55TECHNICALDOCUMENT

COPEMED II WORKING GROUP ON STOCK ASSESSMENT OF P. BOGARAVEO

IN THE STRAIT OF GIBRALTAR

28-29 October 2019Málaga, Spain

CopeMed II Technical Documents Nº55 GCP/INT/028/SPA; GCP/INT/362/EC

COPEMED II WORKING GROUP ON STOCK ASSESSMENT OF P. BOGARAVEO IN THE STRAIT OF GIBRALTAR

October 2019

The conclusions and recommendations given in this document and in other documents in the Co-ordination to Support Fisheries Management in the Western and Central Mediterranean CopeMed II Project series are those considered appropriate at the time of preparation. They may be modified in the light of further knowledge gained in subsequent stages of the Project. The designation employed and the presentation of material in this publication do not imply the expression of any opinion on the part of Food and Agriculture Organization of the United Nations, FAO, the Government of Spain or the Commission of the European Union concerning the legal status of any country, territory, city or area, or concerning the determination of its frontiers or boundaries. This document has been financed by the European Union and the Government of Spain. The views expressed herein can in no way be taken to reflect the official opinion of the European Union or the Government of Spain.

3

Preface The CopeMed II Project on Co-ordination to Support Fisheries Management in the Western and Central Mediterranean is executed by the Food and Agriculture Organization of the United Nations (FAO) and funded by the Government of Spain, represented by the Secretaría General de Pesca (Mº de Agricultura, Pesca y Alimentación, MAPA), and the European Union, represented by the European Commission (EC).The premises of the project at the Subdelegación del Gobierno in Málaga (Spain) are part of the Spanish contribution included in the agreement with the FAO. The overall objective of the project is to contribute to the sustainability of marine fisheries in the Central and Western Mediterranean by providing technical support towards the improvement of monitoring and management of fisheries at national level and by reinforcing scientific collaboration among countries at the sub-regional level. The activities of the project also support countries in the implementation of recommendations agreed at the regional level by the General Fisheries Commission for the Mediterranean (GFCM). Regions covered by CopeMed II are the western and central sub-regions of the Mediterranean. Participating countries are Algeria, France, Italy, Libya, Malta, Morocco, Tunisia and Spain. The main beneficiaries are the fishery policy-makers, managers and fishery administrations in the western and central Mediterranean countries. The project is also contributing to the strengthening of regional collaboration by supporting the participation of the countries in relevant regional scientific organizations, such as the FAO’s General Fisheries Commission for the Mediterranean (GFCM). Secondary beneficiaries include the national research institutes, fishers and fishers’ associations, and industrial organizations.

Project CopeMed II (FAO-FIRF) Subdelegación del Gobierno en Málaga Paseo de Sancha 64, Oficinas 303-304

29016 Málaga España

Tel: (+34) 952 989299 Fax: (+34) 952 989252

e-mail: copemed@fao.org

URL: www.faocopemed.org

4

CopeMed II (GCP/INT/028/SPA – GCP/INT/362/EC) Publications

CopeMed II project publications are issued in the CopeMed Technical Documents series and are related to meetings, missions and research organized or conducted within the framework of the CopeMed II Project. Comments on this document would be welcomed and should be sent to the Project premises:

Project CopeMed II (FAO-FIAF) Subdelegación del Gobierno en Málaga Paseo de Sancha 64, Oficinas 303-304

29016 Málaga (España) copemed@fao.org

For bibliographic purposes this document should be cited as follow: CopeMed II. 2019. Report of the CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar, Malaga, Spain, 28 – 29 October 2019. CopeMed II Technical Documents Nº55 (GCP/INT/028/SPA-GCP/INT/362/EC). 47 pp. Cover picture: Voracera fishing operation in the Strait of Gibraltar. Author. Juan Gil Herrera (IEO-Cadiz)

Preparation of this document

This document is the final version of the report of the ad hoc CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar, Malaga, Spain, 28 – 29 October 2019. All participants (See list in appendix II) have contributed equally to the elaboration of this report.

5

Acknowledgements CopeMed II acknowledges the participation and valuable contributions of all experts from INRH (Morocco), IEO (Spain) in the ad hoc CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar.

CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar, Malaga, Spain, 28 – 29 October 2019. CopeMed II Technical Documents Nº55 (GCP/INT/028/SPA-GCP/INT/362/EC). 47 pp.

ABSTRACT As a follow up of the ongoing open benchmark assessment of blackspot seabream initiated in April 2019 at the request of GFCM-SAC, CopeMed II project organized this “Working Group”. The aim was to ensure all input data are adequately analysed and compiled in advance of the continuation of benchmark assessment in the framework of the GFCM working Groups on Stock Assessment planned for December 2019. This meeting was attended by experts from Morocco (INRH) and Spain (IEO) who analyzed in depth the existing data, the characteristics of the respective fleets, the process of standardization of CPUEs and the update of new biological parameters after new samples have been analyzed in Morocco from July 2018 to October 2019. The group concluded with the establishment of three different fleet segments operating over the same stock. Accordingly, CPUE indices should be split in three and these indices should be standardized for the purpose of assessing the status of the stock. A new set of values for biological parameters were obtained based on a full year biological sampling carried out in Morocco. Length at maturity was confirmed at 34,6 cm of total length for the combined stock.

6

Report of CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar.

Malaga, Spain, 28–29 October 2019

1. Background and objectives The meeting was kindly hosted by the ‘Subdelegación del Gobierno de Málaga’ where CopeMed II premises are located. It was attended by 10 participants from Algeria, Morocco and Spain. The agenda was adopted as presented in Appendix I. The whole list of participants is included as Appendix II.

Copemed II Fishery Expert, Ms Pilar Hernández introduced the background of previous meetings, the objectives and expected outputs of the current meeting that can be summarized as follows: i) Decide about the values of the most relevant biological parameters, ii) Define the possible model runs for GADGET to be carried out during the upcoming benchmark session in Rome 9-14 December. iii) Re-run LCA, BioDyn and Spict with the new adopted parameters; iv) check alternative models for the standardization of CPUEs, to use it as a tuning index 2. Revision of identified issues raised in the last benchmark session (Sète, April

2019) and discussion on the available data for assessment (inputs): The experts in charge of the standardization of CPUEs at the IEO and at the INRH introduced their work.

2.1. CPUE standardization: presentation of Spanish and Moroccan fleet exercise. (Lucía Rueda)

Ms. Lucía Rueda, from IEO presented the work done to standardize the CPUEs during and after the meeting of the SRC-WM in Séte, March 2019. She used a new Spanish series elaborated with information from VMS. This new series included the days with zero landings of P. bogaraveo (2009-2018) and uses the total duration (time) of the daily fishing operations as a better effort measure. Different models were tested which are listed below. Figures 1 and 2 shows a summary of the most fitted ones. Model tested to standardize Spanish series of CPUEs

1. Two Stage Boosted Regression Trees (BRT):

1. Binomial and Log Gaussian CPUE

2. Binomial and Log Gaussian Catch. Haul duration as covariate

2. Catch NB (negative binomial) GLMM: Offset: log(haul duration); Random effect: boat; Crossed/ nested random effects: boat*year

3. CPUE Tweedie family GLMM: Random effect: boat; Crossed/ nested random effects: boat*year

4. Hurdle ( = Delta) Models: Binomial and NB Catch GLMM. Offset: log(haul duration). Random effect: boat;

5. Mixture ( = Zero Inflated) Models:

1. Binomial and Log Gaussian CPUE GLMM. Random effect: boat;

7

2. Binomial and NB Catch GLMM. Offset: log(haul duration). Random effect: boat;

Figure 1: Model comparison – standardized CPUEs

Figure 2: Model comparison – standardized Catches

8

Figure 3 Nominal (red dots) and standardized monthly CPUE index, Spain. Conclusions of the modelling process for the Spanish series:

Indices of abundance (either CPUE or Catches) show a similar a general decreasing trend in the time series with a slight increase in 2015 and decreasing again afterwards which is very similar to the one obtained in the previous benchmark meeting with the combined data of Spain and Morocco and GLM. (Figure 3).

The duration of the haul, when significant, shows a positive effect on the catches/ occurrence of CPUEs.

The length of the boat, when significant, shows a positive effect on the catches/ CPUEs.

At the view of all the results obtained by the different models: It was confirmed that the standardized indices from the first trial performed in in the benchmark session during the Sub-regional Committee meeting in Western Mediterranean, Séte, March 2019, presented a similar trend to the new ones, despite the fact that new ones include days with 0 landings and consider fishing time as effort unit. This additional information on effort was provided by the IEO after the analysis of VMS data in Spain.

2.2. CPUE standardization: presentation of Moroccan fleet exercise. (Mansour Serghini).

The catch data by fishing vessel and by daily sales of each species landed were used to estimate the fishing effort (number of days at sea). First, the commercial data of longline fishery by their corresponding port were compiled with catches by species (in kg), registration numbers, and date of landing for the period 2009 to 2018. The daily sales of each species for one trip were added together by vessel and by port. Then, we built a data matrix with daily sales, registration number, port, date and catch per species as variables. According to the survey conducted in each port, generally a sale of the catch landed by one vessel corresponds to one trip (in general sales are one days). Finally, a new matrix was created with rows correspondent to the different trips. The effort unit is the fishing trip (one day). For the selection of explanatory variables the fishing tactics of longline fishery can be described by a combination of many characteristic variables, the approach is to extract all fishing trips that contributed to the constitution of this fishing tactics and use its significant characteristics variables as explanatory variables in the model to standardize CPUE. In general, the available data guides us to choose the most relevant variables. Nominal CPUEs are calculated based on the positive fishing trips of the

9

Blackspot seabream. As there are other accompanying species in the landings together with the target species, some indication of the associated community must be taken into account. In this case, the choice of the Shannon index is justified better than the integration of any single species as variables Thus, the explanatory variables selected in the model are the power, month, year, length and Shannon index. Finally, blackspot seabream catch, vessel power, length overall, and Shannon index were log(x+1) transformed for use in the final statistical model using Boosted Regression Trees (BRT).

LOG(Blackspot seabream (kg)+1) ~ year + month + LOG(power+1) + LOG(length +1) + LOG(Shannon index +1) Boosted Regression Trees (BRT)

The BRT is an ensemble of method that combine statistical models and machine learning through two algorithms: regression trees are from the classification and regression tree (decision tree) group of models, and boosting build (Machine Learning) (Elith et al., 2008). Handling exploratory variables of different types, treatment of missing data, handling outliers and insensitive to data distribution, dealing with complex nonlinear relations. Moreover, interactions are easily implemented, without concern for potentially complicated calculations of the standardized year effect (De’ath, 2007; Elith et al., 2008; Hastie et al., 2009). In the current study, Blackspot seabream daily catch was standardized using BRT main effect two way interactions (tc=2). The Catch and effort data (32554 fishing’s trips for Morocco) were subdivided on the training and test data sets (70%, 30%) for residual analyses and test prediction model with test data. After the validation of the model, the global data (train data + test data) were used to estimate final standardized CPUE. Results

The number of observations (trips) analyzed by variables combination is summarized in Table 1. For Morocco, a total of 32554 fishing’s trips, which represent an average of 3255 observations per year, were used to compute the standardized CPUE of Blackspot seabream for the period 2009-2019. The BRT models results showed that the relative influence of explanatory variables is ordered according to importance as follows: month (35%), Year (27%), indice_shannon (24%), length_V (8%) and power (7%) (Table 2).

The diagnostic plots showed that both BRT models fit well the data given the normal pattern of the residuals with (RMSE= 0.74 and PED = 25%) (Table 3). The fitted values vs the residuals did not show any particular trend (Figure 4, 5, 6). The computed annual standardized CPUE using the BRT model are summarized in Table 4. The yearly evolution of the standardized CPUE follows the same trend than the nominal CPUEs, indicating a slight declining trend since the early years and slight increase from 2012 until 2015 and declining trend again in the last three years (Figure 7.a) The catch and the effort also show parallels trends between them, differing from CPUE in the early years, when a slight increase is observed, (Figure 7.b).

10

Table 1 Number of observations (trips) by year and month: Morocco Year/Month 4 8 12 2 1 7 6 3 5 11 10 9 Total

2009 124 168 231 61 191 166 161 88 109 329 254 240 2122 2010 72 267 208 50 107 223 262 279 188 82 275 390 2403 2011 143 344 610 120 156 390 320 63 223 276 253 236 3134 2012 408 135 290 175 382 1 247 128 355 122 315 451 3009 2013 123 151 343 108 98 310 247 35 283 305 205 201 2409 2014 236 338 248 175 166 332 459 134 265 251 98 544 3246 2015 382 637 385 91 244 564 567 340 447 197 297 439 4590 2016 493 446 194 131 162 324 638 346 540 233 444 293 4244 2017 379 624 234 220 409 550 563 181 597 375 355 482 4969 2018 125 137 220 181 154 318 191 153 317 200 248 184 2428 Total 2485 3247 2963 1312 2069 3178 3655 1747 3324 2370 2744 3460 32554

Table 2.a Model 1- Blackspot seabream -tc2-BRT-Two-way Interaction: Relative influence of model terms calculated by the contribution of each term in reducing overall model deviance:

Table 3 Brief model summary statistics: Model 1- Blackspot seabream -tc2-BRT-Two-way Interaction: Morocco

Model 1-Blackspot seabream-tc2-BRT-Two-way Interaction

fitting final gbm model with a fixed number of 7300 trees for Blackspot seabream mean total deviance = 0.696 mean residual deviance = 0.52 estimated cv deviance = 0.534 ; se = 0.006 training data correlation = 0.505 cv correlation = 0.482 ; se = 0.006

Percentage of explained deviance 25%

RMSE 0.7438364

Variable Relative influence month 35 YEAR 27 indice_shannon 24 length_V 8 power 7

11

Table 4 Nominal and standardized CPUE using the BRT Two-way interactions: Morocco effort : Number of fishing days Morocco voracera fleet (as a proxy of effort unit)

YEAR Nom.CPUE Stan.CPUE effort Landing (kg)

2009 28.45 34.06 2122 89999 2010 28.21 29.32 2403 93590 2011 26.06 27.89 3134 113331 2012 23.26 23.79 3009 96615 2013 25.55 26.65 2409 81749 2014 25.91 25.95 3246 111395 2015 34.95 33.08 4590 211690 2016 27.72 25.60 4244 152514 2017 27.56 25.00 4969 179618 2018 19.24 19.41 2428 68630

Figure 4 Blackspot seabream -tc2-BRT-Two-way Interaction. QQ plots

12

Figure 5 Blackspot seabream -tc2-BRT-Two-way Interaction. Density functions for the BRT model residuals; values calculated using the default Gaussian kernel

Figure 6 Boxplots of actual log(CPUE+1) data versus model predictions, using the test dataset and train dataset; notches indicate robust estimates of the medians

-3 -2 -1 0 1 2 3

0.0

0.5

1.0

1.5

Model residuals

N = 22787 Bandwidth = 0.07939

De

nsity

13

Figure 7.a Comparisons of the standardized CPUE indices and nominal indices of Blackspot seabream from global models.

Figure 7.b Comparisons of effort and Landing (kg) of Blackspot seabream.

The Group praised the good work of the two teams confirming that the two models presented good fit to the data, and highlighted that in all examples presented there was no relevant difference in the trend of both nominal and standardised CPUE. They also noted some limitations of this second procedure due to the nature of the basic data from Morocco. One of the remarks was that it is not possible to consider positive and negative trips as it is based on sales notes with catch of P. bogaraveo, and therefore the effort is not representative of the total effort deployed by the fleet. The use of VMS data from Moroccan fleet should be tried, but it has to be noted that the frequency of the signal is

14

different from the Spanish one, posing problems to properly identify fishing operations. It was also confirmed that only the longline fleet could be treated, leaving apart the artisanal fleet which represent about 20% of catches. Finally the WG convened to work with the standardised indices separately and also requested to the experts to provide them on a quarterly basis to be input in GADGET. The WG also concluded that some way of standardizing the methodology and the data should be necessary in the near future.

Additional information regarding the Moroccan fleet, stemming from recent studies during the current year were presented by Meryeme Khoukh and Meryeme Benziane from INRH. An in depth analysis of the fleet that targets blackspot seabream in different ports of Northern Morocco showed that two different segments can be identified: longliners and artisanal vessels.

Moreover, recent investigations about the Moroccan fishing activity have noticed differences as well in the characteristics of the fishing gear, the fishing depth and the duration of the fishing operation for artisanal fleet and longliners. Table 5 summarizes the main characteristics of the two fleets.

The two fleets use at least two different but very selective fishing gears. The identified fishing gears are:

- The Bottom longline "Harita" used by the longliners in the port of Tanger. - "Harssya" or “Voracera” as the principal fishing gear used by the artisanal fleet,

and also a traditional hand line "Chambil" which became less common. The bottom longline used by the longliners can reach 2 km length. It holds, two longlines composed by a principal line of about 75-100 m. A number of secondary lines are attached on each principal line and carry hooks. The longliners use an average of 2200 hooks (880-2640) by trip. The reference of the hooks is Nº 3/0. Additionally, fishing depth of this fleet is usually 500 m (300-800 m).

On the other hand, the artisanal fleet fishing gear is composed by a principal line of up to 1 km length and which is wound on a manual winch. The principal line is connected to a main line which length is between 100-200 m, related to secondary lines holding hooks. The boats use an average of 280 (100-720) hooks by trip. The reference of hooks used is N°11. Moreover, the fishing depth of the artisanal fleet is between 200-600 m (mean: 400 m).

15

Table 5: Main characteristics of the two Moroccan fleets and fishing gears (average values).

Moroccan Fleet

LOA GRT HP Depth range (m)

Principal line length

(m)

Main line

Length (m)

Hooks per main line

Number of hooks per

boat

Hooks reference

Number of fishing

hours (Average)

Average

Longliners 14 23 181 300-800 2000 150-200 88/ Line (25 lines)

2200 Nº 3/0 10

Artisanal 4-6 <2T 15-45 200-600 1000 100-200 50-90/ line (4 lines)

280 N° 11 07

16

After discussion on the features of the fleets, the Working group decided that in order to better estimate the fishing effort and standardize the CPUEs, the two fleets will be considered separately (i.e.: longliners and artisanal fleet). For the case of Spain only one fleet targeting the different commercial categories will be considered. 3. Agreement on the assessments info to be used and data preparation (inputs) for

the different approaches (e.g. LCA-YpR, BioDyn, SPiCT, GADGET ) The biological sampling in Morocco has covered already 14 months with the support of CopeMed II what allowed the revision of the preliminary values of biological parameters obtained last March and presented at the SRC-WM in Sète, April 2019. FISAT and different routines in R were used to obtain the parameters Linf, K and t0 from the Von Bertalanffy equation of growth as well as a and b from the length-weight relationship.

In addition, the WG was informed that otoliths readings had been started in Morocco. A total of 105 otoliths have been prepared and digitalized using the software of NIS-Elements, during a training stay of Moroccan PhD student at the IEO of Cadiz (one week in September 2019). These otoliths have been read by only one lector and there is still a need to complete the lecture two more times implicating at least one more experienced lector. Moreover, the important number of the remaining otoliths (600 otoliths) will as well be prepared and the lecture of the rest of otoliths will be performed. The main outcomes of the analysis of the biology of the 947 specimens collected together with the existing information from Spain are showed in appendix III. The agreed values after a thorough discussion are summarised in tables 6 and 7 below.

Table 6: Von Bertalanffy growth parameters, a and b from the length-weigh relationship and M. The green row show the choice of values for further assessments.

Reference Area L K (y-1) t0 a b M

Sobrino & Gil, 2001a Gibraltar Strait 58.00 0.169 -0.67

Previous years Spain (Mark-recapture experiment)

Gibraltar Strait 62,0 0.14 -0,34 0,0087 3,14 0,2

Morocco and Spain combined

0,00867 3,14 0,2

Morroco (2005-2018) (elefan I) Gibraltar Strait 64.2 0.2 -0.68

Morocco and Spain combined Gibraltar Strait 65.84 0.2 - 0.68

0,2

Table 7: Length at maturity (L50, in cm) for the two sexes combined.

Country L50 (cm) sex combined

Spain 2003-2015 (only spawning season) 34,7

Spain + Morocco (only spawning season) 34,6

17

4. Review of previous assessment results (outputs): LCA and Y/R, Biodyn, SPICT

The results of previous assessments were presented and revised by the experts. Table 8 summarizes these previous results as obtained at the Benchmark meeting held in Sète, April 2019. Table 8 - Status of the blackspot seabream stock in the Strait of Gibraltar (S:Spain, M: Morocco)

Model Data series B/BMY F/FMSY BioDyn Aggregated S+M catches 1983-2018 and Spain CPUE

corrected with 0 catch (1995-2018) 0.14 -

BioDyn Aggregated S+M catches 1983-2018 and Spain CPUE corrected with 0 catch (1983-2018)

0.14 -

SPICT Aggregated S+M catches 1983-2018 and Spain CPUE all series corrected (VMS + corrected)

0.22 1.81

SPICT Aggregated S+M catches 1983-2018 and Spain CPUE corrected only 2000 - 2008 (+VMS)

0.18 1.99

LCA Aggregated S + M length frequency distributions (2016 - 2018)

- 2.13

Average 0.17 1.90

5. Conclusions and preparation of Stock Assessment Form (SAF) for the 2019 WGSAD Revision

With the new biological parameters obtained and new standardized CPUE indices, the WG agreed that new tests were going to be performed with all the previously used models: BioDyn, LCA Jones analysis, VPA(VIT) and GADGET between the end of this meeting (29 October) and the end of the forthcoming Benchmark in Rome planend for 14 December 2019.

The Stock Assessment Form partially elaborated as at the end of the current meeting is included as Appendix IV to this report. A final version will be produced after the Benchmark session is closed. CONCLUSIONS

Agreement was reached on:

A set of biological parameters of the species as is shown in tables 6 and 7.

Moroccan vessels should be split in two fleets: long-liners and artisanal boats.

Spanish data must be entered as a single fleet.

Standardized CPUEs indices with the same fleet segmentation as defined above, (if possible) and aggregated quarterly should be used as tuning fleet in GADGET in the absence of surveys data.

The group decided to perform new runs of GADGET, first using catches and effort series and then, as standardization is being completed, use CPUES as a tuning index. This was the list of tasks organised by priority to be completed (on-line) between the end of this meeting and the end of the Benchmark session next December 2019 in Rome.

18

PROPOSED WORKPLAN 1. Merge Spanish fleet in one (from 4 commercial categories) (1983-2018)

2. A new standardized CPUE index combining Spanish + Moroccan fleets (split in two: before and after 2013) will be obtained by quarter for the series of years 2009-2018 and will be provided to run GADGET before the WG in Rome.

3. Separate catches and effort of the Moroccan fleet in two fleets: artisanal and longliners (2001-2018)

4. Moroccan CPUEs should be divided in two fleets, include days with zero catches of P. bogaraveo and be provided by quarter. Consider the proportion of other species in the standardization process.

5. The group recommended to produce a map of fishing grounds in Morocco as the one already existing for Spain.

The Group adopted the conclusions and workplan established and agreed to continue working on-line until 6 December 2019. References Albeare, S.M., 2009. Comparisons of Boosted Regression Tree, GLM And GAM Performance

In The Standardization Of Yellowfin Tuna Catch-Rate Data From The Gulf Of Mexico Lonline Fishery 85.

De’ath, G., 2007. Boosted trees for ecological modeling and prediction. Ecology 88, 243–251. https://doi.org/10.1890/0012-9658(2007)88[243:BTFEMA]2.0.CO;2

Elith, J., Leathwick, J.R., Hastie, T., 2008. A working guide to boosted regression trees. J. Anim. Ecol. 77, 802–813. https://doi.org/10.1111/j.1365-2656.2008.01390.x

Hastie, T., Tibshirani, R., Friedman, J.H., 2009. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer-Verlag, New York 1, 337–387. https://doi.org/10.1007/b94608

19

Appendix I

CopeMed II Working Group on stock assessment of P. bogaraveo in the Strait of Gibraltar.

28-29 October 2019, Málaga, Spain Subdelegación del Gobierno. Paseo de Sancha 64. Sala Caleta

Annotated Agenda

Monday 28 October (9:00-18:00) 1. Background and objectives

2. Revision of identified issues raised in the last benchmark session (Sète, April 2019) and discussion on the available data for assessment (inputs):

2.1. CPUE standardization: presentation of Spanish and Morocco fleets exercises. Feasibility of increase the considered years backwards (before 2009).

2.2. Presentation of the most recent biological information from Moroccan samples (2018-2019) and its parameters estimates: sex ratio and length at maturity. Age and growth from otoliths readings (if available).

3. Agreement on the assessments info to be used and data preparation (inputs) for the different approaches (LCA-YpR, BioDyn, SPiCT, gadget…)

4. Review of assessment results (outputs): LCA and Y/R, Biodyn, SPICT.

Tuesday 29 October (9:00-18:00)

5. Consider (8) gadget models, taking into account the pending benchmark issues:

5.1. Merge the 4 Spanish fleets (by commercial categories) into a single fleet. So attempt models with two fleets (Morocco and Spain).

5.2. Linear fleets fishing effort input: allow the gadget model estimating these years fishing effort.

5.3. Try different Linfinity: fix value (i.e.: 62 cm) and let the model to estimate it. 5.4. Test different likelihood weights (1 and 5000) in the 2 linear fleet components (Morocco

and Spain).

6. Conclusions and preparation of Stock Assessment Form (SAF) for the 2019 WGSAD.

20

Appendix II

List of Participants

ALGERIA Tahar FILALI Centre National de Recherche et de Développement de la Pêche et l’Aquaculture 11 Bd Colonel Amirouche, Bou-Ismail w. de Tipaza Algeria Email: filalitahar@gmail.com Moussa MENNAD Centre National de Recherche et de Développement de la Pêche et l’Aquaculture 11 Bd Colonel Amirouche, Bou-Ismail w. de Tipaza Algeria Email: mennad.moussa@gmail.com MOROCCO Meryem BENZIANE Centre Régional INRH de Tanger Email: me.benziane@gmail.com Meryeme KHOUKH Ingénieur en Halieutique Institut Agronomique et Vétérinaire Hassan II, Rabat Email: khoukh.meryeme@gmail.com Mohamed SELFATI Centre Régional INRH de Nador Email: selfatimohamed@gmail.com Mansour SERGHINI Centre Régional INRH de Casablanca Email: serghini2002@yahoo.com

SPAIN José Luis PEREZ Instituto Español de Oceanografía Puerto Pesquero s/n 29640 Fuengirola, Spain Email: joseluis.perez@ma.ieo.es Juan GIL Instituto Español de Oceanografía Cadiz Email: juan.gil@ieo.es Lucía RUEDA Instituto Español de Oceanografía Cadiz Email: lucia.rueda@ieo.es Beatriz GUIJARRO Instituto Español de Oceanografía Palma de Mallorca Email: beatriz.guijarro@ieo.es FAO Pilar HERNÁNDEZ Fishery Expert Project CopeMed II Policy and Resources Division (FIA) The Food and Agriculture Organization of the United Nations (FAO) Paseo de Sancha 64, despacho 304 29016 Málaga, Spain Email: pilar.hernandez@fao.org

21

Appendix III summary of biological parameters obtained in the two countries

22

23

24

25

26

27

28

29

30

31

Appendix IV

Draft Stock assessment Form (reporting year 2019)

1. Basic Identification Data ................................................................................... 32 2. Stock identification and biological information ............................................... 32 2.1. Stock unit ........................................................................................................... 32 2.2. Growth and maturity ........................................................................................ 34 3. Fisheries information ........................................................................................ 37 3.1. Description of the fleet ...................................................................................... 37 3.2. Historical trends ................................................................................................ 38 3.3. Management regulations .................................................................................... 41 5. Ecological information ...................................................................................... 43 5.1. Protected species potentially affected by the fisheries ......................................... 43 6. Stock Assessment ............................................................................................... 44 6.1. Model assumptions ............................................................................................ 44 6.2. Input data and Parameters ............................................................................... 47

32

1. Basic Identification Data

Scientific name: Common name: ISCAAP Group: Pagellus bogaraveo Blackspot (=red)

seabream, dorade rose, besugo de la pinta, voraz,

الدنيس الوردي

33

1st Geographical sub-area:

2nd Geographical sub-area:

3rd Geographical sub-area:

GSA 01 GSA 03 1st Country 2nd Country 3rd Country

Spain Morocco Stock assessment method: (direct, indirect, combined, none)

1) "BioDyn" based on the Schaeffer production model; 2) Length Cohort Analysis (LCA); 3) Virtual Population Analysis (VPA) from VIT; 4) Length-based Bayesian

biomass estimator (LBB) and 5) Globally applicable Area-Disaggregated General Ecosystem Toolbox (gadget)

Authors: S. Benchoucha2, J. Gil Herrera1, M. Benziane2, L. Rueda1, M. Serghini2, B.T.

Elvarsson4 and P. Hernández5 Affiliation:

1Spanish Institute of Oceanography (IEO), Oceanographic Center of Cadiz. Spain 2National Institute of Fisheries Research (INRH), INRH-Tangier Center. Morocco 3Spanish Institute for of Oceanography (IEO), Oceanographic Center of Malaga. Spain 4Institute of Marine Research (HAFRO), Reykjavik. Iceland CopeMed II project

2. Stock identification and biological information

2.1. Stock unit Blackspot seabream (Pagellus bogaraveo) is found in the NE Atlantic, from South of Norway to Cape Blanc, in the Mediterranean Sea, and in the Azores, Madeira, and Canary Archipelagos (Desbrosses, 1938). Adults inhabit depths ranging around 300-700 m. The vertical distribution of this species varies according to individual size (Desbrosses, 1938; Guegen, 1974; Silva et al., 1994 and Gil, 2006).

This species is one of the most important commercial demersal species in the Strait of Gibraltar area. However, there is not much information available on the stock biology of P. bogaraveo in this narrow site. So, the usual way of stock separation is based in subareas boundaries that offer a better way of recording the available information. A project is currently being conducted by CopeMed II, the TransBoran Project a multidisciplinary approach aiming to study the identity and the boundaries of this stock. Migration patterns have been studied using tagging surveys in the GSA 01 Spanish Southern Mediterranean region and the Strait of Gibraltar area (Gil et al., 2001; Sobrino and Gil, 2001). Since 1997, 7066 individuals were tagged (juveniles + adults) and, at the moment, 545 recaptures were notified. Recaptures from juveniles showed displacements from GSA 01 nursery areas towards the Strait of Gibraltar fishing grounds. However, recaptures from

33

tagged adults did not reflect big displacements, which are limited to feeding movements among the different fishing grounds where the “voracera” fleets works (Gil, 2006). Six main fishing areas (Figure 2.1.1) were identified for the Spanish fleet based on the information provided by the Location and Track System for Andalusia Fishing Vessels (SLSEPA) of the Junta de Andalucía in the period August 2007-December 2009.

Figure 2.1.1- Main fishing grounds of the Spanish blackspot seabream fishery. Information from the Location and Track System for Andalusia Fishing Vessels (SLSEPA) of the Junta de Andalucía.

INRH experts identified the areas V-01, V-02, V-03, V-04 and V-06 as the main important fishing areas for the Moroccan fleets. Based on the available information the area for the joint assessment exercises is delimited around the Strait of Gibraltar, where 90% of the landings come from.

The following two main fishing areas (Figure 2.1.2) were identified in the Strait of Gibraltar area from the investigations with Moroccan fishermen: West of Cap Spartel to the East of Belyounech and Fnideq to Martil. Figure 2.1.2- Map of the main Moroccan fleet fishing grounds. The circles present the most important fishing grounds of the Moroccan longliners and artisanal fleet in the Strait of Gibraltar area.

34

Until now, there was a lack of information on the geographical distribution pattern and the stock (Atlantic and Mediterranean) boundaries of the blackspot seabream population fished in the Strait of Gibraltar.

The main landing ports in Morocco are Tangier, Dikky, Ksar Sghir, Fnideq, M’diq and Belyounech. The main landing ports in Spain are Conil, Tarifa and Algeciras (figure 2.1.3).

Figure 2.1-3. The main landing ports in Morocco and Spain.

2.2. Growth and maturity Blackspot seabream is a species belonging to the Sparidae family. They are bentho-pelagic species, inhabiting depth ranges from 300 to 700m throughout the eastern Atlantic and western Mediterranean. They are hermaphrodites, starting life as males but changing into females at 30 -35 cm, when they are 4 to 6 years old. They grow slowly to a maximum size of 70 cm, weight of 4kg and an age of about 15 years.

Biological parameters used in the assessments were taken from the previous studies because there is not new biological information available. Natural mortality was assumed constant (0.2) for all ages, length classes and years. Parameters estimates on the length-weight relationship (a and b) and the von Bertalanffy growth function (Linf, k and to) are presented in the Tables 2.2-1 and 2.2-1.The information on landings length distribution came from both countries (Spain and Morocco) sampling plan in the North and South region of Strait of Gibraltar. Sampling programs covered the two main landing ports, Tarifa (Spain) and Tangier (Morocco): total length of fish (TL) was measured to the nearest cm. To estimate the demographic structure of the whole catches, length frequency samples were raised to the total landing per fleet (and/or market category) and fishing region.

AlgesirasConil

Tarifa

35

Figure 2.2.1 presents the evolution of the mean length size in the landings in the Strait of Gibraltar area (GSA 01-Spain and GSA 03-Morocco) from 2005 onwards.

36

Table 2.2.2-1: Maximum size, size at first maturity and size at recruitment. Somatic magnitude measured

(LT, LC, etc) Total Length

(cm) References

Sex Fem Mal Combined Reproduction season January-June (Gil J., 2010)

Maximum size observed 62 (Gil J., 2010) Recruitment season

Size at first maturity ±35 ±30* 34.55

Spawning area Male and female values from Strait of Gibraltar area (Gil J., 2010). Combined sex from the CopeMed WK in Octobr 2019.

Recruitment size to the

fishery

Nursery area Shallower bottoms at both sides of the Strait of Gibraltar, mostly Mediterranean one

Table 2.2-2.2: M vector and proportion of matures by size or age (Combined). *ICES WGDEEP Report 2008

Table 2.2.2-3: Growth and length weight model parameters Sex

Units female male Combined Years

Growth model

L∞ cm 65.84

K Year-1 0.2 t0 year -0.68

Data source CopeMed WK, Málaga Octobr 2019

Length weight relationship

a 0.00867 b 3.14

M

(scalar) 0.2

sex ratio

(% females/total) Hermaphrodite

Age Natural mortality* Natural Mortality for VIT Proportion of matures

0 0.2 0.43

0.020

1 0.2 0.28 0.13

2 0.2 0.23 0.49

3 0.2 0.2 0.84

4 0.2 0.18397 0.98

5 0.2 0.165056 0.99

6 0.2 0.17 1.000

7 0.2 0.15615 1.000

8+ 0.2 0.157814 1.000

37

3. Fisheries information

3.1. Description of the fleet Blackspot seabream is one of the principal demersal species targeted in the Strait of Gibraltar for its high commercial value compared to others demersal resources. The fishing hook gears used are known as ‘‘voracera’’ in both countries involved in the fishery (Morocco and Spain).

Spanish fleet: The Spanish fishery targeting blackspot seabream has been developing along the Strait of Gibraltar area (Gil et al., 2000) since the earliest 1980´s. It is almost a mono-specific fishery, with one clear target species which represents the 74% from the total landed species, which constitutes a metier by itself (Silva et al., 2002). The “voracera”, a local mechanized hook line baited with sardine, is the gear used by the fleet from Tarifa and Algeciras ports (see Figure 2.1.1). Fishing is carried out taking advantage of the turnover of the tides in bottoms from 200 to 400 fathoms. The number of hooks by boat is between 200 and 2000. Every boat can only use a maximum of 30 lines per day (each line attached a maximum of 100 hooks, usually ±70) with a maximum legal length of 120 m. The legal dimensions of the hooks are a minimum length of 3.95 ± 0.39 cm and a minimum width of 1.4 ± 0.14 cm. Number of boats decrease in the last years and its mean technical characteristics are: Length= 9.80 meters, GRT= 6.36 and HP= 47.23. Moroccan fleet:

The most important Moroccan fleets targeting blackspot seabream are the longliners mainly based at the port of Tangier and the artisanal fleet of the Strait of Gibraltar area. In the past years, the longliners fleet was more or less stable (78 to 101 vessels). The number of the longliners fleet in 2017 was approximately 94 and 145 artisanal boats. The fishery is carried out at 200-700 m depth and the gear used is the longline known as “voracera”. The number of hooks by boat is between 200 and 2000 and the size of the hooks is between 8 and 11. The blackspot seabream is not the first target species in Moroccan longliners and artisanal fishery. It represents between 18% to 42% in weight and 45 to 56% in commercial value of to the total catches provided by this fleet: the first specie landed by the longliners is Lepidopus caudatus. Some artisanal boats are targeting Pagellus bogaraveo in the Strait of Gibraltar. The mean annual catch of Pagellus bogaraveo in the artisanal feet is about 17 tons. Table 3.1.1-1: Description of operational units exploiting the stock

Country GSA Fleet Segment Fishing Gear Class Group of Target

Species Species

Operational Unit 1* ESP GSA 01 Artisanal Handlines

(“voracera”) Demersal shelf

species Blackspot seabream

Operational Unit 2 MAR GSA 03 Longliners and

artisanal Longlines

(“voracera”) Demersal shelf

species Blackspot seabream

Table 3.1.1-2: Catch, bycatch, discards and effort by operational unit in the reference

38

year

Operational Units*

Fleet (n° of

boats)*

Catch (T or kg

of the species

assessed)

Other species caught (names and weight )

Discards (species

assessed)

Discards (other species

caught) Effort (units)

ESP 01 -HL 54 (2017)

8 T (2018)

Lepidopus caudatus,

Helicolenus dactylopterus, Brama brama, Trachurus spp,

Phycis spp., Polyprion

americanus Thunnus thynnus*

Negligible Unknown

(Fishing days) 1062

(2018)

MAR 03 -LL

94 (2017)

188t (2017)

Lepidopus caudatus Helicolenus dactylopterus Scorpaena scrofa Conger conger Polyprion americanus Xiphias gladius Others

Negligible Unknown

(Fishing days) 3180

(2018)

Total 148 231 4242 *Same boats but different gear (not “voracera” one)

3.2. Historical trends Fishery Information about the Spanish landings were compiled from the two main ports (Tarifa and Algeciras) where Pagellus bogaraveo was landed from 1983 to 2018 (Figure 3.2.1). Landings are distributed in 4 different commercial categories, owing to the wide range of sizes and for market reasons. The trend of the catches shows a big decline in the Spanish fishery, from 700 tons in 2009 to 130 tons in 2013 and only 8 tons in 2018, however this value could be due to a problem of declaration (Figure 3.2-1).

Catches from the Moroccan fisheries were low at the beginning to remain more or less stables for the whole series (Figure 3.2.1). From 2013 onwards it showed an increasing trend setting the highest value in 2015 with 219 tons and decreasing to 68 tons in 2018. The 2010-2018 mean production of this fishing resource is about 276 tons (Table 3.2-2).

Table 3.2.2. Catch (t) and effort (days) in the two countries and combined for 2005-2018

39

GSA _01_ton (Spain) f TOT (days)(Spain) cpue (k/days)_Spain

2005 333.82 4856 68.74 2006 350.42 4955 70.72 2007 365.57 7064 51.75 2008 420.19 8041 52.26 2009 583.16 8892 65.58 2010 369.69 6932 53.33 2011 243.85 5659 43.09 2012 130.03 3638 35.74 2013 70.18 2222 31.59 2014 141.65 3527 40.16 2015 171.90 3384 50.80 2016 103.76 2418 42.91 2017 43.00 1308 32.87 2018 9.00 429 20.98

GSA _03_tons

(Morocco) f (Fishing

days)_Morocco cpue

(k/days)_Morocco 2005 38.95 984 39.59 2006 74.36 1904 39.06 2007 89.08 2560 34.80 2008 75.61 2206 34.27 2009 98.51 2454 40.14 2010 105.27 2786 37.79 2011 136.18 4038 33.72 2012 121.98 3858 31.62 2013 92.43 2820 32.78 2014 117.78 3593 32.78 2015 218.97 4994 43.85 2016 159.03 4614 34.47 2017 187.72 5359 35.03 2018 68.00 3180 21.38

GSAs 1+3 tons f sp+mo CPUE_GSA1-3

2005 372.78 5840.00 63.83 2006 424.78 6859.00 61.93 2007 454.65 9624.00 47.24 2008 495.80 10247.00 48.39 2009 681.67 11346.00 60.08 2010 474.97 9718.00 48.87 2011 380.02 9697.00 39.19

40

2012 252.01 7496.00 33.62 2013 162.61 5042.00 32.25 2014 259.43 7120.00 36.44 2015 390.86 8378.00 46.65 2016 262.79 7032.00 37.37 2017 230.72 6667.00 34.61 2018 77.00 3609.00 21.34

At the start of the series Spanish fishing effort was very high in comparison with the Moroccan. It was about 9000 fishing days in 2009 and declined since 2010 and reached the same level of the Moroccan one in 2014. Moroccan fishing effort has increased and became highest than Spanish one in the last two years (Figure 3.2-1). The Spanish CPUE was high in 2005 (69 kg/fd) and decrease gradually to 21 kg/fd in 2018 but remain highest than Moroccan one for the whole serie except in 2017 where both CPUEs form Morocco and Spain were quite similar. The CPUE for Morocco is stable for the whole serie with small fluctuations but decreases in 2018 with 21 kgs/fd (Figure 3.2-1c).

(a)

(b)

41

(c) Figure 3.2-1: Landings (a), fishing effort (b) and CPUE (c) of Pagellus bogaraveo in GSAs 01 and 03 (Strait of Gibraltar area).

Figure 3.2.2 - Landings and effort by fleet on Pagellus bogaraveo (1983-2017) in the Strait of Gibraltar area (GSAs 01 and 03 - historical series).

3.3. Management regulations Spain (GSA01): A management plan for this fishery was established by the AAA/1589/2012 Order of July 17, establishing a plan for the blackspot seabream fishery in certain areas of the Strait of Gibraltar regulating the area, gear (“voracera”) and the fleet. This plan includes an authorized “voracera” fleet, fishing gear technical characteristics (that was stated above), a seasonal fishery closure between February 1st and March 31st and the regulation of the effort by week. Minimum landing size and the annual Total Allowable Catch (TAC) are related to the EU Regulation a minimum size for blackspot seabream of 33 cm (Total length) currently applies in the Mediterranean and also in the North-East Atlantic since May 2018[Commission Implementing Regulation (EU) 2017/787of 8 May 2017establishing a minimum conservation reference size for red (blackspot) seabream in

42

the North-East Atlantic Ocean].

Morocco (GSA03): The main regulations enforced by Morocco are: the freeze of investment since 1992; the interdiction of fishing beyond 80 m depth in the area between Tangier and Al Hoceima and below 3 miles in the area between Al Hoceima and Saidia., the minimal landing size (25 cm Fork Length, about 28 cm Total Length); trawls mesh size ≥ 50 mm; nets regulations (L = 1000 m, mesh size = 70 mm) and, the protection of areas (marine protected areas) and anti-trawling artificial reefs.

43

4. Fisheries independent information

None

5. Ecological information

5.1. Protected species potentially affected by the fisheries Not relevant for the case of the blackspot seabream fishery of the Strait of Gibraltar, because the fishery do not interact with these kind of species. Anyway the table below shows the list of species which occur in the area included in several protection agreements (Ocaña et al., 2010).

Phylum Species Protection Agreement Chordata Polyprion americanus RL: DD Thunnus thynnus RL: DD /OSPAR: V/ UNCLOS: YES /

BARCOM: III Pagrus pagrus RL: EN Xiphias gladius RL: DD / UNCLOS: YES/ BARCOM:

III Galeorhinus galeus RL: VU / UNCLOS: YES Isurus oxyrhinchus RL: VU / CMS: II / BERN: II /

UNCLOS: YES/ BARCOM: III Cnidaria Caryophyllia spp. CITES: II Lophelia pertusa CITES: II / OSPAR: All Dendrophyllia cornigera CITES: II Dendrophyllia ramea CITES: II Madrepora oculata CITES: II Errina aspera CITES: II / BERN: II (Med.) /

BARCOM: II Echinodermata Ophidiaster ophidianus BERN: II (Med.) / BARCOM: II Paracentrotus lividus BERN: III / BARCOM: III Mollusca Charonia lampas BERN: II / BARCOM: II Ranella olearia BERN: II (Med.) / BARCOM: II Porifera Axinella polypoides BARCOM: II RL: IUCN Red List of Threatened Species: EN (Endangered), VU (Vulnerable), DD (Data Deficient) CITES: Convention on International Trade in Endangered Species of Wild Fauna and Flora: Appendix OSPAR: Convention for the Protection of the marine Environment of the North-East Atlantic: Annex UNCLOS: United Nations Convention on the Law of the Sea - Annex I (highly migratory species) BARCOM: Convention for the Protection of the Marine Environment and the Coastal Region of the Mediterranean (Barcelona Convention): Annex BERN: Convention on the Conservation of European Wildlife and Natural Habitats (Bern Convention): Appendix CMS: Convention on Migratory Species: Appendix

Environmental indexes None. However, the special features of the Strait of Gibraltar area as well as environmental parameters that could affect the stock abundance or the gear catchability might be taken into consideration (i.e. currents´ strength).

44

6. Stock Assessment Five different methods (analytical and global assessments) to evaluate the current status of the stock were attempted to compare the results obtained using the joint data from Morocco and Spain.

The first approach was carried out using a global model "Biodyn" based on the Schaeffer production model. The second model used is a Length Cohort Analysis (LCA) to estimate ad hoc reference points (FMAX and F0.1), Fcurr and a Yield per Recruit, both models developed in excel sheets by Pedro Barros.

The third one is a Virtual Population Analysis (VPA) based on VIT software and NOAA to produce the Yield per recruit analysis and to estimate ad hoc reference points (Fcurr and F0.1). The fourth one is a length-based Bayesian biomass estimator (LBB).

And finally the fifth one was the Globally applicable Area Disaggregated General Ecosystem Toolbox (GADGET).

6.1. Model assumptions For the Biodyn, data and initial parameter estimates should be entered only in the cells colored green. All other cells are either not used, or used to calculate quantities used by the model. Data must be entered for all the data columns colored green, and also for initial values of the parameters. Additionally, the model control settings may be entered (in the cells colored orange. If these control settings are not changed, they may be left at their default values. The non-linear estimation procedures suffer from a number of limitations, of which the most important is probably that the estimates obtained will depend on the start values defined. Therefore, one should try to keep the number of parameters to be estimated non-linearly to the minimum possible values. As a minimum, one must estimate r and K by fitting the model to the data using the solver algorithm. When defining the parameters to estimate, one should as much as possible set constraints (maximum and minimum values) so that the algorithm is limited to reasonable values, defined by the researchers. Use the spreadsheet area of Minimum and Maximum values to define these. For the LCA and Yield per Recruit, the analysis of sizes cohorts (LCA) (Jones, 1984) were used to estimate F current and the exploitation scheme of the fishery in the last years. The analysis of Y/R based on sizes were then used to estimate the reference biological points (BRP) FMAX and F0.1. In VPA, the stock is considered to be composed of several annual cohorts and every cohort of the stock is analyzed and followed separately. It is based on backward calculations through time and ages given knowledge of all ages in the last year and the last age group in all years; by adding the number of individuals lost by fishing and natural mortality during a year to the number of individuals at the end of the year to estimate the number of individuals at the beginning of the year. Virtual Population Analysis (VPA) was attempted using the VIT software (Lleonart and Salat, 1992). VIT is a program created for the analysis of fisheries where information is limited. VIT program was designed to analyze exploited marine populations based on catch data, structured by ages or sizes, from one or several gears. The main assumption is that of the steady state (equilibrium conditions) because the program works with pseudo-

45

cohorts, therefore it is not suitable for historic series. From the catch data with some auxiliary parameters and using VPA, the program rebuilds the population and mortality vectors. After this first step, the user has several analysis tools and reporting options available: obtaining comprehensive VPA results, Yield-per-Recruit analyses based on the fishing mortality (F) vector, analyses of sensitivity to parameters inputs, and transition analyses - outside the equilibrium - due to changes in the pattern of exploitation or recruitment. The stock size estimates, which include recruitment estimates for every year, can be used for a yield per recruit analysis. The use of this software is only recommended when the model is applied to short time series of consecutive annual data and the resulting variation in the estimated stock parameters appears reasonably low (Ratz et al., 2010). Analytical assessment (VPA) requires catch at age numbers. Lengths distributions were transformed into ages by the “slicing technique” implemented in the VIT software. GADGET (Globally applicable Area Disaggregated General Ecosystem Toolbox) is a statistical model of marine ecosystems: it is a forward simulation where the processes are usually modeled/structured dependent on length (but also age can be tracked). In summary, gadget has essentially three components:

1. an ecosystem simulator, 2. a likelihood function that takes the output (from the ecosystem simulator) and

compares the data, 3. and a function minimize (optimization routines to find the best set of the model

parameters values)

Standardization of the CPUE (Catch per Unit of Effort) Catch per Unit of Effort (CPUE) was standardized for the Spanish and Moroccan fleet in order to provide an index of stock abundance for the stock assessment models.

For the Spanish fleet, the CPUE index (2009-2018) was computed as the total catches per day and vessel divided by the total duration of the fishing hauls conducted by the vessel that day. Information on total catches was provided by the authorities from the daily sales in the local fish auctions. Duration of the fishing hauls was estimated from the analysis of the vessel monitoring system (VMS) data. The use of VMS data allowed for taking into account those fishing days with zero catches of blackspot seabream. Preliminary fleet analysis was conducted to remove data from vessels that did not meet minimum criteria to be considered as part of the blackspot seabream fishery fleet. In addition, a spatial analysis of the VMS signals was conducted to identify the major fishing areas for blackspot seabream in the Strait of Gibraltar. Several Generalized Linear Mixed Models (GLMM) for CPUE standardization were attempted including as explanatory variables the temporal component (quarter of the year or year), spatial component (fishing area) and technical characteristics (length of the fishing vessels). The boat was included as a random effect. Different models provided similar results showing an oscillating decreasing trend in the CPUE, which showed the highest values at the beginning of the series (2009) and decreasing afterwards with a slight increase between mid 2014 to mid 2016 and dropping in the last years to values close to zero (Figure 6.1).

46

Figure 6.1.1 Standardized CPUE for the Spanish blackspot seabream fishery in the Strait of Gibraltar. For Morocco the commercial data of longline fishery by their corresponding port were compiled with catches by species (in kg), registration numbers, and date of landing for the period 2009 to 2018. The daily sales of each species for one trip were added together by vessel and by port to built a data matrix with daily sales, registration number, port, date and catch per species as variables. According to the survey conducted in each port, generally a sale of the catch landed by one vessel corresponds to one trip (in general sales are one days). Finally, a new matrix was created with rows correspondent to the different trips. The effort unit is the fishing trip, (one day). Since in the Moroccan data, we do not have the information that indicates that the zero fishing trips of the Blackspot seabream have the same probability of catching the different species landed, therefore, the CPUEs are calculated based on the positive fishing trips of the Blackspot seabream. the explanatory variables selected in the model are the power, month, year, length and Shannon index

the choice of the Shannon index is justified better than the integration of species as variables. Blackspot seabream catch, vessel power and length overall, and Shannon index were log(x+1) transformed for use in the final statistical model using Boosted Regression Trees (BRT).

LOG(Blackspot seabream (kg)+1) ~ year + month + LOG(power+1) + LOG(length +1) + LOG(Shannon index +1).

The yearly evolution of the standardized CPUE follows the same trend than the nominal CPUEs, indicating a slight declining trend since the early years and slight increase from 2012 until 2015 and declining trend again in the last three years (Figure 6.1.2) The catch and the effort also show parallels trends between them, differing from CPUE in the early years, when a slight increase is observed, (Figure 6.1.3).

47

Figure 6.1.2. Comparisons of the standardized CPUE indices and nominal indices of Blackspot seabream from Moroccan data.

Figure 6.1.3 Comparisons of effort and Landing (kg) of Blackspot seabream in Morocco

This Stock Assessment Form will be completed after the Benchmark Assessment is closed and uploaded to the GFCM sharepoint in due time.

Project FAO CopeMed IISubdelegación del Gobierno

Paseo Sancha 64 29016 Malaga - Spain Tel: +34952989299 copemed@fao.org

www.faocopemed.org