Action

Use a square mesh instead of a diamond mesh codend in a trawl net

How is the evidence assessed?
  • Effectiveness
    not assessed
  • Certainty
    not assessed
  • Harms
    not assessed

Study locations

Key messages

COMMUNITY RESPONSE (0 STUDIES)

POPULATION RESPONSE (2 STUDIES)

  • Survival (2 studies): One of two replicated, paired, controlled studies in the Aegean Sea and Bristol Channel found that the short-term survival of two of six fish species was higher after escaping through a square mesh compared to a diamond mesh codend. The other study reported that skate caught in a square mesh codend had a higher overall survival likelihood than those caught in a diamond mesh codend.

BEHAVIOUR (0 STUDIES)

OTHER (25 STUDIES)

About key messages

Key messages provide a descriptive index to studies we have found that test this intervention.

Studies are not directly comparable or of equal value. When making decisions based on this evidence, you should consider factors such as study size, study design, reported metrics and relevance of the study to your situation, rather than simply counting the number of studies that support a particular interpretation.

Supporting evidence from individual studies

  1. A replicated, paired, controlled study in 1990 in an area of pelagic water in the western English Channel, UK (Casey et al. 1992) found no difference in the size composition (selection) of Atlantic mackerel Scomber scombrus in catches from a pelagic trawl with a large square mesh codend, compared to a smaller diamond-mesh codend. Length frequencies of mackerel caught with a 60 mm square mesh codend (range: 8–31 cm, midpoint: 13 cm) were similar to a 40 mm diamond mesh codend (range: 7–32 cm, midpoint: 13 cm). In January–February 1990, mackerel catches were compared from 14 trawl deployments on a commercial fishing vessel: nine with an experimental square mesh codend (60 mm); and five with a conventional diamond mesh codend (40 mm). The experimental codend was 4 m shorter than the conventional codend and had four net panels instead of two. Each trawl type was deployed alternately when mackerel shoals were visible in the water, and all mackerel caught were counted and their lengths measured.

    Study and other actions tested
  2. A replicated, paired, controlled study in 1988–1990 of a bottom fishing ground in the Greenland Sea, north Iceland (Thorsteinsson 1992) reported that shrimp trawl nets with square mesh codends caught less small, unwanted fish than conventional diamond mesh codends. Data were not statistically tested. In two of two comparisons, catch rates of fish aged <1 year were lower in square mesh than diamond mesh codends for Atlantic cod Gadus morhua (square: 2–8 fish/ha, diamond: 6–130 fish/ha) and whiting Merlangius merlangus (square: 4–376 fish/ha, diamond: 27–2,472 fish/ha), and in one case for haddock Melanogrammus aeglefinus (square: 0–457 fish/ha, diamond: 0–1,245 fish/ha). Catches of one- and two-year-old fish were low but were typically lower in square mesh codends (see paper for data). Overall capelin Mallotus villosus catch rates (all ages) were lower in square mesh codends in two of two cases (square: 133–284, diamond: 842–1,104 fish/ha). Target shrimp Pandalus borealis catches were lower in square than diamond mesh codends in three of three comparisons. In 1988 and 1990, catches were compared between square mesh codends (36–37 mm) and conventional diamond mesh codends (36–40 mm) in 11 deployments of the two trawl net types towed side by side for 1 h.

    Study and other actions tested
  3. A replicated, paired, controlled study in 1988–1990 in two offshore bottom fishing grounds in the Northwest Atlantic Ocean, Canada (Walsh et al. 1992) reported that bottom trawl nets with square mesh codends did not improve size selection of long rough dab Hippoglossoides platessoides compared to conventional diamond mesh codends at three different mesh sizes. Data were not statistically tested. The length at which plaice had a 50% chance of escape was lower in square mesh codends at all three mesh sizes tested (130 mm: 31 cm, 140 mm: 31 cm, 155 cm: 32 cm) than diamond mesh (130 mm: 31 cm, 140 mm: 38 cm; 155 cm: 38 cm) and was reported to increase marginally with increasing mesh size. Catches from square and diamond mesh codends of three different mesh sizes were compared during three experimental trials on the Scotian Shelf (140 mm mesh, 31 hauls) and Grand Bank (155 mm mesh, 29 hauls) in October 1988 and on the Grand Bank in March 1990 (130 mm mesh, 32 hauls). All hauls were done using a standard bottom trawl net modified with twin codends: each test codend (square or diamond) on one side towed with a small mesh (39 mm) control codend on the other (sides rotated during each trial). Codend catches from each haul were sorted and plaice were counted, and their lengths measured.

    Study and other actions tested
  4. A replicated, randomized, controlled study in 1993–1994 in two seabed areas in the Aegean Sea, Greece (Petrakis & Stergiou 1997, same experimental set-up as Stergiou et al. 1997 and Stergiou 1999) reported that a bottom trawl with a square mesh codend improved the size selectivity and escape rate of only one of four commercially important bottom fish species, compared to a diamond mesh codend of the same mesh size. Results were not statistically tested. The estimated length at which fish had a 50% chance of escape was greater in 20 mm square mesh codends than 20 mm diamond mesh codends for European hake Merluccius merluccius (15.1 vs 13.8 cm) and both were greater than the 14 mm diamond mesh (4.2 cm); but smaller for blue whiting Micromesistius poutassou (17.0 vs 21.2 cm), poor cod Trisopterus minutus capelanus (11.9 vs 13.7 cm) and four-spot megrim Lepidorhombus boscii (8.5 vs 10.3 cm) (selectivity for these species could not be estimated for the 14 mm diamond mesh codend). The proportion of fish retained in the trawl versus those that escaped was lower in square mesh codends for hake (square: 0.26, diamond: 0.35), but was higher for blue whiting (square: 0.61, diamond: 0.29), poor cod (square: 0.31, diamond: 0.18) and megrim (square: 0.90, diamond: 0.66). Experimental trawl deployments were conducted in the Trikeri Channel in October 1993 (5 stations) and the North Euboikos Gulf in March 1994 (seven stations). A trawl net was randomly assigned either a 20 mm square mesh codend or a 14 mm (the size currently used commercially) or 20 mm diamond mesh codend, and towed for 45–60 min at depths between 73–210 m. Each codend was deployed for 12 hauls. Small mesh (10 mm) covers over the codend sampled the escaping fish catch. Fish from the codend and cover were identified and counted, and their lengths measured.

    Study and other actions tested
  5. A replicated, randomized, controlled study in 1993–1994 in two seabed areas in the Aegean Sea, Greece (Stergiou et al. 1997, same experimental set-up as Petrakis & Stergiou 1997 and Stergiou 1999) found that a bottom trawl net with a square mesh codend did not allow more unwanted individuals and higher number of species (fish and invertebrates) to escape compared to a diamond mesh codend of the same mesh size. In two of two years, the average number of individuals (fish and invertebrates) and species escaping from the codend was similar between square and diamond mesh (individuals, square: 1,653–6,100/h, diamond: 1,486–8,167/h; species, square: 12–15/h, diamond: 12–16/h ). Experimental trawl deployments (using the same experimental set-up as Petrakis & Stergiou, 1997) were conducted in the Trikeri Channel in October 1993 (5 stations) and the North Euboikos Gulf in March 1994 (seven stations). A trawl net was randomly assigned either a 20 mm square mesh codend or 20 mm diamond mesh codend (12 hauls each codend), and towed for 45–60 min at depths between 73–210 m. Small mesh (10 mm) covers over the codend sampled the escaping fish catch. All individuals caught in the covers were identified and counted.

    Study and other actions tested
  6. A replicated, randomized, paired, controlled study in 1988–1990 in fishing grounds in the Pacific Ocean off the west coast USA (Perez-Comas et al. 1998) found that bottom trawls fitted with a square mesh codend did not typically improve the selectivity or reduce the catch of unwanted small, rockfish and flatfish species compared to diamond mesh codends, but increasing the mesh size did, for both designs. The length at which half of fish were likely to escape capture was higher in square mesh codends than diamond mesh codends of the same mesh size for two of five rockfish but none of four flatfish (see paper for individual data – not statistically tested). Increasing the mesh sizes retained fewer undersized fish in both codend types: for four of four rockfish and three of four flatfish in square mesh increased from 114 mm to 127 mm; and for five of five rockfish and three of four flatfish in diamond mesh increased from 114 mm to 127–140 mm (data presented as selectivity curves). In 1988–1990 the West Coast Groundfish Mesh Size survey tested experimental diamond mesh codends with mesh sizes of 76 (chosen as the ‘standard’ for analysis), 114, 127 and 140 mm and square mesh codends of 114 mm and 127 mm. Codends were towed in randomized blocks of two or three codends at a time during each fishing season by commercial trawling vessels, totalling 493 deployments.

    Study and other actions tested
  7. A replicated, randomized, controlled study in 1993–1994 in two seabed areas in the Aegean Sea, Greece (Stergiou 1999, same experimental set-up as Petrakis & Stergiou 1997, and Stergiou et al. 1997) found that found that a bottom trawl net with a square mesh codend retained more fish overall than a diamond mesh codend of the same mesh size but this varied with species shape and size, and both retained fewer fish compared to a diamond mesh codend of smaller mesh size. The average proportion of retained versus escaped catch (for 36 fish species and 1 invertebrate) was higher for the square mesh than diamond mesh of the same size, but lower than smaller diamond mesh (square: 0.63, diamond: 0.49, small diamond: 0.93 retained). In addition, there were differences in the retained proportions of individual fish species between the square and diamond mesh codends of the same mesh depending on fish shape (round- or flatfish) and size (small or large). Experimental trawl deployments were conducted in the Trikeri Channel in October 1993 (5 hauls) and the North Euboikos Gulf in March 1994 (seven hauls). A trawl net was randomly assigned either a 20 mm square mesh codend or a 14 or 20 mm diamond mesh codend, and towed for 45–60 min at depths between 73–210 m. Small mesh (10 mm) covers over the codend sampled the escaping fish catch. All individuals caught in the covers and codends were identified and counted.

    Study and other actions tested
  8. A replicated, paired, controlled study in 1994–1995 in two offshore areas of seabed in the North Atlantic Ocean, Canada (Halliday & Cooper 2000) found that square mesh codends improved the size-selectivity of a trawl net for silver hake Merluccius bilinearis, compared to diamond mesh codends. The estimated length at which 50% of hake were predicted to escape was higher in square mesh codends than diamond, and between square meshes was higher in the larger mesh size (square, 60 mm mesh: 26 cm, 55 mm mesh: 23 cm; diamond, 60 mm mesh: 16–19 cm). Data were collected on two chartered commercial inshore otter trawlers during five experimental surveys in the Emerald and LaHave basins (central Scotian Shelf) between July 1994 and March 1995. During each survey, one experimental codend (one survey each of 55 or 60 mm square mesh, and three surveys of 60 mm diamond mesh with or without a 89 mm chafer section – see paper for gear specifications) was towed on one boat parallel to a small mesh control codend (19 mm) on the other, for a total of 98 valid paired hauls in 180–265 m depth. In all experiments, a size-sorting escape grid was installed in front of the codend. Silver hake catches were subsampled for weight, and fish length (snout to the middle of the tail fin) recorded.

    Study and other actions tested
  9. A replicated, controlled study in 1993 in a fished area of seabed in the Atlantic Ocean off the south coast of Portugal (Campos et al. 2003) found that a square mesh codend improved the size-selectivity of a crustacean trawl net for three of three non-target fish, compared to diamond mesh codends of similar mesh size. The estimated length at which fish had a 50% chance of escape was greater with the square mesh than diamond meshes of increasing sizes for blue whiting Micromesistius poutassou (square: 30 cm, diamond: 23–27 cm). For horse mackerel Trachurus trachurus, the 50% escape length was greater with square mesh than the two smallest diamond mesh codends (square: 22 cm, diamond: 18–20 cm), but similar to the larger diamond mesh codend of 70 mm (22cm). In addition, for European hake Merluccius merluccius, the proportions of escapees below the minimum landing size relative to those retained were improved in the square mesh compared to the two smallest diamond meshes (see paper for data). However, since only very small proportions were retained overall for both the square and the largest diamond mesh, their size-selectivity was not calculable. Data were collected on two surveys in March/April and May 1993 from 133 deployments of a crustacean trawl (1 h) by a research vessel in depths of 152–706 m. A square mesh cod end of 55 mm mesh size (24 hauls), and three diamond mesh cod ends of 55 mm (41 hauls), 60 mm (33 hauls) and 70 mm (35 hauls) were tested. Covers fitted over each codend collected fish escaping through the meshes. Fish in both the codend and cover catches were separately identified, weighed, and total lengths measured.

    Study and other actions tested
  10. A replicated, controlled study in 1992 in a fished area of seabed in the Atlantic Ocean off the southwest coast of Portugal (Campos et al. 2003) reported that changing the configuration of mesh in a bottom trawl net to square from diamond resulted in lower retention and improved size selection of three of four commercial fish species. Across all hauls, the square mesh codend released more smaller and/or undersized individuals of European hake Merluccius merluccius, blue whiting Micromesistius poutassou and horse mackerel Trachurus trachurus than diamond mesh codends, but it was similar for four spot megrim Lepidorhombus boscii (data presented as size frequencies and selectivity curves). The length at which fish have a 50% chance of escape was higher in square mesh for hake (square: 25 cm, diamond: 17–19 cm). Almost all blue whiting escaped from the square mesh codend and all horse mackerel were retained in diamond mesh codends (compared to 40% escape in square mesh) meaning estimates of selectivity could not be calculated. Catch comparison surveys were done by a research vessel in August 1992 using a square mesh codend of 65 mm mesh size (10 hauls) and diamond mesh codends of mesh sizes of 65 mm (13 hauls), 70 mm (18 hauls) and 80 mm (19 hauls). Deployments were of 1 h, at 3.5 kn and in depths of 200–400 m. Covers fitted over each of the codends sampled fish escaping through the meshes. Codend and cover catches were weighed. All total lengths of hake and megrim were measured, and mackerel and whiting lengths sub-sampled.

    Study and other actions tested
  11. A replicated, randomized, paired, controlled study in 2002 in a seabed area in the Clarence Estuary (Tasman Sea), New South Wales, Australia (Broadhurst et al. 2004) found that square mesh codends fitted to prawn trawls reduced the catch numbers of unwanted young fish of three of six commercially important species compared to diamond mesh codends. Average catch numbers of non-target southern herring Herklotsichthys castelnaui, Tasmanian whitebait Lovettia sealii, and pink-breasted siphonfish Siphamia roseigaster were lower in square mesh than diamond mesh codends (square: 0–9 fish, diamond: 17–75 fish) and were similar for catfish Siluriformes, Ramsey’s perchlet Ambassis marianus, and silver biddies Gerreidae (square: 2–4 fish, diamond: 2–9 fish). In March 2002, experimental fishing was done on commercial prawn-trawl fishing grounds in Lake Woolooweyah using a commercial trawler. One of four designs of trawl codend were deployed on one side of a twin trawl, paired on the other side with small mesh control codends, all with Nordmøre escape grids (20 mm bar spacing): two square mesh codends (20 mm mesh, one tapered and one non-tapered), and two diamond mesh codends (40 mm, 100 or 200 meshes circumference). Twenty replicate hauls of each test codend/control were done. All catches were sorted and counted separately.

    Study and other actions tested
  12. A replicated, controlled study in 2005 in two areas of seabed in the Mediterranean Sea off eastern Spain (Bahamon et al. 2006) found that square mesh instead of diamond mesh codends improved the size-selectivity of commercially important European hake Merluccius merluccius, poor cod Trisopterus minutus and greater forkbeard Phycis blennoides in a multi-species bottom trawl fishery. Across all hauls, the selection length (the length at which 50% of fish are predicted to escape) was higher in the square mesh codend for hake (square: 16 cm, diamond: 10 cm), poor cod (square: 13 cm, diamond: 9 cm) and forkbeard (square: 15 cm, diamond: 10 cm). Commercial fishing deployments with both square and diamond mesh codends (40 mm) were conducted in July 2005 on the continental shelf (100 m, 19 hauls) and upper slope (400 m, 9 hauls) of the Catalan Sea. Tow duration was 15–157 minutes. A small mesh cover (15 mm) over each codend sampled escaped fish.

    Study and other actions tested
  13. A replicated, paired, controlled study in 2004 in two seabed areas in the South Pacific Ocean off New South Wales, Australia (Broadhurst et al. 2006) found that square mesh codends in a mixed species bottom trawl fishery reduced the catches of discarded whiting Sillago spp. compared to conventional diamond mesh codends. Data were reported as statistical model results. Results varied between vessels, and in one of three cases the number and weight of total discarded whiting was lower in square mesh codends (35 and 41 mm) than diamond (41 mm/150), but there were no differences between codends for retained total whiting, or species individual categories. For a second vessel, there were no clear reductions in any whiting catches between a 35 mm square mesh and diamond mesh (41 mm/150), but a 41 mm square mesh had lower weight of total retained whiting and lower number and weights of total and retained stout whiting Sillago robusta. For the third vessel, there were no main differences in whiting catches between square (31 mm) and diamond mesh (41 mm/100) codends. In addition, square mesh codends improved selection for stout whiting compared to diamond, and the length at which 50% of fish are predicted to escaped increased with increasing size of the square mesh (35 mm: 14–15 cm, 41 mm: 17–18 mm. Catch data were collected on three commercial prawn trawlers, fishing with the two outer nets of a triple trawl gear configuration, in April-December 2004. Seventy-one paired trawl deployments were carried out in 27–51 m depth using one of five test codends on one side - two square mesh codends (nominal 35 and 45 mm mesh), and three diamond mesh codends (two 41 mm mesh of 100 and 150 mesh circumference and one 45 mm mesh) - and a small mesh (24 mm) control codend on the other. All trawl nets also included a square mesh escape panel. See original study for gear details. The weights, numbers and total lengths of total, retained and discarded stout and red spot whiting Sillago flindersi were recorded.

    Study and other actions tested
  14. A replicated, controlled study in 2002–2003 of deep-water prawn fishing grounds in the Mediterranean Sea, Spain (Guijarro & Massuti 2006, same experimental set-up as Ordines et al. 2006) found that a square mesh codend reduced the amount of discarded fish catch, compared to a conventional diamond mesh codend. The proportion (by weight) of discarded non-commercial species was lower in the square mesh codend (3–11%, of which 80–93% were fish) compared to the diamond mesh codend (7–28%, of which 90–98% were fish), and was similarly decreased for discarded commercial species (square: 2–7%, of which 59–97% were fish; diamond: 7–17%, of which 45–99% were fish). In addition, no differences in commercial retained catch were found between mesh types and overall, the catch composition varied with depth and season. The total catch (weight) comprised fish (teleosts 55%, elasmobranchs 14%), crustaceans (28%), and cephalopods (6%). Catch comparison data were collected by commercial bottom trawler on a main crustacean fishing ground between 251–737 m depths, south of Mallorca. A total of 19 bottom trawl deployments each of square and diamond mesh codends (both 40 mm mesh) were done in September-October 2002 (18 hauls) and May-June 2003 (20 hauls). Deployments were 4.5 h at 2.5 knots. A small mesh (20 mm) cover over the codends sampled the escaping catch. All fish were identified, counted, and length measured.

    Study and other actions tested
  15. A replicated, controlled study in 2002–2003 in two areas of seabed on the continental shelf in the Mediterranean Sea, Spain (Ordines et al. 2006, same experimental set-up as Guijarro & Massuti 2006) found that using a square mesh instead of diamond mesh codend in a multi-species bottom trawl fishery reduced the amount of fish discarded in deeper but not shallower shelf areas, and size-selectivity was improved in both areas. Average catch biomass of total discarded fish (80–90% of which were non-commercial species) was lower in the square mesh codend on the deep shelf (square: 10, diamond: 20 kg/30 min) and similar between mesh shapes on the shallow shelf (square: 6, diamond: 10 kg/30 min). The length at which 50% of fish are predicted to escape, where reported, was higher in the square mesh in both shallow and deep areas for all fish (square: 7–29 cm, diamond: 2–19 cm; see paper for species individual data). Fishing deployments were conducted from a commercial trawler in September–October 2002 and May–June 2003 (same experimental set-up as Guijarro & Massuti, 2006), on the shallow (50–78 m, 12 hauls) and deep (147–189 m, 12 hauls) continental shelf off the Balearic Islands. Twelve hauls were carried out in each area: 6 each of a square and diamond mesh codend (both 40 mm mesh size). A small mesh (20 mm) cover installed over the codends sampled catch escaping through the meshes.

    Study and other actions tested
  16. A replicated, controlled study in 2003 of a fished area of seabed in the northwest Atlantic Ocean off New Hampshire, USA (He 2007) found that the effect of using square mesh instead of diamond mesh codends on the size-selectivity of bottom trawl nets for five important commercial fish species depended on body shape (roundfish or flatfish), but for both square and diamond codends, selectivity increased with larger mesh sizes. For cod Gadus morhua and haddock Melanogrammus aeglefinus there were no differences in the selection length (the length at which 50% of fish are predicted to escape) between square and diamond mesh of the same mesh size (cod, square: 59–69 cm, diamond: 59–66 cm; haddock, square: 54–57 cm, diamond: 55–61 cm); but for three flatfish, the selection lengths were smaller in square mesh codends (long rough dab Hippoglossoides platessoides, square: 33–35 cm, diamond: 39–40 cm; yellowtail flounder Limanda ferruginea, square: 34–38 cm, diamond: 40–42; witch flounder Glyptocephalus cynoglossus, square: 36–40 cm, diamond: 43–46 cm). For all species, an increase in selection length was found with increasing mesh size in both diamond and square mesh (see paper for individual data by species and codend). Catch data was collected in May–July 2003 from 86 fishing vessel deployments in the western Gulf of Maine (33–142 m depths). Five codends were tested: two square mesh (165 and 178 mm), and three diamond mesh (152, 165 and 178 mm). The five codends were tested in a preselected random order, each for up to six consecutive tows, for between 14–20 hauls. A small mesh cover over each codend sampled escaped fish. Codend and cover catches were counted and weighed by species, and fish total lengths measured (sub-sampled where necessary).

    Study and other actions tested
  17. A replicated, controlled study in 2004 in one shallow inshore and one deeper offshore seabed area in the Adriatic Sea, Italy (Sala et al. 2008) found that for five roundfish species, but not one flatfish species, using a square mesh codend improved the size selectivity of a bottom trawl net compared with a diamond mesh codend. Overall, the length at which 50% of fish are predicted to escape was greater in the square mesh codend for European hake Merluccius merluccius (square: 14 cm, diamond: 8 cm), red mullet Mullus barbatus (square: 11 cm, diamond: 8 cm), common pandora Pagellus erythrinus (square: 10 cm, diamond: 8 cm), Mediterranean horse mackerel Trachurus mediterraneus (square: 13 cm, diamond: 10 cm) and poor cod Trisopterus minutus capelanus (square: 11 cm, diamond: 8 cm), but was lower for one flatfish, scaldfish Arnoglossus laterna (square: 8 cm, diamond: 8 cm). Fishing surveys were done by research vessel on two fishing grounds in the Central Adriatic: in August–September 2004 (15–21 m depth, 5nm off Ancona), and in September–October 2004 (70 m depth, Western Pomo pit). Two trawl codends with the same mesh size (38 mm) but different mesh configuration (square and diamond mesh) were fished daily and alternately on the same trawl for a total of 48 deployments (21 shallow, 27 deeper). A small mesh cover (20 mm) attached over each codend collected the escaping fish catch. Catches in both the codends and covers were sampled by species and fish total length.

    Study and other actions tested
  18. A replicated, paired, controlled study in 2007 of two fished areas of seabed in the English Channel off southwest England, UK (Wade et al. 2009) found that beam trawl nets with square instead of diamond mesh codends, reduced the amount of discarded finfish catch. Across both sampling areas, the square mesh codends caught 30–52% fewer discarded finfish than the diamond (square: 1,496–1,830 fish, diamond: 2,124–3,836 fish). By individual fish species/groups, total numbers of four (all roundfish) of the nine most numerous were reduced in one or both areas by 18–80%, while for the rest (all flatfish), there were no differences between codend types for four and one was lower (by 56%) in the square mesh in the inshore area only. In addition, the retained finfish catches were similar between codend types in both areas (square: 943–985 fish, diamond: 948–1,005 fish). Catch comparison trials were done at two separate bottom fishing grounds off the south west coast of England by two commercial beam trawl vessels during 6 d sampling trips in July and August 2007. A total of 24 deployments were made of two beam trawl nets towed simultaneously: one an 80 mm square mesh codend, and one a standard 80 mm diamond mesh codend (see paper for specifications). Catches from both trawl nets were kept separate and divided into discarded and retained portions. Discarded finfish and all retained fish were identified, and their total lengths measured (sub-sampled where necessary).

    Study and other actions tested
  19. A replicated, controlled study in 2008–2009 of an area of seabed in an estuary off the Tasman Sea, Australia (Broadhurst et al. 2010) found that using a square mesh codend in a squid trawl net did not reduce the overall amount of discarded catch (fish and invertebrates), or improve the size selectivity for yellowtail scad Trachurus novaezelandiae and striped seapike Sphyraena obtusata, compared to a diamond mesh codend. Average numbers of total discarded catch (fish and invertebrate species combined – see paper for species caught) were similar for each of two sizes of square mesh codend (29 and 32 mm) compared to a 41 mm diamond mesh codend (29 mm square: 715 ind; 32 mm square: 250 ind; diamond: 300–500 ind). In addition, no statistical differences between square and diamond mesh codends were found in the length at which 50% of fish are predicted to escape for two fish species caught in the covers in sufficient quantities: yellowtail scad (square: 13 cm; diamond: 11–12 cm); and striped seapike (square: 20 cm; diamond: 15 cm). Catches from three different codends (29 and 32 mm square mesh, and 41 mm diamond mesh) were compared on a single-rigged trawler on a commercial squid Loliginidae spp. trawl ground in the Hawkesbury River estuary, in December 2008 and May 2009. All codends also had a 42 mm square mesh escape panel fitted in front. A total of between 10 and 12 deployments (75 min) of each codend were done. A small mesh (18 mm) hooped cover attached over each codend sampled the escaping catch.

    Study and other actions tested
  20. A replicated, paired, controlled study in 2009 of an area of seabed in the Aegean Sea, Turkey (Düzbastılar et al. 2010) found that the short-term survival of two of six fish species after escaping from bottom trawls was higher in square mesh codends compared to diamond mesh codends. Overall, average survival rate was greater in square than diamond mesh codends for escaped red mullet Mullus barbatus (square: 95% of 950 fish; diamond: 81% of 225 fish) and blotched picarel Spicara maena (square: 97% of 460 fish; diamond: 91% of 174 fish). For annular seabream Diplodus annularis (82 fish) and common pandora Pagellus erythrinus (46 fish), survival rate was 100% for both codend types. In addition, average brown comber Serranus hepatus post-escape survival was 97% (of 332 fish) and 95% (of 126 fish) for square and diamond mesh codends, whilst all 355 scaldfish Arnoglossus laterna did not survive. For all species, most mortality occurred in the first 48 h after escape. Six, 15-min experimental bottom trawl deployments were done by research vessel off the southern coast of Yassica Island, Izmir Bay, in October 2009: three using a square mesh and three a diamond mesh codend (both 40 mm). Small mesh (24 mm), hooped detachable covers fitted over each codend collected escaped fish and at the end of each deployment were detached, sealed, and deployed on the seabed. Fish were fed and survival monitored in the anchored covers for seven days by divers.

    Study and other actions tested
  21. A replicated, paired, controlled study in 2009 in bottom fishing grounds in the Bristol Channel, UK (Enever et al. 2010) reported that bottom trawl nets fitted with a square mesh codend caught fewer discarded fish compared to a standard diamond mesh codend, and the overall survival likelihood of skate Rajidae spp. post-capture was improved. Data were not statistically tested. Overall numbers of discarded fish in the 100 mm square mesh codend were 68% lower than the 80 mm diamond mesh codend (square: 2,241 fish, diamond: 7,056 fish), and ranged between 25% to 100% for individual species/groups (see paper for data). The proportion of skate given a good initial ‘health’ score after capture (equal to 86% chance of survival) as a proxy for survival likelihood) was 47% in the square mesh codend and 25% in the diamond mesh codend. Catch data was collected in June/July 2009 on a commercial twin-rigged bottom trawler at 35–65 m depth. Sixteen paired trawl deployments (3–5 kn) were done with an experimental 100 mm square mesh codend towed simultaneously with a conventional 80 mm diamond mesh codend. Separate assessment of the post-capture visual condition and survival of 278 small-eyed skate Raja microocellata was used to determine a three-point ‘health’ scale as an indicator of survival. The scale was used to assess the health of individuals of five skate species (see paper for details) as the nets came aboard, 358 skate from the square mesh and 754 from the diamond mesh codend.

    Study and other actions tested
  22. A replicated, controlled study in 2006–2007 in two fished areas of seabed in the Kattegat and the Skagerrak, Denmark (Frandsen et al. 2010) found that a standard square mesh codend improved the size-selectivity of a bottom trawl net for three roundfish species, but not one flatfish species, compared to a standard diamond mesh codend. Overall, the length at which 50% of fish were predicted to escape was greater in the square mesh compared to the diamond mesh codend for roundfish: Atlantic cod Gadus morhua (26–27 cm vs 15–17 cm), haddock Melanogrammus aeglefinus (26 cm vs 15 cm) and whiting Merlangius merlangus (33 cm vs 18 cm); but it was lower for plaice Pleuronectes platessa (square: 14–15 cm, diamond, 19–20 cm). Catch comparison trials were done on multi-species fishing grounds (Norway lobster Nephrops norvegicus, cod and plaice) on two commercial twin-trawl vessels in September 2006 (18 deployments) and August 2007 (6 deployments). Two codends were tested, towed simultaneously each haul, and interchanged between left and right sides: a commercial square mesh (70 mm) and a commercial diamond mesh (90 mm). Hauls were 1–4 h at 32–184 m depth. Small mesh covers (36 mm) were attached over each codend and collected fish escaping from the upper and lower parts of the codend in separate compartments.

    Study and other actions tested
  23. A replicated, controlled study in 2005 in one area of muddy-sandy seabed in the Adriatic Sea, Italy (Sala & Lucchetti 2010) found that a square mesh codend improved the size selectivity of a prawn trawl net for European hake Merluccius merluccius, blue whiting Micromesistius poutassou and poor cod Trisopterus minutus capelanus, compared to diamond mesh codends of standard and large circumferences. Across both surveys, the estimated lengths at which 50% of fish are predicted to escape were greater in hauls with the square mesh codend than the two diamond mesh codends for three commercially important species: hake (square: 12–16 cm; diamond: 8–11 cm), blue whiting (square: 14–18 cm; diamond: 11–15 cm), and poor cod (square: 10–13 cm; diamond: 6–10 cm). Catch comparison trials were done during two research vessel surveys in the Western Pomo pit (210 m depth; a Norway lobster Nephrops norvegicus fishing ground) in May and September 2005. Three codends were tested, all nominal 40 mm mesh size: a square mesh codend (70 meshes circumference); and two diamond mesh codends, one of conventional circumference (280 meshes) and one larger (326 meshes). Over the two surveys a total of 20 deployments were done with the square mesh codend, and 19 and 13 deployments with the standard and large diamond mesh codends, respectively. A cover attached over each codend sampled the catch escaping through the meshes.

    Study and other actions tested
  24. A replicated, controlled study in 2008–2009 in bottom fishing grounds in the Pacific Ocean, off Chile (Queirolo et al. 2011) found that the effect of crustacean trawl nets fitted with square mesh codends on reducing discarded fish catch, varied with mesh size as well as mesh configuration, compared to a reference diamond mesh codend. In one of two target fisheries for crustaceans, escape rates (by weight) of Chilean hake Merluccius gayi gayi and bigeye flounder Hippoglossina macrops (the two main non-target fish species) were higher in both 70 mm square mesh and diamond mesh (D70) codends than the 56 mm diamond (D56) mesh codend (hake, square: 36%, D70: 16%, D56: 0%; flounder, square: 28%, D70: 17%, D56: 1%). For the other target crustacean fishery, main non-target fish escape rates (by weight) were higher in a 56 mm square mesh codend than a 56 mm diamond mesh codend for eelpout Zoarcidae spp. (99 vs 26%) only, and were similar between the codends for: Chilean hake (3 vs 0%), aconcagua grenadier Coelorinchus aconcagua (17 vs 3%) and cardinalfish Apogonidae spp. (19 vs 4%). However, they were all higher in a 70 mm diamond mesh codend than the 56 mm diamond mesh codend. Retained and escaped catches were compared between four codends of different mesh size (56 or 70 mm) and mesh configuration (square or diamond), and a reference 56 mm diamond mesh codend (see paper for gear specifications). In total, 84 trawl deployments were made in December 2008 in traditional crustacean fishing grounds using commercial vessels. A small mesh (32 mm) cover attached over each codend during deployment collected the escaped catch.

    Study and other actions tested
  25. A replicated, paired, controlled study (year not stated) of a fished area of seabed in the Tasman Sea, Australia (Macbeth et al. 2012) found that a square mesh codend reduced the amounts of discarded total catch (fish and invertebrates) in two target prawn trawl fisheries compared to commercial diamond mesh codends, and the effect on individual categories of discarded fish catch varied between species or the target fishery. Overall numbers of total discarded catch (fish and invertebrates) were reduced by the square mesh codend in both target prawn fisheries, by 48% and 77% (square: 69–382 ind/h, diamond: 287–661 ind/h). For the eastern king prawn Melicertus plebejus fishery, unwanted or undersized catches by number of three of the seven main discarded fish species/categories were reduced in the square mesh codend (by 59–95%), one was higher, and the rest were similar between square and diamond mesh codends. In the fishery targeting school prawns Metapenaeus macleaya, unwanted fish catch of four of the six main discard species/categories were 84–99% lower in the square mesh codend, while there was no difference for the other two between codend types (see paper for individual data). Catch data were collected by observers on seven commercial prawn trawlers operating from four ports off New South Wales: from 42 paired deployments targeting eastern king prawns (41–68 m depth) and 13 targeting school prawns (6–10 m). Each vessel was supplied with a 35 mm square mesh codend with a composite square mesh escape panel to test against the different diamond mesh codends (each with industry-designed square mesh escape panels) being used on each vessel (see paper for gear details). The square mesh codend was towed simultaneously with the industry standard codend on the outer trawls of standard triple-gear trawl configurations. The year the study took place was not reported.

    Study and other actions tested
  26. A replicated, randomized, paired, controlled study in 2002 of a fished area of seabed in the Coral Sea, Australia (Courtney et al. 2014) found that a prawn trawl net with a square mesh codend reduced the overall amount of unwanted non-target catch (fish and invertebrates) compared to a conventional diamond mesh trawl, and the effect on individual fish species varied. Average catch rates by weight of total unwanted catch (fish and invertebrates, seven fish species accounting for 50% by weight) was lower in square mesh compared to diamond mesh codends, both with and without grids (square: 796–908 g/ha, diamond mesh: 1,114–1,150 g/ha). By individual fish species, five of the 40 species analysed had lower catch rates (three by over 90%) in square mesh than diamond mesh codends, without grids (square: 1–115 g/ha, diamond: 5–134 g/ha), one was higher (square: 29 g/ha, diamond: 60 g/ha) and there were no differences between codend types for the rest (see paper for spceis individual data). Over 10 days in July 2002, data were collected from 65 paired trawl deployments on deepwater eastern king prawn fishing grounds off the southeast Queensland coast. Four codends were tested: a 48 mm square mesh with and without a rigid escape grid (turtle excluder device), and a 45 mm diamond mesh codend with and without a grid. Codend designs were randomly assigned to one of the two outer trawl nets of a triple-rigged trawl every 12 hauls and towed simultaneously.

    Study and other actions tested
Please cite as:

Taylor, N., Clarke, L.J., Alliji, K., Barrett, C., McIntyre, R., Smith, R.K., and Sutherland, W.J. (2021) Marine Fish Conservation: Global Evidence for the Effects of Selected Interventions. Synopses of Conservation Evidence Series. University of Cambridge, Cambridge, UK.

Where has this evidence come from?

List of journals searched by synopsis

All the journals searched for all synopses

Marine Fish Conservation

This Action forms part of the Action Synopsis:

Marine Fish Conservation
What Works 2021 cover

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, mammals, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 18

Go to the CE Journal

Discover more on our blog

Our blog contains the latest news and updates from the Conservation Evidence team, the Conservation Evidence Journal, and our global partners in evidence-based conservation.


Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape Programme Red List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Bern wood Supporting Conservation Leaders National Biodiversity Network Sustainability Dashboard Frog Life The international journey of Conservation - Oryx British trust for ornithology Cool Farm Alliance UNEP AWFA Butterfly Conservation People trust for endangered species Vincet Wildlife Trust