatial ecology

1.1.1 Spatial occurrence At the broadest scale, the spatial occurrence of a species simply refers to where an animal is found in space and time, while spatial distribution refers to where a species could potentially be observed (Dingle, 2014, Kramer et al., 1997). It is well established that animals occur in a non- uniform pattern across the globe with some species exhibiting near global distributions and others only occurring in a small localised range (Sims, 2003, Kramer et al., 1997, Morris, 2003, Wallace, 2011). For example, the northern bettong Bettongia tropica is a small marsupial and an ideal species to introduce general concepts in spatial ecology. The bettong is well-studied and only endemic to small areas of north-eastern Australia, Queensland (Figure 1). Its entire range is restricted to three known locations, with an estimated extent of occurrence of 6023 km2 (Bateman et al., 2012, Burbidge and Woinarski, 2016b). Despite the bettong’s solitary and nocturnal nature, direct observation has enabled a greater understanding of the species spatial occurrence in addition to electronic tracking (Vernes and Pope, 2001, Bateman et al., 2012). Different environments however can make establishing the spatial occurrence of an organism more problematic.

As mentioned previously, the ability to directly observe species in the marine environment is especially challenging (Hussey et al., 2015, Cooke et al., 2013a). This often results in gaps in our knowledge. The blue whale Balaenoptera musculus for example is the largest animal known to have existed, is broadly found in all oceans except the Arctic and is absent from some seas such as the Mediterranean, Okhotsk and Bering seas (Figure 2) (Reilly et al., 2008). Yet our knowledge of the full extent of the blue whales spatial occurrence is restricted by the ability to record movements over time in the marine environment (Oleson et al., 2007). In both the bettong and the whale, the spatial occurrence of the species is restricted to particular ranges, though on entirely different scales. There are areas of presence and areas of absence, which raises the question as to why these species are only found in these ranges. What are the factors that link a species to a particular space in time and why are there differences between species?

Figure 1: The spatial occurrence of Bettongia tropica (Burbidge and Woinarski, 2016a)

Figure 2: The spatial occurrence of the blue whale Balaenoptera musculus (in red) (Reilly et al., 2008).

1.1.2 Habitat association The spatial occurrence of a species is usually driven by the adaptation of an animal to a particular set of habitat characteristics and the exploitation of a particular niche. Habitat can be defined as the sum of the resources that are needed by a particular organism to survive (Krausman, 1999). Species may therefore show a tendency to occupy areas with particular abiotic or biotic factors, such as temperature, vegetation type or prey availability (Schlaff et al., 2014). This link between spatial occurrence and environmental factors is known as a habitat association (Kramer et al., 1997).

In the case of the northern bettong B. tropica spatial occurrence is associated with tropical wet forest dominated by eucalyptus and their relatives, where trees are usually taller than 30 m and >1000 mm of rainfall per year. In addition, the bettong is associated with ridges, sparse ground cover, low damage by pigs, low densities of some grasses and high density tree stems (Vernes, 2003). This can be further linked to a specialist mycophagous diet of ‘truffles’ (Johnson and McIlwee, 1997), which in turn are influenced by both rainfall and temperature (Johnson, 1994). These habitat characteristics can often be directly observed and mapped in a terrestrial environment.

In the marine environment, it is difficult directly to observe and quantify habitat types. The marine environment is dynamic and influenced by events from storms to climate change and fishing activity, in addition to the high cost of habitat mapping relative to terrestrial environments (Larkum and West, 1990, Hoegh-Guldberg and Bruno, 2010, McConnaughey et al., 2000, Michelot et al., 2017, Stevens and Connolly, 2004). For many marine species however, habitat associations have been reported. At a Caribbean reef for example, Ginglymostoma cirratum were found in the deep and shallow lagoons rather than other habitats such as ocean reef habitats. The lagoons are thought to be used for mating and nursery grounds in the species and therefore confer reproductive benefits for this species (Pikitch et al., 2005).

These habitat associations however only tell us where there are connections between occurrence and habitat, but not whether an individual is actively making a decision to move towards a habitat (Kramer et al., 1997). Are individuals choosing to be within a particular habitat or are they simply occupying a space through random chance? Zooplankton for instance, may actively swim vertically to shallower habitats at night and deeper in the day (Hays et al., 2001), whereas their horizontal movements are not active choices but represent the movement of currents on which they are carried (Alldredge and Hamner, 1980). Nevertheless, there is evidence that zooplankton move vertically into water layers that can be moving horizontally at different speeds relative to each other, leading to retention or dispersal dependent upon the vertical extent of movements (Queiroga and Blanton, 2005). This then does provide a potential means for zooplankton to choose habitats vertically that then have implications for horizontal dispersal and recruitment.

1.1.3 Habitat selection and habitat preference Mobile species have the capacity to shift their spatial occurrence and habitat associations within an individual’s lifetime and make decisions to move towards habitat that is optimal at a particular time (Krausman, 1999, Hutto, 1985). This process, in which an animal makes behavioural decisions about which habitat to use is known as habitat selection (Kramer et al., 1997). Habitat

22 preference meanwhile is defined as the consequence of habitat selection, which results in the disproportional use of one habitat over another (Hall et al., 1997, Johnson, 1980).

It has been possible to demonstrate that northern bettong B. tropica exhibit habitat selection and habitat preference because individuals have been tracked through space and time (with radio- transmitters) in relation to their habitat. Movement rates in the bettong were higher at night indicating that greater distances were covered within home ranges during nightly foraging, before returning to one of several nests within close proximity of each other (Vernes and Pope, 2001). The movements recorded here clearly demonstrated a choice made by the bettong to occupy its preferred habitat.

Tracking the movement of species through space and time in order to demonstrate habitat selection is less well understood in the marine environment. Constraints result from difficulties in monitoring movements for animals that spend their entire lives submerged and consequently limit the use of technology that can otherwise be deployed in terrestrial systems. The radio transmitters used to track bettongs for example were detected using a two-element antenna and telemetry receiver on foot (Vernes and Pope, 2001). Radio waves however do not propagate sufficently through salt water eliminating their use in the marine environment (Hussey et al., 2015). Historically the spatial occurrence of where and when aquatic organisms could be caught (e.g. in commercial fisheries) were the only available measure of a species movement patterns (Hussey et al., 2015). This knowledge has provided low resolution movement patterns, but do not necessarily indicate whether an individual is actively moving towards a particular habitat (Kramer et al., 1997). However the development of various electronic tags has enabled some insight into habitat selection. For example lemon sharks Negaprion brevirostris tracked with ultrasonic transmitters for 1-153 days were found to show habitat preference for shallow (0-50cm), warm (≥30˚C) waters with a rocky or sandy substrate. It was hypothesised that these sharks selected warmer waters to maintain optimal metabolic performance and selected shallower areas to avoid predators (Morrissey and Gruber, 1993).

The study of habitat selection allows for a greater understanding of the way individual species use their environment and make decisions that optimise their survival. Individual species however rarely occupy a space alone, with multiple species exhibiting overlapping spatial occurrence. Consequently this leads to competition among species for resources. How then might multiple species coexist?

1.1.4 Habitat partitioning The habitat preferences made by a single species are likely to have an impact on other species that are sympatric or broadly occupy the same space (Ross, 1986). Each sympatric species is either predator, prey, a symbiont or a competitor that will affect the movements of other species (Creel and Creel, 2002, Heithaus et al., 2002). To allow these species to coexist, they may use resources, such as habitat differently (Toft, 1985). This resource partitioning is a driver behind the natural control of species diversity (Schoener 1974). The ability to detect habitat partitioning on a micro-scale allows for a greater understanding of fundamental questions in ecology, about how species interact with their environment. Ultimately the differential space use by multiple sympatric species can reduce the competition between species in a relatively small area (Speed et al., 2011, Schoener, 1983, Ross, 1986, Platell et al., 1998).

It has been established that the northern bettong B. tropica has habitat preferences within their range, however another similarly sized mammal, with a similar diet, the northern brown bandicoot Isodon macrourus, overlaps its range. These sympatric species have been shown to use different fine-scale habitat within the space. While the bettong preferred sparse ground cover, low damage by pigs, low densities of some grasses and high density tree stems, the bandicoot preferred dense ground cover, fewer tree stems and greater degrees of pig damage to the habitat (Vernes, 2003). Thus competition between the two species was reduced.

In the marine environment batoids have also been shown to exhibit habitat partitioning in the coastal temperate waters off Australia. Urolophus lobatus (lobed stingaree) and Trygonoptera personata (masked stingaree) were found predominantly in the deeper offshore waters at around 20-35 m, while Urolophus paucimaculatus (sparsely-spotted stingaree) were greater in the southern region, but no depth preference was identified and T. mucosa (western shovelnose stingaree) were found at a single deep inshore site (Platell et al., 1998). It was hypothesised that this separation by depth and site use allows these batoids to broadly use the same coastal region, without reducing their fitness through competition with the other batoids.

There are examples however when two or more species exhibit preference for the same habitat in the same space. If two or more species select the same habitat there are further mechanisms to reduce competition that allow multiple species to co-occur.

1.1.5 Dietary partitioning In addition or as an alternative to habitat partitioning, species may partition food resources to reduce competition. If species consume different prey types or prey of different sizes, species may stably coexist in the same habitat (Ross, 1986). Habitat and food partitioning may not be mutually

24 exclusive, for example if species have different prey preferences and these prey species exhibit habitat partitioning, predators may also become partitioned by habitat.

Sympatric species of gorilla and chimp in the Kahuzi-Biega National Park, Democratic Republic of Congo, exhibit overlap in their diet in terms of their preferred fruits. Yet gorillas tend to consume fruit with longer availability while chimps are inclined to consume fruit that fluctuates seasonally when food is scarce. These differences in so-called ‘fall back’ foods have an impact on chimp and gorilla social interactions and distributions. Gorillas for example tended to maintain their group sizes but increased competition, resulting in a longer daily path length and low site fidelity, while chimps reduced their party sizes when preferred food was scarce and reused the same areas (Yamagiwa and Basabose, 2009).

In marine species it is often difficult to directly observe the food consumed or collect faecal pellets as is possible for many terrestrial studies of diet. Instead marine based studies often rely on stomach contents or isotope analysis to establish diet (Cortés, 1999, Heithaus, 2001, MacNeil et al., 2005). For example, stomach content analysis was used to examine four elasmobranch species in a large sub-tropical bay. The four species were found to partition food resources to reduce competition: Rhinobatus tupus consumed predominantly penaeid prawns and portunid crabs, while the three shark species (Carcharhinus cautus, Negaprion acutidens and Rhizoprionodon acutus) consumed teleosts which also varied between the three species (White and Potter, 2004, White et al., 2004).

Another possibility is that multiple species use the same habitat and consume the same diet. Additional complexity in the spatial ecology of sympatric species can allow species to optimise their fitness while in competition with other species.

1.1.6 Temporal resource partitioning The way animals use resources can change on various temporal scales. Species may occupy a habitat or consume the same prey items, but this may change at different times, including seasonal and diel changes in resource use (Ross, 1986). These dynamic, temporal differences in resource use, potentially by multiple components of an assemblage, can create complex community interactions within a relatively small area (Caveney et al., 1995, Ofarrell, 1974).

In Boulder, Colorado five Myotis species (bats) use the same watering holes to replenish water loss after roosting. In the first 75 min after sunset, some species shifted their visitation patterns to the water hole earlier than the other species (Adams and Thibault, 2006). In marine species, temporal resource partitioning has also been demonstrated. For example, four shark species

25 aggregated in the same area, but some were resident, while others aggregated in the warmer months of the year (Speed et al., 2011). In the same study, Carcharhinus melanopterus (blacktip reef shark) and Carcharhinus amblyrhynchos (grey reef shark) showed diel patterns of site use from 13:00-14:00 hrs, while Negaprion acutidens (sicklefin lemon shark) occupied the same site from 05:00-10:00hrs. By using the same resource at a different time species can further reduce the potential fitness costs of competition over the same resource.

The previous sections have examined concepts that deal with ways in which species and species assemblages occupy space and habitat over time. Moreover, there are other intra-species groups that may use space and habitat differently and can further influence our interpretation of the full extent of species spatial ecology.

1.1.7 Sexual segregation Further to partitioning among species, intra-species groups such as sexes may also partition resources. Sexual segregation is ubiquitous across the animal kingdom and is defined as the separation of a species by sex, either as individuals or in groups and there has been several hypotheses for the drivers of this segregation (Wearmouth and Sims, 2008).

The predation risk hypothesis suggests that one sex is more at risk to predators than the other, which results in one sex using a ‘safer’ habitat than the other, even at the expense of optimal foraging habitat. For example dall’s sheep Ovis dalli dalli segregate by sex because males and females have different reproductive strategies. Males are found farther from cover, at lower altitudes, in gentle terrain, in habitat where their predators occur and where forage density is higher. Meanwhile females occupy areas near cliffs and at higher altitudes with lower forage density (Corti and Shackleton, 2002). In this case the males adopt a risky strategy by occupying habitat in the presence of their predators but gain higher rewards in the form of higher quality food. Conversely the females adopt a safer (less predators), but less rewarding strategy (low quality food) that results in the segregation of the sexes.

Alternatively, the forage selection hypothesis suggests that sex differences in nutrient requirements result in different prey selection and different habitat use. In grey seals, males tend to forage in further offshore areas while females are resident in inshore areas. This segregation was suggested to occur due to differences in males and females prey selection. Males tend to select a diverse range of prey, but with lower energy content and females a narrow range of prey but with higher energy content. Due to the males larger size, they are able to process larger volumes of lower quality prey than the females (Austin et al., 2004).

Another hypothesis is based on differences in activity budget, which suggests sex differences in body size and reproductive investment result in sex-specific activity budgets. Consequently this causes differences in habitat requirements. In the Tibetan gazelle Procapra picticaudata females spent significantly more time feeding and less time on other activities relative to the males, resulting in sexual segregation. Females are considered less efficient than males at digesting food due to sex differences in metabolic rate, body size and mouth size, which requires females to spend more time feeding than males (Li and Jiang, 2008).

In addition the thermal-niche hypothesis suggests fecundity is influenced by temperature and the optimal temperature at which fecundity is optimised may be different between sexes, resulting in differential habitat use. Sexual segregation is exhibited by Atlantic cod Gadus morhua where females remain in deeper warmer water until they are ready to spawn, move up to the bank in cooler waters to join the males during spawning, before returning to deeper warmer waters (Morgan and Trippel, 1996).

Finally, the social factors hypothesis suggests that intra-sexual affinity for the purpose of cooperation of information transfer or intersexual aversion, due to aggression results in sexual segregation. In the catshark Scyliorhinus canicula foraging in males and females was reduced when in the presence of the other sex and it is suggested that this results from male harassment of females inhibiting female foraging, resulting in sexual segregation (Kimber et al., 2009). Courtship behaviours in elasmobranchs are costly to females, frequently resulting in injury and therefore females avoid males, in some cases occupying sub-optimal habitat (Klimley, 1987, Wearmouth et al., 2012).

These examples illustrate that there are multiple drivers of sexual segregation and also that the occurrence of sexual segregation is a widespread phenomenon, from terrestrial to marine species. Furthermore sexual segregation serves to indicate that it is not only differences among species that can cause segregation, but groups within species. Here differences in behavioural strategy between sexes lead to differences in space use, therefore other intra-species groups that exhibit differing behavioural strategies may also result in spatial partitioning.

1.1.8 Resource partitioning among age groups Another important intra-species group that may partition resources are different age classes. This is particularly the case when foraging juveniles are disadvantaged due to experience or physiological limitations such as gape size. Young elephant seals Mirounga leonina for example used smaller areas than older seals, but made more frequent trips to sea. Younger seals remained

27 consistently closer to land resulting in temporal and spatial segregation between age groups (Field et al., 2005).

Differential habitat selection across ontogeny is common in elasmobranchs (Simpfendorfer et al., 2005, Grubbs, 2010, Morrissey and Gruber, 1993), birthing or nursery areas are often visited seasonally by females for oviposition or parturition. These habitat selections by different age groups may reduce intra-species competition between adults and juveniles. Juveniles are likely to be more vulnerable to predators due to their small size, therefore habitat selection by juveniles may occur to avoid predators that adults of the species need not (Lima and Dill, 1990, Morrissey and Gruber, 1993). There is also evidence that increasing body size increases activity space (McNab, 1963) and adults of a species may also have different energy demands that can only be obtained in different habitat to the juveniles (Grubbs, 2010).

The described concepts can be measured to provide a deeper understanding of fine scale spatial ecology across various species. In this thesis, these concepts have been used to investigate the spatial ecology of four benthic marine predators, which have lacked investigation because they inhabit an environment where they are not easily directly observable, are well camouflaged and do not breach the surface.

1.2 Study species

1.2.1 Rajidae resource use The Rajidae are cryptic species, which bury in the sediment for extended periods of time (Greenway et al., 2016) and inhabit coastal turbid waters (Steven, 1936). Despite being commercially important species in the UK (Enever et al., 2009), little is known about their fine- scale distributions, with most detailed studies focussing on one (R. clavata) of the 14 species in UK waters (Hunter et al., 2006). These 14 species occupy coastal waters around the UK (Figure 3), with some species such as R. brachyura occupying southern UK waters, while other species, such as R. clavata occupy all coastal waters around the UK (Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010). Depth ranges have been established for some species, with the most common species found between 0-283 m and a high degree of overlap among the species (Ellis et al. 2005). Fulton (1893) may have been one of the earliest published studies to question the movements of skate and used a method of marking and recapturing to assess the distance travelled whilst at liberty. This study found that the skate were recaptured

28 relatively close to where they were released after up to 278 days. The majority of studies since, which include various species and different locations around in the UK and worldwide, conclude, like Fulton (1893), that skate show high site fidelity and rarely move far from where they were released (Steven, 1936, Templeman, 1984, Walker et al., 1997, King and McFarlane, 2010, Ellis et al., 2011). Through the use of mark and recapture some species have been shown to remain within the release site or within 40 km of their release position; the maximum distance being 61 km for a single R. brachyura (Ellis et al., 2011).

R. clavata is known to make seasonal migrations from the Thames Estuary area, to deeper offshore locations in the winter before returning to shallower inshore areas in the spring and summer (Hunter et al., 2006). However, some of the other species of Rajidae often coexist with R. clavata (Ellis et al., 2011), creating sympatric assemblages in coastal regions (Ellis et al., 2011, Steven, 1936). There is however a paucity of studies that explicitly investigate fine-scale spatial occurrence, habitat selection and habitat partitioning in the UK Rajidae. It is unknown for example how these 14 species may interact and whether within their known depth ranges they overlap or partition resources.

A B

C D

Figure 3: Rajidae spatial occurrence A= Raja brachyura B=Raja clavata C=Raja microocellata D=Raja montagui (Ellis, 2016, Ellis, 2006, Ellis et al., 2009, Ellis et al., 2007)

With regard to feeding ecology and potential resource partitioning by diet, Rajidae are considered to be generalist opportunistic feeders and while some species have dietary preferences, there is a high degree of dietary overlap (Ellis et al., 1996a, Holden and Tucker, 1974). Ontogenetic shifts in diet are commonly reported in Rajidae; for juveniles, mysids and amphipods are important, but with ontogeny there is a shift from such smaller prey items to larger crustaceans, fish and cannibalism (Ajayi, 1982a, Holden and Tucker, 1974, Ellis et al., 1996a, Farias et al., 2006b, Moura et al., 2008).

Evidence for sexual segregation in Rajidae has been implied previously by studies that found a tendency to capture females in trawls (Steven, 1933). Sex ratio bias in trawls was also demonstrated in R. clavata in the Bay of Douarnez, France, where trawls were predominantly male inside the bay and female outside the bay (Rousset, 1990). Other studies have found no evidence for sexual segregation (Martin et al., 2010). There is little evidence for sex differences in prey selection for Rajidae, as studies that have examined stomach contents conclude that they are generally opportunistic feeders (Ellis et al., 1996b). Rajidae in southwest UK are the top predators in the community, and the four skates studied here are larger than any other benthic predators in this coastal assemblage (Ellis et al., 2005a). Consequently Rajidae are unlikely to be subject to significant, if any, predation in the area studied. Thus males or females are less likely to differentially respond to the threat of predators with the consequence of sexual segregation. Over the last 100 years however populations of other elasmobranchs, including possible predators of skates, such as the common skate, Dipturus spp., have declined (Dulvy et al., 2000, Ellis et al., 2005a). Therefore although predator avoidance seems unlikely it is possible that sexual segregation could represent an historic behavioural trait evolved to anticipate the risk of predation. In addition, Rajidae are oviparous species, with females laying eggs in nursery areas, growing larger and maturing later than males (Ellis et al., 2005a). Females may spend time in different habitat to males during oviposition and differences in growth and maturity may cause a separation of the sexes in these species. Furthermore, there is evidence that copulation is costly to female Rajidae (Luer and Gilbert, 1985), which may cause females to avoid males, as is the case in other elasmobranchs (Klimley, 1987, Wearmouth et al., 2012). There is therefore evidence enough to hypothesise that sexual segregation is exhibited by Rajidae, but more investigation is required to clearly demonstrate spatial segregation of the sexes.

Ontogenetic shifts in skate spatial ecology have been suggested previously. Rajidae are oviparous, depositing eggs in nursery areas, with juveniles remaining relatively nearby until maturity (Ellis et al., 2005a). Rajidae do not exhibit parental care after oviposition, therefore adult females are

31 likely to leave these nursery areas resulting in segregation of age classes. Distribution of skate nurseries are broadly suggested to be shallower than adult distributions, but egg laying habits are also poorly understood (Ellis et al., 2005a). Evidence for segregation of these early stages requires further exploration and to define if the nature of the segregation (juveniles shallower than adults) is the same among species.

1.2.2 Fishery history and management of Rajidae In the early 1800s skates appear to be unused by fisheries except for using their meat to bait crab pots. Colloquially, skates were referred to as ‘rabble fish’ by fishermen as they had no value at markets (Stevens 1932). Skate were rejected from fish markets up to the late 1800s; even the word ‘skate’ comes from the Saxon word ‘skitan’ meaning ‘to reject’ (Steven, 1947). Skates and rays were mentioned in a publication in around 1896 as marketable marine fish with many of the common rays and skates used as food (Cunningham, 1896). Towards the end of the century and the beginning of the next, skate fisheries began to increase (Dulvy et al., 2000). After World War One (WW1) (1914-1918) landings of skate increased again along with an increase in price. In 1929 the Devon and Cornwall skate fishery was very valuable, with skates providing 30% of the total value for demersal fishing (Steven, 1932a).

Skates were caught using beam and otter trawls, as well as longline and set nets by the 1930s (Steven, 1932a). Between WW1 and WW2 however there was a large decline in the number of skates and rays caught in the English Channel (Steven, 1947). After WW2 the fishing industry began to expand and post war technology, also allowed more efficient fishing (Thurstan et al., 2010, Smith, 2000). For example the introduction of tickler chains to trawls allowed fishing on grounds that had previously been too rough to exploit (Engelhard, 2009). In addition, the gillnet fisheries previously used twine nets that were heavy, weak and easily visible, but mono-filament nets allowed for more efficient fishing (Bavinck, 2011).

Since the early 1900s scientists have continued to document the state of skate populations and many report continuing declining trends in skate populations. For example, between 1956-1995 there was a reported 80% decline in thornbacks R. clavata in the North Sea (Chevolot et al., 2006). In the 1980s a study was published concerning the local extinction of the common skate in the Irish Sea (Brander, 1981). International landings of skates also declined from 16,000 tonnes in 1945 to 4,000 tonnes in 1995 (Hunter et al., 2006) and in the western English Channel Rajidae have shown declines since the 1950s, primarily due to overfishing (Genner et al., 2010b). There is therefore a clear need to understand skate ecology to provide knowledge in order to manage these vulnerable species.

Management for Rajidae has progressed only more recently. Until 2009 skate landings by the commercial fishery were recorded as ‘skate and ray’ hampering species-specific management (Silva et al., 2012). In 2009 it became mandatory to land ‘skates and rays’ by species, and it also became “prohibited for EU vessels to fish for, to retain on board, to tranship or to land” certain species in specific ICES areas. This protection includes common skate D. batis and white skate Rostroraja alba (Silva et al., 2012, CEC, 2011) principally due to D. batis being IUCN Red List assessed as ‘critically endangered’, and the white skate Rostroraja alba as ‘endangered’. Currently however quotas are largely grouped in allocations (MMO, 2017a). Locally, the Kent Inshore Fisheries Conservation Agency (IFCA) introduced a minimum landing size for skate so that individuals are able to spawn at least once before capture (Whittamore and McCarthy, 2005). Most IFCA districts however have no minimum landing size. Nottage and Perkins (1983) found unrestricted fishing led to 48.6% of skate retained being immature in the Solway Firth.

Hunter et al. (2006) highlight the importance of understanding the movements of skates to help protect and manage them. They suggest the Thames estuary is important for thornback rays R. clavata between March and September as they migrate into the area at this time to breed. Closure of the estuary could result in a significant reduction in skate landings and protect mature females over the spawning season. However, a closure at this time could also result in increased fishing pressure on juvenile thornback rays, which hatch between September and February (Brander and Palmer, 1985) when the fishery is open. This reflects the need for more studies of the complex issues surrounding the best way to protect mobile species and how their movements can effect whether a Marine Protected Area (MPA), for instance, is working.

1.3 Aims and objectives Given that most studies focus on the broad scale spatial occurrence of skates and that an understanding of fine-scale spatial ecology has been overlooked or difficult to investigate, this project focusses on the fine-scale spatial ecology of skates and the potential for resource partitioning among species and intra-species groups as a result. In addition, I explore the interaction of Rajidae with the local and national fisheries.

Specifically, this thesis aims to:

(i) Identify the fine-scale spatial occurrence of four rajid species within a coastal area off the south west coast of the UK and examine whether this spatial occurrence is linked to underlying habitat associations including a standard set of abiotic factors.

(ii) Investigate habitat selection and habitat preference in Rajidae to examine whether rajids actively select and occupy particular areas preferentially. (iii) Explore resource partitioning, of both habitat and diet among species and intra- species groups, such as sex and age groups. (iv) Consider temporal dynamics of habitat selection over seasonal and diel scales, including whether temporal partitioning of resources among and within species may occur. (v) Investigate how habitat partitioning among the species may affect their interaction with the fishery. Furthermore, I explore how these differences may influence the value of existing management measures such as Marine Protected Areas (MPAs) for these mobile species and whether legislation to protect these vulnerable species is effective.

Chapter 2. Spatial and temporal occurrence of four sympatric Rajidae species in coastal waters of the Western English Channel: are there fine-scale habitat differences?

2.1 Introduction The spatial occurrence of a species is considered a fundamental aspect of ecology, playing a role in shaping the evolution of its life history, driving population level processes and species interactions (Elith et al., 2006). Where a species occurs in space may also be linked with particular abiotic or biotic factors, such as depth, temperature, vegetation type or prey availability (Schlaff et al., 2014), and often termed habitat association. Organisms may associate with particular habitats because they are physiologically adapted to an optimal environmental range or because other resources such as prey, mates or shelter are available (Kramer et al., 1997). Habitat selection in animals can be defined as space use that is non-random due to voluntary movements in response to multiple factors, over a spatial scale that can vary from thousands of kilometres to a few centimetres (Sims, 2003, Kramer et al., 1997, Morris, 2003). In addition, habitat selection can be temporally dynamic, when, for example, animals migrate between different habitats at different times of year (Wilson et al., 2001, Dudgeon et al., 2008, Chapman et al., 2015). Species that live in sympatric assemblages, appearing to occupy similar niches, are potentially in direct competition for resources and there is increasing evidence across taxa that habitat selection allows sympatric assemblages to coexist in a relatively small area (Speed et al., 2011, Papastamatiou et al., 2006, Rosenzweig, 1981, Sims, 2003).

An understanding of elasmobranch habitat association is vital for effective management and conservation, because habitat association can affect patterns of distribution and therefore commercial catch rates, potentially leading to overfishing of certain species (Sims, 2003). Such information can provide scientific evidence and advice for the location and designation of marine protected areas that could include multiple species distributions rather than a single species (Speed et al., 2010). Management is especially important for elasmobranchs which are vulnerable to overfishing due to their slow growth, late age at maturity and relatively low fecundity (Dulvy et al., 2014, Field et al., 2009). In the UK, Rajidae species are an important commercially valuable component of mixed demersal fisheries (Hunter et al., 2006). The sympatric nature of Rajidae

35 distributions (Ellis et al., 2005a), in addition to their similar morphologies, likely contributed to commercial fishing vessels grouping these separate species as ‘skate and ray’ in landings. Despite new legislation in 2009, which requires fishers to land Rajidae as separate species, the accuracy of species-specific landings data remains questionable (Simpson and Sims, 2016) and in the most recent quota allocations most Rajidae species remain in the group ‘skates and rays’ (EuropeanCommission, 2016). It is therefore important to increase our understanding of the fine- scale spatial ecology of these vulnerable species and begin to disentangle the differing habitat associations of these frequently grouped sympatric species.

In Europe, fishing surveys, usually in the form of short research trawls, have been used to estimate the spatial distribution of species and fish populations for over 100 years, the first being initiated by the International Council for the Exploration of the Sea (ICES) (Barrett et al., 2004). This ongoing task requires a great deal of effort, time and resources to implement, which results in a sampling design that attempts to cover a large area with the available resources (Gunderson, 1993, Francis, 1984). For example, the International Bottom Trawl Survey (IBTS), aims to investigate the distribution and relative abundance of fish species, with protocols set by ICES (ICES, 2012). Within each ICES area, a survey grid is defined with ICES rectangles of approximately 30 by 30 nautical miles (55.56 km). Vessels from two countries carry out hauls annually, with standardised 30 minute tows. Data is collected in one or two months of the year in each ICES rectangle; for example in the Eastern English Channel, data is collected only in October (Martin et al., 2010). Therefore, the habitats occupied by UK skate species are only known at relatively broad scales of space and time and represent only snapshots of their distributions at specific points in time and space.

The data collected from research surveys form the majority of our knowledge of Rajidae distributions in the UK. The spatio-temporal distribution of several Rajidae species have been mapped for the eastern English Channel e.g. (Martin et al., 2010), and depth ranges for some species have been established (Table 1) (Ellis et al., 2005a). The depth ranges indicate that most of the common species of Rajidae found in the UK have an overlapping coastal distribution, with some species exhibiting a more southerly distribution, such as the blonde ray Raja brachyura (Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010).

Table 1: Rajidae depth ranges and occurrence (Wheeler, 1969b, Whitehead et al., 1984, Ellis et al., 2005a)

Species Depth range(m) Area Occurrence (Wheeler, 1969b) R. brachyura 14-146 Common in inshore In waters less than 100 m, centre Blonde ray waters of southern and of the range for adults is 40 m. western UK Do not tolerate estuarine conditions. Found on sandy ground. Unisex shoals migrate and certain times of the year (unspecified). Females occur in the summer in the English channel. R. clavata 7-192 Widespread Found between 2-60m on a wide Thornback ray variety of habitats (mud, sand shingle and gravel). Inshore migrations during winter, first females followed by males. R. microocellata 0-112 Common in Bristol Inhabits shallow to moderate Smalleyed ray Channel, occasionally depths, confined to sandy in Bristol Channel and grounds, bays and estuaries. Not St Georges Channel widely distributed. Females observed in the summer in the English channel. R. montagui 8-283 Widespread Usually deeper than 25 m, most Spotted ray common between 60-120 m. Found on a wide range of habitats (both sandy and rough grounds). In the English channel from April to July. R. undulata 0-72m Frequent in English Found in moderate to shallow Undulate ray Channel depths, as deep as 200m, not usually inshore. Prefers sandy substrates.

These studies provide important broad-scale estimates of Rajidae distribution, however the ability to detect broad-scale spatial patterns often incurs the cost of overlooking fine-scale details (Wiens, 1989). This fine-scale information is especially important given that Rajidae are mobile, that they have patchy distributions and are thought to remain in relatively localised areas (Steven, 1936). Without assessing finer scale distributions of these species it will be difficult to determine the drivers of habitat selection behaviour of these fish, which are crucial for predicting the likelihood of distributional shifts in relation to predicted environmental changes (Lea et al., 2018).

Rajidae are considered to be generalist feeders and while some species have dietary preferences, there is a high degree of overlap among species (Ellis et al., 1996a). These sympatric Rajidae species, with similar diets are likely in competition for prey, which may lead to resource partitioning. Leading from this I hypothesise that sympatric Rajidae species exhibit finer scale habitat associations than previously described. These distributions may be temporally dynamic, and driven by seasonal behaviour such as egg laying, mating or fluctuations in the abundance of prey. There may also be geographic segregation within a species.

Intra-species habitat selection occurs in elasmobranchs, between sexes and throughout ontogeny, which adds to the complexity of a species’ distribution (Simpfendorfer et al., 2005, Grubbs, 2010). Sexual segregation is ubiquitous across the animal kingdom and is defined as the separation of a species by sex, either as individuals or in groups (Wearmouth and Sims, 2008). Sexual segregation in elasmobranchs is well documented in a number of species including white sharks Carcharodon carcharias (Domeier and Nasby-Lucas, 2012), catsharks Scyliorhinus canicula (Wearmouth et al., 2012) and blacktip reef sharks Carcharhinus melanopterus (Mourier et al., 2013). Segregation of age classes in sharks has also been reported across several species including pigeye sharks Carcharhinus amboinensis (Knip et al., 2011), lemon sharks, Negaprion brevirostris (Guttridge et al., 2009) and blue sharks Prionace glauca (Carvalho et al., 2011). Rajidae are oviparous, depositing eggs in nursery areas, with juveniles remaining relatively close by until maturity (Ellis et al., 2005a). Distribution of skate nurseries are broadly suggested to be shallower than adult distributions, but egg laying habits are also poorly understood (Ellis et al., 2005a). It is likely however that ontogenetic segregation occurs in Rajidae, but more investigation is required to determine if the segregation occurs in all species in the same way.

The current study aims to quantify the frequency of occurrence of four skate species, at five sites within a relatively small local area, that is, at scales smaller than the lowest resolution distance between sample locations in current broad-scale national surveys. On the basis of the broad-scale distribution survey results, I assume that we are sampling within the overlapping distribution

38 range of the four Rajidae species and that differential habitat selection is not exhibited. Thus, I hypothesise that all species should be randomly distributed throughout the area. Alternatively, if Rajidae distribution is non-random and each species is found more frequently in different local sites, it can be inferred that the occurrence of differential habitat selection may be present between four sympatric Rajidae species. I use these results to consider the effects of varying seasonality of catches and differential catches of juveniles compared to adults and between the sexes.

2.2 Methods

2.2.1 Sampling

The study site chosen off Plymouth, UK (50.34°N, 4.28°W) is a complex mosaic of pelagic and benthic habitats with different primary producers including sea grass, kelp, and phytoplankton (Heape, 1888, Parke, 1950), as well as substrates ranging from rock to sand and mud (Figure 1). Between January 2000 and June 2016, three research vessels were used to sample sites located within 30 km of the west end of the Plymouth Breakwater (50.3295°N, 4.1629°W), at a maximum depth of 57 m. The sites used for this study are identified as Whitsand, Bigbury Inner, Bigbury Outer, Standard haul (Station L4) and the lower (marine) reaches of the River Tamar (Table 1and Figure 2). These sites comprised the majority of hauls conducted in the area. The sites are part of a wider fisheries time series that has run for over 100 years and cover areas of different sizes and are different distances apart (Figure 4). This provides an opportunity to compare the occurrence of skates at sites that are adjacent to one another (e.g. Bigbury inner and outer) as well as those further apart (e.g. Whitsand Bay and Bigbury inner). For all hauls since 2000, Rajidae were identified to species level and counts were recorded for each species. The four species most common in the area were analysed in the current study; Raja brachyura, R. clavata, R. montagui and R. microocellata. Sex, length (mm), weight (g) and disc width (mm) were recorded for each individual on routine fishing trawls.

Figure 4: EUNIS substrate types off the coast of Plymouth, UK (European Environment Agency, 2018).

Figure 5: Sampling sites off Plymouth Sound.

Table 2: Details of the sampling sites. Average depth(m) of the site, the maximum depth of the site(m), the number of hauls conducted and the time spent trawling (hours)

Fishing grounds Average Maximum Number of Hrs of depth(m) depth(m) hauls trawling Whitsand 13.43 27.54 280 135.38 Bigbury inner 19.01 30.19 143 53.48 Bigbury outer 34.68 42.78 431 291.32 Standard haul 52.43 57 382 160.06 (L4) River 10 39 63 18.41

Vessels varied in length from 15.4 to 21.5 m (Appendix 1 Table S1). Gear type was either a demersal otter trawl or beam trawl. All nets had an 80mm diamond mesh, but other net dimensions varied between some sample sites, a summary of which can be found in (Appendix 1 Table S2). Of the five sites, three (Whitsand, Bigbury Inner and Bigbury Outer; Figure 5 and Table 2) were sampled with the same demersal otter trawl having the same net dimensions.

Differing gear types and catchability of species can influence catch rates. Firstly, smaller individuals are suggested to have a lower probability of capture in any demersal gear (Walsh, 1992, Fraser et al., 2008). This was tested in this study by examining the overall frequency of each species captured in the study area as a whole using a Chi Square Goodness of fit test. Secondly, if a particular species is more vulnerable to capture than the others, we may expect to find the same species in higher proportions in all sites in the study area. This was tested using a Chi Square Goodness of Fit to compare the proportion of each species caught within the five study sites. Finally, at the Standard haul (Figure 5) site, a larger Otter trawl net was used compared to the other sites, resulting in a larger swept area than the smaller nets. This may result in a relatively greater catch in all species compared to the smaller nets used at other sites. This potential bias was investigated using analysis of percentage presence (one or more) of any Rajidae species in trawls for each site.

To visualise location of capture for each species, all trawl locations between 2000 and 2016 were plotted in Arc GIS (Version 10.2.2) to display trawls with species absence as hollow circles, and warmer colours to indicate increasing presence. To compare catch per unit effort (CPUE) for each

41 species between sample sites, overall CPUE was plotted (sum of species captured/sum hours trawling at each site between 2000 and 2016).

2.2.2 Comparing occurrence of species within and among sites When analysing research trawls, catch per unit effort (CPUE) is frequently used as a proxy for relative abundance. However in the current study the Rajidae trawl data collected were highly variable and contained a large number of zero values. CPUE was therefore deemed inappropriate due to low average CPUE values and high variance. Thus, to test the hypothesis, a Chi Square Goodness of fit was used to test if proportions differed from the expectation of equality predicted by the null hypothesis and to provide an indication of the proportion of each species captured across all the sample sites combined. Equal expected proportions were used in this case because within each sample site, equal effort was given to catching each species assuming a random distribution. A significant result could indicate that (i) there is spatial variation in the distribution of each species or (ii) that one species is generally more frequently caught than another. To resolve this, I compared the proportions of species captured between sites; if the species with the higher proportion differed between sites, it would suggest that variation in spatial distribution was more probable.

The previous analysis was used to indicate whether a species has higher occurrence in an area relative to the other species. The second analysis identified the areas where a species is most abundant, relative to its abundance in other areas. To test this, a Chi Square Goodness of fit was conducted, for each species comparing catches among sample sites. Fishing effort differed between sites, therefore the total duration (minutes) of trawling between 2000 and 2016 was calculated for each sample site and was used as the expected values for this analysis.

2.2.3 Depth of capture Depth was determined for each trawl using the coordinates of the location where the net was shot and interpolated from one second mosaic bathymetry (DEFRA) in Arc GIS (Version 10.2.2). To identify ontogenetic differences in distribution, the skates were separated into two classes (adult, juvenile) using length at maturity data from previous studies (Table 3); where values differed between studies, to be consistent the smallest value was used to identify juveniles.

Table 3: Rajidae length at maturity

Species Length at maturity (mm) Reference R. brachyura 550 (McCully et al., 2012) R. clavata 420 (Nottage and Perkins, 1983) R. microocellata 570 (Ajayi, 1982a) R. montagui 400 (McCully et al., 2012)

To test for differences in capture depth between adults of the four species, analysis of variance (ANOVA) on ranks (Kruskal-Wallis test) with Dunn’s method for post-hoc testing was conducted and Mann-Whitney U tests were conducted on each species to compare juvenile and adult depth of capture. These tests were used due data not meeting the assumptions made by parametric tests (e.g. a zero truncated distribution).

2.2.4 Seasonality analysis To further investigate fine-scale, within-site temporal variability, the sites where species were found at higher frequencies were selected for analysis. Percentage presence (one or more) and absence for each trawl were calculated and plotted for each month. To supplement this analysis, seasonal changes in catch frequency within sites were tested with a Chi squared goodness of fit. Seasons were designated as follows; spring= February, March and April, summer= May, June and July, autumn= August, September and October, winter= November, December and January. The fishing effort varied between seasons, therefore expected proportions were based on the effort (hours trawling) in each sample site and in each season.

2.2.5 Sex ratio analysis If skate sexes are randomly distributed across this coastal zone then there should not be significant intra species differences in sex ratio. To test where sex ratio was significantly different to equality (1:1) across the whole study site, counts for each sex within a species were analysed with a Chi Squared goodness of fit test.

2.3 Results Overall, catches of the four species across the entire study site were significantly different from equality [2(3, N=1554) =571.318, p<0.001]. Catches of R. montagui were significantly higher than expected, making up 51.16% of the total catch of skates (Figure 8A).

Catches made during the study period indicate the very low occurrence of three of the four species at the L4 site (Figure 6). Only three R. brachyura and one R. microocellata were captured within this site in 16 years. In contrast, R. brachyura catches are very focussed spatially, occurring solely around Bigbury Outer. R. clavata appears to have a more widespread distribution, being captured at a similar frequency to the other species across a broader range of sites.

Catch per hour analysis indicated the magnitude of differences in catch rate, even between nearby sample sites (Figure 7). For example Bigbury Outer CPUE was much higher for both R. brachyura and R. montagui. The difference in catch per hour was more than 10 fold higher for R. brachyura and R. montagui at the Bigbury Outer site than Bigbury Inner, even though Bigbury Inner and Bigbury Outer are adjacent sites, with the centre points of each site only 4.8 km apart.

Figure 6: Capture locations for four species 2000-2016. Hollow circles represent empty hauls, colours indicate presence. A=R. brachyura, B=R. clavata, C=R. microocellata, D=R. montagui

10 R. brachyura R. clavata 8 R. microocellata R. montagui

Catch per hour per Catch

0 River Whitsand Bigbury inner Bigbury Outer Standard haul Grounds Figure 7: CPUE at each of the five sites during the study period 2000-2016. Note the magnitude of difference in CPUE between Bigbury Outer and Bigbury Inner which are adjacent sites (their centre points are only 4.831km apart).

2.3.1 Comparing occurrence of species within and among sites The results suggested that the four skate species were not found in equal proportions within every sample site (Table 4, Figure 8B-F) and that the species that were most predominant also differed between sites. In Bigbury Outer and Inner R. montagui made up the highest proportion of the catches (67% and 41% respectively), however at L4 and in the River, R. clavata were proportionally higher than the other species (65% and 88%). In Whitsand, R. microocellata made up the highest proportion (47%) of catches.

100 A B

Proportion captured (%) captured Proportion 20

100 C R. brachyura R. clavata R. microocellataR. montagui D R. brachyura R. clavata R. microocellata R. montagui Species Species 80

Proportion captured (%) captured Proportion 20

100 E R. brachyuraR. clavataR. microocellataR. montagui F

80 Species

Proportion captured (%) captured Proportion 20

R. clavata R. clavata R. brachyura R. montagui R. brachyura R. montagui R. microocellata R. microocellata

Species Species

Figure 8: Within site difference in species counts. A) All sites combined B) Bigbury Inner) C) Standard haul D) Whitsand E) River F) Bigbury Outer

Table 4: Within site Chi Square Goodness of Fit test results

Fish ground Chi square goodness of fit Dominant species Bigbury Outer 2(3, N=980)=1007.766, p<0.001 R. montagui Bigbury Inner 2(3, N=117)=26.76, p<0.001 R. montagui Standard haul (L4) 2(3, N=133)=139.09, p<0.001 R. clavata Whitsand 2(3, N=290)=91.41, p<0.001 R. microocellata River 2(3, N=34)=73.06, p<0.001 R. clavata

Results indicate that catches of all species differed significantly between sites (Figure 9 and Table 5). R. clavata were caught more frequently than expected in the River and Bigbury Outer, but less than expected in hauls at L4. R. montagui was clearly most frequently caught in Bigbury Outer and far less than expected in all other areas, while R. microocellata was caught most frequently in Whitsand and very rarely in the Standard haul area (L4). Finally, R. brachyura was most frequently caught in Bigbury Outer and rarely caught on other grounds. Interestingly, there was a complete absence of R. brachyura in the river sample sites and every species were found in greater numbers than expected in Bigbury Outer.

Table 5: Between site differences Chi Square Goodness of Fit test results.

Species Chi Square Goodness of Fit Grounds most frequently caught R. brachyura 2(4, N=270)=1348.89, p<0.001 Bigbury Outer R. clavata 2(4, N=269)=89.80, p<0.001 River R. microocellata 2(4, N=220)=225.10, p<0.001 Whitsand R. montagui 2(4, N=795)=3764.41, p<0.001 Bigbury Outer

Figure 9: Between site differences in counts. a) R. brachyura b) R. clavata c) R. microocellata d) R. montagui 49

2.3.2 Depth of capture There were significant differences between the depths of adults of some of the four species (ANOVA on Ranks, Kruskal-Wallis, p<0.001), however not between R. brachyura and R. montagui, and R. montagui and R. clavata. On average, R. brachyura and R. montagui were found at 35 m, R. clavata at 32 m and R. microocellata much shallower at 15 m (Figure 10).

There were also significant differences between adult and juvenile depth of capture for every species (Mann-Whitney U, p<0.001) except R. microocellata (p=0.257). In R. montagui and R. brachyura, juveniles were found shallower than adults, but R. clavata juveniles were found deeper than adults.

0 R. brachyura R. clavata R. microocellata R. montagui

Depth (m) Depth 40

60 Adult Juvenile Adult Juvenile Adult Juvenile Adult Juvenile Species/age Figure 10: Difference in depth of capture between four species and age classes. The box represents 25th and 75th percentiles, whiskers, 5th and 95th percentiles, horizontal line is the median

2.3.3 Seasonality analysis The analysis of the proportion of presence and absence of the four species in different sites (Figure 11), suggested that Rajidae occurrence varies across the year depending on the species and the sample site. The highest presence for any species was R. montagui in Bigbury Outer in July, at 86.7% occurrence. Whilst R. clavata was the most common species in the Standard haul site, its highest occurrence was only in 28.1% of hauls in September. Meanwhile, R. microocellata was most common in Whitsand with its highest occurrence in March (40%). The overall presence of any Rajidae species in a trawl varied between areas (Table 6). Across all sites there was only 39% presence of any Rajidae species, but when divided into sites, presence varied from extremely low, at 17% in Standard haul (L4), to 87% at Bigbury Outer.

100

Precent presence Precent 30

100 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 90 Month 80

Percent presence Percent 30

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Month

Figure 11: % presence/absence across the year. A) R. brachyura at Bigbury Outer B) R. clavata at Standard haul C) R. microocellata at Whitsand D) R. montagui at Bigbury Outer. Colour represent seasons, blue= winter, yellow= spring, orange= summer, red= autumn.

Table 6: Overall % presence of all Rajidae species in hauls/sites

Site % presence of Rajidae spp. All hauls 39% Whitsand 34% River 71% Bigbury Outer 87% Bigbury Inner 52% Standard Haul (L4) 17%

Seasonal variations in catch abundance were identified in all species (Table 7) although differences were detected between sites. Catches of R. montagui in Bigbury Outer were significantly higher in summer (Figure 12c) while R. brachyura were captured more frequently than expected in autumn (Figure 12e). R .clavata in Whitsand were captured more frequently in summer (Figure 12b) and in the River they were captured more frequently in spring (Figure 12d), however it is important to note that no sampling took place in the River during winter. R. microocellata in Whitsand Bay were captured more frequently in winter and spring than other seasons (Figure 12a).

Table 7: Seasonal differences in catches, Chi Squared Goodness of Fit results

Species Fishing grounds Chi Squared Goodness Peak season of fit R. montagui Bigbury Outer 2(3, N=656)=38.53, Summer p<0.001 R. brachyura Bigbury Outer 2(3, N=228)=86.59, Autumn p<0.001 R. clavata River 2(3, N=28)=17.52, Spring (NO FISHING IN p<0.001 WINTER) R. microocellata Whitsand 2(3, N=136)=68.31, Winter/Spring p<0.001 R. clavata Whitsand 2(3, N=76)=11.39, Summer p=0.01

Expected 400 a Observed b 100

300

C 200 50

100 25

0 0 Category Spring Summer Autumn Winter Category Spring Summer Autumn Winter

14 60 60 c d e 40 12 40

10 45 45 30 30

s s

t t

n 8

n n

u u

u o o 30

C o

30 o

C C C 6 C 20 20

4 15 15 10 10 2

0 0 0 0 0 Category Spring Summer Category AutumnSpring CategorySummerSpring AutumnSummer WinterAutumn Winter Category Spring CategorySummerSpring AutumnSummer WinterAutumn Winter

Figure 12: Seasonal difference in captures by site and species. a) R. brachyura in Bigbury Outer b) R. montagui at Bigbury Outer c) R. clavata in the River d) R. microocellata at Whitsand e) R. clavata at Whitsand

2.3.4 Sex ratio analysis The sex ratio for all fishing grounds combined indicated a bias in sex ratio towards females for 3 out of 4 species, only R. montagui showed no significant biases towards either sex (Table 8, Figure 10). (Figure 13, Table 8).

Table 8: Sex ratio in catches, Chi Squared Goodness of Fit results

Species Chi squared Goodness of fit Bias R. brachyura 2(1, N=138)=7.42, p=0.006 Female R. clavata 2(1, N=148)=29.43, p<0.001 Female R. microocellata 2(1, N=135)=6.23, p=0.013 Female R. montagui 2(1, N=276)=1.45, p=0.229 Equal

100%

90%

80%

70%

60%

50% Female Male 40%

30%

20%

10%

0% R. brachyura R. clavata R. microocellata R. montagui

Figure 13: Sex ratio in the catches for four Rajidae species. There is a significant bias towards females in R. brachyura, R. clavata and R. microocellata.

2.4 Discussion It has been suggested that only by understanding spatial distributions through time can we estimate abundance in an unbiased manner (Gunderson, 1993). Previous studies have provided estimates of broad-scale distributions for various Rajidae species, indicating a coastal distribution with overlapping depth ranges (Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010). However the current study focuses on an area of this coastal zone, within the previously described depth ranges of all species and suggests that four species of Rajidae and subsets of each species (sex and age group) exhibit pronounced differences in the local areas they occupy within this zone.

2.4.1 Rajidae habitat associations At each of the five sampled sites, different species were not captured in equal proportions and the species with the highest proportion differed between sites, indicating a non-random distribution. R. montagui was the most dominant species in Bigbury Inner and Bigbury Outer, R. clavata were highest in Standard haul (Station L4) and within the River, while R. microocellata were the predominant species in Whitsand Bay. Interestingly, R. brachyura was caught in relatively low numbers at every site compared to other species. Results from the between site analysis reveals that although R. clavata were the predominant species in Standard haul, it is not the site where they were most commonly caught, since they were most frequently caught in the River and at Bigbury Outer. The association of R. clavata with the River was expected, as previous studies have demonstrated an association of this species with estuarine habitats (Hunter et al., 2005b). In this sampling area, it appears only R. clavata have a preference for estuarine habitat. All species were found in greater numbers than expected at the Bigbury Outer site. For R. brachyura and R. montagui both Bigbury sites appeared especially important, because they were rarely caught at any other site sampled. In contrast, R. microocellata were caught most frequently in Whitsand Bay relative to other sites.

Despite clear differences in catch rates of skate species across the different sites the results should be interpreted with caution because at some of the sites different gear was used and different species may have different catchability. Different gear types and net dimensions, as well as the behaviours of different species when faced with trawl gear, have been shown to influence their catchability (Rijnsdorp et al., 1996, Fraser et al., 2008, Harley and Myers, 2001, Winger et al., 2004, Fraser et al., 2007). The practicalities of standardising gear, especially in long-term data collection spanning decades is particularly challenging, and therefore many long-term studies of fish distribution and abundance involve multiple gear types (Ellis et al., 2005a, Knijn et al., 1993, Martin et al., 2010, Walker and Heessen, 1996, Walker and Hislop, 1998). However, Rajidae

55 species have similar morphology and are vulnerable to trawling due to their passive behaviour when faced with a trawl (Queirolo et al., 2012). Consequently the four species studied here are more likely to exhibit similar catchability. In addition, all four species are caught by commercial fisheries and scientific surveys using both beam and otter trawls (Ellis et al., 2005a, Silva et al., 2012, Ellis et al., 2011), indicating that all species can be caught by these gear types. The complete or near absence of a species in a trawl is likely a true representation of species presence or absence in an area. It was hypothesised that a smaller species, in this case R. montagui, may have a lower probability of capture. However across the entire area R. montagui had the highest frequency of capture, though the majority were captured at a specific site. In addition, it was hypothesised that the larger Otter trawl used in the Standard haul area would result in larger number of Rajidae captured, due to a larger ‘swept area’. Contrasting with this expectation, the majority of hauls at this site were devoid of any Rajidae species, indeed, only 17% of trawls contained one or more Rajidae individual. In addition, three out of five sample sites in the current study used exactly the same gear and within these sites there were differences in the most commonly captured species. This provides further evidence that the non-random distributions of these species in this small-scale coastal zone were due to differences in distribution rather than catchability.

This observed uneven distribution of skates appears to occur on a relatively fine scale. The distance between the closest sites sampled was only 4.8 km (between the centre point of the sites), yet there was over tenfold difference in CPUE of R. montagui and R. brachyura between these sites. It was striking that it was necessary to move only 17.2 km to catch a different species at a higher frequency, e.g. R. clavata at the Standard haul site. The overall chance of capturing one or more Rajidae spp. within these two nearby sites (Bigbury Outer 87%; Standard haul 17%) highlights the difference between them.

Depth of capture can also be an indicator of benthic species distributions and has been used to delineate benthic coastal assemblage structure (Garces et al., 2006, Silvestre et al., 2003, Gray and Otway, 1994, Blaber et al., 1994, Connell and Lincoln-Smith, 1999). In the current study, depth of capture differed significantly between some species, whilst others overlapped. R. brachyura and R. montagui had a similar average depth of capture, while R. microocellata were caught at significantly shallower depths than the other species. R. clavata were caught deeper, but at a wider range of depths. In most cases these results support a recent electronic tagging study in the same coastal region, which found R. brachyura and R. montagui had deeper weekly mean depths (around 35 m) than R. microocellata (17 m) (Humphries et al., 2016c). The current

56 study adds a horizontal component to the previous tagging study and suggests that not only do R. brachyura and R. montagui share a similar depth preference, but also that they occupy at least one geographic area in common (Bigbury Outer). The current results do suggest however, greater differences in R. microocellata and R. clavata horizontal distributions than the tagging study could detect. This fine-scale habitat partitioning likely reduces competition (Ross, 1986) among these predators, within a small coastal zone.

The available habitat maps (Figure 4) were not fine enough scale to make any links to substrate preferences. For example R. montagui and R. brachyura were most commonly found at the Bigbury Outer site, which, according to the habitat maps, is a large area of high energy circa- littoral rock. However this area is trawled by demersal gear, both commercially and by this research survey, suggesting that this area contains some soft bottom substrates which have not been detected in the EUNIS habitat survey. Finer scale habitat maps are required to establish if the four species show fine-scale associations with particular substrates.

2.4.2 Temporal habitat association The results also suggest high seasonal variability in the catches of a species within each site. R. montagui in Bigbury Outer for example, were caught more frequently in summer, while R. brachyura at the same site were caught most frequently in autumn and summer. In Whitsand, R. microocellata were most frequently captured in winter and spring, while R. clavata were more commonly captured in summer and to a lesser extent spring. Finally, R. clavata within the River were most commonly caught in spring. These seasonal differences in catches suggest that all species exhibit some seasonal differences in occurrence, presumably due to movements into and away from these areas at those times. In this analysis of trawl data, it was not possible to identify individuals to assess whether the same individuals are returning to the site each year in a particular season. However, previous number tagging of R. clavata in the same area off Plymouth demonstrates a high degree of site fidelity (Steven, 1936) and suggests that the same individuals may return to the same sites seasonally.

Furthermore, there are variations in capture within seasons at each site across the year. For example, for R. montagui at Bigbury Outer, during the month where catches were most frequent, a proportion of the trawls were devoid of R. montagui. Another example is R. brachyura in Bigbury Outer, where July showed the peak percentage presence, yet 33.3% of trawls were devoid of R. brachyura. Previous studies have shown that demersal fish have the capacity to avoid trawls, and, when multiple trawls are conducted over the same grounds, disturbance may cause Rajidae spp. to avoid the area (Ona and Godø, 1990, De Robertis and Handegard, 2012).

Conversely, it has also previously been demonstrated that trawling can encourage fish to move into the trawled area. Trawlers discarding unwanted fish that are likely to be dead or damaged can encourage scavengers or opportunistic feeders to move into the area (Ramsay et al., 1998, Groenewold and Fonds, 2000 (Kaiser and Ramsay, 1997). R. montagui for example has been demonstrated to move into recently trawled areas to scavenge discarded fish or damaged benthic invertebrates (Olaso et al., 2002). Overall the results of the current study indicate some fine-scale temporal movement and aggregation behaviour within the small sample sites such as Bigbury Outer, that is either natural or due to human influences.

2.4.3 Ontogenetic shifts in habitat associations Differences in the depth of capture between age classes were found in three out of the four species. In R. montagui and R. brachyura, immature individuals were caught at significantly shallower depths than mature individuals. No significant difference was found between adult and juvenile depth of capture in R. microocellata, while juvenile R. clavata were caught at deeper habitats than adults. Generally, previous studies have suggested that these four Rajidae species have nursery grounds shallower than adult preferred habitats (Ellis et al., 2005a). This concurs with the current results for two of the four species. For R. clavata however, the current study is different to previous evidence. It may be the case that different populations of R. clavata exhibit different strategies with regard to egg laying, or, simply that environmental factors, such as substrate type, are more important for the deposition of eggs than depth. It is important to note that the catchability of juveniles, especially neonates, is likely to be greatly reduced in any fishing survey, due to the mesh size allowing the escape of small fish and thus results should be interpreted cautiously. In addition, unlike the current study, research trawls rarely take place in rivers, which likely misses sampling an important habitat for R. clavata adults.

2.4.4 Sexual segregation in Rajidae Results indicate significant sex biases towards female in three out of four species for all trawls combined. This tendency to capture females in trawls has been previously noted, at least as early as the late 1800s (Steven, 1933). Off Plymouth for example, numerous occurrences of female sex bias in Rajidae spp. were recorded with many examples of hauls containing 100% female Rajidae (Steven, 1933). Sex ratio bias in trawls was also demonstrated in R. clavata in the Bay of Douarnez, France, where trawls were predominantly male inside the bay, whilst outside the bay females were predominant in the summer (Rousset, 1990). Conversely, no sexual segregation of R. clavata was detected during another study focusing on the English Channel, though it was suggested that sexual segregation may occur at finer scales than the surveys conducted (Martin et al., 2010). Four hypotheses could be considered to explain the female sex bias in the current

58 study. First, there may be a greater proportion of females than males in the population, although sex ratio at birth is considered to be 1:1 (Ellis and Shackley, 1995). Second, there may be differential juvenile survival rates between males and females, which translates into a bias in the adult population; it has for example been demonstrated that males are more likely to die after capture in a trawl than females (Enever et al., 2009). Third, differences in male/female behaviour may also cause differences in catch rates. Male plaice (Pleuronectes platessa) for example are more active than females during their spawning season which has been suggested as an explanation for higher catches of male plaice (Solmundsson et al., 2003). Finally, males and females may be sexually segregated and therefore one sex is captured in greater numbers depending on the fishing location. Sexual segregation is a common occurrence in elasmobranchs (Mucientes et al., 2009, Klimley, 1987, Sims, 2005, Sims et al., 2001, Domeier and Nasby-Lucas, 2012) and is therefore most likely the main driver behind the current findings. Given the persistence of studies finding a female bias in trawls since the 30s and the ubiquitous nature of sexual segregation in elasmobranchs, spatial segregation seems to be the most plausible driver of the current findings.

2.4.5 Management implications Differential habitat association by species, age class or sex has important implications for management, for example, if females are more easily captured due to sexual segregation, population declines can be exacerbated if fishing is focused on those habitats. This is illustrated by the collapse of the spurdog Squalus acanthias fishery in south west UK, which has been attributed in part to the bias in capture of females in the late 1920s (Wearmouth and Sims, 2008). It is also likely that, due to differential habitat association, certain species may be more vulnerable to capture than others. Hypothetically, R. clavata preference for estuarine habitat may provide protection from larger trawlers that work offshore, while R. brachyura prefer to occupy deeper areas more prone to intense trawl fisheries. Due to the practice of grouping Rajidae spp. as ‘skate and ray’ in fisheries landings, this bias toward catching a particular species can mask potential declines in the species studied. For instance, the decline of the common skate (Dipturus batis complex) (Griffiths et al., 2010) was not identified until the 1980s when the species was already considered locally extinct (Brander, 1981).

2.4.6 Conclusion Collectively the results presented here indicate differential inter- and intra-species occurrence in different areas off Plymouth and that distributions appear complex at relatively fine spatial and temporal scales. Over a study area that is small, relative to other research surveys such as those conducted by IBTS, there was a significant difference in the occurrence of the four studied species

59 within and between the study sites. Juveniles and adults were caught at different depths and there were also biases in sex ratio in trawls for three out of four species.

Chapter 3. The spatial ecology of Rajidae from mark-recapture tagging and its implications for assessing fishery interactions and efficacy of Marine Protected Areas

3.1 Introduction The exploitation of the ocean for fish has impacted the marine environment for thousands of years, with many fish stocks driven to low levels during the last few decades (Pauly et al., 2002). With the industrialisation of the fishing industry in the 20th century, improved technology allowed more efficient extraction of fish, as well as access to previously unexploited fishing grounds (Pauly et al., 2002). Unsustainable fishing has led to the decline of many of the oceans larger, k-selected species, especially elasmobranchs (Fernandes et al., 2017, Pauly et al., 1998, Brander, 1981). K-selected species are considered to be vulnerable to overfishing because they are late maturing and have relatively low fecundity (Dulvy et al., 2014, Field et al., 2009). To stop, or indeed reverse declining trends, fisheries managers employ a wide range of measures, such as quotas, minimum landing sizes, closed areas and/or marine protected areas (Cochrane, 2002). All of these management measures require an understanding of target or bycatch species and their ecosystem ecology, as well as information about fisheries that target them. This is by no means an easy task given the scale of global fisheries.

Often the ecology of a target species and the behaviour of the fishers are inextricably linked. Fishers, for example, know from previous experience and shared knowledge (often cultural) where certain species are found and the types of fishing gear that will best capture them. For instance, a recent study found that there was significant overlap between the space use hotspots of blue sharks Prionace glauca and fishing vessels in the north-east Atlantic, with co-occurrence near oceanic fronts (Queiroz et al., 2016). An understanding of where such hotspots occur can be used as a focus for the management of this species. Importantly however, different species exhibit unique ecological traits which will influence their probability of capture (Alós et al., 2012, Gallagher et al., 2014, Cooke et al., 2013b). This can lead to the exploitation of certain species in a sympatric assemblage, while others are relatively unaffected. Understanding the movements and spatial dynamics of species distributions can aid management, for instance by closing fishing grounds that encompass the movements of a species. In addition, species that exhibit complex spatial movements or show site fidelity raise unique challenges for management of the species

61 and consequently accurate data are required to implement effective management (Bass et al., 2016).

3.1.1 The UK Rajidae fishery The main elasmobranch fisheries in the UK are for the Rajidae, of which there are 14 species in UK waters (Silva et al., 2012). Rajidae, like other elasmobranchs, are late maturing and have low fecundity and are therefore vulnerable to over-exploitation (Dulvy et al., 2014, Field et al., 2009). Accurate monitoring and management of these species is therefore essential. In the early 1800s skates were not commercially viable in the UK, but towards the end of the century and the beginning of the 20th century, skate fisheries began to expand (Steven, 1932a, Dulvy et al., 2000). By 1929 the Devon and Cornwall skate fishery was very valuable as a relatively cheap source of animal protein for human consumption, with skates providing 30% of the total value for demersal fishing (Steven, 1932a). This increase in demand and value of Rajidae species therefore increased the necessity for increased monitoring and management.

3.1.2 Monitoring and management Monitoring of the main species fished by commercial fisheries is conducted on a global scale by the Food and Agriculture Organisation (FAO) (Fisheries and Aquaculture department, 2017) and by the Marine Management Organisation (MMO) and Marine Scotland in the UK (MMO, 2014). However, until 2009 the main commercially valuable Rajidae were grouped together as ‘skates and rays’ in UK fisheries data (Silva et al., 2012). This has hampered an ability to understand the fishery and the species they catch and therefore apply appropriate management. It is likely that poor recording of species landings masked severe declines in some species, such as Dipturus batis and Raja clavata in certain regions (Walker and Heessen, 1996, Dulvy et al., 2000, Brander, 1981, Chevolot et al., 2006).

In the UK, commercial fisheries use demersal trawls, longlines or static nets to catch Rajidae (Silva et al., 2012) and therefore management of the fishery can be complex. There are few management measures in place for Rajidae aside from the prohibition of some endangered species, such as the common skate D. batis (CEC, 2011). There are, for example, no minimum landing sizes for Rajidae except within the Kent and Essex IFCA region (MMO, 2017b, KEIFCA, 2017). In the most recent fisheries quotas, in most cases Rajidae are still given a grouped quota (MMO, 2017a) despite fishers landing Rajidae as individual species since 2009 (Silva et al., 2012). Recent analyses however indicate the accuracy of this data remains questionable (Simpson and Sims, 2016).

Marine protected areas (MPAs) are another method used for the conservation and management of marine species in the UK. MPAs placed within UK territorial waters are areas of sea and coast that are protected by legislation from damage and disturbance and there are a variety of MPAs with a range of protection levels and codes of conduct. For example, within the study site investigated here (Plymouth UK), there are primarily areas restricting mobile fishing gear such as trawling (Figure 14), but static gear such as netting is permitted (IFCA, 2014). Furthermore, although the Whitsand and Looe Marine Conservation Zone (MCZ) has been designated, to date there are no bylaws to restrict activities; a restriction on mobile gear has been proposed but currently has not been passed (IFCA, 2017). Largely these protected areas were designated to protect reef habitats and sessile organisms such as the pink sea fan Eunicella verrucosa (JNCC & Natural England, 2016). The restriction of mobile gear has been shown to be effective in conserving these sessile organisms (Pikesley et al., 2016, Howarth et al., 2015). However, the efficacy of these MPAs for protecting mobile species is largely unknown. In addition, the gear types that are permitted within a MPA, such as static netting, may still capture large numbers of mobile species. To assess the efficacy of MPAs and other management techniques for the conservation of Rajidae, an understanding of the fishery and of their ecology, especially movement ecology, is fundamental.

Figure 14: Marine protected areas within the study area. 1 The Lizard, 2 Plymouth Sound and Estuaries, 3 Fal and Helford, 4 South Devon Shore Dock, 5 Blackstone Point, 6 Polruan to Polperro, 7 Lyme Bay and Torbay, 8 Lizard Point, 9 Start Point to Plymouth Sound and Eddystone. Red area is the designated Whitsand and Looe Bay MCZ.

3.1.3 Rajidae ecology The grouping of Rajidae in UK landings data is likely due to their broadly similar morphology and their general coastal sympatric distribution (Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010). However, Rajidae species are known to exhibit different

63 ecological traits, such as age at maturity (Gallagher et al., 2005, McCully et al., 2012), fecundity (Koop, 2005, Holden, 1975), egg laying seasons (Koop, 2005) and, importantly, their spatial preferences and site fidelity within the coastal area (see Chapter 1) (Humphries et al., 2016c). These differences in the ecology of each species, as well as the behaviour of the fishery are likely to have an effect on species catchability and consequent vulnerability.

Previous studies also suggest that Rajidae make seasonal migrations. R. clavata in the North Sea for example, moved from deeper waters in the autumn and winter to shallower areas during the spring (Hunter et al., 2005b). This seasonal movement also has the potential to affect catchability at different times of the year, especially at times when individuals aggregate. In addition, in Chapter 1 the four species were also captured seasonally at particular sites. An understanding of these seasons of increased catchability can be used to inform management measures, such as temporal closures of a fishery (Guénette et al., 1998).

To further understand the behaviour of the Rajidae fishery and the spatial ecology of the species themselves, the current study used the capture-mark-recapture method using long-lasting numbered tags attached to individual skates. This method has been used for over a hundred years to track movements and migrations, growth rates, abundance and mortality of animals including fish (Kohler and Turner, 2001, Bolle et al., 2005, McFarlane et al., 1990, Fulton, 1893, Steven, 1936, Ellis et al., 2011). Mark-recapture studies of fish often rely on the recaptures made entirely by the commercial fishery. This attracts the criticism that the recaptures reflect the distribution or behaviour of the fishing fleet, rather than the true distribution of the species in question. Recaptures are unlikely to be made in areas where there is low effort (Bolle et al., 2005) and particular gear types may capture particular species more efficiently than others (Gunderson, 1993). Despite these limitations data can provide information about the behaviour of the fishery, as well as differences in ecology between the species (Quinn and Brodeur, 1991).

The current study aims to understand the Rajidae fishery off the coast of Plymouth UK, using mark-recapture of four Rajidae species in the area (Raja clavata, R. brachyura, R. microocellata and R. montagui) in relation to fisheries and MPAs. I hypothesise that due to differences in spatial occurrence found in Chapter 2, the four species will vary in their interaction with the fishery and that the efficacy of the MPAs will be different for each species because of their underlying ecological differences and catchability. The potential differences in location of capture, mobility between species and differences in seasonal catchability throughout the year are investigated. In addition, the gear types used to catch different species, the proportion of each species

64 recaptured, the rate of recapture and the difference in the number of days at liberty between species are examined.

3.2 Method Four common Rajidae species, Raja clavata, R. brachyura, R. microocellata and R. montagui were caught, tagged and released from 2009 to 2016 using beam or otter trawls off the coast of Plymouth SW UK. Vessels were either the Plymouth Quest or RV Sepia and all tows lasted less than 40 minutes. Serially numbered Peterson discs (10 mm width) were attached to skates via tagging using the methods described by Wearmouth & Sims (2009a). Briefly, tags were secured in stainless steel wire with a pointed stainless steel wire attachment that was passed through a Peterson disc and then through the centre of the pectoral fin. A second Peterson disc was placed onto the length of wire on the opposite (ventral) side of the skate before turning a series of rounds into the length of wire remaining to secure it. Tagging was completed in less than two minutes, including weighing and the measurement of total length and disc width. Skates were immediately transferred to on-board aquaria with fresh running sea water for observation prior to release. A total of 439 skates were tagged and release location was recorded. Tagged skates were returned through the local fishery for rewards and fishers provided recapture location. Not all fishers could provide latitude and longitude and these recaptures were removed from the analysis.

3.2.1 Ethics statement All tagging procedures were approved by the Marine Biological Association Animal Welfare and Ethical Review Body (AWERB) and licensed by the UK Home Office under the Animals (Scientific Procedures) Act 1986.

3.2.2 Analysis To test if recapture numbers were equal across species and to compare the rate of recapture (number recaptured within x days at liberty), a Chi-Squared Goodness of fit test was used with expected values calculated from the number of individuals released for each species. To further consider the rate of recapture, the percentage of a species recaptured was plotted against days at liberty. Differences in the number of days at liberty between species were tested with a Negative Binomial Regression Model in R (Version 1.0.143) used to account for over-dispersed count data.

To visualise the distribution of recaptures for each species, release and recapture positions were plotted in ARC GIS (ARC MAP 10.2.2) along with one second mosaic bathymetry (Department for Environment, Food & Rural Affairs DEFRA). Kernel density estimations were determined with release and recapture positions to provide an indication of spatial aggregation of recapture locations for each species. To further test for differences in species distribution and movement behaviour, depth of recapture was interpolated from the bathymetry layer and displacement between release and recapture was measured as the crow flies. Differences in the depth of recapture between species were analysed with ANOVA on Ranks (Kruskal Wallis). The proportion of recaptures caught at <10km, between 10 and 30km and >30km from release were also calculated for each species. These categories were selected to reflect the findings of other studies, which found that the majority of skates were recaptured within their release site (<10 km), fewer individuals move distances of 30 km and very few individuals move distances greater than 60 km (Steven, 1936, Ellis et al., 2011). To test for seasonality of recaptures, a Chi-Squared Goodness of fit test was used. Seasons were designated to match local seasonal temperature shifts as follows: spring, February, March and April; summer, May, June and July; autumn, August, September and October; winter, November, December and January. Expected proportions were used based on the number of individuals at liberty in each season.

From data on the method of recapture, a Chi Square Goodness of fit test was used to test if the four species were caught equally by trawl, with expected values calculated from the number of a species released.

To assess the efficacy of the current spatial management within the study site, ARC GIS was used to calculate the proportion of recaptures made within the boundary of a MPA for each species, as well as the methods of recapture. I also calculated the proportion of recaptures that would not have been made if the Whitsand and Looe MCZ bylaw was in place, allowing an assessment of the potential efficacy of restricting mobile gear for the conservation of Rajidae. In addition I investigate the number of recaptures that would not have been made if additional management strategies were in place, such as restricting static netting or additional MPAs.

3.3 Results Between 2009 and 2016 a total of 439 skates were tagged and 156 were returned (35.53%) after a maximum time at liberty of up to 1052 days for a R. clavata (Table 9). There were no significant differences in days at liberty between species (Negative Binomial Regression, df= 3, P=0.18) and

there were no species differences in the number of individuals recaptured (N=154, df=3, 2=2.83, P=0.41). However, the rate of recapture differed between species. R. brachyura had the fastest rate of recapture and R. clavata the slowest (Figure 15). Chi-Squared goodness fit tests suggested differences in the rate of recapture (N=752, df=3, 2=192.309, p<0.001) with 20% of R. montagui and R. clavata caught and returned within approximately 280 days, while R. brachyura had the fastest rate of recapture with 20% caught within 60 days. In addition, R. montagui had an initial burst of recaptures similar to the other species, but thereafter the rate decreased, only to be followed by a further burst of recaptures between 281 to 354 days post-release.

Figure 15: Left- total number recaptured during the study period. Expected numbers are based on the total number released. There was no significant difference between species in the number released. Right- the rate of recapture, where R. brachyura are recaptured more quickly than the other species. Note the increase in recaptures of R. montagui after 281 days.

Table 9: Quantifying the sampling effort and return rates

Species Total Total % Average Standard

released recaptured recaptured days at deviation liberty of days at liberty R. brachyura 92 40 43.48 131.58 172.28 R. clavata 155 45 29.03 160.49 141.39 R. microocellata 90 33 36.67 170 207.01 R. montagui 102 38 37.26 220.2 162.08

Total 439 156 35.54

3.3.1 Rajidae spatial ecology Release and recapture positions were similar for each species and these locations were not evenly distributed across the area. Instead, spatial aggregations of recapture (hotspots) occurred but were different among species based on kernel density estimations. R. brachyura were caught in a focussed area around the outer Bigbury Bay area (Figure 16), as were R. montagui (Figure 19). R. microocellata were recaptured primarily in an area to the west of Plymouth Sound in Whitsand Bay and the inner section of Bigbury Bay (Figure 18). R. clavata was the only species to have recapture hotspots inside Plymouth Sound at Cawsand Bay (Figure 17). The four species, split into two groups with regard to depth of recapture (Figure 20). R. montagui and R. brachyura were recaptured at significantly (H=47.22, df=3, P<0.001) deeper depths (median 31.5 m and 32 m respectively), than R. clavata and R. microocellata (median 12 m and 13.5m respectively).

Figure 16: Kernel density estimation of R. brachyura, Top= Release, Bottom=Recapture. Hotspots of recapture were in the outer section Bigbury Bay.

Figure 17: Kernel density estimation of R. clavata, Top= Release, Bottom=Recapture. Hotspots of recapture were in the estuary, in the Cawsands Bay area.

Figure 18: Kernel density estimation of R. microocellata, Top= Release, Bottom=Recapture. Note the wider dispersal of recapture hotspots in this species. Recaptures were primarily in the Whitsand Bay area.

Figure 19: Kernel density estimation of R. montagui, Top= Release, Bottom=Recapture. Hotspots of recapture were in the Bigbury Bay area.

Figure 20: Differences in depth (m) of capture between species. Box represents 25th and 75th percentile, whiskers 5th 9th percentiles and middle horizontal line is the median. R. clavata and R. microocellata were caught at significantly shallower depths than R. brachyura and R. montagui (P<0.001).

Individual R. microocellata were recaptured further from their release point, given days at liberty, than R. clavata and R. montagui (Kruskal Wallis ANOVA on Ranks, H=14.95, df=3, P=0.002) (Figure 22). Higher percentages (18.2%) of R. microocellata were caught >30 km from their release point, while R. clavata were caught no further than 20.09 km from their release point (Figure 21). There were several individuals of interest, including one R. clavata that was at liberty for the longest period of 1052 days, yet only moved 1.48 km from its release point. Furthermore, a single R. microocellata moved the greatest distance of 82.6 km in 466 days and one R. montagui that exhibited the fastest rate of dispersal and moved 1.01 km per day at liberty over 69 days (Table 10).

100%

90%

80%

70%

60%

50% >30km 10-30km Percentage 40% <10km 30%

20%

10%

0% Raja brachyura Raja clavata Raja microocellata Raja montagui Species

Figure 21: Proportion of recaptures at <10km, 10-30km and >30km from release. R. clavata were predominantly caught within 10km of where they were released, while in R. microocellata a higher proportion were caught greater displacements.

Figure 22: Displacement as the crow flies. Green=release, red=recapture. A- R. brachyura, B- R. montagui, C- R. clavata, D- R. microocellata. Note the difference between C and D, where R. clavata (C) dispersed less than R. microocellata (D).

Table 10: Displacement metrics for each species

Species Median displacement Maximum Distance between (km) displacement (km) furthest recaptures (km) R. brachyura 6.5 43.19 71.03 R. clavata 5.3 20.09 27 R. microocellata 13.24 82.6 134.83 R. montagui 4.48 69.73 96.81

3.3.2 Seasonality of recaptures Tagging effort differed across seasons and between species. R. brachyura and R. montagui were predominantly tagged in summer, R. clavata were tagged mainly in summer and winter, while for R. microocellata tagging effort was higher in winter and spring. Recaptures differed significantly across seasons in every case (Table 11). Recaptures were made more frequently in summer for R. clavata and R. microocellata, while R. montagui were recaptured more frequently in spring and R. brachyura had higher recaptures in autumn (Figure 23).

Table 11: Chi-Squared results of differences in season of recapture

Species N df 2 P Peak season of recapture R. brachyura 43 3 7.21 0.066 Autumn R. clavata 97 3 14.97 0.002 Summer R. microocellata 56 3 7.84 0.049 Summer R. montagui 43 3 10.44 0.015 Spring

18 40 Expected Expected 16 A B Observed Observed 14

d 30 d

e e

r r

u 12 u

t t

p p

a 10 a

c c

e e

r 20

r 8 r

e e

b b

6 m

u u

N 4 N 10

0 0 Category Spring Summer Autumn Winter Category Spring Summer Autumn Winter

20 Expected 20 C Expected Observed D Observed

s s

e e

r 15 r 15

u u

t t

p p

a a

c c

e e

r r

f 10 f 10

o o

r r

e e

b b

m m

u 5 u

N 5

0 0 Category Spring Summer Autumn Winter Category Spring Summer Autumn Winter Figure 23: Seasons of recaptures. A. R. brachyura B. R. clavata C. R. microocellata D. R. montagui

3.3.3 Method of recapture R. brachyura was predominantly recaptured by trawl, while R. clavata were captured by static nets (Figure 24). Overall 56.94% were re-captured by trawl, 40.28% by nets, 1.39% by longline and 1.39% by angling, however R. microocellata was the only species to be captured by longlines and anglers.

100% 90%

80%

70% 60% 50% Trawl 40% Nets

30% Longline Proportioncaptured 20% Angling 10% 0% Raja brachyura Raja clavata Raja Raja montagui microocellata Species

Figure 24: Proportion of species captured by fishing gear type. Note the high proportion of R. clavata caught in nets relative to R. brachyura which were more frequently caught by trawl.

3.3.4 Recaptures in MPAs Results show that 37% of recaptures were located inside MPAs (Figure 25) and this proportion differed between species (Table 12). These recaptures were predominately made by static nets and secondarily by trawls (Table 12). All of those caught by trawl were within the boundary of the Whitsand MCZ, which had been designated but the bylaw had yet to be passed. If the bylaw had been passed 12% of recaptures would not have occurred at that time. Furthermore, static netting within MPAs accounted for 24% of recaptures, which, if restricted, would have decreased recapture by almost a quarter. An area between two protected areas (Figure 26) appears to be a hotspot for Rajidae, especially R. montagui and R. brachyura. If this area was also closed to trawls and netting a further 28% or 45 individuals would not have been recaptured within the MPA.

Figure 25: Rajidae recapture locations. MPA boundaries marked in blue.

Table 12: Proportion of a species recaptured within the boundary of a designated MPA

Species % of recaptures inside a protected area R. brachyura 18 R. clavata 53.33 R. microocellata 39 R. montagui 34.2

Table 13: Of those recaptured in an MPA, the proportion of different gears types

Method Percentage Count

Longline 1.75% 1 Nets 64.91% 37 Trawl 33.33% 19

Figure 26: Potential new MPA (yellow) containing 28% of all recaptured Rajidae species.

3.4 Discussion Variation in the spatial ecology of sympatric species in addition to variation in the behaviour of the fishery can cause differences in catchability between species, resulting in a requirement for species-specific management measures (Bertrand et al., 2004, van Oostenbrugge et al., 2008, Ulrich et al., 2002). In the current study, mark-recapture data from four sympatric species returned by the commercial fishery were used to investigate skate spatial ecology, the interaction between fish and fisher and the efficacy of current spatial management.

The overall recapture rate in the current study was relatively high (35% overall) compared to other elasmobranch studies. For example, a review of 52 tag and recapture studies, containing 72 species of shark, 55% of reported recapture rates were <5% (Kohler and Turner, 2001). However other studies of Rajidae in the UK are more variable. For example, a tagging study of Rajidae in the western English Channel near the Island of Jersey reported a 17.1% return rate (Ellis et al., 2011). Meanwhile, a R. clavata tagging study in Thames area reported a 51% return rate (Hunter et al., 2006) and a tagging study in the Plymouth area reported a 33% return rate (Steven, 1936). There are multiple causes of variation in recapture rates between such varied species, which are reviewed in Kohler and Turner (2001). However there are three main causes for the recapture rates in the Plymouth area. First, despite relatively low commercial fishing effort in the shallow coastal areas, recaptures may reflect the high fishing pressure in the Plymouth area compared to most of the UK (Figure 27). The fishing pressure in the Thames area is even higher and may reflect the high number of recaptures in Hunter et al. (2006). Second, tagging has been conducted in the area since at least the 1930s (Steven, 1932a) and fishers may be more aware of the reward system in place compared to other areas and are therefore more likely to return recaptured tags.

Third, Rajidae are generally considered bycatch in demersal trawls (Silva et al., 2012), although in the Plymouth area there may be some seasonal targeting of skates. For instance, some of the fishers returning tags in this study stated that they target skate and refer to their static gear as ‘ray nets’ (S. Simpson, personal communication). This targeting of skates may lead to higher recaptures than fisheries in which elasmobranchs are bycatch. It is likely that a combination of these factors have resulted in the recapture rate found here.

Figure 27: Relative fishing effort by trawls (sightings per unit effort). Data and method found at Vanstaen and Breen 2014.

3.4.1 Rajidae spatial ecology The four Rajidae species at the focus of the current study were not captured equally across the study site and kernel density estimation indicates that the four species’ recaptures were aggregated in different areas (i.e. species recapture hotspots). R. brachyura and R. montagui were recaptured predominantly in an overlapping area within Bigbury Bay, while R. microocellata were recaptured primarily in Whitsand Bay and R. clavata within Plymouth Sound and Whitsand Bay. In

78 addition, the four species were caught at different depths; R. microocellata and R. clavata were recaptured in shallower areas than R. montagui and R. brachyura, with the four species grouping into two pairs. This result supports a recent tagging study that used electronic depth-logging tags in the same area, which also found these four species split into two pairs in terms of their depth occupancy (Humphries et al., 2016c). In the eastern North Pacific, a skate assemblage containing five main species were also found in areas of high aggregation, that were distinct among species, and depth was an important characteristic of their preferred habitat (Bizzarro et al., 2014). Therefore, kernel density estimation, depth of capture and the previous tagging studies collectively demonstrates that habitat partitioning was exhibited among the four sympatric species in the current study.

Rajidae recapture locations were very similar to release locations, suggesting that these individuals did not move very far while at liberty. Indeed, although the maximum distance moved by any individual was 82.6 km, the highest median displacement for any species was only 13.24 km. The current results are similar to those in a previous mark and recapture study monitoring four rajids (Raja brachyura, R. clavata, R. microocellata and R. undulata) where the majority were recaptured within 40 km, with a maximum displacement of 61 km for a single R. brachyura (Ellis et al., 2011). The problem with using the mark-recapture method is that it cannot be known what the movements of an individual between these two known points were, or the fate of individuals that are not recaptured. It is possible that a small number of individuals in the population may be more mobile (Vøllestad et al., 2012, Rodríguez, 2002) but were not recaptured because they moved away to areas outside the operation of skate fisheries. A genetic study of R. clavata found little differentiation between distant populations, suggesting that some individuals must migrate over wider areas (Chevolot et al., 2006). However, with such a high proportion of rajids recaptured so close to their release site, the current study provides additional evidence for site fidelity of skates to small-scale areas in habitats they may actively select.

Though all species exhibit site fidelity, the degree of dispersal between tagging and recapture was different between species. R. clavata for example were caught no more than 20 km from their release point which supports previous studies suggesting high site fidelity of this species in the Plymouth region (Steven, 1936). Conversely, R. microocellata had a higher displacement (up to 82.6 km straight line distance), with 18.2 % of the recaptures being >30 km from the release site. Given that there were no significant differences in days at liberty between the species and no significant difference in number recaptured, it appears that R. microocellata individuals may disperse further than R. clavata and is indicative of a higher degree of population mixing. Further

79 genetics studies, similar to Chevolot et al. (2006), sampling R. microocellata around the coast would help to examine the question of wider dispersal and population mixing in this species.

3.4.2 Seasonality of recapture Results from the current study suggest that the four rajids had different seasons where recaptures were higher. Two out of the four species had higher rates of recapture during the summer, while R. brachyura were more frequently recaptured in the autumn and R. montagui in the spring. The increased recapture of R. clavata in the spring and summer, predominantly in the estuary area, is similar to a study that found R. clavata were migrating into the Thames estuary during the spring (Hunter et al., 2005b). There is likely to be a bias in commercial fishing effort at particular times of year, for example in the summer when conditions are improved for more frequent fishing. Nonetheless we found differences in captures in the same areas at the same time in different species. For instance, R. montagui and R. brachyura were caught and tagged at the same site in summer, R. brachyura were then recaptured in autumn indicating commercial fishing effort at the site. R. montagui however had the highest season of recapture in spring, long after the peak release season. This is also reflected in the rate of recapture, where after an initial burst of recaptures the rate slows, before increasing again after 281 days. It is therefore more likely that these seasonal recaptures reflect movement of the population to and from particular sites that are seasonally exploited. This observation was further supported by Steven (1932a), who also suggested that R. clavata make inshore movements in the spring. Based on these results temporal management of these species may be an efficient method to ensure protection at important times of aggregation.

3.4.3 Method of recapture Overall, the four species were not recaptured by the same gear types, most clearly illustrated in the difference between R. clavata, which were recaptured predominantly by static nets, while R. brachyura were more frequently recaptured by trawl. These results are supported by a previous study in the UK, which also suggested R. brachyura was captured more frequently by beam trawls whilst R. clavata was more predominant in nets (Silva et al., 2012). This is likely due to the aforementioned habitat partitioning in these species. R. clavata are associated with the estuary area where mobile gear is prohibited in Plymouth and most other estuaries in the UK, and are therefore more likely to be caught in static nets. In contrast R. brachyura are captured in deeper locations where trawling is perhaps more likely. Understanding that Rajidae species are captured by different gear types may increase the complexity of management, but it is important to implement the most effective way to conserve these threatened species. For example MPAs restricting mobile gear to protect reef habitats and sessile organisms such as the pink sea fan

Eunicella verrucosa (JNCC & Natural England, 2016) have resulted in the effective conservation of these organisms (Pikesley et al., 2016, Howarth et al., 2015). Prohibiting or management of vessels with different gear types can therefore offer different protection for each skate species and could allow more targeted and efficient management of vulnerable skates.

3.4.4 Efficacy of existing MPAs for Rajidae Habitat partitioning between these four species has the potential to influence their catchability and the efficacy of management measures such as MPAs. Overall 37% of recaptures were made inside designated MPAs, but this proportion was different between species. R. clavata for example were caught more frequently in MPAs (53%) than R. brachyura (18%). These results provide important insights into the potential efficacy of MPAs. First, MPAs off Plymouth do not eliminate the capture of skates due to the continued use of static gear. For instance, 65% of recaptures within MPAs were made with static nets, which are legal within these MPAs off Plymouth (IFCA, 2014). If static gear were also restricted in these areas, a further 24% of recaptures would not have been made. Second, the prohibition of trawling could significantly reduce the chance of catching skates for species that occupy an area within its boundaries (Hunter et al., 2006). If for example the planned bylaw restricting mobile gear in Whitsand and Looe MCZ was to be passed, 12% of recaptures would not have been made. It is however important to note that by restricting mobile gear in the area, the use of other static gear may increase due to the decrease in likelihood of damage by mobile gear. Indeed fishers using static gear tend to show a positive attitude towards MPAs because they are less affected or have more access to the fishing grounds (Pita et al., 2013). Third, efficacy of the local MPAs is different for each species; R. brachyura were caught in equal numbers to the other species but caught most frequently outside the boundaries of MPAs. Worryingly the R. brachyura has the largest body size among species found off Plymouth, as wells as the highest value and are therefore perhaps the most vulnerable to overfishing. Interestingly the gap between the Start Point to Plymouth Sound MPA (Bigbury area) was a hotspot for recapture for skates, especially R. brachyura and R. montagui. If this area was also closed to mobile and static gear a further 28% of recaptures would not have been made. Other studies have also provided tailored advice based on the spatial preferences of the flapper skate Dipturus intermedia on the west coast of Scotland (Neat et al., 2014), demonstrating that MPAs could be used to protect mobile species that show site fidelity.

These results suggest that MPAs in this area have the potential to further protect these mobile but philopatric species, with simple changes to the areas already designated and by expanding existing areas. If, for example, the Whitsand and Looe MCZ bylaw restricting mobile gear was to be passed, and netting in designated MPAs is prohibited in addition to this additional Bigbury area

81 closure, 65% of recaptures would not have occurred in these areas. Individuals may of course move out of these protected areas, although the philopatric nature of these species and the reduced disturbance makes this less likely.

While MPAs can be beneficial, some argue that they are not the solution for every species, especially mobile species (Kaiser, 2005). There is evidence to suggest that spatial closures simply displace the fishery or move the problem elsewhere (Halpern et al., 2004, Abbott and Haynie, 2012). However there is also evidence to suggest that spill over from recovering populations within the boundaries of MPAs can benefit the surrounding unprotected areas (Halpern et al., 2009). A further drawback of protecting a particular species with spatial closures is that other fisheries and stakeholders in the area can be affected (Agardy et al., 2011). Temporal spatial closures can be used by fisheries managers to resolve this conflict between stakeholders, protecting species at critical times, while allowing fishers to exploit the area at other times. This approach was also explored by (Hunter et al., 2006) that suggested seasonal spatial closures would result in significantly less landings of R. clavata. The results for the current study for example, suggest that for the established MPAs in the area, restricting static gear during the spring could be an optimal method of protecting R. clavata and R. montagui. Moving forward, management that recognises species differences in spatial ecology and requires implementation of species-specific landings and data recording will ultimately help conservation of these vulnerable species.

3.4.5 Conclusion In summary, four rajids off the coast of Plymouth were found to exhibit apparent habitat partitioning and differences in movement pattern that had an impact on their commercial catchability. Each species were found in different habitats, with different displacements, were recaptured in different seasons and by different gear types. Management of these four species therefore requires different approaches, and existing management, in this case MPAs, varied in its efficacy depending on the species. Overall existing MPAs did not stop the commercial capture of any species within their boundaries; they do however likely reduce the capture of rajids, especially R. clavata. Efficacy could be increased by restricting static nets and expanding the boundaries into Rajidae capture hotspots. R. brachyura is perhaps the least likely to benefit from existing MPAs with its deeper habitat occupancy indicating that R. brachyura habitat falls outside current MPAs.

Chapter 4. Investigating fine-scale habitat selection, spatial partitioning and sexual segregation in Rajidae using passive acoustic telemetry

4.1 Introduction Habitat association simply links the spatial occurrence of a species to a particular feature of the environment, such as depth. Inferring habitat selection however, requires the observation of an individual though space and time to provide evidence of an individual making active decisions to remain or move towards a particular habitat. Identifying habitat selection is particularly challenging in the marine environment, where direct observation is often difficult and where species are cryptic, do not breach the surface and are often too small for large tags (Hussey et al., 2015, Cooke et al., 2013a). In the previous two chapters, habitat associations and apparent partitioning among four species of Rajidae from research surveys and recapture data were identified. These findings were limited to establishing habitat associations because research survey data for instance can only provide one known location in time, whilst mark and recapture provides two points in time. Hence, with these data alone it is difficult to demonstrate active habitat selection (Figure 28).

Fishing data Mark-recapture Acoustic data telemetry

Figure 28: The line is the animals track, the red dots are known locations. How much of an individual’s movement can we sample using three different techniques? Fishing survey data can only provide a single known location and we cannot infer active selection, while a passive acoustic array can detect movement at a finer scale and therefore active movement towards particular areas.

The current chapter uses passive acoustic telemetry to track tagged individuals resulting in multiple known locations through time and allowing for a more detailed exploration of habitat selection and resource partitioning. Passive acoustic monitoring using a spatial array of acoustic receivers allows observers to collect fine spatial scale, long-term movement data without the need for the physical recovery of the tags themselves. Stationary hydrophone receivers are placed on the seabed or suspended in the water column and can detect the presence or absence of animals tagged with acoustic transmitters within a detection range. This range is variable depending on the environmental conditions surrounding the receiver, including water depth, water flow and even biofouling on the receiver (Kessel et al., 2014). Espinoza et al. (2011) for example, reports a maximum detection range of radius of 350-900 m in an estuary, while Melnychuk and Walters (2010) reports a typical detection range of 400-500 m in river, estuary and marine conditions. Acoustic arrays provide a higher resolution of distribution through time than conventional tagging (e.g. Chapter 3), are only limited in resolution by the number of receivers placed in the study area, but they also enable tracking of multiple individuals at the same time. In addition, once the animals are tagged and the array is deployed, there are fewer requirements for researchers physically to remain at the site, unlike for active tracking for example (Heupel et al., 2006).

Previous studies have successfully demonstrated the use of passive acoustic receiver arrays to identify habitat selection in multiple mobile species. Tiger sharks Galeocerdo cuvier and Galapagos sharks Carcharhinus galapagensis have been tracked using acoustic telemetry and it was suggested that tiger sharks occupied shallow lagoons more than Galapagos sharks (Meyer et al., 2010). The flapper skate Dipturus intermedia were also tracked using an acoustic array and it was revealed that over half of the tagged individuals were resident to the study site over a period of months (Neat et al., 2015). The Rajidae at the focus of the current study are cryptic species, which bury in the sediment for extended periods of time (Greenway et al., 2016) and inhabit coastal turbid waters (Steven, 1936). These attributes, in addition to the site fidelity exhibited by the Rajidae (Ellis et al., 2011), suggests that these species may be ideal for using passive acoustic telemetry to help understand spatial dynamics, fine-scale habitat selection and resource partitioning.

In addition to apparent spatial partitioning between rajid species off Plymouth, Chapter 2 also identified a biased sex ratio towards females in catches which suggests the occurrence of sexual segregation. Sexual segregation is ubiquitous across the animal kingdom and is defined as the separation of a species by sex, either as individuals or in groups (Wearmouth and Sims, 2008). Sexual segregation in elasmobranchs is well documented in a number of species including white

84 sharks Carcharodon carcharias (Domeier and Nasby-Lucas, 2012), catsharks Scyliorhinus canicula (Wearmouth et al., 2012) and blacktip reef sharks Carcharhinus melanopterus (Mourier et al., 2013). However, there is conflicting evidence for sexual segregation in Rajidae (Martin et al., 2010, Steven, 1933, Rousset, 1990); these studies used fishing surveys to investigate distribution, which as mentioned previously limits presence or absence to one location in time. Therefore the spatial movements underpinning patterns of sexual segregation remain to be determined.

The present study aims to investigate habitat selection and spatial partitioning among four species of Rajidae off the south west coast of the UK, using passive acoustic telemetry. With network analysis we investigate whether movement patterns were random within the receiver array or if there was higher occupancy and movement within a particular part of the array, thereby indicating habitat selection. We also compare these movements between both species and sexes to identify spatial or habitat partitioning.

4.2 Method

During the study period between 2010 and 2017 there were two deployments of six Vemco hydrophone receivers (six Vemco VR3, six VR4; Vemco, Canada). All receivers were deployed by the Marine Biological Association (MBA) using its research vessels onto the seabed attached to large, weighted metal tripods (Figure 29). Six VR3 receivers were deployed in close formation within the Whitsand bay area in 2010 in depths between 19-25 m (the Whitsand array), a further six VR4 receivers were deployed in 2014, spread across an area ranging from the Eddystone to Bigbury bay, at depths between 29-53 m (the Plymouth array) (Table 14, Figure 30). All receivers were positioned on flat substrates to ensure maximum detection ranges and within the boundary of a Marine Protected Area (a SAC, Special Area of Conservation), primarily to avoid potential damage from mobile fishing gear used commonly in adjacent areas. Receivers were capable of identifying and recording the identification number and time stamp of coded transmitters and communicated remotely to a surface modem to transmit the stored data to a surface unit on demand (VEMCO, 2013). Surface range testing conducted by the MBA indicate that receivers in the array have a maximum detection range between 408-648 meters and the average maximum detection range was 502.5 m (±105.62 m) (Figure 31).

Figure 29: One of 12 receiver deployments on large tripod landers off the coast of Plymouth. The data-logging acoustic receiver is arrowed.

Table 14: Receiver array depths and names

Array Receiver Depth (m)

Whitsand L2 -20.17 Whitsand L3 -19.38 Whitsand L4 -24.74 Whitsand L5 -24.47 Whitsand L6 -20.83 Whitsand L1 -19.44 Plymouth M1 -53.27 Plymouth M2 -30.13 Plymouth M3 -46.61 Plymouth M4 -29.71 Plymouth M5 -52.21 Plymouth M6 -30.40

Figure 30: Location of the Whitsand and Plymouth arrays off the south west coast, Plymouth, UK. Bottom left highlights the Whitsand array in more detail.

Figure 31: Range testing in Whitsand array. Red circles are the receiver positions and arrows show the distance (m) of detection ranges.

Four common Rajidae species; R. brachyura, R. clavata, R. microocellata and R. montagui were caught, tagged and released (see Chapter 3, Figure 16, 17, 18, 19 for release positions) from 2014 to 2016 using beam or otter trawls off the coast of Plymouth, southwest UK. Vessels were either the Plymouth Quest or RV Sepia and all tows lasted less than 40 min. Serially numbered 10mm Peterson discs were attached to skates using the methods described by Wearmouth & Sims (2009a). Briefly, tags were secured in stainless steel wire with a pointed stainless steel wire attachment that was passed through a Peterson disc and then through the centre of the pectoral fin. A second Peterson disc was placed onto the length of wire on the opposite (ventral) side of the skate before turning a series of rounds into the length of wire remaining. Tagging took less than two minutes, including the measurement of total length and disc width of the individual. Skates were immediately transferred to aquaria with fresh running sea water for observation prior to release. A total of 201 Rajidae were released with Vemco V13 tags during the study period (Table 15). The tags used for all fish were V13 69 kHz transmitters that were 36mm long, weighing 11 g in air and 6 g in water with an estimated lifespan of 1,993 days.

Table 15: Number of acoustic tags released by species and sex

Species Male Female Average (±StdDev) of Average (±StdDev) of Residency Index Liberty (days) R. brachyura 15 16 21.86 ±48.79 124.68 ±92.81

R. clavata 28 26 26.09 ±32.15 96.68 ±131.93

R. microocellata 10 34 38.15 ±46.51 103.44 ±141.78

R. montagui 30 42 21.21 ±36.55 182.23 ±208.52

Data recorded by the 12 receivers were uploaded remotely with a Vemco deck box and surface modem. Uploaded detections were imported into a Microsoft Access (Microsoft Corporation, Redmond, USA) database, which assigned transmitter detections to the appropriate individual and receiver locations, and filtered out detections that did not match an active tag or receivers. Receiver time corrections were also made during this import process and were calculated from the difference between the receiver and PC clock at the time of download, assuming linear receiver clock drift.

4.2.1 Acoustic networks In network theory, connected systems are comprised of nodes connected by edges. Nodes can represent anything from mobile individual organisms or static physical locations while edges represent the associations between those nodes creating a network (Jacoby et al., 2012). In the present study nodes are the 12 static acoustic receivers and the edges represent the movement of an individual between two receivers within the spatial array.

Movements between receivers were used to construct a network displaying the movement of animals through space that can be used to test for habitat use within the array. There were no time frame restrictions set on connections due to these species moving relatively slowly and the wide spacing of the Plymouth array. To test whether the observed movements of the four species differed from random, random networks were generated. Due to spatial restrictions of the nodes in the network, connections between particular receivers are more likely than others because they are closer together. Random networks therefore accounted for this spatial bias in the network structure as follows: for each individual within a species the first detection at the first receiver was kept, then a maximum swimming distance was calculated based on the time between detections and an average swim speed that was calculated separately from the detections for each individual. Two receivers were then selected at random from those within the maximum swimming distance and the closer of the two was assigned as the next location. If no receiver was found in range then the current receiver was assigned. This was repeated for the entire track, for each individual and then repeated 100 times for each track to provide the mean random network metrics. Network metrics include ‘node density’, which measures the extent of the array occupied, and ‘edge density’ which provides a measure of mobility within the network. The values of both metrics ranged from 0–1. These random networks were tested against the observed networks using a Wilcoxons matched pairs signed-rank test. Species where the observed network differed from random were compared to identify differences in receiver occupancy and identify potential habitat partitioning in the four Rajidae species.

To identify whether species spent time equally within the array and whether the different species moved to other parts of the array equally, the total number of detections and connections were compared between species within the Whitsand array, the Plymouth array and between arrays with a Chi-squared Goodness of fit test. Expected values were based on the total number of days transmitter signals (pings) were detected within that array, or in other words the total number of days where there was a least one detection. The number of connections between the two arrays were also investigated with the same test, but with the expected values based on the total number of ping days across the entire array. In addition, the same analysis was applied to

89 investigate the difference in the number of detections by males and females within a species, with expected values based on total number of ping days for each sex.

4.2.2 Spatial overlap analysis To further investigate shared occupancy between pairs of species, a shared occupancy matrix was computed with counts of detections for each 0.5 km. To provide a numerical comparison of the matrices an overlap coefficient was computed using the following equation (Horn, 1966, Rijnsdorp et al., 1998):

Where Paj is the occupancy value (count of detections) for species a in grid cell j and Pbj the corresponding value for species b. These coefficients range from 0.0 (no overlap) to 1.0 (matching occupancy) and provide a way in which to determine similarity between grid occupancy constructed for the four different species. To investigate the statistical significance of the overlap coefficients a null model was used with a Monte-Carlo approach summarised in Humphries et al. (2016c). The same grid occupancy analysis was conducted to investigate whether different species overlapped within the arrays at the same time, that is, how frequently pairs of species and sexes within a species occupied the same 0.5 km grid squares in a week long period.

4.3 Results During the study period a total of 201 Rajidae were released with acoustic transmitters, providing 17,288 tracking days (Table 15). The mean track length was 127.76 days and the longest track was recorded for R. montagui at 831 days. Most tags were deployed during 2015 and this is reflected in the number of detections within the acoustic array (Figure 32). Of those tags released, 66.67% of tags were detected by the acoustic array.

Raja brachyura Raja clavata Raja microocellata Raja montagui

Individual

01/10 01/11 01/12 01/13 01/14 01/15 01/16 01/17 01/18 Date (mm/yy) Figure 32: Tag detections over time from 2010 to 2017.

4.3.1 Species detections For each species the observed network patterns (both node and edge metrics) were significantly different from the randomised acoustic networks (Figure 33) (P<0.001 Wilcoxon signed rank matched pairs test Appendix 2, S Table 3 and 4).

Figure 33: Observed networks left, Example randomised networks on the right A= R. brachyura B= R. clavata C= R. microocellata D= R. montagui. Bottom left box highlights the Whitsand array in more detail. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections.

Considering each array separately, the number of detections was not equally distributed between species [Whitsand 2(3, N=42512) =16894.8, p<0.001, Plymouth array 2(3, N=22180) =36267.8, p<0.001]. For the Whitsand array, R. clavata were detected more than expected, but both R. montagui and R. brachyura were detected less. Conversely in the Plymouth array, R. montagui and R. brachyura were detected more frequently than expected, while R. clavata and R. microocellata were detected less (Figure 34 and 35).

Within the Whitsand array species differed in the number of connections they made between two receivers [2(3, N=4184) =2232.68, p<0.001]. R. microocellata and R. clavata both made more connections than the expected values, while R. brachyura and R. montagui were less. In the Plymouth array species again did not make equal numbers of connections between receivers (2(3, N=39) =257.876, p<0.001). R. brachyura made significantly more connections than expected, while the other 3 species connected less than indicated with random networks (Figure 34 and 35).

100%

90%

80%

70%

60% R. brachyura 50% R. clavata 40% R. microocellata R. montagui 30%

20%

10%

0% Plymouth Whitsand Plymouth Whitsand Between array detections array connections array detections connections connections

Figure 34: Proportion of detections or connections by species at the Plymouth and Whitsand array and connections between the 2 arrays.

Detections were not equal among receivers and those with the highest detections differed between species. For R. brachyura, 87.78% of detections occurred at M4 in the Bigbury Bay area, while for R. montagui, 60% of detections were at M3 and 22% at M6. In R. clavata, 31% of detections occurred at L3 and another 31% of detections at L6. Finally in R. microocellata, 41% were detected at L6 and 29% at L1 (Figure 35).

4.3.2 Spatial overlap R. clavata and R. brachyura were the only species to have a significantly lower overlap coefficient than the randomised coefficient, or in other words the species pair overlap less than expected by random chance (Table 16). Conversely, R. montagui overlapped all other species more than expected by chance. Similarly, R. microocellata overlapped R. clavata significantly more than expected by chance. Finally the overlap between R. brachyura and R. microocellata was not significantly different to the randomised overlap coefficient.

Table 16: Observed versus randomised overlap coefficient. ↓= Overlap coefficient was lower than expected by chance, → not significantly different than expected by chance ↑ higher than expected by chance.

R. clavata R. microocellata R. montagui R. brachyura p-value 0.049 0.402 0.994 Overlap coefficient 0.044↓ 0.054→ 0.163↑ Mean randomised 0.066 0.063 0.048 coefficient R. clavata p-value 0.954 0.950 Overlap coefficient 0.184↑ 0.184↑ Mean randomised 0.045 0.046 coefficient R. microocellata p-value 0.946 Overlap coefficient 0.146↑ Mean randomised 0.039 coefficient

Despite some overlap between the species, there was much less temporal overlap (Figure 36, Figure 37, Figure 38). As expected from the previous results, R. clavata and R. brachyura rarely overlapped and this was similar for R. brachyura versus R. microocellata and R. microocellata versus R. montagui. There was slightly more overlap with R. montagui and R. brachyura, but perhaps the most interesting was the overlap pattern across time with R. microocellata and R. clavata, which peaked in the winter months.

5 1.2 5 1.0 1.0 4 4 0.8 0.8 3 3 0.6 0.6 2 2 0.4 0.4

Coefficient of overlap 1 Coefficient of overlap 1 0.2 0.2

Maximum R. clavata R. occupancy Maximum Maximum R. brachyura occupancy brachyura R. Maximum 0 0.0 0 0.0

1 1 Col 13 vs Col 14

2 2

3 3

4 4

Maximum R. montagui occupancy montagui R. Maximum

Maximum R. microocellata occupancy microocellata R. Maximum 5 5 10 11 12 13 14 15 16 17 10 11 12 13 14 15 16 17 18 Year (20yy) Year (20yy)

Figure 36: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. brachyura Right= R. montagui vs R. clavata.

5 1.2 5 1.2

1.0 4 1.0 4

0.8 0.8 3 3 0.6 0.6 2 2 0.4 0.4

Coefficient of overlap 1 1 Coefficient of overlap 0.2 0.2

Maximum R. clavata R. occupancy Maximum 0 0.0 occupancy brachyura R. Maximum 0 0.0

1 1

2 2

3 3

4 4

Maximum R. clavata R. occupancy Maximum

Maximum R. microocellata occupancy microocellata R. Maximum 5 5 10 11 12 13 14 15 16 17 18 10 11 12 13 14 15 16 17 18 Year (20yy) Year (20yy)

Figure 37: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. clavata Right= R. clavata vs R. brachyura.

5 1.2 5 1.2

1.0 1.0 4 4

0.8 0.8 3 3 0.6 0.6 2 2 0.4 0.4

Coefficient of overlap 1 Coefficient of overlap 0.2 1 2D Graph 5 0.2

Maximum R. montagui occupancy montagui R. Maximum

Maximum R. montagui occupancy montagui R. Maximum 0 0.0 0 0.0

1 1

2 2

3 3

4 4

Maximum R. brachyura occupancy brachyura R. Maximum

Maximum R. microocellata occupancy microocellata R. Maximum 5 5 10 11 12 13 14 15 16 17 18 11 12 13 14 15 16 17 18 Year (20yy) Year (20yy)

Figure 38: Spatial overlap between species pairs over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata vs R. montagui Right= R. brachyura vs R. montagui.

4.3.3 Differences in the detection of sexes In every species there were differences in detections between males and females (Figure 39, Figure 40, Figure 41). In R. microocellata [2(1, N=24634) =747.05, p<0.01], females were detected more frequently than males. Both sexes were detected predominantly in the Whitsand array. For R. brachyura [2(1, N=3348) =179.61, p<0.001], males were detected more frequently than expected and females less. Males were recorded predominantly in the west and south of the Plymouth array, and while females overlapped males in the west, they also made connections to the Whitsand array in the east. In R. montagui [2(1, N=17793) =6305.93, p<0.001], females were detected more than expected and males less. Finally in R. clavata [2(1, N=29541) =2226.4, p<0.001], females were detected more than expected and males less. While there were only a few detections of R. clavata males, when they were detected they overlapped the females in the west end of the Whitsand array.

Figure 39: Acoustic networks comparing sexes. A= R. brachyura B= R. clavata. Left = female, Right = male. Bottom left box highlights the Whitsand array in more detail where required. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections.

Figure 40: Acoustic networks comparing sexes. A= R. microocellata B= R. montagui. Left = female, Right = male. Bottom left box highlights the Whitsand array in more detail where required. Warmer colours indicate more connections/detections. Stars indicate nodes with zero detections.

100

100% 90% 80% 70% 60% 50% Male 40% Female 30% 20% 10% 0% R. brachyura R. clavata R. microocellata R. montagui

Figure 41: Proportion of detections by sex for each species, normalised by total number of days with at least one ping to account for tagging bias.

R. clavata and R. microocellata sexes both had high overlap and when the males were detected it was generally at the same place as the females (Figure 42and Figure 43). However, there were times when females were detected, but males were not, as expected from the significantly lower number of male detections. R. montagui males and females overlapped during the summer months (May-July) but there were times when both males and females were detected, with an overlap coefficient of zero (Figure 43). The same pattern exists for R. brachyura, although there was less overlap between males and females (Figure 42).

101

3.5 1.2 5 1.2

3.0 1.0 1.0 4

2.5 0.8 0.8 3 2.0 0.6 0.6 1.5 2 0.4 0.4

Coefficient of overlap of Coefficient 1.0 overlap of Coefficient 1 0.2 0.2

0.5 occupancy male clavata R. Maximum

Maximum R. brachyura male occupancy male brachyura R. Maximum

0.0 0.0 0 0.0

0.5 1

1.0

2 1.5

2.0 3

2.5 4 3.0

Maximum R. clavata female occupancy female clavata R. Maximum

Maximum R. brachyura female occupancy female brachyura R. Maximum

3.5 5 06/15 09/15 12/15 03/16 06/16 01/10 01/11 01/12 01/13 01/14 01/15 01/16 01/17 01/18 Date (mm/yy) Date (mm/yy)

Figure 42: Spatial overlap between sexes over time. Red lines indicate the timing and extent of overlap between species. Left= R. brachyura, Right= R. clavata.

5 1.2 5 1.2

1.0 1.0 4 4

0.8 0.8 3 3

0.6 0.6

2 2 0.4 0.4

Coefficient of overlap of Coefficient

1 1 0.2 0.2

Maximum R. montagui male occupancy male montagui R. Maximum

Maximum R. microocellata male occupancy male microocellata R. Maximum 0 0.0 0 0.0

1 1

2 2

3 3

4 4

Maximum R. montagui female occupancy female montagui R. Maximum

Maximum R. microocellata female occupancy female microocellata R. Maximum 5 5 01/10 01/11 01/12 01/13 01/14 01/15 01/16 01/17 01/11 01/12 01/13 01/14 01/15 01/16 01/17 01/18 Date (mm/yy) Date (mm/yy)

Figure 43: Spatial overlap between sexes over time. Red lines indicate the timing and extent of overlap between species. Left= R. microocellata Right= R. montagui.

102

4.4 Discussion

4.4.1 Habitat selection in Rajidae Habitat selection is defined as space use that is non-random due to voluntary movements by a species (Sims, 2003, Kramer et al., 1997, Morris, 2003). The networks analysed in this study represent the voluntary movements of four species and in every species, observed network metrics were significantly different from the random networks. In the Plymouth array, receiver M1 was never visited for example, M5 very occasionally, while M4 and M6 and most of the Whitsand array received a higher number of detections (presence of skates). In a recent study, depth was a factor in habitat preference in Rajidae; indeed, the maximum mean depth for any of these four species was 36.4 m, suggesting preference for shallow coastal habitats (Humphries et al., 2016c). This may explain why the receivers M1 at 53 m and M5 at 52.21 m were rarely visited. It was also found that species had preferences for particular nodes, with for example R. brachyura having 87.78% of detections at M4. The findings demonstrate collectively that the distribution of Rajidae is non-random and infers active selection towards particular areas (receiver nodes), consistent with habitat selection on a fine scale.

There are several hypotheses that may explain the occurrence of fine-scale habitat selection in Rajidae. Firstly, the study area is a complex ecosystem, influenced by an estuary and varying habitat types, including sandy to rocky substrates (European Environment Agency, 2018) which may play a role in the habitat selection of these species. Substrate has been reported as an important driver of Rajidae distribution, with many species associated with sandy or muddy substrate (Martin et al., 2010, Damalas et al., 2016) which they may be actively selecting when moving within their preferred areas of the acoustic array. Second, prey availability can drive habitat selection; tiger sharks for example have been shown to preferentially occupy shallower seagrass habitats where their prey is most abundant (Heithaus et al., 2002). Similarly the distribution of the small spotted catshark in the Mediterranean was associated with the availability (abundance) of finfish and cephalopod prey (Navarro et al., 2016). The four Rajidae species in this study are broadly opportunistic feeders, and while every species is capable of eating both crustaceans and fish, some species appear to exhibit preferences. For example R. clavata and R. montagui showing preference for crustaceans, while R. microocellata and R. brachyura prefer teleosts (Ellis et al., 1996a, Forman and Dunn, 2012). It is likely that particular habitats within the study site are more likely to contain a greater abundance of preferred prey

103 items, such as sand eels or swimming crabs, and this may account for the observed habitat selection.

A third possibility is that reproductive state is important, as it can be another known driver of habitat selection in elasmobranchs. Nurse sharks, for example, use specific sites for mating and parturition behaviour (Carrier and Pratt, 1998). Rajidae are oviparous species and are thought to use shallower nursery grounds for oviposition (Ellis et al., 2005a) and this may explain movements between specific habitats found in the current study, for example R. brachyura moving between landers M4 and M6. Finally, competition from other benthic predators, including plaice or other skate species may lead to each species having a different habitat selection (Papastamatiou et al., 2006). It is likely that a combination of suitable soft substrates for burying, the availability of sand eels or crabs, a need to find mates and deposit eggs drive the habitat selection found in the current study.

4.4.2 Resource partitioning in Rajidae A major driver of habitat partitioning between species is competitive exclusion; by occupying different habitat, sympatric species can separate resources and avoid potential competitors (Papastamatiou et al., 2006). The sympatric Rajidae species at the focus of the current study are morphologically similar and are considered to be opportunistic generalist feeders (Ellis et al., 1996a), which may mean they exhibit habitat partitioning to reduce competition.

This study found that the selection of habitats were different among species. There were differences between species networks and the proportion of time spent at receivers. Receivers detected R. clavata and R. microocellata at the shallower receivers (19.4-24.7 m) in Whitsand bay array, while R. montagui and R. brachyura were detected more frequently in the deeper Plymouth array (29.7-53.3 m), splitting the four species into two pairs. This species pairing has previously been demonstrated with depth-logging electronic tags (Humphries et al 2016). However in the current study, which can pinpoint a geographic location, R. montagui were more frequently detected at lander M3 (60%), while R. brachyura were predominantly detected (87.78%) at M4 (M3 and M4 are 9.9 km apart), suggesting horizontal spatial partitioning in this pair.

Spatial overlap analysis however found species pairs with differing degrees of spatial overlap. For example R. clavata and R. brachyura had a lower overlap coefficient, whilst R. brachyura and R. montagui were more overlapped than expected by chance. However, when investigating space use over time, there was less overlap between species. R. brachyura and R. montagui for example only show some overlap during the summer months and R. brachyura and R. microocellata were very rarely overlapped in the same grid square at the same time. Detections of R. microocellata

104 and R. clavata overlapped within the Whitsand array, however there may be temporal habitat partitioning between the pair. Spatial overlap analysis suggests that there may be short peaks of overlap in winter, followed by periods where one or both species are detected by the array but with no spatial overlap. Care however must be taken when interpreting these results. Investigating timing of detections for example could be biased by a large release of tagged animals. Nonetheless, this potential bias may be mitigated by the observation that large batches of tag releases occur at a time when skates were caught in a particular area, which may be entirely representative of the individuals present in space and time. In addition, we can only detect tags where we chose to position the receivers and because the array is relatively limited spatially, overlaps between species may result from the constraint of only being able to sample at 12 locations. Despite these limitations, it is clear that these four sympatric species exhibit fine- scale temporal habitat partitioning within the study area.

It is interesting to note that density-dependent habitat selection suggests that in an expanding population individuals will occupy suboptimal habitat. Conversely, in a decreasing population a species will move to occupy only the optimal part of their range (Blanchard et al., 2005, Sutherland, 1983). In addition, fishing pressure may also have acted to remove skates frequently from optimal habitats where fisheries co-occur, so that they are generally observed to occur more frequently in sub-optimal habitats. This may act further to reduce the overlap between the species. Over the last century Rajidae populations have declined, for example between 1956-1995 there was a reported 80% decline in thornbacks R. clavata in the North Sea and 45% in the Irish Sea between 1988-1997 (Walker and Heessen, 1996, Dulvy et al., 2000, Chevolot et al., 2006). It is possible that the habitat partitioning observed in the current study is more apparent because these Rajidae populations have declined and retracted into areas of the highest habitat suitability. Common skate (Dipturus batis complex) is an example of a species that were once common but are now locally extinct due to overfishing (Brander, 1981). Hence, they may now occupy refugia where areas of low fishing effort and preferred habitat co-occur (Shephard et al., 2012). Common skate seem to occupy much deeper habitat (Wearmouth and Sims, 2009b) than the species studied here, which suggests habitat partitioning among species. Alternatively, this could be the result of the retraction of the common skate range into refugia, masking their historic spatial occurrence and exaggerating habitat partitioning in skates. On balance, it is likely that the population declines in Rajidae have exaggerated the habitat partitioning observed in the study area relative to the historic baseline ecosystem, for which there is very limited information.

In addition, understanding this habitat partitioning among Rajidae species is also important to consider in a commercial fisheries context, because habitat partitioning has an impact on the

105 catchability and therefore the vulnerability of each species (Mucientes et al., 2009). As explored in Chapter 3, it is clear that when habitat partitioning exists some species are caught by different gear types and different species may occupy habitats that vary in fishing pressure. Quotas for Rajidae species are still largely set by fisheries managers as a grouped quota and not treated as individual species. When particular species have higher catchability due to their habitat selection, grouped quotas may allow the further decline of Rajidae.

4.4.3 Differences in movement among species The connections observed in the networks are a proxy for the movements made by individuals between different geographical locations. The four rajid species differed in their movements within the array. Within the Plymouth array connections were predominantly made by R. brachyura and the connections within this array are much wider than the Whitsand array. In addition, long distance connections between the arrays were made primarily by R. brachyura and R. microocellata. R. montagui were detected more frequently within the Plymouth array than R. brachyura, but made fewer connections, indicating that R. montagui exhibit a higher degree of residency than R. brachyura. The differences in the extent of their movement may reflect prey availability, with the preferential piscivores (R. microocellata and R. brachyura)(Ellis et al., 1996a) searching over wider areas for their prey. R. clavata are associated with both marine and brackish waters and are considered to be the skate with the widest distribution of the four species (Ellis et al., 2005a, Hunter et al., 2006), yet individuals tracked in this study did not make long distance movements within the array. Conversely, R. brachyura is a species considered to have a patchy and narrower range in UK waters (Martin et al., 2010), and were observed here to make long distance movements within the array at least. The wider habitat tolerance of R. clavata may allow this species to exploit a range of habitats in close proximity and therefore it may require only to move short distances during habitat selection. In contrast R. brachyura, with its narrower habitat preference, may therefore have to move greater distances to find these preferred (less generalised) habitats. Overall the results indicate a difference in the movement strategies of these four species.

4.4.4 Sexual segregation in Rajidae In every species there were biases between sexes in the detections made within the array. In R. microocellata, R. montagui and R. clavata there were significantly more detections from females, while in R. brachyura more males were detected. Evidence for sexual segregation has been implied previously by studies that found a tendency to capture females in trawls (Steven, 1933). Sex ratio bias in trawls was also demonstrated in R. clavata in the Bay of Douarnez, France, where trawl catches for this species were predominantly male-biased inside the bay and female outside

106 the bay (Rousset, 1990). The results of the current study suggest that males and females do not consistently occupy the same fine-scale habitat which confirms active sexual segregation in these species.

Two of five main hypotheses (Wearmouth and Sims, 2008) are likely explanations for sexual segregation in Rajidae. The activity budget hypothesis suggests that sex differences in body size and reproductive investment result in sex-specific activity budgets resulting in different habitat requirements. The Rajidae are oviparous species, with females laying eggs in nursery areas and growing larger and maturing later than males. Females may therefore spend time in different habitat to males during oviposition, and growth and maturity may cause a separation of the sexes in these species. Additionally, the social factors hypothesis suggests sexual segregation is caused by intra-sexual affinity for the purpose of cooperation of information transfer or intersexual aversion, due to aggression. Elasmobranch males tend to exhibit aggressive behaviour towards females during mating, which is likely costly to females, which often results in female avoidance of males (Kimber et al., 2009, Wearmouth et al., 2012). Female Rajidae also have the capacity to store sperm thereby reducing the need to mate regularly and enabling females to avoid males for longer periods (Luer and Gilbert, 1985). Currently there is not enough evidence to conclude which of these two main hypotheses are more likely, indeed it is plausible that both oviposition by females and female avoidance of males play a role in the sexual segregation observed here.

A key limitation of the acoustic array used in this study is the number of receivers in the array and therefore the spatial coverage of detection. This resulted in a third of tagged skates never being detected by the array. In addition, as is the case with all passive acoustic arrays, when transmitters are outside the detection range of the receivers, we cannot know where the animal is. This reduces our ability to record fine scale movements and decreases the resolution of the spatial and temporal data recorded. The placement of additional receivers is possible; however in this heavily fished region (see Chapter 3, Figure 27) receivers should be placed within the boundaries of MPAs to reduce potential damage by mobile fishing gear. Furthermore the detection ranges of the receivers can change depending on multiple environmental conditions (Kessel et al., 2014), likely resulting in seasonal changes in the coverage of the array. The data and conclusions from such studies must therefore been interpreted with some caution. Nevertheless, these limitations were countered to some degree by the passive acoustic array at the focus of the current study being successfully deployed and staying operational over a long period, with the Whitsand array remaining in position and detecting tags for approximately seven years. The array was still in place and receivers were able to communicate remotely, even after the severe storms

107 of winter 2014-2015 (Met Office, 2015), suggesting that the use of an acoustic array is possible in a high energy ecosystem over long periods..

4.4.5 Conclusion The findings suggest habitat selection and partitioning among these four rajid species, as well as intra-species sexual segregation, revealed by long-term acoustic telemetry. The higher spatial and temporal resolution provided by the passive acoustic array relative to mark and recapture and fishing surveys, allows for a more robust and fine scale picture of the spatial distribution of rajids. For example, R. brachyura and R. montagui appear to exhibit a greater degree of fine-scale separation than previously demonstrated in Chapters 2 and 3.

108

Chapter 5. Diel vertical migration and central place foraging in benthic predators

This work was published as: Humphries NE, Simpson SJ, Sims DW. Diel vertical migration and central place foraging in benthic predators. Marine Ecology Progress Series. 2017; 582:163-80. With Humphries NE and Simpson SJ as joint first author. It has been adapted for this thesis.

5.1 Introduction

Diel vertical migration (DVM) is commonly observed in pelagic animals from plankton to elasmobranchs (Hays, 2003, Sims et al., 2005, Andrews et al., 2009, Matern et al., 2000) and it is suggested to be the largest daily movement of animals on the planet (Hays, 2003). Animals exhibiting DVM typically move from shallow water at night to deeper water in the day. For prey animals such as copepods and krill DVM is hypothesised to result from predator avoidance, enabling organisms to shelter from visual predators in darker deeper waters during daytime (Hays, 2003, Lampert, 1989). The study of DVM in large pelagic vertebrates such as tuna and sharks, has been permitted by the arrival of electronic tagging (Evans et al., 2014, Goodyear et al., 2008), (Walli et al., 2009, Schaefer and Fuller, 2002, Humphries et al., 2016a), (Sims et al., 2005, Shepard et al., 2006, Queiroz et al., 2012). It is hypothesised that DVM in predators results from foraging trips, with predators following vertically migrating prey (Hays, 2003, Queiroz et al., 2010, Sims et al., 2005). While the proximate driver of DVM is generally agreed to be predator avoidance or prey tracking, the ultimate cause of the movement is considered to be changing light levels (Lampert, 1989, Mehner et al., 2007, Mehner, 2012). In bigeye tuna Thunnus obesus DVM has been equated to central place foraging, where the tuna forage below the thermocline in deeper water during the day and return to shallower water at night (Humphries et al., 2016a, Schaefer and Fuller, 2010). Although typically DVM is a movement from shallow water at night to deep water during the day and is therefore termed normal DVM (nDVM), reverse DVM (rDVM, deeper at night, shallower during the day) has also been observed in zooplankton, sharks and ocean sunfish Mola mola (Irigoien et al., 2004, Sims et al., 2005, Pade et al., 2009, Queiroz et al., 2012). Basking sharks Ceteorhinus maximus rDVM has been observed and is attributed to their following the rDVM exhibited by their prey Calanus helgolandicus (Sims et al., 2005) which in turn were responding to the presence of their predators, the chaetognaths (arrow worm). DVM therefore represents a complex and flexible response to the movements of either predator or prey, often triggered by changing light levels.

109

Recently the occurrence of DVM has recently been expanded to include not only pelagic but also benthic organisms. Studies have examined differing patterns of vertical migration, with some benthic organisms moving from the benthos into the water column (Aguzzi et al., 2015) and conversely some pelagic organisms moving towards the benthos (Zouhiri and Dauvin, 1996), resulting in links between traditionally separate vertical habitats (Aguzzi et al., 2015, Sutton, 2013). While these studies have revealed differences in DVM behaviour, DVM in benthic predators, such as skates, is unexpected because they are generally evolved for benthic living. Nonetheless, DVM has been observed in benthic species, such as decapod crustaceans (Aguzzi and Company, 2010), catsharks, Scyliorhinus canicula (Sims et al., 2006) and, in freshwater lakes, burbot Lota lota (Cott et al., 2015, Harrison et al., 2013). However, in these species DVM does not represent a vertical movement into the water column, rather an inshore-offshore (nektobenthic displacement) movement along the sea (or lake) bed (Aguzzi and Company, 2010). In freshwater systems, studies have used the term diel bank migration (DBM), to define this nektobenthic movement (e.g. Cott et al., 2015) however this seems less appropriate for marine studies. Instead, to distinguish the movements studied presently from pelagic DVM we will adopt the term nektobenthic DVM throughout (Aguzzi and Company, 2010).

In male S. canicula, nektobenthic DVM has been attributed to behavioural thermoregulation to improve bioenergetic efficiency (Sims et al., 2006), where the sharks hunt in shallow warmer water and rest and digest in deeper cooler water. It is known from previous studies, however, that the skates studied here do not perform behavioural thermoregulation (Humphries et al., 2016b).

Skates are strongly adapted benthic mesopredators with dorso-ventrally flattened bodies, negative buoyancy, ventrally located mouths and dorsal gill openings (spiracles) that permit the animals to respire whilst buried in sediment. Skates have electrosensitive receptors (ampullae of Lorenzini) that are used during foraging for cryptic benthic prey, such as sand eels (Ammodytes spp.) or crustaceans (Wueringer et al., 2012, Collin, 2012, Kimber et al., 2013).

The association of Rajidae with the benthos is further confirmed by dietary studies which show that the main prey items were benthic invertebrates or benthic or demersal teleosts (Šantić et al., 2012, Farias et al., 2006a, Ellis et al., 1996c, Kadri et al., 2014, Ajayi, 1982b, Steven, 1932b, Pinnegar, 2014, Ajayi, 1982a). Pelagic species were only found in less than 2% of Rajidae stomachs (Pinnegar, 2014, Farias et al., 2006a) and are likely to have been scavenged from species that are discarded by the commercial fishery. Given these benthic associations it seems unlikely that DVM would occur in skates. However, there is little information concerning the fine scale movements and activity patterns of these animals. Currently all the species studied presently, excluding R.

110 montagui, are listed as near threatened by the International Union for the Conservation of Nature (IUCN) Red List and consequently an understanding of population abundances, movements and preferred habitats is fundamental given reported declines in many skate populations (Genner et al., 2010a, Simpson et al., 2011, Walker and Hislop, 1998, Brander, 1981). Vertical movements that are potentially connected to DVM have been reported in electronically tagged common skate, Dipturus batis, however only half the individuals tracked exhibited vertical movements and the behaviour was not investigated in detail (Wearmouth and Sims, 2009a). It is therefore possible that nektobenthic DVM, also occurs in skates. To investigate this hypothesis, depth time series data from 89 electronic-tagged individuals of four Rajidae species were analysed to identify the presence of DVM, to describe possible these events and to determine possible drivers. Our main interests were events that suggested central place foraging, a behaviour not previously reported in rajids. In a pelagic species, such as tuna, central place foraging is associated with depth. However, the benthos is made of a complex mosaic of habitats and if skates are returning to a depth, then it is possible that this represents some preferred foraging or refuging habitat. This chapter presents novel observations of DVM in four sympatric species of skates, Raja brachyura, R. clavata, R. microocellata and R. montagui and tests hypotheses to account for this behaviour. This research investigates whether the observed DVM represents nektobenthic rather than pelagic movements, and investigates the purpose and nature of the events in the context of central place foraging.

5.2 Methods

5.2.1 Tags and tagging Fish were captured during research surveys in the Western English Channel (WEC) between Whitsand Bay (50.34 N, 4.28 W) and Bigbury Bay (50.26 N, 3.89 W) and between July 2008 and May 2013. Fish were tagged using either of two types of CTL G5 (www.cefastechnology.co.uk, UK), or Star Oddi (star-oddi.com; Star-Oddi, Iceland) data storage tags (DSTs). Standard G5 DSTs were 31 mm long by 8 mm diameter and weighed 1 g in water, whereas long-life DSTs measured 35.5 mm long x 11.5 mm diameter and weighed 2.1 g in water. Star Oddi DSTs were 39.4 mm by 13 mm and weighed 5 g in water. DSTs record temperature from 2 to 34 °C (accuracy 0.1 °C, resolution 0.03 °C) and pressure to a depth of 100 m or 200 m (accuracy 1%, resolution 0.04%). DSTs were programmed to record depth at 20 s, 30 s or 1 min intervals and temperature every 10 min. DSTs were attached to skates via Peterson disc tagging using the methods described by Wearmouth & Sims (2009a) and described in Chapter 2. DSTs were returned through the

111 commercial fisheries with a reward of £50 for return of the DST, the fish and for information about the size of the fish and where it was captured.

5.2.2 Data analysis To increase the accuracy of the recorded depths in the time series the tidal signal was attenuated using a correction factor taken from the Plymouth tide gauge data recorded at Devonport (50.3684°N, 4.1852°W). The general pattern of the DVM events was unknown at the beginning of data analysis and only through an accidental observation of a vertical movement in R. clavata, such as those shown in Figure 44a, were the events of interest revealed. Consequently automated identification of events within the time series data, which would require a priori parameterisation or assumptions was not appropriate (e.g. Adachi et al., 2016). Automatic identification of individual events would also have been hindered and masked the complexity of the vertical displacements during the events which make the investigation of changes in behaviour, such as from ascent to plateau, more difficult. Instead, software was developed by the Marine Biological Association (MBA) to manually identify events by using mouse clicks to delineate four points for each event (event start time, plateau start time, plateau end time and event end time)(Figure 44b). The intervals between these four times were defined as ascent, plateau and descent and overall as an event, with the time between events termed deep. To view and mark more easily the time-series data were under-sampled to 1 in 10 by selecting every 10th data point. This provided a clearer view of longer events than if the data were displayed at the original 20 s resolution, making marking times simpler. When the end of an event was marked the delineated times of an event were recorded to a Microsoft Access™ database. When all the events were identified in a particular time-series the software used stored event times to recalculate events, using tidally corrected but non-under-sampled time-series data to accurately calculate metrics to describe the events. Metrics include ascent/descent deltas (overall vertical displacements), speeds and durations; plateau delta, duration and activity (sum of and mean vertical displacements) and pre- event activity (activity in the hour prior to the event start). Furthermore a straightness index (SI) for each delineated phase was calculated using L0/L1 where L0 is the delta and L1 is the sum of step lengths. The SI is therefore a value between 0 and 1, where 1 is a straight movement and lower values represent increased tortuosity. To test for significant differences in the straightness index of the ascent and descent within a species, a Mann-Whitney-U test was performed.

112

5 a 10

12:00 18:00 00:00 06:00 12:00 18:00 00:00 06:00 12:00 18:00 00:00 06:00 12:00 5 b 10

Depth 20

12:00 18:00 00:00 06:00 12:00 c d e 10

11:00 18:00 01:00 08:00 11:00 18:00 01:00 08:00 12:00 19:00 02:00 09:00 16:00 Time of day Figure 44: Example DVM plots Depth time-series plots from a thornback ray (tag B0342) showing examples of DVM and other movements not studied here. a) Three events showing movement from deeper water (blue dashed line, 26 m, note the clear tidal signal) to a shallower depth (green dashed line, 10 m); b) A single DVM event from (a) with the tidal signal removed, showing the delineated event times (red circles); c) an event representing a longer term movement from deep (blue line) to shallower water (green line); d) a reverse event, with movement from shallow to deeper water and back again; e) a complex movement pattern resulting in a transition to deeper waters. Events such as those shown in c-e, are not classed as central place foraging events and are not further analysed.

During the manual identification, all vertical migration events that could be clearly identified were marked. This preliminary examination of the time-series distinguished 2354 events, however, the events of interest in this study were those which represent clear, repeated, diel excursions from deep to shallower water and, importantly, to suggest activity akin to central place foraging. These events were therefore filtered from the database for further analysis using the following parameters: ascent delta > 5 m, overall event delta < 2.5 m and event duration between 1 and 18

113 hours. This filter selects events where the skate moved at least 5 m shallower, remains there for between 1 and 18 hours and returns to a depth within 2.5 m of the starting depth. This filtering removed events such as events representing a longer term movement from deep to shallower water (Figure 42 c), reverse events (Figure 42 d), with movement from shallow to deeper water and back again and complex movement patterns resulting in a transition to deeper waters (Figure 42 e). Events such as these are not classed as central place foraging events and are not further analysed. Appendix 3 supplementary Figure S1 shows the range of values observed for the 2354 event deltas (the difference in depth between the start and end of the event) and the ascent deltas and indicates the cut off values for the events selected for further analysis.

To investigate differences in temperature between the deep phase and the plateau phase, average temperatures for the deep phase were calculated using the tag-recorded temperatures at the event start and end times, while the average temperature of the plateau phase was calculated from all data points between the plateau start and end points. Average vertical displacements (step lengths) or activity rates were calculated from data points up to one hour before the start of the event (pre-event activity) and for all data between the plateau start and end points (plateau activity). When an event was preceded by another event within less than an hour, all available points from the end of the previous event, to the start of the next were used for the pre-event activity calculation. Both differences in temperature and activity between deep and plateau phases were tested with a Mann-Whitney-U test.

Vertical movements seen in these time-series are either the vertical component of a 3- dimensional movement, or are the result of the relationship between horizontal movements and the complexity of the benthos. These vertical movements are therefore not a true measure of activity but are a reasonable proxy given that, generally increased activity will result in increased vertical displacements.

It is possible that a higher occurrence of events is associated with environmental factors or certain geographic locations. While the actual locations of the animals throughout the tracking period are unknown precisely off Plymouth, the depth is known accurately and can be used as a proxy for geographic location, as deeper depths are generally associated with locations further offshore in the study area. To identify peaks in the frequency of occurrence in relation to the start depth, histograms were plotted to show the count of event start depths in 10m depth bins (0-80 m) by species and sex.

To investigate seasonal peaks in event frequency, expected seasonal counts of events by species and sex were calculated using the total number of events for the year (for species and sex) and

114 the proportion of individuals tracked in each season. The expected counts therefore assumed an even distribution of events throughout the year, allowing for differences in the number of individuals tracked in each season. To provide more robust statistical analysis, counts were pooled into winter (December, January and February), spring (March to May), summer (June to August) and autumn (September to November). These months were selected using tag recorded temperatures such that winter represented the coldest three months, spring the warming months, summer the warmest and autumn the three cooling months. The statistical significance of differences between the observed and expected counts was determined using Chi-Squared Goodness of Fit test.

All statistical tests were performed using SigmaPlot 12.5 (Systat Software, San Jose, CA) or MiniTab 15.1 (State College, PA: Minitab, Inc., www.minitab.com). Because in most cases data were found not to be normally distributed, the statistical tests employed to determine significant differences in, for example, metrics such as ascent delta, were non-parametric tests such as Mann-Whitney rank sum, or Kruskal-Wallis One Way Analysis of Variance on Ranks.

5.3 Results

5.3.1 Summary Of the 179 tags deployed, 92 (51.4 %) were returned with 89 having valid data with a total of 12,585 tracked days (Table S1). A total of 2354 events were identified from which 674 were filtered for further analysis (Table 17). The number of events recorded for each individual was found to be variable in all species (Figure 45a). Although R. brachyura and R. montagui have fewer individuals, fewer events and fewer days tracked, the average number of DVM events per day are very similar in all species, with no significant differences found (Kruskal-Wallis One Way Analysis of Variance on Ranks, p=0.334), suggesting an adequate sample of events had been recorded. Consequently, there is an expected correlation between the number of days tracked and the number of events recorded (linear regression: R. brachyura R2=0.123, p=0.29; R. clavata R2=0.313, p=0.002; R. microocellata R2=0.474, p<0.001; R. montagui R2=0.248, p=0.173; Figure 45b-e). Notably, the 20 tracks where no events were selected were very short, with an average length of 49 days (range 2-154), compared to an average length of 166 days (range 14-418) for those with events. Differences in the number of events for each sex were significant only in R. brachyura and R. clavata, where more events were recorded for males than females (Table 18). No significant correlation was found between the length of the individuals and the number of events (Appendix 3 Supplementary Figure S2).

115

Table 17: Frequency of DVM events per day tracked by species Individuals No with No of Days DVM per day DVM events tracked tracked R. brachyura 12 11 62 1215 0.051 R. clavata 43 28 294 5037 0.058 R. microocellata 24 21 240 3474 0.069 R. montagui 10 9 78 1795 0.043 Total 89 69 674 11521

a) No of events per individual

Count

R. brachyura R. clavata R. microocellata R. montagui Species Figure 45: Frequency of events per individual a) The histogram shows a count of the selected events recorded for each of the 69 individuals. It is clear that this behaviour is not performed consistently by all individuals. For R. brachyura and R. montagui in particular the behaviour is frequent in only a few individuals.

Table 18: Χ 2 results from the frequency by sex analysis Species Χ 2 N df P R. brachyura 12.33 62 1 <0.001 R. clavata 31.41 294 1 <0.001 R. microocellata 2.5 240 1 0.114 R. montagui 2.77 78 1 0.096

116

5.3.2 Movement analysis To understand of how the animals were moving during these events, the ascent, plateau and descent phase movements were analysed for each species (Figure 46 and Table 19). Ascent speeds (Figure 46a) ranged from 0.53 to 3.3 cm/s; descent speeds ranged from 0.13 to 5.5 cm/s. No significant differences were found between the ascent and descent rates (Table 19).

No significant differences were found between species in the ascent SI values (Kruskal-Wallis One Way Analysis of Variance on Ranks, p=0.224) (Figure 46b). A significant difference was found between the descent SI values between R. clavata and R. microocellata (Kruskal-Wallis One Way Analysis of Variance on Ranks, p=0.012), however, no significant differences were found between the intra-species ascent and descent values, except for R. microocellata (Table 19). In all cases the SI values suggest a tortuous rather than a straight vertical movement.

Except R. montagui the SI index for the plateau phase was significantly lower than for the ascent or descent phases. For R. montagui although the median value of the plateau phase SI was comparable with the other species, there was a much greater range of values, suggesting greater individual variability. The descent phase SI for R. montagui also had a greater range of values and some that were much lower than the other species.

Table 19: Summarised statistical results A summary of the statistical tests, showing the test performed for each analysis and the associated p values. LR: Linear regression; MWRS: Mann-Whitney Rank Sum. Significant p- values are highlighted.

Species Measure Test brachyura clavata microocellata montagui R2=0.123 R2=0.313 R2=0.474 R2=0.24 Events per day tracked LR p=0.29 p=0.002 p<0.001 p=0.173 Ascent vs descent rates MWRS p=0.743 p=0.915 p=0.669 p=0.427 Ascent vs descent SI MWRS p=0.162 p=0.138 p=0.049 p=0.222 Plateau vs deep activity MWRS p<0.001 p=0.001 p<0.001 p=0.005 Plateau vs commuting time MWRS p=0.158 p<0.001 p<0.001 p=0.659 Deep vs shallow temp. MWRS p=0.793 p=0.652 p=0.841 p=0.959

There was significantly more vertical activity during plateau phases than in deep phases (Figure 46c, Table 19). This supports the contention that the plateau phase represents active foraging in shallower water while the deep, inter-event phase represents resting or refuging.

117

In all cases, more time was spent in the plateau phase than the combined ascent and descent phases (Figure 46d), which ranged from 43% of the plateau phase (R. microocellata) to 73% (R. brachyura). Differences were significant however only for R. clavata and R. microocellata (Table 19). Average time spent in the plateau phase ranged from 4.07 to 4.89 hours which represents between 58 and 70% of the total event duration.

a) Ascent vs descent speeds Ascent b) Straightness indices Ascent Descent Plateau 1.0 Descent 0.03 0.8

)

-1 0.02 0.6

index 0.4

Straightness Speed (ms Speed 0.01 0.2

0.00 0.0 10 c) Plateau vs deep activity Plateau d) Plateau vs travel time Plateau 0.8 Deep Travelling 8

0.6 6

0.4 4

Activity

(hours)

Duration

(meanlength) step 2 0.2

0 0.0 R. brachyura R. clavata R. microocellata R. montagui R. brachyura R. clavata R. microocellata R. montagui Species Species Figure 46: Event movement analysis a) Average ascent/descent speeds. No significant differences were found between the ascent and descent speeds of any species; b) Ascent, plateau and descent straightness indices. While there are no significant differences between the ascent and descent straightness indices, differences between the ascent/descent and plateau indices are significant in all species except for R. montagui; c) Plateau vs deep. Activity during the plateau phase is, in all cases, significantly greater than in the deep phase; d) Plateau vs travel time. The plot shows the time spent in the plateau phase compared to the total time spent travelling between the deep and plateau depths. In all cases, except R. montagui, more time is clearly spent in the plateau phase than in the combined ascent and descent phases.

118

5.3.3 Temperature as a possible driver The average deep and plateau temperatures were analysed to determine whether the skates were migrating to deeper colder waters for energetic advantages (Figure 47a). However, no significant temperature differences were found in any case (Table 19). Interestingly, R. clavata experienced significantly warmer temperatures than the other species during both the plateau and deep phases (Kruskal-Wallis One Way Analysis of Variance on Ranks, p=0.009).

5.3.4 Depth ranges The average depths of the deep and plateau phases were analysed in order to identify species differences (Figure 47b). Significant differences were found in the deep phase across species (Kruskal-Wallis One Way Analysis of Variance on Ranks, p<0.001). A pairwise multiple comparison (Dunn’s method) indicated significant differences (p<0.05) between the two pairs R. brachyura / R. montagui and R. clavata / R. microocellata. However, differences in the shallower, plateau depth, although significant (Kruskal-Wallis One Way Analysis of Variance on Ranks, p=0.04) were less pronounced, with pairwise comparisons only being significant between R. montagui and R. clavata / R. microocellata. There was therefore more overlap in plateau depths between R. brachyura and R. clavata / R. microocellata, with R. montagui occupying deeper depths during the plateau phase.

0 a) Plateau vs deep temperatures Plateau b) Plateau vs deep depths Plateau 16 Deep Deep 10

14 20

12 40

Depth (m)

Temperature °C 10 50

8 70

R. brachyura R. clavata R. microocellata R. montagui R. brachyura R. clavata R. microocellata R. montagui Species Species Figure 47: Plateau vs deep temperatures and depths a) Plateau temperature vs deep temperatures. No significant differences were found except for R. clavata where both temperatures were significantly higher than the other species; b) Differences in the deeper depths are as expected from Humphries et al. (2016b). Differences in the shallower, plateau depths are less pronounced, suggesting a greater overlap in putative foraging depths than expected.

119

The analysis of event frequency and start depth showed no clear overall pattern (Figure 48). However, significant differences were found between male and female start depths in both R. montagui and R. clavata (p=<0.001 in both cases, Kruskal-Wallis One Way Analysis of Variance on Ranks). In R. montagui, females started events from deeper depths than males, while R. clavata females started events shallower than males. In R. brachyura and R. clavata the peak in events coincides with the preferred depth, as identified by Humphries et al. (2016b), however with R. microocellata and R. montagui the peak in event frequency occurs in shallower waters. These plots also show a significantly greater propensity for DVM in male R. brachyura and R. montagui (N=674, df=1, Χ2=26.47, p < 0.001).

0.030 a) R. brachyura b) R. clavata 0.05 Female Male 0.025 0.04 0.020 0.03 0.015

0.02 0.010

0.01 0.005

10-20 20-30 30-40 40-50 50-60c) R. microocellata60-70 70-80 10-20 20-30 30-40 40-50 50-60 d) R.60-70 montagui70-80 0.035 0.025 X Data X Data 0.030 0.020

Events per day tracked per day Events

0.025 tracked per day Events 0.015 0.020

0.015 0.010 0.010 0.005 0.005

10-20 20-30 30-40 40-50 50-60 60-70 70-80 10-20 20-30 30-40 40-50 50-60 60-70 70-80 Event start depth(m)

Figure 48: Frequency of events by start depth for species and sex. The vertical dashed line indicates the preferred mean depth for each species (from Humphries et al., 2016b).

5.3.5 Timing of events To determine whether light was a possible driver for the movements the event start and end times were analysed to give counts of event starts and ends per hour. The results (Figure 49) were plotted together with the local times of sunrise and sunset. A clear correlation was observed

120 between the time of sunset and the start of the events, with event start times being in most cases just after sunset. Event end times were correlated with dawn and occurred mainly after sunrise.

24 Start 24 End

20 20

16 16

12 12

Hour ofHour the day 8 8

4 4

0 0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Scale

Month Figure 49: Timing of events in relation to sunrise and sunset times Throughout most of the year events can be seen to start (green circles) just after sunset and to end (blue circles) just after sunrise. Bubble size indicates the number of events recorded, single events have been omitted for clarity.

5.3.6 Seasonality of events The seasonal event frequency differed significantly from the expectation of randomness in all but three cases (female R. brachyura and R. microocellata and male R. montagui) (Table 20). Figure 50 shows the differences between observed and expected event frequencies by species, sex and season. Peaks in event frequency occur throughout the year for different species and sexes; in particular in spring for R. brachyura males, summer for R. clavata (male and female) and R. microocellata males and autumn for R. montagui females. It is interesting that peaks do not occur in winter in any case and that lows occur in spring for R. microocellata females and R. montagui males. Female R. brachyura and R. montagui and male R. microocellata exhibit the most consistent event frequency throughout the year.

121

a) R. brachyura b) R. clavata Female 40 15 Male

10 30

5 20

0 10 Observed events - Expected

0 -5

-10 -10

20 c) R. microocellata d) R. montagui

5 10

Observed - Expected events Observed 0 0

-5 -10

-10 -20

Winter Spring Summer Autumn Winter Spring Summer Autumn Season Figure 50: Observed – Expected event counts by species, sex and season The plot highlights the differences between the observed and expected seasonal frequency of events, where bars above the line indicate higher, and below the line lower, frequencies than expected if there was no seasonal pattern. The expected counts are those determined by the Χ 2 analysis.

Table 20: Χ 2 results from the seasonal analysis, significant results highlighted The results confirm that in most cases the event frequency has seasonal highs and lows that differ significantly from the expectation of randomness.

Species Sex Χ 2 N df P R. brachyura Female 4.3667 25 3 0.224 R. brachyura Male 42.0378 37 3 <0.001 R. clavata Female 21.2486 120 3 <0.001 R. clavata Male 80.8832 174 3 <0.001 R. microocellata Female 34.9508 156 3 <0.001 R. microocellata Male 0.8225 84 3 0.844 R. montagui Female 2.2066 24 3 0.531 R. montagui Male 21.5625 54 3 <0.001

122

5.4 Discussion

Movements, including longer term shifts in depth to areas further inshore or offshore, are expected in active benthic predators such as skates as part of seasonal migrations (Hunter et al., 2005a, Hunter et al., 2005d). Regular diel vertical migrations (DVM) however, from deep resting or refuging areas to shallow foraging areas are unexpected. Nonetheless it is clear that all the rajid species in this study do so and in a similar way.

5.4.1 Benthic rather than pelagic movement Both the specialised morphology and the benthic diet indicate that skates are strongly adapted to a benthic lifestyle. While foraging, skates use their crura to ‘walk’ along the sea bed and manoeuvre to search for buried prey using their electroreceptive sense and only to change to swimming as an escape response when alarmed (Koester and Spirito, 2003, Lucifora and Vassallo, 2002). At high swimming speeds (around 2 body lengths/s) energetic costs increase significantly (Di Santo et al., 2017) and consequently sustained high speed swimming, as might be required for pelagic hunting, is unlikely. The lower energetic cost of ‘walking’, in comparison to even slower swimming (1 BL/s) explains why most observed movements along the bottom involve walking using crura rather than swimming using undulations of the pectoral fins (Di Santo et al., 2017, Macesic and Kajiura, 2010). Additionally, as walking involves fewer muscles, it provides a quieter electrical environment while foraging (Macesic and Kajiura, 2010, Di Santo et al., 2017).

The speed of the ascent and descent phases also suggests that the inward and outward migrations are likely the result of skates traversing the sea bed at ‘walking pace’. Firstly, the speeds, in terms of vertical displacements, are in all cases very low, with median ascent and descent speeds being < 1 cm/s (Figure 46a). Median ‘walking’ speeds have been shown to be around 14.5 cm/s, or about one third body length per second (Koester and Spirito, 2003). If the skates studied here are moving over the sea bed it is expected that the maximum rate of vertical displacement would be significantly less than the rate of horizontal displacement. The low measured speeds are therefore expected, however at more than an order of magnitude less than ‘walking’ speed the results do suggest that the skate’s horizontal speeds are also slow, suggesting not only that the migrations are benthic but also that they may be meandering rather than directed.

The analysis of tortuosity (SI index) of the ascent and descent phases again suggests that the skates are not making directed movements typical of the DVM of pelagic predators (Humphries et al., 2016a, Schaefer and Fuller, 2010, Shepard et al., 2006). Instead, the high tortuosity suggests benthic movement over complex sea bed topology, possibly opportunistically searching and

123 foraging. In this way the movements are different from DVM in pelagic predators but clearly resemble those of other benthic predators such as burbot Lota lota or catsharks Scyliorhinus canicula (Gorman et al., 2012, Cott et al., 2015, Harrison et al., 2013, Sims et al., 2006).

Finally, it is important to note that the average duration of a DVM event is 7.3 hours, which would represent a considerable expenditure of energy if the entire event was pelagic rather than benthic.

Consequently, the majority of movements vertical displacements recorded by the tags occur due to horizontal benthic movements over complex topology and therefore it is reasonable to refer to the events subsequently as nektobenthic DVM.

5.4.2 Nektobenthic DVM as possible central place foraging Central place foraging is typically observed when animals have conflicting requirements for foraging and other behaviours. A common example is nesting sea birds, such as northern gannets, Morus bassanus or black browed albatross, Diomedes melanophrys, where birds need safe terrestrial locations for nesting, yet forage in open seas at long distances from the colony (Patrick et al., 2014, Weimerskirch et al., 1997). In other cases, foraging areas can only be visited temporarily, as is the case with air breathing marine predators such as seals (Burns et al., 2008) or penguins (Wilson et al., 1993). For many animals, for example Cape fur seals, Arctocephalus pusillus pusillus (De Vos et al., 2015) it is predation risk that affects foraging behaviour. In these examples animals forage in areas in which they cannot permanently reside and therefore must incur the significant cost of movement between these areas or the risk of predation. The skates studied in this work are temperate benthic predators that are unlikely, as adults, to be subject to significant predation pressure in the area studied (Scharf et al., 2000). It was therefore expected that the depths or habitats occupied would be those that offer the best feeding opportunities (Krebs and Davies, 1997). While considerable movement resulting from exploration was expected, it was not expected that skates would leave foraging areas and return to apparent refuge areas in the way frequently observed with known central place foragers (e.g. Lawton, 1987).

Due to the similarity of the events selected and the DVM of pelagic predators, the hypothesis explored here is that these movements represent a form of central place foraging and several lines of evidence support this. For example, the difference in activity between the deep and plateau phases suggests that during the deep phase skates exhibited limited movement and were resting, while during the plateau phase they were more actively exploring the sea bed. Further support for central place foraging is provided by the presence of a clear tidal signal observed in the deep phase of many of the tracks but which is rarely evident in the plateau phase. A clear tidal

124 signal indicates that an individual is remaining at a relatively constant depth on or close to the sea bed. Some teleosts, such as cod Gadus morhua, can remain neutrally buoyant at the same depth above the benthos, allowing a tidal signal to be detected by tags (Pedersen et al., 2008). Skates lack swim bladders, are negatively buoyant and therefore cannot maintain a depth above the seabed without active swimming. During the deeper phase the skates are therefore most likely resting on the sea bed and, if they are moving, it is with considerably less activity then during the shallow plateau phase. Given that skate’s often bury themselves in soft sediment (Kotwicki and Weinberg, 2005), resting on the sea bed would seem to be the most likely behaviour to be occurring during the low activity deep phase observed between nektobenthic DVM events. Consequently it is reasonable to conclude that the deep phase represents resting or refuging behaviour, while the plateau phase represents foraging.

5.4.3 Potential drivers of nektobenthic DVM There are several hypotheses that could be considered as possible drivers of the nektobenthic DVM behaviour, for example, behavioural thermoregulation, predator avoidance, prey tracking or simply the avoidance of light. A potential driver for skate nektobenthic DVM appears to be the light levels, with most events beginning just after sunset and ending shortly after sunrise. There is little direct evidence in the literature for nocturnal activity patterns in skates, except for Hove and Moss (1997), Auster (1995) and some trawling studies (e.g. Appa Rao and Krishnamoorthi, 1982, Casey and Myers, 1998a) where catches of skates were greater at night. However, it is clear from this study that the nektobenthic DVM events, and the higher rates of activity, occur mainly during night time hours. Casey and Myers (1998a) suggest that the increased catchability of skates at night is due to them being unable to visually detect the oncoming trawl net. However, skate eyes are well adapted for low light and nocturnal activity (Murphy and Howland, 1990). A possibility, suggested by this study, is that skates performing nektobenthic DVM are moving to areas for daytime resting where trawling is not undertaken. Indeed, diel differences in distribution have been previously recorded by video transect survey in the little skate Raja erinacea, with microhabitat association being observed during the day but with more random and dispersed distributions during the night, consistent with resting or refuging during daylight and active foraging and searching during the night (Auster et al., 1995).

It is interesting that the regularity of the nektobenthic DVM events observed here is quite unlike that frequently observed for pelagic marine predators, such as swordfish (e.g. Xiphius gladius) where it is usual for DVM to be performed by every individual every day (Schaefer and Fuller, 2010, Evans et al., 2014). The skates studied here only performed sporadic nektobenthic DVM. From the sporadic nature of the events it would appear that while light is a trigger for the onset or

125 end of the movements, it may not be the causal driver (Ringelberg and Van Gool, 2003). In other words, while changing light levels might trigger the start or end of an event, the underlying motivation is more likely to be the need to find more suitable foraging habitat (at the start of event) or refuging habitat (at the end). The skates may therefore be responding to changing needs for environmental conditions (prey availability, substrate type) or internal states (hunger, satiation, exhaustion) on a case by case basis, with the timing of the response triggered by a light. There must therefore be differences, between the area in which the animals prefer to forage and areas they prefer when resting, with this driving the motivation to move. One possibility is that the animals prefer a particular substrate (sand, mud or gravel) when burying for resting, but this offers few opportunities for foraging. However, there is a gap in our knowledge regarding preferred substrates, as skates are reported as being associated with a range of habitats, from mud and sand through to coarse gravel (Wheeler, 1969a, Fock, 2014). Precisely which habitats are preferred for different activities remains unknown.

While there was no common seasonal pattern for the onset of nektobenthic DVM behaviour, there was less activity in winter for all species and increased activity for R. clavata, R. microocellata and R. montagui in the warmer months of summer and autumn. Indeed for R. clavata nektobenthic DVM was mainly observed in summer months which likely explains the higher overall temperatures recorded for the events in this species (Figure 47a). With the exception of R. brachyura, it is possible that the observed increase in nektobenthic DVM is a result of these species being more active overall in warmer water. Warmer water temperature will increase metabolic rate of ectothermic skates and consequently increase the requirement for food, possibly driving increased foraging activity. However, it is not clear why large predators such as skates should choose not to remain in favourable foraging grounds, but instead return each day to deeper waters.

One possible explanation for the spatial separation between foraging and resting phases could be energy conservation through behavioural thermoregulation. In the catshark Scyliorhinus canicula behavioural thermoregulation was suggested to be the driver of nektobenthic DVM (Sims et al., 2006). Catsharks hunted in warmer, shallow waters at night, but returned to deep, cooler waters at night to rest and digest (Sims et al., 2006). In contrast, blacktip reef sharks, Carcharhinus melanoptris, have been found to move into shallower water during daytime in order to warm and possibly increase rates of digestion, but hunt during the cool early evening (Papastamatiou et al., 2015). Temperatures recorded for the skates studied here however showed no significant differences between the deep and shallow areas occupied (Figure 47a). The reason for a lack of a relation between temperature and depth is most likely the strong tidal mixing that occurs in these

126 coastal waters, making them different from stratified pelagic waters (Pingree and Griffiths, 1978). Therefore, behavioural thermoregulation is unlikely to be a driver for the activity observed here.

The return to deeper, darker waters during daylight hours could be motivated by predator avoidance, as is the case with many pelagic fish (e.g. Sutton and Hopkins, 1996). However, the four skates studied here are considerably larger than any other benthic predators in this coastal assemblage (Ellis et al., 2005c) and are unlikely to be influenced by predation in the area studied. Consequently they are less likely to choose protection from predators over foraging. Over the last 100 years populations of other elasmobranchs, including possible predators of skates (such as the common skate, Dipturus spp.) have declined precipitously (Ellis et al., 2005b, Dulvy et al., 2000). Therefore, although predator avoidance seems unlikely, it is possible that the observed nektobenthic DVM represents a vestigial behavioural trait evolved to reduce the risk of predation (Mehner et al., 2007). However, as discussed by Mehner (2007), this behaviour would be expected to produce a more consistent pattern of movement than is observed here.

Given that the skates are one of the apex predators in this coastal, demersal ecosystem, another possibility is that these movements are being exhibited to follow migrating prey. Skates are opportunistic generalist predators with diet comprised of mainly fish and crustaceans (Ellis et al., 1996c, Pinnegar, 2014, Šantić et al., 2012, Farias et al., 2006a) with typical prey being crabs (e.g. Liocarcinus, Carcinus spp.) or sand eels (e.g. Ammodytes marinus). Sand eels are nocturnal and spend the autumn and winter buried in sediments (Greenstreet et al., 2010). This might make them easy prey for skates that can use their electroreception to detect them (Wueringer, 2012). Interestingly, it is the two species (R. microocellata and R. montagui) with a greater preference for fish in their diets (Wearmouth et al., 2014a, Pinnegar, 2014) that perform more nektobenthic DVM in autumn. In the North Sea, there is evidence of sand eels performing DVM, occupying pelagic habitats during the day and returning to shallower waters at night (van der Kooij et al., 2008). If sand eels perform similar migrations in the area studied here, then this might represent a possible driver for the observed events. Similarly, crabs (e.g. Liocarcinus depurator) have been shown to have diel activity rhythms, burying in substrate during the day and emerging and becoming active at night (Aguzzi et al., 2009, Aguzzi et al., 2015), along with other endobenthic and nektobenthic animals (Aguzzi and Company, 2010). The availability of such prey might be a driver of either inshore migrations or a trigger for increased activity in the skates. However, at present there remains no clear reason why the skates would then return to deeper water after each foraging bout.

Lastly, the events could possibly be linked to egg-laying; Rajidae are oviparous, laying eggs which are commonly found in high density in nursery beds within spatially restricted coastal areas, often

127 linked to substrate type (Love et al., 2008, Ellis et al., 2011, Serra-Pereira et al., 2014). The preferred nursery sites are located in shallower depths than the preferred depths in these species, and therefore movement to and from nursery sites might generate diel patterns such as those observed here (Ellis et al., 2005c). In some species the peak in egg laying frequency coincides with the peak in event frequency (Koop, 2005, Holden, 1975, Ellis and Shackley, 1995). In R. brachyura for example, the peak in DVM frequency is April- September, overlapping the period of peak egg laying (Koop 2005), suggesting a possible link between nektobenthic DVM behaviour and egg deposition. However, as previously noted, the events are observed more frequently in males, and while males could be following egg laying females to gain access to mating opportunities, the coincidence between nektobenthic DVM and egg laying peaks is not sufficiently consistent to support this hypothesis.

Light level, coupled with intrinsic behavioural rhythms, therefore remains the only clear trigger for the behaviour. This hypothesis is challenged, however, by skates frequently spending extended periods of time in shallow water. Skates often performed a migration to shallower water and remained there, despite the increased light levels during August and September and having previously performed regular nektobenthic DVM.

5.4.4 Conclusions The difficulties in obtaining horizontal movements through space and time, presents a significant obstacle in our understanding of the movement ecology of these important species. Through a detailed analysis of high resolution depth time-series data this chapter has identified novel diel inshore-offshore migrations that likely result in the animals having a wider dispersal and larger area of occupancy than might otherwise have been considered by other methods, such as those used in the previous chapters. The limitations of the depth time-series data hamper attempts to determine the factors that drive these rajids to perform these migrations; although some potential drivers, such as behavioural thermoregulation, can be discounted. From the patterns of activity observed here, and the comparisons with nektobenthic DVM in other marine predators, it seems most likely that the movements represent foraging excursions. Nektobenthic DVM observed also provides evidence for possible central place foraging by the skate species. This study cannot provide any clear motivation for central place foraging in skates however. The observations presented here suggest that skates are mobile at finer scales than were previously expected, making the effectiveness of spatial protected areas or fishing controls more difficult to accurately define. The diel mobility patterns of skates suggest that protected areas should be large enough to encompass these movements and that fishing controls should consider the nature of the nocturnal activity which likely makes the skates more susceptible to capture during

128 the night in either trawls or set nets. Finally, the common occurrence of nektobenthic DVM in the skates studied here suggests that nektobenthic DVM might be a common behaviour in many other benthic organisms, unobserved and unreported as yet because of the difficulties in determining the fine scale movements of temperate benthic animals.

129

130

Chapter 6. Resource partitioning and ontogenetic shifts in the trophic geography of sympatric skates (Rajidae) inferred from eye lens isotopic signatures

6.1 Introduction Species living in sympatric assemblages may partition their environment through the differential use of resources such as habitat and prey type (Schoener, 1983, Ross, 1986, Platell et al., 1998). Such resource partitioning can reduce competition between individuals of different species or age classes, potentially influencing population and community interactions and resource availability (Ross, 1986). Habitat partitioning for example will affect the spatial distribution of a species, while differences in prey preferences can allow predators to coexist (Speed et al., 2011). Resources may also be partitioned temporally, i.e. different species using the same habitat or consuming the same prey but at different times (Schoener, 1974). Resource partitioning is therefore expected to occur in assemblages containing several predators with similar diets and overlapping distributions, even if on a fine-scale.

The stable isotope compositions of carbon and nitrogen in animal tissues are routinely used as tracers for tropho-spatial ecology (Domi et al., 2005). Carbon isotope ratios vary between sources of primary production, which frequently differ between habitats, such as inshore and offshore zones and can therefore be used to infer feeding provenance (Burns et al., 1998, Craig, 1953, DeNiro and Epstein, 1978, Hobson et al., 1994). Variations in carbon isotope ratios among locations occur due to different modes of photosynthesis by primary producers, differences in the carbon isotope ratios in dissolved organic carbon or by differential fractionation during photosynthetic uptake and assimilation (Barnes et al., 2009). The extent of isotopic fractionation of carbon during marine photosynthesis is mainly controlled by the rate of dissolved organic carbon uptake, but growth rate, species differences and nutritional status also play a role. The concentration and isotopic composition of dissolved inorganic carbon and phytoplankton growth rates are also influenced by temperature which leads to predictable latitudinal gradients in δ13C values of marine phytoplankton (Trueman et al., 2012, Barnes et al., 2009, Magozzi et al., 2017).

Nitrogen isotope ratios in tissues of marine consumers also vary spatially and are controlled by the size of the nitrogen pool, the nutrient source at the base of the food web and the mechanism by which nitrogen is integrated into the tissues (Trueman et al., 2012, Jennings and Warr, 2003). Nitrate availability is also influenced by temperature and therefore has an effect on nitrogen

131 isotope ratios. Other influences on nitrogen ratios include coastal proximity and pollution (Cabana and Rasmussen, 1996, Cole et al., 2004).

Preferential excretion of 12C and 14N leads to enrichment of consumer tissues in the heavier isotope (13C and 15N) compared to their prey. The combination of spatial and physiological differences in isotopic ratios at the base of the food web, and relatively predictable levels of trophic enrichment (of around 1‰ and 3.0‰ in carbon and nitrogen respectively), provide the basis for reconstruction of trophic level diet sources and food web structure from stable isotope compositions (Bearhop et al., 2004, Inger and Bearhop, 2008, DeNiro and Epstein, 1981, Burns et al., 1998). However the combination of combined spatial and trophic influences on consumer tissue isotopes also complicates their interpretation. The term ‘trophic geography’ (Bird et al. (2018) refers to the integrated effects of differences in space and diet type. Two animals with different tissue isotopic compositions therefore may be separated by diet type, area of foraging or both – and can be defined as showing different trophic geographies, implying resource and/or habitat partitioning.

As tissues grow and turnover at different rates, the selection of different tissues for isotope analysis can provide information about an individual’s trophic geography over varying time scales (Bearhop et al., 2004, Hobson and Clark, 1992). Muscle tissue for example can provide information about foraging location and trophic level over longer periods (a few months to a year), while liver or blood samples can provide the same information over a much shorter time period (weeks) (Malpica-Cruz et al., 2012, Tieszen et al., 1983). Metabolically inert tissues retain the isotopic composition of the organism at the time of tissue formation (Wallace et al., 2014, Schell et al., 1989, Hobson, 1999). Metabolically inert tissues that grow in increments across the life of animal can therefore provide a biochemical record of an individual’s trophic geography throughout ontogeny, and tissues such as otoliths provide a wealth of biochemical information relating to fish growth, movement, stock structure and life history ecology (Gillanders and Kingsford, 2000, Trueman et al., 2012). Groups such as the Elasmobranchs however, do not possess large calcified otoliths (Cailliet et al., 1986) and most do not have fin spines that can be used as tissue samples in stable isotope analysis. While elasmobranch vertebrae have been used to recover isotopic life history records (Estrada et al., 2006), sequential sampling of elasmobranch vertebrate can be challenging, particularly in smaller species (Quaeck, 2017).

The eye lenses of vertebrates are also metabolically inert issues, which grow over the lifetime of an animal by circumferential addition of layers of protein (laminae) to the lens surface (Dove and Kingsford, 1998, Horwitz, 1992, Smelser, 1965). The lens begins to grow during the early days of

132 embryonic development (Grainger et al., 1992), with subsequent layers added on top. This leaves a chronological biochemical record that can potentially be used to determine foraging behaviour throughout ontogeny, with the lens core being the earliest days of the animal’s life (Hunsicker et al., 2010, Wallace et al., 2014). In oviparous species, the core of the eye lens therefore forms from proteins assimilated by the mother during egg provisioning (or gestation in the case of placental animals). After hatching or at the onset of independent feeding, the isotopic composition of subsequent layers of lens protein reflect the isotopic expression of foraging (Olin et al., 2011). The use of eye lens tissue is relatively novel in isotope analysis; however eye lenses have been successfully used to determine the feeding ontogeny of the commander squid Berryteuthis magister (Hunsicker et al., 2010), and several teleost and chondrichthyan species (Wallace et al., 2014, Quaeck, 2017).

The Rajidae (skates) are commercially valuable mesopredatory elasmobranchs that form sympatric assemblages in coastal regions and are generally considered to exhibit site fidelity (Steven, 1936). Studies of their broad-scale spatial ecology in northeast Atlantic waters around the UK have provided important snapshots of Rajidae distribution (Martin et al., 2010) and depth ranges for some species have been established using fishing surveys (Ellis et al., 2005a). These depth ranges suggest that most of the common species of Rajidae in the UK have an overlapping coastal distribution, with some species such as R. brachyura exhibiting a more southerly distribution (Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010).

While broad-scale spatial patterns of distribution are relatively well-established, fine-scale details of how sympatric species utilise shared space are less clear (Wiens, 1989). This fine-scale information is especially important given that Rajidae are mobile, have patchy distributions, and are thought to remain in relatively localised coastal areas (Steven, 1936, Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998, Martin et al., 2010). Recently, an electronic-tag release study revealed habitat partitioning between four species of rajid skates in a small coastal area (Humphries et al., 2016c). However, the archival pressure data-logging tags used provided only depth preferences and there may be further habitat partitioning between the species within their preferred depth range which is missed by depth-only recording tags. Furthermore, habitat partitioning between juveniles and adults is often missed in tagging studies and fishing surveys due to limitations on the size of individuals that can be tagged, and because neonates can escape through fishing nets with wider meshes. Our understanding of the distribution of juvenile Rajidae is therefore limited (Ellis et al., 2005a), even though rajids are known to be oviparous and are

133 thought to deposit eggs in nursery areas, where the juveniles remain until adulthood. Locations of skate nurseries around the UK have been proposed to be broadly in shallower waters than adult distributions (Ellis et al., 2005a) suggesting that habitat partitioning may be an important factor separating juveniles and adults, and perhaps further spatial differences between species.

With regard to feeding ecology and potential resource partitioning by diet, Rajidae are considered to be generalist opportunistic feeders and while some species have dietary preferences, there is a high degree of dietary overlap (Ellis et al., 1996a, Holden and Tucker, 1974). Ontogenetic shifts in diet are commonly reported in Rajidae; for juveniles, mysids and amphipods are important, but with increasing size and age there is a shift from such smaller prey items to larger crustaceans, fish and cannibalism (Ajayi, 1982a, Holden and Tucker, 1974, Ellis et al., 1996a, Farias et al., 2006b, Moura et al., 2008).

To identify trophic level and dietary preferences in Rajidae, studies have largely used stomach contents analysis (Ajayi, 1982a, Holden and Tucker, 1974, Ellis et al., 1996a, Farias et al., 2006b, Moura et al., 2008). These methods can work well for adults of a species, however juveniles are often excluded, especially newly hatched individuals, as they may be missed by commercial fisheries and may be too small to perform reliable stomach contents analysis (Moura et al., 2008, Coelho and Erzini, 2006). Juveniles are frequently analysed as broad groups (Ellis et al., 1996a), often simply before and after maturity (Farias et al., 2006b), which potentially misses ontogenetic shifts in the early life of a rajid. Establishing trophic level through stomach contents analysis also has several caveats including the identification of partially digested prey, differential rates of digestion of prey types and regurgitation of prey on capture (Estrada et al., 2006). To determine trophic ecology through ontogeny with stomach contents analysis, large numbers of individuals at different life stages must be sampled, which may be ethically problematic in species’ populations that are considered threatened. The success of juveniles in a population is vital for the survival of a species and governs the rate at which a population may recover from exploitation pressure or environmental perturbation. This is especially the case for threatened skates which are late maturing and have relatively low fecundity (Dulvy et al., 2014, Field et al., 2009).

Stable isotope analysis of Rajidae eye lenses therefore has the potential to provide an insight into juvenile feeding ecology, an insight that has progressed slowly using traditional methods. We aimed to investigate the use of eye lenses in the study of tropho-spatial ecology of Rajidae. We provide the first life history isotope records for sympatric Rajidae derived from eye lens proteins, and test the hypothesis that resource partitioning occurs between the four species and between juveniles and adults within species. Our results therefore provide an insight into any ontogenetic

134 shifts in feeding ecology of four skate species that may have been missed by stomach contents studies.

6.2 Methods This study is based on four sympatric skate species, spotted ray Raja montagui, small eyed ray R. microocellata, blonde ray R. brachyura and thornback ray R. clavata, which are relatively common Rajidae species within the study site off Plymouth, UK (Figure 51). As part of a tagging programme aimed at revealing skate movements and migrations, archival data-logging tags were returned via the local commercial trawl and net fisheries for a reward (£50 per tag) together with information concerning its capture location and date of recapture. Occasionally however, fishers returned a tag still attached to a dead skate, which serendipitously enabled us to sample eye lenses.

Length, disc width and weight were recorded for returned skates and tissues were removed for isotope analysis. Whole eyes were removed from 50 dead individuals and muscle samples were removed from 47 individuals (Table 21). Muscle was removed from the base of the tail, avoiding the skin and both muscle and eye tissues were stored at -20˚C until processing took place. Despite described differences in the maximum attainable length for each of the four species (McCully et al., 2012), there were no significant differences in total length between individuals of species sampled in the current study (ANOVA on ranks, Kruskal-Wallis, H=6.707, df=3, P=0.082).

Table 21: Summary of sample sizes and median capture lengths for each species

Species Muscle Male/Female/ Eye lens Eye lens Mean eye Median tissue Unknown individuals samples lens total length samples (Muscle tissue samples of mature samples) per individuals individual (mm) R. brachyura 12 6/6 14 97 6.92 743 R. microocellata 11 3/6/2 8 49 6.13 671 R. montagui 11 5/6 11 66 6 636 R. clavata 13 2/11 17 111 6.53 774

Total 47 16/29/2 50 323 6.46

135

Figure 51: Study location off Plymouth, south-west UK. Filled circles represent the results from a tagging study (green= release, red=recapture) defining the spatial extent of the study area.

In the laboratory, whole eyes were thawed and a small incision made in the cornea to remove the eye lens. The outer gelatinous layer of the lens could not be sampled systematically and was removed prior to analysis. Under a microscope, layers of the lens were removed (delamination) using forceps and a scalpel, which were cleaned using ethanol between each layer to avoid contamination. The lens diameter (excluding the hydrated outer layer) was measured using Vernier calipers before each layer of the lens was removed. Between 5 and 9 layers (mean = 6.46) were removed from each lens depending on the starting diameter. Hence, 50 eyes produced 323 layers which were stored individually in sample tubes and frozen.

Samples were freeze-dried for at least 6 hours, then both muscle and eye lens tissues were pulverised, weighed using a microbalance to between 0.65-0.75 mg and sealed into tin capsules. Isotope ratios of carbon and nitrogen were determined at the SEAPORT stable isotope facility at the University of Southampton using an Isoprime100 isotope ratio mass spectrometer (Isoprime Ltd, Cheadle, UK) coupled to a Vario ISOTOPE select elemental analyser (Elementar, Cheadle, UK). All stable isotope data are reported in delta (δ) notation. The internationally certified standards (US Geological Survey), Glutamic acid (SGA40) and Protein (Casein) Standard OAS (Batch no. 114859) were run at intervals between the eye lens samples to determine analytical accuracy and precision.

136

Eleven eyes were also removed from neonate skates (6 R. montagui, 5 R. microocellata) that died after hatching from eggs laid in experimental aquaria by adult females tagged as part of a previous study (Humphries et al., 2016c). Eye lenses were measured using Vernier calipers to provide an estimate of the eye lens diameter at hatching, which could be used to identify the likely position within each lens sampled from adult skates that corresponds to hatching.

6.2.1 Statistical analysis The relationship between eye lens outer diameter and body length was tested by species using ANCOVA. An ANOVA on ranks (Kruskal-Wallis) was conducted to test for differences in isotopic compositions among species in three grouped ontogenetic phases (core, post hatching and adult). The ‘core’ was classified as tissue from diameters smaller than the average diameter of the neonate eye lenses, the ‘post hatch’ group contained the remaining eye lens tissue and the ‘adult’ group contained the muscle tissue data.

Within-group isotopic variability was quantified using Bayesian standard ellipse areas (SEAc) for each of the ontogenetic phases (core, post hatch and adult) using the SIBER package in R (version 2.14.1) (R Development Core Team, 2015, Jackson et al., 2011). Isotopic niche area and overlap (‰2) were estimated based on 100 000 posterior draws of the SEAc. Sequential samples of eye lens proteins recovered from eye lenses were analysed to generate isotopic time series (life history profiles). To characterise temporal patterns in δ15N and δ13C values across eye lens diameters, Generalized Additive Mixed Models (GAMM) were run in R with the mgcv package (Version 1.8-18). The weighted Akaike Information Criterion (wAIC) and the deviance explained was used to select the appropriate model. In the models, species and eye lens diameter were used as fixed effects and individual as a random effect to account for the repeated measures sample design.

To test for differences in the magnitude of ontogenetic variation in δ15N values among species, the difference between maximum and minimum eye lens δ15N values was calculated for each individual. All comparisons were explored using ANOVA with Tukeys post-hoc multiple comparison tests. All analyses were conducted in R (Version 1.0.143).

6.3 Results

6.3.1 Allometry and the use of Rajidae eye lens tissues in stable isotope analysis There was a significant positive correlation between eye lens diameter and total length (F= 95.135, df=1, P<0.001) (Figure 52). Nevertheless, accurate measurement of the full diameter of

137 lenses was difficult due to the hydrated gelatinous outer layer, particularly after freezing and thawing. Consequently, measured lens diameters showed a large variation, and while it was clear that lens diameter increased consistently, and probably linearly with body size, it was not possible to estimate accurately adult size (or age) from lens diameters in the case of the four Rajidae species. The neonate eye lenses were less hydrated than the adults and the outer layers were retained, therefore the observed range in eye lens diameters in neonates (1.75-2.5 mm) was likely more representative and provides a good estimate of eye lens diameter at hatching (2.11 mm). Eye lens diameter values <2.11 mm were therefore considered to be the ‘core’ of the eye, where the isotopic composition of the lens core represents maternal foraging during egg provisioning. In general eye lenses delaminated well with, on average, about 6 layers removed per individual, which provided a large enough sample for analysis as well as a relatively fine-scale isotopic life history.

Figure 52: The linear relationship between the total length (mm) of an individual on recapture and their maximum eye lens diameter (mm). Standard error is represented by the grey areas. Each data point plotted denotes an individual skate.

6.3.2 Isotopic differences among species at throughout ontogeny No significant differences in δ15N values were found among species during the adult phase (P=0.115). There were, however, differences among species during the core and juvenile phases (Figure 53, Table 22). R. montagui in particular showed lower δ15N values, especially in lens core regions. Lens core δ13C values did not differ among species (P=0.27), although there were differences detected between the adult and juvenile phases (Figure 54, Table 22).

138

Figure 53: Differences in δ15N values among species at 3 ontogenetic stages. The box represents the 25th and 75th percentile, the whiskers are the 5thand 95th percentiles and the horizontal line is the median. No difference was found in the adult phase, but differences occurred in both the juvenile and pre-hatch phase.

Figure 54: Difference in δ13C among species at 3 ontogenetic stages. The box represents the 25th and 75th percentile, the whiskers are the 5th and 95th percentiles and the horizontal line is the

139 median. No significant differences were found in the pre-hatch phase, but differences were found in the adult and juvenile phases.

Table 22: Statistical differences in δ15N and δ13C values among species at 3 ontogenetic stages

Stable Species comparison Core Juvenile Adult Isotope P values P values P values δ13C R. clavata - R. brachyura 0.0450 * < 0.001 * R. microocellata - R. brachyura 0.27 0.0832 0.01811 * R. montagui - R. brachyura No significant 0.6774 0.00924 * R. microocellata - R. clavata differences 0.9863 0.00788 * R. montagui - R. clavata 0.0026 * 0.00570 * R. montagui - R. microocellata 0.0093 * 0.99986 δ15N R. clavata - R. brachyura 0.9625 <0.001 * R. microocellata - R. brachyura 0.6562 0.467 0.115 R. montagui - R. brachyura 0.0744 <0.001 * No significant R. microocellata - R. clavata 0.8807 0.387 differences R. montagui - R. clavata 0.0302 * <0.001 * R. montagui - R. microocellata 0.0116 * <0.001 *

6.3.3 Standard ellipse area (SEAc) analysis In the adult phase, isotopic SEAc values suggested that R. clavata has a significantly broader isotopic niche than R. montagui and R. microocellata. Overlap indices suggested distinct isotopic niches for all species, especially R. brachyura where no overlap with any other species was indicated (A-Figure 55; Table 3).

In the juvenile phase, R. clavata had a larger SEAc value than R. brachyura, while R. microocellata had a larger SEAc than R. montagui. We also found greater isotopic overlap between all species during the juvenile phase, the greatest difference between species was R. montagui and R. clavata with an index of 0.23 (B-Figure 55; Table 23).

In the lens core reflecting tissue synthesis before hatching, there were no significant differences in SEAc among the four species, but R. montagui had a distinct isotopic niche, with an index of 0 overlap with every other species (C-Figure 55; Table 23).

140

Figure 55: Standard ellipse areas and convex hulls by species. A= Adults/muscle tissue, B=Juvenile/eye lens tissue, C=Pre-hatching/eye lens core

141

Table 23: The difference and the overlap in standard ellipse area (SEAc) for each stage of development

Stage SEAc differences P value Overlap SEAc Overlap indices Muscle (Adult) R. clavata & R. 0.016 R. montagui & R. brachyura 0 montagui R. microocellata & R. 0.13 montagui R. clavata & R. 0.05 R. montagui & R .clavata 0.14 microocellata R. clavata & R. microocellata 0.1 R. clavata & R. brachyura 0 R. brachyura & R. 0 microocellata

Eye lens R. brachyura & R. 0.03 R. montagui & R. brachyura 0.58 (Juvenile) clavata R. microocellata & R. 0.35 montagui 0.23 R. microocellata & R. 0.05 R. montagui & R. clavata 0.98 montagui R. clavata & R. microocellata 1.14 R. clavata & R. brachyura 0.85 R. brachyura & R. microocellata

Eye lens core No significant R. montagui & R. brachyura 0 <2.11mm (Pre- differences R. microocellata & R. 0 hatching) montagui R. montagui & R. clavata 0 R. clavata & R. microocellata 1.4 R. clavata & R. brachyura 0.83 R. brachyura & R. 0.60 microocellata

142

6.3.4 δ15N and δ13C and eye lens diameter Results from GAMMs showed species and eye lens diameter significantly influenced isotopic composition: δ13C (p= 0.048 and p<0.001 respectively; 50.7% deviance explained) and δ15N (p<0.001 for both species and diameter; 73% deviance explained). δ15N and δ13C values in the core of the lens (reflecting maternal foraging during egg provisioning) were higher than in the two post-hatch phases. Both δ15N and δ13C values reach a lifetime minimum at around 2-2.5 mm lens diameter in every species (δ15N Figure 56, δ13C Figure 57) and was more pronounced in δ15N values. After this minimum, values increased consistently with increasing eye diameter, and thus body size. Adult muscle tissue δ15N values were not significantly different to lens core values in R. microocellata, R. clavata and R. brachyura, but were higher in R. montagui. In both cases wAIC values were lower in the model with the individual as a random effect and in both cases were significant (p<0.001) (Figure 58, Figure 59).

143

Figure 56: δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases. Note the similar pattern across all species, where δ15N reaches a low point shortly after hatching.

144

Figure 57: δ13C values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases.

145

Figure 58: Individual level δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases.

146

Figure 59: Individual level δ15N values across eye lens diameter and in muscle tissue. Red lines delineate ‘core’, ‘juvenile’ and ‘adult’ phases.

147

6.3.5 Magnitude of δ15N difference in eye lens tissue R. montagui had the most consistent lifetime δ15N values across eye lens tissue, with lifetime mean δ15N values significantly lower than seen in R. brachyura (P<0.001) and R. microocellata (P<0.001), which also gave the greatest ontogenetic change in δ15N across the eye lens (Figure 60, Table 24).

Change in in Change 1

R. brachyura R. clavata R. microocellata R. montagui

Species

15 15 Figure 60: The magnitude of change in δ N in the eye lens tissues (i.e.  Nmax-min). The box represents the 25th and 75th percentile, the whiskers are the 5th and 95th percentiles and the horizontal line is the median. Note the higher magnitude of difference in R. microocellata compared to R. montagui.

15 Table 24: Average  Nmax-min for each species

Species Mean Std Dev R. brachyura 2.338 0.792 R. clavata 1.777 0.712 R. microocellata 2.513 0.802 R. montagui 1.117 0.561

148

6.4 Discussion The use of vertebrate eye lens tissues as a host for stable isotope analysis is relatively novel, but has been demonstrated in several species (Wallace et al., 2014, Hunsicker et al., 2010, Quaeck, 2017). The current study is the first to use eye lens proteins from Rajidae as a host for stable isotope analysis, and examined potential resource partitioning and juvenile life histories in four sympatric species.

Eye lenses of skates were, on average, delaminated into around 6 layers per individual, providing a relatively detailed isotopic life history, particularly during the juvenile stages. The core of the lens represents protein assimilated by the mother during egg provisioning and therefore provides a unique cross-generational biochemical link. Relating the position of a sample within a lens to the body size of the individual (and thus linking isotopic samples to individual body size and age) depends on recovery of precise allometric growth relationships between lens diameter and body size. Difficulties associated with measuring lens diameters in defrosted lenses precluded accurate reconstruction of body size from lens diameters. In future, it would be beneficial if lens-body length calibrations were based upon fresh rather than frozen eye lens samples. Overall however, the eye lenses provided effective incremental tissues and have the potential to be used to compare trophic geographies of Rajidae across ontogeny.

6.4.1 Comparative trophic geography and resource partitioning Carbon and nitrogen isotope ratios in the muscle tissue of adults of four Rajidae species imply distinct isotopic niches for all species, especially R. brachyura, in which there was no overlap in SEAc. Variation between species occurred primarily in the carbon isotope ratios, suggesting that the four species were partitioned by habitat rather than differences in their trophic ecology. A recent tagging study in the same location also partitioned the four species into two pairs, suggesting a preferred shallower depth and an inshore preference of R. clavata and R. microocellata and deeper depths or offshore preference of R. brachyura and R. montagui (Humphries et al., 2016c). Collectively, these findings provide support for the existence of fine- scale habitat partitioning between these four sympatric species.

The current study however, suggests further partitioning between species than could be inferred from depth archival tags. R. clavata and R. microocellata had distinct isotopic niches, but used similar shallow-water habitats in the tagging study. Of the four species investigated in this study, R. clavata is the only species that routinely associates with estuarine habitats (Hunter et al., 2005b, Hunter et al., 2005c), as well as marine and more varied substrates (Steven, 1932a). R. clavata also had a significantly wider adult isotopic niche than R. microocellata and R. montagui.

149

This observed difference in the current study is supported by data of the reported prey of R. clavata, which ranges from entirely crustaceans in Carmarthen Bay (Ajayi, 1982a) to a more piscivorous diet off the coast of Africa (Ebert et al., 1991, Smale and Cowley, 1992). In addition, studies have shown that off the African coast, R. clavata had a more diverse diet than any of 14 other species of skate (Ebert et al., 1991). Combined isotopic and tagging methods provide strong evidence for a wider niche and more individual variation in trophic geography in adult R. clavata relative to the other rajid species studied here.

The eye lens tissues that represented the juvenile phase indicated a greater overlap in isotopic niche between species than during the adult phase. This is perhaps a result of the large proportion of time that this portion of the eye lens represents, which also captures the greatest degree of change or shifts in trophic geography during ontogeny. Despite this, there were distinctions in SEAc between R. montagui and R. clavata during this phase. Significant differences were found in both δ15N and δ13C values among species. Most notably, δ15N values in R. montagui were significantly lower than all other species during the juvenile phase. Differences in trophic ecology between species may be a result of differences in specialisation, preference or availability of prey items between species. Such differences may reduce competition for sympatric species whose young occupy overlapping habitats (Espinoza et al., 2012). Differences in spatial ecology between adult and juvenile Rajidae have previously been suggested, where nursery grounds are shallower than adult preferred habitat (Ellis et al., 2005a). The current study suggests however, that not all juveniles of these four species are foraging in the same habitat. Overall, the differences found in the current study suggest the occurrence of both habitat partitioning and trophic ecology play a role in partitioning these species at this life stage.

6.4.2 Maternal egg provisioning In Rajidae, embryonic eye lens proteins provide a biochemical record of the isotopic composition of food consumed by mothers during egg provisioning. Subsequently, all individuals contain a biochemical record of maternal foraging in the lens core and of their own foraging post-hatching or birth in subsequent layers of the eye lens (Olin et al., 2011). Therefore the lower nitrogen isotope values found in the core eye lens compared to the remaining lens tissues in R. montagui, likely represent a difference in the feeding ecology of the mothers. This is especially interesting because in the adult tissues there were no significant differences in nitrogen isotope ratios between species, perhaps suggesting that R. montagui females change their feeding behaviour during egg provisioning.

150

6.4.3 Trophic ecology across ontogeny In all species, δ15N values were higher in the eye lens core. Here, enriched values in the core of the eye represent the embryonic phase where, when individuals rely on the egg yolk provided by the mother for nutrition. After hatching (lens diameters >2.11 mm) values decrease reaching their lowest level at 2.0-2.5 mm, after which there a gradual increase in 15N in all species across ontogeny. Several studies have demonstrated a dietary shift in Rajidae from amphipods and crustaceans to fish across ontogeny (Ebert and Bizzarro, 2007, Smale and Cowley, 1992, Ellis et al., 1996a, Ajayi, 1982a) and the analysis of individual-level isotopic life histories recovered from lens proteins in the current study suggested that such ontogenetic diet shifts were evident in all four skate species. A commonly cited hypothesis to explain ontogenetic shifts in diet refers to the development in feeding apparatus (Costalago et al., 2012, MacNeill and Brandt, 1990). For example, the menhaden Brevoortia tyrannus shift from using teeth to using gill rakers for primary feeding during ontogeny, which coincides with a shift from a carnivorous to an omnivorous diet (June and Carlson, 1971). Whilst Rajidae do not undergo such a dramatic shift in feeding apparatus, their gape size and the strength of the jaw both increase with ontogeny in the same way as with many elasmobranchs (Motta and Huber, 2004). Another hypothesis for dietary shifts across ontogeny is that behavioural shifts cause a change in consumed prey, for instance, switches in prey type due to changing densities of prey in the habitat they occupy (Costalago et al., 2012, Tanaka et al., 2006). It is also likely that larger fish are more experienced or are more efficient predators than juveniles (Espinoza et al., 2012). In Rajidae, there is some evidence that oviposition occurs in nursery areas separate from adults (Ellis et al., 2005a), where the density of prey types is likely different. Overall, there appears to be evidence of both physiological and behavioural characteristics, such as gape size and foraging experience, which drive the ontogenetic shifts in diet observed in the current study.

Across ontogeny there was significant individual variation in δ15N, potentially representing the opportunistic or generalist nature of individuals from an early age. Across ontogeny there were also significant differences among species in the magnitude of change in δ15N values. R. microocellata had the largest difference in δ15N followed by R. brachyura. Interestingly, both of these species are suggested to have an adult diet containing more fish, while the other 2 species diets contain more crab (Ellis et al., 1996a, Ajayi, 1982a), indicating that this greater magnitude of change potentially reflects this shift to a more piscivorous diet by these two species.

6.4.4 Spatial ecology across ontogeny Carbon isotope ratios also varied across eye lens diameter, in a similar pattern to nitrogen isotopes, starting at a high value in the egg phase, reaching minimum values on hatching and

151 increasing across ontogeny. Differential habitat selection across ontogeny is common in elasmobranchs (Simpfendorfer et al., 2005, Grubbs, 2010) where birthing or nursery areas are often visited seasonally by females for oviposition or parturition. These habitat selections by different age groups may reduce intra-species competition between adults and juveniles. Juveniles are likely to be more vulnerable to predators, due to their smaller size, therefore habitat selection by juveniles may occur to avoid predators that adults of the species need not (Lima and Dill, 1990). There is also evidence that increasing body size increases activity space (McNab, 1963) and adults of a species may also have different energy demands that can only be obtained in different habitat to the juveniles (Grubbs, 2010). In this study, the change in carbon isotope ratios across the eye lens suggest that juveniles are partitioned by habitat from adults, and that juveniles disperse or migrate during development to different habitats.

6.4.5 Conclusions and implications The stable isotope composition of eye lens proteins provides promising insight into trophic geography and resource partitioning in Rajidae. Retrospective analyses from eye lens proteins allowed a comparative assessment of tropho-spatial ecology during cryptic early juvenile phases and into foraging behaviour during maternal egg provisioning. With future analysis of the development of Rajidae eye lenses and the rate at which eye lens layers are deposited as the individual grows, a greater understanding of the time frame a layer of eye lens represents will help to improve prediction about when important ontogenetic shifts may be occurring temporally. Equally, a greater understanding of the complex fine-scale habitat mosaic within this and other study areas, and with the construction of dynamic isoscapes, will all help to determine the likely habitats of these species.

Inter- and intra-species resource partitioning has important implications for management and conservation of skates because certain age groups or species may have differences in catchability. There are for example, no minimum size restrictions for commercial harvesting of Rajidae species in most of the UK’s waters, consequently, juveniles remain particularly vulnerable to capture. Additionally, and perhaps more likely, fishers target the higher value, largest individuals (that also have the highest reproductive output), which can also have consequences for the population such as reduction in population growth rate, fecundity and offspring quality (Fenberg and Roy, 2008). Furthermore, if fishers already understand this resource partitioning and can target particular Rajidae species, predictably those commanding a higher price, overfishing can potentially go unnoticed. Due to the grouping of Rajidae species in fishing quotas, this has already been realised by the local extinction of the Dipturus batis complex (Brander, 1981).

152

Chapter 7. Are critically endangered fish back on the menu? Analysis of U.K. fisheries data suggest post-ban landings of prohibited skates in European waters

This chapter was published as: Simpson SJ, Sims DW. Are critically endangered fish back on the menu? Analysis of UK fisheries data suggest post-ban landings of prohibited skates in European waters. Marine Policy. 2016;69:42-51.

Abstract

Skates (Rajidae) have been commercially exploited in Europe for hundreds of years with some species’ abundances declining dramatically during the twentieth century. In 2009 it became “prohibited for EU vessels to target, retain, tranship or land” certain species in some ICES areas, including the critically endangered common skate and the endangered white skate. To examine compliance with skate bans the official UK landings data for 2011-2014 were analysed. Surprisingly, it was found that after the ban prohibited species were still reported landed in UK ports, including 9.6 tonnes of common skate during 2011-2014. The majority of reported landings of common and white skate were from northern UK waters and landed into northern UK ports. Although past landings could not be validated as being actual prohibited species, the landings’ patterns found reflect known abundance distributions that suggest actual landings were made, rather than sporadic occurrence across ports that would be evident if landings were solely due to systematic misidentification or data entry errors. Nevertheless, misreporting and data entry errors could not be discounted as factors contributing to the recorded landings of prohibited species. These findings raise questions about the efficacy of current systems to police skate landings to ensure prohibited species remain protected. By identifying UK ports with the highest apparent landings of prohibited species and those still landing species grouped as ‘skates and rays’, these results may aid authorities in allocating limited resources more effectively to reduce landings, misreporting and data errors of prohibited species, and increase species-specific landing compliance.

153

7.1 Introduction Humans have exploited fish for thousands of years (Lotze and Worm, 2009) and have had a major impact on key species as well as their ecosystems (Jackson, 2008, Worm et al., 2009, Worm and Branch, 2012). Since the industrialisation of fishing in the late 19th and early 20th centuries, fishing has caused depletions of many species that have, in numerous cases, been masked by increasing catch efficiencies enabled by advances in technology, geographic expansion of fishing ranges and the exploitation of previously rejected species (Pauly et al., 2002). Prior to industrialised fisheries there appeared little or no need to collect catch data and to manage a longstanding traditional human food source, which at that time was thought to be inexhaustible (Daniel and Minot, 1954, Sims and Southward, 2006).

Despite anecdotal evidence suggesting a rapid increase in marine fishing ca.1000 A.D. in Europe, fisheries statistics were first collected about 110 years ago by the newly formed International Council for the Exploration of the Sea (ICES) (Barrett et al., 2004). With these data, investigations assessed the impact of fishing and were used to inform advice on sustainable levels of fishing for specific species. From these long-term records it has been documented for example that in England and Wales annual demersal fish landings from bottom trawl catches have significantly declined since the industrialisation of fishing in the 19th century (Thurstan et al., 2010).

On a global scale stock collapses due to overfishing have been well documented for some commercially important fish species, such as Atlantic cod in Canada (Cook, 1998) and Pacific anchovies (Lluch-Belda et al., 1989), but many other marked declines in abundances of large fish species have gone largely unnoticed (Brander, 1981, Jackson et al., 2001, Myers and Worm, 2005, Casey and Myers, 1998b). There are several examples of longstanding, unregulated exploitation of large fish leading to dramatic declines, particularly so among the elasmobranchs (sharks, skates and rays). Elasmobranchs have life-history characteristics that make them vulnerable to overfishing, including slow growth, late age at maturity and low fecundity, making them less resilient than bony fishes to overexploitation (Dulvy et al., 2014, Field et al., 2009). According to the International Union for the Conservation of Nature (IUCN) Red List of threatened species, a quarter of all assessed sharks, skates and rays are thought to be ‘threatened’ due to overfishing. Of the seven most threatened families, five are skates and rays, with an increasing global catch of elasmobranchs now being made up of more skates and rays than sharks (Field et al., 2009, Dulvy et al., 2014).

In the north-east Atlantic Ocean and in the UK in particular, the main commercial interest for elasmobranchs is the family Rajidae (skates), of which there are 16 principal species. Prior to the

154 expansion of marine fisheries in the 20th century, skates were of low value in the UK and were often rejected from fish markets (Steven, 1932a). However, by the beginning of the 1900s they became an increasingly important fishery, notably around the southern coast of England where they made up the highest quantity and value of any species group within the fishery (Dulvy et al., 2000, Steven, 1932a). In the 1930s, during investigations of the catches of skates in fish markets in south-west England, it was noted that it was difficult to assess which species were of importance to the fishery because individuals were not landed as species but instead under the broad group ‘skate and ray’ (Steven, 1936). Despite this early observation foreseeing the difficulties of accurate assessment without species data, it was not until 2009 that it became mandatory in European waters to land skates as species-specific groups rather than as ‘skates and ray’ (Silva et al., 2012). During this period of increasing fishing pressure and unmonitored species catches (ca. 1900-2009), several species of skates declined in abundance. For example, in the late 19th century, common skate (Dipturus batis) were abundant in the waters around the UK and were caught throughout the year (Heape, 1887, Herdman and Dawson, 1902). By the 1920s there were reports that former areas of abundance in shallower coastal zones were now devoid of common skate (Clark, 1922), but during the 1930s fishermen were still landing significant quantities of D. batis from deeper waters (Steven, 1932a). However, by 1981 it was reported that D. batis had been extirpated from its former range due to overfishing. Indeed, records from > 800 trawls in the Irish Sea by the Ministry of Agriculture, Fisheries and Food (MAFF) in the 1970s showed no common skate were caught (Brander, 1981).

In addition to mandatory landing of ‘skates and rays’ by species after 2009, it became “prohibited for EU vessels to fish for, to retain on board, to tranship or to land” certain species in specific ICES areas. This protection includes common skate (D. batis) and white skate (Rostroraja alba) (Silva et al., 2012, CEC, 2011) principally due to D. batis being IUCN Red List assessed as ‘critically endangered’, and the white skate Rostroraja alba as ‘endangered’. Importantly, recent studies used morphometric and molecular genetic markers to demonstrate that there were cryptic species of common skate (D. batis), with two species in the north-east Atlantic having distinct but overlapping distributions (Griffiths et al., 2010, Iglésias et al., 2010). However, UK landings data groups these two species (D. batis species-complex) into one ‘common skate’ group that will, in this study, be referred to as such. The undulate ray (Raja undulata) was also a prohibited species from 2009, however in 2015 the IUCN Red List assessment for European species downgraded its classification to ‘near threatened’ (Nieto et al., 2015), essentially opening up the fishery for this species once more.

155

The Marine Management Organisation (MMO) and Marine Scotland are the authorities responsible for the enforcement of marine regulations including landing of restricted species in England and Wales, and in Scotland respectively. The MMO record data on the fish landings made at the ports, including both weight and value, which are collected from fishermen’s log books and market sales notes. These agencies can also have representatives based at fish markets around the UK that inspect catches landed at market and those held in market cold stores. Data are then checked and verified by port staff as well as database managers and statisticians at the data input and archiving stages (MMO, 2014). For data to support fisheries management measures reliably it is essential that landings and discard data are recorded accurately. This is especially important because landings data are widely used to inform and support the development and delivery of government decision-making at the UK and European level to enact components of the European Union Common Fisheries Policy. This includes contribution to stock assessment for estimation of total allowable catches (TACs), quota management, effort control and fleet management (MMO, 2014). These data are also crucial to ongoing assessment of whether particular management policies are effective for sustainable exploitation of European fish stocks.

In a previous study the species composition of skates in UK commercial landings and discards was examined between 2007 and 2010, a period spanning the implementation of the bans (Silva et al., 2012). The latter study concluded that reported landings of prohibited species had decreased after 2009, in line with conservation measures (Silva et al., 2012). In the current study it was investigated whether the landings of prohibited skates have further declined toward zero, as would be expected if bans are being adhered to and are being policed effectively. Therefore, to investigate the effectiveness of the 2009 changes for skate landings in the UK with respect to prohibited species and the need for landings of species-specific groups, data from 2011-2014 were obtained from the UK MMO for analysis. The expectation was that if the restrictions in place are effective, monitored and enforced, with sufficient resources available for error checking, data should be categorised as individual species and none of the prohibited species should appear in the data (Sims and Simpson, 2015).

7.2 Method

Species-specific skate and ray data were obtained by written request from the UK MMO. The data were provided on 26th January 2015 and comprised data for UK flagged vessels landing into the UK and abroad, and foreign flagged vessels landing into the UK over the period from 2011 to 2014 inclusive. The data provided included landings of species in addition to the grouping ‘skates and

156 rays’. The dataset also included ICES area of capture, Food and Agriculture Organisation of the United Nations (FAO) area of capture, port where a landing was made and the live weight (metric tonnes) and value (£) of the landed catch. Live weight data were mapped in ArcGIS (10.2.2) according to ICES area and port. Relative quantities of common skate were also calculated to investigate whether higher landings of this species in northern ICES areas were a function of the higher overall landings from these areas. For each ICES area total common skate landings in 2014 were divided by the total skate and ray landings in 2014 for that respective ICES area. ‘Skate and ray’ landings by port were only mapped for ports when total landings were greater than 5 tonnes. The 2014 data were considered ‘provisional’ by the MMO at the time the analysis was undertaken.

7.3 Results

7.3.1 Prohibited species Between 2011 and 2014, 9.6 tonnes of common skate (D. batis-species complex) were reported as landed all around the UK (Figure 61). There were higher landings in the north western and eastern ICES areas VIa (2.43 t; north-west Scotland) and IVb (1.89 t; central North Sea), respectively. Ports with particularly high total reports of landings of common skate were Scrabster (1.4 t), Mallaig (2.3 t), Peterhead (1.6 t), Oban (0.7 t), and Portavogie (0.7 t), all in Scotland, and Exmouth (0.6 t) in south-west England (Figure 62). Landings of common skate did not necessarily occur at all ports in every year. For example, Scrabster reported no landings in 2014 whereas Oban and Portavogie reported their highest landings of common skate in 2014 (0.4 t and 0.6 t respectively).

The general pattern of higher recorded landings of common skate from northern ICES areas, e.g. VIa north-west Scotland (Figure 61), were not dependent on the higher overall landings of skates and rays made into northern UK ports. Rather, the landings of common skate from northern areas remained relatively higher than those from more southerly ICES areas even after accounting for the total skate and ray landings made from each area (Figure 63). This pattern indicates landings of common skate were not distributed randomly around UK ports, but appeared to reflect latitudinal abundance differences.

The reported landings of white skate were 17.89 t in ICES area VIa (north-west Scotland), whereas in IVa (northern North Sea) the reported landings were higher at 29.49 t. In the latter area, however, it was not prohibited to fish for, retain or land white skate (

157

Figure 64), indicating that the reported landings of white skate were of the same magnitude in weight irrespective of whether a ban for that species in that area was in place. In terms of ports, Mallaig (12.3 t) had relatively high quantities of prohibited landings of white skate, with lower numbers in other ports around the UK (Figure 65). Mallaig had its first year of zero reported landings of white skate in 2014, in contrast to Kinlochbervie in the same ICES area which had its highest reported landings in 2014 (0.5 t).

Figure 61 Reported landings (tonnes) of common skate by ICES area, 2011-2014, in UK MMO data.

158

Figure 62: Quantity (tonnes) of common skate reported landed by port, 2011-2014. Only the highest landing ports (>0.085t) are named.

Figure 63: The relative quantity (tonnes) of common skate landed by ICES area 2014.

159

Figure 64: Reported landings (tonnes) of white skate by ICES area 2011-2014. Hatched areas are not prohibited from landing white skate.

Figure 65: Reported landings (tonnes) of white skate landed by port, 2011-2014. Hatched areas are not prohibited from landing white skate. Only the highest landing ports are named (>0.233).

160

7.3.2 Other species Undulate ray (R. undulata) was also a prohibited species over the time period covered by the data. Indeed, only minor landings of this species were reported in Newlyn in 2012 (1.6 kg), which likely represented a single individual. The data also report some species landings which are not prohibited but seem less reliable based on their species range. Data show that Arctic skate (Amblyraja hyperborea) were landed off the southern coast of England between 2011 and 2014. The areas of note are IVc (southern North Sea; 1.59 t), VIId (eastern English Channel; 0.13 t) and VIIe (western English Channel; 0.13 t). Norwegian skate (Dipturus nidarosiensis) (1.1 kg), likely a single individual, was reported landed in area VIId (eastern English Channel).

7.3.3 Skates and rays Overall, the MMO landings data records that 769.6 tonnes of ‘skates and rays’ were landed as one group between 2011 and 2014 in all areas of the UK (Figure 66). The amount of the former ‘skates and rays’ group landed as species was 96% in 2014, with that remaining as ‘skates and rays’ amounting to 133 tonnes. The areas with the highest total landings during 2011-2014 were VIa (north-west Scotland; 154.03 tonnes) and IVa (north-east Scotland; 287.60 t). For the ‘skates and rays’ landing group, the highest landing ports were Peterhead (148.4 t) and Scrabster (163.4 t) followed by Lervick (41.5 t), Lochinver (40.7 t) (all in Scotland), and Padstow (34.7 t) in south-west England (Figure 67). For Padstow, the majority of this total (34.6 t) was landed in 2011, however since that time landings ascribed to the ‘skates and rays’ group have been very low. For Scrabster, the highest landing port for ‘skates and rays’ landings have increased during the period, indeed almost doubling from 2011 (37.0 t) to 2014 (64.4 t). For the other ports mentioned here, all showed some decrease in this landing group, although landings remained substantial. The lowest was Lochinver recording 1.6 tonnes in 2014.

161

Figure 66: Skates and rays (tonnes) landed in the UK ICES areas 2011-2014. Landings are from UK and foreign vessels landing into the UK and UK vessels landing abroad.

Figure 67: Quantity (tonnes) of 'skate and ray' landed by port 2011-2014. Only ports with landings >5t are displayed and only the highest landing ports (>10t) are named.

162

7.4 Discussion

This study reveals that prohibited skate species were recorded as landed all around the UK between 2011-2014 following the bans in 2009. The recorded landings were not distributed evenly, but instead, some areas and ports reported notably higher landings than others. Overall, the areas in the north of the UK reported higher landings of both common skate and white skate. Exmouth in the south also reported a relatively high number of common skate in the landings data (0.56 t between 2012 and 2014). These data indicate annual landings of prohibited species were still being made, or being recorded as made, across the UK at a time when bans for these species were in place within European waters. As well as the possibility that these may be actual landings of the prohibited skate species, there may also be factors that contribute to errors in reporting that mean these data may not represent actual landings. Therefore, three possible explanations are proposed for the occurrence of prohibited species in the UK landings statistics: (i) prohibited skate species were being caught, retained, landed and sold as the correctly named species; (ii) misidentification of skate species means no actual prohibited species were being caught, landed and sold; and, (iii) data entry errors at ports or elsewhere were occurring that mean no actual prohibited species were being caught, landed and sold. These principal possibilities are discussed in turn prior to making some conclusions based upon the available information presented here and that found elsewhere in the literature.

7.4.1 Are prohibited, endangered skates being landed into UK ports from fishing areas where bans are in place? A recent study reported that the prohibited D. batis-species complex was recorded in both commercial and observer data as having been landed in the UK following capture in the central and northern North sea areas (Silva et al., 2012). For example, the observer programme in the central and northern North Sea recorded 2.1 t of D. batis-complex being retained from the catch during 2010, after the 2009 ban was in place, a quantity that was higher than the 0.3 t reported as landed by the commercial otter trawl fleet from those areas in 2010. Our results confirm these prior landings by showing for the period 2011-2014 that common skate were still being recorded as landed by fishers after 2010 and that this quantity appears to be not insignificant. This study found that 9.6 t of common skate were recorded landed in UK ports between 2011 and 2014. If an average-sized individual common skate is considered to be between 3.5 kg (Iglésias et al., 2010) and 33.17 kg (Wearmouth and Sims, 2009b) in total body mass, this estimates that between 289 and 2,743 individuals were landed (mean, 72-686 per year) in 2011-2014. Furthermore, it was evident that the apparent landing pattern was not random, with most recorded landings occurring in northern UK ports and caught within northern UK sea areas (see Figs 1 and 2). This

163 pattern of catches and landings of common skate appears to be consistent with their currently known centres of abundance within their distributional range, which are thought to be greater in northern UK waters (Griffiths et al., 2010). Clearly, this northern bias in common skate landings in 2011-2014 in the raw data may be a consequence of a higher number of skates generally being landed into northern ports as opposed to southern ones. However, even after accounting for the higher landings of skate species generally made from northern UK ICES areas, the current study still found the recorded landings of common skate to be relatively higher in the north of the UK (areas VIa, IVa and IVb), a pattern that would be expected if they were in higher abundance there (Figure 3). This implies that the recorded landings of common skate in 2011-2014 reflect the expected patterns of landings based on abundance and distribution. In support of this, common skate are occasionally reported from VIIa (Irish Sea), VIIf (Bristol Channel) and IVb (central North Sea), though it is suggested that its range is now limited to VIa (north-west Scotland) and the VIIh (Celtic Sea) (Dulvy and Ungaro, 2006). That the relative landings of common skate reflect their reported latitudinal abundance trends argues against the pattern being largely due to misidentification of skates by fishers or officials, or due to erroneous data inputs occurring more often in northern areas, given that these types of errors should theoretically be equally likely in all areas and ports. Therefore, our results cannot entirely discount the possibility that common skate have actually been retained and landed into the UK in at least four of the years after 2009 when the ban came in force.

Despite these recorded landings in official data, there appears at present to be no evidence of common skate products entering the UK retail chain. Griffiths et al (Griffiths et al., 2013) analysed DNA sequences in tissue from 98 skate wings purchased in retail outlets, such as supermarkets and restaurants, but found no evidence for the presence of prohibited or vulnerable skates for sale. This result may be a consequence of sample size however, since it may be expected that very few individuals of critically endangered species are likely to be sampled in markets or food outlets because they are naturally at low abundance and hence few are landed compared to other species. This assertion is supported by considering common skate numbers compared to total skate landed in 2014 for example, which estimates there would be a 0.054% chance of sampling a common skate, equivalent to finding one common skate for every 1,852 skates examined. Therefore, nearly two thousand individual skate would need to be tissue sampled for DNA before a single positive identification is likely statistically, even if they are entering the retail chain. Therefore, the possibility that common skate are entering the retail chain cannot be discounted on the basis of forensic studies undertaken to date.

164

7.4.2 Are fishers misidentifying or misreporting prohibited skate species? Species misidentification is a major potential problem in skate fisheries that can contribute in important ways to confusion with interpreting prohibited species landings data. For instance, a recent study using molecular genetic markers found that in supermarkets where skate pectoral fins (marketed as ‘skate wings’) were labelled with a species name, 33% of the labels were incorrect (Griffiths et al., 2013). Therefore, it seems misidentification of skate species occurs frequently and is being introduced somewhere along the retail chain from the point of skate capture to the location of sale to consumers. Of course, once the skate wings have been processed (skinned), it becomes much more difficult for retailers to identify the species correctly without molecular genetic analysis. Moreover, UK skate species are also difficult to identify even when alive. The spot pattern and/or colouration that is often used by fishers to distinguish between skate species can be highly variable within a species which enhances the problems of easy identification soon after capture (Iglésias et al., 2010). The problem of persistent misidentification is well illustrated by two examples. One recent study (Silva et al., 2012) found significant discrepancies in the quantity of skate reported in commercial landings and that recorded by observers. In the North Sea in 2010 it was found that commercial otter trawlers reported 3.4 % of their skate catch as the spotted ray Raja montagui, whereas observers on otter trawlers in the same year reported that 50.9% was R. montagui (Silva et al., 2012). In addition to that investigation, in this study the presence of Arctic and Norwegian skate on the south coast of the UK in the landings data would also suggest ongoing issues with misidentification. Our investigation found recorded landings data supporting one or two individual Norwegian skate and a significant quantity of Arctic skate being ‘landed’ on the southern coast of England. However, the southern coast of England does not fall within the distributional range of these species, so it would seem highly unlikely that these species were in fact caught there or subsequently landed nearby. It is more likely that the individuals landed were misidentified or were incorrectly entered into the landings data. It is evident that correct identification is a significant and continuing problem in reporting landings.

Misidentification can take two forms. It could simply be unintentional error on the part of the fishers when faced with individuals from different species that look similar. This may happen frequently for more common species and remain undetected because the misidentifications are effectively lost among the large quantities of correctly identified individuals. However, equally, there may be intentional misidentification that manifests as misreporting. For example, at the fish market each of the main species of skate landed and traded (thornback ray Raja clavata; blonde ray, R. brachyura; spotted ray, R. montagui; cuckoo ray, Leucoraja naevis; and smalleyed ray, R.

165 microocellata) has a separate price for species and for size class. Although prices fluctuate, blonde ray generally obtains the highest value, with thornback ray obtaining the lowest. Therefore, fishers may have an incentive to misreport a species for one most likely to obtain a higher price. Indeed, other studies have shown that misidentification has occurred purposefully in order to obtain a higher price or to hide the capture of a restricted species (Faunce, 2011, Marko et al., 2004). The discrepancy in common skate reported as landed by fishers (0.3 t in 2010) and that recorded by observers as retained species (2.1 t in 2010) in the central and northern North Sea (Silva et al., 2012) could be explained by intentional misreporting of common skate as blonde ray, for example. However, this does not explain in the context of the current study why a fisher would record a common skate in the logbook of catches. It seems unlikely that a fisher would identify a common skate (rightly or wrongly), and regardless, attempt to land it for sale as common skate when a ban is known to be in operation for the species. Furthermore, the common skate is prohibited so there should in effect be no price for it, thus it would seem more likely that a fisher would log it as a different species if their intention was to command a higher price. However, the recorded landings of common skate in 2011-2014 in this study were officially reported as having a monetary value of £10,456, implying that common skate were openly landed and sold as common skate unless of course these data were entirely incorrect (see section 3).

There is good reason to assume that misidentifications involving prohibited species should have a greater chance of being detected. It is not only the fishers that are involved in the process of catching fish right through to data collection and input by the management authorities, but there are other stages at which identification errors could be corrected. The fish merchants employ staff that sort and grade fish so that it can be priced according to species. The UK authoritative agencies also visit fish markets and cold stores to verify the catches. The skate are then sold to buyers that often prefer one species over another, because some species are easier to process than others, hence the higher price for blonde ray for example. There are then multiple steps with data cross checks and data validation that occur at the port with the fisheries’ authorities and also their central database statisticians (MMO, 2014). Given the number of steps involved from fish capture to identification and data entry, it is possible that apparent landings of prohibited species would most likely be checked and contraventions identified at the ports. Therefore, it seems unlikely that systematic misidentifications of a prohibited species, mainly in northern UK ports, can account for the relatively large quantities of these rare species appearing in official UK landings statistics.

166

7.4.3 Are systematic errors in data input being made? Fishers input data into their log books, and in port the market agents then sort and grade species, as well as providing sales notes for the fish sold. The observers appointed by authoritative agencies inspect catches and enter the species they find onto data collection forms. There are many stages in this process where human error in data input could be introduced. Simply an incorrect box ticked could cause errors in allocations of landings to individual species. The entire chain from fishers to regulatory agencies’ data input uses codes to identify each species. These codes are based on species scientific names, although the landings data uses common names. For example, RJB is the market code for common skate Dipturus batis, however it is easy to appreciate that this may be confused with blonde ray Raja brachyura, whose code is actually RJH.

The scale of erroneous data entry appears to be significant. For example the MMO have reported that 80% of all electronic logs for fishing vessels in 2013 had to be amended due to incorrect information (MMO, 2014). This only serves to indicate the large potential for error in data input that can be introduced from the very start of the data chain and right up to data transfer from fishers to fisheries managers. As part of data validation and review, statisticians also check for unlikely combinations of ICES area and species landing. However, the records found in this study after the data was provided to us on 26/01/2015 showing the occurrence of Arctic skate in ICES VIIe (western English Channel) in 2011 and 2012, suggests that likely errors are not being identified and corrected/deleted in a time frame that is relevant to management needs. For example, the 2011 data that held the likely errors noted above were still being sent to researchers in 2015. A previous study using recent UK landings data also questioned the accuracy of some of the commercial landings data, where skate species were apparently caught outside of their natural range (Silva et al., 2012).

7.4.4 Concluding Remarks In the above discussion some of the data and arguments relevant to each possible cause of the reported landings of prohibited skate present in the UK official landings data has been set out. It was not possible to determine precisely which of the factors was largely responsible for the apparent illegal landings of protected species because there was no way to identify post hoc what actual species made up the landings reported. It is very likely that these three possible explanations are not mutually exclusive, making identifying the cause even more difficult. Nevertheless, there does appear to be support from this study and from a previous investigation (Silva et al., 2012) for the conclusion that common skate were landed and sold in UK ports after the 2009 ban. It is also possible that captured common skate were misidentified or misreported

167 as being other skate species, while errors in allocating and recording market codes, and other data entry errors prior to finalising the official landings data may also play a significant role in misrepresenting landings of prohibited skate species.

Regardless of which of the three principal explanations was the most likely to account for prohibited skate landings, there appears to be a lack of official investigation to determine the origin of the apparent prohibited landings and to correct them where necessary. For example, this study was provided with possible error-laden data by the regulatory authority some 3 years after it first appeared in the UK official fish landings statistics. The persistence of erroneous data in official records may reflect unequal resources available at ports across the UK for early detection of prohibited skate landings, misidentifications, misreporting and data entry errors. Misreporting is more difficult to identify generally in mixed catches of skates but should be possible with sufficient surveillance. The UK spends significant public funds on the monitoring and enforcement of fisheries regulations, yet it seems that some potential errors that can be straightforward to check are not only still occurring but are remaining within the official statistics for at least three years.

There has been significant progress however in greater reporting of skate catches according to species rather than the generic ‘skates and rays’ grouping. The current study supports the findings of Silva et al (Silva et al., 2012), in that since the 2009 regulations were implemented improvements have been made in terms of landing species-specific skates. Silva et al (Silva et al., 2012) report that in 2010, 92% were landed as species specific. By the end of 2014 our analysis shows that this figure had risen to 96%. However, the 4% that were not landed as species represent 133.3 tonnes of unknown species. Therefore it remains possible that prohibited, endangered skates may make up some of this grouping.

Furthermore, there have been some improvements to reduce landings within the ‘skates and rays’ group between 2011 and 2014 at certain ports, but such reductions are not consistent across the UK. ICES areas and certain ports in northern Scotland appear to be relatively high in landing ‘skates and rays’ grouped. Although it is possible that this could occur because of a lack of fisheries enforcement offices in these more remote areas, it was evident that there are Marine Scotland compliance offices based at all of the main ports including those with relatively high landings of prohibited species, such as Mallaig, Scrabster and Fraserburgh. This suggests more needs to be done to enforce the landing of skates in species-specific groups, not only to reduce the potential for prohibited species to be included in landings of ‘skates and rays’, but to improve the accuracy of fisheries management advice for skate species.

168

In summary, this study draws attention to the recorded landings of prohibited skate species in each year from 2011-2014 since the European Union ban was put into effect in UK waters in 2009. That common skate have actually been openly landed and sold in UK ports since 2009 could not be entirely discounted. This possibility emphasises the need for greater efforts to enforce the ban across major UK fishing ports if these endangered species of fish are to be adequately protected according to the management measures put in place to safeguard their populations.

169

170

Chapter 8. General discussion

The work presented in this thesis investigates the spatial occurrence, habitat associations, habitat selection and resource partitioning (spatial and food partitioning) in sympatric Rajidae species. Furthermore, intra-species partitioning was demonstrated between the sexes and age groups and interactions with the commercial fishery and existing management measures were investigated. This general discussion explores concepts in spatial ecology using R. brachyura as a focus species and then discusses the wider impacts, limitations and future work stemming from this thesis.

8.1 The spatial ecology of Rajidae

8.1.1 Spatial occurrence It was clear from multiple data collection approaches that all species exhibited finer scale spatial occurrence in the study area than demonstrated in previous studies (Martin et al., 2010, Ellis et al., 2005a, Walker and Heessen, 1996, Walker and Hislop, 1998). Figure 68 contains the data from Chapters 2, 3 and 4, as well as Humphries et al. (2016c), for R. brachyura and clearly indicates that the Bigbury Bay site is an area of aggregation for this species. Depth ranges established by Humphries et al. (2016c), together with the data from research surveys, recaptures by the commercial fishery and detections and movements within the acoustic array, all overlap within this Bigbury bay area. There is a conspicuous absence of R. brachyura in the estuary and in the L4 area (as shown in Chapter 2, 3 and 4). Indeed, both R. montagui and R. microocellata were also largely absent from the L4 area and R. clavata were only present in low numbers. In previous studies other skate species are also found in discrete areas of high aggregation (Bizzarro et al., 2014, Neat et al., 2015, Hunter et al., 2006). This patchy, fine-scale spatial occurrence may therefore be a general characteristic of the Rajidae. Consequently, this suggests that in other regions of the UK R. brachyura may also show this patchy, higher density spatial occurrence. Patchy populations have been demonstrated to result in sampling bias (Miller and Ambrose, 2000). In Chapter 2 for example, adjacent sampling sites were entirely different in terms of their composition and frequency of capture for each species. It is therefore important to consider the way in which these species are sampled to account for such spatial complexities in future studies aimed at identifying population trends from potentially limited surveys.

8.1.2 Habitat associations The strong site fidelity for the Bigbury Bay area and occupancy in the 30-40 m depth range suggest that the characteristics of the habitat in this site must be important for R. brachyura.

171

Habitat associations were more challenging to define for Rajidae species, mainly due to the lack of fine-scale habitat data in the area. R. brachyura would have been associated with a rocky substrate, if the EUNIS habitat data (European Environment Agency, 2018) was used. However, the MBA research survey uses a trawl in this area and through personal observation it was evident that this is a soft bottomed, potentially coarse sand habitat (Figure 69). This is supported by the boundaries of the designated MPA which were delineated to protect reef habitat, leaving a strip between the MPAs that is currently unprotected legally. It is therefore likely that this area of spatial occurrence is associated with soft substrate rather than reef. Furthermore, Ellis et al. (1996b) demonstrated from stomach contents that R. brachyura showed a preference for ammodytids (sand eels), which are also associated with soft sediments (Greenstreet et al., 2010). It was also clear that R. brachyura are not associated with the brackish, estuarine habitats, that R. clavata occupy and also that temperature was not a factor in habitat association at the spatiotemporal scales investigated. Moreover, the stable isotope analyses in Chapter 7 revealed that R. brachyura occupy a distinct isotopic niche, which suggests that they occupy different habitat to the other species and are therefore associated with particular habitats.

172

Figure 68: Multiple R. brachyura spatial datasets. The blue area represents the average depth range (plus standard error) (Humphries et al., 2016c), red circles represent locations of recapture (dispersed in ARC GIS to visualise multiple point in the same location) (Chapter 3), triangles represent research survey data (Chapter 2), pentagons and lines represent acoustic networks (Chapter 4). Note the high overlap of all data in the Bigbury Bay area. 173

Figure 69: The Bigbury Bay area. Red lines are boundaries of MPAs, triangles represent R. brachyura in Chapter 2. Black lines represent 2 meter bathymetry contour lines. DEFRA bathymetry highlights the complex bathymetry inside the MPAs (designated to protect reef features) and the relatively simple bathymetry between the two MPA boundaries which may indicate a sandy habitat.

8.1.3 Habitat selection and preference Chapter 4 demonstrates that all species including R. brachyura exhibit habitat selection and preference, as I was able to track individuals of each species through space and time. The acoustic telemetry networks analysed in represent the voluntary movements of four species and in every species, observed network metrics were significantly different from random. R. brachyura spent a greater proportion of time within the Bigbury Bay area and made active movements within or towards the area. It is possible that a combination of suitable soft substrates for burying, the availability prey such as sand eels or crabs, a need to find mates and deposit eggs drive the habitat selection suggested by the current study. By spending a greater proportion of time in a different habitat to another predator, a species can reduce the competition between them (Ross, 1986) and therefore it is also possible that different habitat preferences facilitates the coexistence of these predators living in this coastal zone.

174

8.1.4 Resource partitioning Resource partitioning was most clear between R. brachyura and R. clavata and there is some evidence for different behavioural strategies between these species. R. brachyura occupy deeper depths, with habitat preference for a particular area. Meanwhile R. clavata occupy shallower depths, but have a comparatively broad distribution, being found in the estuary habitats as well as marine and were the main species present in the L4 area (Figure 70). In addition, R. clavata had a broader isotopic niche than R. brachyura (Chapter 5), indicative of greater individual variation in R. clavata in both diet and habitat use. These findings indicate that R. clavata may be considered a habitat generalist while R. brachyura is a specialist relative to R. clavata. These differing strategies are reflected in dietary analysis studies. R. brachyura consumes a greater proportion of fish, while R. clavata is more of a generalist, but with a greater proportion of crustaceans (Ellis et al., 1996a, Holden and Tucker, 1974, Ajayi, 1982a, Ebert et al., 1991, Smale and Cowley, 1992). This dietary difference has in turn been linked to more fundamental differences in behavioural ecology such as how each species structures searches for prey (Wearmouth et al., 2014b). The generalist nature of R. clavata may also explain why in Chapter 3 this species exhibited greater site fidelity than the other species. R. clavata may be able to exploit a range of habitats in close proximity and a wider range of food resources, potentially enabling them to remain for longer within less extensive areas. Equally, a specialist such as R. brachyura or R. microocellata may spend extended periods in preferred habitat but as a specialist may be forced to move greater distances to find preferred habitats when conditions change, hence the larger distances moved (Chapter 3).

8.1.5 Temporal resource partitioning Temporal space use on different scales was important in the spatial ecology of all four species including R. brachyura. There appears to be a strong association to the Bigbury area in the summer months, as evidenced by peak captures in the research survey and by recaptures in the commercial fishery. In addition, there appears to be temporal resource partitioning by R. brachyura and R. montagui within this area. R. montagui appear to use the Bigbury Bay area in spring and summer while R. brachyura use the area in summer and autumn. This temporal resource partitioning may allow R. montagui and R. brachyura to use the same habitat by overlapping less in the main times each species occupy the area, thereby potentially reducing competition between these predators.

175

Figure 70: Multiple R. clavata spatial datasets. The blue area represents the average depth range (plus standard error) (Humphries et al., 2016c), red circles represent locations of recapture (dispersed in ARC GIS to visualise multiple point in the same location) (Chapter 3), triangles represent research survey data (Chapter 2), pentagons and lines represent acoustic networks (Chapter 4). Note the wider distribution of data from the different data sets relative to R. brachyura.

176

8.1.6 Interaction with the commercial fishery Understanding this habitat partitioning among Rajidae species is also important to consider in a commercial fisheries context, because habitat partitioning has an impact on the catchability and therefore the vulnerability of each species (Mucientes et al., 2009). The partitioning observed between rajid species also appears to influence species interactions with the commercial fishery. Chapter 3 highlights a number of differences between the species, including location of capture and the gear used. R. brachyura was caught more frequently by trawls than the other species, potentially because of its deeper habitat preference. However, this does not explain why R. montagui, that have the same deeper habitat preference, were not also caught more frequently by trawls. This difference between these two species may be explained by the temporal resource partitioning found between R. brachyura and R. montagui in Chapter 2 and 3. R. montagui had a tendency to occupy the Bigbury Outer site in the spring and summer, while R. brachyura occupy the area summer and autumn.

The preference for a narrower range of habitat by R. brachyura compared to other rajids potentially leaves this species vulnerable to habitat damage and over-exploitation. R. brachyura have the highest value of the four rajids (Personal communication, Plymouth Trawler Agents 10/06/2015), potentially resulting in a targeted fishery. Conversely, understanding these specific habitat preferences and site fidelity could result in a spatially targeted approach to their management, as has been suggested for the flapper skate and thornback rays (Neat et al., 2014, Hunter et al., 2006). Chapter 3 for example suggests that the Bigbury site between the two MPAs could be closed to protect the aggregation hotspot that R. brachyura actively select. In addition, hypothetically, R. clavata preference for estuarine habitat may provide protection from larger trawlers that work offshore.

Differential habitat association by age class and sex also has important implications for management, for example, if females are more easily captured due to sexual segregation, population declines could be exacerbated if fishing is focused on those habitats. In Chapter 2 there were sex biases in trawls, particularly in R. clavata and in Chapter 4 there were biases towards females in the number of detections within the array. Rajidae may therefore be vulnerable in the same way to commercial fishing. This is illustrated by the collapse of the spurdog Squalus acanthias fishery in south west UK, which has been attributed in part to the bias in capture of females in the late 1920s (Wearmouth and Sims, 2008).

While progress has been made with regard to the management of these vulnerable species, Chapter 7 suggested that the 2009 management measures of ensuring species specific landings of

177

Rajidae species may not currently be enforced efficiently. Nevertheless, this thesis suggests possible avenues for future management and potential improvements to existing measures. For example, R. brachyura could be more efficiently protected by the closure of the area between the two existing protected areas (See Chapter 3, or Figure 69) or by restricting static gear in established MPAs. Alternatively – or additionally – temporal closure of the fishery within the MPA during the summer and autumn months could also reduce their capture in this local area. This could be further replicated around the coast of the UK where MPAs already exist. For example, R. brachyura are known to occur further along the south coast of the UK, near the Isle of Wight (Martin et al., 2010), where there are also multiple established MPAs (JNCC 2017). The efficacy of the MPAs could be improved by replicating the conclusions that have arisen from the current study, such as restricting static gear in MPAs or seasonal closures. Furthermore, enforcement of existing legislation could be improved by investigating ports at which there were high levels of grouped ‘skate and ray’ landings, such as Peterhead and Scrabster in Scotland.

Species that spend their lives occupying relatively fine spatial scales should also be sampled on finer scales, as not doing so will likely result in sampling biases. The idea that studies should be conducted at the appropriate or multiple spatial scales is well established in terrestrial ecology (Wiens, 1989), but is often lacking in marine ecology due to logistical difficulties that to surpass become prohibitively expensive. Nevertheless, if we are to understand and manage these species effectively, understanding the fine scale Rajidae distribution is fundamental and there is a clear need to investigate the ways in which these species may differ in their interaction with the commercial fishery.

8.2 Wider implications The fine-scale habitat preferences exhibited by the four Raja species found here may have an impact on the ecosystem where they are present. Rajidae bury in the sediment for extended periods of time (Greenway et al., 2016) so it is therefore likely that within their spatial occurrence they impact sediment processes through bioturbation. Bioturbation is the shaping and modifying of the abiotic and biotic characteristics of this habitat, usually through behaviours such as burrowing, foraging or digging (Kristensen et al., 2012). Other batoids have been shown to rework 41% of sediment over 1 km2 in one year (O'Shea et al., 2012). This allows oxygen to enter deeper into sediments, extending the area of nitrification (Gilbert et al., 1995) and can create habitat for

178 other species (O'Shea et al., 2012). The spatial occurrence of skates in preferred areas may therefore impact the biochemical processes and the presence of other species in those areas.

Skates may not only create habitat for other species, but also prey upon and compete with others. Skates are opportunistic feeders (Ellis et al., 1996a, Holden and Tucker, 1974) that co-exist with other predators, such as plaice Pleuronectes platessa. These species likely influence one another through competitive exclusion or by reducing populations. For example, the feeding behaviour of rays has been shown to negatively impact bivalve populations (Smith and Merriner, 1985) and the reduce the growth of seagrass (Orth, 1975). Skates, as top predators in the sites studied presently must therefore play an important role in shaping the life histories of other species in the habitat they occupy.

Given the importance of the influence of skate on the wider ecosystem, their removal may have a significant impact on the structure and function of these habitats. The removal of top predators has been well studied and is commonly cited to result in trophic cascades (Heithaus et al., 2008). Although strongly debated the collapse of cod Gadus morhua populations in Canadian waters for example, may have resulted in increases in herring, shrimp and crab, which could potentially also have led to a shift in the phytoplankton community (Worm and Myers, 2003). The sustained removal of skates therefore may cause trophic cascades that have far reaching and potentially unseen consequences that could affect not only the wider ecosystem but also the commercial fishery prevalent in the area. With limited baseline data however, it is difficult to assess what the impact of the removal of skates would be. Well known examples of trophic cascades involving elasmobranchs (Myers et al., 2007) have been demonstrated to be far more nuanced than previously concluded (Grubbs et al., 2016). The removal of skates could provide a competitive release for other predators such as plaice, which may limit the likelihood of trophic cascades (Parsons, 1992). Marine ecosystems are complex and highly connected, which leads to multiple possible consequences when changes in the trophic structure occur (Hughes et al., 2005) making predicting the outcome of predator removal challenging. Indeed, even with a highly relevant example of the removal of a top predator, the common skate (Brander, 1981), there is limited information about what effect this has had on the ecosystem. This is a result of a lack of baseline data, concerning diet, spatial ecology and population dynamics and indicates that the collection of baseline data such as that presented in this thesis will be fundamental to assessing ecosystem consequences of overfishing.

This thesis also adds to the growing evidence that despite skates being morphologically similar and broadly categorised as generalists, there are marked differences in their biology. This includes

179 egg laying season (Koop, 2005), fecundity, egg incubation times (Bottari et al., 2013, Holden, 1975, Koop, 2005, Ellis and Shackley, 1995), maximum size (McCully et al., 2012, Gallagher et al., 2005) and age at maturity (McCully et al., 2012, Nottage and Perkins, 1983). In addition to these reported differences among species we can add differences in fine-scale habitat preferences and temporal aggregations. These differences in biology among species likely result in different impacts on the wider ecosystem. The most obvious is perhaps that R. clavata are the only species to occupy the estuary and therefore are the only rajid species likely to make an impact on the estuarine habitats. Consequently, the removal of R. clavata has the potential to change the dynamics of the estuary ecosystem, whereas the removal of R. brachyura for example is unlikely to have a direct impact on the estuary.

This thesis has implied an important link between fine-scale movement data and the ability to inform effective species-specific management for elasmobranchs. It is becoming increasingly apparent that all fisheries data from commercial and research surveys have spatial trends (Booth, 2000). When these spatial trends are overlooked relative abundance estimates can be incorrect (Swartzman et al., 1992), which consequently can lead to inaccurate conclusions concerning a species growth, movement, reproduction and feeding patterns. There are now multiple studies that demonstrate the scales of spatial overlap between elasmobranchs and fisheries that help identify catch reduction tools (e.g. quotas, protected areas ) and where these will be most effective to aid the effective management towards a sustainable fishery (Queiroz et al., 2016, Lea et al., 2016, Neat et al., 2015, Punt et al., 2000, Barnett et al., 2012). The results presented in this research therefore contribute to our increasing knowledge of species-specific spatial ecology and how these data may inform potential management strategies.

8.3 Overall limitations Aside from those mentioned in the data chapters of this thesis, there were some overall limitations of the study. Every individual sampled in this research came from research trawls in the same set of locations across a relatively small area off Plymouth. This has several consequences that may limit the interpretation of the data. Although the sampling regime used in this research provided fine scale temporal and spatial data, the scale of sampling in the overall sample site was relatively constrained. This may result in missing other patches of rajid occupation, in for example deeper sites, which are likely to affect the interpretation of depth preferences for some species.

180

In addition, fishing gear has been demonstrated to show bias for catching particular personality traits (Andersen et al., 2017, Diaz Pauli et al., 2015). For example, guppy Poecilia reticulata individuals that exhibited more exploratory behaviours were able to escape controlled laboratory ‘trawls’ more readily than non-exploratory individuals (Diaz Pauli et al., 2015). Sampling by trawls may therefore result in a bias towards individuals with particular personalities and therefore sampling may not represent the entire population. If there are skates that exhibit more exploratory behaviour, that can escape the trawl, it is possible that they could also move further. It cannot be discounted therefore that this study focused primarily on individuals exhibiting greater philopatric behaviour.

We are also limited in defining habitat associations and selection by a lack of fine-scale environmental data. The only ‘continuous’ fine-scale data available was bathymetry. The Bigbury Bay area was particularly important due to the high numbers of Rajidae captured there, however habitat data was not accurate enough to make inferences about fine-scale habitat characteristics at the site. It would be useful to collect environmental data at a scale that applies to potential habitat preferences. Given that Rajidae occurrence was so different at sites that were only <5 km apart, 1 km resolution environmental spatial layers of potential habitat features such as substrate, vegetation and salinity would allow for a more in depth study of habitat selection. Furthermore, despite advances in tagging technology and analytical methods, we are still limited in our ability to record high spatial and temporal resolution tracks for species that do not breach the surface. This is unlike the situation for large pelagic fishes carrying GPS transmitters that surface frequently, enabling detailed fine and large-scale movements to be recorded accurately (Sims et al., 2009).

8.4 Future directions

8.4.1 Acoustic array Chapter 4 illustrates the clear potential of an acoustic array to track benthic organisms in a temperate and turbid environment. However, the array used off Plymouth is limited in receiver number and spatial extent. This array would benefit from expansion both through new receivers and through potential collaboration with other institutes. Acoustic arrays are becoming increasingly common in UK waters and therefore the area over which we could detect any tagged animal is increasing. For skates specifically, curtains of receivers across estuary mouths may reveal the movements of R. clavata into and out of estuaries, while increasing the number of receivers at known Rajidae aggregation sites, such as the Bigbury site, would aid our interpretation of habitat

181 preferences at this site. The longer dispersal of R. microocellata (Chapter 3) could be fully explored with an acoustic array that extended around the coast. Furthermore, with a large network of receivers and increasing the number of tagged individuals it would be possible to obtain higher resolution movement data and answer some key ecological questions (Hays et al., 2016). For instance, how might climate change, including increased storminess and warming seas, affect the distribution of skates which have been demonstrated to show site fidelity? Cuckoo rays Leucoraja naevus in the North Sea are thought to have moved into deeper waters as temperatures have warmed and other species such as cod Gadus morhua have been suggested to have shifted northward in distribution (Perry et al., 2005). With an expanded network of acoustic receivers these shifts could be detected on a fine-scale. In addition, established acoustic arrays can be used to detect any tagged species and therefore can be used to explore fundamental ecological questions such as interactions among species, including predator-prey interactions. Highly extensive acoustic arrays have also been used to establish core distributions for threatened animals and to establish effective spatial protection for mobile species (Lea et al., 2016).

8.4.2 Accelerometer / magnetometer Accelerometers are devices that when attached to animals can measure both orientation and changes in movement (Yoda et al., 2001). They can therefore provide information on the frequency and occurrence of a wide range of behaviours (Watanabe et al., 2008). Acceleration can be recorded in 3 axes: the dorso–ventral, anterior–posterior and lateral axes and therefore any change in movements made by an individual can be detected ((Shepard et al., 2008). This method has been used in marine species, including penguins, to investigate feeding ecology whereby prey captures were detected by the signal of head acceleration relative to body acceleration. Accelerometers indicated that penguins captured two krill per second in swarms (Watanabe and Takahashi, 2013). If successful, the application of an accelerometer type tag to a skate could provide detail at the very least about an individual’s body posture, allowing an inference for example of how much a skate swims in the water column or along the seabed.

In combination with magnetometers, accelerometers provide data which can be used to examine headings and calculate high resolution tracks. through dead reckoning (Wilson et al., 1991). Dead reckoning involves calculating a track using magnetometer headings multiplied by the speed, to calculate estimated distance travelled between course deviations, to reconstruct a track that is spatially coordinated geographically (Sims and Quayle, 1998). Dead reckoning often requires GPS location data to correct for drift, due to currents for example, which would not be possible for skates. Using the accelerometer to estimate dynamic body acceleration (summing the acceleration from all three orthogonal axes) as a proxy for energy expenditure (Qasem et al.,

182

2012) would enable further questions to be asked about foraging efficiency. The method has been demonstrated in some marine species including penguins (Shiomi et al., 2008, WILSON et al., 1991) and whales (Wensveen et al., 2015). However, the new tag combination of a magnetometer and accelerometer provides a potential means to obtain fine-scale tracks of skates. Moreover, passive acoustic telemetry with an array of receiver locations could provide the geographical positions needed to help calibrate a dead-reckoned track reconstructed by these means. High resolution tracks are vital to fill in gaps in time between acoustic detections for example and fully explore the fine-scale spatial ecology of skates. This combination of technologies (acoustic and accelerometry) will be a new frontier in benthic fish ecology that is likely to lead to rapid advances in understanding over the next decade.

8.4.3 Maternal offspring biochemical transfer and eye lens allometry The successful use of eye lens tissues to reconstruct individual trophic geography of skates across ontogeny in Chapter 6 opens up potential areas for further investigation. For example, the core of the eye lens in skate represents the egg phase of an individual’s lifetime. This is also a biochemical record of the isotopic composition of food consumed by mothers during egg provisioning (Olin et al., 2011). This transfer of nutrients has the potential to be further investigated. If egg laying females were captured, the isotopic composition of the female’s tissues, the egg yolk and the juvenile could be analysed to examine the nature of biochemical transfer between mother and offspring. An understanding of this transfer could enable the use of eye lens tissue to examine the trophic geography of two generations of skate, including the individual in question and its mother. This may for example reveal more about the nursery grounds of juveniles, whether females occupy different locations during oviposition and what food females consume during this period. It may also provide a greater understanding of the nature of the transfer of nutrients between mother and offspring, for example to quantify nutrient allocation from endogenous (internal energy stores) and exogenous (recently assimilated prey) sources e.g. (Hobson et al., 2004).

In addition, there are several gaps in our understanding that, with further investigation, could aid the interpretation of this research. First, to understand at what age the ontogenetic shifts in trophic-geography found in Chapter 6 occur, one would need to understand the relationship between eye lens diameter and age. Future studies could measure the volume of the eye, rather than the diameter, so that the hydrated outer layers could be included, as well as using fresh eye lenses rather than frozen. Controlled feeding experiments would also be beneficial to understand the biochemical transfer between prey and predator.

183

8.4.4 Investigating landings Chapter 6 presented some evidence concerning the apparent landings of prohibited species, but this topic clearly requires further investigation. It could not be fully concluded whether the landings data represented actual landings of prohibited species, were misidentifications by fishers or were simply errors in data input. To fully understand the underlying cause of the presence of prohibited species in commercial landings data, the ports with the highest landings could be targeted for further investigation. Relatively high landings of prohibited species were reported at the Scottish ports of Mallaig, Scrabster and Fraserburgh and using the DNA barcoding methods outlined in Griffiths et al. (2013), skate tissue could be tested at these ports and confirm whether prohibited skate species are actually being landed, or if these were largely errors in data input and/or misidentification.

8.4.5 Fisheries Chapter 2 suggests that the differences in spatial ecology between four species of skate results in differences in the method of capture in commercial fisheries, as well as the season they were captured in. To further investigate whether spatial ecology plays a role in the way a species interacts with the fishery, it would be useful to understand the fine-scale spatial patterns of different gear types used in commercial fisheries. Currently AIS data provides locations of vessels over 15 m, which excludes smaller vessels, which tend to operate inshore and are more likely to be catching Rajidae species. In addition our understanding of skate the fine-scale spatial ecology of skates in other locations around the UK for example would help investigate this link between fishers and their catch. Furthermore data concerning the gear used to catch a species could be collected at ports at the time of landing to provide further information regarding the potential differences in gear type used to capture each species.

184

Additional information

Appendix 1: Chapter 2 supplementary information

S Table 1: Gear types and trawl detail for each site

Survey grounds Gear type Beam Length Headline Foot Cod- (m) length rope end (m) length mesh (m) size (mm) River Beam Trawl 4 3.8 4.5 60 Whitsand Demersal otter trawl N/A 14 18 50 Standard haul Demersal otter trawl N/A 19.4 21 60 Bigbury inner & outer Demersal otter trawl N/A 14 18 50

S Table 2: Research vessel details

Vessel Vessel Engine Length Power (m) MBA 15.4 257 kW Sepia Quest 21.5 470 kW Squilla 19.74m 328kw

Appendix 2: Chapter 4 supplementary information

S Table 3: Statistical results from Wilcoxon signed ranked pairs test, comparing observed to random node metrics.

Species Observed Mean random V P n node density node density

R. brachyura 0.67 0.79 3690 <0.001 100

185

R. clavata 0.92 1 5050 <0.001 100

R. microocellata 0.75 0.995 5050 <0.001 100

R. montagui 0.91 0.97 2880 <0.001 100

S Table 4: Statistical results from Wilcoxons signed ranked pairs test, comparing observed to random edge metrics.

Species Observed Mean random V P n edge density edge density

R. brachyura 0.197 0.104 0 <0.001 100

R. clavata 0.29 0.92 5050 <0.001 100

R. 0.303 0.66 5050 <0.001 100 microocellata

R. montagui 0.17 0.15 566 <0.001 100

186

Appendix 3: Chapter 5 supplementary and publication

1000 700 Event deltas Ascent deltas 600 800

500

600 400

Count

400 Count 300

200 200

100

0 0 -50 -40 -30 -20 -10 0 10 20 30 40 50 -60 -50 -40 -30 -20 -10 0 10 20 Event delta (m) Ascent delta (m) 1200 Event duration 1000

800

600

Count

400

200

0 0 10 20 30 40 50 60 70 80 90 100 Event duration (h) Figure S1: Observed event deltas, ascent deltas and event durations for all events initially extracted. The red lines indicate the cut off for events during filtering.

187

a) R. brachyura b) R. clavata 0.3

0.2

0.1

0.0 c) R. microocellata d) R. montagui 0.3

Events per day tracked day per Events 0.2

0.1

0.0

500 600 700 800 900 500 600 700 800 900

Length (mm) Figure S2: No significant correlation between individual length and the number of events per day tracked.

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

Appendix 4: Chapter 6 supplementary information

Overall difference in δ15N and δ13C in eye lens tissue between species

Overall eye lens tissue δ15N values differed significantly between Raja montagui and all other species (S Table 5). R. montagui eye lenses contained siginificantly lower values of δ15N than the other 3 species which were not significantly different from each other, except R. brachyura and R. clavata, the latter having a significantly higher δ15N value (Figure). Raja montagui had overall eye lens tissue δ13C values significantly lower than R. clavata and R. microocellata but not R. brachyura (S Table 5) (Figure). R. brachyura was not significantly different from any other species.

207

S Table 5 Statistical results-differences in overall eye lens δ15N values between species

Species comparison P value Direction of difference R. brachyura and R. clavata 0.004 R. clavata higher than R brachyura R. microocellata and R. 0.154 N/A brachyura R. montagui and R. brachyura <0.001 R. brachyura higher than R. montagui R. microocellata and R. clavata 0.932 N/A R. microocellata and R. montagui <0.001 R. microocellata higher than R. montagui R. montagui and R. clavata <0.001 R. clavata higher than R. montagui

S Table 6: Statistical results-differences in overall eye lens δ13C values between species

Species comparisons P value Direction of difference R. montagui and R. clavata <0.001 R. clavata higher values than R. montagui R. montagui and R. 0.002 R. microocellata higher values than R. microocellata montagui

208

Appendix 5: Chapter 7 publications

209

210

211

212

213

214

215

216

217

218

219

220

References

ABBOTT, J. K. & HAYNIE, A. C. 2012. What are we protecting? Fisher behavior and the unintended consequences of spatial closures as a fishery management tool. Ecological Applications, 22, 762-777. ADACHI, T., COSTA, D. P., ROBINSON, P. W., PETERSON, S. H., YAMAMICHI, M., NAITO, Y. & TAKAHASHI, A. 2016. Searching for prey in a three-dimensional environment: hierarchical movements enhance foraging success in northern elephant seals. Functional Ecology, 361–369. ADAMS, R. & THIBAULT, K. 2006. Temporal resource partitioning by bats at water holes. Journal of Zoology, 270, 466-472. AGARDY, T., DI SCIARA, G. N. & CHRISTIE, P. 2011. Mind the gap: addressing the shortcomings of marine protected areas through large scale marine spatial planning. Marine Policy, 35, 226-232. AGUZZI, J. & COMPANY, J. B. 2010. Chronobiology of deep-water decapod crustaceans on continental margins. In: LESSER, M. (ed.) Advances in Marine Biology, Vol 58. San Diego: Elsevier Academic Press Inc. AGUZZI, J., COMPANY, J. B. & GARCÍA, J. A. 2009. Ontogenetic and gender-modulated behavioural rhythms in the deep-water decapods Liocarcinus depurator (Brachyura: Portunidae), Munida tenuimana and Munida intermedia (Anomura: Galatheidae). Marine Ecology, 30, 93-105. AGUZZI, J., SBRAGAGLIA, V., TECCHIO, S., NAVARRO, J. & COMPANY, J. B. 2015. Rhythmic behaviour of marine benthopelagic species and the synchronous dynamics of benthic communities. Deep-Sea Research Part I-Oceanographic Research Papers, 95, 1-11. AJAYI, T. 1982a. Food and feeding habits of Raja species (Batoidei) in Carmarthen Bay, Bristol channel. Journal of the Marine Biological Association of the United Kingdom, 62, 215-223. ALLDREDGE, A. L. & HAMNER, W. M. 1980. Recurring aggregation of zooplankton by a tidal current. Estuarine and Coastal Marine Science, 10, 31-37. ALÓS, J., PALMER, M. & ARLINGHAUS, R. 2012. Consistent selection towards low activity phenotypes when catchability depends on encounters among human predators and fish. PLoS One, 7, e48030. ANDERSEN, K. H., MARTY, L. & ARLINGHAUS, R. 2017. Evolution of boldness and life history in response to selective harvesting. Canadian Journal of Fisheries and Aquatic Sciences, 1-11. ANDREWS, K. S., WILLIAMS, G. D., FARRER, D., TOLIMIERI, N., HARVEY, C. J., BARGMANN, G. & LEVIN, P. S. 2009. Diel activity patterns of sixgill sharks, Hexanchus griseus: the ups and downs of an apex predator. Animal Behaviour, 78, 525-536. APPA RAO, T. & KRISHNAMOORTHI, B. 1982. Diurnal variation in the catches of demersal fishes in the northwest region of bay of bengal india during 1959-1960. Indian Journal of Fisheries, 29, 134-143. AUSTER, P. J., MALATESTA, R. J. & LAROSA, S. C. 1995. Patterns of microhabitat utilization by mobile megafauna on the southern New-England (USA) continental-shelf and slope. Marine ecology progress series, 127, 77-85. AUSTIN, D., BOWEN, W. & MCMILLAN, J. 2004. Intraspecific variation in movement patterns: modeling individual behaviour in a large marine predator. Oikos, 105, 15-30. BARNES, C., JENNINGS, S. & BARRY, J. T. 2009. Environmental correlates of large-scale spatial variation in the δ 13 C of marine animals. Estuarine, Coastal and Shelf Science, 81, 368- 374. BARNETT, A., ABRANTES, K. G., SEYMOUR, J. & FITZPATRICK, R. 2012. Residency and spatial use by reef sharks of an isolated seamount and its implications for conservation. PLoS One, 7, e36574.

221

BARRETT, J. H., LOCKER, A. M. & ROBERTS, C. M. 2004. The origins of intensive marine fishing in medieval Europe: the English evidence. Proceedings of the Royal Society of London B: Biological Sciences, 271, 2417-2421. BASS, N. C., MOURIER, J., KNOTT, N. A., DAY, J., GUTTRIDGE, T. & BROWN, C. 2016. Long-term migration patterns and bisexual philopatry in a benthic shark species. Marine and Freshwater Research. BATEMAN, B. L., VANDERWAL, J., WILLIAMS, S. E. & JOHNSON, C. N. 2012. Biotic interactions influence the projected distribution of a specialist mammal under climate change. Diversity and Distributions, 18, 861-872. BAVINCK, M. 2011. The megaengineering of ocean fisheries: A century of expansion and rapidly closing frontiers. Engineering Earth. Springer. BEARHOP, S., ADAMS, C. E., WALDRON, S., FULLER, R. A. & MACLEOD, H. 2004. Determining trophic niche width: a novel approach using stable isotope analysis. Journal of Animal Ecology, 73, 1007-1012. BERTRAND, S., DÍAZ, E. & ÑIQUEN, M. 2004. Interactions between fish and fisher's spatial distribution and behaviour: an empirical study of the anchovy (Engraulis ringens) fishery of Peru. ICES Journal of Marine Science, 61, 1127-1136. BIRD, C. S., VERÍSSIMO, A., MAGOZZI, S., ABRANTES, K. G., AGUILAR, A., AL-REASI, H., BARNETT, A., BETHEA, D. M., BIAIS, G. & BORRELL, A. 2018. A global perspective on the trophic geography of sharks. Nature Ecology & Evolution, 1. BIZZARRO, J. J., BROMS, K. M., LOGSDON, M. G., EBERT, D. A., YOKLAVICH, M. M., KUHNZ, L. A. & SUMMERS, A. P. 2014. Spatial segregation in eastern north Pacific skate assemblages. BLABER, S., BREWER, D. & HARRIS, A. 1994. Distribution, biomass and community structure of demersal fishes of the Gulf of Carpentaria, Australia. Marine and Freshwater Research, 45, 375-396. BLANCHARD, J. L., MILLS, C., JENNINGS, S., FOX, C. J., RACKHAM, B. D., EASTWOOD, P. D. & O'BRIEN, C. M. 2005. Distribution abundance relationships for North Sea Atlantic cod (Gadus morhua): observation versus theory. Canadian Journal of Fisheries and Aquatic Sciences, 62, 2001-2009. BOLLE, L. J., HUNTER, E., RIJNSDORP, A. D., PASTOORS, M. A., METCALFE, J. D. & REYNOLDS, J. D. 2005. Do tagging experiments tell the truth? Using electronic tags to evaluate conventional tagging data. ICES Journal of Marine Science: Journal du Conseil, 62, 236- 246. BOOTH, A. J. 2000. Incorporating the spatial component of fisheries data into stock assessment models. ICES Journal of Marine Science, 57, 858-865. BOTTARI, T., RINELLI, P., BIANCHINI, M. L. & RAGONESE, S. 2013. Stock identification of Raja clavata L.(Chondrichthyes, Rajidae) in two contiguous areas of the Mediterranean. Hydrobiologia, 703, 215-224. BRANDER, K. 1981. Disappearance of common skate, Raia batis, from the Irish Sea. Nature, 290. BRANDER, K. & PALMER, D. 1985. Growth rate of Raia clavata in the Northeast Irish Sea. Journal du Conseil, 42, 125-128. BURBIDGE, A. & WOINARSKI, J. 2016a. Notoryctes typhlops [Online]. Available: http://dx.doi.org/10.2305/IUCN.UK.2016-1.RLTS.T14879A21965004.en. [Accessed 15/02/2018. BURBIDGE, A. A. & WOINARSKI, J. 2016b. Bettongia tropica (errata version published in 2017) [Online]. IUCN Red List of Threatened Species 2016: e.T2787A115063540. Available: http://dx.doi.org/10.2305/IUCN.UK.2016-3.RLTS.T2787A21961037.en. [Accessed 16 March 2018. BURNS, J., TRUMBLE, S., CASTELLINI, M. & TESTA, J. 1998. The diet of Weddell seals in McMurdo Sound, Antarctica as determined from scat collections and stable isotope analysis. Polar Biology, 19, 272-282.

222

BURNS, J. M., HINDELL, M. A., BRADSHAW, C. J. A. & COSTA, D. P. 2008. Fine-scale habitat selection of crabeater seals as determined by diving behavior. Deep Sea Research Part II: Topical Studies in Oceanography, 55, 500-514. CABANA, G. & RASMUSSEN, J. B. 1996. Comparison of aquatic food chains using nitrogen isotopes. Proceedings of the National Academy of Sciences, 93, 10844-10847. CAILLIET, G. M., RADTKE, R. & WELDEN, B. Elasmobranch age determination and verification: a review. Indo-Pacific Fish Biology: proceedings of the second international conference on Indo-Pacific Fishes, 1986. Ichthyological Society of Japan, Tokyo, 345-360. CARRIER, J. C. & PRATT, H. L. 1998. Habitat management and closure of a nurse shark breeding and nursery ground. Fisheries Research, 39, 209-213. CARVALHO, F. C., MURIE, D. J., HAZIN, F. H., HAZIN, H. G., LEITE-MOURATO, B. & BURGESS, G. H. 2011. Spatial predictions of blue shark (Prionace glauca) catch rate and catch probability of juveniles in the Southwest Atlantic. ICES Journal of Marine Science: Journal du Conseil, 68, 890-900. CASEY, J. M. & MYERS, R. A. 1998a. Diel variation in trawl catchability: is it as clear as day and night? Canadian Journal of Fisheries and Aquatic Sciences, 55, 2329-2340. CASEY, J. M. & MYERS, R. A. 1998b. Near extinction of a large, widely distributed fish. Science, 281, 690-692. CAVENEY, S., SCHOLTZ, C. H. & MCINTYRE, P. 1995. Patterns of daily flight activity in onitine dung beetles (Scarabaeinae, onitini). Oecologia, 103, 444-452. CEC 2011. Council Regulation (EU) No 57/2011 of 18 January 2011 fixing for 2011 the fishing opportunities for certain fish stocks and groups of fish stocks, applicable in EU waters and, for EU vessels, in certain non-EU waters. Official Journal of the European Communities L24, 1–125. CHAPMAN, D. D., FELDHEIM, K. A., PAPASTAMATIOU, Y. P. & HUETER, R. E. 2015. There and back again: a review of residency and return migrations in sharks, with implications for population structure and management. Annual review of marine science, 7, 547-570. CHEVOLOT, M., ELLIS, J. R., HOARAU, G., RIJNSDORP, A. D., STAM, W. T. & OLSEN, J. L. 2006. Population structure of the thornback ray (< i> Raja clavata L.) in British waters. Journal of Sea Research, 56, 305-316. CLARK, R. S. 1922. Rays and Skates (Raiœ) No. 1.—Egg-Capsules and Young. Journal of the Marine Biological Association of the United Kingdom (New Series), 12, 578-643. COCHRANE, K. L. 2002. A fishery manager's guidebook: management measures and their application, Food & Agriculture Org. COELHO, R. & ERZINI, K. 2006. Reproductive aspects of the undulate ray, Raja undulata, from the south coast of Portugal. Fisheries Research, 81, 80-85. COLE, M. L., VALIELA, I., KROEGER, K. D., TOMASKY, G. L., CEBRIAN, J., WIGAND, C., MCKINNEY, R. A., GRADY, S. P. & CARVALHO DA SILVA, M. H. 2004. Assessment of a δ N Isotopic Method to Indicate Anthropogenic Eutrophication in Aquatic Ecosystems. Journal of Environmental Quality, 33, 124-132. COLLIN, S. P. 2012. The Neuroecology of Cartilaginous Fishes: Sensory Strategies for Survival. Brain Behavior and Evolution, 80, 80-96. CONNELL, S. & LINCOLN-SMITH, M. 1999. Depth and the structure of assemblages of demersal fish: experimental trawling along a temperate coast. Estuarine, Coastal and Shelf Science, 48, 483-495. COOK, R. 1998. A sustainability criterion for the exploitation of North Sea cod. ICES Journal of Marine Science: Journal du Conseil, 55, 1061-1070. COOKE, S. J., MIDWOOD, J. D., THIEM, J. D., KLIMLEY, P., LUCAS, M. C., THORSTAD, E. B., EILER, J., HOLBROOK, C. & EBNER, B. C. 2013a. Tracking animals in freshwater with electronic tags: past, present and future. Animal Biotelemetry, 1, 5.

223

COOKE, S. J., SACK, L., FRANKLIN, C. E., FARRELL, A. P., BEARDALL, J., WIKELSKI, M. & CHOWN, S. L. 2013b. What is conservation physiology? Perspectives on an increasingly integrated and essential science. Conservation Physiology, 1. CORTÉS, E. 1999. Standardized diet compositions and trophic levels of sharks. ICES Journal of Marine Science, 56, 707-717. CORTI, P. & SHACKLETON, D. M. 2002. Relationship between predation-risk factors and sexual segregation in Dall's sheep (Ovis dalli dalli). Canadian Journal of Zoology, 80, 2108-2117. COSTALAGO, D., NAVARRO, J., ÁLVAREZ-CALLEJA, I. & PALOMERA, I. 2012. Ontogenetic and seasonal changes in the feeding habits and trophic levels of two small pelagic fish species. Marine Ecology Progress Series, 460, 169-181. COTT, P. A., GUZZO, M. M., CHAPELSKY, A. J., MILNE, S. W. & BLANCHFIELD, P. J. 2015. Diel bank migration of Burbot (Lota lota). Hydrobiologia, 757, 3-20. CRAIG, H. 1953. The geochemistry of the stable carbon isotopes. Geochimica et cosmochimica acta, 3, 53-92. CREEL, S. & CREEL, N. M. 2002. The African wild dog: behavior, ecology, and conservation, Princeton University Press. CUNNINGHAM, J. 1896. Marketable Marine Fishes. London. DAMALAS, D., KATSANEVAKIS, S., MARAVELIAS, C., KARAGEORGIS, A., ANAGNOSTOU, C., TSITSIKA, E. & PAPACONSTANTINOU, C. 2016. Habitat preferences and spatiotemporal distribution of four benthic skates in the eastern Mediterranean Sea. DANIEL, H. & MINOT, F. 1954. The inexhaustible sea, Dodd, Mead. DE ROBERTIS, A. & HANDEGARD, N. O. 2012. Fish avoidance of research vessels and the efficacy of noise-reduced vessels: a review. ICES Journal of Marine Science: Journal du Conseil, fss155. DE VOS, A., O'RIAIN, M. J., MEYER, M. A., KOTZE, P. G. H. & KOCK, A. A. 2015. Behavior of Cape fur seals (Arctocephalus pusillus pusillus) in response to spatial variation in white shark (Carcharodon carcharias) predation risk. Marine Mammal Science, 31, 1234-1251. DENIRO, M. J. & EPSTEIN, S. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et cosmochimica acta, 42, 495-506. DENIRO, M. J. & EPSTEIN, S. 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et cosmochimica acta, 45, 341-351. DI SANTO, V., BLEVINS, E. L. & LAUDER, G. V. 2017. Batoid locomotion: effects of speed on pectoral fin deformation in the little skate, Leucoraja erinacea. Journal of Experimental Biology, 220, 705-712. DIAZ PAULI, B., WIECH, M., HEINO, M. & UTNE‐PALM, A. C. 2015. Opposite selection on behavioural types by active and passive fishing gears in a simulated guppy Poecilia reticulata fishery. Journal of Fish Biology, 86, 1030-1045. DINGLE, H. 2014. Migration: the biology of life on the move, Oxford University Press, USA. DOMEIER, M. L. & NASBY-LUCAS, N. 2012. Sex-specific migration patterns and sexual segregation of adult white sharks, Carcharodon carcharias, in the Northeastern Pacific. Global Perspectives on the Biology and Life History of the White Shark’.(Ed. ML Domeier.) pp, 133-146. DOMI, N., BOUQUEGNEAU, J.-M. & DAS, K. 2005. Feeding ecology of five commercial shark species of the Celtic Sea through stable isotope and trace metal analysis. Marine Environmental Research, 60, 551-569. DOVE, S. & KINGSFORD, M. 1998. Use of otoliths and eye lenses for measuring trace-metal incorporation in fishes: a biogeographic study. Marine Biology, 130, 377-387. DUDGEON, C. L., NOAD, M. J. & LANYON, J. M. 2008. Abundance and demography of a seasonal aggregation of zebra sharks Stegostoma fasciatum. Marine Ecology Progress Series, 368, 269-281.

224

DULVY, N. K., FOWLER, S. L., MUSICK, J. A., CAVANAGH, R. D., KYNE, P. M., HARRISON, L. R., CARLSON, J. K., DAVIDSON, L. N., FORDHAM, S. V. & FRANCIS, M. P. 2014. Extinction risk and conservation of the world’s sharks and rays. Elife, 3, e00590. DULVY, N. K., METCALFE, J. D., GLANVILLE, J., PAWSON, M. G. & REYNOLDS, J. D. 2000. Fishery stability, local extinctions, and shifts in community structure in skates. Conservation Biology, 14, 283-293. DULVY, N. K., NOTOBARTOLO DI SCIARA, G., SERENA, F., TINITI, F., & UNGARO, N., MANCUSI, C., ELLIS, J. 2006. Dipturus Batis 2008 IUCN Red List of Threatened Species [Online]. Available: www.iucnredlist.org. EBERT, D., COWLEY, P. & COMPAGNO, L. 1991. A preliminary investigation of the feeding ecology of skates (Batoidea: Rajidae) off the west coast of southern Africa. South African Journal of Marine Science, 10, 71-81. EBERT, D. A. & BIZZARRO, J. J. 2007. Standardized diet compositions and trophic levels of skates (Chondrichthyes: Rajiformes: Rajoidei). Environmental Biology of fishes, 80, 221-237. ELITH, J., GRAHAM, C. H., ANDERSON, R. P., DUDÍK, M., FERRIER, S., GUISAN, A., HIJMANS, R. J., HUETTMANN, F., LEATHWICK, J. R. & LEHMANN, A. 2006. Novel methods improve prediction of species' distributions from occurrence data. Ecography, 129-151. ELLIS, J., CRUZ-MARTINEZ, A., RACKHAM, B. & ROGERS, S. 2005a. The distribution of chondrichthyan fishes around the British Isles and implications for conservation. Journal of Northwest Atlantic Fishery Science, 35, 195-213. ELLIS, J., DULVY, N. K., JENNINGS, S., PARKER-HUMPHREYS, M. & ROGERS, S. I. 2005b. Assessing the status of demersal elasmobranchs in UK waters: a review. Journal of the Marine Biological Association of the United Kingdom, 85, 1025-1047. ELLIS, J., MOREL, G., BURT, G. & BOSSY, S. 2011. Preliminary observations on the life history and movements of skates (Rajidae) around the Island of Jersey, western English Channel. Journal of the Marine Biological Association of the United Kingdom, 91, 1185-1192. ELLIS, J., PAWSON, M. & SHACKLEY, S. 1996a. The comparative feeding ecology of six species of shark and four species of ray (Elasmobranchii) in the north-east Atlantic. Journal of the Marine Biological Association of the United Kingdom, 76, 89-106. ELLIS, J. & SHACKLEY, S. 1995. Observations on egg‐laying in the thornback ray. Journal of Fish Biology, 46, 903-904. ELLIS, J. R. 2006. Raja microocellata [Online]. The IUCN Red List of Threatened Species 2006: e.T39400A10225571. Available: http://dx.doi.org/10.2305/IUCN.UK.2006.RLTS.T39400A10225571.en. [Accessed 25/03/2018. ELLIS, J. R. 2016. Raja clavata [Online]. The IUCN Red List of Threatened Species 2016: e.T39399A103110667. Available: http://dx.doi.org/10.2305/IUCN.UK.2016- 3.RLTS.T39399A103110667.en. [Accessed 25/03/2018. ELLIS, J. R., CRUZ-MARTINEZ, A., RACKHAM, B. D. & ROGERS, S. I. 2005c. The distribution of chondrichthyan fishes around the British Isles and implications for conservation. Journal of Northwest Atlantic Fishery Science, 35, 195-213. ELLIS, J. R., UNGARO, N., SERENA, F., DULVY, N. K., TINTI, F., BERTOZZI, M., PASOLINI, P., MANCUSI, C. & NOARBARTOLO DI SCIARA, G. 2007. Raja montagui [Online]. The IUCN Red List of Threatened Species 2007: e.T63146A12623141. Available: http://dx.doi.org/10.2305/IUCN.UK.2007.RLTS.T63146A12623141.en [Accessed 25/03/2018. ELLIS, J. R., UNGARO, N., SERENA, F., DULVY, N. K., TINTI, F., BERTOZZI, M., PASOLINI, P., MANCUSI, C. & NOARBARTOLO DI SCIARA, G. 2009. Raja brachyura [Online]. The IUCN Red List of Threatened Species 2009: e.T161691A5481210. Available: http://dx.doi.org/10.2305/IUCN.UK.2009-2.RLTS.T161691A5481210.en [Accessed 25/03/2018.

225

ENEVER, R., CATCHPOLE, T., ELLIS, J. & GRANT, A. 2009. The survival of skates (Rajidae) caught by demersal trawlers fishing in UK waters. Fisheries Research, 97, 72-76. ENGELHARD, G. H. 2009. One hundred and twenty years of change in fishing power of English North Sea trawlers. Advances in Fisheries Science: 50 years on from Beverton and Holt, 1- 25. ESPINOZA, M., CLARKE, T., VILLALOBOS‐ROJAS, F. & WEHRTMANN, I. 2012. Ontogenetic dietary shifts and feeding ecology of the rasptail skate Raja velezi and the brown smoothhound shark Mustelus henlei along the Pacific coast of Costa Rica, Central America. Journal of Fish Biology, 81, 1578-1595. ESPINOZA, M., FARRUGIA, T. J. & LOWE, C. G. 2011. Habitat use, movements and site fidelity of the gray smooth-hound shark (Mustelus californicus Gill 1863) in a newly restored southern California estuary. Journal of Experimental Marine Biology and Ecology, 401, 63- 74. ESTRADA, J. A., RICE, A. N., NATANSON, L. J. & SKOMAL, G. B. 2006. Use of isotopic analysis of vertebrae in reconstructing ontogenetic feeding ecology in white sharks. Ecology, 87, 829- 834. EUROPEAN ENVIRONMENT AGENCY. 2018. The European Nature Information System, EUNIS [Online]. Available: https://eunis.eea.europa.eu/about [Accessed 23/03/2018. EUROPEANCOMMISSION. 2016. Updates to Commission Proposal COM (2016) 698, fixing for 2017 the fishing opportunities for certain fish stocks and groups of fish stocks, applicable in Union waters and, for Union fishing vessels, in certain non-Union waters. COMMISSION SERVICES NON PAPER [Online]. Available: https://ec.europa.eu/fisheries/sites/fisheries/files/2016-11-15-com-2016-698-non- paper.pdf.

EVANS, K., ABASCAL, F., KOLODY, D., SIPPEL, T., HOLDSWORTH, J. & MARU, P. 2014. The horizontal and vertical dynamics of swordfish in the South Pacific Ocean. Journal of Experimental Marine Biology and Ecology, 450, 55-67.

FARIAS, I., FIGUEIREDO, I., MOURA, T., GORDO, L. S., NEVES, A. & SERRA-PEREIRA, B. 2006a. Diet comparison of four ray species (Raja clavata, Raja brachyura, Raja montagui and Leucoraja naevus) caught along the Portuguese continental shelf. Aquatic Living Resources, 19, 105-114. FAUNCE, C. H. 2011. A comparison between industry and observer catch compositions within the Gulf of Alaska rockfish fishery. ICES Journal of Marine Science: Journal du Conseil, fsr093. FENBERG, P. B. & ROY, K. 2008. Ecological and evolutionary consequences of size‐selective harvesting: how much do we know? Molecular Ecology, 17, 209-220. FERNANDES, P. G., RALPH, G. M., NIETO, A., CRIADO, M. G., VASILAKOPOULOS, P., MARAVELIAS, C. D., COOK, R. M., POLLOM, R. A., KOVAČIĆ, M. & POLLARD, D. 2017. Coherent assessments of Europe’s marine fishes show regional divergence and megafauna loss. Nature Ecology & Evolution, 1, 0170. FIELD, I. C., BRADSHAW, C. J., BURTON, H. R., SUMNER, M. D. & HINDELL, M. A. 2005. Resource partitioning through oceanic segregation of foraging juvenile southern elephant seals (Mirounga leonina). Oecologia, 142, 127-135. FIELD, I. C., MEEKAN, M. G., BUCKWORTH, R. C. & BRADSHAW, C. J. 2009. Susceptibility of sharks, rays and chimaeras to global extinction. Advances in marine biology, 56, 275-363. FISHERIES AND AQUACULTURE DEPARTMENT. 2017. About us - Fisheries and Aquaculture Department. In: FAO Fisheries and Aquaculture Department [online] [Online]. Rome: FAO. [Accessed 05/09/2017. FOCK, H. O. 2014. Patterns of extirpation. I. Changes in habitat use by thornback rays Raja clavata in the German Bight for 1902-1908, 1930-1932, and 1991-2009. Endangered Species Research, 25, 197-207.

226

FORMAN, J. & DUNN, M. 2012. Diet and scavenging habits of the smooth skate Dipturus innominatus. Journal of Fish Biology, 80, 1546-1562. FRANCIS, R. 1984. An adaptive strategy for stratified random trawl surveys. New Zealand Journal of Marine and Freshwater Research, 18, 59-71. FRASER, H. M., GREENSTREET, S. P. R., FRYER, R. J. & PIET, G. J. 2008. Mapping spatial variation in demersal fish species diversity and composition in the North Sea: accounting for species- and size-related catchability in survey trawls. ICES Journal of Marine Science, 65, 531-538. FRASER, H. M., GREENSTREET, S. P. R. & PIET, G. J. 2007. Taking account of catchability in groundfish survey trawls: implications for estimating demersal fish biomass. ICES Journal of Marine Science, 64, 1800-1819. FULTON, T. W. 1893. An experimental investigation on the migration and rate of growth of the food fishes. Rep. Fish. Bd. Scot, 11, 176-196. GALLAGHER, A., SERAFY, J., COOKE, S. & HAMMERSCHLAG, N. 2014. Physiological stress response, reflex impairment, and survival of five sympatric shark species following experimental capture and release. Marine Ecology Progress Series, 496, 207-218. GALLAGHER, M., NOLAN, C. & JEAL, F. 2005. Age, growth and maturity of the commercial ray species from the Irish Sea. J Northwest Atl Fish Sci, 35, 47-66. GARCES, L., STOBUTZKI, I., ALIAS, M., CAMPOS, W., KOONGCHAI, N., LACHICA-ALINO, L., MUSTAFA, G., NURHAKIM, S., SRINATH, M. & SILVESTRE, G. 2006. Spatial structure of demersal fish assemblages in South and Southeast Asia and implications for fisheries management. Fisheries Research, 78, 143-157. GENNER, M. J., SIMS, D. W., SOUTHWARD, A. J., BUDD, G. C., MASTERSON, P., MCHUGH, M., RENDLE, P., SOUTHALL, E. J., WEARMOUTH, V. J. & HAWKINS, S. J. 2010a. Body size- dependent responses of a marine fish assemblage to climate change and fishing over a century-long scale. Global Change Biology, 16, 517-527. GILBERT, F., BONIN, P. & STORA, G. 1995. Effect of bioturbation on denitrification in a marine sediment from the West Mediterranean littoral. Hydrobiologia, 304, 49-58. GILLANDERS, B. M. & KINGSFORD, M. J. 2000. Elemental fingerprints of otoliths of fish may distinguish estuarine'nursery'habitats. Marine Ecology Progress Series, 201, 273-286. GOODYEAR, C. P., LUO, J. G., PRINCE, E. D., HOOLIHAN, J. P., SNODGRASS, D., ORBESEN, E. S. & SERAFY, J. E. 2008. Vertical habitat use of Atlantic blue marlin Makaira nigricans: interaction with pelagic longline gear. Marine Ecology-Progress Series, 365, 233-245. GORMAN, O. T., YULE, D. L. & STOCKWELL, J. D. 2012. Habitat use by fishes of Lake Superior. I. Diel patterns of habitat use in nearshore and offshore waters of the Apostle Islands region. Aquatic Ecosystem Health & Management, 15, 333-354. GRAINGER, R., HENRY, J., SAHA, M. & SERVETNICK, M. 1992. Recent progress on the mechanisms of embryonic lens formation. Eye, 6, 117-122. GRAY, C. & OTWAY, N. 1994. Spatial and temporal differences in assemblages of demersal fishes on the inner continental shelf off Sydney, south-eastern Australia. Marine and Freshwater Research, 45, 665-676. GREENSTREET, S. P. R., HOLLAND, G. J., GUIREY, E. J., ARMSTRONG, E., FRASER, H. M. & GIBB, I. M. 2010. Combining hydroacoustic seabed survey and grab sampling techniques to assess "local" sandeel population abundance. ICES Journal of Marine Science, 67, 971-984. GREENWAY, E., JONES, K. S. & COOKE, G. M. 2016. Environmental enrichment in captive juvenile thornback rays, Raja clavata (Linnaeus 1758). Applied Animal Behaviour Science, 182, 86- 93. GRIFFITHS, A. M., MILLER, D. D., EGAN, A., FOX, J., GREENFIELD, A. & MARIANI, S. 2013. DNA barcoding unveils skate (Chondrichthyes: Rajidae) species diversity in ‘ray’products sold across Ireland and the UK. PeerJ, 1, e129. GRIFFITHS, A. M., SIMS, D. W., COTTERELL, S. P., EL NAGAR, A., ELLIS, J. R., LYNGHAMMAR, A., MCHUGH, M., NEAT, F. C., PADE, N. G. & QUEIROZ, N. 2010. Molecular markers reveal

227

spatially segregated cryptic species in a critically endangered fish, the common skate (Dipturus batis). Proceedings of the Royal Society of London B: Biological Sciences, rspb20092111. GRUBBS, R. D. 2010. Ontogenetic shifts in movements and habitat use. Sharks and their relatives II: Biodiversity, adaptive physiology, and conservation. CRC Press, Boca Raton, p319-350. GRUBBS, R. D., CARLSON, J. K., ROMINE, J. G., CURTIS, T. H., MCELROY, W. D., MCCANDLESS, C. T., COTTON, C. F. & MUSICK, J. A. 2016. Critical assessment and ramifications of a purported marine trophic cascade. Scientific reports, 6, 20970. GUÉNETTE, S., LAUCK, T. & CLARK, C. 1998. Marine reserves: from Beverton and Holt to the present. Reviews in Fish Biology and Fisheries, 8, 251-272. GUNDERSON, D. R. 1993. Surveys of fisheries resources, John Wiley & Sons. GUTTRIDGE, T., GRUBER, S., GLEDHILL, K., CROFT, D., SIMS, D. & KRAUSE, J. 2009. Social preferences of juvenile lemon sharks, Negaprion brevirostris. Animal Behaviour, 78, 543- 548. HALL, L. S., KRAUSMAN, P. R. & MORRISON, M. L. 1997. The habitat concept and a plea for standard terminology. Wildlife Society Bulletin, 173-182. HALPERN, B. S., GAINES, S. D. & WARNER, R. R. 2004. Confounding effects of the export of production and the displacement of fishing effort from marine reserves. Ecological Applications, 14, 1248-1256. HALPERN, B. S., LESTER, S. E. & KELLNER, J. B. 2009. Spillover from marine reserves and the replenishment of fished stocks. Environmental conservation, 36, 268-276. HARLEY, S. J. & MYERS, R. A. 2001. Hierarchical Bayesian models of length-specific catchability of research trawl surveys. Canadian Journal of Fisheries and Aquatic Sciences, 58, 1569- 1584. HARRISON, P. M., GUTOWSKY, L. F. G., MARTINS, E. G., PATTERSON, D. A., LEAKE, A., COOKE, S. J. & POWER, M. 2013. Diel vertical migration of adult burbot: a dynamic trade-off among feeding opportunity, predation avoidance, and bioenergetic gain. Canadian Journal of Fisheries and Aquatic Sciences, 70, 1765-1774. HAYS, G., KENNEDY, H. & FROST, B. 2001. Individual variability in diel vertical migration of a marine copepod: why some individuals remain at depth when others migrate. Limnology and Oceanography, 46, 2050-2054. HAYS, G. C. 2003. A review of the adaptive significance and ecosystem consequences of zooplankton diel vertical migrations. Hydrobiologia, 503, 163-170. HAYS, G. C., FERREIRA, L. C., SEQUEIRA, A. M., MEEKAN, M. G., DUARTE, C. M., BAILEY, H., BAILLEUL, F., BOWEN, W. D., CALEY, M. J. & COSTA, D. P. 2016. Key questions in marine megafauna movement ecology. Trends in ecology & evolution, 31, 463-475. HEAPE, W. 1887. Notes on the fishing industry of Plymouth. Journal of the Marine Biological Association of the United Kingdom, 1, 45-95. HEAPE, W. 1888. Preliminary report upon the fauna and flora of Plymouth Sound. Journal of the Marine Biological Association of the United Kingdom, 1, 153-193. HEIN, L., VAN KOPPEN, K., DE GROOT, R. S. & VAN IERLAND, E. C. 2006. Spatial scales, stakeholders and the valuation of ecosystem services. Ecological economics, 57, 209-228. HEITHAUS, M., DILL, L., MARSHALL, G. & BUHLEIER, B. 2002. Habitat use and foraging behavior of tiger sharks (Galeocerdo cuvier) in a seagrass ecosystem. Marine Biology, 140, 237-248. HEITHAUS, M. R. 2001. The biology of tiger sharks, Galeocerdo cuvier, in Shark Bay, Western Australia: sex ratio, size distribution, diet, and seasonal changes in catch rates. Environmental Biology of fishes, 61, 25-36. HEITHAUS, M. R., FRID, A., WIRSING, A. J. & WORM, B. 2008. Predicting ecological consequences of marine top predator declines. Trends in ecology & evolution, 23, 202-210. HERDMAN, W. A. & DAWSON, R. A. 1902. Fishes and Fisheries of the Irish Sea: And Especially of the Lancashire and Western Sea-fisheries District, George Philip & Son.

228

HEUPEL, M., SEMMENS, J. M. & HOBDAY, A. 2006. Automated acoustic tracking of aquatic animals: scales, design and deployment of listening station arrays. Marine and Freshwater Research, 57, 1-13. HOBSON, K., ATWELL, L., WASSENAAR, L. & YERKES, T. 2004. Estimating endogenous nutrient allocations to reproduction in redhead ducks: a dual isotope approach using δD and δ13C measurements of female and egg tissues. Functional ecology, 18, 737-745. HOBSON, K. A. 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia, 120, 314-326. HOBSON, K. A. & CLARK, R. G. 1992. Assessing avian diets using stable isotopes .1. turnover of c- 13 in tissues. Condor, 94, 181-188. HOBSON, K. A., PIATT, J. F. & PITOCCHELLI, J. 1994. Using stable isotopes to determine seabird trophic relationships. Journal of Animal Ecology, 63, 786-798. HOEGH-GULDBERG, O. & BRUNO, J. F. 2010. The impact of climate change on the world’s marine ecosystems. Science, 328, 1523-1528. HOLDEN, M. 1975. The fecundity of Raja clavata in British waters. Journal du Conseil, 36, 110-118. HOLDEN, M. & TUCKER, R. 1974. The food of Raja clavata Linnaeus 1758, Raja montagui Fowler 1910, Raja naevus Müller and Henle 1841 and Raja brachyura Lafont 1873 in British waters. Journal du Conseil, 35, 189-193. HORN, H. S. 1966. Measurements of ‘overlap’ in comparative ecological studies. American Naturalist, 100, 419:424. HORWITZ, J. 1992. Alpha-crystallin can function as a molecular chaperone. Proceedings of the National Academy of Sciences, 89, 10449-10453. HOVE, J. R. & MOSS, S. A. 1997. Effect of MS-222 on response to light and rate of metabolism of the little skate Raja erinacea. Marine Biology, 128, 579-583. HOWARTH, L. M., PICKUP, S. E., EVANS, L. E., CROSS, T. J., HAWKINS, J. P., ROBERTS, C. M. & STEWART, B. D. 2015. Sessile and mobile components of a benthic ecosystem display mixed trends within a temperate marine reserve. Marine Environmental Research, 107, 8- 23. HUGHES, T. P., BELLWOOD, D. R., FOLKE, C., STENECK, R. S. & WILSON, J. 2005. New paradigms for supporting the resilience of marine ecosystems. Trends in ecology & evolution, 20, 380- 386. HUMPHRIES, N. E., QUEIROZ, N., DYER, J. R., PADE, N. G., MUSYL, M. K., SCHAEFER, K. M., FULLER, D. W., BRUNNSCHWEILER, J. M., DOYLE, T. K. & HOUGHTON, J. D. 2010. Environmental context explains Lévy and Brownian movement patterns of marine predators. Nature, 465, 1066. HUMPHRIES, N. E., SCHAEFER, K. M., FULLER, D. W., PHILLIPS, G. E. M., WILDING, C. & SIMS, D. W. 2016a. Scale-dependent to scale-free: daily behavioural switching and optimized searching in a marine predator. Animal Behaviour, 113, 189-201. HUMPHRIES, N. E., SIMPSON, S. J., WEARMOUTH, V. J. & SIMS, D. W. 2016b. Two's company, three's a crowd: fine-scale habitat partitioning by depth among sympatric species of marine mesopredator. Marine Ecology Progress Series, 561, 173-187. HUNSICKER, M. E., ESSINGTON, T. E., AYDIN, K. Y. & ISHIDA, B. 2010. Predatory role of the commander squid Berryteuthis magister in the eastern Bering Sea: insights from stable isotopes and food habits. Marine Ecology Progress Series, 415, 91-108. HUNTER, E., BERRY, F., BUCKLEY, A. A., STEWART, C. & METCALFE, J. D. 2006. Seasonal migration of thornback rays and implications for closure management. Journal of Applied Ecology, 43, 710-720. HUNTER, E., BUCKLEY, A. A., STEWART, C. & METCALFE, J. D. 2005a. Repeated seasonal migration by a thornback ray in the southern North Sea. Journal of the Marine Biological Association of the United Kingdom, 85, 1199-1200.

229

HUNTER, E., BUCKLEY, A., STEWART, C. & METCALFE, J. 2005b. Migratory behaviour of the thornback ray, Raja clavata, in the southern North Sea. Journal of the Marine Biological Association of the United Kingdom, 85, 1095-1105. HUNTER, E., BUCKLEY, A. A., STEWART, C. & METCALFE, J. D. 2005d. Migratory behaviour of the thornback ray Raja clavata, in the southern North Sea. Journal of the Marine Biological Association of the United Kingdom, 85, 1095-1105. HUSSEY, N. E., KESSEL, S. T., AARESTRUP, K., COOKE, S. J., COWLEY, P. D., FISK, A. T., HARCOURT, R. G., HOLLAND, K. N., IVERSON, S. J. & KOCIK, J. F. 2015. Aquatic animal telemetry: a panoramic window into the underwater world. Science, 348, 1255642. HUTTO, R. L. 1985. Habitat selection by nonbreeding, migratory land. Habitat selection in birds, 455. ICES 2012. Manual for the International Bottom Trawl Surveys. Series of ICES Survey Protocols, SISP 1-IBTS VIII., 68 pp. IFCA. 2014. Start Point to Plymouth Sound an Eddystone cSAC [Online]. IFCA. Available: https://secure.toolkitfiles.co.uk/clients/15340/sitedata/Misc/Annex_5_St_pt_to_Ply_Edd y_.pdf [Accessed 05/09/2017. IFCA. 2017. January 2017: Consultation opens for Whitsand and Looe Bay MCZ (Fishing Restrictions) Byelaw 2017 [Online]. Available: http://www.cornwall-ifca.gov.uk/january- 2017-consultation-opens-for-whitsand-and-l [Accessed 03/10/2017. IGLÉSIAS, S. P., TOULHOAT, L. & SELLOS, D. Y. 2010. Taxonomic confusion and market mislabelling of threatened skates: important consequences for their conservation status. Aquatic Conservation: Marine and Freshwater Ecosystems, 20, 319-333. INGER, R. & BEARHOP, S. 2008. Applications of stable isotope analyses to avian ecology. Ibis, 150, 447-461. IRIGOIEN, X., CONWAY, D. V. P. & HARRIS, R. P. 2004. Flexible diel vertical migration behaviour of zooplankton in the Irish Sea. Marine ecology progress series, 267, 85-97. JACKSON, A. L., INGER, R., PARNELL, A. C. & BEARHOP, S. 2011. Comparing isotopic niche widths among and within communities: SIBER–Stable Isotope Bayesian Ellipses in R. Journal of Animal Ecology, 80, 595-602. JACKSON, J. B. 2008. Ecological extinction and evolution in the brave new ocean. Proceedings of the National Academy of Sciences, 105, 11458-11465. JACKSON, J. B., KIRBY, M. X., BERGER, W. H., BJORNDAL, K. A., BOTSFORD, L. W., BOURQUE, B. J., BRADBURY, R. H., COOKE, R., ERLANDSON, J. & ESTES, J. A. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science, 293, 629-637. JACOBY, D. M., CROFT, D. P. & SIMS, D. W. 2012. Social behaviour in sharks and rays: analysis, patterns and implications for conservation. Fish and Fisheries, 13, 399-417. JENNINGS, S. & WARR, K. 2003. Environmental correlates of large-scale spatial variation in the δ15N of marine animals. Marine Biology, 142, 1131-1140. JNCC & NATURAL ENGLAND. 2016. Review of the MCZ features of conservation importance [Online]. Peterborough. Available: http://jncc.defra.gov.uk/pdf/20160512_MCZReviewFOCI_v7.0.pdf 2017]. JOHNSON, C. 1994. Fruiting of hypogeous fungi in dry sclerophyll forest in Tasmania, Australia: seasonal variation and annual production. Mycological research, 98, 1173-1182. JOHNSON, C. & MCILWEE, A. 1997. Ecology of the northern bettong, Bettongia tropica, a tropical mycophagist. Wildlife Research, 24, 549-559. JOHNSON, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology, 61, 65-71. JUNE, F. & CARLSON, F. 1971. Food of young Atlantic menhaden, Brevoortia tyrannus, in relation to metamorphosis. Fish. Bull, 68, 493-512. KADRI, H., MAROUANI, S., BRADAI, M. N. & BOUAÏN, A. 2014. Diet and feeding strategy of thornback ray, Raja clavata (Chondrichthyes: Rajidae) from the Gulf of Gabes (Tunisia—

230

Central Mediterranean Sea). Journal of the Marine Biological Association of the United Kingdom, 94, 1509-1516. KAISER, M. & RAMSAY, K. 1997. Opportunistic feeding by dabs within areas of trawl disturbance: possible implications for increased survival. Marine Ecology Progress Series. KAISER, M. J. 2005. Are marine protected areas a red herring or fisheries panacea? Canadian Journal of Fisheries and Aquatic Sciences, 62, 1194-1199. KEIFCA. 2017. Minimum Sizes [Online]. IFCA. Available: http://www.kentandessex-ifca.gov.uk/i- want-to-find-out-about/regulations/minimum-sizes/ [Accessed 05/09/2017. KESSEL, S., COOKE, S., HEUPEL, M., HUSSEY, N., SIMPFENDORFER, C., VAGLE, S. & FISK, A. 2014. A review of detection range testing in aquatic passive acoustic telemetry studies. Reviews in Fish Biology and Fisheries, 24, 199-218. KIMBER, J., SIMS, D., BELLAMY, P. & GILL, A. 2013. Elasmobranch cognitive ability: using electroreceptive foraging behaviour to demonstrate learning, habituation and memory in a benthic shark. Animal Cognition, 1-11. KIMBER, J. A., SIMS, D. W., BELLAMY, P. H. & GILL, A. B. 2009. Male–female interactions affect foraging behaviour within groups of small-spotted catshark, Scyliorhinus canicula. Animal Behaviour, 77, 1435-1440. KING, J. R. & MCFARLANE, G. A. 2010. Movement patterns and growth estimates of big skate (< i> Raja binoculata) based on tag-recapture data. Fisheries Research, 101, 50-59. KLAASSEN, R. H., NOLET, B. A. & BANKERT, D. 2006. Movement of foraging tundra swans explained by spatial pattern in cryptic food densities. Ecology, 87, 2244-2254. KLIMLEY, A. P. 1987. The determinants of sexual segregation in the scalloped hammerhead shark, Sphyrna lewini. Environmental Biology of fishes, 18, 27-40. KNIJN, R., BOON, T., HEESEN, H. & HISLOP, J. 1993. Atlas of North Sea Fishes–ICES Cooperative Research Report No. 194. KNIP, D. M., HEUPEL, M. R., SIMPFENDORFER, C. A., TOBIN, A. J. & MOLONEY, J. 2011. Ontogenetic shifts in movement and habitat use of juvenile pigeye sharks Carcharhinus amboinensis in a tropical nearshore region. Marine Ecology Progress Series, 425, 233-246. KOESTER, D. M. & SPIRITO, C. P. 2003. Punting: An unusual mode of locomotion in the little skate, Leucoraja erinacea (Chondrichthyes : Rajidae). Copeia, 553-561. KOHLER, N. E. & TURNER, P. A. 2001. Shark tagging: a review of conventional methods and studies. The behavior and sensory biology of elasmobranch fishes: an anthology in memory of Donald Richard Nelson. Springer. KOOP, J. H. 2005. Reproduction of captive Raja spp. in the Dolfinarium Harderwijk. Journal of the Marine Biological Association of the United Kingdom, 85, 1201-1202. KOTWICKI, S. & WEINBERG, K. L. 2005. Estimating capture probability of a survey bottom trawl for Bering Sea skates (Bathyraja spp.) and other fish. Alaska Fishery Research Bulletin, 11, 135-145. KRAMER, D. L., RANGELEY, R. W. & CHAPMAN, L. J. 1997. Habitat selection: patterns of spatial distribution from behavioural decisions. Behavioural ecology of teleost fishes, 37-80. KRAUSMAN, P. R. 1999. Some basic principles of habitat use. Grazing behavior of livestock and wildlife, 70, 85-90. KREBS, J. R. & DAVIES, N. B. 1997. Behavioural Ecology: An Evolutionary Approach, Oxford, Blackwell. KRISTENSEN, E., PENHA-LOPES, G., DELEFOSSE, M., VALDEMARSEN, T., QUINTANA, C. O. & BANTA, G. T. 2012. What is bioturbation? The need for a precise definition for fauna in aquatic sciences. Marine Ecology Progress Series, 446, 285-302. LAMPERT, W. 1989. The adaptive significance of diel vertical migration of zooplankton. Functional Ecology, 3, 21-27. LARKUM, A. & WEST, R. 1990. Long-term changes of seagrass meadows in Botany Bay, Australia. Aquatic Botany, 37, 55-70.

231

LAWTON, P. 1987. Diel activity and foraging behavior of juvenile american lobsters, Homarus americanus. Canadian Journal of Fisheries and Aquatic Sciences, 44, 1195-1205. LEA, J. S., HUMPHRIES, N. E., VON BRANDIS, R. G., CLARKE, C. R. & SIMS, D. W. Acoustic telemetry and network analysis reveal the space use of multiple reef predators and enhance marine protected area design. Proc. R. Soc. B, 2016. The Royal Society, 20160717. LEA, J. S., WETHERBEE, B. M., SOUSA, L. L., AMING, C., BURNIE, N., HUMPHRIES, N. E., QUEIROZ, N., HARVEY, G. M., SIMS, D. W. & SHIVJI, M. S. 2018. Ontogenetic partial migration is associated with environmental drivers and influences fisheries interactions in a marine predator. ICES Journal of Marine Science. LI, Z. & JIANG, Z. 2008. Sexual segregation in Tibetan gazelle: a test of the activity budget hypothesis. Journal of Zoology, 274, 327-331. LIMA, S. L. & DILL, L. M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology, 68, 619-640. LLUCH-BELDA, D., CRAWFORD, R., KAWASAKI, T., MACCALL, A., PARRISH, R., SCHWARTZLOSE, R. & SMITH, P. 1989. World-wide fluctuations of sardine and anchovy stocks: the regime problem. South African Journal of Marine Science, 8, 195-205. LOTZE, H. K. & WORM, B. 2009. Historical baselines for large marine animals. Trends in ecology & evolution, 24, 254-262. LOVE, M. S., SCHROEDER, D. M., SNOOK, L., YORK, A. & COCHRANE, G. 2008. All their eggs in one basket: a rocky reef nursery for the longnose skate (Raja rhina Jordan & Gilbert, 1880) in the southern California Bight. Fishery Bulletin, 106, 471-475. LUCIFORA, L. O. & VASSALLO, A. I. 2002. Walking in skates (Chondrichthyes, Rajidae): anatomy, behaviour and analogies to tetrapod locomotion. Biological Journal of the Linnean Society, 77, 35-41. LUER, C. A. & GILBERT, P. W. 1985. Mating behavior, egg deposition, incubation period, and hatching in the clearnose skate, Raja eglanteria. Environmental Biology of fishes, 13, 161- 171. MACESIC, L. J. & KAJIURA, S. M. 2010. Comparative Punting Kinematics and Pelvic Fin Musculature of Benthic Batoids. Journal of Morphology, 271, 1219-1228. MACNEIL, M. A., SKOMAL, G. B. & FISK, A. T. 2005. Stable isotopes from multiple tissues reveal diet switching in sharks. Marine Ecology Progress Series, 302, 199-206. MACNEILL, D. & BRANDT, S. 1990. Ontogenetic shifts in gill-raker morphology and predicted prey capture efficiency of the alewife, Alosa pseudoharengus. Copeia, 164-171. MAGOZZI, S., YOOL, A., VANDER ZANDEN, H., WUNDER, M. & TRUEMAN, C. 2017. Using ocean models to predict spatial and temporal variation in marine carbon isotopes. Ecosphere, 8. MALPICA-CRUZ, L., HERZKA, S. Z., SOSA-NISHIZAKI, O. & LAZO, J. P. 2012. Tissue-specific isotope trophic discrimination factors and turnover rates in a marine elasmobranch: empirical and modeling results. Canadian Journal of Fisheries and Aquatic Sciences, 69, 551-564. MANGI, S. C. & AUSTEN, M. C. 2008. Perceptions of stakeholders towards objectives and zoning of marine-protected areas in southern Europe. Journal for Nature Conservation, 16, 271-280. MARKO, P. B., LEE, S. C., RICE, A. M., GRAMLING, J. M., FITZHENRY, T. M., MCALISTER, J. S., HARPER, G. R. & MORAN, A. L. 2004. Fisheries: mislabelling of a depleted reef fish. Nature, 430, 309-310. MARTIN, C., VAZ, S., ELLIS, J., COPPIN, F., LE ROY, D. & CARPENTIER, A. 2010. Spatio-temporal patterns in demersal elasmobranchs from trawl surveys in the eastern English Channel (1988-2008). Marine Ecology-progress Series, 417, 211-228. MATERN, S. A., CECH, J. J. & HOPKINS, T. E. 2000. Diel movements of bat rays, Myliobatis californica, in Tomales Bay, California: Evidence for behavioral thermoregulation ? Environmental Biology of Fishes, 58, 173-182.

232

MCCONNAUGHEY, R., MIER, K. & DEW, C. 2000. An examination of chronic trawling effects on soft-bottom benthos of the eastern Bering Sea. ICES Journal of Marine Science, 57, 1377- 1388. MCCULLY, S. R., SCOTT, F. & ELLIS, J. R. 2012. Lengths at maturity and conversion factors for skates (Rajidae) around the British Isles, with an analysis of data in the literature. ICES Journal of Marine Science: Journal du Conseil, 69, 1812-1822. MCFARLANE, G., WYDOSKI, R. S. & PRINCE, E. D. External tags and marks. Historical review of the development of external tags and marks. Amer. Fish. Soc. Symp, 1990. 9-29. MCNAB, B. K. 1963. Bioenergetics and the determination of home range size. The American Naturalist, 97, 133-140. MEHNER, T. 2012. Diel vertical migration of freshwater fishes - proximate triggers, ultimate causes and research perspectives. Freshwater Biology, 57, 1342-1359. MEHNER, T., KASPRZAK, P. & HOLKER, F. 2007. Exploring ultimate hypotheses to predict diel vertical migrations in coregonid fish. Canadian Journal of Fisheries and Aquatic Sciences, 64, 874-886. MELNYCHUK, M. C. & WALTERS, C. J. 2010. Estimating detection probabilities of tagged fish migrating past fixed receiver stations using only local information. Canadian Journal of Fisheries and Aquatic Sciences, 67, 641-658. MET OFFICE. 2015. Winter storms, January to February 2014 [Online]. Available: https://www.metoffice.gov.uk/climate/uk/interesting/2014-janwind [Accessed 26/03/2018. MEYER, C. G., PAPASTAMATIOU, Y. P. & HOLLAND, K. N. 2010. A multiple instrument approach to quantifying the movement patterns and habitat use of tiger (Galeocerdo cuvier) and Galapagos sharks (Carcharhinus galapagensis) at French Frigate Shoals, Hawaii. Marine Biology, 157, 1857-1868. MICHELOT, C., PINAUD, D., FORTIN, M., MAES, P., CALLARD, B., LEICHER, M. & BARBRAUD, C. 2017. Seasonal variation in coastal marine habitat use by the European shag: Insights from fine scale habitat selection modeling and diet. Deep Sea Research Part II: Topical Studies in Oceanography. MILLER, A. W. & AMBROSE, R. F. 2000. Sampling patchy distributions: comparison of sampling designs in rocky intertidal habitats. Marine Ecology Progress Series, 1-14. MMO. 2014. Fishing data collection, coverage, processing and revisions [Online]. Marine Management Organisation. Available: https://www.gov.uk/guidance/fishing-activity-and- landings-data-collection-and-processing#page-navigation [Accessed 24/11/2015. MMO. 2017a. Current catch limits: September 2017 [Online]. MMO. Available: https://www.gov.uk/government/publications/current-catch-limits-over-10-metre-non- sector-pool/current-catch-limits-over-10-metre-non-sector-pool-august-2016 [Accessed 06/09/2017. MMO. 2017b. Minimum Conservation Reference Sizes [Online]. MMO. [Accessed 05/09/2017. MOORTER, B., BUNNEFELD, N., PANZACCHI, M., ROLANDSEN, C. M., SOLBERG, E. J. & SÆTHER, B. E. 2013. Understanding scales of movement: animals ride waves and ripples of environmental change. Journal of Animal Ecology, 82, 770-780. MORGAN, M. & TRIPPEL, E. 1996. Skewed sex ratios in spawning shoals of Atlantic cod (Gadus morhua). ICES Journal of Marine Science, 53, 820-826. MORRIS, D. W. 2003. Toward an ecological synthesis: a case for habitat selection. Oecologia, 136, 1-13. MORRISSEY, J. F. & GRUBER, S. H. 1993. Habitat selection by juvenile lemon sharks, Negaprion brevirostris. Environmental Biology of fishes, 38, 311-319. MOTTA, P. J. & HUBER, D. R. 2004. Prey capture behavior and feeding mechanics of elasmobranchs. Biology of sharks and their relatives, 1, 153-209.

233

MOURA, T., FIGUEIREDO, I., FARIAS, I., SERRA-PEREIRA, B., NEVES, A., BORGES, M. D. F. & SERRANO GORDO, L. 2008. Ontogenetic dietary shift and feeding strategy of Raja undulata Lacepède, 1802 (Chondrichthyes: Rajidae) on the Portuguese continental shelf. Scientia Marina, 72, 311-318. MOURIER, J., MILLS, S. & PLANES, S. 2013. Population structure, spatial distribution and life‐ history traits of blacktip reef sharks Carcharhinus melanopterus. Journal of Fish Biology, 82, 979-993. MUCIENTES, G. R., QUEIROZ, N., SOUSA, L. L., TARROSO, P. & SIMS, D. W. 2009. Sexual segregation of pelagic sharks and the potential threat from fisheries. Biology Letters, rsbl. 2008.0761. MURPHY, C. J. & HOWLAND, H. C. 1990. The functional-significance of crescent-shaped pupils and multiple pupillary apertures. Journal of Experimental Zoology, 22-28. MYERS, R. A., BAUM, J. K., SHEPHERD, T. D., POWERS, S. P. & PETERSON, C. H. 2007. Cascading effects of the loss of apex predatory sharks from a coastal ocean. Science, 315, 1846- 1850. MYERS, R. A. & WORM, B. 2005. Extinction, survival or recovery of large predatory fishes. Philosophical Transactions of the Royal Society B: Biological Sciences, 360, 13-20. NATHAN, R., GETZ, W. M., REVILLA, E., HOLYOAK, M., KADMON, R., SALTZ, D. & SMOUSE, P. E. 2008. A movement ecology paradigm for unifying organismal movement research. Proceedings of the National Academy of Sciences, 105, 19052-19059. NAVARRO, J., CARDADOR, L., FERNÁNDEZ, Á. M., BELLIDO, J. M. & COLL, M. 2016. Differences in the relative roles of environment, prey availability and human activity in the spatial distribution of two marine mesopredators living in highly exploited ecosystems. Journal of biogeography, 43, 440-450. . NEAT, F., PINTO, C., BURRETT, I., COWIE, L., TRAVIS, J., THORBURN, J., GIBB, F. & WRIGHT, P. J. 2015. Site fidelity, survival and conservation options for the threatened flapper skate (Dipturus cf. intermedia). Aquatic Conservation: Marine and Freshwater Ecosystems, 25, 6-20. NIETO, A., RALPH, G., COMEROS-RAYNAL, M., HEESSEN, H. & RIJNSDORP, A. 2015. European red list of marine fishes, Publications Office of the European Union. NOTTAGE, A. & PERKINS, E. 1983. Growth and maturation of roker Raja clavata L. in the Solway Firth. Journal of Fish Biology, 23, 43-48. O'SHEA, O. R., THUMS, M., VAN KEULEN, M. & MEEKAN, M. 2012. Bioturbation by stingrays at Ningaloo Reef, Western Australia. Marine and Freshwater Research, 63, 189-197. OFARRELL, M. J. 1974. SEASONAL ACTIVITY PATTERNS OF RODENTS IN A SAGEBRUSH COMMUNITY. Journal of Mammalogy, 55, 809-823. OLASO, I., SÁNCHEZ, F., RODRÍGUEZ-CABELLO, C. & VELASCO, F. 2002. The feeding behaviour of some demersal fish species in response to artificial discarding. Scientia Marina, 66, 301- 311. OLESON, E. M., CALAMBOKIDIS, J., BARLOW, J. & HILDEBRAND, J. A. 2007. Blue whale visual and acoustic encounter rates in the Southern California Bight. Marine Mammal Science, 23, 574-597. OLIN, J. A., HUSSEY, N. E., FRITTS, M., HEUPEL, M. R., SIMPFENDORFER, C. A., POULAKIS, G. R. & FISK, A. T. 2011. Maternal meddling in neonatal sharks: implications for interpreting stable isotopes in young animals. Rapid Communications in Mass Spectrometry, 25, 1008- 1016. ONA, E. & GODØ, O. R. Fish reaction to trawling noise: the significance for trawl sampling. 1990. ICES. ORTH, R. J. 1975. Destruction of eelgrass, Zostera marina, by the cownose ray, Rhinoptera bonasus, in the Chesapeake Bay. Chesapeake Science, 16, 205-208.

234

PADE, N. G., QUEIROZ, N., HUMPHRIES, N. E., WITT, M. J., JONES, C. S., NOBLE, L. R. & SIMS, D. W. 2009. First results from satellite-linked archival tagging of porbeagle shark, Lamna nasus: Area fidelity, wider-scale movements and plasticity in diel depth changes. Journal of Experimental Marine Biology and Ecology, 370, 64-74. PAPASTAMATIOU, Y. P., WATANABE, Y. Y., BRADLEY, D., DEE, L. E., WENG, K., LOWE, C. G. & CASELLE, J. E. 2015. Drivers of Daily Routines in an Ectothermic Marine Predator: Hunt Warm, Rest Warmer? PLoS ONE, 10. PAPASTAMATIOU, Y. P., WETHERBEE, B. M., LOWE, C. G. & CROW, G. L. 2006. Distribution and diet of four species of carcharhinid shark in the Hawaiian Islands: evidence for resource partitioning and competitive exclusion. Marine Ecology Progress Series, 320, 239-251. PARKE, M. 1950. Algal records for the Plymouth region. Journal of the Marine Biological Association of the United Kingdom, 29, 257-261. PARSONS, T. 1992. The removal of marine predators by fisheries and the impact of trophic structure. Marine Pollution Bulletin, 25, 51-53. PATRICK, S. C., BEARHOP, S., GRÉMILLET, D., LESCROËL, A., GRECIAN, W. J., BODEY, T. W., HAMER, K. C., WAKEFIELD, E., LE NUZ, M. & VOTIER, S. C. 2014. Individual differences in searching behaviour and spatial foraging consistency in a central place marine predator. Oikos, 123, 33-40. PAULY, D., CHRISTENSEN, V., DALSGAARD, J., FROESE, R. & TORRES, F. 1998. Fishing down marine food webs. Science, 279, 860-863. PAULY, D., CHRISTENSEN, V., GUÉNETTE, S., PITCHER, T. J., SUMAILA, U. R., WALTERS, C. J., WATSON, R. & ZELLER, D. 2002. Towards sustainability in world fisheries. Nature, 418, 689-695. PEDERSEN, M. W., RIGHTON, D., THYGESEN, U. H., ANDERSEN, K. H. & MADSEN, H. 2008. Geolocation of North Sea cod (Gadus morhua) using hidden Markov models and behavioural switching. Canadian Journal of Fisheries and Aquatic Sciences, 65, 2367-2377. PERRY, A. L., LOW, P. J., ELLIS, J. R. & REYNOLDS, J. D. 2005. Climate change and distribution shifts in marine fishes. Science, 308, 1912-1915. PIKESLEY, S. K., GODLEY, B. J., LATHAM, H., RICHARDSON, P. B., ROBSON, L. M., SOLANDT, J.-L., TRUNDLE, C., WOOD, C. & WITT, M. J. 2016. Pink sea fans (Eunicella verrucosa) as indicators of the spatial efficacy of Marine Protected Areas in southwest UK coastal waters. Marine Policy, 64, 38-45. PIKITCH, E. K., CHAPMAN, D. D., BABCOCK, E. A. & SHIVJI, M. S. 2005. Habitat use and demographic population structure of elasmobranchs at a Caribbean atoll (Glover’s Reef, Belize). Marine Ecology Progress Series, 302, 187-197. PINGREE, R. D. & GRIFFITHS, D. K. 1978. Tidal fronts on the shelf seas around the British Isles. Journal of Geophysical Research: Oceans, 83, 4615-4622. PINNEGAR, J. K. 2014. DAPSTOM - An Integrated Database & Portal for Fish Stomach Records. In: CENTRE FOR ENVIRONMENT, F. A. S., LOWESTOFT, UK. (ed.). PITA, C., THEODOSSIOU, I. & PIERCE, G. J. 2013. The perceptions of Scottish inshore fishers about marine protected areas. Marine Policy, 37, 254-263. PLATELL, M., POTTER, I. & CLARKE, K. 1998. Resource partitioning by four species of elasmobranchs (Batoidea: Urolophidae) in coastal waters of temperate Australia. Marine Biology, 131, 719-734. PUNT, A. E., WALKER, T. I., TAYLOR, B. L. & PRIBAC, F. 2000. Standardization of catch and effort data in a spatially-structured shark fishery. Fisheries Research, 45, 129-145. PYKE, G. H., PULLIAM, H. R. & CHARNOV, E. L. 1977. Optimal foraging: a selective review of theory and tests. The quarterly review of biology, 52, 137-154. QASEM, L., CARDEW, A., WILSON, A., GRIFFITHS, I., HALSEY, L. G., SHEPARD, E. L., GLEISS, A. C. & WILSON, R. 2012. Tri-axial dynamic acceleration as a proxy for animal energy expenditure; should we be summing values or calculating the vector? PLoS One, 7, e31187.

235

QUAECK, K. 2017. Stable isotope analysis of fish eye lenses:reconstructionof ontogenetic trends in spatial and trophic ecology of elasmobranchs and deep water teleosts. Doctor of Philosophy, University of Southampton. QUEIROGA, H. & BLANTON, J. 2005. Interactions between behaviour and physical forcing in the control of horizontal transport of decapod crustacean larvae. Advances in marine biology, 47, 107-214. QUEIROLO, D., GAETE, E., MONTENEGRO, I., SORIGUER, M. & ERZINI, K. 2012. Behaviour of fish by‐catch in the mouth of a crustacean trawl. Journal of Fish Biology, 80, 2517-2527. QUEIROZ, N., HUMPHRIES, N. E., MUCIENTES, G., HAMMERSCHLAG, N., LIMA, F. P., SCALES, K. L., MILLER, P. I., SOUSA, L. L., SEABRA, R. & SIMS, D. W. 2016. Ocean-wide tracking of pelagic sharks reveals extent of overlap with longline fishing hotspots. Proceedings of the National Academy of Sciences, 113, 1582-1587. QUEIROZ, N., HUMPHRIES, N. E., NOBLE, L. R., SANTOS, A. M. & SIMS, D. W. 2010. Short-term movements and diving behaviour of satellite-tracked blue sharks Prionace glauca in the northeastern Atlantic Ocean. Marine Ecology Progress Series, 406, 265-279. QUEIROZ, N., HUMPHRIES, N. E., NOBLE, L. R., SANTOS, A. N. M. & SIMS, D. W. 2012. Spatial Dynamics and Expanded Vertical Niche of Blue Sharks in Oceanographic Fronts Reveal Habitat Targets for Conservation. PLoS ONE, 7, e32374. QUINN, T. P. & BRODEUR, R. D. 1991. Intra-specific variations in the movement patterns of marine animals. American zoologist, 231-241. R DEVELOPMENT CORE TEAM 2015. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. REILLY, S. B., BANNISTER, J. L., BEST, P. B., BROWN, M., BROWNELL, J. R. L., BUTTERWORTH, D. S., CLAPHAM, P. J., COOKE, J., DONOVAN, G. P., URBÁN, J. & ZERBINI, A. N. 2008. Balaenoptera musculus [Online]. The IUCN Red List of Threatened Species 2008: e.T2477A9447146. . Available: http://dx.doi.org/10.2305/IUCN.UK.2008.RLTS.T2477A9447146.en. [Accessed 17 March 2018. RIJNSDORP, A. D., BUYS, A. M., STORBECK, F. & VISSER, E. G. 1998. Micro-scale distribution of beam trawl effort in the southern North Sea between 1993 and 1996 in relation to the trawling frequency of the sea bed and the impact on benthic organisms. Ices Journal of Marine Science, 55, 403-419. RIJNSDORP, A. D., VAN LEEUWEN, P. I., DAAN, N. & HEESSEN, H. J. 1996. Changes in abundance of demersal fish species in the North Sea between 1906–1909 and 1990–1995. ICES Journal of Marine Science: Journal du Conseil, 53, 1054-1062. RINGELBERG, J. & VAN GOOL, E. 2003. On the combined analysis of proximate and ultimate aspects in diel vertical migration (DVM) research. Hydrobiologia, 491, 85-90. RODRÍGUEZ, M. A. 2002. Restricted movement in stream fish: the paradigm is incomplete, not lost. Ecology, 83, 1-13. ROSENZWEIG, M. L. 1981. A theory of habitat selection. Ecology, 62, 327-335. ROSS, S. T. 1986. Resource partitioning in fish assemblages: a review of field studies. Copeia, 352- 388. ROUSSET, J. 1990. Population structure of thornback rays Raja clavata and their movements in the Bay of Douarnenez. Journal of the Marine Biological Association of the United Kingdom, 70, 261-268. ŠANTIĆ, M., RAĐA, B. & PALLAORO, A. 2012. Diet and feeding strategy of thornback ray Raja clavata. Journal of Fish Biology, 81, 1070-1084. SCHAEFER, K. M. & FULLER, D. W. 2002. Movements, behavior, and habitat selection of bigeye tuna (Thunnus obesus) in the eastern equatorial Pacific, ascertained through archival tags. Fishery Bulletin, 100, 765-788.

236

SCHAEFER, K. M. & FULLER, D. W. 2010. Vertical movements, behavior, and habitat of bigeye tuna (Thunnus obesus) in the equatorial eastern Pacific Ocean, ascertained from archival tag data. Marine Biology, 157, 2625-2642. SCHARF, F. S., JUANES, F. & ROUNTREE, R. A. 2000. Predator size - prey size relationships of marine fish predators: interspecific variation and effects of ontogeny and body size on trophic-niche breadth. Marine Ecology Progress Series, 208, 229-248. SCHELL, D., SAUPE, S. & HAUBENSTOCK, N. 1989. Bowhead whale (Balaena mysticetus) growth and feeding as estimated by δ 13 C techniques. Marine Biology, 103, 433-443. SCHLAFF, A. M., HEUPEL, M. R. & SIMPFENDORFER, C. A. 2014. Influence of environmental factors on shark and ray movement, behaviour and habitat use: a review. Reviews in Fish Biology and Fisheries, 24, 1089-1103. SCHOENER, T. W. 1974. The compression hypothesis and temporal resource partitioning. Proceedings of the National Academy of Sciences, 71, 4169-4172. SCHOENER, T. W. 1983. Field Experiments on Interspecific Competition. The American Naturalist, 122, 240-285. SERRA-PEREIRA, B., ERZINI, K., MAIA, C. & FIGUEIREDO, I. 2014. Identification of Potential Essential Fish Habitats for Skates Based on Fishers' Knowledge. Environmental Management, 53, 985-998. SHEPARD, E. L., WILSON, R. P., HALSEY, L. G., QUINTANA, F., LAICH, A. G., GLEISS, A. C., LIEBSCH, N., MYERS, A. E. & NORMAN, B. 2008. Derivation of body motion via appropriate smoothing of acceleration data. Aquatic Biology, 4, 235-241. SHEPARD, E. L. C., AHMED, M. Z., SOUTHALL, E. J., WITT, M. J., METCALFE, J. D. & SIMS, D. W. 2006. Diel and tidal rhythms in diving behaviour of pelagic sharks identified by signal processing of archival tagging data. Marine Ecology-Progress Series, 328, 205-213. SHEPHARD, S., GERRITSEN, H., KAISER, M. J. & REID, D. G. 2012. Spatial heterogeneity in fishing creates de facto refugia for endangered Celtic Sea elasmobranchs. PLoS One, 7, e49307. SHIOMI, K., SATO, K., MITAMURA, H., ARAI, N., NAITO, Y. & PONGANIS, P. J. 2008. Effect of ocean current on the dead-reckoning estimation of 3-D dive paths of emperor penguins. Aquatic Biology, 3, 265-270. SILVA, J., ELLIS, J. & CATCHPOLE, T. 2012. Species composition of skates (Rajidae) in commercial fisheries around the British Isles and their discarding patterns. Journal of Fish Biology, 80, 1678-1703. SILVESTRE, G. T., GARCES, L. R., STOBUTZKI, I., AHMED, M., VALMONTE-SANTOS, R., LUNA, C. Z. & ZHOU, W. South and South-East Asian coastal fisheries: their status and directions for improved management: conference synopsis and recommendations. Assessment, management and future directions for coastal fisheries in Asian countries. Worldfish Center Conference Proceedings, 2003. 1-40. SIMPFENDORFER, C. A., FREITAS, G. G., WILEY, T. R. & HEUPEL, M. R. 2005. Distribution and habitat partitioning of immature bull sharks (Carcharhinus leucas) in a southwest Florida estuary. Estuaries, 28, 78-85. SIMPSON, S. D., JENNINGS, S., JOHNSON, M. P., BLANCHARD, J. L., SCHON, P. J., SIMS, D. W. & GENNER, M. J. 2011. Continental Shelf-Wide Response of a Fish Assemblage to Rapid Warming of the Sea. Current Biology, 21, 1565-1570. SIMPSON, S. J. & SIMS, D. W. 2016. Are critically endangered fish back on the menu? Analysis of UK fisheries data suggest post-ban landings of prohibited skates in European waters. Marine Policy, 69, 42-51. SIMS, D. 2003. Tractable models for testing theories about natural strategies: foraging behaviour and habitat selection of free‐ranging sharks. Journal of Fish Biology, 63, 53-73. SIMS, D., NASH, J. & MORRITT, D. 2001. Movements and activity of male and female dogfish in a tidal sea lough: alternative behavioural strategies and apparent sexual segregation. Marine Biology, 139, 1165-1175.

237

SIMS, D. W. 2005. Differences in habitat selection and reproductive strategies of male and female sharks. Sexual segregation in vertebrates: Ecology of the two sexes, 127-147. SIMS, D. W. & QUAYLE, V. A. 1998. Selective foraging behaviour of basking sharks on zooplankton in a small-scale front. Nature, 393, 460. SIMS, D. W., QUEIROZ, N., HUMPHRIES, N. E., LIMA, F. P. & HAYS, G. C. 2009. Long-term GPS tracking of ocean sunfish Mola mola offers a new direction in fish monitoring. PLoS One, 4, e7351. SIMS, D. W. & SIMPSON, S. J. 2015. Fisheries: Better policing for fishy catch data. Nature, 520, 623-623. SIMS, D. W., SOUTHALL, E. J., TARLING, G. A. & METCALFE, J. D. 2005. Habitat-specific normal and reverse diel vertical migration in the plankton-feeding basking shark. Journal of Animal Ecology, 74, 755-761. SIMS, D. W. & SOUTHWARD, A. J. 2006. Dwindling fish numbers already of concern in 1883. Nature, 439, 660-660. SIMS, D. W., WEARMOUTH, V. J., SOUTHALL, E. J., HILL, J. M., MOORE, P., RAWLINSON, K., HUTCHINSON, N., BUDD, G. C., RIGHTON, D., METCALFE, J., NASH, J. P. & MORRITT, D. 2006. Hunt warm, rest cool: bioenergetic strategy underlying diel vertical migration of a benthic shark. Journal of Animal Ecology, 75, 176-190. SMALE, M. & COWLEY, P. 1992. The feeding ecology of skates (Batoidea: Rajidae) off the Cape south coast, South Africa. South African Journal of Marine Science, 12, 823-834. SMELSER, G. K. 1965. Embryology and morphology of the lens. Investigative Ophthalmology & Visual Science, 4, 398-410. SMITH, H. D. 2000. The industrialisation of the world ocean. Ocean & Coastal Management, 43, 11-28. SMITH, J. W. & MERRINER, J. V. 1985. Food habits and feeding behavior of the cownose ray, Rhinoptera bonasus, in lower Chesapeake Bay. Estuaries, 8, 305-310. SOLMUNDSSON, J., KARLSSON, H. & PALSSON, J. 2003. Sexual differences in spawning behaviour and catchability of plaice (Pleuronectes platessa) west of Iceland. Fisheries Research, 61, 57-71. SPEED, C. W., FIELD, I. C., MEEKAN, M. G. & BRADSHAW, C. J. 2010. Complexities of coastal shark movements and their implications for management. Marine Ecology Progress Series, 408, 275-293. SPEED, C. W., MEEKAN, M. G., FIELD, I. C., MCMAHON, C. R., STEVENS, J. D., MCGREGOR, F., HUVENEERS, C., BERGER, Y. & BRADSHAW, C. J. 2011. Spatial and temporal movement patterns of a multi-species coastal reef shark aggregation. Marine Ecology Progress Series, 429, 261-275. STEVEN, G. 1932a. Rays and skates of Devon and Cornwall. II. A study of the fishery; with notes on the occurrence, migrations and habits of the species. Journal of the Marine Biological Association of the United Kingdom (New Series), 18, 1-33. STEVEN, G. 1933. Rays and skates of Devon and Cornwall. III. The proportions of the sexes in nature and in commercial landings, and their significance to the fishery. Journal of the Marine Biological Association of the United Kingdom, 18, 611-626. STEVEN, G. 1936. Migrations and growth of the thornback ray (Raia clavata L.). Journal of the Marine Biological Association of the United Kingdom (New Series), 20, 605-614. STEVEN, G. 1947. The British raiidae. Science Progress (1933-), 35, 220-236. STEVENS, T. & CONNOLLY, R. M. 2004. Testing the utility of abiotic surrogates for marine habitat mapping at scales relevant to management. Biological Conservation, 119, 351-362. SUTHERLAND, W. J. 1983. Aggregation and theideal free'distribution. The Journal of Animal Ecology, 821-828. SUTTON, T. T. 2013. Vertical ecology of the pelagic ocean: classical patterns and new perspectives. Journal of Fish Biology, 83, 1508-1527.

238

SUTTON, T. T. & HOPKINS, T. L. 1996. Trophic ecology of the stomiid (Pisces: Stomiidae) fish assemblage of the eastern Gulf of Mexico: Strategies, selectivity and impact of a top mesopelagic predator group. Marine Biology, 127, 179-192. SWARTZMAN, G., HUANG, C. & KALUZNY, S. 1992. Spatial analysis of Bering Sea groundfish survey data using generalized additive models. Canadian Journal of Fisheries and Aquatic Sciences, 49, 1366-1378. TANAKA, H., AOKI, I. & OHSHIMO, S. 2006. Feeding habits and gill raker morphology of three planktivorous pelagic fish species off the coast of northern and western Kyushu in summer. Journal of Fish Biology, 68, 1041-1061. TEMPLEMAN, W. 1984. Migrations of thorny skate, Raja radiata, tagged in the Newfoundland area. J. Northw. Atl. Fish. Sci, 5, 55-64. THURSTAN, R. H., BROCKINGTON, S. & ROBERTS, C. M. 2010. The effects of 118 years of industrial fishing on UK bottom trawl fisheries. Nature Communications, 1, 15. TIESZEN, L. L., BOUTTON, T. W., TESDAHL, K. G. & SLADE, N. A. 1983. Fractionation and turnover of stable carbon isotopes in animal tissues: implications for δ13C analysis of diet. Oecologia, 57, 32-37. TOFT, C. A. 1985. Resource partitioning in amphibians and reptiles. Copeia, 1-21. TRUEMAN, C., MACKENZIE, K. & PALMER, M. 2012. Identifying migrations in marine fishes through stable‐isotope analysis. Journal of Fish Biology, 81, 826-847. ULRICH, C., PASCOE, S., SPARRE, P. J., WILDE, J.-W. D. & MARCHAL, P. 2002. Influence of trends in fishing power on bioeconomics in the North Sea flatfish fishery regulated by catches or by effort quotas. Canadian Journal of Fisheries and Aquatic Sciences, 59, 829-843. VAN DER KOOIJ, J., SCOTT, B. E. & MACKINSON, S. 2008. The effects of environmental factors on daytime sandeel distribution and abundance on the Dogger Bank. Journal of Sea Research, 60, 201-209. VAN OOSTENBRUGGE, H. J., POWELL, J. P., SMIT, J. P., POOS, J.-J., KRAAK, S. B. & BUISMAN, E. F. 2008. Linking catchability and fisher behaviour under effort management. Aquatic Living Resources, 21, 265-273. VEMCO. 2013. VR4-UWM Datasheet (5022-03) [Online]. Available: https://vemco.com/wp- content/uploads/2014/01/vr4-uwm.pdf [Accessed 27/11/2017. VERNES, K. 2003. Fine‐scale habitat preferences and habitat partitioning by three mycophagous mammals in tropical wet sclerophyll forest, north‐eastern Australia. Austral Ecology, 28, 471-479. VERNES, K. & POPE, L. C. 2001. Stability of nest range, home range and movement of the northern bettong (Bettongia tropica) following moderate-intensity fire in a tropical woodland, north-eastern Queensland. Wildlife Research, 28, 141-150. VØLLESTAD, L. A., SERBEZOV, D., BASS, A., BERNATCHEZ, L., OLSEN, E. M. & TAUGBØL, A. 2012. Small-scale dispersal and population structure in stream-living brown trout (Salmo trutta) inferred by mark–recapture, pedigree reconstruction, and population genetics. Canadian Journal of Fisheries and Aquatic Sciences, 69, 1513-1524. WALKER, P. & HEESSEN, H. 1996. Long-term changes in ray populations in the North Sea. ICES Journal of Marine Science: Journal du Conseil, 53, 1085-1093. WALKER, P. & HISLOP, J. 1998. Sensitive skates or resilient rays? Spatial and temporal shifts in ray species composition in the central and north-western North Sea between 1930 and the present day. ICES Journal of Marine Science: Journal du Conseil, 55, 392-402. WALKER, P., HOWLETT, G. & MILLNER, R. 1997. Distribution, movement and stock structure of three ray species in the North Sea and eastern English Channel. ICES Journal of Marine Science: Journal du Conseil, 54, 797-808. WALLACE, A. A., HOLLANDER, D. J. & PEEBLES, E. B. 2014. Stable isotopes in fish eye lenses as potential recorders of trophic and geographic history. PLoS One, 9, e108935.

239

WALLACE, A. R. 2011. The geographical distribution of animals: with a study of the relations of living and extinct faunas as elucidating the past changes of the earth's surface, Cambridge University Press. WALLI, A., TEO, S. L. H., BOUSTANY, A., FARWELL, C. J., WILLIAMS, T., DEWAR, H., PRINCE, E. & BLOCK, B. A. 2009. Seasonal Movements, Aggregations and Diving Behavior of Atlantic Bluefin Tuna (Thunnus thynnus) Revealed with Archival Tags. PLoS ONE, 4. WALSH, S. J. 1992. Size-Dependent Selection at the Footgear of a Groundfish Survey Trawl. North American Journal of Fisheries Management, 12, 625-633. WARD, T., VANDERKLIFT, M., NICHOLLS, A. & KENCHINGTON, R. 1999. Selecting marine reserves using habitats and species assemblages as surrogates for biological diversity. Ecological Applications, 9, 691-698. WATANABE, N., SAKANOUE, S., KAWAMURA, K. & KOZAKAI, T. 2008. Development of an automatic classification system for eating, ruminating and resting behavior of cattle using an accelerometer. Grassland science, 54, 231-237. WATANABE, Y. Y. & TAKAHASHI, A. 2013. Linking animal-borne video to accelerometers reveals prey capture variability. Proceedings of the National Academy of Sciences, 110, 2199- 2204. WEARMOUTH, V. J., MCHUGH, M. J., HUMPHRIES, N. E., NAEGELEN, A., AHMED, M. Z., SOUTHALL, E. J., REYNOLDS, A. M. & SIMS, D. W. 2014a. Scaling laws of ambush predator ‘waiting’ behaviour are tuned to a common ecology. Proceedings of the Royal Society of London, Series B: Biological Sciences, 281. WEARMOUTH, V. J. & SIMS, D. W. 2008. Sexual segregation in marine fish, reptiles, birds and mammals: behaviour patterns, mechanisms and conservation implications. Advances in marine biology, 54, 107-170. WEARMOUTH, V. J. & SIMS, D. W. 2009a. Movement and behaviour patterns of the critically endangered common skate Dipturus batis revealed by electronic tagging. Journal of Experimental Marine Biology and Ecology, 380, 77-87. WEARMOUTH, V. J., SOUTHALL, E. J., MORRITT, D., THOMPSON, R. C., CUTHILL, I. C., PARTRIDGE, J. C. & SIMS, D. W. 2012. Year-round sexual harassment as a behavioral mediator of vertebrate population dynamics. Ecological Monographs, 82, 351-366. WEIMERSKIRCH, H., MOUGEY, T. & HINDERMEYER, X. 1997. Foraging and provisioning strategies of black-browed albatrosses in relation to the requirements of the chick: natural variation and experimental study. Behavioral Ecology, 8, 635-643. WENSVEEN, P. J., THOMAS, L. & MILLER, P. J. 2015. A path reconstruction method integrating dead-reckoning and position fixes applied to humpback whales. Movement ecology, 3, 31. WHEELER, A. 1969a. The Fishes of the British Isles and North-West Europe, London, Macmillan.

WHITE, W., PLATELL, M. & POTTER, I. 2004. Comparisons between the diets of four abundant species of elasmobranchs in a subtropical embayment: implications for resource partitioning. Marine Biology, 144, 439-448. WHITE, W. & POTTER, I. 2004. Habitat partitioning among four elasmobranch species in nearshore, shallow waters of a subtropical embayment in Western Australia. Marine Biology, 145, 1023-1032. WHITEHEAD, P., BAUCHOT, M., HUREAU, J., NIELSEN, J. & TORTONESE, E. 1984. Fishes of the north-eastern Atlantic and the Mediterranean. v. 1. WHITTAMORE, J. & MCCARTHY, I. 2005. The population biology of the thornback ray, Raja clavata in Caernarfon Bay, north Wales. Journal of the Marine Biological Association of the United Kingdom, 85, 1089-1094. WIENS, J. A. 1989. Spatial scaling in ecology. Functional ecology, 3, 385-397.

240

WILSON, R. P., PUETZ, K., BOST, C. A., CULIK, B. M., BANNASCH, R., REINS, T. & ADELUNG, D. 1993. Diel dive depth in penguins in relation to diel vertical migration of prey - whose dinner by candlelight. Marine ecology progress series, 94, 101-104. WILSON, R. P., WILSON, M.-P. T., LINK, R., MEMPEL, H. & ADAMS, N. J. 1991. Determination of movements of African penguins Spheniscus demersus using a compass system: dead reckoning may be an alternative to telemetry. Journal of Experimental Biology, 157, 557- 564. WILSON, S. G., TAYLOR, J. G. & PEARCE, A. F. 2001. The seasonal aggregation of whale sharks at Ningaloo Reef, Western Australia: currents, migrations and the El Nino/Southern Oscillation. Environmental Biology of fishes, 61, 1-11. WINGER, P. D., WALSH, S. J., HE, P. & BROWN, J. A. 2004. Simulating trawl herding in flatfish: the role of fish length in behaviour and swimming characteristics. ICES Journal of Marine Science: Journal du Conseil, 61, 1179-1185. WORM, B. & BRANCH, T. A. 2012. The future of fish. Trends in ecology & evolution, 27, 594-599. WORM, B., HILBORN, R., BAUM, J. K., BRANCH, T. A., COLLIE, J. S., COSTELLO, C., FOGARTY, M. J., FULTON, E. A., HUTCHINGS, J. A. & JENNINGS, S. 2009. Rebuilding global fisheries. Science, 325, 578-585. WORM, B. & MYERS, R. A. 2003. Meta‐analysis of cod–shrimp interactions reveals top‐down control in oceanic food webs. Ecology, 84, 162-173. WUERINGER, B. E. 2012. Electroreception in Elasmobranchs: Sawfish as a Case Study. Brain Behavior and Evolution, 80, 97-107. WUERINGER, B. E., JNR, L. S., KAJIURA, S. M., TIBBETTS, I. R., HART, N. S. & COLLIN, S. P. 2012. Electric Field Detection in Sawfish and Shovelnose Rays. Plos One, 7. YAMAGIWA, J. & BASABOSE, A. K. 2009. Fallback foods and dietary partitioning among Pan and Gorilla. American Journal of Physical Anthropology, 140, 739-750. YODA, K., NAITO, Y., SATO, K., TAKAHASHI, A., NISHIKAWA, J., ROPERT-COUDERT, Y., KURITA, M. & LE MAHO, Y. 2001. A new technique for monitoring the behaviour of free-ranging Adelie penguins. Journal of Experimental Biology, 204, 685-690. ZOUHIRI, S. & DAUVIN, J. C. 1996. Diel changes of the Benthic Boundary Layer macrofauna over coarse sand sediment in the western English Channel. Oceanologica Acta, 19, 141-153.

241