Stock Assessment Form of Hake (M. Merluccius) in Combined GSA

Stock Assessment Form of Hake

(M. merluccius) in combined

GSA 12-16 Reference Year: 2007-2016

Reporting Year: 2017

Merluccius merluccius, is an important demersal target species of the commercial fisheries in the central Mediterranean Sea (GFCM-GSA 12-16). In this area, European hake is fished by 5 fishing fleet components: Italian coastal trawlers, Italian distant trawlers, Tunisian trawlers, Maltese trawlers. Tunisian and Italian artisanal vessel and Annual landings of hake for 2013-2016 was about 1400 t from Italian about 2000 t Tunisian trawlers. Hake is the main commercial by-catch species of deep water shrimp fisheries and a target species for vessels using longlines and gillnets. Trawlers catching hake exploit a highly diversified species assemblage, the main other commercial species being deep water rose shrimp (Parapenaeus longirostris, generally main target species), catches range mostly between 8 and 70 cm total length (TL). The minimum landing size is 20 cm total length (TL). The assessment, performed by extended survivor analysis is ( XSA) integrated with trawl survey data, shown a state of overfishing with high level of relative stock abundance. The status of the stock assessment was confirmed also by a Gadget assessment.

1 Stock Assessment Form version 1.0 (January 2014) Uploader: Widien Khoufi

Stock assessment form

Contents 1 Basic Identification Data...... Error! Bookmark not defined. 2 Stock identification and biological information ...... 5 2.1 Stock unit ...... 5 2.2 Growth and maturity ...... 5 3 Fisheries information ...... 8 3.1 Description of the fleet ...... 8 3.2 Historical trends ...... 10 3.3 Management regulations ...... 19 3.4 Reference points ...... 21 4 Fisheries independent information ...... 22 4.1 MEDITS Trawl Survey ...... 22 4.1.1 Brief description of the direct method used ...... 22 4.1.2 Spatial distribution of the resources ...... 30 4.1.3 Historical trends ...... 30 5 Stock Assessment...... 34 5.1 Extended Survivor Analysis (XSA) ...... 34 5.1.1 Model assumptions ...... 34 5.1.2 Scripts ...... 35 5.1.3 Input data and parameters ...... 35 5.1.4 Results ...... 37 5.1.5 Robustness analysis ...... 34 5.1.6 Retrospective analysis ...... 35 5.1.7 Assessment quality ...... 36 5.2 Alternative models: GADGET ...... 36 5.2.1 Input data and Parameters (Gadget)...... 37 6 Stock predictions ...... 45 6.1 Short term predictions (XSA) ...... 45 7 Draft scientific advice ...... 46 7.1 Explanation of codes ...... 47

1 Basic Identification Data

Scientific name: Common name: ISCAAP Group: Merluccius merluccius [European hake] [HKE] 1st Geographical sub-area: 2nd Geographical sub-area: 3rd Geographical sub-area: [GSA_12] [GSA_13] [GSA_14]

4th Geographical sub-area: 5th Geographical sub-area: 6th Geographical sub-area:

[GSA_15] [GSA_16]

1st Country 2nd Country 3rd Country [TUNISIA] [ITALY] [MALTA]

4th Country 5th Country 6th Country

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

Indirect Method: XSA analysis

Authors: 1 2 2 1 2 4 W. Khoufi , V. Gancitano , G. Milisensa , S. Ben Meriem , F. Colloca , E. Arneri , L. 4 Ceriola ,

1 3 2 O. Jarboui , R. Micallef , F. Fiorentino

Affiliation: 1Institut National des Sciences et Technologies de la Mer (INSTM), Tunisia; 2IAMC-CNR Mazara del Vallo, Italy; 3MSDEC, Malta, 4Food and Agriculture Organization of the United Nations (FAO), MedSudMed Project, Viale delle Terme di Caracalla, 00153 Rome, Italy.

The ISSCAAP code is assigned according to the FAO 'International Standard Statistical Classification for Aquatic Animals and Plants' (ISSCAAP) which divides commercial species into 50 groups on the basis of their taxonomic, ecological and economic characteristics. This can be provided by the GFCM secretariat if needed. A list of groups can be found here: http://www.fao.org/fishery/collection/asfis/en

Direct methods (you can choose more than one):

- Acoustics survey

- Egg production survey

- Trawl survey

- SURBA

- Other (please specify)

Indirect method (you can choose more than one):

- ICA

- VPA - LCA

- AMCI

- XSA

- Biomass models

- Length based models

- Other (please specify)

Combined method: you can choose both a direct and an indirect method and the name of the combined method (please specify)

2 Stock identification and biological information 2.1 Stock unit The European hake (Merluccius merluccius, L. 1758) has a wide distribution from the coastal grounds to the high depth (up to 1000m), the larger individuals preferring the deepest waters. The stock structure of hake in the Strait of Sicily (south-central Mediterranean Sea) is still under investigation. Levi et al., (1994) compared the growth of M. merluccius from several Mediterranean areas and found a similar pattern in individuals from the northern side of the Strait of Sicily (GSA15 and 16) and those caught in the Gulf of Gabès (GSA 14). Ben Meriem and Gharbi (1996) indicated a similar historical trend in CPUE in north and south Tunisia (GSA 12, 13 and 14). Lo Brutto et al., (1998) have also found no evidence of genetic subdivisions or significant differences in allelic frequencies, between samples near Sicily and those from the north Tunisia. More recently, Levi et al., (2004) applied electrophoretic, morphometric and growth analyses to test the hypothesis of the existence of a unique stock of hake in the south-central Mediterranean, which includes part of the North African continental shelf off the Tunisian coast and the shelf off the southern Sicilian coast. Milano et al., (2014), working with genetic markers able to identify fine population structure (SNPs), found a main difference between hake inhabiting Western, Central and Eastern Mediterranean. According to the multidimensional definition of stock followed in the STOCKMED project (Fiorentino et al., 2015), European hake of the Strait of Sicily would belong to a single a population unit, extending from the Ligurian-North Tyrrhenian Sea (GSA 9) to eastern Ionian (GSA19).

2.2 Growth and maturity M. merluccius is considered a long living species reaching at least 25 year old (Vitale et al., 2016). Hake growth is still a matter of active debate in fishery biology, with some Authors proposing the so called low growth and other the fast growth. On the basis of a preliminary validation of age at size with bomb radiocarbon analysis (Vitale et al., 2016), the so-called low growth hypothesis is considered as the most reliable for the species in the south-central Mediterranean.

Table 2.2-1: Maximum size, size at first maturity and size at recruitment.

Somatic magnitude measured Units (TL in cm)

Reproduction season all the year round

Sex Fem Mal Combined peak winter

Maximum Recruitment season size observed 90 58 All years

Size at first 28.35 Spawning area On outer shelf and maturity upper slope

Recruitment size to the fishery 8 Nursery area Identified

Table 2-2.2: M vector from Prodbiom (Abella et al., 2009) and proportion of matures by age (sex combined).

Size/A Natural mortality Proportion of matures

0 1.38 0.01

1 0.56 0.16

2 0.27 0.61

3 0.22 0.93

4 0.19 0.98

5 0.18 1.00

6 0.17 1.00

7 0.16 1.00

8 0.16 1.00

Table 2-3: Growth and length weight model parameters.

Sex

Units female male Combined Years

L∞ cm 100

K 0.116 Growth model

t0 -0.6

Data source Tunisian and Italian data

Length weight a 0.004 relationship b 3.15

M 0.28

(scalar)

sex ratio (% females/total) 0.48

3 Fisheries information

3.1 Description of the fleet European hake, Merluccius merluccius, is an important demersal species for commercial fisheries in the Strait of Sicily (GFCM-GSAs12-16, south-central Mediterranean Sea). It is the main accessory species of trawling targeting deep water rose shrimp as well as an important target species for artisanal vessels using longlines and gillnets on the outer shelf and upper slope. In the south-central Mediterranean, European hake is exploited by 6 main fishing fleet segments: Italian coastal trawlers, Italian distant trawlers, Tunisian trawlers, Maltese trawlers, Italian and Tunisian vessels using fixed nets. Sicilian coastal trawlers (length overall, LOA, between 12 and 24 m) operate mainly on short-distance fishing trips, which range from 1 to 2 days at sea, and fishing taking place on the outer shelf and upper slope. With 290 registered vessels, this is the largest component of the fleet fishing deep water rose shrimp in 2015. Sicilian trawlers over 24 m in length perform longer fishing trips, which may last up to 4 weeks. These vessels operate offshore, in both Italian and international waters of the south-central Mediterranean. Due to the progressive shift in target species of such trawlers from deep water rose shrimps and Norway lobster to red shrimps, about 75 out 114 can be considered as targeted on P. longirostris in 2015. In Maltese Islands small vessels measuring 12 to 24 m in length target rose shrimp very close to land (around 6 km from the coast) at a depth of around 200 m. The number of trawlers targeting rose shrimp decreased from 13 in 2011 to 10 in 2015. The activity is mainly carried out in winter, when the weather does not allow to fish in deeper waters. Tunisian trawl vessels which target rose shrimp measure around 24 m in length, and operate primarily in Northern Tunisia where 90% of the country’s total P. longirostris catches originate. The great majority of these catches are landed in the town of Bizerte and Kelibia. The number of Tunisian trawlers targeting rose shrimp has increased from 40 in 1996 to around 71 in 2015.

Table 3-1: Description of operational units exploiting the stock.

Group of Target Species Country GSA Fleet Segment Fishing Gear Species Class

Operational E - Trawl (12-24 34 - Demersal slope species metres) Unit 1* ITA 16 03 - Trawls HKE

Operational F - Trawl (>24 metres) 34 - Demersal slope species Unit 2 ITA 16 03 - Trawls HKE

Operational F - Trawl (>24 metres) 34 - Demersal slope species Unit 3 TUN 12-13-14 03 - Trawls HKE

Operational H - Gillnet (12-24 metres) Unit 4 TUN 12-13-14 07-Nets Demersal species HKE

Operational E - Trawl (12-24 Unit 5 34 - Demersal slope species MLT 15 metres 03 - Trawls HKE

Table 3.1-2: Catch and effort by operational unit in 2016. Hake catch Other species Discards (tons) (other Fleet (n° caught Hake discards species Effort of (names and caught) (units) Operational Units* boats)* (tons) weight )

ITA 99 E 03 34 - HKE 290 913 tons

ITA 99 F 03 34 - HKE 123 288.9 134.2 tons

TUN 99 F 03 34 - HKE 1439.4 tons

TUN 99 H 07 - HKE 72 tons

MLT 99 E03 34-HKE 10 20.7 tons

Total 2734 134.2 tons

3.2 Historical trends Time series analysis with figures showing the observed trends in landings and LFD.

Figure 3.2.1 - Catch of HKE from 2007 to 2016 in the Strait of Sicily, Central Mediterranean (GSA 12,13,14,15 and 16) by fleet (Italian coastal and distant trawlers, Tunisian trawlers and gill nets, Maltese coastal trawlers) and country.

Figure 3.2.2 – Fishing effort from 2004 to 2015 in the Strait of Sicily, Central Mediterranean (GSA 15 and 16) by fleet distinguished by country and LOA.

2000 ITA_Distant_mean_2010- 2016 25

1800 TUN_Trawl_mean_2010- 2016 ITA_Coast_mean_2010- 2016 1600 TUN_Net_mean_2010- 2016 20

1400 MLT_Trawl_mean_2010- 2016

1200 15

1000

800 10

600

400 5

200

0 0

5 10 15 20 25 30 35 40 45 50 55 60 65 70 TL (c Figure 3.2.4 – Length frequency distributions mean (LFD) by fleet and country, combined sex.

3.3 Management regulations In the south-central Mediterranean, as in other Mediterranean areas, fishery management is based on control of capacity (number of fishing licenses), fishing effort (days at sea, number of vessels), and technical measures (cod-end mesh size for trawlers, temporary or permanent area closures and fish-size limits). . In Italian (south Sicily) waters, a medium term management plan for 2008-2013 has been agreed for Italian trawlers in the Strait of Sicily. Italian Management Fishery Plans (IFMP) was mainly based on a fleet reduction of 25% of the capacity obtained in two steps. The first (12.5%) from 2008 to 2010, and the second (12.5%) from 2011 to 2013. A trawling ban of 30 day per year is adopted, normally in late summer early autumn. In addition, the Mediterranean Regulation EC 1967 of 21 December 2006 fixed a minimum harvest size of 20 cm TL and a minimum mesh size of 40 mm square or 50 mm diamond for EU bottom trawling vessels (i.e. Italian and Maltese trawlers). In 2015, Malta had 14 trawlers that operated on a full-time basis. A preliminary analysis of the capacity of the fleet in the Sub-regional Committee report for the Central Mediterranean (SRC_CM) showed that there was a 39% reduction from 2011 to 2015 (7 permanent and 2 temporary). Fishing effort and capacity in the 25 nautical miles fisheries management zone are being managed by limiting vessel sizes, as well as total vessel engine powers (EC 813/2004; EC 1967/2006). Trawling is allowed within this designated conservation area, however only by vessels not exceeding an overall length of 24m and only within designated areas. Such vessels fishing in the management zone hold a special fishing permit in accordance with

19 Article 7 of Regulation (EC) No 1224/2009, and are included in a list containing their external marking and vessel's Community fleet register number (CFR) to be provided to the Commission annually by the Member States concerned (EC 813/2004).

In Tunisia, fisheries management targets trawlers generally setting a number of measures affecting also hake related fisheries. Trawling is not permitted within 3 nautical miles of the coast and at less than 50m depth in GSAs 12-14. Moreover, in GSA 14, a three-month closed season for trawling (from July to September) is in place. The objective of the measure is to protect recruits of a large number of species. In addition, a trawling ban of 11 months is in place in the Gulf of Tunis (GSA 12) and trawling in this area is allowed only during July. Also, minimum landing size of 20 cm total length in Tunisia has been established. The GFCM in 2015 adopted a recommendation to protect juvenile European hake in the entire Mediterranean Sea (REC.CM-GFCM/40/2016/5 establishing a minimum conservation reference size for European hake in the Mediterranean Sea). In addition, owing to the importance of the deep water rose shrimp fishery for the costal countries in the, area the CFGM adopted a series of recommendations targeting deep water rose shrimp and hake fisheries: i) REC.CM-GFCM/39/2015/2 on the establishment of a set of minimum standards for bottom trawling fisheries of demersal stocks in the Strait of Sicily, pending the development and adoption of a multiannual management plan; ii) REC.CM-GFCM/40/2016/4 establishing a multiannual management plan for the fisheries exploiting European hake and deep- water rose shrimp in the Strait of Sicily (GSA 12 to 16); and iii) REC.CM-GFCM/41/2017/8 on an international joint inspection and surveillance scheme outside the waters under national jurisdiction of the Strait of Sicily (geographical subareas 12 to 16). In order to protect juveniles, three Fishery Restricted Areas (FRAs) where bottom trawling is prohibited were established in the northern sector of the Strait of Sicily (GSA 15 and 16) based on the REC.CM-GFCM/40/2016/4. Periodic scientific studies assessing the effectiveness of the measure. More studies should be carried out to identify nursery areas of hake in GSAs other than 15 and 16, in order to evaluate the possibility of proposing additional FRAs to protect nursery areas throughout the subregion.

20 3.4 Reference points

Table 3.4-1: List of reference points and empirical reference values in 2016.

Limit Target Reference Reference point/empirical point/empirical Indicator reference value Value reference value Value Comments

SSB

F F0.1 0.2

CPUE

Index of Biomass at sea

21 4 Fisheries independent information

4.1 MEDITS Trawl Survey In order to collect fisheries independent data, which is a requirement of the EU DCF (Council Regulation 199/2008, Commission Regulation 665/2008, Commission Decision EC 949/2008 and Commission Decision 93/2010), the MEDITS international trawl survey is carried out in GSAs 15 & 16 on an annual basis. In July 2011 an inter-calibration experiment was carried out to standardize MEDITS indices from GSAs 15-16 with those of Tunisian surveys.

4.1.1 Brief description of the direct method used Distribution, abundance and demographic information of the stock at sea derived from data collected during the standard bottom trawl surveys carried out annually in the northern sector of the Strait of Sicily from 1994 to 2015 in spring/early summer within the MEDITS (MEDiterranean International Bottom Trawl-Surveys) program, included in the European Data Collection Framework (DCF). A total of 45 hauls in GSA 15 and 120 hauls in GSA 16 were performed yearly. The bottom trawl surveys covered an area of about 10,580 km2 in GSA 15 and 45,000 km2 in GSA 16 within a depth-range 10-800 m in both areas. The sampling design is random stratified with allocation of hauls proportional to strata extension (depth strata: 10-50 m, 51-100 m, 101-200 m, 201-500 m, 501-800 m). Roughly the same haul positions were kept each year. A GOC 73 gear is used with mesh size in the cod-end 20 mm opening and the vertical opening of the mouth of 2.4-2.9 m. More details on the MEDITS protocol is reported in the MEDITS-Handbook. Version n. 7 (2013).

Direct methods: trawl based abundance indices

Table 4.1-1: Trawl survey basic information (GSA 16)

Survey MEDITS Trawler/RV TRAWLER

Sampling s eason MAY-JULY

Sampling design Stratified with number of haul by stratum proportional to stratum surface (see MEDITS-Handbook. Version n. 7, 2013, MEDITS Working Group: 120 pp..)

Sampler (g ear Bottom trawl made of four panels (IFREMER reference GOC 73) used)

Cod –end mesh 10 mm mesh size, which corresponds to ~ 20 mm of mesh opening size as opening in mm

Investigated depth 10-800m range (m)

22 Table 4.1-2: Trawl survey sampling area and number of hauls (GSA 16 in 2016).

Strat Total surface Trawlable surface Swept area Number of (km2) (km2) (km2) hauls

a 2979 11

b 5943 23

c 5563 21

6972 27

e 9927 38

Total 31384 120

Figure 4.1.1 - Map of hauls positions in the Strait of Sicily (GSA 16).

Figure 4.1.2 Map of hauls positions in GSA 15.

24 Table 4.1-3: Trawl survey abundance and biomass results (GSA 16)

Depth kg per km2 CV N per km2 CV kg CV N per km2 CV (GSA 16) (GSA 16) (GSA 16) (GSA 16) perkm2(GS (GSA 15) (GSA 15) (GSA 15) Stratum A 15) 10-800 m

1994 34.9 20.4 1249 45.2

1995 26.3 27.6 549 42.2

1996 15.4 21.8 334 25.8

1997 21.9 24.3 927 33.2

1998 15.8 16.8 355 28.5

1999 17.0 23.6 432 37.4

2000 24.5 22.4 710 27.6

2001 18.0 17.8 333 24.7

2002 20.6 22.8 427 28.8

2003 21.7 29.3 804 49.8

2004 28.8 27.1 613 29.9

2005 49.1 34.2 1635 42.3 38.5 462.66

2006 37.0 21.0 1258 43.0 31.8 432.29

2007 35.2 22.1 1586 38.3 44.1 559.8

2008 38.4 38.6 928 32.3 51.6 949.2

2009 35.0 27.2 1172 31.7 46.3 807.9

2010 38.7 26.2 507 27.3 25.2 275.3

2011 30.6 28.8 826 38.7 31.9 820.0

2012 46.5 21.0 1795 30.9 41.0 897.3

2013 54.3 23.8 995.6 27.8 42.3 777.3

2014 44.37 31.4 1001 50 49.3 906.6

2015 31.47 30.93 729 43 33.8 465.6

2016 25.7 30.2 492 27 28.65 3.74 743.6 28.6

25 Direct methods: trawl based length/age structure of population at sea Slicing method Length structure was sliced by knife edge approach.

Table 4.1-4: Hake MEDITS abundance index (n km-2/100) by age class (GSA 16)

N (Total or sex Year combined) by Age class in thousand 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

0 42.44 19.64 27.55 7.21 20.80 47.50 16.78 19.68 22.65 19.19

1 4.67 8.18 6.51 5.91 3.38 5.52 11.88 10.06 6.93 9.83

2 0.75 1.16 1.24 2.04 0.77 1.18 1.57 1.06 2.36 1.23

3 0.22 0.26 0.31 0.36 0.27 0.26 0.39 0.25 0.42 1.64

4 0.11 0.08 0.11 0.11 0.10 0.08 0.12 0.09 0.26 0.65

5 0.04 0.03 0.04 0.04 0.06 0.04 0.06 0.06 0.14 0.19

6 0.02 0.02 0.02 0.01 0.03 0.01 0.02 0.04 0.04 0.11

7 0.01 0.02 0.00 0.01 0.02 0.00 0.02 0.01 0.04 0.08

8+ 0.00 0.01 0.00 0.01 0.01 0.00 0.01 0.02 0.02 0.04

Total 48.27 29.40 35.77 15.70 25.44 54.60 30.86 31.26 32.86 31.76

26 Table 4.1-5 Sex ratio by length class in mm of hake from MEDITS in GSA 16. All year combined.

TL (mm) Sr TL (mm) Sr

100 0.94 460 0.99

120 0.42 480 0.99

140 0.46 500 1.00

160 0.48 520 1.00

180 0.48 540 1.00

200 0.48 560 1.00

220 0.47 580 1.00

240 0.51 600 1.00

260 0.51 620 1.00

280 0.65 640 1.00

300 0.75 660 1.00

320 0.80 700 1.00

340 0.87 720 1.00

360 0.85 740 1.00

380 0.91 760 1.00

400 0.93 780 1.00

420 0.88 800 1.00

440 0.98

27 Direct methods: trawl based Recruitment analysis (GSA 16) Table 4.1-6: Trawl surveys, recruitment analysis summary

Survey MEDITS Trawler/RV TRAWLER

Survey season May- July

Cod – end mesh size as opening in mm 20

Investigated depth range (m) 10- 800

Recruitment season and peak All year around (months ) Age at fishing- grounds recruitment 0

Length at fishing- grounds recruitment 6 cm TL

28 Table 4.1-7: Trawl surveys: recruitment analysis results

2 Years Area in km2 N of recruit per km CV

1994 1019 48.2

1995 399 52.7

1996 223 33.8

1997 796 35.5

1998 264 33.3

1999 330 44.1

2000 572 31.3

2001 204 33.7

2002 230 33.7

2003 637 55.2

2004 405 34.3

2005 1230 43.6

2006 934 56.1

2007 1365 42.4

2008 593 33.0

2009 874 37.2

2010 212 37.9

2011 626 43.7

2012 1447 35.1

2013 446 39.0

2014 518 46.0

2015 267 39

2016 206 43

29 4.1.2 Spatial distribution of the resources Data collected during MEDITS trawl survey were used to map the distribution of stable nurseries in GSA 15 and 16.

Fig. 4.1.2.1- Temporal persistence of nursery areas of hake in GSAs 15-16, from MEDISEH - MAREA project.

4.1.3 Historical trends The trends in biomass and density of European hake (HKE) during the MEDITS survey in GSA 15 and GSA 16 showed large fluctuations (Fig. 4.1.3.1). The current relative stock biomass (kg km-2) is below the 66° percentile of the time series (Fig. 4.1.3.2).

60.0 BI_GSA16 B33 percentile B66 percentile

50.0

40.0

30.0

20.0

10.0

0.0

Figure 4.1.3. 2: Biomass index (BI_GSA16), Biomass rd and th percentile from Medits survey in GSA 16

32 The length structures of hake in MEDITS 2007-2016 are shown in Fig. 4.1.3.4

a Trawl survey 2007-2015 GSA 15

35000

30000

25000

20000

15000

10000

5000 0 0 20 40 60 80 100 TL (cm)

2007 2008 2009 2010 2011

2012 2013 2014 2015 2016

b Trawl survey 2007- 2015 16 GSA 300

250

200

150

100

0 0 10 20 30 40 50 60 70 80 TL ( 2007 2008 2009 2010 2011

2012 2013 2014 2015 2016

Figure 4.1.3.4 – a) Length frequency distributions (LFD) from MEDITS survey in GSA 15.b) Length frequency distributions LFD from MEDITS survey in GSA 16. Sex combined.

33 5 Stock Assessment An Extended Survivor Analysis (XSA) assessment was carried out using official trawl catch data (landings and discards, EU data collection framework) collected in GSAs 15-16 and GSAs 12-14 (Tunisia) in the period 2007 - 2016. The XSA was tuned using Medits survey data (2007- 2016) from GSA 15 and GSA 16. The natural mortality M was estimated by Prodbiom’s method. The annual size of the landings, as well as Medits data was converted into the number at age by knife edge slicing. The stock status of European hake was also assessed using the Gadget (Area Disaggregated General Ecosystem Toolbox).

5.1 Extended Survivor Analysis (XSA) 5.1.1 Model assumptions Darby and Flatman (1994) outlined the XSA algorithm as performing the following steps: (1) a cohort analysis of the total catch-at-age data to produce estimates of population abundance- at-age, and total fishing mortalities; (2) adjustment of the CPUE values for the period of fishing defined using the alpha and beta parameters in the fleet tuning file, into CPUE values that would have been recorded if the fleet had fished only at the beginning of the year. The adjusted values are directly comparable with the population abundances at the beginning of the year; (3) calculation of fleet-based estimates of population abundance-at-age from the adjusted CPUE values and fleet catchabilities; (4) calculation of a least squares estimate (weighted mean) of the terminal population (survivors at the end of the final assessment year) for each cohort in the tuning range using the fleet-derived estimates of population abundance-at-age. These terminal populations are used to initiate the Cohort analysis in the next iteration. The process iterates until the convergence criteria described for ad hoc tuning are achieved. Various options are available for catchability analysis, time series weighting and shrinkage of the weighted estimates.

34 5.1.2 Scripts

5.1.3 Input data and parameters For analytical models: catch matrix in ages. It’s included discard. Table 5.1.3.1 - Catch matrix by year and age.

0 1 2 3 4 5 6 7 8

2007 320.1 31537 6279 1181 241 86 83 17 14

2008 237.4 23829 6109 1297 224 62 48 16 13

2009 1029.9 29922 6007 1180 367 176 154 47 21

2010 663.6 21690 6678 1448 345 251 140 47 20

2011 709.7 20992 4192 1119 250 157 82 32 17

2012 1754.3 15839 7351 1821 431 171 92 94 77

2013 1890.3 15479 8129 2002 498 189 87 88 74

2014 2930.4 18641 9657 2426 602 236 121 137 122

2015 2354.8 13657 7403 2288 556 634 310 100 158

2016 600.1 10650.6 6983 1845 717 658 380 138 162

Table 5.1.3.2 - Catch weight matrix by year and age.

0 1 2 3 4 5 6 7 8

2007 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2008 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2009 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2010 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2011 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2012 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2013 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2014 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

2015 0.013 0.073 0.206 0.415 0.696 1.035 1.420 1.836 2.269

35 Table 5.1.3.3 – Tuning data by year and age.

Tuning data at age

0 1 2 3 4 5 6 7 8

2007 42.439 4.67 0.752 0.221 0.107 0.043 0.018 0.012 0.003

2008 19.643 8.178 1.162 0.264 0.077 0.028 0.022 0.017 0.011

2009 27.549 6.505 1.236 0.31 0.106 0.041 0.018 0 0.001

2010 7.212 5.91 2.038 0.356 0.107 0.042 0.014 0.013 0.008

2011 20.8 3.382 0.765 0.274 0.101 0.056 0.031 0.018 0.009

2012 47.496 5.524 1.179 0.259 0.084 0.037 0.013 0.004 0.004

2013 16.782 11.877 1.569 0.389 0.123 0.064 0.02 0.021 0.011

2014 19.68 10.058 1.061 0.252 0.092 0.055 0.035 0.01 0.017

2015 22.65 6.93 2.36 0.42 0.26 0.14 0.04 0.04 0.02

2016 19.19 9.83 1.64 0.65 0.19 0.11 0.08 0.04 0.03

Table 5.1.3.4 – Natural Mortality at age estimated by Prodbiom method.

Group 0 Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 Group 8

1.38 0.56 0.27 0.21 0.19 0.18 0.17 0.16 0.16

Table5.1.3.5 – Proportional of maturity specimens by age.

Group 0 Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 Group 8

0.01 0.16 0.61 0.93 0.98 1.0 1.0 1.0 1.0

36 5.1.4 Results

XSA was run using different setting of shrinkage on fishing mortality (0.5, 1.0, 1.5 and 2.0), rage (0, 1), qage (1, 2, 3, 4), shk.age (2,3), shk.years (3).

The assessment was based on the same settings of the 2016 assessment: xsa_control <- FLXSA.control (x=NULL, tol=1e-09, maxit=30, min.nse=0.3, fse=1, rage=0, qage=1, shk.n=TRUE, shk.f=TRUE, shk.yrs=3, shk.ages=3, window=100, tsrange=20, tspower=3, vpa=FALSE) The XSA run results with shrinkage = 1 and catchability = 1 years, was adopted as final model (Fig. 5.1.4.1) based on both residuals and retrospective analysis.

In table 5.1.4.1 are reported the XSA results in terms of spawning stock biomass (SSB), total biomass (TB) and annual recruitment. SSB increased in 2011-2014 to decrease since then. Recruitment is continuously decreasing since 2008 with lowest value observed in 2016.

Figure 5.1.4.1 - Final XSA model for European hake in GSAs 12-16.

37 Table 5.1.4.1 – XSA estimates of spawning stock biomass (SSB), total biomass (TB) and recruitment (REC).

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

2975,2 2997,3 3270,8 3400,1 3021,7 4250,9 4885,6 5873,1 5724,7 4971,5 SSB (tons)

TB (tons) 8839 8368 8287 8292 7928 9351 9983 10407 9496 7959

REC (in 160660 168255 138861 158853 145936 151514 142523 110210 125483 74482 milions)

Fishing mortality values at age are reported in table 5.1.4.2. Mean F (F1-5) is declining since

2007 achieving the lowest value in 2016: F1-5 = 0.61. The average of the last three year is

(2014-2016) is F1-5 = 0.74.

Table 5.1.4.2 – F at age by XSA.

F at age 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

0 0,003 0,002 0,01 0,01 0,01 0,02 0,02 0,04 0,03 0,01

1 1,22 1,15 1,06 1,05 0,73 0,56 0,52 0,77 0,70 0,28

2 1,15 1,33 1,07 1,26 0,72 0,86 0,88 1,10 1,12 0,84

3 1,01 0,91 0,79 1,10 0,71 0,90 0,64 0,81 0,88 0,66

4 1,06 0,55 0,53 0,66 0,51 0,69 0,67 0,42 0,41 0,51

5 1,09 0,94 0,81 1,04 0,67 0,82 0,73 0,83 1,00 0,75

6 1,09 0,94 0,81 1,04 0,67 0,82 0,73 0,83 1,00 0,75

Fbar(1-5) 1,11 0,98 0,85 1,02 0,67 0,77 0,69 0,79 0,82 0,61

33 Yield per Recruit

A yield per recruit analysis was carried using the FLBRP library (FLR) to calculate F0.1.

The estimated F01 value was 0.11 below the F01 value included in the GFCM management

plan (F01 =0.2).

Figure 5.1.4.2 - Yield per recruit analysis.

5.1.5 Robustness analysis Residuals at age of the Medits survey are reported in fig. 51.5.1.

Figure 5.1.5.1 - Residual at age of Medits survey.

34 5.1.6 Retrospective analysis Retrospective analyses showed rather consistent results with no major pattern in the data (Fig. 5.1.6.1).

Figure 5.1.6.1 - Retrospective analysis.

35 5.1.7 Assessment quality Based on Medits residuals and retrospective analysis the XSA results appear consistent and in line with the XSA assessment carried out in 2016 (Fig. 5.1.7.1).

Figure 5.1.7.1 - Final XSA model obtained in 2016 for hake in GSA 12-16

5.2 Alternative models: GADGET Gadget model (Beagley and Howell, 2004) is a parametric forward simulation model of an ecosystem, typically consisting of various fish populations, fleets and their interactions. In to the model the features of the ecosystem are: - One or more species, each of which may be split into multiple components (stocks); - Multiple areas with migration between areas; - Predation between and within species; - Growth; - Maturation; - Reproduction and recruitment; - Multiple commercial and survey fleets taking catches from the populations. - Different fisheries selectivities.

36 The Gadget framework consists of three parts, a model to simulate the ecosystem, a statistical functions to compare the model output to data and search algorithms to optimize the model parameters. Gadget works by running an internal forward projection model based on many parameters describing the ecosystem, and then comparing the output from this model to observed measurements to get a likelihood score. The model ecosystem parameters can then be adjusted, and the model re-run, until an optimum is found, which corresponds to the model with the lowest likelihood score. This iterative, computationally intensive process is handled within Gadget, using a robust minimization algorithm.

It was already used in the Mediterranean (hake in GSA 9: Bartolino et al., 2011) and currently adopted for the assessment of hake (Merluccius merluccius) in Ices divisions 8.c and 9.a (Southern Stock of hake). European hake (HKE) population is defined by 2 cm total length groups. The year is divided into four quarters. HKE age range in between 0 and 8, this latter used as plus group. Recruitment take place in spring and summer. Parameters of VBGF are Linf=100 (fixed) and K=0.116 estimated. Natural mortality was assumed as a vector using the PRODBIOM approach (Abella et al., 1997) ad hoc implemented in R.

5.2.1 Input data and Parameters (Gadget) The data included in the model: • Length based • Time: 2002-2016 • Five fleets: commercial trawlers ITA-MLT and TUN, small scale ITA and TUN • Medits survey fleet • Recruitment spring and summer • Fleet selectivity curve for HKE created ad hoc Parameters: • Linf=100.0 TL (fixed), K=0.12-0.16 (estimated) • M: 1.6 0.57 0.33 0.25 0.21 0.19 0.17 0.16 • Recruitments (estimated) • Maturity: 0.04 0.15 0.36 0.56 0.86 0.98 1.0 1.0 1.0 • Selectivity parameters (estimated)

Figure 5.2.1.1 - Double logistic selectivity curve with a right tail: a new function implemented. A specific selectivity curve for each fleet modeled

Figure 5.2.1.2 -. Ad-hoc fleet selectivity curves for Italian and Tunisian trawlers, Italian and Tunisian artisanal and Medit survey.

38 a) b)

Figure 5.2.1.3 - HKE in GSAs 12-16. Model simulated (black line) and observed (grey line) size distributions of trawlers catch (a: Italian trawlers; b: Tunisian trawlers). Medits fitting

Figure 5.2.1.4 - HKE in GSAs 12-16. Medits index (n Km-2) by length group: <20cm, 20-30cm, 30-40cm, >40mm ( Simulated; Observed).

43 Catch&biomass Fishing mortality

Figure 5.2.1.5 - HKE in GSAs 12-16. Catch, biomass and Fishing mortality (F1-6+) estimated by Gadget

44 6 Stock predictions 6.1 Short term predictions (XSA) A deterministic short term prediction for the period 2016 to 2019 was performed using the FLR routines provided by JRC and based on the results of the XSA assessment done during the GFCM WG. The input parameters were the same used for the XSA stock assessment. An average of the last three years has been used for weight at age, maturity at age and F at age. Recruitment (age 0) has been estimated from the population results as geometric mean of the last 3 years (190 million individuals). To achieve FMSY hake catch in 2018 should not exceed 1033 tons.

Table 6.1.1 – GSA 12-16 HKE - Short term forecast in different F scenarios.

Change SSB Change Catch 2016- 2017-2019 Ffactor Fbar Catch_2016 Catch_2017 Catch_2018 Catch_2019 SSB_2018 SSB_2019 2017 (%) (%)

Zero catch 0 0 2868 3609 0 0 4323 7976 84 -100

Status quo 1 0,732 2868 3609 2955 2794 4323 3938 -9 3

F0.1 0,273 0,2 2868 3609 1033 1498 4323 6539 51 -64

0,1 0,073 2868 3609 403 652 4323 7413 71 -86 0,2 0,146 2868 3609 777 1179 4323 6893 59 -73

0,3 0,220 2868 3609 1124 1602 4323 6414 48 -61

0,4 0,293 2868 3609 1446 1937 4323 5971 38 -50

0,5 0,366 2868 3609 1746 2199 4323 5562 29 -39

0,6 0,439 2868 3609 2024 2402 4323 5184 20 -29

0,7 0,513 2868 3609 2282 2554 4323 4835 12 -20

0,8 0,586 2868 3609 2522 2666 4323 4512 4 -12

0,9 0,659 2868 3609 2746 2744 4323 4214 -3 -4 1,1 0,805 2868 3609 3149 2822 4323 3682 -15 10

1,2 0,879 2868 3609 3330 2832 4323 3445 -20 16

1,3 0,952 2868 3609 3498 2828 4323 3226 -25 22

1,4 1,025 2868 3609 3656 2812 4323 3023 -30 27

1,5 1,098 2868 3609 3803 2787 4323 2835 -34 33

1,6 1,172 2868 3609 3940 2756 4323 2660 -38 37

Different 1,7 1,245 2868 3609 4069 2719 4323 2498 -42 42 scenario 1,8 1,318 2868 3609 4189 2678 4323 2348 -46 46

1,9 1,391 2868 3609 4301 2635 4323 2209 -49 50

2 1,464 2868 3609 4407 2590 4323 2079 -52 54

45 7 Draft scientific advice

Based on Indicator Analytical Current Empirical Trend Stock reference value from reference value (2007- Status point the analysis 2016)

Fishing Fishing F0.1 = 0.20 Fcurr =0.73 FC/F0.1=3.6 N OH mortality mortality OH

Fishing D effort

Catch

Stock Relative 24.8 33th percentile N Ol abundance Biomass 66th percentile (Medits kg Bcurrent km-2) 35.1

25.7

SSB 3263 33th percentile

66th percentile 4847 SSBcurrent 4972

Recruitment 74482 million Rec 2016

Final Diagnosis The ratio Fcurr/F0.1 is above to 1.66 in both case (F0.1 = 0.12; F0.1=0.2), the stock is in high overfishing with relative low biomass SSB.

The diagnoses, is based on analytical and empirical references.

46 7.1 Explanation of codes Trend categories

1) N - No trend 2) I - Increasing 3) D – Decreasing 4) C - Cyclic

Stock Status Based on Fishing mortality related indicators

1) N - Not known or uncertain – Not much information is available to make a judgment; 2) U - undeveloped or new fishery - Believed to have a significant potential for expansion in total production; 3) S - Sustainable exploitation- fishing mortality or effort below an agreed fishing mortality or effort based Reference Point; 4) IO –In Overfishing status– fishing mortality or effort above the value of the agreed fishing mortality or effort based Reference Point. An agreed range of overfishing levels is provided;

Range of Overfishing levels based on fishery reference points

In order to assess the level of overfishing status when F0.1 from a Y/R model is used as LRP, the following operational approach is proposed:

• If Fc*/F0.1 is below or equal to 1.33 the stock is in (OL): Low overfishing

• If the Fc/F0.1 is between 1.33 and 1.66 the stock is in (OI): Intermediate overfishing

• If the Fc/F0.1 is equal or above to 1.66 the stock is in (OH): High overfishing *Fc is current level of F

5) C- Collapsed- no or very few catches;

Based on Stock related indicators

1) N - Not known or uncertain: Not much information is available to make a judgment

47 2) S - Sustainably exploited: Standing stock above an agreed biomass based Reference Point; 3) O - Overexploited: Standing stock below the value of the agreed biomass based Reference Point. An agreed range of overexploited status is provided;

Empirical Reference framework for the relative level of stock biomass index

• Relative low biomass: Values lower than or equal to 33rd percentile of biomass index in the time series

(OL)

th • Relative intermediate biomass: Values falling within this limit and 66 percentile (OI)

th • Relative high biomass: Values higher than the 66 percentile (OH) 4) D – Depleted: Standing stock is at lowest historical levels, irrespective of the amount of fishing effort exerted; 5) R –Recovering: Biomass are increasing after having been depleted from a previous period;

Agreed definitions as per SAC Glossary

Overfished (or overexploited) - A stock is considered to be overfished when its abundance is below an agreed biomass based reference target point, like B0.1 or BMSY. To apply this denomination, it should be assumed that the current state of the stock (in biomass) arises from the application of excessive fishing pressure in previous years. This classification is independent of the current level of fishing mortality. Stock subjected to overfishing (or overexploitation) - A stock is subjected to overfishing if the fishing mortality applied to it exceeds the one it can sustainably stand, for a longer period. In other words, the current fishing mortality exceeds the fishing mortality that, if applied during a long period, under stable conditions, would lead the stock abundance to the reference point of the target abundance (either in terms of biomass or numbers)