ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on December 4, 2008
ICES Journal of Marine Science: Journal du Conseil 2009 66(1):72-81; doi:10.1093/icesjms/fsn187
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article appears in the following ICES Journal of Marine Science issue: European Symposium on Marine Protected Areas as a Tool for Fisheries Management and Ecosystem Conservation [View the issue table of contents]
Investigating the consequences of Marine Protected Areas for the South African deep-water hake (Merluccius paradoxus) resource
Marine Resource Assessment and Management Group, Department of Mathematics and Applied Mathematics, University of Cape Town, Rondebosch 7701, South Africa
Correspondence to C. T. T. Edwards: tel: +27 21 650 2336; fax: +27 21 686 0477; e-mail: charles.edwards{at}uct.ac.za.
Edwards, C. T. T., Rademeyer, R. A., Butterworth, D. S., and Plagányi, É. E. 2009. Investigating the consequences of Marine Protected Areas for the South African deep-water hake (Merluccius paradoxus) resource. – ICES Journal of Marine Science, 66: 72–81.Protected areas are often touted as important management tools to mitigate the uncertainty inherent in marine ecosystems, and thereby improve the long-term prospects for sustainable resource use. However, although they certainly play an important role in conservation, their usefulness in improving fishery yields is contentious. We present a simulation model that explores spatial closure options, and apply it to the demersal hake trawl fishery off South Africa. The model is based on the age-structured approach used for current assessments, representing the dynamics of the deep-water hake Merluccius paradoxus within a zonally disaggregated spatial system. Fitting the model to two zones, which demarcate a potential closed area from the remaining fished area, we investigate the consequences that such a protected area could have for the fishery. Our model suggests that area closures would have a negligible benefit for the fishery, regardless of the level of hake movement between areas. This is likely the result of the model's simplicity, and we suggest additional factors that should be considered to quantify the impact of Marine Protected Areas on the fishery more reliably.
Keywords: hake, Marine Protected Areas, South Africa, spatial model, trawl
Received 7 November 2007; accepted 30 May 2008; advance access publication 4 December 2008.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
There is widespread consensus that Marine Protected Areas (MPAs) provide an effective means of supporting the biodiversity within their boundaries (Halpern and Warner, 2002; Halpern, 2003; Willis et al., 2003), and in South Africa, recent legislative initiatives have promoted the implementation of MPAs as a means of achieving conservation targets (Lombard et al., 2004). The implications for local fisheries, however, are unclear, reflecting current controversy over the benefits of MPAs for fishery management (Hilborn et al., 2004; Walters et al., 2007).
Within the context of fishery management, MPAs are intended to insure against unintentional overexploitation (Roberts and Polunin, 1993; Clark, 1996; Botsford et al., 1997; Lauck et al., 1998), which may result from the inherent uncertainty involved in predicting fishery dynamics and designing (and implementing) effective management strategies. Proponents of MPAs often suggest that they have a long-term beneficial effect on fishery productivity (Auster and Shackell, 2000; Gell and Roberts, 2003), primarily as a result of improved habitat quality. MPAs may also provide protection for spawning aggregations or allow the age structure of resident populations to recover. If fecundity per unit mass increases with age (Birkeland and Dayton, 2005), then a shift towards older fish in the protected population may enhance recruitment to the fishery (Berkeley et al., 2004). However, these arguments depend on a restricted range of assumptions (Hilborn, et al., 2004). Primarily, adult fish populations must be viscous enough to benefit from the protection that the MPA provides. If they are too mobile, then the anticipated benefits of an MPA are negated (Polacheck, 1990; Walters et al., 2007).
Although in some circumstances, proposed fishery benefits may be considered secondary to the primary motivations of ecosystem conservation, it is important to recognize that MPAs may have negative consequences for both the fishery and the conservation-orientated objectives behind their implementation. Spatial exclusion will likely lead to inefficient fishing practices, as fishing is forced onto less-productive grounds. This may result in increased levels of overall effort and the targeting of previously untrawled areas (ICES, 2001; Dinmore et al., 2003). As a result, it may even impede the recovery of exploitable biomass in the areas open to fishing (Hobday et al., 2005). There may also be negative economic consequences, because an MPA can reduce the fishery's sustainable value (Walters and Bonfil, 1999; Holland, 2000), exacerbate annual fluctuations in recruitment (Larson and Julian, 1999), and increase the level of uncertainty surrounding assessments of the resource (Field et al., 2006).
If MPAs are to be advanced based on their proposed fishery benefits, then their implications for the fishery must be quantitatively assessed (Walters and Bonfil, 1999; Holland, 2000; Walters et al., 2007). Here, we investigate the consequences of area closures for the hake fishery, South Africa's economically most important marine resource. The fishery consists of trawl (both deep-sea and inshore), longline and handline fleets, and targets two species, the shallow-water Merluccius capensis and deep-water hake Merluccius paradoxus. Hake was originally a bycatch of the Agulhas sole (Austroglossus pectoralis) fishery, but landings increased steadily from the early 20th century, reaching a peak of nearly 300 000 t year–1 in the early 1970s. Catches subsequently declined to 120 000–150 000 t after the establishment of South Africa's Exclusive Economic Zone (EEZ) in 1977, which led to the immediate exclusion of most foreign vessels, and have remained relatively stable since.
A spatial, age-structured population model of the hake resource, based on the current stock assessment methodology (Rademeyer, 2004; Butterworth and Rademeyer, 2005), was fitted to available data. This allowed the investigation of potential consequences of an area closure to fisheries, within a realistic framework. The focus was on the trawl fleet only, which is responsible for 
90% of the commercial catch, and M. paradoxus, which is this fleet's current primary target (Rademeyer and Butterworth, 2006).
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
Data
Commercial catch and effort data for the hake fishery are reported for 20' x 20' rectangles (minutes of latitude and longitude), which form the grid of the South African EEZ. Commercial data for the trawl fleet (consisting of deep-sea and inshore fleets plus a few foreign vessels) were available from 1978 to 2005 and were provided by Marine and Coastal Management (MCM), South Africa. This was used to calculate the nominal catch per unit effort (cpue) over years for each grid. In addition, species-specific length frequency data were available from observers placed on commercial vessels by Capricorn Fisheries Monitoring, South Africa, since 2002. Quarterly survey data assigned to commercial grids were also available from MCM from 1983 onwards, providing both a swept-area-based abundance index and length frequency data.
The data represented a total of 421 grids. Commercial and survey catch-rate data were available from 358 and 254 grids, and observer and survey length frequency data from 149 and 178 grids, each respectively.
Species-specific age–length keys and mass-at-age keys were obtained directly from MCM. Age–length keys for M. paradoxus were available from 1988 to 2000 (excluding 1989 and 1998) and were applied to the survey length frequency data from this period. No age–length keys were available for the years covered by the commercial catch-at-age data (2002–2005). If an age–length key was not available for a particular year, then an age–length relationship was used to predict catches-at-length (see Appendix), to which the catch-at-length data from that year were fitted directly.
Commercial species split
Because the two hake species are morphologically very similar, only species-combined commercial catch data are collected. However, species-specific commercial observer length frequency data were available. Observers sampled individual fish at random for species and length (D. Japp, pers. comm.). This information (nlspp) was therefore used to approximate the species split of the commercial catch for each grid.
The weight (W) of each length category (l) for each species was calculated using the published length (in centimetres) to weight (in grammes) relationship with 
paradoxus = 0.00615, βparadoxus = 3.046, capensis = 0.00505, and βcapensis = 3.113 (Punt and Leslie, 1991). Summing over all 140 length categories gave the proportionate catch per species Cspp for each grid
|
| (1) |
|
| (2) |
Observer data were not available for all the fished grids. To calculate the species split of catches in these grids, it was therefore necessary to use the relationship between species split and depth apparent from surveys (Gaylard and Bergh, 2004). Grids were divided into 14 50-m depth categories, and the observer data summed over all grids in each category. These summed data were then allocated to grids in the same depth category for which no observer data were available, to calculate the species split as described above.
Historical catches
Spatially disaggregated commercial catch data were available for relatively recent years only, whereas fishing records of total catch for each fleet extend back to 1917 (Rademeyer, 2004). It was therefore necessary to assign these past catches spatially across grids in a manner that fits with assumed historical changes in fishing distribution.
For changes in patterns of trawl fishing, we assumed that the early fishery was concentrated on abundant fish stocks on the Agulhas Bank, but then progressively moved offshore as the shallow-water resource was reduced and technology improved to allow fishing in deeper waters. This would have resulted in a change in the species composition of the catch, because spatial distributions of the two hake species can be differentiated by depth, with M. paradoxus occupying deeper water than M. capensis. Currently, most of the catch is taken from 300–400 m.
The spatial allocation of past catches (before 1978) therefore proceeded as follows.
- The proportion of the total yearly catch was calculated for each grid and averaged over the period 1978–1988.
- The annual allocation coefficient ry was defined for each past year.
- Catch proportions for each past year were assumed to be the same per grid as the 1978–1988 average, but scaled up or down depending on the depth. For grids of 250 m or deeper, the proportionate catch was scaled down (multiplied by 1 – ry), and for grids shallower than 250 m, the proportionate catch was scaled up by an equal amount (multiplied by 1 + ry).
- Proportionate yearly catch was normalized across grids to sum to 1.
- Proportionate catch per grid for each past year was finally used to assign historical catches spatially.
|
The model
To investigate resource dynamics under different spatial management scenarios, we developed a model based on the age-structured population model currently used for the stock assessment, modified to represent spatial variations in abundance. The equations used are based on those described by Rademeyer (2004). The model considers a single species (M. paradoxus) and a single commercial trawl fleet. We assume single instantaneous spawning and fishing events at the beginning and middle of the year, respectively. Similarly, movement between zones is assumed to occur as an instantaneous event at the beginning of the year, before spawning.
The model allows the resource to be divided into a number of spatially defined zones, with specified age-dependent movements between them. Each zone z has its own carrying capacity Kzsp and selectivity parameters. Natural mortality and steepness of the Beverton–Holt stock–recruitment function were considered to be the same across zones. The model is completely described in the Appendix. It was fitted to data from 1917 to 2005 within a maximum likelihood framework, using AD Model Builder (Otter Research Ltd), ensuring convergence in each case.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
Overall dynamics
To obtain an overall impression of biomass dynamics, the model was first fitted to a single zone consisting of the entire fished area within the South African EEZ. Fits to input data were good (data not shown). It was apparent from the results (Figure 2) that the precise assumption regarding the historical species split of catches (Figure 1) had little impact on dynamics. Owing to the lack of sensitivity shown, only Scenario 3 was implemented for all subsequent analyses. Parameter estimates for Scenario 3 are given in Table 1. Population trends predicted by the model differed somewhat from those from the standard stock assessment model (Rademeyer and Butterworth, 2006) that underlies scientific advice on resource management, being notably more optimistic. An exact match was not expected, however, given a number of differences in model structure and the data fitted. In particular, the nominal commercial cpue used here is increasing, whereas the cpue series used in the stock assessment, which is GLM-standardized for factors such as vessel, is declining (Glazer and Butterworth, 2002; Glazer, 2007). This may be partly responsible for the higher estimated Ksp and hence better current population status. Future analyses will seek to derive and use a standardized cpue series.
|
|
Consequences of a single closed area
A single zone (Zone 1) was selected as a proposed MPA, consisting of six contiguous commercial grids that have been identified by industry consultants as candidates for protection owing to the heavy trawling intensities in that region (D. Japp, pers. comm.). Zone 1 covers an area of
40 x 60 nautical miles. The remaining fished area was considered as a second zone (Zone 2). Biomass projections extended to 2020, illustrating the consequences of closing the proposed MPA to fishing. The annual catch taken from the fishery during the period of projection was set equal to 100 000 t (approximating the commercial M. paradoxus catch for 2005). Under closure, the catch tonnage that would have been taken from within the MPA was assigned to Zone 2. It is implicitly assumed that any redistribution of fishing effort does not lead to changes in the total catch of M. paradoxus (i.e. that fishing effort excluded from the MPA continues to be directed towards fishing grounds with a similar species split of catches and is adjusted to a level that maintains the total catch).
The model was first fitted to data from each zone, assuming a low movement rate, specifying that 10% of individuals within Zone 1 move to Zone 2 each year. Movement proportions were assumed to be the same for all ages. Biomass dynamics under this assumption are shown in Figure 3, with parameter estimates in Table 2. To illustrate the consequences of higher levels of movement, a rate of 80% was then specified and the model refitted (Figure 4; Table 3). Assuming that all fish in Zone 1 start the year at its centre and move according to a diffusion process, rates of 10% and 80% correspond to average annual movements of 23 and 118 nautical miles, respectively. Note that the movement parameter 
a was specified for Zone 1 only, with the values for Zone 2 obtained by use of Equation (A6; see Appendix).
|
|
|
|
When the level of movement between the zones was low, there was a clear benefit of closure for the population within the MPA (Figure 3), and a corresponding lesser rate of population growth in the remaining area. At higher rates of movement, the predicted benefits to the protected population decreased (Figure 4), illustrating that the outcome of MPA implementation depends on mobility of the population. This is consistent with the results from similar analyses (Gerber et al., 2003).
If MPAs are to be used as a fishery management tool, then it is important to consider the dynamics of the whole system. We therefore summed the biomasses in both zones for each movement rate. The consequences of area closure for the overall abundance of the hake resource are negligible (Figure 5). In other words, the benefits to the protected population are almost exactly cancelled by the increased fishing mortality in the surrounding zone. Therefore, the spatial displacement of catch to outside the MPA has no real consequence, irrespective of the rate of movement assumed.
|
|
Discussion and conclusions |
|---|
|
TopIntroductionMethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
In response to an increasing number of calls for the establishment of MPAs worldwide, there has been a recent proliferation of models that have attempted to quantify the frequently argued fishery benefits (Gerber et al., 2003; Pelletier and Mahévas, 2005). However, the possibility of drawing general conclusions is often limited by model sensitivity to assumptions regarding the specific biological and socio-economic systems on which they focus (Holland, 2002). There is, therefore, a need to develop spatial models that can be fitted to available data (Pelletier and Mahévas, 2005) and assessment models that are robust to changes in the spatial distribution of fishing effort (Field et al., 2006). This investigation was intended to provide a preliminary insight into the consequences of MPA implementation for one of South Africa's most important marine resources, using a model based on the current stock assessment methodology and fitted to data. It represents a first step towards understanding whether or not MPAs are likely to be a useful fishery management tool in this context.
The model results presented here suggest that an area closure of the type examined is likely to yield no net benefit to the hake fishery. Even if an MPA is beneficial to the population enclosed, its implementation is unlikely to lead to enhancement of the resource as a whole. However, before this point is overemphasized, it is important to consider the limitations of the model applied here. Because natural mortality and recruitment parameters are shared by Zones 1 and 2, and because selectivity curves in the two zones are similar (data not shown), the sole driver of comparative dynamics is catch. Reallocating the catch from one zone to the other will result in compensatory changes in dynamics of a nearly equivalent magnitude in each zone (Figure 5). Net benefits stemming from the implementation of an MPA are only likely to accrue if the exclusion of fishing from a particular area results in a benefit to the protected population that is larger than that expected from a simple reduction in catch, and additional consequences for the unprotected population are relatively small. For example, benthic habitat recovery within an MPA may lead to lower juvenile mortality rates (Tupper and Boutilier, 1995). If increased or spatially redistributed effort external to the MPA does not lead to an equivalent level of habitat damage, and assuming appropriate levels of movement or larval dispersal, then the MPA is likely to enhance growth of the resource, at least to some extent. Identifying when such conditions are met is crucial to the design of MPAs, if they are to be used successfully in fishery management.
The detrimental impacts of bottom trawling on benthic communities are well known (Jennings and Kaiser, 1998; Thrush and Dayton, 2002; Rodwell et al., 2003), and the associated loss of habitat has been proposed as probably contributing to dwindling catch rates for fish (Botsford et al., 1997). However, a link between the ecological state of the benthos and productivity of resident fish populations has yet to be established in the literature (Ocean Studies Board, 2002). In an effort to curb the perceived ecological costs of trawling, the governments of New Zealand and the USA recently established areas closed to this type of fishing. The restoration of benthic communities is a primary motive for the establishment of MPAs in South Africa (Lombard et al., 2004). However, work in the North Sea illustrates that it cannot be assumed that area closures will necessarily lead to improvements in the wider ecosystem (Dinmore et al., 2003; Hiddink et al., 2006). Therefore, it is important that, if indeed there is a link between the state of the benthos and population productivity, the potential fishery consequences are examined, although naturally the confirmation of a link may be difficult to establish and may require experimental management (e.g. Sainsbury et al., 1997). The model presented here will be developed to examine this question within a simulation framework.
It may also be necessary to evaluate the economic consequences of an MPA. It has been assumed here that on closure of a fished area, total catch of the trawl fleet is unchanged. However, because the closed area considered is a region with high catch rates, it is likely that exclusion will require an increase in overall effort to catch the same amount from the remaining fished area. This will have economic consequences that will require evaluation if the fishery consequences of a closed area are to be fully understood. The importance of such an economic component has been highlighted in recent local discussions concerning implementation of an MPA at Prince Edward Island, where economic viability of the toothfish fishery is already marginal.
Our results illustrate that, for the model considered, although protected populations are likely to recover if movement rates are low, the net benefit of MPA implementation to the resource as a whole, and hence the fishery, is negligible. This highlights the importance of considering additional components of the system, such as the habitat damage of trawling together with any consequent negative impact on a resource, possible small-scale stock structure, and age-dependent fecundity, if the consequences of area closure are to be fully evaluated before establishment of an MPA.
|
Appendix: Population model for M. paradoxus and associated parameter estimation |
|---|
|
TopIntroductionMethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
Estimated model parameters
Parameters estimated during the model fit are listed in Table A1.
|
Natural mortality-at-age a is given by
|
| (A1) |
M = 6(M2 – M5) and 
M = M2 – M/3.
Selectivity-at-age for both the trawl fleet and survey vessels is represented by
|
| (A2) |
Different survey selectivity curves are assumed, depending on the location of the survey, specified as either off the west coast of South Africa (conducted in summer and winter) or the south coast (conducted in spring and autumn).
Population dynamics
Numbers-at-age
Numbers-at-age for a particular zone and year (Nzya) are modelled as
|
| (A3) |
Movement
Movement represents the number of individuals in a given zone that move to an adjoining zone in 1 year. Because each zone has its own separate pristine biomass, migration from zone z is specified as a proportion of the numbers present at Kzsp and scaled so that, when the overall population is at equilibrium, the numbers entering and leaving a particular zone are equal.
The migration matrix for age a and for Z zones is specified as follows:
|
| (A4) |
ija is the proportion of individuals that move from z = i to z = j, and N0 the equilibrium numbers-at-age. So that migration into and out of zone z sums to 0 at Kzsp, 
a must satisfy the condition
|
| (A5) |
We simplify by assuming paired migration rates, so that
|
| (A6) |
|
| (A7) |
migration parameters, representing the upper or lower diagonal of the migration matrix (excluding the diagonal itself). By ensuring that Equation (A6) is satisfied, a is completely described. |
| (A8) |
|
| (A9) |
a) were specified for 1 
a 5 as model input. Values for lower and higher ages were set equal to values for a = 1 and a = 5, respectively.
Pristine conditions
The population in 1917 (the first year for which historical catch data are available) is assumed to be at equilibrium. To calculate numbers-at-age in 1917, we first estimate the proportionate age structure at equilibrium relative to age 0 (P0):
|
| (A10) |
It is then necessary to estimate the equilibrium numbers-at-age (N0za) by scaling P0za using the estimated pristine recruitment (R0). This is done through the following steps:
|
| (A11) |
|
| (A12) |
|
| (A13) |
Recruitment
Recruitment to age 0 is calculated using the Beverton–Holt stock–recruitment relationship:
|
| (A14) |
|
|
|
|
Biomass
Estimates of biomass are obtained using independently estimated start- and mid-year mass-at-age relationships given by wa and wa+1/2, respectively (Punt and Leslie, 1991). Spawning biomass is estimated assuming knife-edge maturity, with all individuals older than 4 years mature and equally fecund:
|
| (A15) |
|
| (A16) |
Available survey biomass is given by
|
| (A17) |
|
| (A18) |
Fished proportion
The proportion of the resource harvested (F) is estimated directly from the data, using the catches by weight (W) for each grid g:
|
| (A19) |
Catches-at-age
Commercial catches-at-age are evaluated as
|
| (A20) |
|
| (A21) |
|
| (A22) |
Catches-at-length
Length-at-age is described by the von Bertalanffy growth equation:
|
| (A23) |
Published parameter values were used (Punt and Leslie, 1991), with l
= 219.4 cm, 
= 0.049 year–1, and a0 = –0.914 years. Owing to differences in the growth rates of individual fish, the length distribution of fish of age a is modelled by a normal distribution, truncated to three standard deviations in both directions (N*), with the standard deviation increasing with age:
|
| (A24) |
This allows calculation of the age-to-length conversion matrix A, which describes the proportion of fish of age a that are in length category l. It then follows that:
|
| (A25) |
Maximum likelihood parameter estimation
Abundance indices
The commercial and survey catch rates are treated in an analogous fashion. Observation error with a lognormal distribution is assumed so that catch rate (I) for grid (g) and year (y) is given by
|
| (A26) |
and 
.
Because biomass is estimated at the level of the zone, we assume it to be uniformly distributed across grids in that zone, so that
|
| (A27) |
1 for all g 
z.
Assuming that qg = qz, it follows that
|
| (A28) |
We therefore calculate the observation error for each grid as
|
| (A29) |
z and Igy > 0.
Zero Igy values were associated with zero effort, so provided no information. The negative log-likelihood is then given by
|
| (A30) |
We assume a constant 
across grids in a particular zone (
) and for all years. Both 
and 
are obtained analytically from their maximum likelihood values, so that
|
| (A31) |
|
| (A32) |
1 for all g z and Igy > 0.
Age and length frequency data
The model is fitted to both catch-at-age data and directly to catch-at-length data for years where age–length keys are not available. Likelihood calculations for both types of data are analogous. Here we describe fitting to catch-at-age (survey) data.
We again assume observation error with a lognormal distribution, so that the proportion of the catch (p) for each age (a), grid (g), and year (y) is given by
|
| (A33) |
gyaNadj (0,
g2) and 
is assumed to follow an adjusted normal distribution to take account of sampling variability across ages [see Equation (A35)]. Because catches are evaluated at the level of the zone [see Equation (A20)], predicted catches for each zone are assumed to be distributed uniformly across grids within that zone, so that catch in each grid is equal to the mean catch across grids. This gives
|
| (A34) |
z.
The contribution to the negative log-likelihood is given by
|
| (A35) |
for all g z and pgya >0. Zero pgya values refer to grids for which no catch-at-length data were available.
is again assumed to be constant across grids, with 
estimated analytically from its maximum likelihood value:
|
| (A36) |
1 for all g z and pgya > 0.
Summation of the likelihood contributions
Finally, the various contributions to the negative log-likelihood are summed:
|
| (A37) |
|
Acknowledgements |
|---|
The contributions of Dave Japp (Capricorn Fisheries Monitoring) and Robin Leslie, Tracey Fairweather, and Jean Glazer (Marine and Coastal Management, South Africa) in providing the data and associated background information for the project are gratefully acknowledged. We also thank Colin Attwood (Department of Zoology, University of Cape Town) and Kerry Sink (South African National Biodiversity Institute) for useful discussions. Comments and suggestions by two reviewers (Trevor Hutton and Daniel Howell) are also acknowledged. This study was funded by the National Research Foundation of South Africa (CTTE and EEP), which also supported CE's attendance at the MPA Symposium 2007, Murcia, Spain, and Marine and Coastal Management (RR).
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussion and conclusionsAppendix: Population model for...References |
|---|
-
Auster P., Shackell N. Marine protected areas for the temperate and boreal Northwest Atlantic: the potential for sustainable fisheries and conservation of biodiversity. Northeastern Naturalist (2000) 7:419–434.[CrossRef]
Berkeley S., Hixon M., Larson R., Love M. Fisheries sustainability via protection of age structure and spatial distribution of fish populations. Fisheries (2004) 29:23–32.
Birkeland C., Dayton P. The importance in fishery management of leaving the big ones. Trends in Ecology and Evolution (2005) 20:356–358.[CrossRef]
Botsford L., Castilla J., Peterson C. The management of fisheries and marine ecosystems. Science (1997) 277:509–515.
Butterworth D., Rademeyer R. Sustainable management initiatives for the Southern African hake fisheries over recent years. Bulletin of Marine Science (2005) 76:287–320.[Web of Science]
Clark C. Marine reserves and the precautionary management of fisheries. Ecological Applications (1996) 6:369–370.[CrossRef][Web of Science]
Dinmore T. A., Duplisea D. E., Rackham B. D., Maxwell D. L., Jennings S. Impact of a large-scale area closure on patterns of fishing disturbance and the consequences for benthic communities. ICES Journal of Marine Science (2003) 60:371–380.
Field J., Punt A., Methot R., Thomson C. Does MPA mean major problem for assessments? Considering the consequences of place-based management systems. Fish and Fisheries (2006) 7:284–302.[CrossRef][Web of Science]
Gaylard J., Bergh M. A species splitting mechanism for application to the commercial hake catch data 1978 to 2003. (2004) 8. Marine and Coastal Management Document WG/09/04/D: H: 21.
Gell F., Roberts C. The fishery effects of marine reserves and fishery closures. (2003) Washington, DC: WWF-US. Technical Report.
Gerber L., Botsford L., Hastings A., Possingham H., Gaines S., Palumbi S., Andelman S. Population models for marine reserve design: a retrospective and prospective synthesis. Ecological Applications (2003) 13(Suppl):S47–S64.[CrossRef][Web of Science]
Glazer J. P. Updated hake GLM-standardized cpue series. (2007) 22. Marine and Coastal Management Document 2007:WG-Dem:H:06.
Glazer J. P., Butterworth D. S. GLM-based standardisation of the catch per unit effort series for the South African West Coast hake, focusing on adjustments for targeting other species. South African Journal of Marine Science (2002) 24:323–339.
Halpern B. The impact of marine reserves: do reserves work and does reserve size matter? Ecological Applications (2003) 13(Suppl):S117–S137.[CrossRef][Web of Science]
Halpern B., Warner R. Marine reserves have rapid and lasting effects. Ecology Letters (2002) 5:361–366.[CrossRef][Web of Science]
Hiddink J. G., Hutton T., Jennings S., Kaiser M. Predicting the effects of area closures and fishing effort restrictions on the production, biomass, and species richness of benthic invertebrate communities. ICES Journal of Marine Science (2006) 63:822–830.
Hilborn R., Stokes K., Maguire J., Smith T., Botsford L., Mangel M., Orensanz J., et al. When can marine reserves improve fisheries management? Ocean and Coastal Management (2004) 47:197–205.[CrossRef]
Hobday D., Punt A., Smith D. Modelling the effects of Marine Protected Areas (MPAs) on the southern rock lobster (Jasus edwardsii) fishery of Victoria, Australia. New Zealand Journal of Marine and Freshwater Research (2005) 39:675–686.[Web of Science]
Holland D. A bioeconomic model of marine sanctuaries on Georges Bank. Canadian Journal of Fisheries and Aquatic Sciences (2000) 57:1307–1319.
Holland D. Integrating marine protected areas into models for fishery assessment and management. Natural Resource Modelling (2002) 25:369–386.
ICES. Effort allocation of the Dutch beam trawl fleet in response to a temporarily closed area in the North Sea. (2001) 17. ICES Document CM 2001/N: 01.
Jennings S., Kaiser M. The effects of fishing on marine ecosystems. Advances in Marine Biology (1998) 34:201–352.[CrossRef][Web of Science]
Larson R., Julian R. Spatial and temporal genetic patchiness in marine populations and their implications of fisheries management. CalCOFI Report (1999) 40:94–99.
Lauck T., Clark C., Mangel M., Munro G. Implementing the precautionary principle in fisheries management through marine reserves. Ecological Applications (1998) 8:72–78.
Lombard A. T., Strauss T., Harris J., Sink K., Attwood C., Hutchings L. South African National Spatial Biodiversity Assessment 2004 Volume 4: Marine Component. (2004) Pretoria: South African National Biodiversity Institute.
Ocean Studies Board. Effects of trawling and dredging on seafloor habitat. (2002) Washington, DC: Division of Earth and Life Sciences, National Research Council.
Pelletier D., Mahévas S. Spatially explicit fisheries simulation models for policy evaluation. Fish and Fisheries (2005) 6:307–349.[CrossRef][Web of Science]
Polacheck T. Year round closed areas as a management tool. Natural Resource Modelling (1990) 4:327–353.
Punt A., Leslie R. Estimates of some biological parameters for the Cape hakes off the South African West coast. South African Journal of Marine Science (1991) 10:271–284.
Rademeyer R. A. Assessment and management procedures for the hake stocks off Southern Africa. (2004) University of Cape Town. MSc thesis.
Rademeyer R. A., Butterworth D. S. Updated reference set for the South African Merluccius paradoxus and M. capensis resources and projections under a series of candidate OMPs. (2006) 27. Marine and Coastal Management Document WG/09/06/DH: 33.
Roberts C. M., Polunin N. V. C. Marine reserves: simple solutions to managing complex fisheries. Ambio (1993) 22:363–368.
Rodwell L. D., Barbier E. B., Roberts C. M., McClanahan T. R. The importance of habitat quality for marine reserve-fishery linkages. Canadian Journal of Fisheries and Aquatic Sciences (2003) 60:171–181.
Sainsbury K. J., Campbell R. A., Lindholm R., Whitelaw W. Experimental management of an Australian multispecies fishery: examining the possibility of trawl-induced habitat modification. American Fisheries Society Symposium (1997) 20:107–112.[Medline]
Thrush S. F., Dayton P. K. Disturbance to marine benthic habitats by trawling and dredging: implications for marine biodiversity. Annual Review of Ecology and Systematics (2002) 33:449–473.[CrossRef][Web of Science]
Tupper M., Boutilier R. G. Effects of habitat on settlement, growth and postsettlement survival of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences (1995) 52:1834–1841.
Walters C., Bonfil R. Multispecies spatial assessment models for the British Columbia groundfish trawl fishery. Canadian Journal of Fisheries and Aquatic Sciences (1999) 56:601–628.
Walters C., Hilborn R., Parrish R. An equilibrium model for predicting the efficacy of marine protected areas in coastal environments. Canadian Journal of Fisheries and Aquatic Sciences (2007) 64:1009–1018.
Willis T. J., Millar R. B., Babcock R. C., Tolimieri N. Burdens of evidence and the benefits of marine reserves: putting Descartes before des horse? Environmental Conservation (2003) 30:97–103.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||