A knowledge platform to inform on the effects of trawling on benthic communities

Accepted Manuscript A knowledge platform to inform on the effects of trawling on benthic communities Alba Muntadas, Michel Lample, Montserrat Demestre, Johanna Ballé-Béganton, Silvia de Juan, Francesc Maynou, Denis Bailly PII:

S0272-7714(17)30003-3

DOI:

10.1016/j.ecss.2017.01.001

Reference:

YECSS 5359

To appear in:

Estuarine, Coastal and Shelf Science

Received Date: 16 March 2015 Revised Date:

5 September 2016

Accepted Date: 3 January 2017

Please cite this article as: Muntadas, A., Lample, M., Demestre, M., Ballé-Béganton, J., de Juan, S., Maynou, F., Bailly, D., A knowledge platform to inform on the effects of trawling on benthic communities, Estuarine, Coastal and Shelf Science (2017), doi: 10.1016/j.ecss.2017.01.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

1

A knowledge platform to inform on the effects of trawling on benthic

2

communities Alba Muntadas*1, Michel Lample2, Montserrat Demestre1, Johanna Ballé-Béganton2, Silvia de Juan3, Francesc Maynou1 and Denis Bailly2

5 6

1

RI PT

3 4

Institut de Ciències del Mar. Passeig Marítim de la Barceloneta 37-49.08003, Barcelona 2

7 8 9 10 11 12

Université de Brest, UEB, UMR AMURE, 12 rue du Kergoat, CS 93837, 29238 Brest Cedex 3, France

13 14 15

*Corresponding autor: Alba Muntadas. Institut de Ciències del Mar. Passeig Marítim de la Barceloneta 37-49. 08003 Barcelona. [email protected]. Tel: +34 93 230 95 00-FAX: +34 93 230 95 55

3

16 GRAPHICAL ABSTRACT

18 19

AC C

EP

TE D

17

M AN U

SC

Center for Marine Conservation, Estación Costera de Investigaciones MarinasFacultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Casilla 114D, Santiago, Chile.

ABSTRACT

20

For a successful implementation of an Ecosystem Approach to Fisheries (EAF)

21

management, it is necessary that all stakeholders involved in fisheries management

22

are aware of the implications of fishing impacts on ecosystems and agree with the

23

adopted measures to mitigate these impacts. In this context, there is a need for tools to 1

ACCEPTED MANUSCRIPT

share knowledge on the ecosystem effects of fisheries among these stakeholders.

25

When managing bottom trawl fisheries under an EAF framework, one of the main

26

concerns is the direct and indirect consequences of trawling impacts on benthic

27

ecosystems. We developed a platform using the ExtendSim® software with a user-

28

friendly interface that combines a simulation model based on existing knowledge, data

29

collection and representation of predicted trawling impacts on the seabed. The platform

30

aims to be a deliberation support tool for fisheries’ stakeholders and, simultaneously,

31

raise public awareness of the need for good benthic community knowledge to

32

appropriately inform EAF management plans. The simulation procedure assumes that

33

trawling affects benthic communities with an intensity that depends on the level of

34

fishing effort exerted on benthic communities and on the habitat characteristics (i.e.

35

sediment grain size). Data to build the simulation comes from epifaunal samples from

36

18 study sites located in Mediterranean continental shelves subjected to different levels

37

of fishing effort. In this work, we present the simulation outputs of a 50% fishing effort

38

increase (and decrease) in four of the study sites which cover different habitats and

39

different levels of fishing effort. We discuss the platform strengths and weaknesses and

40

potential future developments.

41

KEYWORDS: Benthos, Biological Traits Analysis, trawling, Ecosystem Approach to

42

Fisheries, simulation platform, fishing effort, Mediterranean

43

1. INTRODUCTION

46

AC C

EP

TE D

M AN U

SC

RI PT

24

47

an Ecosystem Approach to Fisheries (EAF) management, which takes into account the

48

consequences of fisheries on target and non-target species and habitats (Carter, 2013;

49

Lucthman et al. 2008; Garcia & Cochrane, 2005). This change of management

44 45

During the past decades, a societal concern for the environmental consequences of

trawling has arisen, particularly bottom trawling. This concern has led to a change in fisheries management policies from a target resource-oriented management towards

2

ACCEPTED MANUSCRIPT

paradigm does not only entail a wider environmental scope but also a higher

51

involvement of stakeholders in the process. For example, while only scientists and

52

politicians are involved in the decision process of the species´ stock-based

53

management framework, managers, fishermen and NGOs are also involved in the EAF

54

management process (Carter, 2013). Consequently, EAF needs to deal with a more

55

complex multi-actor scenario.

56

The different stakeholders that might play a role in EAF management are of diverse

57

backgrounds and interests, which may lead to inevitably trade-offs in decision making

58

(Carter, 2013; Daw and Gray, 2005; Garcia and Cochrane, 2005; Toonen and Mol,

59

2013). For instance, while NGOs claim that no-take zones must be included in spatial

60

management plans, fishermen are obviously reluctant to do so (Toonen and Mol, 2013).

61

Moreover, scientific advice for management normally involves large uncertainties which

62

can lead to lack of credibility, and therefore absence of support by other stakeholders

63

(Daw and Gray, 2005). However, for management to be successful, it is imperative that

64

all the stakeholders involved in the process are in common ground. Hence, appropriate

65

tools to communicate scientific knowledge supporting decision making are needed for a

66

successful EAF, and this is the main purpose of the knowledge platform presented in

67

this work.

68

The scientific community has already developed different models that can simulate

69

management strategies under different EAF scenarios, and that could advice policy

71 72

SC

M AN U

TE D

EP

AC C

70

RI PT

50

makers, e.g., end-to-end models such as Ecopath with Ecosim (EwE) (Christensen and Walkers, 2004; Corrales et al., 2015), Object-oriented Simulator of Marine ecOSystem Exploitation (OSMOSE) (Travers et al., 2010), Atlantis (Fulton et al., 2015) or European

73

Regional Seas Ecosystem Model (ERSEM) (Allen and Clarke, 2007; Petihakis et al.,

74

2007). However, the seafloor dimension, including demersal species (e.g. species

75

targeted by bottom trawling) and their relation with non-target species and habitats, is

76

poorly developed in these models (Rose et al., 2010). 3

The most acute impacts caused by bottom trawl fisheries are those exerted on benthic

78

organisms and habitats (Thrush and Dayton, 2002). These changes in habitat structure

79

may also affect target species’ populations, as demersal fish need particular habitat

80

characteristics to fulfil their life cycles (Auster and Langton, 1998). Benthic ecosystem

81

models are hence needed to evaluate trawl impacts under different management

82

scenarios, and some examples can be found in the literature (Dichmont et al., 2008;

83

Duplisea et al., 2002; Fujioka, 2006; Lundquist et al., 2010; Queirós et al., 2006).

84

However, despite these benthic ecosystem models are less complex than end-to-end

85

models, the outputs are strongly science-oriented and still have a number of

86

parameters that might be difficult to understand by the non-scientific community.

87

Therefore, efforts should be driven towards building deliberation support tools to

88

disseminate scientific information among the stakeholders In this context, a database

89

that provides information for both scientific and management purposes about benthic

90

species and habitat sensitivities is already available at www.marlin.ac.uk, which also

91

reports trawling effects on benthic communities (Hiscock and Tyler-Walters, 2006).

92

However, this website does not allow users to run scenario modelling.

93

With the aim of providing a tool that illustrates trawling impacts on the seafloor in a

94

visual and understandable way, we introduce an interactive knowledge platform.

95

Additionally, the platform allows the user to test the potential effects of an effort change

96

on benthic community structure. This way of sharing knowledge and scenario

97

100 101

Mongruel et al., 2013; Tomlinson et al., this issue). The objective of the knowledge

102

platform we present here is twofold:

98 99

EP

TE D

M AN U

SC

RI PT

77

AC C

ACCEPTED MANUSCRIPT

simulation among stakeholders has proved to be a good approach to deal with other complex socio-ecological systems, such as freshwater management, microbiological contamination in coastal zones or the interaction of jellyfish with fisheries and tourism (Ballé-Béganton et al., 2010; Ballé-Béganton et al., 2012; Mongruel et al., 2011;

4

a) to compile knowledge about changes on benthic community structure caused by

104

bottom trawling in a dynamic and visual way.

105

b) to offer a simulation tool that allows the user to test how a change in fishing effort

106

would potentially affect the benthic community in a fishing ground.

107

By achieving these objectives, we also aim at raising society, and particularly decision

108

makers’, awareness about the importance of good scientific knowledge on benthic

109

ecosystems to improve the management of trawl fishing activities.

110

2. MATERIALS AND METHODS

111

The knowledge platform has been developed using the ExtendSim® software

112

(ExtendSim 9).

113

2.1. Study areas

114

To build the platform we used an epifaunal data set collected from 7 areas in the

115

Mediterranean Sea: three in the Catalan coast (in the north-western Mediterranean -

116

CAT): Cap de Creus (CC), Medes (M) and Ebre Delta (D); one in Cabo de Palos (in the

117

south-western Mediterranean, CP); one in the Ligurian Sea coast (LC); one in the

118

Adriatic Sea coast (AC); one in the Ionian Sea coast (IC) (Fig. 1). CC, CP and IC areas

119

were further divided in 3 different sites that were subjected to different levels of fishing

120

effort (Low or no effort, Medium and High, hereafter CCL, CCM and CCH for Cap de

121

124 125

Area. In total, we used data from 18 sites, each comprising between 2 and 4 km2.

126

The LC, IC and D areas are characterised by muddy sediments (99.5%, 96.6% and

127

93.3% of mud content respectively), while CC and AC had sandy-mud sediments (40.3%

122 123

EP

TE D

M AN U

SC

RI PT

103

AC C

ACCEPTED MANUSCRIPT

Creus, CPL, CPM and CPH for Cabo de Palos and ICL, ICM and ICH for Ionian sea

coast). The LC area was divided in two sites with low and medium effort (hereafter LCL and LCM) (Fig. 5). For further information on the definition of fishing effort levels see Demestre et al. (2015). Medes and CCL sites were within a no-take Marine Protected

5

ACCEPTED MANUSCRIPT

and 63.8% of sand content respectively). The CP area was characterized by maërl

129

beds protruding within sandy-mud bottoms (de Juan et al., 2013). The M area was

130

characterised by coastal detritic mud.

131

2.2. Fishing effort estimation

132

Two different approaches aimed at quantifying the fishing effort at each site were

133

available in our study areas: 1) fishermen interviews along with information from

134

fisheries associations, and 2) Side-scan sonar surveys (SSS) (Table 1). The former

135

provides first-hand information about fleet movement in the fishing grounds, which

136

allowed us to estimate fishing effort as Gross Tonnage*days at sea/month. The SSS

137

approach allowed the estimation of trawl marks density on the seafloor, considering

138

their total length per unit of surface area within each site (trawl tracks length/km2) (see

139

Demestre et al. 2015 for further details).

140

2.3. Data collection

141

All samples were collected between spring and early summer in different years: June-

142

July 2003 for D and AC, May 2007 for M and June 2009 for the other areas. Epifauna

143

was sampled with an experimental dredge with a 2 m iron-frame aperture, and a 1 cm

144

cod-end. Six samples were randomly collected at each site, the dredge towed for 15

145

minutes with a constant speed of 2.3 knots and a scanmar device was attached to the

146

dredge to ensure continuous contact with the seabed. Only 3 replicates were collected

148 149

SC

M AN U

TE D

EP

AC C

147

RI PT

128

in CCL site and 2 in ICL due to bad weather conditions. Epifaunal organisms were generally identified to species level, and the number of individuals for each species was recorded and standardized to 1000 m2.

150

2.4. Biological Traits Analysis

151

In order to represent the benthic community in a simple and understandable way, we

152

chose a Biological Traits Analysis (BTA) approach instead of a species approach, i.e.,

153

species’ morphological, behavioural and feeding characteristics were used instead of 6

the species identity. The BTA was preferred over the species approach due to several

155

advantages of the former. First, species composition is highly influenced by local

156

environmental conditions, which can mask the consequences of fishing impact. As

157

different species may share the same traits, the trait approach allows overcoming

158

regional differences in species composition. For instance, Bremner et al. (2003) found

159

a geographical gradient in species´ composition along the English channel due to tidal

160

action, sand transport and temperature gradient. However, this gradient disappeared

161

with the BTA approach that highlighted functionality patterns that might be related with

162

anthropogenic activities. Additionally, this approach proved successful in identifying

163

fishing effects on benthic communities (Bremner, 2008; de Juan et al., 2007; Tillin et al.,

164

2006), assess the Good Environmental Status of benthic communities (de Juan and

165

Demestre, 2012), and can be used as an indicator of ecosystem functioning (Muntadas

166

et al. 2015; Muntadas et al. 2016). Moreover, this approach reduces the number of

167

variables and thus simplifies the analysis (i.e. the species dataset is reduced from 300

168

to 22 variables in the corresponding biological traits data set). Despite the advantages

169

of using a BTA approach, we emphasize the importance of the underlying species

170

composition to understand the implications of the traits composition on the ecosystem

171

functioning and the functional diversity (Muntadas et al. 2015; Muntadas et al. 2016)

172

Five traits exhibiting different behavioural and morphological categories were chosen to

173

characterize the epifaunal benthic community (Fig. 2). These traits (position on the

174

177

178

Traits were assigned to each species based on available literature, experts' knowledge

179

and information from the BIOTIC database (http://www.marlin.ac.uk/biotic/) (see

180

Muntadas et al. 2015 for further details). In order to obtain the trait abundance per

175 176

EP

TE D

M AN U

SC

RI PT

154

AC C

ACCEPTED MANUSCRIPT

seabed, feeding mode, motility, size and other attributes, e.g. fragility, vermiform, hard shell), were selected because they are easily assigned to epibenthic species and are essential attributes for an indicator to represent community responses to trawling activity (de Juan and Demestre, 2012; de Juan et al., 2009) (Fig. 2).

7

ACCEPTED MANUSCRIPT

replicate, the abundance of all species in a replica exhibiting a particular trait category

182

were added (see de Juan et al., 2007 for further details on the approach).

183

2.5. Statistical approach

184

For each trait, we built a predictive model based on multiple regressions linking trait

185

abundance to trawling effort and to sediment composition in each site.

186

a) Predictor of trait abundance

187

The abundance of each trait category per replicate (Tij, where i is an index of site and j

188

is an index of replicate within each site) was standardized to a 0-1 scale by dividing the

189

abundance of the trait by the sum of the abundance of all trait categories in that

190

replicate. For each trait, the standardized trait data was used along with sediment sand

191

content and SSS effort data to estimate regression coefficients in the following

192

expression, by means of classical least squares regression:

SC

M AN U

Tij=AS· SSSij + ASa · Sandij + B

(1)

TE D

193

RI PT

181

Where As is the regression coefficient for fishing effort computed as trawl tracks

195

density/km2 (SSS) over a soft bottom with a specific sand content (Sand, %) and Asa is

196

the regression coefficient for Sand. We chose the SSS effort estimates instead of the

197

fishermen interviews because, as SSS provides an actual picture of the trawl tracks on

198

the seabed (Collie et al., 1997; Friedlander et al., 1998), it better reflects the potential

199

effects of fishing effort on epibenthic community (Demestre et al. 2015). Sand content

201 202

AC C

200

EP

194

was included as a surrogate of the sediment composition (i.e. grain size), an important factor when evaluating fishing effects on benthic communities, as different sediment

types will be differently affected by bottom trawling (Kaiser et al., 2006; Muntadas et al.,

203

2015).

204

AS, ASa and B are the regression coefficients estimated for each trait category in figure

205

2 (Table 2). The hypothesis of correlation r between each trait Tij and effort index SSS

8

ACCEPTED MANUSCRIPT

equal 0, was tested using Student’s t-test. For the traits “small” and “predator” we could

207

not accept this hypothesis (p-value are 0.72 and 0.56 respectively) and hence the

208

expression 1 could not be applied to these traits and they were marked as

209

“undetermined” in the simulation outputs, i.e. changes in the abundance of these traits

210

does not depend on effort variation (Table 2). For some other traits (deep burrowing,

211

surface burrowing, deposit feeding, high motile and flexible) this hypothesis remains

212

questionable because the p-value was in [0.10, 0.30] range, but we considered the

213

expression 1 still valid for these traits under the practical significance hypothesis

214

(Gibbons and Pratt 1975) (Table 2). However, these low correlations suggest that

215

these traits are little impacted by fishing effort.

216

An additional dataset can be loaded to the platform by the user, hence adding a new

217

study site of the users’ interest in which an effort change might be tested. If SSS data is

218

not available for this additional site, needed as an input for expression 1, other

219

estimates of fishing effort can be used indirectly. As fishermen interviews are more

220

straightforward to obtain, a regression model that allows the calculation of effort as

221

trawl tracks density/km2 (SSS) from the Gross Tonnage*days at sea/month estimation

222

was defined using the available estimations in our study sites (fishing effort values are

223

here divided by 100000 to scale them to the same order of magnitude as the traits

224

variables):

225

Trawl tracks density/ km2 = 1.76* GT*days at sea/month + 0.57

227

SC

M AN U

TE D

EP

(2)

AC C

226

RI PT

206

All the regression parameters in expressions 1 and 2 were calculated using the software SPAD 2003 version 5.6.

228

Finally, the simulation tool calculates the trait category abundance in a given site under

229

the new fishing effort (T*) considering the current trait category abundance (T) using

230

the following expression:

231

T*= T+ (Test*-Test)

(3) 9

ACCEPTED MANUSCRIPT

Where Test and Test* are the estimated values of a given trait category for the current

233

and the new effort respectively, calculated by the expression (1). T and T* are,

234

respectively, the current and new estimated abundance of the trait category in that site.

235

Using Test* and Test we obtain a gradient (Test*-Test), which is assumed to be caused by

236

effort change and that is used to modify the present trait category abundance T. T* is

237

the value represented in the simulation outputs

238

b) Condition to pressure index (CPI)

239

In order to provide an overview of the average presence of vulnerable/resistant traits

240

on a site, we developed a “condition to pressure” index calculated by the following

241

expression:

SC

Index= (2+2* RR+R-V-2*VV)/4

M AN U

242

RI PT

232

(4)

Where RR is the sum of the abundance of all the very resistant traits in a site (traits in

244

green in figure 2), R is the sum of the abundance of all the resistant traits in the same

245

site (traits in orange in figure 2), V the sum of the abundance of all vulnerable traits

246

(traits in yellow in figure 2) and VV the sum of the abundance of all very vulnerable

247

traits (traits in red in figure 2).

248

This formula provides a value comprised between 0 and 1, 0 meaning that all the traits

249

in the site are very vulnerable to trawling and 1 meaning that all the traits are very

250

resistant to trawling. For example, a site dominated by sessile, fragile, emergent and

252 253

EP

AC C

251

TE D

243

filter feeding organisms such as gorgonians or bryozoans would have a CPI close to 1, while a site dominated by motile scavengers such as sea stars or crabs would gave a CPI close to 0. In our case study, as the Medes site is located within a MPA, we can

254

take its CPI (0.227) as a reference value. Therefore, the importance of this index

255

resides in the possibility to represent with a single number the overall benthic

256

community status under a given trawling intensity.

10

ACCEPTED MANUSCRIPT

257

Similar to the trait abundance estimation case (expression 3), and in order to estimate

258

the CPI under a different fishing effort (CPI*), the platform uses the following

259

expression:

260

CPI*= CPI + (CPIest* - CPIest)

261

Where CPIest and CPIest* are the estimated CPI values for the current and the new effort

262

values respectively, calculated by the expression (4) using Test and Test* values

263

respectively. CPI and CPI* are the current and new estimated CPI values. Using CPIest*

264

and CPIest we obtain a gradient (CPIest* - CPIest), which is assumed to be caused by

265

effort change and that is used to modify the present CPI value.

266

3. RESULTS

267

3.1. The knowledge platform

268

The platform is designed in four blocks (Fig. 3), aimed at disseminating scientific

269

knowledge about bottom trawling effects on benthic communities in an understandable

270

and visual way. The first block contains the platform description, where the user will

271

find a brief explanation of the objectives and contents of this platform. A second block

272

contains the platform methodology. Within this second block, the user will find three

273

sections analogous to the previous sections 2.2, 2.4, 2.5 in this manuscript (i.e. fishing

274

effort estimation, biological traits approach and statistical approach). In the platform

275

section “biological traits approach”, besides the biological traits’ methodology, the user

277 278

RI PT

SC

M AN U

TE D

EP

AC C

276

(5)

will find detailed information on the data used to build the simulations (Fig. 4). The third block describes the 18 study sites located in the Mediterranean where the sampling was carried out (see section 2.2 in this manuscript). Moreover, it provides detailed

279

information about the epifaunal organisms identified in each site, as well as a

280

community comparison among sites (see an example in Fig. 5). This information is

281

aimed at visualizing the variable structure of benthic communities subjected to different

282

fishing effort regimes. This comparison might help transferring knowledge on the 11

ACCEPTED MANUSCRIPT

effects of trawling on the seafloor and raise awareness amongst stakeholders of the

284

ecosystem impact of this fishing gear.

285

The fourth block corresponds to the simulation tool, an example of which is provided in

286

section 3.2 of this manuscript. In this block there is also a tool that enables the user to

287

load data from their area of interest. The user needs to enter 1) the standardised

288

abundance of each trait category following the methodology explained in section 2.3 of

289

this manuscript (an example with a small species set is provided in the platform’s

290

Materials and Methods block), 2) the percentage of sand content in the sediment and 3)

291

an estimation of the trawling effort in their area of interest (computed as trawl tracks

292

density/km2 or as GT*days at sea). By following these three steps, a new site is created

293

in the platform’s database, therefore, it can be selected by the simulation tool. Hence,

294

the user can test potential changes on benthic community structure due to a fishing

295

effort change in their area of interest.

296

These four main blocks are not totally independent as the user may need information

297

on, for example, the trait approach (block 2) when exploring community structure in a

298

given area (block 3). In such cases, the platform provides an information button to

299

move quickly from one block to another one (Fig. 6).

300

Additionally, the platform has the capacity to communicate with other software such as

301

Adobe Acrobat to display the PDF files. This allows including further information about

302

the topics displayed in each block (Fig. 6).

304

SC

M AN U

TE D

EP

AC C

303

RI PT

283

3.2. Simulation block outputs The simulation tool displayed within the simulation block offers the possibility to choose

305

one of the study sites, modify the fishing effort and run the simulation to visualize the

306

potential effect of the effort change on the benthic community structure from the

307

selected site. The regression model described in section 2.5. performs the simulation. If,

12

ACCEPTED MANUSCRIPT

as previously mentioned, the user creates a new study site by entering a data set from

309

their area of interest, this site can be selected to perform the simulation on it.

310

As an example, figures 7 and 8 show the simulation results for a 50% increase and a

311

50% decrease, respectively, in fishing effort in the sites D, ICL, CCH and CPM. The

312

effort decrease might be attained by a reduction of days at sea or by a reduction of the

313

fleet size, whereas the effort increase might be achieved by increasing the days at sea

314

(e.g., by fishing more days per week or by eliminating closed seasons). The site D was

315

chosen for being the muddiest site holding the highest effort (99.5% mud; 116835 trawl

316

tracks length/km2). The sites ICL, CCH and CPM are subjected to moderate trawling

317

effort (86615, 81114 and 54199 trawl tracks length/km2 respectively) and have

318

heterogeneous sediment (96.3%, 61.6% and 32.5% of mud respectively). Therefore,

319

the simulation outputs of these four sites will be useful for comparing 1) the response of

320

benthic communities inhabiting similar habitats under different effort regimes (D and

321

ICL) and 2) the response of benthic communities from different habitats under similar

322

effort regimes (ICL, CCH and CPM).

323

As expected, generally the very vulnerable traits decreased with an increase in the

324

effort and vice-versa. Regarding the very resistant traits, the opposite happened; these

325

traits predominantly increased with an increase in the effort and vice-versa. However,

326

some unexpected outcomes arose. Surprisingly, the trait “large size” increased when

327

augmenting the fishing effort and decreased when reducing effort in all cases. Also,

329 330

SC

M AN U

TE D

EP

AC C

328

RI PT

308

“fragile”, a very vulnerable trait, increased when the effort was raised and decreased when the effort was diminished at all sites with the exception of D site. Another inconsistency was detected in D site, as when reducing the effort almost all very

331

vulnerable traits (except for “filter feeding”) decreased; however, these very vulnerable

332

traits also decreased when the effort increased. On the other hand, the very resistant

333

trait “vermiform” decreased in all sites when augmenting the effort. The trait “deep

13

ACCEPTED MANUSCRIPT

burrowing” decreased in both cases, when augmenting and when reducing the fishing

335

effort in CPM site (Fig. 7 and 8).

336

The computed index, CPI, increased in all sites when augmenting the effort and

337

decreased when reducing the effort, but the magnitude of increase was higher in the

338

less muddy sites, while the magnitude of decrease was higher in the muddier sites

339

(Table 3).

340

4. DISCUSSION

341

The knowledge-based platform was built by combining the capacity to represent

342

information in an understandable way with simulation capacities (Ballé-Beganton et al.

343

2010). These characteristics allowed fulfilling the two main objectives of the platform: to

344

communicate trawling impacts on the seafloor in a dynamic and visual way, and to

345

simulate the effects of an effort change on the benthic community structure. Moreover,

346

the platform is able to interface with other software (e.g., PDF reader, GIS and other

347

executables), which offers the possibility to provide information through different

348

interfaces. Having these characteristics, this platform may be used as a deliberation

349

support tool for fishery managers and decision-makers. But, in order to be used by

350

stakeholders, it is important that the platform is also characterised by portability.

351

Portability is attained by the open access of a demo version of ExtendSim®, which can

352

load and run a full model. The portability to an Apple Mac version of ExtendSim® has

353

not been attempted, but this might be done with minor adjustments. Therefore, the

356

AC C

EP

TE D

M AN U

SC

RI PT

334

357

Despite the large amount of available data, a model forecasting the effects of fishing

358

effort on one biological trait cannot be envisaged yet. The primary reason for this, lies

359

in the lack of temporal data for measuring the impact of an increasing or decreasing

354 355

main strength of this platform is its ability to disseminate scientific results in a

comprehensible way and, hence, its potential to be used as a tool to share information among fisheries’ stakeholders.

14

ACCEPTED MANUSCRIPT

trawling effort on the benthic community within each study site. But we can argue that

361

all traits were found in all sites, the abundance of the categories within them varying

362

depending on the effort level, the sediment and their vulnerability to trawling. Moreover,

363

the trawling activities were similar in all sites. In consequence, there are elements that

364

argue in favor of a kind of ergodicity that should overcome the lack of temporal data.

365

Nevertheless, our model cannot be considered as a model for prediction and

366

forecasting of trawling effects on benthic communities, but as a tool to display trends of

367

potential changes on benthic communities caused by trawling. Consequently, the

368

model objective is not to forecast specific responses of biological traits to effort

369

changes, and it should be framed in the proper perspective of a pragmatic

370

communication and information exchange tool among fisheries’ actors.

371

Simulation outputs of a fishing effort increase generally revealed a vulnerable traits’

372

abundance decrease and a resistant traits’ abundance increase. However, contrary to

373

what we expected, fragile organisms, which are known to be vulnerable to fishing (de

374

Juan et al., 2009), increased their density when increasing effort and decreased when

375

diminishing effort. Nevertheless, the category fragile organisms include some bivalves

376

and sea urchins that may also exhibit resistant traits such as small sizes or burrowing

377

behaviour. Moreover, the trait “fragile”, despite being “very vulnerable” to trawl fishing,

378

was not the main trait driving differences between fished and unfished areas (generally,

379

“filter feeding” and “sessile” were the very vulnerable traits driving differences between

380

effort levels) (Muntadas et al., 2014). Besides, the correlation of the trait “fragile” with

383

AC C

EP

TE D

M AN U

SC

RI PT

360

384

such as sea stars and fish. As mentioned in section 2.4., we emphasize the importance

385

of the knowledge on the species composition behind each set of biological traits in

386

order to understand the wider ecosystem implications of a variable trait composition

387

(Palumbi et al., 2009; Snelgrove et al., 2014). It is also worth to mention that

381 382

fishing effort was not very strong (see section 2.5). On the other hand, the increase in the abundance of the trait “large” size related with an increasing effort, might be due to the increase in the abundance of large mobile/highly motile scavengers and predators,

15

ACCEPTED MANUSCRIPT

unexpected results might be also due to correlation between traits, for example filter

389

feeders tend to be sessile organisms in our study sites (Muntadas et al. 2015). These

390

correlations arise due to our limited species set and may change if other sites with

391

different species are added to the platform, for example filter feeders might be

392

organisms with mobility like crinoidea or bivalve scallops.

393

Moreover, it is important to stress that it is the relative abundance of all traits, and not

394

the abundance of each trait category individually, that allows to assess the overall

395

community status (de Juan and Demestre, 2012). In this context, the simulation tool

396

calculates the Condition to Pressure Index (CPI) to provide an overview of the

397

community status based on the combination of all traits exhibited by the community

398

being assessed.

399

According to the CPI, a fishing effort increase would mainly affect the sandy-mud CCH

400

site and the maërl site CPM. Actually, fishing is known to have strong negative effects

401

on biogenic habitats such as the maërl beds in the CP area (Collie et al., 2009; Kaiser

402

et al., 2006). On the other hand, the slight increase of CPI in muddy bottom

403

communities (D and ICL) may reflect that these intensively trawled communities have

404

already attained an alternative state dominated by organisms with low vulnerability to

405

trawling (de Juan et al., 2009; Thrush and Dayton, 2010). Otherwise, benthic

406

communities inhabiting muddy sediments showed to recover the most vulnerable traits

407

when the fishing effort decreased, despite being still far from the scenario observed in

408

Medes site (MPA). These results suggest a higher recovery potential of the species

411

AC C

EP

TE D

M AN U

SC

RI PT

388

412

as Nemertesia antenina or Suberites carnosus, grew rapidly when fishing effort

413

decreased. However, it is worth to bear in mind that most of our study sites had muddy

414

sediments, thus predicted effects for these habitats would be more consistent than for

415

gravel/sand sites. Predicted effects might be improved with the addition of sites in the

409 410

that characterise muddy bottoms, compared to the species that exhibit the same

vulnerable traits in sandier areas. These results are consistent with Strain et al. (2012)

who found that some sessile filter feeding species associated to muddy bottoms, such

16

platform database that cover a wide range of fishing effort levels and sediment type.

417

These new sites could be used to recalculate the regression parameters and thus

418

increase their robustness.

419

It is worth to bear in mind that the simulation provided by the platform does not take

420

into account recovery by migration of adult organisms from neighboring sites, or by

421

larval dispersal at larger scales (de Juan et al., 2014). This omission could partly

422

explain why the abundance of almost all very vulnerable traits in D site decreased

423

when reducing the effort. The original abundance of the traits “emergent”, “sessile” and

424

“fragile” in the D site (i.e. trait abundance at current/old effort) was 0, hence, no

425

recovery could take place without migration. This weakness is a very important aspect

426

to take into account for further development of the simulation tool, as migration from

427

nearby and potentially less damaged patches plays a very important role in community

428

recovery (Lambert et al., 2014; Lundquist et al., 2010). However, the existence of these

429

less damaged patches must be recorded and, to our knowledge, in D site the area

430

surrounding the study site is heavily trawled (de Juan et al. 2009). This might also

431

explain why the abundance of deep-burrowing organisms did not increase in CPM

432

when fishing effort increased, as the original abundance of this trait was also 0.

433

Moreover, maërl sediments such as those that characterized CPM do not favor the

434

settlement of this type of organisms (Barberá et al., 2012; Bremner et al., 2006). These

435

results strengthen previous evidence for an habitat-dependent response to trawling (de

436

439 440

despite increasing the accuracy of outputs, would also make the parametrization more

441

complex. Including recovery would imply to add the species’ dispersal traits (i.e. type of

442

larvae or reproduction frequency), which are unknown for many species (Tyler et al.,

437 438

EP

TE D

M AN U

SC

RI PT

416

AC C

ACCEPTED MANUSCRIPT

Juan et al., 2013). But as mentioned earlier, in order to detect consistent patterns for

each habitat type, in the future more sites should be added to the model. Therefore, the platform’s results are rather conservative as, by taking migration into account, the

recovery potential could be higher. However, the inclusion of recovery by migration,

17

ACCEPTED MANUSCRIPT

2012), and the community patches’ distribution at medium scales (de Juan et al., 2014;

444

Lundquist et al., 2010). Our model, despite providing coarser estimations, is potentially

445

applicable to any soft-bottom trawled area, as no previous knowledge of community

446

distribution is required.

447

An additional platform’s strength is the option for adding a new study site from the

448

user’s area of interest, as long as epifauna data, sediment sand content and effort

449

estimation had been recorded. The potential effects of a trawling effort change on the

450

benthic community in the new area of interest can be tested using the regression

451

model explained in section 2.5. Currently, a link to connect this platform with R

452

(www.R-project.org) is under development. This connection would allow the

453

recalculation of the regression model parameters by using both, current data and data

454

from the new study sites included by the user. The more sites are added, the wider will

455

be the range of variables used in the parameters’ estimation which, as previously

456

mentioned, will increase the simulation robustness.

457

An additional future development considers the demersal fish community. Benthic

458

community status plays an essential role for the life-cycle of target species’ stocks,

459

hence, fisheries management should take into account of this compartment

460

(Archambault et al, this issue). For example, Muntadas et al. (2014) observed that the

461

functional trait composition of benthic communities was tightly linked with demersal fish

462

species. Therefore, in line with the aim of being a support deliberation tool, further

464 465

SC

M AN U

TE D

EP

AC C

463

RI PT

443

development of the platform should include the link benthos-commercial species. In

this context, the additional inclusion of an infauna compartment in the simulation would provide important information, as many target species feed on infauna (Hiddink et al.,

466

2008; Labropoulou et al., 1997). These further developments would improve the

467

information given to stakeholders and would contribute to generate knowledge to

468

advance towards an integrated assessment of trawled ecosystems.

469

ACKNOWLEDGEMENTS 18

ACCEPTED MANUSCRIPT

This work was funded by the research project COMSOM (CTM2008-04617/MAR),the

471

project MEDAS (CTM2007-29845-E/MAR), the EU project RESPONSE (Q5RS-2002-

472

00787) and by the EU Seventh Framework Program for research, technological

473

development and demonstration (FP7/2007-2013) within the Ocean of Tomorrow call

474

under Grant Agreement No.266445-VECTORS. The crew of the RV Garcia del Cid is

475

thanked for their help during the “8 Veda” cruises of RESPONSE, and the cruise

476

(CTM2008-04206-E/MAR). Alba Muntadas was supported by a CSIC JAE_predoc

477

grant cofounded by the FSE (European Social Funds).

478

REFERENCES

479 480

Allen J, Clarke K (2007) Effects of demersal trawling on ecosystem functioning in the North Sea: a modelling study. Mar Ecol Prog Ser 336:63–75

481 482 483

Archambault B, Rivot E, Savina M, Le Pape O (this issue) Disentangling the impacts of natural (hydroclimate) and anthropogenic (habitat degradation and fishing pressure) stressors occurring along the life cycle of an exploited flatfish population

484 485

Auster PJ, Langton RW (1998) The Effects of Fishing on Fish Habitat. Am Fish Soc Symp 22:150–187

486 487 488 489

Ballé-béganton J, Bailly D, Lample M, Mongruel R, Prou J, Réthoret H, Bacher C, Agùndez JAP (2012) Building an Integrated Model for Freshwater Allocation with Local Managers in a Coastal Area. Proc 2012 Int Congr Environ Model Software Sixth Bienn Meet

490 491 492 493

Ballé-Béganton J, Lample M, Bacher C, Fiandrino A, Guillard V, Laugier T, Mongruel R, Pérez Agúndez JA (2010) A modelling platform for complex socio- ecosystems : an application to freshwater management in coastal zones. Proc 2010 Int Congr Environ Model Software Fifth Bienn Meet

494 495 496

Barberá C, Moranta J, Ordines F, Ramón M, Mesa A, Díaz-Valdés M, Grau a. M, Massutí E (2012) Biodiversity and habitat mapping of Menorca Channel (western Mediterranean): implications for conservation. Biodivers Conserv 21:701–728

SC M AN U

TE D

EP

AC C

497 498

RI PT

470

Bremner J (2008) Species’ traits and ecological functioning in marine conservation and management. J Exp Mar Bio Ecol 366:37–47

499 500

Bremner J, Rogers S, Frid C (2003) Assessing functional diversity in marine benthic ecosystems: a comparison of approaches. Mar Ecol Prog Ser 254:11–25

501 502

Bremner J, Rogers S, Frid C (2006) Matching biological traits to environmental conditions in marine benthic ecosystems. J Mar Syst 60:302–316

19

ACCEPTED MANUSCRIPT

Carter CA (2013) Constructing sustainability in EU fisheries: Re-drawing the boundary between science and politics? Environ Sci Policy 30:26–35

505 506

Christensen V, Walters CJ (2004) Ecopath with Ecosim: methods, capabilities and limitations. Ecol Modell 172:109–139

507 508

Collie JS, Escanero GA, Valentine PC (1997) Effects of bottom fishing on the benthic megafauna. Mar Ecol Prog Ser 155:159–172

509 510 511

Collie JS, Hermsen JM, Valentine PC (2009) Recolonization of gravel habitats on Georges Bank (northwest Atlantic). Deep Res Part II Top Stud Oceanogr 56:1847–1855

512 513 514

Corrales X, Coll M, Tecchio S, Bellido JM, Fernández ÁM, Palomera I (2015) Ecosystem structure and fishing impacts in the northwestern Mediterranean Sea using a food web model within a comparative approach. J Mar Syst 148:183–199

515 516 517

Daw T, Gray T (2005) Fisheries science and sustainability in international policy: a study of failure in the European Union’s Common Fisheries Policy. Mar Policy 29:189–197

518 519 520

de Juan S, Thrush S, Demestre M (2007) Functional changes as indicators of trawling disturbance on a benthic community located in a fishing ground (NW Mediterranean Sea). Mar Ecol Prog Ser 334:117–129

521 522 523

de Juan S, Demestre M, Thrush S (2009) Defining ecological indicators of trawling disturbance when everywhere that can be fished is fished: A Mediterranean case study. Mar Policy 33:472–478

524 525

de Juan S, Demestre M (2012) A Trawl Disturbance Indicator to quantify large scale fishing impact on benthic ecosystems. Ecol Indic 18:183–190

526 527 528

de Juan S, Iacono C Lo, Demestre M (2013) Benthic habitat characterisation of softbottom continental shelves: Integration of acoustic surveys, benthic samples and trawling disturbance intensity. Estuar Coast Shelf Sci 117:199–209

529 530 531

de Juan S, Thrush S, Hewitt J, Halliday J, Lohrer A (2014) Cumulative degradation in estuaries: contribution of individual species to community recovery. Mar Ecol Prog Ser 510:25–38

536 537 538 539 540 541 542

SC

M AN U

TE D

EP

AC C

532 533 534 535

RI PT

503 504

Demestre M, Muntadas A, de Juan S, Mitilineou C, Sartor P, Mas J, Kavadas S, Martín J (2015) The need for fine-scale assessment of trawl fishing effort to inform ecosystem approach to fisheries: exploring three data sources in Mediterranean trawling grounds. Mar Pol 62:134-143. doi: 10.1016/j.marpol.2015.09.012

Dichmont CM, Deng A, Punt AE, Ellis N, Venables WN, Kompas T, Ye Y, Zhou S, Bishop J (2008) Beyond biological performance measures in management strategy evaluation: Bringing in economics and the effects of trawling on the benthos. Fish Res 94:238–250 Duplisea DE, Jennings S, Warr KJ, Dinmore TA (2002) A size-based model of the impacts of bottom trawling on benthic community structure. Can J Fish Aquat Sci 1795:1785–1795 20

ACCEPTED MANUSCRIPT

ExtendSim 9 by Imagine That Inc. Product overview available via [accessed march 2015]

545 546 547

Friedlander AM, Boehlert GW, Field ME, Mason JE, Gardner JV, Dartnell P (1998) Sidescan-sonar mapping of benthic trawl marks on the shelf and slope off Eureka , California. Fish Bull 97:786–801

548 549

Fujioka JT (2006) A model for evaluating fishing impacts on habitat and comparing fishing closure strategies. Can J Fish Aquat Sci 63:2330–2342

550 551

Fulton E, Smith A, Punt A (2005) Which ecological indicators can robustly detect effects of fishing? ICES J Mar Sci 62:540–551

552 553

Garcia S, Cochrane K (2005) Ecosystem approach to fisheries: a review of implementation guidelines. ICES J Mar Sci 62:311–318

554 555

Gibbons JD, Pratt JW (1975) P-values : Interpretation and Methodology. Am Stat 29:20–25

556 557 558

Gregory RS, Anderson JT (1997) Substrate selection and use of protective cover by juvenile Atlantic cod Gadus morhua in inshore waters of Newfoundland. Mar Ecol Prog Ser 146:9–20

559 560 561

Hiddink JG, Jennings S, Kaiser MJ, Queirós AM, Duplisea DE, Piet GJ (2006) Cumulative impacts of seabed trawl disturbance on benthic biomass , production , and species richness in different habitats. Can J Fish Aquat Sci 736:721–736

562 563 564

Hiddink JG, Rijnsdorp AD, Piet G (2008) Can bottom trawling disturbance increase food production for a commercial fish species? Can J Fish Aquat Sci 65:1393– 1401

565 566 567

Hiscock K, Tyler-walters H (2006) Assessing the sensitivity of seabed species and biotopes – the Marine Life Information Network ( MarLIN ). Hydrobiologia 555:309–320

568 569 570

Kaiser M, Clarke K, Hinz H, Austen M, Somerfield P, Karakassis I (2006) Global analysis of response and recovery of benthic biota to fishing. Mar Ecol Prog Ser 311:1–14

571 572 573

Labropoulou M, Machias A, Tsimenides N, Eleftheriou A (1997) Feeding habits and ontogenetic diet shift of the striped red mullet, Mullus surmuletus Linnaeus, 1758. Fish Res 31:257–267

574 575 576

AC C

EP

TE D

M AN U

SC

RI PT

543 544

577 578 579

Lundquist C, Thrush S, Coco G, Hewitt J (2010) Interactions between disturbance and dispersal reduce persistence thresholds in a benthic community. Mar Ecol Prog Ser 413:217–228

Lambert GI, Jennings S, Kaiser MJ, Davies TW, Hiddink JG (2014) Quantifying recovery rates and resilience of seabed habitats impacted by bottom fishing (A Punt, Ed.). J Appl Ecol 51:1326–1336

21

ACCEPTED MANUSCRIPT

Mongruel R, Prou J, Ballé-béganton J, Lample M, Vanhoutte-brunier A, Réthoret H, Antonio J, Agúndez P, Vernier F, Bordenave P, Bacher C (2011) Modeling Soft Institutional Change and the Improvement of Freshwater. Ecol Soc 16:15

583 584 585 586 587

Mongruel R, Vanhoutte-Brunier A, Fiandrino A, Valette F, Ballé-Béganton J, Pérez Agúndez JA, Gallai N, Derolez V, Roussel S, Lample M, Laugier T (2013) Why, how, and how far should microbiological contamination in a coastal zone be mitigated? An application of the systems approach to the Thau lagoon (France). J Environ Manage 118:55–71

588 589 590

Muntadas A, Demestre M, de Juan S, Frid CLJ (2014) Trawling disturbance on benthic ecosystems and consequences on commercial species: a northwestern Mediterranean case study. Sci Mar 78S1:53–65

591 592 593

Muntadas A, de Juan S, Demestre M (2015) Integrating the provision of ecosystem services and trawl fisheries for the management of the marine environment. Sci Total Environ 506-507:594–603

594 595

Muntadas A, de Juan S, Demestre M (2016) Assessing functional redundancy in chronically trawled benthic communities. Ecol Ind 61: 882-892.

596 597 598 599

Palumbi SR, Sandifer P a, Allan JD, Beck MW, Fautin DG, Fogarty MJ, Halpern BS, Incze LS, Leong J-A, Norse E, Stachowicz JJ, Wall DH (2009) Managing for ocean biodiversity to sustain marine ecosystem services. Front Ecol Environ 7:204–211

600 601 602 603 604

Petihakis G, Petihakis G, Smith CJ, Smith CJ, Triantafyllou G, Triantafyllou G, Sourlantzis G, Sourlantzis G, Papadopoulou K-N, Papadopoulou K-N, Pollani a, Pollani a, Korres G, Korres G (2007) Scenario testing of fisheries management strategies using a high resolution ERSEM POM ecosystem model. ICES J Mar Sci 64:1627–1640

605 606 607

Pinnegar J, Sinerchia M, Palacz A, Hufnagl M, Girardin R, Peck MA, Nielsen JR, Kempf A, Marchal P (this issue) Application of the holistic modelling framework ATLANTIS to European Shelf Seas: lessons learnt and continuing challenges.

608 609 610 611 612 613

Rose K, Allen JI, Artioli Y, Barange M, Blackford J, Carlotti F, Cropp R, Daewel U, Edwards K, Flynn K, Hill S, HilleRisLambers R, Huse G, Mackinson S, Megrey B, Moll A, Rivkin R, Salihoglu B, Schrum C, Shannon L, Shin Y-J, Smith SL, Smith C, Solidoro C, John M, Zhou M (2010) End-To-End Models for the Analysis of Marine Ecosystems: Challenges, Issues, and Next Steps. Mar Coast Fish Dyn Manag Ecosyst Sci 2:115–130

SC

M AN U

TE D

EP

AC C

614 615

RI PT

580 581 582

Snelgrove PVR, Thrush SF, Wall DH, Norkko A (2014) Real world biodiversity– ecosystem functioning: a seafloor perspective. Trends Ecol Evol 29:398–405

616 617 618

Strain EMA, Allcock AL, Goodwin CE, Maggs CA, Picton BE, Roberts D (2012) The long-term impacts of fisheries on epifaunal assemblage function and structure, in a Special Area of Conservation. J Sea Res 67:58–68

619 620

Thrush SF, Dayton PK (2002) Disturbance to marine benthic habitats by trawling and dredging: Implications for Marine Biodiversity. Annu Rev Ecol Syst 33:449–473

22

ACCEPTED MANUSCRIPT

Thrush SF, Dayton PK (2010) What Can Ecology Contribute to Ecosystem-Based Management? Ann Rev Mar Sci 2:419–441

623 624 625

Tillin H, Hiddink J, Jennings S, Kaiser M (2006) Chronic bottom trawling alters the functional composition of benthic invertebrate communities on a sea-basin scale. Mar Ecol Prog Ser 318:31–45

626 627 628

Tomlinson B, Maynou F, Sabatés A, Fuentes V, Canepa A, Sastre S (this issue) Systems approach modelling of the interactive effects of fisheries, jellyfish and tourism in the Catalan coast

629 630 631

Toonen HM, Mol APJ (2013) Putting sustainable fisheries on the map? Establishing notake zones for North Sea plaice fisheries through MSC certification. Mar Policy 37:294–304

632 633 634

Travers M, Watermeyer K, Shannon LJ, Shin Y-J (2010) Changes in food web structure under scenarios of overfishing in the southern Benguela: Comparison of the Ecosim and OSMOSE modelling approaches. J Mar Syst 79:101–111

635 636 637 638

Tyler EHM, Somerfield PJ, Berghe E Vanden, Bremner J, Jackson E, Langmead O, Palomares MLD, Webb TJ (2012) Extensive gaps and biases in our knowledge of a well-known fauna: Implications for integrating biological traits into macroecology. Glob Ecol Biogeogr 21:922–934

639 640 641

Verweij MC, Van Densen WLT , Mol AJP (2010) The tower of Babel: Different perceptions and controversies on change and status of North Sea fish stocks in multi-stakeholder settings. Mar Policy 34:522–533

M AN U

SC

RI PT

621 622

645 646 647 648 649 650

EP

644

AC C

643

TE D

642

651 652 653

23

ACCEPTED MANUSCRIPT

Table 1. Fishing effort estimations in the sites included in the platform.

656

657

658

659

660

sea/month)

lenght/km2)

LCL

6256

87542

LCM

7930

96954

ICL

5593

86615

ICM

10603

102017

ICH

17255

122131

CCL

0

22393

CCM

1323

63171

CCH

1623

81114

CPL

1125

CPM

17889

CPH

22881

AC

21893

D

45125

M

0

M AN U

SC

RI PT

SSS (Trawl tracks

48248

54199

137243 60772

116835 0

AC C

655

Fishermen data (GT*days at

EP

Site

TE D

654

661

662

24

ACCEPTED MANUSCRIPT

663

Table 2. Regression coefficients and correlations obtained from Expression 1 applied

664

to each biological trait in figure 2. * p<0.05, ** p <0.01, *** p<0.001, ns no significant Trait

SSS

Sand correlation

ASa

Emergent

-0.064

7.04***

-0.042

Surface

0.038

3.45**

0.132

Surface Burrowing

0.013

1.10ns

-0.092

Deep burrowing

0.011

1.18ns

0.006

Filter feeding

-0.073

4.94***

-0.094

4.04***

Deposit feeding

0.009

1.05ns

0.010

0.73ns

Predator

-0.003

0.59ns

0.003

0.45ns

Scavenger

0.067

5.26***

0.073

3.68***

Sessile

-0.0504

4.86***

-0.037

2.29*

Sedentary

-0.038

2.98**

-0.056

5.86**

Motile

0.073

5.36***

-0.013

0.59ns

0.0129

1.09ns

0.104

5.68***

0.038

3.82***

-0.054

3.55***

-0.033

2.24*

0.061

2.65**

-0.0063

0.35ns

-0.012

0.45ns

0.0039

2.24*

0.004

1.59ns

No protection

-0.046

2.39*

-0.021

0.68ns

Flexible

0.003

1.29ns

0.004

1.64ns

Strong

-0.0242

1.69ns

0.095

4.27***

Regeneration

0.0558

5.38***

-0.050

3.49***

Vermiform

-0.0093

2.27*

-0.002

0.31ns

Hard shell

0.0233

2.78**

-0.003

0.26ns

Large

AC C

Fragile

EP

Medium Small

2.96**

7.66***

4.88*** 0.42ns

SC

M AN U

TE D

Highly motile

correlation

RI PT

As

665 25

ACCEPTED MANUSCRIPT

666

Table 3. Condition to Pressure Index values for D, ICL, CCH and CPM sites under the

667

current effort pressure and for 50% increase and 50% decrease of the current effort. Current effort

50% effort decrease

D

0.686

0.643

0.504

ICL

0.654

0.630

0.531

CCH

0.652

0.552

0.533

CPM

0.664

0.541

0.523

SC

668

RI PT

50% effort increase

AC C

EP

TE D

M AN U

669

26

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

HIGHLIGHTS ● Understandable deliberation support tools to share fisheries’ knowledge are needed. ● The presented platform interface is easy to use by any fisheries’ decision actor. ● The platform shows trawling impact on benthic communities in a comprehensible way.

RI PT

● The platform provides a simulation tool to assess fishing effort changes’ effects.

AC C

EP

TE D

M AN U

SC

● The user may test their dataset form their community of interest.