ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on September 18, 2007
ICES Journal of Marine Science: Journal du Conseil 2007 64(9):1702-1709; doi:10.1093/icesjms/fsm136
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gillnet mesh selectivity of the sandbar shark (Carcharhinus plumbeus): implications for fisheries management
1 Department of Fisheries, Government of Western Australia, WA Fisheries and Marine Research Laboratories, PO Box 20, North Beach, WA 6920, Australia
2 Centre for Ecosystem Management, School of Natural Sciences, Edith Cowan University, 100 Joondalup Drive, Joondalup, WA 6027, Australia
3 School of Earth and Environmental Sciences, James Cook University, Townsville, QLD 4811, Australia
Correspondence to R. B. McAuley: tel: +61 8 9203 0210; fax: +61 8 9203 0111; e-mail: rmcauley{at}fish.wa.gov.au
McAuley, R. B., Simpfendorfer, C. A., and Wright, I. W. 2007. Gillnet mesh selectivity of the sandbar shark (Carcharhinus plumbeus): implications for fisheries management. – ICES Journal of Marine Science, 64.Gillnet mesh selectivity parameters for the sandbar shark (Carcharhinus plumbeus) were estimated from catches taken by an experimental net of six panels of mesh, varying in size from 10.2 to 25.4 cm. The length selectivity of each mesh size was described by five different models. According to model deviance values, the four models based on the SELECT method of estimation provided better fits to the data than the gamma model previously applied to sharks. Lengths at maximum selectivity were estimated to be between 5.3 and 7.0xstretched mesh size. The breadth of the selectivity curves was greater than have been reported for most species of shark. Lognormal and normal curve forms yielded the lowest model deviance and were judged to provide the best fits to the data. Peak selectivity of the commercially utilized mesh sizes was generally estimated to be greater than the observed modal length class of the commercial C. plumbeus catch. This suggests that a relatively high abundance of smaller sharks in the study area offsets gear selectivity effects in determining the size composition of commercial catches. These results have important implications for the recovery of this overexploited stock and also for managing international gillnet fisheries for the species.
Keywords: Carcharhinus plumbeus, gillnet, mesh selectivity, SELECT, shark
Received 12 April 2006; accepted 29 July 2007; advance access publication 18 September 2007.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
Understanding the size-selectivity characteristics of fishing gear is fundamental to interpreting catch data accurately, determining the size structure of fish populations, and assessing the effects of fishing on exploited stocks (Hamley, 1975; Kirkwood and Walker, 1986; Millar and Fryer, 1999). Because of the size-selective nature of gillnets, which are widely used by fisheries that target sharks, mesh size regulations can be an effective tool for managing the size composition of catches. Several authors have suggested that shark stocks are most resilient to exploitation when fishing mortality is concentrated on a small number of age classes and that this "gauntlet" effect may be particularly important for fisheries exploiting more K-selected species (Au and Smith, 1997; Stevens et al., 1997; Walker, 1998; Simpfendorfer, 1999; Brewster-Geisz and Miller, 2000; Prince, 2005; McAuley et al., 2007a). Therefore, knowledge of how shark catches are influenced by the size selectivity of gillnets may be important for developing sustainable harvest strategies, improving economic efficiency, and, where necessary, providing management options for stock recovery. However, despite their numerous potential benefits, mesh-selectivity parameters are currently only available for a handful of shark species (Kirkwood and Walker, 1986; McLoughlin and Stevens, 1994; Simpfendorfer and Unsworth, 1998; Carlson and Cortés, 2003).
The sandbar shark, Carcharhinus plumbeus, is a carcharhinid shark of medium size with a circumglobal distribution in tropical and temperate coastal and adjacent oceanic waters (Compagno, 1984; Last and Stevens, 1994). Because of its relatively great abundance nearshore, high quality flesh, and large fins, C. plumbeus populations support significant target fisheries around the world, and are often commercially important components of non-target fisheries’ bycatch (Bonfil, 1994; Joung and Chen, 1995; Castro et al., 1999; Chan A Shing, 1999; Joung et al., 2004; McVean et al., 2006). However, C. plumbeus has among the lowest intrinsic population growth rates yet estimated for any shark species and is therefore both highly susceptible to overexploitation and very slow to recover from stock depletion (Hoff, 1990; Sminkey and Musick, 1996; Cortés, 1998, 1999; Smith et al., 1998; McAuley et al., 2007a). Therefore, fisheries that exploit sandbar sharks and biologically similar species require careful management to preclude stock depletion, the potential for associated ecological consequences, and long-term economic loss to fisheries.
In Australia, C. plumbeus is distributed on both east and west coasts (Last and Stevens, 1994), but it is apparently uncommon north of 16°S on either coast (McAuley et al., 2007b; R. Pillans, CSIRO Marine and Atmospheric Research, pers. comm.) and in South Australian waters. It is therefore considered to be represented by two distinct regional populations. Juvenile sandbar sharks are the primary catch component of a multispecies demersal gillnet fishery operating off the lower west coast of Australia (McAuley and Simpfendorfer, 2003; McAuley, 2006a; McAuley et al., 2006, 2007a, b). Minimum permitted mesh sizes in this fishery are either 16.25 or 17.50 cm, depending on the management area, but there is currently no upper mesh size specification for the fishery. Until recently, adult C. plumbeus were also targeted by a demersal longline shark fishery off northwestern Australia. However, in response to evidence that escalating longline fishing effort was causing unsustainable exploitation of the breeding stock (McAuley et al., 2007a), the northern longline fishery was excluded from the range of the population (McAuley, 2006b). As nearly all other Western Australian commercial fisheries are prohibited from landing elasmobranchs, the only significant remaining source of C. plumbeus fishing mortality in Western Australia is the previously described demersal gillnet fishery. The mean estimated annual C. plumbeus catch by that fishery was 160 t (live weight) between 1995/1996 and 2004/2005, with an approximate value of $840 000 (McAuley, 2006a; unpublished data). To ensure sustainably managed continuing C. plumbeus catches by this fishery, a better understanding of the stock’s vulnerability to gillnets is necessary. In particular, an evaluation of current and alternative gillnet harvest strategies is required to provide management options for recovering the breeding stock and for responding to the reductions in recruitment that are expected as a result of the recent period of overexploitation.
As the mesh-selectivity characteristics of C. plumbeus have not been determined previously, the primary aim of this study was to estimate the length selectivity parameters of the gillnet mesh sizes used to target the species in the Western Australian fishery. However, although the relationship between fish and mesh size can potentially take multiple forms (Hamley, 1975; Millar and Holst, 1997; Millar and Fryer, 1999; Bromaghin, 2005), previous mesh-selectivity studies involving sharks have all assumed a gamma (right-skewed) form of selection and a single method for estimating its parameters (Kirkwood and Walker, 1986). Therefore, the secondary aim of this research was to investigate the suitability of a variety of selectivity curve forms and an alternative method of estimation, which have not previously been applied to sharks.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Data collection
Comparative C. plumbeus catch data were collected from experimental gillnets of six different mesh sizes (10.2, 14.0, 15.2, 17.8, 22.4, and 25.4 cm stretched). Nets were constructed to commercial fishery specifications, using commonly available nylon monofilament webbing, hung with a coefficient of 0.66 between a float-core headline and a lead-core groundline. The number of meshes between the top and bottom of each net panel (i.e. the "mesh drop") was varied so that all panels were of approximately equal depth (Table 1). Nets were deployed from commercial gillnet vessels 83 times in waters off the lower west coast of Western Australia (Figure 1) during summer and autumn between February 2001 and June 2003. The experimental net was set demersally in conjunction with commercial nets in depths ranging from 9 to 111 m. Panels were randomly ordered, each time the net was deployed on a new vessel to reduce the potential effect of sharks’ preference for particular areas of the net. Soak times ranged from 6 to 27 h, with a mean of 19 h.
|
|
The fork length (FL) of each shark captured in the experimental net was measured as the straight line distance between the tip of the snout and the fork in the caudal fin. As male and female sharks exhibited no discernable morphological differences, catch data were not separated by sex, nor was the initial mechanism of capture specified (i.e. whether sharks had been "gilled" or "rolled" in the net). Catch data therefore included all sharks recovered onto the vessel.
Data analysis
Experimental net catches of C. plumbeus were pooled into 5-cm FL classes for each mesh size. Five families of mesh selectivity curve were fitted to the midpoint of each length class (lj), according to the methods of Kirkwood and Walker (1986) and Millar and Fryer (1999). As defined in this study, all models share the following assumptions: (i) each panel has equal fishing power; (ii) the length at maximum selectivity of each panel is proportional to mesh size; (iii) catches within each length class are independent observations from a Poisson distribution; and (iv) sampling is equal across all length classes. However, the former method is also constrained by an additional assumption (v) that variances within this family of gamma selectivity models are equal.
The Kirkwood and Walker (1986) method uses a maximum likelihood approach to fit a gamma probability distribution to the catch data for each mesh size, such that:
|
| (1) |
and ß):
|
| (2) |
Values of and ß were derived from maximum likelihood estimates of the constants 
1, which relates the mode (ß) of each selection curve to mesh size (mi), and 2, which controls the shape and spread of the selectivity curve. To satisfy assumptions (ii) and (v) above, and ß were constrained as follows:
|
| (3) |
|
| (4) |
Maximum likelihood estimates of 1 and 2 were obtained using the non-linear optimization routine in Microsoft Excel (and confirmed using Matlab fminsearch, which gave exactly the same results), from the log-likelihood function:
|
| (5) |
To estimate the levels of uncertainty in parameter estimation, the observed catches by each mesh size were randomly resampled (with replacement) to generate 500 bootstrapped datasets of catch by mesh size with the same sample numbers as the observed catches. The model was refitted to each bootstrapped dataset, using randomly generated starting values of 1 and 2. For comparison of these results with other shark studies, the mean of the values maximizing the likelihood function was calculated, and 95% confidence intervals (CIs) were derived from the 2.5 and 97.5 percentiles of these estimates.
The methods described by Millar and Fryer (1999) were used to fit four selectivity curve forms (Table 2) to the catch data using the gillnetfunctions package for R (Millar, 2003). Parameters for two forms of normally distributed (fixed spread and spread proportional to mesh size), lognormal, and gamma selection curves were estimated, by fitting the general log-linear model:
|
| (6) |
ij is the expected catch of sharks of length class j by mesh size i, and f1(mi, j) and f2(mi, j) are the selectivity functions of mi and j (which are given in the right column of Table 2). Factor (j) denotes that a length class is fitted as a factor in the model. Like the two gamma models, the normal (with proportional) and lognormal models observe geometric symmetry, whereas the second normal curve has fixed spread and is therefore not geometrically symmetric.
|
The use of maximum likelihood estimation to fit these models to the proportions of the total catch (for each length class) taken by each mesh size is known as SELECT (derived from Share Each Length Class Total). Although the Kirkwood and Walker (1986) model uses the same general approach, for ease of reference, the term SELECT is only used for the four models described by Millar and Fryer (1999).
The deviance of the fitted models from the observed data was calculated both as the sum of the squared residual values (referred to as model deviance) and the sum of the absolute residual values (referred to as absolute deviance).
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The experimental net caught 229 C. plumbeus, of which length measurements were obtained for 222 (Table 3). All nets caught a wide range of size classes, but the mean length of catches generally increased with mesh size. The exception was the catch from the 10.2 cm mesh, which had a higher mean length (86.5 cm FL) than the next largest 14.0 cm mesh (79.0 cm FL). Size frequency distributions from the five smallest mesh sizes all exhibited some degree of right skew. Although adult-sized sharks (127–136 cm FL, McAuley et al., 2006) were caught in all panels, sharks <75 cm FL were not caught by either the 22.4 or the 25.4 cm panels.
|
The modes (i.e. lengths at maximum selectivity) of the selection curves were estimated to be between 5.3 and 7.0xthe stretched mesh size of each panel (Figure 2). Point estimates of the five fitted models indicated that maximum selectivity of the 17.8 cm mesh size commonly used in the Western Australian C. plumbeus fishery was attained at between 95 and 124 cm FL. Although the commercially important mesh size (16.5 cm) was not tested empirically, lengths at maximum selectivity for this mesh size were estimated to be between 88 and 115 cm FL by substituting the model-fitted parameter estimates into each respective selection curve equation.
|
Residual plots do not indicate any obvious biases or lack of fit among mesh-specific selectivity models, apart perhaps from a slight under-representation of the largest size classes in the smallest mesh, and over-representation of 90–100 cm size classes in the largest mesh (Figure 3). Although absolute deviance values were similar for all five models, model deviance values indicated that better fits were obtained from the four SELECT models than from the Kirkwood and Walker (1986) gamma model (Table 4). Of the models fitted using the SELECT method, marginally better fits were obtained from the lognormal and normal (fixed spread) models than were obtained from the gamma and, particularly, the normal (proportional spread) forms.
|
|
According to the results obtained from the SELECT models, the modes of the commercially important 16.5 and 17.8 cm selection curves (96.1–115.3 and 103.7–124.3 cm FL, respectively) were larger than the modal size class (80–85 cm) of C. plumbeus catches by the Western Australian commercial gillnet fishery (McAuley et al. 2007b; Figure 4). This difference was more pronounced for the normal (proportional spread) and gamma curves than for the lognormal and normal (fixed spread). The mode of the observed commercial catch was almost precisely the same as that estimated by the Kirkwood and Walker (1986) model for the 16.5 cm mesh size. There was a secondary mode in the observed commercial catch (at 135–140 cm FL) that was not described by any of the unimodal selectivity curve forms tested in this study.
|
The relative contact rates estimated from the mesh selectivity parameters closely matched the observed size frequency of commercial C. plumbeus catches for length classes >95 cm FL (Figure 4). However, commercial catches of smaller length classes, particularly those in the 70–95 cm FL range, were noticeably larger than the 16.5 and 17.8 cm selectivity curves and derived contact rates suggested they should be.
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Although the methods described by Kirkwood and Walker (1986) have become a standard approach for investigating the gillnet mesh selectivity characteristics of shark stocks, results from this study suggest that more suitable models may be available for this purpose. Visual inspection of the residual plots and comparison of the absolute deviance values suggest that the Kirkwood and Walker (1986) model provided an adequate (arguably superior) fit to the observed catch data than the four SELECT models described by Millar and Fryer (1999). However, not only do the model deviance values indicate better overall fits from the four SELECT models but also that, of these, lognormal and normal (with fixed spread) curves provided better descriptions of the data than either of the gamma models tested. Therefore, one of the key assumptions of the Kirkwood and Walker (1986) method, that mesh selectivity is best represented by a gamma probability distribution, may not always be valid for sharks. Although the assumption of right-skewed selectivity does seem intuitively reasonable, the suitability of other right-skewed curve forms for modelling mesh selectivity of sharks has not previously been assessed. Kirkwood and Walker (1986) did, however, compare their gamma selectivity model with an alternative (normal) curve form, but found that this resulted in a much worse fit than achieved by the gamma model.
There also appears limited justification for the Kirkwood and Walker (1986) model assumption of constant variance across mesh sizes, despite its general acceptance in the shark literature. Not only is the assumption difficult to assess directly, as acknowledged by the original authors, but it also imposes a possibly unnecessary constraint on the model’s ability to fit empirical data. Another issue relating to the flexibility of the Kirkwood and Walker (1986) method was identified, and this could also have compromised the accuracy of results obtained from this model. Each family of net selectivity curves in the Kirkwood and Walker (1986) scheme is defined and controlled by two parameters: 1, which is a constant of proportionality determining the mode of the selectivity curve for a given mesh size, and 2, which simultaneously controls the shape (skewness) of the curves and the growth rate of the spread of the curves relative to mode. For a family of curves with a narrow range of mesh sizes (proportionally speaking), there is little conflict between the competing incompatible controls described by 2. Under those circumstances, the Kirkwood and Walker (1986) procedure will produce a family of curves which is very accurate, even for the smallest and largest mesh sizes. However, for a family with a broad range of mesh sizes (proportionally speaking), the selectivity curves at the extremities may reflect the conflict between spread growth and shape by not following the observed data very closely.
Notwithstanding the issues discussed earlier, as the analytical methods and experimental gear used in previous shark mesh selectivity studies were broadly similar (although McLoughlin and Stevens, 1994, used four mesh sizes), the parameters estimated for C. plumbeus by the Kirkwood and Walker (1986) model may be compared with those for other shark species in relative terms. Estimates of 1 were similar to those previously reported for other Carcharhinus species (Table 5). This observation is consistent with those of Simpfendorfer and Unsworth (1998) and Carlson and Cortés (2003), who suggested that this parameter is likely to reflect morphological similarities between congeneric species. As C. plumbeus has a moderately broad and blunt snout compared with the other Carcharhinus species, it was unsurprising that 1 estimates were somewhat lower (i.e. smaller length at maximum selectivity) than has been reported for species with more slender and pointed snouts (C. acronotus, C. isodon, C. tilstoni, Furgaleus macki, Mustelus antarcticus, and Rhizoprionodon terranovae). However, the larger 1 values reported for C. obscurus and C. sorrah do not fit this pattern readily, because those species have similarly blunt snouts.
|
The family of selection curves derived from the Kirkwood and Walker (1986) method had narrower selectivity profiles than those obtained from the four SELECT models. However, all families of curve were considerably broader than reported for most other sharks, and this is reflected in the relatively high estimates of
2 (Table 5). Previously reported 2 estimates for sharks were generally 3–4 times smaller than those obtained for C. plumbeus in this study. However, the estimate for C. tilstoni (McLoughlin and Stevens, 1994) was ten times smaller, and those for R. terranovae and C. isodon (Carlson and Cortés, 2003) were slightly larger. Although variability in the data caused by the relatively small sample size could potentially have contributed to the breadth of the selection curves estimated for C. plumbeus, this seems unlikely because the 95% CIs (and standard errors) of 2 parameter estimates are reasonably discrete. Instead, it appears that C. plumbeus has genuinely broader mesh selectivity characteristics than other shark species for which mesh selectivity has been studied. Although the reasons for this are unclear, it may possibly be a consequence of this species’ proportionally much larger pectoral and first dorsal fins, which may increase the retention rates of smaller length classes. Therefore, although the selectivity of mesh sizes being used in the Western Australian commercial gillnet fishery will certainly affect the size composition of C. plumbeus catches, their effects may not be as strong as for species such as C. obscurus and F. macki, which are also important components of the catch by the fishery. However, given the previously discussed reservations about how mesh selectivity was determined for the last two species, it would be beneficial to re-examine their mesh selectivity characteristics using alternative techniques to confirm this conclusion.
Although the observed size frequency distribution of the commercial C. plumbeus catch could have been explained by fitting a bimodal selectivity curve (such as the binormal selectivity model described by Millar and Fryer, 1999) to the experimental net data, this was not considered appropriate because the commercial catch is thought to reflect aspects of both stock structure and mesh selectivity. McAuley et al. (2007b) reported that juvenile sharks <
100 cm FL are most commonly distributed throughout the study area and that larger juveniles and adults are found to the north. The primary modal size class of the observed commercial catch (85–90 cm FL) was smaller than the estimated lengths at maximum selectivity of all the commercially important 16.5 and 17.8 cm mesh size selectivity curves (apart from the 16.5 cm curve estimated from the Kirkwood and Walker, 1986, method). The estimated relative contact rates (which might be considered as a proxy for the size frequency of the fished population) also suggest that the most commonly encountered size classes were between 70 and 95 cm FL. Therefore, it can be inferred that a relatively high abundance of sharks <95 cm FL within the study area offset the described mesh selectivity effects (to varying degrees for each model) in determining the mode of the commercial fishery’s C. plumbeus catch. The mesh sizes used in this fishery therefore appear to be simultaneously limiting catches of smaller length classes and selecting for larger juveniles (i.e. between 96 and 124 cm FL).
The smaller secondary mode in the commercial catch (at between 130 and 150 cm FL) can be explained by the incidental capture of adult sharks during their annual southerly migration into the study area (McAuley et al., 2007b). On the basis of personal observations, most of these larger sharks were likely to have been entangled by "rolling" themselves in the nets or by breaking meshes and becoming gilled in the bigger openings. Therefore, catches of these size classes are thought to be largely independent of mesh size. As such, their inclusion in the comparative catch data used for this study may have biased the model fits and contributed to the relative breadth of the estimated selectivity curves. However, because determining the initial means of capture can be highly subjective, particularly under the time constraints of the commercial fishing environments in which these data were obtained, there was considered to be even greater potential for bias in trying to exclude these data.
Demographic analysis of the Western Australian C. plumbeus stock (McAuley et al., 2007a) and other studies (notably, Stevens et al., 1997; Walker, 1998; Simpfendorfer, 1999; Prince, 2005) have highlighted potential sustainability benefits in restricting shark fishing mortality to a limited number of age classes. Simpfendorfer (1999), Prince (2005), and McAuley et al. (in press) further pointed out that, for stocks of more K-selected shark species such as C. plumbeus, there are additional benefits in restricting fishing mortality to the youngest juvenile age classes. However, in addition to the gillnet fishery targeting juvenile C. plumbeus off the lower west coast of Western Australia and small levels of bycatch in non-target fisheries, the development of a longline fishery targeting adults off the north western coast during the late 1990s, resulted in fishing mortality across a wide range of age classes (McAuley et al., 2007a). The intent of excluding the northern longline shark fishery from a large portion of northwest Western Australia and prohibiting all non-target fisheries from landing shark bycatch was therefore to restrict C. plumbeus catches solely to the temperate gillnet fishery, where they can be more carefully managed.
Although mesh selectivity effects may be less pronounced for C. plumbeus than for other species of shark, results from the current study have obvious implications for managing the recovery of this stock and developing more sustainable future harvest strategies. First, restricting the fishery’s minimum mesh size to 16.5 cm (or smaller) would be expected to reduce catches of 95+ cm size classes, which may have benefits for stock recovery and future sustainability. Second, restricting the maximum permitted mesh size would reduce the fishery’s ability to target larger size classes, which might occur in response to the currently declining catch rate of smaller juveniles (McAuley, 2006a) and/or the relatively high value of fins from larger sharks. However, such changes may have little effect on reducing catches of migratory adults (>130 cm FL), because catches of these larger sharks are likely to be largely independent of mesh size. Although the mesh selectivity parameters estimated in this study primarily provide the information needed to quantify the outcomes of such management changes and for improving assessments of this stock, it is also hoped that the application and review of alternative gillnet mesh selectivity assessment techniques may be of benefit to shark fishery scientists and managers internationally.
|
Acknowledgements |
|---|
We acknowledge the Fisheries Research and Development Corporation who provided funding for this and related research, and Gordon Lymn of the WA Department of Fisheries who designed, constructed, and maintained the experimental gillnets (and who has provided enormous assistance to the shark research programme for many years). We also wish to thank the skippers and crews of the FVs "Tee Nee Dee" and "San Margo", from which the experimental net was deployed: Carlo and Jason Gullotti, and Brad Warnock, Jamie Thornton, John Smythe, and Darren Sappelli. Without their assistance, this study would not have been possible. We also thank Ryan Ashworth, Justin Chidlow, and Ben Sale of the Western Australian Department of Fisheries’ Shark Research Section for their assistance in the collection of data. Finally we thank Rod Lenanton, Steve Newman, Dan Gaughan, and Peter Stephenson from the Western Australian Fisheries and Marine Research Laboratory, Glenn Hyndes at Edith Cowan University, and the anonymous reviewers for their valued advice during preparation of this manuscript.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Au D. W., Smith S. E. A demographic method with population density compensation for estimating productivity and yield per recruit of the leopard shark (Triakis semifasciata). Canadian Journal of Fisheries and Aquatic Sciences (1997) 54:415–420.
Bonfil R. Overview of world elasmobranch fisheries. FAO Fisheries Technical Paper 341. (1994) 119.
Brewster-Geisz K. K., Miller T. J. Management of the sandbar shark, Carcharhinus plumbeus: implications of a stage-based model. Fishery Bulletin US (2000) 98:236–249.
Bromaghin J. F. A versatile net selectivity model, with application to Pacific salmon and freshwater species of the Yukon River, Alaska. Fisheries Research (2005) 74:157–168.[CrossRef][Web of Science]
Carlson J. K., Cortés E. Gillnet selectivity of small coastal sharks off the southeastern United States. Fisheries Research (2003) 60:405–414.[CrossRef][Web of Science]
Castro J. I., Woodley C. M., Brudek R. L. A preliminary evaluation of the status of shark species. FAO Fisheries Technical Paper (1999) 380:72.
Chan A Shing C. Shark fisheries in the Caribbean: the status of their management including issues of concern in Trinidad and Tobago, Guyana and Dominica. In: Case Studies of the Management of Elasmobranch Fisheries—Shotton R., ed. (1999) 904–920. FAO Fisheries Technical Paper, 378.
Compagno L. J. V. FAO Species Catalogue. 4. Sharks of the world. An annotated and illustrated catalogue of sharks known to date. FAO Fisheries Synopsis (1984) 125.
Cortés E. Demographic analysis as an aid in shark stock assessment and management. Fisheries Research (1998) 39:199–208.[CrossRef][Web of Science]
Cortés E. A stochastic stage-based population model of the sandbar shark in the western North Atlantic. In: Life in the Slow Lane. Ecology and Conservation of Long-Lived Marine Animals—Musick J. A., ed. (1999) 115–136. American Fisheries Society Symposium, 23.
Hamley J. M. Review of gillnet selectivity. Journal of the Fisheries Research Board of Canada (1975) 32:1943–1969.[Web of Science]
Hoff T. B. Conservation and management of the western North Atlantic shark resource based on the life history strategy limitations of sandbar sharks. (1990) Newark: University of Delaware. PhD dissertation.
Joung S. J., Chen C. T. Reproduction in the sandbar shark, Carcharhinus plumbeus, in the waters off northeastern Taiwan. Copeia (1995) 3:659–665.
Joung S. J., Liao Y. Y., Chen C. T. Age and growth of sandbar shark, Carcharhinus plumbeus, in northeastern Taiwan waters. Fisheries Research (2004) 70:83–96.[CrossRef][Web of Science]
Kirkwood G. P., Walker T. I. Gill net selectivities for gummy shark, Mustelus antarcticus Günther, taken in south-eastern Australian waters. Australian Journal of Marine and Freshwater Research (1986) 37:689–697.[CrossRef]
Last P. R., Stevens J. D. Sharks and Rays of Australia. (1994) Melbourne, Australia: CSIRO. 513.
McAuley R. Demersal gillnet and demersal longline fisheries status report. In: State of the Fisheries Report 2005/06—Fletcher W. J., Head F., eds. (2006) a. Perth, Western Australia: Department of Fisheries. 212–220.
McAuley R. Northern shark fisheries status report. In: State of the Fisheries Report 2005/06—Fletcher W. J., Head F., eds. (2006) b. Perth, Western Australia: Department of Fisheries. 165–169.
McAuley R., Simpfendorfer C. Catch composition of the Western Australian temperate demersal gillnet and demersal longline fisheries, 1994–1999. In: Fisheries Research Report, 146. (2003) Perth, Western Australia: Department of Fisheries. 78.
McAuley R. B., Simpfendorfer C. A., Hall N. G. A method for evaluating the impacts of fishing mortality and stochastic influences on the demography of two long-lived shark stocks. ICES Journal of Marine Science, (2007) 64(a):1710–1722.
McAuley R. B., Simpfendorfer C. A., Hyndes G. A., Allison R. R., Chidlow J. A., Newman S. J., Lenanton R. C. J. Validated age and growth of the sandbar shark. In: Carcharhinus plumbeus (2006) 77. Environmental Biology of Fishes. 385–400. (Nardo, 1827) in the waters off Western Australia.
McAuley R. B., Simpfendorfer C. A., Hyndes G. A., Lenanton R. C. J. Distribution and reproductive biology of the sandbar shark. In: Carcharhinus plumbeus (2007) 58, b. Marine and Freshwater Research. 116–126. (Nardo, 1827) in Western Australian Waters.
McLoughlin K. J., Stevens J. D. Gill-net mesh selectivities for two species of commercial carcharhinid sharks taken in northern Australia. Australian Journal of Marine and Freshwater Research (1994) 45:521–534.[CrossRef]
McVean A. R., Walker R. C. J., Fanning E. The traditional shark fisheries of southwest Madagascar: a study in the Toliara region. Fisheries Research (2006) 82:280–289.[CrossRef][Web of Science]
Millar R. R CODE for fitting SELECT models to gillnet data. (2003) http://www.stat.auckland.ac.nz/~millar/selectware/R/gillnets/gillnetfunctions.R.
Millar R. B. Untangling the confusion surrounding the estimation of gillnet selectivity. Canadian Journal of Fisheries and Aquatic Sciences (2000) 57:507–511.
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]
Millar R. B., Holst R. Estimation of gillnet and hook selectivity using log-linear models. ICES Journal of Marine Science (1997) 54:471–477.
Prince J. D. Gauntlet fisheries for elasmobranchs – the secret of sustainable shark fisheries. Journal of Northwest Atlantic Fishery Science (2005) 35:407–416.
Simpfendorfer C. A. Demographic analysis of the dusky shark fishery in southwestern Australia. American Fisheries Symposium (1999) 23:149–160.
Simpfendorfer C. A., Unsworth P. Gill-net mesh of dusky sharks (Carcharhinus obscurus) and whiskery sharks (Furgaleus macki) from south-western Australia. Australian Journal of Marine and Freshwater Research (1998) 49:713–718.[CrossRef][Web of Science]
Sminkey T., Musick J. Demographic analysis of the sandbar shark, Carcharhinus plumbeus, in the western North Atlantic. Fishery Bulletin US (1996) 94:341–347.
Smith S. E., Au D. W., Show C. Intrinsic rebound potentials of 26 species of Pacific sharks. Marine and Freshwater Research (1998) 49:663–678.[CrossRef][Web of Science]
Stevens J. D., Walker T. I., Simpfendorfer C. A. Are southern Australian shark fisheries sustainable? In: Developing and Sustaining World Fisheries Resources. The State of Science and Management. Proceedings of the 2nd World Fisheries Congress—Hancock D. A., Smith D. C., Grant A., Beumer J. P., eds. (1997) Melbourne: CSIRO. 62–66.
Walker T. I. Can shark resources be harvested sustainably? A question revisited with a review of shark fisheries. Marine and Freshwater Research (1998) 49:553–572.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||