Artículos

 

Effects of Nordmøre-grid angles, profiles and other industry-developed modifications on catches in an Australian penaeid-trawl fishery

 

Efectos de ángulos y perfiles de las rejillas Nordmøre, y otras modificaciones desarrolladas por la industria, sobre las capturas de una pesquería australiana de peneidos

 

Matt K. Broadhurst*, Damian J. Young and Cristiana Damiano

 

* NSW Fisheries Conservation Technology Unit National Marine Science Centre PO Box J321 Coffs Harbour, NSW 2450 Australia. *E-mail: mbroadhurst@nmsc.edu.au

 

Recibido en julio de 2003;
aceptado en septiembre de 2003.

 

Abstract

In response to claims that conventional Nordmøre-grids used in the Clarence River penaeid-trawl fishery were occasionally being fouled or blocked during fishing, two experiments were done to examine the utility of modified grids with different installation angles, sizes and profiles, and alternative industry-developed, behavioral-type bycatch reduction devices (BRDs). In experiment 1, a conventional Nordmøre-grid (600 mm in length, installed at 45°) was compared against a longer grid (1200 mm in length, installed at 21°) and two locally-developed, behavioral-type BRDs. While slightly more difficult to operate than the conventional design, the longer Nordmøre-grid retained similar quantities of penaeids and bycatch; in contrast, both behavioral-type BRDs retained significantly (up to five times) more bycatch and were considered ineffective for use in this fishery. In experiment 2, the conventional Nordmøre-grid was tested against three modified Nordmøre-grids, including a 600-mm-long grid with a curved profile and installed at 45°, and two grids, both 900 mm in length and installed at 28°, but with flat and curved profiles. Only the curved 900-mm Nordmøre-grid caught significantly fewer penaeids and bycatch, although both of the curved-profile grids appeared to facilitate the release of debris from the trawl. The results are discussed in terms of fishery-specific factors that probably contributed towards the performances of the various BRDs.

Key words: bycatch reduction, shrimp, prawn, selectivity, Nordmøre-grid.

 

Resumen

En respuesta al reclamo de que las rejillas Nordmøre convencionales usadas en la pesquería de peneidos en el Río Clarence (NSW, Australia) se obstruían o bloqueaban durante las operaciones de pesca, se llevaron a cabo dos experimentos para examinar la eficacia de rejillas modificadas (con diferentes ángulos de instalación, tamaños y perfiles) y de dispositivos de tipo conductual para la reducción de la fauna acompañante desarrollados por la industria. En el experimento 1, se comparó una rejilla Nordmøre convencional (600 mm de longitud, instalada a 45°) con una rejilla Nordmøre más larga (1200 mm de longitud, instalada a 21°) y con dos dispositivos conductuales desarrollados localmente. Aunque ligeramente más difícil de operar que la rejilla Nordmøre convencional, la rejilla más larga retuvo cantidades similares de peneidos y fauna acompañante; sin embargo, los dos dispositivos conductuales retuvieron considerablemente más fauna acompañante (hasta cinco veces más), por lo que fueron considerados ineficaces para este tipo de pesquería. En el experimento 2, se comparó la rejilla Nordmøre convencional con tres rejillas Nordmøre modificadas: la primera de 600 mm de longitud con perfil curvo e instalada a 45°, y las otras dos de 900 mm de longitud e instaladas a 28°, pero una con perfil plano y la otra con perfil curvo. Sólo la rejilla Nordmøre de 900 mm y perfil curvo capturó significativamente menos peneidos y fauna acompañante; no obstante, ambas rejillas con perfil curvo parecían facilitar la descarga de basura de la red de arrastre. Se discuten los resultados considerando factores específicos de la pesquería que probablemente influyeron en el comportamiento de los diversos dispositivos utilizados.

Palabras clave: reducción de la fauna acompañante, camarón, langostino, selectividad, rejilla Nordmøre.

 

Introduction

The capture and mortality of non-target organisms (termed bycatch) from prawn/shrimp trawling has received substantial attention over the past 20 years (for reviews, see Saila, 1983; Andrew and Pepperell, 1992; Alverson et al., 1994). In most fisheries, the negative perceptions associated with this issue have been mitigated through the legislation of physical modifications to the codends of trawls, collectively termed "bycatch reduction devices" or BRDs (for a review, see Broadhurst, 2000). Inherent variability among the characteristics of different fisheries has resulted in a plethora of BRDs, although all can be classified into two categories. The first includes those designed to mechanically exclude organisms larger than the targeted species, based on differences in their size (e.g., Isaksen et al., 1992; Broadhurst and Kennelly, 1996a); these designs comprise some sort of inclined grid that directs bycatch out of trawls through the bottom or top of the codend. The second category of BRDs have strategically-positioned openings that function by exploiting differences in species' behavior and physiology (e.g., Rulifson et al., 1992; Rogers et al., 1997; Broadhurst et al., 2002a) and are mostly used to exclude fish similar in size or smaller than the targeted species (Broadhurst, 2000).

BRDs from both categories are legislated for use in penaeid-trawl fisheries throughout New South Wales (NSW), Australia. Specifically, a rigid mechanical-type BRD called the Nordmøre-grid (sensu Isaksen et al., 1992) is used in estuaries and is effective in reducing up to 90% of unwanted bycatch, with no reduction in the catches of the targeted school prawn, Metapenaeus macleayi, or eastern king prawn, Penaeus plebejus (Broadhurst and Kennelly, 1996a). Various behavioral-type BRDs are similarly efficient in reducing the catches of small, unwanted fish from trawlers targeting eastern king prawns and other commercially-important species in oceanic waters (Broadhurst and Kennelly, 1996b; Broadhurst et al., 2002a). These BRDs are simple, involving large rectangular- and triangular-shaped openings or composite panels of square-shaped mesh located in the tops of codends. An important and common factor among these designs is their legislated positioning at less than 1.2 m anterior to the last row of meshes in the codend. At this location, the displacement of water in front of the catch greatly facilitates the escape of individuals of key fish species, but not the targeted penaeids (Broadhurst et al., 2002a).

There is a facility for the ongoing refinement and testing by commercial fishers of BRDs with the appropriate legislation in NSW. This has occurred across many areas, but particularly in the Clarence River trawl fishery (Broadhurst, 2000). Fishers working in this estuary have reported that, while there are few operational problems with the Nordmøre-grid, occasionally it can be fouled or blocked by organisms and/or weed and debris. Several fishers have addressed this problem by substituting the Nordmøre-grid with locally-designed, behavioral-type BRDs. These designs are similar in concept to those developed in other local (e.g., Broadhurst et al., 2002a) and overseas (e.g., Rulifson et al., 1992; Rogers et al., 1997) fisheries and comprise simple components which are not fouled like inclined grids. Although the use of these new industry-developed BRDs is widespread, no formal information is available on their effectiveness in reducing bycatch in the Clarence River.

The fouling and/or blocking of mechanical-type BRDs during trawling is a problem in several fisheries (e.g., Renaud et al., 1993; Larsen, 1996; Broadhurst et al., 2002b) and is attributed to the grid angle, which typically ranges between 35° and 50°, but more commonly is 45° (Christian and Harrington, 1987; Isaksen et al., 1992; Madsen and Hansen, 2001; Broadhurst et al., 2002b). Few published studies have examined the effects on catches associated with lowering the angles of grids and/or altering their profiles. Of the work that has been done, Isaksen et al. (1992) suggested that Nordmøre-grids installed in North Sea shrimp, Pandalus borealis, trawls at angles below 35° and above 50° had greater shrimp loss and blockage, respectively. More recently, Larsen (1996) suggested that grids installed at angles below 35° are less likely to be fouled owing to an increase in the movement of water (and therefore catch) through the escape opening. Further, because the dimensions of grids need to correspond to defined codend fishing circumferences (frequently approximated as 0.35 x the mesh opening x the number of meshes in circumference), lower angles require longer grids and therefore greater surface areas, which may facilitate the movement of catch and its separation (Riedel and DeAlteris, 1995).

Given the above, our aims in this paper were to formally compare the relative effectiveness of industry-developed, behavioral-type BRDs and lower angles, larger sizes and different profiles of the Nordmøre-grid on bycatch reduction in the Clarence River penaeid-trawl fishery. These aims were addressed in two experiments. The first involved a comparison of the existing conventional Nordmøre-grid installed at 45°, a long Nordmøre grid installed at 21° and two locally-developed, behavioral-type BRDs. The second experiment involved testing the same conventional design of Nordmøre-grid against more practical-sized grids (with and without different profiles).

 

Materials and methods

The two experiments were done in the Clarence River (29°26' S, 153°22' E) during February and March 2003 using a chartered commercial penaeid trawler (10 m in length). Two Florida Flyer trawls, each with a headline length of 7.32 m, were rigged in a standard twin-gear configuration (one on each side of the vessel) and towed at approximately 1.2 m s^-1 over sandy bottoms in depths ranging from 2 to 6 m. Both trawls were rigged with zippers (Buraschi S146R, 1.45 m in length) to facilitate the changing of codends.

All codends comprised two sections (for detailed descriptions on terminology used see Anon, 1978). The posterior section was made from 24-mm (all mesh sizes refer to stretched mesh opening) black knotless polyamide (PA) mesh netting (2.5-mm diameter [Φ] braided twine) hung on the bar (i.e., square-shaped mesh with a bar length of 12 mm) and measured 65 and 100 bars in length and circumference, respectively ((fig. 1a, b). The anterior section was made from 42-mm green knotted polyethylene (PE) netting (2.5-mm Φ twisted twine), 100 meshes in circumference and contained the various BRDs (see below). Zippers were attached to the leading edge of each anterior section.

Experiment 1: the conventional Nordmøre-grid vs a modified Nordmøre-grid and two behavioral-type BRDs

Four BRDs were constructed and installed into the anterior codend sections described above. The first and second designs were termed the "600 and 1200 Nordmøre-grids" (fig. 1a, b) øre-grid had the same profile, but was twice as long as the 600 Nordmøre-grid and sewn into the codend at an angle of approximately 21° (fig. 1b). This was the longest, practical-sized grid that could be used in the trawls. The third and fourth BRDs were industry-developed, behavioral-type designs, termed the "diamond" and "Clarence quality panel". Both BRDs had a diamond-shaped opening (11 x 11 bars) located on the top of the codend, 1.1 m in front of the last row of meshes in the posterior section (fig. 1c, d). The Clarence quality panel BRD had a second triangular opening (13 bars in length) located in front of the diamond opening and an oblique, wedge-shaped guiding panel sewn from the top of the codend to the bottom, terminating immediately below the diamond-shaped opening (fig. 1d). Commercial fishers hypothesized that this oblique guiding panel directed all organisms to the bottom of the codend and created an area of reduced flow that allowed fish to maintain position in the vicinity of the diamond and triangular openings, thereby increasing their probability of escape.

The 1200 Nordmøre-grid, diamond and Clarence quality panel BRDs were compared against the conventional 600 Nordmøre-grid in independent, paired hauls (using the twin-rigged trawls). The position and order of each codend was determined randomly and used in normal commercial hauls of 20-min duration between 0700 and 1500 h each day. Two replicate hauls of each treatment BRD against the 600 Nordmøre-grid were done on each day, providing a total of 10 replicate hauls for each pair over 5 days.

Experiment 2: the conventional Nordmøre-grid vs three modified Nordmøre-grids

Three modified designs of Nordmøre-grid, with different profiles and angles, were constructed and compared to the conventional 600 Nordmøre-grid (fig. 2). Like those tested in experiment 1, all grids were made from 10-mm Φ aluminum rod and with a bar spacing of 20 mm. The first modified grid was termed the "curved 600 Nordmøre-grid" and contained bars that had a flat profile throughout their lower 70% and an elevated and curved (radius of 50 mm) profile through their upper 30% (figs. 2a, 3). The upper curved section of the grid was designed to facilitate the removal of objects from the codend by preventing them from fouling across the upper frame of the grid. Like the 600 Nordmøre-grid, this grid was sewn into the codend at approximately 45°. The "900 and curved 900 Nordmøre-grids" were similar to the 600 and curved 600 Nordmøre-grids described above, but longer (i.e., 900 mm) and sewn into the codend at approximately 28° (fig. 2b, c). As in experiment 1, these three modified Nordmøre-grids were compared against the 600 Nordmøre-grid in independent, paired hauls. Two replicate tows of each modified grid and the 600 Nordmøre-grid were done on each day, providing a total of 14 replicate hauls over 7 days.

Data collected and statistical analyses

The data collected from all hauls included: the number and weight of total penaeids; the numbers and weights of school and eastern king prawns and a subsample (approximately 100 individuals per tow) of the carapace lengths (CL to the nearest 1 mm) of school prawns; the number and weight of bycatch; and the numbers of all commercially- and/or recreationally-important fish and their fork lengths (FL, to the nearest 5 mm). Where there were sufficient data for variables (defined as at least one individual in each of 8 replicate tows), these were analyzed with two-tailed, paired t-tests (P < 0.05). In experiment 1, catches of the three key fish species (yellowfin bream, Acantopagrus australis; silver biddy, Gerres subfasciatus; and southern herring, Herklotsichthys castelnaui) were combined to provide a larger data set. To examine the relative effectiveness of the various BRDs, the differences in catches (between each treatment BRD and the 600 Nordmøre-grid) for those variables that were non-zero in all tows were ln (x + 10² or 10⁴) transformed, tested for heteroscedasticity using Cochran's test and analyzed using appropriate analyses of variance (ANOVA) (BRDs and days were treated as fixed and random factors, respectively). Significant differences detected in these analyses were investigated by Student-Newman-Keuls (SNK) multiple comparisons of means. Where analyses provided similar results for weights and numbers of variables, only data concerning weights were included to conserve space. Size-frequencies of prawns and fish were combined across all tows and where there were sufficient data (>40 individuals in each codend), they were compared with two-sample Kolmogorov-Smirnov tests (P ≤ 0.05).

 

Results

Experiment 1: the conventional Nordmøre-grid vs a modified Nordmøre-grid and two behavioral-type BRDs

No significant differences were detected in the catches of penaeids between the 600 Nordmøre-grid and the three treatment BRDs (paired t-tests, P > 0.05, n = 10) (fig. 4a-c). Compared to the 600 Nordmøre-grid, the diamond and Clarence quality panel BRDs retained significantly greater weights (paired t-tests, P < 0.05 and 0.01, respectively, n = 10) and numbers (paired t-tests, P < 0.05, n = 10) of bycatch (mean catches up to five times greater) (fig. 4d, e). The diamond BRD also retained significantly more key fish combined (bream, southern herring and silver biddy) than did the 600 Nordmøre-grid (up to three times more, paired t-test, P < 0.05, n = 10) and although not significant, the Clarence quality panel BRD caught relatively similar numbers of these fish (paired t-test, P > 0.05, n = 10) (fig. 4f). ANOVA of the differences in catches between the treatment BRDs and the 600 Nordmøre-grid detected a significant interaction for the weight of bycatch and differences in the numbers of bycatch for the main effects of BRDs and days. Appropriate SNK tests of these means revealed no definitive order for the interaction, but showed that the Clarence quality panel and diamond BRDs retained similar numbers of bycatch, which were significantly greater than the 1200 Nordmøre-grid (P < 0.05) (fig. 4e; table 1); (fig. 3).

Two sample Kolmogorov-Smirnov tests failed to detect significant differences in the relative size-frequency distributions of school prawns retained in the codends containing the treatment BRDs and the 600 Nordmøre-grid (P > 0.05). The size-frequency compositions of bream retained in the Clarence quality panel and diamond BRDs and southern herring retained in the 1200 Nordmøre-grid and diamond BRD were significantly different to those retained in the 600 Nordmøre-grid (P < 0.05) (fig. 5a, b). In all cases, both behavioral-type BRDs retained proportionally more larger-sized fish. In contrast, the 1200 Nordmøre-grid retained significantly fewer large southern herring than did the 600 Nordmøre-grid (fig. 5b).

Experiment 2: the conventional Nordmøre-grid vs three modified Nordmøre-grids

Compared to the 600 Nordmøre-grid, the curved 900 Nordmøre-grid retained a significantly lower weight of school prawns (means reduced by 15.7%; paired t-test, P < 0.05, n = 14), number and weight of bycatch (by 33.4% and 25%, respectively; paired t-test, P < 0.05, n = 14) and number of silver biddy (by 42.5%; paired t-test, P < 0.01, n = 14) (fig. 6a-d, h). Paired t-tests failed to detect any other significant differences between the 600 and various treatment Nordmøre-grids (P > 0.05) (fig. 6). ANOVA of the differences between the 600 and treatment Nordmøre-grids detected significant differences in the numbers and weights of school prawns and the numbers of bycatch for the main effect of BRDs (table 2). SNK tests of these means showed that the curved 600 Nordmøre-grid retained significantly more school prawns and bycatch than the curved 900 Nordmøre-grid, but similar quantities as the 900 Nordmøre-grid (P < 0.05; fig. 6a-c). No significant differences were detected between the 900 and the curved 900 Nordmøre-grids for the catches of school prawns, while the curved 900 Nordmøre-grid retained less numbers of bycatch (fig. 6c).

Two sample Kolmogorov-Smirnov tests failed to detect any significant differences in size-frequency distributions of fish between the treatment Nordmøre-grids and the 600 Nordmøre-grid (P > 0.05). Differences were detected for school prawns, with the curved 900 and curved 600 Nordmøre-grids retaining significantly fewer larger-sized individuals (P < 0.05) (fig. 7).

 

Discussion

The results presented here support previous studies which have shown that mechanical-type BRDs generally are more effective at separating organisms than BRDs that exploit differences in species' behavior and/or physiology (e.g., Broadhurst et al., 1996; Madsen and Hansen, 2001). Further, by demonstrating that the majority of modifications to the size, installation angle and profile of the Nordmøre-grid did not negatively affect its performance, we have illustrated the robustness of this mechanical-type BRD for consistently excluding unwanted organisms, while maintaining catches of the targeted penaeids.

Compared to the behavioral-type BRDs examined in experiment 1, the commercially used 600 Nordmøre-grid retained up to 80% less total bycatch and 67% fewer individuals of the key fish combined (i.e., yellowfin bream, silver biddy and southern herring) (fig. 4d-f). These sorts of differences are similar to those observed during several previous experiments where the Nordmøre-grid was tested against codends with no BRD (e.g., Broadhurst et al., 1996; Broadhurst and Kennelly, 1996a). As an example, Broadhurst and Kennelly (1996a) showed that the 600 Nordmøre-grid significantly reduced the weight of total bycatch and number of bream by 76% and 67%, respectively. Such comparable differences suggest that few organisms (if any) escaped from the diamond and Clarence quality panel BRDs examined during experiment 1. This conclusion is supported by Broadhurst et al. (1996), who showed that few individuals escaped through large-mesh panels located at similar positions in the codends of Clarence River trawlers, but conflicts with the demonstrated utility of similar BRDs in other local and overseas fisheries (e.g., Broadhurst and Kennelly, 1996b; Rogers et al., 1997; Broadhurst et al., 2002a, 2002b).

The poor relative performance of the diamond and Clarence quality panel BRDs may be attributed to species- and operational-specific factors associated with the Clarence River fishery. Specifically, Broadhurst et al. (1996) suggested that some fish, such as yellowfin bream, may not display an active escape response upwards in codends, effectively precluding their escape through the openings in either BRD. A related hypothesis involves several operational characteristics of Clarence River trawlers that include a short tow duration and fast retrieval of trawls. Observations made in other studies have revealed that some small fish accumulate and maintain position in front of the catch during towing and then, owing to changes in tactile and visual stimuli, collectively escape through the openings of behavioral-type BRDs when the vessel is stopped and the trawl is slowly winched to the surface (e.g., Watson, 1989; Rogers et al., 1997). Clarence River trawlers retrieve their nets very quickly, from shallow water (mostly <10 m) and with no delay in haul back. This would limit the opportunity for any swimming fish to escape through openings in the codend.

These sorts of operational factors did not affect the performance of the Nordmøre-grid that consistently reduced bycatch and maintained catches of penaeids across most configurations and all angles of installation. These observations are comparable to those made by Larsen (1996) in a related study comparing a conventional Nordmøre-grid installed at approximately 48° with a longer grid installed at approximately 23° in North Sea shrimp trawls. In this earlier work, the longer grid similarly maintained catches of the targeted shrimp, but slightly reduced the bycatch. This was attributed to the larger area (allowing more opportunities for shrimp to pass between the bars) and a faster flow of water along the low-angled grid (facilitating the movement of fish through the escape exit). The 900 and 1200 Nordmøre-grids examined here did not similarly exclude more bycatch than the conventional grid, but this may be explained by the configuration of the guiding panel. Unlike the Nordmøre-grids used in North Sea shrimp trawls that have a guiding funnel which terminates up to 1 m anterior to the base of the grid, the designs used in the Clarence River have a panel that extends to the surface of the grid and is weighted with chain links (fig. 3). This means that all organisms exit the guiding panel against the surface of the grid and are sorted immediately, irrespective of the grid angle and any ancillary effects on water flow. Moving the posterior opening of the guiding opening forwards and/or removing the chain links would probably prevent small fish being forced through the bars and allow more to escape, but this may correspond to some loss of penaeids (Isaksen et al., 1992).

Like the changes to the angle and surface area of grids discussed above, altering the profile to include a curved upper section designed to prevent weed and debris from accumulating had no effect on catches retained in 600 Nordmøre-grid. However, the same modification to the 900 Nordmøre-grid resulted in significant and similar reductions in the weights of bycatch and school prawns (by 25% and 15%, respectively compared to the 600 Nordmøre-grid; fig. 6). Most fish escaping from this grid did so across all sizes, since Kolmogorov-Smirnov tests failed to detect any differences between BRDs. In contrast, more larger-sized school prawns appear to have escaped from the curved 900 Nordmøre-grid (fig. 7). In the absence of underwater video or direct observations, it is difficult to postulate reasons for these reductions in catches. One possible explanation is that this grid was not sufficiently buoyed (i.e., 6 x 100 mm floats) to compensate for the extra aluminum required in its construction and it may have twisted during fishing, randomly releasing some of the catch.

Despite the above, neither of the curved Nordmøre-grids had weed or objects fouled along their top edges (unlike the conventional design), and so commercial fishers have begun trialing various similar, low-angled and modified-profile grids. Such ongoing development and refinement to the Nordmøre-grid should eventually address the requirements of the Clarence River fishers. It is also likely that, while any subtle modifications may be specific to this particular fishery, the basic concepts examined here may warrant investigation in other fisheries that use similar mechanical-type BRDs.

 

Acknowledgements

Thanks are extended to Jennifer Rowling for technical assistance; Roger Larsen and Don and Barry Johnson for their advice and expertise; and Steve Kennelly, Charles Gray and Paul O'Conner for critically reviewing the manuscript.

 

References

Alverson, D.L., Freeberg, M.H., Murawski, S.A. and Pope, J.G. (1994). A global assessment of fisheries bycatch and discards. FAO Fish. Tech. Pap. No. 339, 233 pp.         [ Links ]

Andrew, N.L. and Pepperell, J.G. (1992). The by-catch of shrimp trawl fisheries. Oceanogr. Mar. Biol. Ann. Rev., 30: 527-565.         [ Links ]

Anon (1978). FAO Catalogue of Fishing Gear Designs. Fishing News Books, Surrey, England, 160 pp.         [ Links ]

Broadhurst, M.K. (2000). Modifications to reduce bycatch in prawn trawls: A review and framework for development. Rev. Fish Biol. Fish., 10: 27-60.         [ Links ]

Broadhurst, M.K. and Kennelly, S.J. (1996a). Rigid and flexible separator-panels in trawls that reduce the by-catch of small fish in the Clarence River prawn-trawl fishery, Australia. Mar. Freshwater Res., 47: 991-998.         [ Links ]

Broadhurst, M.K. and Kennelly, S.J. (1996b). Effects of the circumference of codends and a new design of square-mesh panel in reducing unwanted by-catch in the New South Wales oceanic prawn-trawl fishery, Australia. Fish. Res., 27: 203-214.         [ Links ]

Broadhurst, M.K., Kennelly, S.J. and Isaksen, B. (1996). Assessments of modified codends that reduce the by-catch of fish in two estuarine prawn-trawl fisheries in New South Wales, Australia. Fish. Res., 27: 89-111.         [ Links ]

Broadhurst, M.K., Kennelly, S.J. and O'Doherty, G. (1997). Specifications for the construction and installation of two by-catch reducing devices (BRDs) used in New South Wales prawn-trawl fisheries. Mar. Freshwater Res., 48: 485-489.         [ Links ]

Broadhurst, M.K., Kennelly, S.J. and Cray, C.A. (2002a). Optimal positioning and design of behavioural-type by-catch reduction devices involving square-mesh panels in penaeid prawn-trawl codends. Mar. Freshwater Res., 53: 813-823.         [ Links ]

Broadhurst, M.K., Kangas, M.I., Daminao, C., Bickford, S.A. and Kennelly, S.J. (2002b). Using composite square-mesh panels and the Nordmøre-grid to reduce bycatch in the Shark Bay prawn-trawl fishery, Western Australia. Fish. Res., 58: 349-365.         [ Links ]

Christian, P. and Harrington, D. (1987). Loggerhead turtle, finfish and shrimp retention studies on four excluder devices (TEDs). In: Proc. Nongame and Endangered Wildlife Symposium, 8-10 September 1987, Georgia DNR, Social Circle, GA, pp. 114-127.         [ Links ]

Isaksen, B., Valdemarsen, J.W., Larsen, R.B. and Karlsen, L. (1992). Reduction of fish by-catch in shrimp trawl using a rigid separator grid in the aft belly. Fish. Res., 13: 335-352.         [ Links ]

Larsen, R.B. (1996). Experiments with a new, larger type of fish/ shrimp separator grid with comparisons to the standard Nordmøre-grid. In: ICES CM 1996/B:1, Report of the study group on grid (grate) sorting systems in trawls, beam trawls, and seine nets, Woods Hole, Massachusetts, 13-14 April, 1996, pp. 68-80.         [ Links ]

Madsen, N. and Hansen, K.E. (2001). Danish experiments with a grid system tested in the North Sea shrimp fishery. Fish. Res., 52: 203-216.         [ Links ]

Renaud, M., Gitschlag, G., Klima, E., Shah, A., Koi, D. and Nance, J. (1993). Loss of shrimp by turtle excluder devices (TEDs) in coastal waters of the United States, North Carolina to Texas: March 1988-August 1990. Fish. Bull., 91: 129-137.         [ Links ]

Riedel, R. and DeAlteris, J. (1995). Factors affecting hydrodynamic performance of the Nordmøre grate system: A bycatch reduction device used in the Gulf of Maine shrimp fishery. Fish. Res., 24: 181-198.         [ Links ]

Rogers, D.R., Rogers, B.D., de Silva, J.A. and Wright, V.L. (1997). Effectiveness of four industry-developed bycatch reduction devices in Louisiana's inshore waters. Fish. Bull., 96: 552-565.         [ Links ]

Rulifson, R.A., Murray, J.D. and Bahen, J.J. (1992). Finfish catch reduction in South Atlantic shrimp trawls using three designs of by-catch reduction devices. Fisheries, 17: 9-19.         [ Links ]

Saila, S.B. (1983). Importance and assessment of discards in commercial fisheries. FAO Fish. Circ. No. 765, 62 pp.         [ Links ]

Watson, J. (1989). Fish behavior and trawl design: Potential for selective trawl development. In: C.M. Campbell (ed.), Proc. World Symposium on Fishing Gear and Fishing Vessels. Marine Institute, St. Johns, Canada, pp. 25-29.         [ Links ]