ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on April 1, 2008
ICES Journal of Marine Science: Journal du Conseil 2008 65(4):687-696; doi:10.1093/icesjms/fsn038
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Developing a beam trawl for sampling estuarine fish and crustaceans: assessment of a codend cover and effects of different sizes of mesh in the body and codend
1 NSW Department of Primary Industries, Cronulla Fisheries Research Centre of Excellence, PO Box 21, Cronulla, NSW 2230, Australia
2 NSW Department of Primary Industries, Fisheries Conservation Technology Unit, PO Box J321, Coffs Harbour, NSW 2450, Australia
Correspondence to D. Rotherham: tel: +61 2 95278411; fax: +61 2 95278576; e-mail: douglas.rotherham{at}dpi.nsw.gov.au
Rotherham, D., Broadhurst, M. K., Gray, C. A., and Johnson, D. D. 2008. Developing a beam trawl for sampling estuarine fish and crustaceans: assessment of a codend cover and effects of different sizes of mesh in the body and codend. – ICES Journal of Marine Science, 65: 687–696.An experiment was carried out in the Clarence River (New South Wales, Australia) to test the hypotheses that fish and crustacean catches in an experimental beam trawl were affected by a codend cover and the sizes of mesh in the body and codend. The cover had no obvious effects on the catches retained in the codend. Similarly, in comparisons between trawl bodies made from 26- and 41-mm diamond-shaped mesh, there were no differences in the assemblages of fish caught, or in the mean numbers entering the codends. For one species of fish (Acanthopagrus australis), however, there were differences in the proportions caught between the trawl bodies across different size classes. There was also some evidence to suggest that mesh size in the body of the trawl influenced the size selection of school prawns (Metapenaeus macleayi). For most finfish, there were no differences in catches between codends made from 20-mm and from 29-mm mesh hung on the bar (i.e. square-shaped mesh). In contrast, mesh size in the codend was important for the size selectivity of school prawns, with smaller carapace lengths at 50% retention in the 20-mm codend. We conclude that use of a 41-mm mesh in the body and a 20-mm square mesh in the codend of the beam trawl would be appropriate for future sampling with this gear in estuaries of New South Wales. A similar experimental approach to ours is needed in adapting the beam trawl to estuaries in other parts of the world, or in developing other types of research trawl.
Keywords: beam trawl, body effect, cover effect, fishery-independent surveys, sampling gear
Received 19 February 2007; accepted 8 February 2008; advance access publication 1 April 2008.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Conventional otter trawls have been used widely to sample populations and assemblages of demersal fish and crustaceans (Gunderson, 1993; Kennelly et al., 1993; Korsbrekke et al., 2001; Petrakis et al., 2001). Nevertheless, the use of these gears for scientific sampling is problematic because (i) catches of fish are often enhanced by the herding effects of the otter boards and rigging, and (ii) the horizontal opening of trawls can be highly variable, particularly in scaled-down adaptations of commercial gears (Gunderson and Ellis, 1986; Wardle, 1986). Although otter trawls are used in estuaries throughout the world (Stokesbury et al., 1999; Araújo et al., 2002; Richardson et al., 2006), full-scale commercial gears typically require large vessels, which often precludes sampling in relatively shallow water (West, 2002).
As sampling tools, beam trawls overcome many of the inherent inadequacies of otter trawls. Their rigid opening ensures a constant swept area, and the absence of otter boards limits any herding of fish (Gunderson and Ellis, 1986; Gunderson, 1993). Scaled-down versions of these gears also remain effective when deployed from small boats, and are ideal for sampling in shallow estuarine habitats. Consequently, small beam trawls (
1-m horizontal opening and 1–2-mm mesh opening) have been used extensively to sample juvenile fish and epifaunal crustaceans in seagrass beds (Gray and Bell, 1986; Loneragan et al., 1995; Guest et al., 2003). Although not commonly used in estuaries, larger beam trawls (3-m horizontal opening), specifically developed to target a wider size- and species-range of icthyofauna, have potential as sampling tools for both ecological and fishery-independent studies across a range of habitats and depths.
Understanding the effects of different configurations of gear is essential for the development of standardized sampling tools that are optimal, reliable, and cost-efficient (Kennelly and Craig, 1989; Rotherham et al., 2007). The size and configuration of mesh, especially in the codend, strongly affects the efficiency and selectivity of towed gears, and has been the focus of extensive research aimed at reducing unwanted catches in commercial otter trawl fisheries (Wileman et al., 1996; Millar and Fryer, 1999; Broadhurst, 2000). Comparatively fewer studies have quantified similar changes to other towed gears, such as beam trawls, used for research sampling (e.g. Mous et al., 2002). There is also limited information on the effects of changing the size of mesh used in other sections of towed gears on their retained catches (DeAlteris et al., 1990; Broadhurst et al., 2000, 2005), which has implications for both improving gear selectivity and developing useful and cost-effective sampling gears.
The general experimental procedures for assessing changes to the sizes or configurations of mesh used in towed gears involve either towing the treatment and control configurations to be examined in alternate or simultaneous hauls, or attaching fine-meshed covers to the treatment gears to retain the escaping organisms (Pope et al., 1975; Wileman et al., 1996). The covered-codend technique is generally preferred, providing that the cover does not affect either the performance of the gear or the behaviour of the targeted organisms (Wileman et al., 1996). Despite the widespread and routine use of codend covers on both towed and static gears, most studies have ignored the potential for any confounding effects on the efficiency or selectivity of the gear (but see Madsen and Holst, 2002; Macbeth et al., 2005). Nevertheless, such effects are real (Wileman et al., 1996), so require examination in any study that seeks to estimate selection using this methodology.
Here, we investigate the utility of a beam trawl (3-m horizontal opening) for sampling key species of estuarine fish and crustaceans. Specifically, we first tested the effects of a codend cover on the catching efficiencies of a beam trawl rigged with two different sizes of mesh in both the body and the codend. We then examined the effects of different sizes of mesh in these sections of the gear on the size and species selection of fish and crustaceans in the codend. The results from this experimental work were then used towards developing an appropriately configured beam trawl for future research surveys.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Study site and construction of gears
The study was done in the Clarence River, New South Wales (NSW; 29°27'S 153°12'E; see West, 2002, for a description of the study area) in November and December 2005 using a chartered commercial prawn trawler (10 m long), rigged to tow two trawls in a standard twin-gear configuration. Two identical, stainless-steel beam trawl frames, (3x0.8 m; Figure 1) were constructed, along with two trawl bodies of identical metric dimensions (headropes, 3.7 m; groundropes, 4.1 m), twine characteristics (green polyethylene [PE], 3-strand, twisted twine of
1.1 mm diameter) and mesh orientation (diamond-shaped), but different nominal mesh sizes (26 and 41 mm; Figure 2a).
|
|
The 41-mm mesh was chosen because it is similar to the minimum size of mesh used in commercial prawn trawls in estuaries of NSW (
40 mm) and is known to retain key species of fish and crustaceans of different size (Liggins and Kennelly, 1996; West, 2002). Although sizes of mesh other than 26 and 41 mm were available, differences in twine diameters precluded their comparison between trawl bodies. We also considered sizes of mesh smaller than 26 mm in the body of the trawl to be unsuitable because (i) sorting times would increase through retention of large numbers of very small, unwanted taxa (e.g. Acetes shrimp), and (ii) the trawl would be more difficult to tow through the water, which would increase the use of fuel and potentially affect the proper operation of the codend and cover. Buraschi S146R zippers 1-m long were sewn into the posterior ends of the trawl bodies to facilitate changing the codends and covers. Four codends (all 1-m long and in circumference) were constructed from knotless, polyamide (PA) netting (braided twine, 2.5 mm diameter) hung on the bar (i.e. square-shaped) (Figures 1 and 2b). As the geometry of conventional diamond-shaped mesh can be affected by several parameters, including the volume or weight of the catch (Robertson, 1986), square mesh was used so that each codend remained a fixed sampling unit. Two of the codends were made from 20-mm mesh (100 x 100 bars) and the other two from 29-mm mesh (70 x 70 bars) (Figure 2b). A zipper 1-m long was sewn into the anterior edge of each codend to allow attachment and removal from the trawl bodies.
Two identical, hooped covers were designed to fit over the codends (Figure 1). Both covers were made from 16-mm knotless PA, diamond-shaped mesh (0.9 mm diameter, 3-strand, twisted twine) and measured 200 meshes in both the transverse (T) and normal (N) directions. Three aluminium hoops (600 mm diameter) were distributed along the length of the cover immediately adjacent to the codend (Figure 1). Zippers were used to attach the anterior edge of the covers to the trawl bodies.
Experimental procedure
At the start of the study, the 26- and 41-mm trawl bodies were randomly assigned to either side of the trawler (to eliminate any potential biases). Four codend configurations (20- and 29-mm codends, with and without covers) were tested for each trawl body (Table 1). For each deployment, two identical examples of one, randomly selected codend configuration were zippered to the 26- and 41-mm trawl bodies, then towed for 10 min at a speed of 1.2 m s–1. Identical codends were attached to the different trawl bodies during each deployment so as to preclude the need for additional experiments (which are costly and may harm or kill more organisms than necessary), if it was later found that the cover had no significant effects on retained catches. Data from the covered codends could then be used to test the effects of mesh size in the body (26 vs. 41 mm, twin trawl) and codend (20 vs. 29 mm, alternate haul) on catches retained in the beam trawl.
|
The testing order of the codend configurations was randomized in blocks (each block consisting of one replicate of the four configurations). On each day, we attempted four replicate deployments of each treatment configuration (i.e. 4 blocks x 4 codends). On the first day, however, only three replicate blocks were completed, resulting in a total of 15 replicates of each treatment configuration for each trawl body.
After each tow, the contents of the codends and covers (where appropriate) were emptied onto a partitioned tray and sorted by species. Data collected included the total numbers of individuals of each species, and the sizes of economically important fish (fork length, FL, to the nearest 0.5 cm), crabs (carapace width, to the nearest mm), and prawns (carapace length, CL, to the nearest millimetre). When the catch of prawns in a tow was large, the total weight and a subsample of 150 prawns were used to estimate total numbers, then CL measurements taken.
Analyses of data
Two-factor, orthogonal analyses of variance (ANOVA) were used to test the null hypotheses of no differences in retained catches (analysed for variables that included the total numbers of individuals and species, and selected abundant species) between (i) codend configurations (covered vs. uncovered; fixed factor) and days (random factor), separately for each codend (20 and 29 mm) attached to each trawl body, and (ii) covered codends (20 vs. 29 mm; fixed factor) and days (random factor) separately for each trawl body (26 and 41 mm). To ensure balanced analyses, data from day 1 were excluded.
Before ANOVAs, data were ln(x+1) transformed, to model treatment effects as approximately multiplicative. Data for prawns (Metapenaeus macleayi) from the comparison between 20- and 29-mm codends were, however, expressed as the proportion retained in each codend (i.e. not passing through the meshes of the codend and into the cover) and transformed to arcsin of the square root of the proportion. All transformed data were then tested for heteroscedasticity using Cochran’s test. For data that remained heterogeneous, 
was set to 0.01 to reduce the risk of type 1 error (Underwood, 1997). Multiple comparisons among means were carried out with Student–Newman–Keuls (SNK) tests. Differences between days were noted, but not investigated further. Where interaction terms were not significant at p > 0.25, they were pooled with the residual to increase the power of the test for the main effect of codend configuration.
Paired t-tests (two-tailed) were used to test the null hypothesis of no differences in the numbers and species of fish (total numbers of fish and species and selected abundant species) caught by the two trawl bodies (26 vs. 41 mm) attached to each of the covered codends (20 and 29 mm). Catches were expressed as totals in the codends and covers.
Differences in the structures of fish assemblages between (i) codend configurations (covered vs. uncovered), separately for each codend (20 and 29 mm) attached to each trawl body (26 and 41 mm), (ii) trawl bodies (26 vs. 41 mm), separately for each covered codend (20 and 29 mm), and (iii) codends (20 vs. 29 mm), separately for each trawl body (26 and 41 mm) were investigated using non-parametric multivariate analyses from the PRIMER 5 package (Version 5.2.2, PRIMER-E Ltd, 2001). Data were log(x+1) transformed to reduce the influence of abundant species and ordination plots generated from non-metric multidimensional scaling (nMDS) of Bray–Curtis similarity matrices. Analyses of similarities (ANOSIM) were used to test the a priori hypothesis that the structure of fish communities differed between covered and uncovered codends, and trawl bodies and codends with different sizes of mesh.
Differences in the size frequency distributions for abundant, economically important species were tested across all appropriate combinations of treatments using Kolmogorov–Smirnov (K–S) tests. For each replicate haul of a trawl configuration with a covered codend (see Table 1), the size frequencies of prawns (M. macleayi) caught in each replicate of the codend and cover were scaled where necessary, then vertically stacked (see Millar et al., 2004). Parametric selection curves (logistic and Richards) were then fitted using maximum likelihood. To account for between-haul variation, standard errors of the selectivity parameter estimates (i.e. CL at 50% retention, L50, and selection range, SR) were corrected using the replicate estimate of dispersion (Millar and Fryer, 1999; Millar et al., 2004). The most appropriate model was determined by likelihood-ratio tests and visual examination of residual plots. The bivariate form of Walds F-test was used to test for differences between selection curves for comparisons of mesh size in the (i) trawl body for each codend (26B–20CC vs. 41B–20CC; 26B–29CC vs. 41B–29CC), and (ii) codend for each trawl body (26B–20CC vs. 26B–29CC; 41B–20CC vs. 41B–29CC). To preserve an overall type I error rate of 5%, a Bonferroni corrected p-value of 0.0125 (0.05/4) was used for each comparison.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Numbers of fish and species
In all, 8241 fish and crustaceans (5634 and 2607 individuals in the codends and covers, respectively), consisting of >26 species, were caught during the experiment (Table 2). Four species accounted for >90% of the total catch: M. macleayi (65%), Gerres subfasciatus (11%), Herklotsichthys castelnaui (7%), and Acanthopagrus australis (7%). Catches in the codend covers mainly consisted of M. macleayi (
95%) and Ambassid spp. (2%). Only small numbers (<20 individuals) and proportions (<0.1%) of most economically important finfish species passed through the meshes in the codends and into the covers. These were mainly small fish, including juveniles of species with low dorsal profiles (e.g. Hyperlophus vittatus, Pomatomus saltatrix, and H. castelnaui). For other finfish species (e.g. A. australis and Rhabdosargus sarba), no individuals passed into the covers.
|
Assessments of the codend cover
ANOVA did not detect any significant differences in retained catches between the covered and uncovered codends attached to either of the trawl bodies, for any of the variables examined (Table 3). Similarly, there were no differences in the structures of fish assemblages between gear configurations (covered and uncovered codends) and days for any combinations of the trawl bodies and codends (ANOSIM, p > 0.25). Ordinations generated from nMDS scaling supported these results, because samples did not separate into distinct groupings for any codend or net body. For brevity, ordination plots are not shown for these or for any other multivariate analyses below. Always, stress values of the ordinations ranged between 0.13 and 0.18, which are generally considered adequate for reliable representations of data (Clarke, 1993).
|
K–S tests comparing the size frequencies of abundant economically important fish (G. subfasciatus, A. australis, and H. castelnaui) and crustaceans (M. macleayi) between covered and uncovered codends detected significant differences (Table 4). Nevertheless, results were inconsistent among species and across the different codends and trawl bodies (Table 4). For example, (i) the size frequency distributions and proportions of M. macleayi were variable between treatments (Figure 3a–c), (ii) the uncovered and covered codends caught greater proportions of small and large individuals of A. australis, respectively, but only for the 29-mm codend attached to the 26-mm trawl body (Figure 3d), and (iii) the uncovered codends caught greater proportions of small G. subfasciatus and H. castelnaui than the covered codends, but only for the 26-mm trawl body (Figure 3e; for brevity, data are only shown for G. subfasciatus).
|
|
Effects of mesh size in the body of the beam trawl
There were no significant differences in catches retained in the codends and covers between the 26- and 41-mm trawl bodies, for any of the variables examined (paired t-tests, p > 0.05; Table 5). These results were also consistent for both the small- and large-meshed codends. Multivariate analyses failed to detect significant differences between the trawl bodies for either codend (ANOSIM, p > 0.25). These results were supported by a lack of species groupings in the two nMDS ordinations.
|
K–S tests comparing the size frequency distributions of A. australis from the 26- and 41-mm trawl bodies detected significant differences for the 29-mm codend (the 26- and 41-mm bodies retained greater proportions of small and larger fish, respectively; Figure 4). No other significant differences were detected (p > 0.05).
|
Effects of mesh size in the codend
ANOVA detected a significant difference between codends (20 vs. 29 mm) for the mean number of H. castelnaui caught with the 26-mm trawl body (Table 6). Subsequent SNK tests showed that catches of H. castelnaui were significantly greater in the 20-mm codend (Figure 5). No other differences between codends were detected by ANOVA for the remaining variables analysed (Table 6; Figure 5). There was, however, a significant difference between days for G. subfasciatus (41-mm trawl body).
|
|
There were no significant differences in the assemblages of fish and crustaceans between codends for either trawl body (ANOSIM, p > 0.05) and there were no clear groupings of samples in nMDS ordinations. K–S tests comparing size frequency distributions of fish between codends detected significant differences in the size compositions of A. australis for the 41-mm trawl body. There were, however, no clear patterns in the range of sizes or proportions of fish caught for this species (not shown for brevity). No other significant differences were detected by K–S tests (p > 0.05).
Size selection of M. macleayi
Size-selectivity curves were successfully converged for each codend attached to each trawl body. The logistic model was used always because there were no significant reductions in deviances with the Richards model (likelihood-ratio tests; p > 0.05). The selection curves representing the comparison between codends for each net body (i.e. 26B–20CC vs. 26B–29CC, and 41B–20CC vs. 41B–29CC) were significantly different (Wald’s tests, p < 0.001; Figure 6, Table 7). There were, however, no significant differences between selection curves representing the comparison between trawl bodies for each codend (i.e. 26B–20CC vs. 41B–20CC, and 26B–29CC vs. 41B–29CC; Wald’s tests, p > 0.01, Bonferroni corrected). Nevertheless, although the estimated L50s for school prawns caught with different trawl bodies were similar for both the 20- and 29-mm codends, the SRs of the 26-mm trawl bodies were slightly wider than for the 41-mm trawl bodies (Figure 6, Table 7).
|
|
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Our study has provided evidence to support the utility of the covered codend method for assessing the selectivity of an experimental beam trawl. Equally important, by testing hypotheses concerning the sizes of mesh used in the beam trawl we have contributed towards (i) understanding the influence of different sections of this gear on its selectivity, (ii) demonstrating the practical importance of pilot studies for developing cost-effective and reliable sampling tools, and (iii) developing an appropriate configuration for sampling estuarine fish and crustaceans in southeastern Australia.
The use of rigid hoops in covernets can alleviate the physical masking of codend meshes (Wileman et al., 1996). Nevertheless, the presence of hooped covers can still potentially affect organisms entering the trawl and escaping through codend meshes owing to visual stimuli, changes in water flow through the gear, or both (Dahm et al., 2002; Macbeth et al., 2005). For example, any cover-induced reduction in the flow of water through a trawl may: (i) result in fewer numbers of smaller organisms being washed through the codend meshes and into the cover, and (ii) because of concomitant increases in anteriorly displaced water, allow larger actively swimming fish to avoid capture. Such effects may manifest as differences in the numbers, proportions, and sizes of individuals caught in comparisons between covered and uncovered codends.
We found no differences between covered and uncovered codends for the mean numbers of fish and species retained, or the structures of their assemblages. Further, differences in the size frequency distributions of abundant economically important species were inconsistent among treatments and species, which probably reflected spatial and temporal variability among replicate tows, rather than the mechanisms described above. These results indicate that the cover had minimal, if any, effects on the sampling performance of the beam trawl, irrespective of the sizes of mesh in the body or codend. Similar results have been reported in the few studies that have included assessments of cover effects on other towed (Madsen and Holst, 2002) and static (Macbeth et al., 2005) gears.
Changing the size of mesh in the body of the beam trawl had no effect on catches entering the codend for most of the variables examined. Similarly, there were no differences in the structure of assemblages between trawl bodies. For one species (A. australis), however, greater proportions of larger individuals were retained in the 29-mm codend when it was attached to the 41-mm trawl body. This result may indicate that the 26-mm trawl body enhanced avoidance reactions in larger individuals of this species, possibly owing to more visual impact, some reduction in relative water flow through the meshes, or both. Alternatively, the difference may reflect natural between-haul variation, given that (i) the smaller-meshed trawl body also caught greater proportions of smaller individuals of A. australis (which were between 8 and 13 cm FL and too large to escape through the meshes of the 41-mm trawl body), and (ii) differences were inconsistent between small- and large-meshed codends.
Several of the few previous studies comparing mesh sizes in the bodies of trawls have demonstrated no differences in the retained catches of mobile finfish (e.g. DeAlteris et al., 1990; Broadhurst et al., 2005). Nevertheless, some results are inconsistent between different studies and have been related to species-specific behavioural responses of fish or fishery-specific operational characteristics of the gears (Broadhurst et al., 2000, 2005). In contrast, several studies have demonstrated significant influences of the bodies of both otter and beam trawls on the size and species selection of crustaceans (e.g. Thorsteinsson, 1981; Hillis and Earley, 1982; Dremière et al., 1999; Polet, 2000).
We found no differences in the selectivity parameters of M. macleayi between codends attached to the 26- and 41-mm trawl bodies examined here, although SRs were slightly wider for both codends when they were attached to the 26-mm trawl body. As for other penaeids (e.g. Penaeus latisulcatus; Broadhurst et al., 2000), this result probably reflected some influence of the trawl body on the size selection of M. macleayi. Further research, perhaps using covers or pockets placed strategically over the body of the trawl (e.g. Polet, 2000), is required to validate these observations. The estimates of L50 that we obtained for M. macleayi were, however, similar to those estimated for this species using the same codends (i.e. 20- and 29-mm square mesh) attached to other types of gear (e.g. single and twin otter trawls; Broadhurst et al., 2004; Macbeth et al., 2004).
It is generally considered that most of the selection in towed gears occurs in the codend (DeAlteris et al., 1990; Wileman et al., 1996), and many studies have investigated the differences in catches between codends comprising different sizes and configurations of mesh (Broadhurst, 2000). In our study, mesh size in the codend was important for the size selectivity of school prawns, evidenced by the significant differences in parameter vectors, and in particular the lower estimate of L50 for the 20-mm codend. There were, however, no differences in the structure of assemblages between codends. Moreover, we found that, for most finfish, both codends were virtually non-selective for the sizes caught at the sampled sites, because only few individuals escaped through codend meshes and were retained in the cover. Therefore, the difference between codends for numbers of H. castelnaui (a species known to form large schools; Kuiter, 1996) was most likely the result of spatial and temporal variability among replicate tows.
We conclude that mesh sizes of 41 mm in the body and 20 mm in the codend of a beam trawl are appropriate for sampling the relative abundance, diversity, and sizes of fish and crustaceans caught in this study. The species caught in this short-term experiment are a typical subset of the fauna caught in towed gears in estuaries of NSW (Liggins and Kennelly, 1996; West, 2002). The beam trawl developed here may not, however, be appropriate for sampling species that (i) were not encountered during the experiment, and (ii) occur in estuaries elsewhere in the world. In fact, representative sampling of multispecies assemblages often requires several methods (Olin and Malinen, 2003), which should be developed for each specific region using properly designed pilot experiments (Andrew and Mapstone, 1987; Underwood, 1997; Rotherham et al., 2007). Our pilot work in developing fishery-independent sampling tools in estuaries of NSW has also included the use of multi-mesh gillnets (Gray et al., 2005; Rotherham et al., 2006).
Testing the effects of different configurations of gear is only one part of a broader strategy for developing scientific sampling tools for fishery-independent surveys (see Rotherham et al., 2007). Before starting large-scale, long-term surveys with a beam trawl, additional experiments are needed to test the effects of different sampling practices (e.g. tow duration, diel period) on retained catches. Moreover, determining optimal levels of replication for future surveys requires an understanding of spatial and temporal variation in fish fauna, across appropriate scales and strata.
|
Acknowledgements |
|---|
This work is part of a cooperative study between the Wild Fisheries Programme of the NSW Department of Primary Industries, the Centre for Research on Ecological Impacts of Coastal Cities at the University of Sydney, and the Australian Fisheries Research and Development Corporation (Project 2002/059). Sampling was carried out under NSW DPI Animal Care and Ethics Authority 02/15. We thank Alan Bodycote for constructing the beam trawls and for the use of his vessel and expertise, and Will Macbeth and Russell Millar for assisting with analyses and for providing helpful comments that improved the manuscript. Steve Kennelly, along with the anonymous referees, are also thanked for their input.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Andrew N. L., Mapstone B. D. Sampling and the description of spatial pattern in marine ecology. Oceanography and Marine Biology: An Annual Review (1987) 25:39–90.
Araújo F. G., De Azevedo M. C. C., Silva M. A., Pessanha A. L. M., Gomes I. D., Da Cruz-Filho A. G. Environmental influences on the demersal fish assemblages in the Sepetiba Bay, Brazil. Estuaries (2002) 25:441–450.[Web of Science]
Broadhurst M. K. Modifications to reduce by-catch in prawn-trawls: a review and framework for development. Reviews in Fish Biology and Fisheries (2000) 10:27–60.[CrossRef][Web of Science]
Broadhurst M. K., McShane P. E., Larsen R. B. Effects of twine diameter and mesh size in the body of prawn trawls on bycatch in Gulf St Vincent, Australia. Fishery Bulletin US (2000) 98:463–473.
Broadhurst M. K., Millar R. B., Kennelly S. J., Macbeth W. G., Young D. J., Gray C. A. Selectivity of conventional diamond- and novel square-mesh codends in an Australian estuarine penaeid-trawl fishery. Fisheries Research (2004) 67:183–194.[CrossRef][Web of Science]
Broadhurst M. K., Young D. J., Gray C. A., Wooden M. E. L. Improving selection in south eastern Australian whiting (Sillago spp.) trawls: effects of modifying the body, extension and codend. Scientia Marina (2005) 69:301–311.[Web of Science]
Clarke K. R. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology (1993) 18:117–143.[CrossRef][Web of Science]
Dahm E., Wienbeck H., West C. W., Valdemarsen J. W., O’Neill F. G. On the influence of towing speed and gear size on the selective properties of bottom trawls. Fisheries Research (2002) 55:103–119.[CrossRef][Web of Science]
DeAlteris J. T., Castro K. M., Testaverde S. A. Effect of mesh size in the body of a bottom trawl on the catch retained in the codend. (1990) Proceedings of the Fisheries Conservation Engineering Workshop: Rhode Island. 12–18.
Dremière P. V., Fiorentini L., Cosimi G., Leonori I., Sala A., Spagnolo A. Escapement from the main body of the bottom trawl used for the Mediterranean international trawl survey (MEDITS). Aquatic Living Resources (1999) 12:207–217.[CrossRef][Web of Science]
Gray C. A., Bell J. D. Consequences of two common techniques for sampling vagile macrofauna associated with the seagrass Zostera capricorni. Marine Ecology Progress Series (1986) 28:43–48.[CrossRef][Web of Science]
Gray C. A., Jones M. V., Rotherham D., Broadhurst M. K., Johnson D. D., Barnes L. M. Utility and efficiency of multi-mesh gill nets and trammel nets for sampling assemblages and populations of estuarine fish. Marine and Freshwater Research (2005) 56:1077–1088.[CrossRef][Web of Science]
Guest M. A., Connolly R. M., Loneragan N. R. Seine nets and beam trawls compared by day and night for sampling fish and crustaceans in shallow seagrass habitat. Fisheries Research (2003) 64:185–196.[CrossRef][Web of Science]
Gunderson D. R. Surveys of Fisheries Resources. (1993) New York: John Wiley. 248.
Gunderson D. R., Ellis I. E. Development of a plumb staff beam trawl for sampling demersal fauna. Fisheries Research (1986) 4:35–41.[CrossRef][Web of Science]
Hillis J. P., Earley J. Selectivity in Nephrops trawl. ICES Document CM 1982/B: 19. (1982) 11.
Kennelly S. J., Craig J. M. Effects of trap design, independence of traps and bait on sampling populations of spanner crabs Ranina ranina. Marine Ecology Progress Series (1989) 51:49–56.[CrossRef][Web of Science]
Kennelly S. J., Graham K. J., Montgomery S. S., Andrew N. L., Brett P. A. Variance and cost–benefit analyses to determine optimal duration of tows and levels of replication for sampling relative abundances of species using demersal trawling. Fisheries Research (1993) 16:51–67.[Medline]
Korsbrekke K., Sigbjørn M., Nakken O., Pennington M. A survey-based assessment of the northeast Arctic cod stock. ICES Journal of Marine Science (2001) 58:763–769.
Kuiter R. H. Guide to Sea Fishes of Australia: A Comprehensive Reference for Divers and Fishermen. (1996) Sydney: New Holland Publishers. 434.
Liggins G. W., Kennelly S. J. By-catch from prawn trawling in the Clarence River estuary, New South Wales, Australia. Fisheries Research (1996) 25:347–367.[CrossRef][Web of Science]
Loneragan N. R., Wang Y. G., Kenyon R. A., Staples D. J., Vance D. J., Heales D. S. Estimating the efficiency of a small beam trawl for sampling tiger prawns Penaeus esculentus and P. semisulcatus in seagrass by removal experiments. Marine Ecology Progress Series (1995) 118:139–148.[CrossRef][Web of Science]
Macbeth W. G., Broadhurst M. K., Millar R. B. The utility of square-mesh codends in Hawkesbury River penaeid trawls. Ecological Management and Restoration (2004) 5:221–224.[CrossRef]
Macbeth W. G., Broadhurst M. K., Millar R. B. Improving selectivity in an Australian penaeid stow-net fishery. Bulletin of Marine Science (2005) 76:647–660.[Web of Science]
Madsen N., Holst R. Assessment of the cover effect in trawl codend selectivity experiments. Fisheries Research (2002) 56:289–301.[CrossRef][Web of Science]
Millar R. B., Broadhurst M. K., Macbeth W. G. Modelling between-haul variability in the size selectivity of trawls. Fisheries Research (2004) 67:171–181.[CrossRef][Web of Science]
Millar R. B., Fryer R. J. Estimating the size-selection curves of towed gears, traps, nets and hooks. Reviews in Fish Biology and Fisheries (1999) 9:89–116.[CrossRef][Web of Science]
Mous P. J., van Densen W. L. T., Machiels M. A. M. The effect of smaller mesh sizes on catching larger fish with trawls. Fisheries Research (2002) 54:171–179.[CrossRef][Web of Science]
Olin M., Malinen T. Comparison of gillnet and trawl in diurnal fish community sampling. Hydrobiologia (2003) 506–509:443–449.
Petrakis G., Maclennan D. N., Newton A. W. Day–night and depth effects on catch rates during trawl surveys in the North Sea. ICES Journal of Marine Science (2001) 58:50–60.
Polet H. Codend and whole trawl selectivity of a shrimp beam trawl used in the North Sea. Fisheries Research (2000) 48:167–183.[CrossRef][Web of Science]
Pope J. A., Margetts A. R., Hamley J. M., Akyuz E. F. Manual of methods for fish stock assessment. 3. Selectivity of fishing gear. FAO Fisheries Technical Report (1975) 41:65.
Richardson N., Whitfield A. K., Paterson A. W. The influence of selected environmental parameters on the distribution of the dominant demersal fishes in the Kariega Estuary channel, South Africa. African Zoology (2006) 41:89–102.[CrossRef][Web of Science]
Robertson J. H. B. Design and construction of square mesh codends. Scottish Fisheries Information Pamphlet (1986) 12:10.
Rotherham D., Gray C. A., Broadhurst M. K., Johnson D. D., Barnes L. M., Jones M. V. Sampling estuarine fish using multi-mesh gill nets: effects of panel length and soak and setting times. Journal of Experimental Marine Biology and Ecology (2006) 331:226–239.[CrossRef][Web of Science]
Rotherham D., Underwood A. J., Chapman M. G., Gray C. A. A strategy for developing scientific sampling tools for fishery-independent surveys of estuarine fish in New South Wales, Australia. ICES Journal of Marine Science (2007) 64:1512–1516.
Stokesbury K. D. E., Bichy J. B., Ross S. W. Selectivity and efficiency of two otter trawls used to assess estuarine fish and macroinvertebrate populations in North Carolina. Estuaries (1999) 22:882–888.[CrossRef][Web of Science]
Thornsteinsson G. The effect of net slack in the sidepanels of shrimp trawls on the size distribution of the catch. ICES Document CM 1981/B: 5. (1981) 9.
Underwood A. J. Experiments in Ecology: Their Logical Design and Interpretation using Analysis of Variance. (1997) Cambridge, UK: Cambridge University Press. 504.
Wardle C. S. Fish behaviour and fishing gear. In: The Behaviour of Teleost Fishes—Pitcher T. J., ed. (1986) Baltimore: John Hopkins University Press. 463–495. 553.
West R. J. Comparison of fish and shrimp trawls for sampling deep-water estuarine fish in a large coastal river in eastern Australia. Fisheries Research (2002) 54:409–417.[CrossRef][Web of Science]
Wileman D. A., Ferro R. S. T., Fonteyne R., Millar R. B. Manual of methods of measuring the selectivity of towed gears. ICES Cooperative Research Report (1996) 215:126.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||