Is square-mesh better selective than larger mesh? A

Fisheries Research 161 (2015) 182–190

Contents lists available at ScienceDirect

Fisheries Research journal homepage: www.elsevier.com/locate/fishres

Is square-mesh better selective than larger mesh? A perspective on the management for Mediterranean trawl fisheries Antonello Sala a,∗ , Alessandro Lucchetti a , Anna Perdichizzi b , Bent Herrmann c , Paola Rinelli b a Consiglio Nazionale delle Ricerche (CNR), Istituto di Scienze Marine (ISMAR) [National Research Council (CNR), Institute of Marine Sciences (ISMAR)], Largo Fiera della Pesca, 60125 Ancona, Italy b Consiglio Nazionale delle Ricerche (CNR), Istituto per l’Ambiente Marino Costiero (IAMC) [National Research Council (CNR), Institute of Coastal Marine Environment (IAMC)], Sp. San Raineri 86, 98122 Messina, Italy c SINTEF Fisheries and Aquaculture, Fishing Gear Technology, Willemoesvej 2, 9850 Hirtshals, Denmark

a r t i c l e

i n f o

Article history: Received 27 November 2013 Received in revised form 23 July 2014 Accepted 25 July 2014 Handling Editor P. He Keywords: Selectivity Trawl fishery Diamond-mesh Square-mesh Aristaeomorpha foliacea Mullus barbatus

a b s t r a c t Relatively little scientific work has been done to assess the selectivity of square-mesh codends in the highly variable multi-species conditions in the Italian trawl fisheries. Therefore this study was initiated to investigate the effect of using square-mesh and larger diamond-mesh codends on size selectivity of deepwater red shrimp (Aristaeomorpha foliacea) and red mullet (Mullus barbatus), with possible implication in future management measures. Four different codends were used in the sea trials. When size selection estimates are applied for management issues it is not sufficient only to consider the mean size selection parameters. It also needs to consider the effect of between-haul variations in the selection process. In the current study, potential consequences of the between-haul variation on the selection curves have been considered by applying a pooled curve with a double bootstrap approach. The results attained in the present study indicate that the use of 40 mm square-mesh codend results in a 50% retention length (L50) similar to that of the 50 mm diamond-mesh codend, but with a steeper selection range (SR). For red mullet, both 50 mm diamond-mesh and 40 mm square-mesh codends lead to an L50 that is higher than the minimum landing size (MLS, 11 cm). Deepwater red shrimp is not subject to any MLS and may always be landed legally. However, large amounts of juveniles of deepwater red shrimp have been retained in all codends. Council Regulation (EC) No. 1967/2006 called for a discard reduction policy in waters under the jurisdiction of the European Union. As demonstrated in the current paper, to simultaneously improve the size selectivity of fish and shrimp species can be difficult due to large differences in their morphological characteristics, therefore it may question whether more sophisticated alternative of selective devices, such as grids or square-mesh panels, could be implemented in some Mediterranean fisheries. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Italian trawl fisheries land a large number and a variety of commercially important species using small diamond-mesh codends, which tend to retain almost all animals. In order to reduce fishing mortality and discards of dying marine organisms, the improvement of the trawl selectivity is therefore of prime importance. For example, Council Regulation (EC) No. 1967/2006 requested an increase of diamond-mesh size from 40 mm to 50 mm or the use of 40 mm square-mesh in the codend. The selectivity performance

∗ Corresponding author. Tel.: +39 071 2078841; fax: +39 071 55313. E-mail address: [email protected] (A. Sala). http://dx.doi.org/10.1016/j.fishres.2014.07.011 0165-7836/© 2014 Elsevier B.V. All rights reserved.

of Italian bottom trawls has been widely investigated (Lucchetti, 2008; Lucchetti and Sala, 2012; Sala et al., 2004, 2006, 2007a,b; Sala and Lucchetti, 2010, 2011), however relatively little scientific work has been done to assess the selectivity of square-mesh codends in the highly variable multi-species Italian trawl fisheries (Sala et al., 2008). Deepwater red shrimp (Aristaeomorpha foliacea) are widely represented in the slope bottoms (400–800 m) of the South Tyrrhenian Sea and Strait of Sicily, and are actively fished by both Italian and foreign fleets, mainly from the Northern Africa countries. This deepwater fishery can be nearly considered as mono-specific, given that red shrimp represent the most abundant and valuable catch (Ragonese et al., 2002). Red mullet (Mullus barbatus) is one of the most important resources for shelf demersal fishery (down

A. Sala et al. / Fisheries Research 161 (2015) 182–190

to 200 m) in the Mediterranean. The current investigation was restricted to the red mullet inhabiting the Sicilian side, because biological features of those living on the Tunisian shelf suggest that they belong to another population (Levi et al., 1994). Most of the Mediterranean studies on the selectivity for deepwater red shrimp focus on diamond-mesh codends (Ragonese et al., 1994, 2002; D’Onghia et al., 1998; Carlucci et al., 2006). Currently there are only three studies that investigated the effect of square-meshes in the codend on the selectivity for shrimp species in Mediterranean. In these studies, selectivity parameters for the 40 mm square-mesh were estimated for Aristeus antennatus, Parapenaeus longirostris and Plesionika martia by Guijarro and Massutí (2006) for P. longirostris by Sala et al. (2008) and for A. foliacea, A. antennatus, P. longirostris and P. martia by Deval et al. (2009). The codend tested by Deval et al. (2009) was entirely made of knotted polyethylene (PE). No study has examined the effect of using knotless polyamide (PA) square-mesh codends on size selectivity for deepwater red shrimp. This study investigates the effect of using square-mesh netting and larger diamond-meshes in the codend on size selectivity of deepwater red shrimp and red mullet, two of the most important commercial species in this area, with possible implication on management measures for the fisheries.

2. Materials and methods 2.1. Sea trials Selectivity trials were conducted on a commercial vessel on two different fishing grounds of the South Tyrrhenian normally exploited by local fishermen. The gear employed in the sea trials was a typical Italian commercial trawl net (Prat et al., 2008; Eigaard et al., 2011a) entirely made of knotless PA netting. All rigging components of the gear were identical with those commonly adopted in commercial practice in the Italian trawl fisheries (Dremière et al., 1999; Fiorentini et al., 1999, 2004; Sala et al., 2007c; Notti et al., 2013). Gear performance (i.e., door spread, horizontal and vertical net openings) was measured during all hauls using a SCANMAR system. A laptop automatically controlled data acquisition and provided for real time correct system functioning through a customized software (Brˇcic´ et al., 2014). The main goal of these measurements was to obtain for each haul detailed, real time data on gear performance. Four different codends were used in the sea trials. The nominal mesh sizes tested were 44 mm and 54 mm and two different mesh configurations were associated with each mesh size: diamond- and square-mesh. The codend mesh sizes were measured using an ICES mesh gauge (ICES, 1962) with 4 kg tension while the netting was wet. The actual mesh sizes of the four tested codends were: 45.15 ± 0.75 mm (DM44-320) and 45.95 ± 0.39 mm (SM44-160) for the 44 mm diamond- and square-mesh respectively; and 54.70 ± 0.86 mm (DM54-256) and 56.90 ± 0.45 mm (SM54-128) for the larger 54 mm diamond- and square-mesh respectively (Table 1). The four codends were designed to have almost the same circumference during fishing (approximately 3.5 m). Generally, underwater observations have shown the diamond-meshes in the codend of demersal trawls tend on average to open circumferentially only about 15–25% of the mesh size (Robertson, 1986, 1993). Therefore, the circumferential length during fishing, clf (mm), with a diamond-mesh codend would not be expected to exceed clfDM = 0.25 × ms × nmw, where ms (mm) and nmw are the mesh size and the number of meshes in the codend circumference, respectively. On the other hand, the circumferential length of a square-mesh codend will be nearly clfSM = 0.50 × ms × nbw, where nbw is the number of bars in the circumference.

183

Notice we have used that the square-mesh bar length is half the diamond-mesh size. In order to accurately matching the circumference of the diamond- and square-mesh codends, we must therefore calculate the number of meshes in circumference (or bars) of the square-mesh codend to produce similar values for clf as for the diamond-mesh codends while accounting for the differences in circumferential openness of the meshes for these two codend types. Considering the mesh sizes and number of meshes in circumference of the two tested diamond-mesh codends (Table 1), we find: clfDM44−320 = 0.25 × 45.15 × 320 = 3612 mm

(1)

clfDM54−256 = 0.25 × 54.70 × 256 = 3501 mm

(2)

Therefore, to find out the number of bars of the square-mesh codends to match at such circumferences, it is necessary to divide each clf of the diamond-mesh codends, (1) and (2), by the bar length of the square-mesh codends: nbwSM44−160 =

3612 = 157 0.5 × 45.95

(3)

nbwSM54−128 =

3501 = 123 0.5 × 56.90

(4)

These numbers in (3) and (4) have been adjusted to 160 and 128, as reported in Table 1, in order to have identical rigging rules to those commonly adopted in commercial practices in the Italian demersal trawl fisheries. Size selectivity data were collected using the covered codend technique (Fig. 1), where a cover with a nominal mesh size of 20 mm was supported by circular hoops to keep it clear of the codend and minimize masking effects (Wileman et al., 1996). The cover was approximately 1.5 times larger in circumference and longer than the codend, as recommended by Stewart and Robertson (1985). For each haul, catches from trawl codend and cover were handled separately on board and weighed. Total length (TL, cm) for red mullet and carapace length (CL, mm) for deepwater red shrimp were measured respectively to the nearest 0.5 cm and 1.0 mm below in randomly selected sub-samples. The subsampling ratios were then calculated by dividing the sub-sample weight by the total weight in the codend and cover separately. 2.2. Estimation of mean selection curve Analysis was conducted separately for deepwater red shrimp and red mullet following the procedure described in Annex I. The selectivity results for each species were derived only from the first or the second cruise within the same area. When size selection estimates are to be applied for managements issues it is not sufficient just to consider the mean size selection obtained using the method described in Fryer (1991). One also needs to consider the effect of between-haul variations in the selection process (Frandsen et al., 2011). For example, even if mean SR is small (e.g. steep mean size selection curve) if there is a considerable between-haul variation in L50 then for the deployment over hauls hypothetically some hauls could lead to the catch of many small fish while others could lead to the loss of big fish. Such effect is not included in the Fryer mean size selection curve as it is represented by the between-haul variation matrix, see Fryer (1991) for details. Further different gear configurations can have different between-haul variations in selection process and for management questions it is also important to consider this effect when evaluating the benefit from applying a certain gear type (Wienbeck et al., 2011). One way to include the effect of between-haul variations into a single selection curve is to make what Millar (1993) called a “fishery selection curve” (Sistiaga et al., 2010; Herrmann et al., 2012). This can be obtained by pooling hauls, as described in Annex I and as in the current study it has been decided for estimating the selection

184

A. Sala et al. / Fisheries Research 161 (2015) 182–190

Table 1 Characteristics of the four codends (DM44-320, SM44-160, DM54-256, SM54-128) tested during the sea trials: nominal mesh size (nms); netting material (Mat.); measured mesh size (ms); twine linear density (tld), twine diameter (td); number of diamond-meshes in the codend circumference and in length (nmw and nml, respectively); number of bars in the circumference and in length for the square-mesh codend (nbw and nbl, respectively); longitudinal codend length (lcl); circumferential length during fishing (clf). The acronyms of codend name provide the mesh configuration (DM: diamond- or SM: square-mesh), the nominal mesh size (DMxx or SMxx) and the number of meshes around (-xxx). The numbers of meshes around do not include selvedge meshes. Mean values and standard deviations enclosed in parenthesis. Parameter/codend

DM44-320

SM44-160

DM54-256

SM54-128

nms [mm] Mat.

44 PA

44 PA

54 PA

54 PA

ms [mm] tld [tex] td [mm] nmw, nbw nml, nbl lcl [m] clf [mm]

45.15 (0.75) 4801 (39) 3.08 (0.01) 320 120 5.42 3612

45.95 (0.39) 4801 (39) 3.08 (0.01) 160 240 5.51 3676

54.70 (0.86) 4912 (37) 3.13 (0.02) 256 96 5.25 3501

56.90 (0.45) 4912 (37) 3.13 (0.02) 128 192 5.46 3642

Fig. 1. Schematic view of the codend and cover arrangement. A cover with a nominal mesh opening of 20 mm is supported by circular hoops to keep it clear of the codend and minimize masking effects. The cover was approximately 1.3–1.5 times larger and longer than the four tested codends.

curves for the different gear designs. Such curves will often have a less steep curvature since between-haul variations in L50 will be incorporated into the selection range (Fig. 2). Based on using Fryer (1991) mean selection curve alone, the expected catch of small fish would have been very low as the mean curves gives a retention probability close to 0. However, due to the between-haul variation, in some hauls the trawl might have caught a measurable fraction. Using a pooled curve with the double bootstrap method accounts for such potential effects in the estimated selection curve. An approach of this type can led to selection curves that are more suitable for management related evaluations because they better account for the average deployment pattern (Fig. 2). 2.3. Estimation of fractions retained below and above minimum landing size (MLS) In the Mediterranean Sea, red mullet is subject to an MLS of 11 cm (EC Reg. 1967/2006). On the contrary, deepwater red shrimp is not subject to any MLS and may always be landed legally. Based on the proportion of fully mature gonads by size class, Perdichizzi et al. (2012) estimated 42 mm CL of 50% size at first maturity. Therefore, in the current paper, a virtual MLS of deepwater red shrimp has been set accordingly.

Fig. 2. Hypothetical effect of the between-haul variation in the selection process. The stippled distribution is an average fish size structure. In the case of small SRs (e.g. steep size selection curves) with a considerable between-haul variation in L50, hypothetically some hauls could lead to the catch of many small fish while others could lead to the loss of big fish. The thin grey stipple curves represent simulated individual haul selection curves assuming relative steep selection hauls (small SR) but with a considerable between-haul variation in L50. The thick black curve represents the mean selection curve estimated by Fryer (1991).

A. Sala et al. / Fisheries Research 161 (2015) 182–190

185

The average length-integrated percentage of fractions below and above MLS in number of individuals and the ratio between number below and above (nP− , nP+ , nRatio) retained in each of the codends were calculated for each species. This was done by summing the number of individuals retained that were below and above MLS for each codend type. This sum was then divided by the total number of individuals in this size fraction for each specific codend to obtain the average fraction. Thus, the fractions were estimated using the following formulae:

   ncDjl /qCDj j lMLS  nP+ = 100 ×  



(5)

   ncDjl /qCDj j lMLS

ncDjl /qCDj

The two-compartments data format meant that, for each haul (j), counted numbers of deepwater red shrimp and red mullet of each length class l in compartment cover CV (nCVjl ) and in compartment codend CD (nCDjl ) were available. Likewise, we known the information on the length-independent sub-sampling fractions qCVj and qCDj , applied in the individual hauls both for the deepwater red shrimp and red mullet in the cover and in the codend. All terms are described in detail in Annex I. nP− gives a quick estimate of how large a fraction of the number of individuals below MLS a specific codend catch. It thus gives an indication if fishing is a problem with a specific codend on the population structure of the species available at the location where the fishing is conducted. nP− should preferable be low. Opposite does nP+ . It gives an indication on the retention efficiency of the population above MLS for the specific codend while considering the size structure of the population fished on. If the species is a target species it is preferable that nP+ is high (close to 100). nRatio give the number of individuals below MLS retained per individual above MLS retained. Thus for the size selectivity for the codend to be well adjusted for the MLS and considering the population fished on, nRatio should be low (close to zero). The above indicators were based on number of individuals but since the value of catch is more related to weight similar indicators based on weight are also estimated:

   wl × (ncDjl /qCDj ) j lMLS    wl × (ncDjl /qCDj ) j lMLS

(6)

wl × (ncDjl /qCDj )

where the weight wl , for individual belonging to length class l, have been estimated by: wl = a × lb

(7)

To estimate the uncertainty in nP− , nP+ , nRatio, wP− , wP+ and wRatio for each species, considering both the effect of betweenhaul variation and uncertainty related to within-haul variation, the double bootstrapping method, described in Annex I, have been used to estimate the bca “Efron percentile” 95 confidence limits. The analysis was conducted using the software tool SELNET as described in Annex I. 3. Results During the sea trials each codend was fished daily on the same trawl. Details of the hauls performed are reported in Table 2

Fig. 3. Size selection of deepwater red shrimp (Aristaeomorpha foliacea) in the four codends tested during the sea trials. The four codends differ in mesh configuration (DM: diamond- or SM: square-mesh) and in the mesh opening (MS44: 44 mm or MS54: 54 mm). Cross symbols represent the experimental data; thick curve indicate the fitted size selection curves; dotted curves describe the 95% confidence limits for the fitted size selection curves; vertical dash-dot line represents the MLS; thin continuous curves represent the population of shrimp entering the codend while thin dotted curves show the population of shrimp retained in the codends.

Overall, 29 valid hauls were carried out in the first cruise and 28 in the second. The parameter estimates of each codend obtained from analysis are shown in Table 3. Mean selectivity curves, as well as the length structure of shrimp and fish populations entering the codends and those retained using the different mesh sizes and configurations are shown in Figs. 3 and 4. For both the species, the L50 estimates demonstrate a change both with the use of square-mesh and mesh size increase (Table 3 and Fig. 5). Another point of interest is the different between-haul variation in the four codends. For both the analyzed species, the DM44-320 codend had very poor selectivity and did not provide sufficient length classes in particular hauls and/or a sufficient number of fish or shrimp both retained and released; therefore, some hauls have been discarded. For deepwater red shrimp, similar L50s were obtained with DM54256 and SM44-160 (about 21 mm) but with around 30% lower SR with square-mesh (SM44-160: 2.60 mm; DM54-256: 8.73 mm). A significantly lower retention of shrimp between 10–30 mm CL when using larger mesh size (Figs. 3 and 5) results in a reduction of 31–33% in number (nP− ) and 17–18% in weight (wP− ) of shrimps below MLS compared with those entering the trawl (Table 3). The reduced retention of square-mesh compared to the diamond-mesh caused a less pronounced improvement in the discard rates, with a reduction in number and weight of shrimps retained below MLS of 3–4% and 17–18% respectively (Table 3). Furthermore, the reduction in discard is not accomplished by an expected loss in catches of shrimps above the MLS both in number and weight (nP+ and wP+ in Table 3). Red mullet was the most abundant fish species and, for several hauls, the catch was more than 70% of the total catch weight. L50 increased by 49–54% and 35–36% with the use of square-mesh and larger diamond-mesh size respectively (DM44-320: 8.58 cm; SM44-160: 13.20 cm; DM54-256: 11.63 cm; SM54-128: 17.28 cm). Only mesh size had a significant effect on SR (p < 0.001), in

186

A. Sala et al. / Fisheries Research 161 (2015) 182–190

Table 2 Summary of the selectivity trials conducted on the South Tyrrhenian Sea in the course of two subsequent cruises (FC) on two different grounds normally exploited by Italian trawl fishermen. The first cruise (FC1) took place from 20 to 27 of May 2006 at about 520 m of mean depth, approximately 25 nm off Gulf of Patti (Sicily), and the second from 18 to 23 of October 2004, ashore at a mean depth of about 80 m. ID: haul code; CT: codend type; HD: haul duration; BSD: bottom sea depth; CDC: codend catch; CVC: cover catch. TOT refers to the total catch, while SPE is the catch of deepwater red shrimp (Aristaeomorpha foliacea) in the FC1 and red mullet (Mullus barbatus) in the FC2 cruise. The acronyms of codend name provide the mesh configuration (DM: diamond- or SM: square-mesh), the nominal mesh opening (DMxx or SMxx) and the number of meshes around (-xxx). FC

ID

CT

Date

Towing time

BSD [m]

Start

End

HD

CDC [kg] TOT

CVC [kg] SPE

TOT

SPE

FC1 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617

DM44-320 DM44-320 DM44-320 DM44-320 SM44-160 SM44-160 SM44-160 SM44-160 DM54-256 DM54-256 DM54-256 DM54-256 SM54-128 SM54-128 SM54-128 SM54-128 DM44-320 DM44-320 DM44-320 DM44-320 SM44-160 SM44-160 SM44-160 DM54-256 DM54-256 DM54-256 SM54-128 SM54-128 SM54-128

5/20/2006 5/20/2006 5/20/2006 5/20/2006 5/21/2006 5/21/2006 5/21/2006 5/21/2006 5/22/2006 5/22/2006 5/22/2006 5/22/2006 5/23/2006 5/23/2006 5/23/2006 5/23/2006 5/24/2006 5/24/2006 5/24/2006 5/24/2006 5/26/2006 5/26/2006 5/26/2006 5/26/2006 5/27/2006 5/27/2006 5/27/2006 5/27/2006 5/27/2006

8:49 11:16 14:31 16:45 7:52 10:18 13:58 16:18 8:14 10:40 14:42 16:57 8:12 10:28 14:56 17:41 7:24 9:43 13:22 15:44 7:30 9:48 12:12 15:24 6:30 8:47 12:03 14:18 16:40

10:04 12:40 15:44 18:09 9:04 11:41 15:15 17:45 9:28 12:08 15:57 18:23 9:25 11:52 16:12 18:41 8:37 11:08 14:45 17:13 8:46 11:15 13:30 16:54 7:45 10:13 13:20 15:47 17:54

1:15 1:24 1:13 1:24 1:12 1:23 1:17 1:27 1:14 1:28 1:15 1:26 1:13 1:24 1:16 1:00 1:13 1:25 1:23 1:29 1:16 1:27 1:18 1:30 1:15 1:26 1:17 1:29 1:14

543 478 604 417 520 457 622 468 548 424 627 496 289 431 629 459 549 472 647 527 563 464 632 492 559 460 669 499 634

46.16 12.92 8.70 85.26 18.98 17.17 12.48 19.67 25.81 32.33 11.83 15.61 14.18 9.90 9.64 8.10 20.77 14.90 25.46 23.01 11.93 15.58 23.05 57.40 15.76 17.69 11.36 24.50 13.40

2.70 1.90 2.80 2.80 1.72 1.90 6.20 5.50 2.60 1.00 4.00 3.60 0.90 0.50 3.70 3.60 7.40 1.69 6.20 4.80 5.40 2.80 6.00 5.40 5.80 3.50 4.80 5.80 6.60

4.63 0.87 0.39 4.20 17.14 19.76 3.54 2.65 19.13 16.62 1.96 5.38 55.05 25.46 11.95 1.78 1.54 2.44 0.86 2.59 5.53 2.90 10.62 12.32 5.39 6.90 5.20 22.27 5.44

0.02 0.08 – 0.05 0.03 0.20 – 0.10 0.54 0.60 0.45 0.16 0.50 0.02 0.22 0.30 0.02 0.01 – – – 0.01 0.04 0.09 1.20 0.46 0.60 0.08 1.12

1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721

DM44-320 DM44-320 DM44-320 DM44-320 SM44-160 SM44-160 SM44-160 SM44-160 SM44-160 DM54-256 DM54-256 DM54-256 DM54-256 DM54-256 SM54-128 SM54-128 SM54-128 SM54-128 SM54-128 DM44-320 DM44-320 DM44-320 SM44-160 SM44-160 DM54-256 DM54-256 SM54-128 SM54-128

10/18/2006 10/18/2006 10/18/2006 10/18/2006 10/19/2006 10/19/2006 10/19/2006 10/19/2006 10/19/2006 10/20/2006 10/20/2006 10/20/2006 10/20/2006 10/20/2006 10/21/2006 10/21/2006 10/21/2006 10/21/2006 10/21/2006 10/22/2006 10/22/2006 10/22/2006 10/22/2006 10/22/2006 10/23/2006 10/23/2006 10/23/2006 10/23/2006

10:47 13:22 16:01 17:23 8:21 9:58 11:27 14:27 16:05 8:17 9:37 11:00 12:14 15:23 8:26 10:00 11:18 12:23 15:33 8:13 9:44 11:05 12:47 15:57 8:35 10:05 13:19 14:42

11:21 13:52 16:31 17:53 8:51 10:28 11:59 14:57 16:35 8:47 10:07 11:30 12:44 15:48 8:57 10:30 11:48 12:53 16:03 8:43 10:14 11:35 13:17 16:27 9:05 10:35 13:49 15:12

0:34 0:30 0:30 0:30 0:30 0:30 0:32 0:30 0:30 0:30 0:30 0:30 0:30 0:25 0:31 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30 0:30

73 101 96 74 65 95 93 95 72 65 96 93 97 69 70 94 94 98 71 69 97 94 93 75 66 90 94 69

247.97 10.98 34.49 90.20 85.25 64.28 145.03 34.80 72.60 52.28 72.78 43.50 21.80 43.61 46.41 23.32 11.64 32.67 55.42 109.98 27.23 50.00 40.10 59.10 78.14 52.26 33.80 43.01

81.60 2.85 2.60 10.50 22.50 4.60 4.00 8.14 22.00 9.20 6.00 2.70 2.70 15.00 9.90 3.50 2.00 3.75 6.80 1.95 2.55 10.20 4.20 8.40 8.60 6.45 9.68

4.39 11.75 3.46 5.61 21.92 19.65 30.27 16.79 23.71 16.26 16.70 21.26 14.05 20.72 29.72 49.60 18.73 29.41 37.10 3.95 3.37 5.34 43.70 19.50 25.84 20.06 36.28 26.63

1.40 0.70 0.85 1.40 6.00 1.00 1.60 1.05 8.70 4.35 0.90 0.90 0.90 6.60 17.60 6.40 3.90 1.80 9.80 3.53 0.95 1.05 9.40 3.80 3.45 4.00 10.00 –

FC2

particular, by changing from DM44-320 to DM54-256. SR increased from around 1.51–3.41 cm and from 1.73 cm to 4.30 cm for square-meshes (Table 3). The codend which performed best was the SM44-160, with the best L50-SR combination (Fig. 5). Noticeably, compared to the DM54-256, the SM44-160 had a significant better retention indicated by no overlap of the 95% confidence

bands of the between-haul variation (Fig. 5). The SM44-160 codend had the lowest retention of undersized red mullet, however, relevant reductions of legal sized fish also occurred (nP− and nP+ in Table 3). Compare to the other codends, the loss of legal sized individuals occurred in the SM44-160 might directly translate into an increase in effort but not in an additional catch of undersized red

A. Sala et al. / Fisheries Research 161 (2015) 182–190

187

Table 3 Direct estimate of the selectivity parameters for deepwater red shrimp (Aristaeomorpha foliacea) and red mullet (Mullus barbatus) in the four codends (DM44-320, SM44-160, DM54-256, SM54-128) tested during the sea trials. The acronyms of codend name provide the mesh configuration (DM: diamond- or SM: square-mesh), the nominal mesh opening (DMxx or SMxx) and the number of meshes around (-xxx). Mean values (in bold), respective standard deviations (in parenthesis) and Efron percentile 95 confidence limits (in italics) of the retention length at 50% (L50), selection range (SR) and the length-integrated percentage of fractions below and above MLS in number of individuals and the ratio between number below and above (nP− , nP+ , nRatio) retained in each of the codends. Since the value of catch is more related to weight similar indicators based on weight (wP− , wP+ , wRatio) have been also estimated. L50 and SR are in mm for deepwater red shrimp and in cm for red mullet. Codend/parameter

L50

SR p-Value DOF Deviance nP−

nP+

nRatio

wP−

wP+

wRatio

ARISFOL

MULLBAR

DM44-320

SM44-160

DM54-256

SM54-128

DM44-320

SM44-160

DM54-256

SM54-128

0.10 (2.86) 0.10–11.55 14.10 (3.24) 4.80–19.36 0.299 32 35.7 97.73 (1.39) 94.05–99.35 100.00 (0.00) 100.00–100.00 2.820 (0.49) 2.24–4.12 98.05 (1.36) 94.35–99.60 100 (0.00) 100.00–100.00 0.782 (0.12) 0.65–1.12

20.71 (0.88) 18.90–22.25 2.60 (0.66) 1.48–4.01 0.980 39 23.0 88.49 (4.84) 75.12–94.82 100.00 (0.00) 100.00–100.00 1.650 (0.30) 1.29–2.51 94.48 (2.35) 88.07–97.65 100 (0.00) 100.00–100.0 0.479 (0.09) 0.37–0.73

21.52 (22.45) 16.89–25.60 8.73 (1.73) 5.88–12.85 0.000 39 85.2 66.77 (10.33) 41.23–82.46 99.77 (0.32) 98.41–100.00 2.413 (0.52) 1.50–3.49 80.44 (7.10) 61.28–89.91 99.8 (0.28) 98.60–100.00 0.765 (0.14) 0.54–1.08

27.47 (1.11) 25.47–29.78 5.63 (0.74) 4.44–7.46 0.998 46 23.6 55.45 (9.60) 39.63–76.76 100.00 (0.00) 100.00–100.00 0.841 (0.18) 0.57–1.28 77.12 (7.12) 62.67–89.81 100 (0.00) 100.00–100.00 0.367 (0.08) 0.25–0.55

8.58 (0.25) 8.17–9.13 1.51 (0.26) 0.99–1.96 0.000 40 105.9 21.08 (5.78) 9.27–31.66 99.83 (0.12) 99.33–99.93 0.129 (0.15) 0.07–0.59 30.87 (8.58) 13.45–45.19 99.93 (0.06) 99.61–99.97 0.019 (0.01) 0.01–0.06

13.20 (0.23) 12.74–13.62 1.73 (0.25) 1.31–2.27 0.000 36 184.4 0.69 (0.41) 0.00–1.53 64.00 (5.67) 53.37–75.82 0.017 (0.02) 0.00–0.07 1.53 (0.94) 0.00–3.55 81.05 (3.87) 72.51–88.13 0.003 (0.00) 0.00–0.01

11.63 (0.46) 10.91–12.80 3.41 (0.38) 2.82–4.37 0.000 37 112.6 6.89 (1.69) 4.23–10.87 80.53 (3.80) 73.15–88.21 0.169 (0.07) 0.08–0.34 8.89 (1.76) 5.78–12.65 88.62 (3.28) 81.40–94.36 0.020 (0.01) 0.01–0.03

17.28 (0.51) 16.55–18.75 4.30 (0.66) 3.21–5.94 0.000 39 142.8 0.84 (0.49) 0.00–1.98 32.77 (4.44) 25.55–42.45 0.022 (0.02) 0.00–0.07 1.00 (0.70) 0.00–2.89 48.71 (5.60) 39.13–60.37 0.002 (0.00) 0.00–0.01

Fig. 4. Size selection of red mullet (Mullus barbatus) in the four codends tested during the sea trials. The four codends differ in mesh configuration (DM: diamondor SM: square-mesh) and in the mesh opening (MS44: 44 mm or MS54: 54 mm). Cross symbols represent the experimental data; thick curve indicate the fitted size selection curves; dotted curves describe the 95% confidence limits for the fitted size selection curves; vertical dash-dot line represents the MLS; thin continuous curves represent the population of fish entering the codend while thin dotted curves show the population of fish retained in the codends.

Fig. 5. Mean REML values (residual maximum likelihood) and between-haul variation of selection range (SR) plotted against the 50% retention length (L50), based on the lowest value of Akaike’s Information Criterion (AIC) for deepwater red shrimp (ARISFOL, Aristaeomorpha foliacea) and red mullet (MULLBAR, Mullus barbatus) in the four codends (DM44-320, SM44-160, DM54-256, SM54-128) tested during the sea trials. The acronyms of codend name provide the mesh configuration (DM: diamondor SM: square-mesh), the nominal mesh opening (DMxx or SMxx) and the number of meshes around (-xxx). L50 and SR are in mm for deepwater red shrimp and in cm for red mullet.

188

A. Sala et al. / Fisheries Research 161 (2015) 182–190

mullet. This is quantified by the lowest discards to landings ratio (nRatio in Table 3).

4. Discussion The Regulation No. 1343/2011 of the European Parliament and Council of 13 December 2011 has allowed for clarification of Article 9 of the Council Regulation (EC) No. 1967/2006 concerning management measures for the sustainable exploitation of fishery resources in the Mediterranean Sea. The main intention behind the amendment was to clarify that the request of ship owners to use a diamond-mesh net of 50 mm in the codend, instead of a 40-mm square-mesh net, must be justified on the basis of selectivity criteria and not, for example, socio-economic constraints. Inspired by the necessity to provide a scientific justification, experiments have been conducted in order to investigate whether both larger mesh size and square-mesh could equally improve codend size selection. Therefore, four codends with two different mesh sizes across two mesh configurations (square- and diamond-mesh) have been tested. For this reason the current study is able to infer in general of the effects of mesh size and mesh configuration on the codend size selectivity. From the literature (see Sala et al., 2008, for a critical review) it is known that the increase in mesh size produces both an increment in L50 and an unwanted increment in SR (Sala and Lucchetti, 2011). The results attained in the current study indicate that the use of 40 mm square-mesh codend produces a L50 similar to the 50 mm diamond-mesh codend and a steeper selection curve with a smaller SR. It is also worth mentioning the very low estimated L50 of 0.10 mm for deepwater red shrimp with 44 mm diamond-mesh, which does not mean that individuals down near that size have been caught. The small values for the parameters in this case L50 simply suggest that the retention likelihood is very high for all sizes present. When such small L50 is commented, it is important to consider the uncertainty of the value (CI) which is very wide. Practically the current results are not stating that the L50 is necessary that small, however a specific value is needed to be presented else the selection curve cannot be produced, what is important is that the curve is not extrapolated below the length class range available. Beside differences in mean size selection, different codend designs may also differ in their sensitivity to other factors acting within the fishing process (Sala and Lucchetti, 2011). This could lead to differences in between-haul variation for different codend designs. According to Herrmann and O’Neill (2005) and Wienbeck et al. (2011) one relevant factor could be the variation in the shape of the codend during different stages of the fishing process linked to randomness of fish arrival to the codend during the fishing process. Selectivity parameters of diamond-meshes estimated in this study for deepwater red shrimp are in agreement with those obtained from the Sicily Channel (Ragonese et al., 1994, 2001, 2002) and the Ionian Sea along the Calabrian coast (D’Onghia et al., 1998, 2003). Currently, only Deval et al. (2009) addresses the effect of square-mesh on selectivity of deepwater red shrimp, however they tested codend entirely made up of knotted PE and not knotless PA netting, as in the current study, therefore for such species there is no information available to compare with the current study results. The Council Regulation (EC) No. 1967/2006, concerning management measures for the sustainable exploitation of fishery resources in the Mediterranean, established the MSL of marine organisms. To simultaneously improve the size selectivity of fish and shrimp species can be difficult due to large differences in their morphological characteristics. The current experiment has found mismatch between the codend selectivity and the MLS, which is larger for shrimp than for fish. For red mullet, both 50 mm diamondmesh size and 40 mm square-mesh codends lead to an L50 that is

higher than the MLS (11 cm). However, large amounts of juvenile deepwater red shrimp will be retained in all codends, as the size at 50% maturity ranges between 39 mm and 47 mm CL (Cau et al., 1994; D’Onghia et al., 1998; Carlucci et al., 2006; Perdichizzi et al., 2012). The results presented in this study have potential implications for fisheries management in different Mediterranean areas. In the Strait of Sicily, Italian and non-EU trawl fleets (mostly from North Africa countries) exploit the same shared fishing grounds, therefore common management rules should be applied (Lucchetti and Sala, 2010). Council Regulation (EC) No. 1967/2006 called for a discard reduction policy in waters under the jurisdiction of the European Union and as it is often permitted the use of a relatively wide range of codend characteristics in many management areas. Based on the current results, it may be questioned whether more sophisticated alternative of selective devices, such as grids or square-mesh panels, could be used by some fisheries and have therefore some broader value in this context (Sala et al., 2011). Acknowledgements This work was partially funded by the research project MARE (Sicily POR-SFOP 2000/2006, Measure 4.17b, Decree D.D.G. 385/Pesca 09/07/2004) and the Flagship Project RITMARE “The Italian Research for the Sea”, coordinated by the Italian National Research Council and funded by the Italian Ministry of Education, University and Research within the 2011–2013 National Research Programme. We would like to thank our colleagues Vito Palumbo, Domenico Micucci (CNR-ISMAR, Italy) and Francesco Perdichizzi (CNR-IAMC) for their fieldwork. Finally, we are grateful to the editor and the two anonymous reviewers for their helpful suggestions, constructive comments and critical reading, which we feel has improved our manuscript. Annex I. Estimation of codend size selectivity The experimental design applied during the sea trials enabled analysis of the collected catch data as two-compartment data (binominal data). The fish were either retained by the cover (CV) over the codend or by the codend (CD) itself to estimate the size selection in the different codends (i.e., length-dependent retention probability). The two-compartments data format meant that, for each haul (j), counted numbers of deepwater red shrimp and red mullet of each length class l in compartment CV (nCVjl ) and in compartment CD (nCDjl ) were available. In addition, information on the length-independent sub-sampling fractions qCVj and qCDj , applied in the individual hauls for the deepwater red shrimp and red mullet in the cover and in the codend, were respectively available. The probability of finding an individual (deepwater red shrimp and red mullet, respectively) with length l in compartment CD in haul j given that it is found in one of the compartments is expressed by the function rj (l), which quantifies the length-dependent retention probability within the codend. The purpose of the analysis is to estimate the values of this function for all relevant sizes of deepwater red shrimp and red mullet respectively for each codend investigated. The values of rj (l) can be expected to vary between-hauls with the same codend (Fryer, 1991), although this variation was of no specific interest for this study. Instead, we were interested in the length-dependent values of r(l) averaged over hauls with the same codend because this would provide information about the average consequences for the size selection process of applying the specific codend to the fishery. Thus, it has been assumed that the size selective performance of the codends for the group of hauls conducted was representative of how the codends would perform in a commercial fishery (Millar, 1993; Sistiaga et al., 2010).

A. Sala et al. / Fisheries Research 161 (2015) 182–190

Estimation of the average size selection over hauls for a specific codend rav (l) involves pooling the raised data from the different hauls with this codend. According to Fryer (1991), simply pooling data over hauls could lead to underestimation of the uncertainties in the size selection process due to the potential between-haul variation. This problem was resolved by using a double bootstrapping technique that accounts for both within- and between-haul variation in the selection process. To account for between-haul variation an outer bootstrap resample with replacement from the group of hauls was included in the procedure. Within each resampled haul, the data for each length class was bootstrapped in an inner bootstrap with replacement to account for within-haul variation. The inner resampling of the data in each length class was performed prior to the raising of the data, to avoid underestimation of the within-haul variation. For each case analyzed (specific species for specific codends), 10,000 bootstrap repetitions were conducted to estimate the bca (bias corrected and accelerated) percentile 95% confidence limits (Efron, 1982; Chernick, 2007). Because this technique is similar to the one applied by Sistiaga et al. (2010), Eigaard et al. (2011b), Herrmann et al. (2012), and Madsen et al. (2012), it is not described further here. We write rav (l,v), where v is a vector consisting of the parameters of the model. The purpose of the analysis is to estimate the values of the parameter v that make experimental data (averaged over hauls) most likely to be observed, assuming that the model is able to describe the data sufficiently well. Thus, function (8) was minimized, which is equivalent to maximizing the likelihood for the observed data:

  nCDjl j

qCDj

l

× ln(rav (l, v)) +

nCVjl qCVj

 × ln(1.0 − rav (l, v))

(8)

where the summations are over hauls j for one specific codend type and length classes l. For each codend and species individually, the Logit model to describe rav (l,v) was used. The formulas for this model and the calculation of the selection parameters, L50 and SR, are described below (9). Further details on this size selection model can be found in Wileman et al. (1996). Logit:

rav (l, v) = L50 = − SR =

exp(v1 + v2 × l) 1 + exp(v1 + v2 × l)

v1 v2

(9)

ln (9)

v2

Evaluating the ability of the model to describe the data sufficiently well using (8) and (9) is based on calculating the corresponding p-value, which expresses the likelihood to obtain at least as big a discrepancy between the fitted model and the observed experimental data by coincidence. Therefore, for the fitted model to be a candidate to model the size selection data, this p-value should not be below 0.05 (Wileman et al., 1996). Model deviance versus degree of freedom also can be applied in the model evaluation (Wileman et al., 1996). Size selectivity was analyzed using the software SELNET following the methodological recommendations in Wileman et al. (1996). SELNET offers a variety of size selection models and methods for analysis, including the double bootstrap technique described above. Additional information about the SELNET software can be obtained by consulting Sistiaga et al. (2010), Eigaard et al. (2011b), Frandsen et al. (2011), Wienbeck et al. (2011), Madsen et al. (2012) and Herrmann et al. (2012).

189

References ´ J., Grilli, F., Sala, A., 2014. Influence of sea-bottom temperature and depth on Brˇcic, swept area estimation from trawl surveys. Sci. Mar. 78 (1), 135–142. Carlucci, R., D’Onghia, G., Sion, L., Maiorano, P., Tursi, A., 2006. Selectivity parameters and size at maturity in deep-water shrimps, Aristaeomorpha foliacea (Risso, 1827) and Aristeus antennatus (Risso, 1816), from the North-Western Ionian Sea. Hydrobiologia 557, 145–154. Cau, A., Sabatini, A., Murenu, M., Follesa, M.C., Cuccu, D., 1994. Considerazioni sullo stato di sfruttamento delle risorse demersali (Mari di Sardegna). Atti Seminario sulla regolazione dello sforzo di pesca. Biol. Mar. Mediterr. 1 (2), 67–76. Chernick, M.R., 2007. Bootstrap Methods: A Guide for Practitioners and Researchers. Wiley Series in Probability and Statistics, second ed. Wiley, New York, 400 pp. Council Regulation (EC) no. 1967/2006 of 21 December 2006, concerning management measures for the sustainable exploitation of fishery resources in the Mediterranean Sea, amending Regulation (EEC) no 2847/93 and repealing Regulation (EC) no 1626/94., 2006. Off. J. Eur. Union L 409. D’Onghia, G., Mastrototaro, F., Maiorano, P., Basanisi, M., 1998. Selettività della rete a strascico utilizzata sui fondi di scarpata (250–750 m) dello Ionio (Mediterraneo Centrale). Biol. Mar. Mediterr. 5 (2), 437–448. D’Onghia, G., Carlucci, R., Maiorano, P., Panza, M., 2003. Discards from deep-water bottom trawling in the eastern-central Mediterranean Sea and effects of mesh size changes. J. Northwest Atl. Fish. Soc. 31, 245–261. Deval, M.C., Bök, T., Ates¸, C., Ulutürk, T., Tosuno˘glu, Z., 2009. Comparison of the size selectivity of diamond (PA) and square (PE) mesh codends for deepwater crustacean species in the Antalya Bay, eastern Mediterranean. J. Appl. Ichthyol. 25 (4), 372–380. Dremière, P.Y., Fiorentini, L., Cosimi, G., Leonori, I., Sala, A., Spagnolo, A., 1999. Escapement from the main body of the bottom trawl used for the Mediterranean International Trawl Survey (MEDITS). Aquat. Living Resour. 12 (3), 207–217. Efron, B., 1982. The Jackknife, the Bootstrap and Other Resampling Plans. SIAM Monograph No. 38. CBSM-NSF, Philadelphia. Eigaard, O.R., Rihan, D., Graham, N., Sala, A., Zachariassen, K., 2011a. Improving fishing effort descriptors: modelling engine power and gear-size relations of five European trawl fleets. Fish. Res. 110, 39–46. Eigaard, O.R., Herrmann, B., Nielsen, J.R., 2011b. Influence of grid orientation and time of day on grid sorting in a small-meshed trawl fishery for Norway pout (Trisopterus esmarkii). Aquat. Living Resour. 25, 15–26. Fiorentini, L., Dremière, P.Y., Leonori, I., Sala, A., Palumbo, V., 1999. Efficiency of the bottom trawl used for the Mediterranean International Trawl Survey (MEDITS). Aquat. Living Resour. 12 (3), 187–205. Fiorentini, L., Sala, A., Hansen, K., Cosimi, G., Palumbo, V., 2004. Comparison between model testing and full-scale trials of new trawl design for Italian bottom fisheries. Fish. Sci. 70, 349–359. Frandsen, R.P., Herrmann, B., Madsen, N., Krag, L.A., 2011. Development of a codend concept to improve size selectivity of Nephrops (Nephrops norvegicus) in a multispecies fishery. Fish. Res. 111, 116–126. Fryer, R.J., 1991. A model of between-haul variation in selectivity. ICES J. Mar. Sci. 48, 281–290. Guijarro, B., Massutí, E., 2006. Selectivity of diamond- and square-mesh codends in the deepwater crustacean trawl fishery off the Balearic Island (western Mediterranean). ICES J. Mar. Sci. 63, 52–67. Herrmann, B., O’Neill, F.G., 2005. Theoretical study of the between-haul variation of haddock selectivity in a diamond mesh codend. Fish. Res. 74, 243–252. Herrmann, B., Sistiaga, M., Nielsen, K.N., Larsen, R.B., 2012. Understanding the size selectivity of redfish (Sebastes spp.) in North Atlantic trawl codends. J. Northwest Atl. Fish. Sci. 44, 1–13. ICES, 1962. Procés-Verbal de la Réunion 1961. Conseil Permanent International pour l’Exploration de la Mer. Recommendation B, vol. 2., pp. 62–63. Levi, D., Andreoli, M.G., Arneri, E., Giannetti, G., Rizzo, P., 1994. Otolith reading as a tool for stock identification. Fish. Res. 20, 97–107. Lucchetti, A., 2008. Comparison of diamond- and square-mesh codends in the hake (Merluccius merluccius L. 1758) trawl fishery of the Adriatic Sea (central Mediterranean). Sci. Mar. 72 (3), 451–460. Lucchetti, A., Sala, A., 2010. An overview of loggerhead sea turtle (Caretta caretta) bycatch and technical mitigation measures in the Mediterranean Sea. Rev. Fish Biol. Fish. 20 (2), 141–161. Lucchetti, A., Sala, A., 2012. Impact and performance of Mediterranean fishing gear by side-scan sonar technology. Can. J. Fish. Aquat. Sci. 69 (11), 1806–1816. Madsen, N., Herrmann, B., Frandsen, R.P., Krag, L.A., 2012. Comparing selectivity of a standard and turned mesh T90 codend during towing and haul-back. Aquat. Living Resour. 25 (3), 231–240. Millar, R.B., 1993. Incorporation of between-haul variation using bootstrapping and nonparametric estimation of selection curves. Fish. Bull. 91, 564–572. ´ J., Sala, A., 2013. Technical specifications of MediterNotti, E., De Carlo, F., Brˇcic, ranean trawl gears. In: Paschen (Ed.), Contributions on the Theory of Fishing Gears and Related Marine Systems. Proceedings of the 11th International Workshop on Methods for the Development and Evaluation of Maritime Technologies, Rostock, 9–12 October 2013, vol. 8, pp. 163–169, ISBN: 978-3-8440-2251-3, 0945-0874. Perdichizzi, A., Pirrera, L., Micale, V., Muglia, U., Rinelli, R., 2012. A histological study of ovarian development in the Giant Red Shrimp Aristaeomorpha foliacea (Crustacea: Decapoda: Aristeidae) from the Southern Tyrrhenian Sea (Western Mediterranean). Sci. World J., 9, http://dx.doi.org/10.1100/2012/289608, Article ID 289608.

190

A. Sala et al. / Fisheries Research 161 (2015) 182–190

Prat, J., Antonijuan, J., Folch, A., Sala, A., Lucchetti, A., Sardà, F., Manuel, A., 2008. A simplified model of the interaction of the trawl warps, the otterboards and netting drag. Fish. Res. 94, 109–117. Ragonese, S., Bianchini, M.L., Di Stefano, L., Campagnuolo, S., Bertolino, F., 1994. The Selectivity and the Assessment of the Coefficient of Retention of the Trawl Net used for Red Shrimp Fishing (Aristaeomorpha foliacea, Risso, (1827) and Aristeus antennatus, Risso, 1816: Crustacea – Aristeidae) in the Sicilian Channel (Central Mediterranean Sea). Final Report (Contract MED 92/010 EU). Ragonese, S., Zagra, M., Di Stefano, L., Bianchini, M.L., 2001. Effect of codend mesh size on the performance of the deepwater bottom trawl used in the red shrimp fishery in the Strait of Sicily (Mediterranean Sea). Hydrobiologia 449, 279–291. Ragonese, S., Bianchini, M.L., Di Stefano, L., 2002. Trawl cod-end selectivity for deepwater red shrimp (Aristaeomorpha foliacea, Risso 1827) in the Strait of Sicily (Mediterranean sea). Fish. Res. 57, 131–144. Regulation (EU) Nr. 1343/2011 of the European Parliament and of the Council of 13 December 2011 on certain provisions for fishing in the GFCM (General Fisheries Commission for the Mediterranean) agreement area and amending Council Regulation (EC) no. 1967/2006 concerning management measures for the sustainable exploitation of fishery resources in the Mediterranean Sea., 2011. Off. J. Eur. Union. L 347/44. Robertson, J.H.B., 1986. Design and construction of square mesh cod-ends. Scott. Fish. Inf. Pam. 12, 1–10. Robertson, J.H.B., 1993. Design and fitting of square mesh windows in whitefish and prawn trawls and seine nets. Scott. Fish. Inf. Pam. 20, 1–8. Sala, A., Lucchetti, A., 2010. The effect of mesh configuration and codend circumference on selectivity in the Mediterranean trawl Nephrops fishery. Fish. Res. 103, 63–72. Sala, A., Lucchetti, A., 2011. Effect of mesh size and codend circumference on selectivity in the Mediterranean demersal trawl fisheries. Fish. Res. 110, 252–258.

Sala, A., Lucchetti, A., Buglioni, G., 2004. The change in physical properties of some nylon (PA) netting samples before and after use. Fish. Res. 69, 181–188. Sala, A., Priour, D., Herrmann, B., 2006. Experimental and theoretical study of red mullet (Mullus barbatus) selectivity in codends of Mediterranean bottom trawls. Aquat. Living Resour. 19, 317–327. Sala, A., Lucchetti, A., Buglioni, G., 2007a. The influence of twine thickness on the size selectivity of polyamide codends in a Mediterranean bottom trawl. Fish. Res. 83, 192–203. Sala, A., O’Neill, F.G., Buglioni, G., Lucchetti, A., Palumbo, V., Fryer, R.J., 2007b. Experimental method for quantifying resistance to the opening of netting panels. ICES J. Mar. Sci. 64, 1573–1578. Sala, A., Lucchetti, A., Palumbo, V., Hansen, K., 2007c. Energy saving trawl in Mediterranean demersal fisheries. In: Guedes, Soares, Kolev (Eds.), Ocean Engineering and Coastal Resources. Taylor & Francis Group, London, pp. 961–964, ISBN: 978-0-415-45523-7. Sala, A., Lucchetti, A., Piccinetti, C., Ferretti, M., 2008. Size selection by diamond- and square-mesh codends in multi-species Mediterranean demersal trawl fisheries. Fish. Res. 93, 8–21. Sala, A., Lucchetti, A., Affronte, M., 2011. Effects of turtle excluder devices on bycatch and discards reduction in the demersal fisheries of Mediterranean Sea. Aquat. Living Resour. 24, 183–192. Sistiaga, M., Herrmann, B., Grimaldo, E., Larsen, R., 2010. Assessment of dual selection in grid based selectivity systems. Fish. Res. 105, 187–199. Stewart, P.A.M., Robertson, J.H.B., 1985. Small mesh covers. Scot. Fish. Res. Rep. 32, 11 pages, ISSN: 0308-8022. Wienbeck, H., Herrmann, B., Moderhak, W., Stepputtis, D., 2011. Effect of netting direction and number of meshes around on size selection in the codend for Baltic cod (Gadus morhua). Fish. Res. 109, 80–88. Wileman, D., Ferro, R.S.T., Fonteyne, R., Millar, R.B., 1996. Manual of Methods of Measuring the Selectivity of Towed Fishing Gear. ICES Coop. Res. Rep. No. 215, 126 pp.