© 2006 International Council for the Exploration of the Sea
The Irish Sea cod recovery plan: some lessons learned
a Fisheries Science Services, Marine Institute Galway Technology Park, Galway, Ireland, United Kingdom
b Department of Zoology, Ecology and Plant Science, University College Cork Cork, Ireland, United Kingdom
*Correspondence to C. J. Kelly: tel: +44 353 91 730 400; fax: +44 353 91 730 470. e-mail: ciaran.kelly{at}marine.ie.
Historically, cod has been one of the most important fish stocks in the North Atlantic. Recent stock collapses have been attributed to overfishing, and in February 2000 the European Commission established a closed area in the Irish Sea as part of a general recovery plan. The recovery plan was further revised and implemented between 2001 and 2005. However, the recovery plan has not provided the expected benefit, and the stock is still thought to be below the safe limit of Blim = 6000 t. We use stochastic simulations to investigate possible stock trajectories over a seven-year period from 1999 to 2005 under various scenarios of fishing mortality. Comparing the results of simulations with reality, it is clear that more drastic action is required if the stock is to recover in the medium term. The recovery plan was not explicitly designed to deal with uncertainty in the system, and this, we believe, resulted in the failure to meet the recovery plan objectives.
Keywords: closed areas, Irish Sea cod, recovery plans, stochastic simulations
Received 5 April 2005; accepted 13 December 2005.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionReferences |
|---|
The cod (Gadus morhua) is one of the most widely recognized, well-researched fish in the North Atlantic, with a documented history of exploitation extending for more than 1000 years (Kurlansky, 1997). In most of the areas where it is found, cod are important top predators (Essington and Hansson, 2004; Savenkoff et al., 2004). In the Irish Sea (ICES Area VIIa), where the species is close to its southern limit of distribution, cod mature at 2–3 years of age (ICES, 2003a; Armstrong et al., 2004), and have been exploited by targeted and mixed-species commercial fisheries for several centuries (Kurlansky, 1997). Cod are an important source of revenue; in 2002 the value of the Irish commercial quota was around
7 million (Fisheries Science Services, 2003). As with many North Atlantic cod stocks, Irish Sea cod have experienced a continuous decline in spawning-stock biomass (SSB) over the past 20 years (Cook et al., 1997; Christensen et al., 2003; Myers and Worm, 2003). The biomass was estimated to be so low in 1999 that ICES (the International Council for the Exploration of the Sea) advised that the stock was in danger of collapse, and recommended that a recovery plan be put in place (ICES, 2001). In February 2000, the European Commission established measures to aid recovery (Anon., 2000a). These measures initially included two closed areas in the eastern and western Irish Sea to provide the maximum possible protection during the spawning season and to maximize egg production of the existing stock. The closed areas were based on the putative spawning grounds at peak spawning time (14 February–30 April; ICES, 2003b). The closures applied to all fishing activities, excepting derogations for Norway lobster (Nephrops norvegicus) trawls and beam trawlers, which were permitted to fish in defined "boxes" within the closed areas (Figure 1a). Because of the initial restrictions, some sections of the fleet felt unfairly constrained, and many fishing representative groups lobbied for mitigating measures that would not reduce the opportunity of fishing for other species.
|
Additional measures were adopted in November 2000 (Anon., 2000b), banning various technical specifications of towed nets. The regulation stated in its preamble "The closure to protect cod should, therefore, be established in such a way that fisheries for Norway lobster, shrimps and flatfish, should not be significantly diminished while minimising risk to cod." The extent of the closed area and derogations for fishing within the area were amended in February 2001 (Anon., 2001a), limiting the closure to the eastern Irish Sea and permitting two types of fishing within the reduced closed area through derogation (Figure 1b). This was again further amended in July 2001 (Anon., 2001b), permitting the use of double twine no greater than 4 mm in the construction of the codend of the trawls.
The recovery plan was further specified in 2004 (Anon., 2004), setting a target biomass for the stock of Bpa = 10 000 t and establishing procedures for the setting of the total allowable catch (TAC). These procedures were designed to ensure a 30% annual increase in SSB (relative to the most recent assessment estimates of stock size) and to limit TAC changes to 15%.
During that time there were consultations with fishers, but there were no compensation packages for those disadvantaged by the scheme, as in the case of the Canadian Northwest Atlantic cod fishery plan (Rice et al., 2003). As such, the Irish Sea cod recovery plan relied for its success on the reduction of quotas, the closure of spawning grounds, and technical gear regulations (ICES, 2003a).
The result of the recovery plan in terms of landings was an initial decrease, followed by increased landings and unreported catches. In terms of SSB, the "recovery" did not yield the expected gain, and some six years on, the stock is still well below Bpa (10 000 t), and is likely also to be below Blim (6000 t; see Table 1, Figure 2). The question is, therefore, why did the recovery plan not work as expected?
|
|
In this paper we attempt to answer this question by investigating hindcast projections of the Irish Sea cod stock from 1999 to 2005 using stochastic simulations under particular hypothetical scenarios relating to different levels of target fishing mortality. We demonstrate that to achieve a noticeable recovery of the stock requires a drastic cut in fishing mortality, which the recovery plans so far implemented have not produced, so unless such action is taken the stock is unlikely to recover in the short or medium term. We also discuss the design of the recovery plan and its objectives, and argue that, in its current form, it does not sufficiently deal with system uncertainty, which is likely to have contributed to its apparent failure.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Hindcast projection simulations were made with the F-PRESS simulation tool, although we stress that results are not dependent on the software and can easily be replicated using any stochastic stock projection tool. F-PRESS is an in-house single-species non-spatial stochastic stock projection tool designed to evaluate management plans and harvest control rules (HCRs; ICES, 2005a). The simulation is programmed in the R environment (R Development Core Team, 2003) and source code and user/technical manuals are available from the authors. We do not use the simulation tool explicitly to attempt to evaluate all the separate parts of the recovery plan (e.g. closed area, technical controls, and effort regulation). Instead, we look at several more general hypothetical harvest strategies based on a target level of fishing mortality, and compare these results with reality to demonstrate the inadequacies of the recovery plan in its current form.
Operating model
To project our simulated Irish Sea cod stock we use the standard Baranov (1918) model for an age-structured population with exponential mortality used in most VPAs (e.g. Beverton and Holt, 1957; Fuiman and Werner, 2002):
|
| (1) |
The underlying population dynamics model requires standard ICES data on stock characteristics, such as weight-, maturity-, and natural and fishing mortality-at-age, as well as CV values for these parameters so that realistic stochasticity can be introduced. These data are available from the Working Group on Northern Shelf Demersal Stocks (WGNSDS) as XSA output (ICES, 2004b, 2005b), and are given in Table 2.
|
Any forward projection of a fishery must include some element of stochasticity (ICES, 2004a, 2005a) to account for the high levels of variability in stock and environmental dynamics, as well as the uncertainty in assessments and the errors in data (Patterson et al., 2001). In our projection, all at-age stock parameters are randomized at the start of each year (population numbers are randomized at the start of 1999 only) by drawing new values from a Normal distribution with constant mean and variance, e.g. N(Z,
2), where Z is the assessment value of the parameter, and 2 is the variance calculated from the assessment CV of the parameter, as given in Table 2. Note that although F-at-age values are randomized each year, we do not assume any correlation between changes in relative F-at-age values and changes in TAC.
Similarly, the annual TAC in the projection is also made stochastic (to account for variations in, for example, fishing effort, catchability, or environmental factors) and may also be subject to implementation bias. If biased, the actual TAC is randomly drawn from N(Z* BZ, (BZ)2 2), where Z is the original TAC, BZ a relative measure of the bias (such that BZ = 1 corresponds to no bias), and 2 corresponds to the required level of introduced random noise.
In all random draws, negative values are rejected and redrawn, so the distributions we use are "truncated" Normal distributions. This method is effectively a "white-noise" process (e.g. Grimmett and Stirzaker, 2001), and no serial autocorrelation is assumed.
Recruitment
Recruitment is simulated using a "stochastic Ricker function" based on the standard Ricker function (Ricker, 1954):
|
| (2) |
and ß are parameters estimated by fitting the function to historical data, and R is the expected number of recruits. Annual recruitment is simulated by calculating an expected value of R using Equation (2) and the virtual SSB level, then randomly drawing the actual R value from a Normal distribution (in the same way that stock characteristics are randomized in the operating model). For Irish Sea cod for the period 1968–2003, the CV in the recruitment values is approximately 0.6 (data from WGNSDS; ICES, 2004b), and we use this value when drawing random recruitment. Typical values for recruitment given by this method are shown in Figure 3; they illustrate that although the method is simple [compared with the time-series residual model of Needle et al. (2003)], it produces qualitatively realistic recruitment behaviour. Recruitment models in the literature have been subject to varying degrees of criticism (Needle, 2002), for instance for not properly taking into account the environmental effects such as sea temperature (Planque and Frédou, 1999; O'Brien et al., 2000; Planque et al., 2003), temporal cycles such as the North Atlantic Oscillation (NAO; Brander and Mohn, 2004), plankton levels (Beaugrand et al., 2003), cultivation/depensation effects (Myers et al., 1995; Walters and Kitchell, 2001), or other environmental factors. However, we only attempt to replicate qualitative recruitment behaviour in order to compare the effectiveness of different management strategies in the short and medium term, and a simple model suffices. A more complex model would be needed for longer-term projections if, for example, a direct link between sea temperature and recruitment was demonstrated, and that sea temperature is expected to change significantly in the longer term (Clark et al., 2003).
|
As the Ricker function in Equation (2) is based on historically observed data, any results where the stock is at high SSB outside the range of historical data (e.g. >30 000 t) should be treated with scepticism.
Assessment model
The ICES Working Group on Methods of Fish Stock Assessments (WGMG; ICES, 2004a) and the Study Group on Management Strategies (SGMAS; ICES, 2005a) suggest three possible ways of modelling the assessment process in a fishery simulation model: (i) perfect assessment is assumed and the exact values of the virtual stock characteristics (e.g. SSB, 
) are used in any management model; (ii) a quantifiable level of bias and random error is added to the virtual data; and (iii) the full ICES assessment software package is used with the virtual data and "real" assessment error and bias is replicated (this process is known as "assessment feedback"). In our projections we either use method (i) or method (ii), where management is based on the observed . An advantage of method (ii) over the "assessment feedback" approach is that we are able to control and quantify the exact level of bias and random error in our virtual assessment. For method (ii), bias and random error are introduced to the observed by randomly drawing new values from N(Z* BZ, (BZ)2 2).
Harvesting scenarios
Simulations were run using the F-PRESS tool under four hypothetical scenarios of harvest strategy and different levels of assessment/implementation error. The scenarios all use a simple harvest control rule (HCR; ICES, 2005a) to manage the level of exploitation of the virtual stock. No attempt is made explicitly to simulate the recovery plan procedures introduced in 2004 (Anon., 2004), where the target is a 30% increase in SSB each year, for reasons that are discussed later. Instead, it is assumed that the management objective is to achieve a target level of fishing mortality.
A TAC is set each year (initial TAC = 4000 t in 1999, slightly less than the WGNSDS estimates, see Table 1) based on the observed level of relative to the desired target fishing mortality FT. The general form of the HCR is
|
| (3) |
*n–1 is the observed fishing mortality in the previous year (this may be different from the actual virtual n–1 value depending upon whether the simulated assessment has random error and bias). Essentially most standard ICES stocks are managed in a similar manner; annual TACs are adjusted depending on the perceived level of fishing mortality and stock size relative to some target level of fishing mortality and stock size. No attempt is made to find an optimal HCR; the aim is rather to replicate the traditional ICES management process in a qualitative manner.
For all scenarios (HCR 1–HCR 4), a range of target fishing mortality values is used (0.1–1.5) and 1000 seven-year projections are completed for each value of FT. The details of each scenario are given in Table 3. Scenario HCR 1 is simply the general form of the HCR given in Equation (2) with no restriction on annual change in TAC, while perfect assessment and no bias (but a level of random error: CV = 0.1) in implementation/enforcement is assumed. HCR 2 is the same as HCR 1, except that a constraint of a maximum year-on-year change in TAC of 15% is applied. HCR 3 is also the same as HCR 1, but in this case there is assumed to be a bias of 20% and random error (CV = 0.2) in the assessment-estimated value of virtual SSB and the n–1 values (bias is such that SSB is overestimated and n–1 is underestimated), and an upward implementation bias of 20% is applied to catches. HCR 4 is the same as HCR 3, except that a constraint of a maximum year-on-year change in TAC of 15% is applied. Our assumed levels of assessment bias and error are conservative; in reality they may be much higher, as can be seen in retrospective assessment errors; see Figure 1 and ICES (2003b, 2004b). The level of implementation bias is also likely to be conservative, although this is harder to quantify if catch data are unreliable (ICES, 2004b).
|
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
From the results shown in Figure 4 and Table 4, it is clear that the stock is in its healthiest state when the target fishing mortality, FT, is lowest (as would be expected). We should remain sceptical of the extremely high SSB values (>30 000 t) at the lowest levels of fishing mortality because they lie outside the range of historical observations on which the recruitment model is based. Care should also be taken in interpreting the results. The trend of the mean final SSB is of most interest, but the spread of the distribution of SSB values from the projections should also be considered. Table 4 shows that for certain levels of target fishing mortality, all harvest strategies had projections that resulted in "extinction" of the stock (even at the lowest level of target fishing mortality for HCR 2 and HCR 4).
|
|
Comparing SSB trends in HCR 1 and HCR 3 (strategies with no limit on annual TAC changes), the introduction of conservative levels of bias and random error in the assessment and implementation results in lower expected final SSB values, although the difference is not great (Figure 4). However, it is the difference when comparing HCR 1 with HCR 2, and HCR 3 with HCR 4 that is of most interest. The introduction of a 15% limit to the annual changes in TAC results in a massive impairment to the likely recovery of the stock (assuming low target fishing mortality). In particular, the 15% limit combined with the conservative assessment/implementation bias (HCR 4) results in a stock that is seemingly likely to be "stuck" below Bpa irrespective of the target fishing mortality.
These results are also confirmed in Figure 5, in which it is clear that with HCR 1 and HCR 3 there is a low probability of SSB being below Bpa (similar results hold for Blim) at the end of the projection period when the target fishing mortality is also low. However, the results for HCR 2 and HCR 4 illustrate that even with a low target fishing mortality, there is still a significant probability that the stock will be below Bpa at the end of the projection. For all harvest strategies, there is a high probability of being below Blim or Bpa at the end of the projection when the target fishing mortality is "medium" (FT = 0.8) or "high" (FT = 1.4). It is interesting that, with HCR 2, there is slightly less risk of being below Blim and Bpa at high levels of target fishing mortality than with HCR 1 (p(SSB < Bpa) = 0.71 for HCR 2 and p(SSB < Bpa) = 0.89 for HCR 1, when FT = 1.4). This is simply because the annual TAC change limit works to stop exploitation increasing too quickly. However, when conservative levels of bias and error are introduced (HCR 4), there is great risk to the stock at all levels of target fishing mortality.
|
The optimal catch is obtained at a low-to-medium level of target fishing mortality in HCR 1 and HCR 3, but mean catches are much lower and more independent of the target fishing mortality for HCR 2 and HCR 4, in which there is a limit on TAC change (Figure 6). It is noteworthy that the optimal catch in HCR 1 is obtained when FT
0.7, close to the Fpa = 0.72 value recommended by ICES as the limit level of fishing mortality (ICES, 2004b). However, when conservative bias and error are included (HCR 3), the optimal catch is obtained at a lower level of target fishing mortality (FT 0.45). Our results suggest that including the 15% limit on changes in TAC will actually reduce the overall yield (e.g. from Table 4: cf. mean catch = 4970 t for HCR 1 with FT = 0.8; mean catch = 3000 t for HCR 2 with the same FT). Moreover, from Table 4 it is clear that only in the case of perfect assessment and enforcement (HCR 2) will the 15% limit act to reduce yield variability. Including conservative bias and error (HCR 4), the variability in yield is higher and in average yield lower for all levels of target fishing mortality (23% larger mean CV in catch and 67% lower mean catch when FT = 0.8) than if the 15% limit was not in place (HCR 3).
|
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The latest stock assessment for Irish Sea cod (ICES, 2005b) is highly uncertain owing to problems with unreliable data, but SSB is thought to be still well below Blim = 6000 t, and the stock is classed as being "at reduced reproductive capacity and being harvested unsustainably" and as "overexploited". From 1999 to 2004, the SSB of the stock is thought to have been below Blim every year, while the average
2–4 over the same period was approximately 1.3. In fact, the Irish Sea cod stock at the end of 2004 seemed to be in a state similar to that in 1999, so the recovery plan seems to have had little effect. This is a point to bear in mind when considering the simulation results discussed below: our conclusions drawn from hindcast projections for the period 1999–2005 are likely to be equally valid for forecasting the future of the stock from 2005 on. The simulations presented do not attempt explicitly to evaluate the constituent elements of the Irish Sea cod recovery plan (e.g. closed area, technical measures, and effort regulations). Instead, we attempt to demonstrate elements of the dynamics of Irish Sea cod, and to show how the stock is likely to respond to different levels of fishing mortality implemented using a simple generic harvesting strategy. As the projections are based on an underlying stochastic model, it is possible to illustrate how the probability of stock recovery is affected by conservative levels of bias and random error in the simulated assessment, implementation, and enforcement, or by adding constraints on annual TAC changes.
Comparing the projected SSB levels and probability of stock recovery (Figures 4 and 5, Table 4) with the actual perceived state of the stock (ICES, 2005b), it seems unsurprising that the stock has not recovered given the high levels of fishing mortality over the period 1999–2005. Even the most optimistic results from the HCR simulations presented here (HCR 1: perfect assessment and enforcement) show that unless fishing mortality is reduced significantly (by about 33%), so the probability of recovery will be low (<50%; Figure 5). When a realistic (in fact highly conservative) level of 20% bias is introduced to the simulated assessment and implementation (HCR 3), the probability of recovery is even less, suggesting that fishing mortality would need to be cut by more than half. The introduction of bias in the HCR simulation shows how sensitive the stock recovery is to the assessment error or the exceeding of quotas. In reality, levels of bias and error in the assessment may be much higher (retrospective errors in assessment estimates of the most recent SSB can be 60% or higher, see ICES, 2003b, 2004b, 2005b, and Figure 2), and TACs have frequently been exceeded, particularly when drastic cuts in quota have been implemented (ICES, 2003b, 2004b). Further, the simulations explicitly do not include discards because the relation between changes in discard rates and TAC or effort levels in cod fisheries remains uncertain (Myers et al., 1997; Borges et al., 2005). We assume also that there are no other mixed fisheries, ecosystem or environmental factors, but the reality is likely to be more complex, and immediate stock recovery may not actually be possible (Hutchings, 2000). This could be due to other organisms moving into the "ecological niche" previously occupied by cod, leading to a cultivation/depensation effect (Walters and Kitchell, 2001). Taking these points into consideration means that our simulation results have to be considered optimistic.
The situation is worsened by introducing a 15% limit on annual TAC changes (HCR 2 and HCR 4). Whereas the introduction of a TAC change limitation has some merit in stabilizing yields in healthy stocks, its use in the situation of stock recovery is questionable. A TAC limitation only decreases risk to the stock where the relative target fishing mortality is too high (compare HCR 1 and HCR 2 in Figure 5). However, when there is bias in the assessment and implementation, the TAC change limitation decreases the probability of recovery and gives mean yields lower and slightly more variable than when no limit is applied (Table 4). The addition of a mechanism to stabilize yield can actually result in higher annual variability in catch and a lower average catch. Effectively, the TAC change limit stops highly reactive management. For a stock at a (safe) stable SSB level, highly reactive management is not particularly desirable, and the priority is stability in yield. However, when the stock is in a recovery state, it is essential that fishing mortality be controlled quickly to counteract, for example, environmental or biological effects such as recruitment failure. It is clear that the priority for a recovery stock should be flexibility of management and reduction of exploitation, rather than stability in yield (Caddy and Agnew, 2003).
The Irish Sea cod recovery plan included a 15% TAC limit similar to that in HCR 3 and HCR 4, although clauses were included such that the rule could be overridden if the stock was thought to be below Blim or fishing mortality above Fpa = 0.72 (Anon., 2004). However, the agreed TAC actually increased from 1950 t in 2003 [where SSB < Blim, 2–4 < Fpa and ICES advice (ICES, 2003b) was to close all fisheries for cod] to 2150 t in 2004 and 2005. It remains to be seen how recovery proceeds if the objectives of the agreed management plan do not appear to be adhered to.
Rather than a target fishing mortality (as simulated) the primary objective of the recovery plan is "to set a TAC that achieves a 30% increase in the SSB from one year to the next". However, we believe that this objective is fundamentally flawed for reasons we outline below. In fact, when discussing the Irish Sea cod recovery plan, ICES (2005b) states, "The management plan has not been evaluated by ICES ... the plan requires annual predictions of spawning stock size, which is not available given the recent poor catch data a plan that does not require such a precision should be considered." As discussed above, the retrospective year-on-year errors in assessment estimates of SSB are high, and problems with recent catch data (ICES, 2004b, 2005b) mean that these errors are unlikely to be mitigated in the short term. This means that it is almost impossible to be certain that a 30% increase in SSB has been observed until retrospective errors are removed, so it is hard to see if the plan is working until years later. An easily measurable objective is a key point for the success of any recovery plan (Powers, 2003), and this appears to be lacking in the case of the Irish Sea cod recovery plan. Further, even with a perfect SSB estimate to calculate the TAC and (seemingly) produce a 30% increase in SSB, there is still great uncertainty as to whether this will actually happen given both implementation bias and error (exceeding of quotas) and the impact of environmental factors. As stated by ICES (2005b), an alternative strategy is required, and a better strategy could be to aim for a low target fishing mortality (and to reduce TACs accordingly), as illustrated in the simulations here.
Assuming that a low target fishing mortality is a more sensible objective, there are still problems with the constituent elements of the Irish Sea cod recovery plan and their apparent failure significantly to reduce fishing mortality or to aid stock recovery. The recovery plan did not set out to compensate fishers for their loss of revenue caused by reduced opportunities to fish. Moreover, it was not an aim of the recovery plan in its initial formulation to reduce effort (ICES, 2003b), and in fact the plan stated that opportunities to fish other species "should not be significantly diminished" (Anon., 2000a). Therefore, if an area was closed for a period, it would be natural to assume that some effort would be displaced either temporally or spatially. There is some evidence for this, with the midwater trawl fishery increasing its effort in the North Channel part of Area VIIa (ICES, 2003b). Such effort displacement would undermine the benefit of a closed area in reducing exploitation. There were also increases in twin-trawling for Nephrops in the closed area (under derogation). Both Nephrops trawlers and beam trawlers (which were allowed into the closed areas under derogations) had cod bycatch (ICES, 2003b, 2004b). Although bycatch limits were specified in the recovery plan (Anon., 2001a), discard data were unavailable across all fleets (ICES, 2003b, 2004b). This lack of data prevents a complete analysis of the impact of the derogated fisheries. However, it is likely that the potential of the closed areas in reducing exploitation was eroded somewhat by creating derogations for other fishing enterprises, whose impact could not be evaluated fully. This highlights the lack of another important aspect in the Irish Sea cod recovery plan: namely, that sufficient data be collected to allow evaluation of the impact of the plan on the stock.
The initial recovery plan, included closed areas to "allow as many cod as possible to spawn" as part of the plan to rebuild the stock (Anon., 2000a). This strategy implies a relationship between reproductive potential and recruitment. However, such a relationship is far from assured. Even the obvious link between reproductive potential and SSB is subject to interannual fluctuation through egg viability (Marshall et al., 1998). The forward link to recruitment is additionally thought to be subject to factors influencing the survival and growth of larvae and juveniles, including temperature, primary production, and predation (e.g. Planque and Frédou, 1999; Beaugrand et al., 2003). Given the complexity of this relationship, protecting spawning offers only a tentative probability of increased recruitment. Even if the closed areas did effect 100% protection of spawning cod, the benefit to the stock in terms of recovery would be subject to the prevailing environmental and ecological conditions (and, of course, the exploitation rate). Such a measure to protect spawning to increase recruitment would therefore seem at best passive, and at worst ineffective. Horwood et al. (1998) discuss how use of a closed area to protect the spawning stock may actually be counterproductive, because fishing mortality can be displaced onto juvenile fish outside the spawning grounds. They suggest that better use of a closed area would be permanently to close the nursery areas of the stock, so reducing discards of juvenile fish. However, even closed nursery areas have been hard to evaluate, and results are not unequivocally positive (Pastoors et al., 2000).
Even with complete exclusion of all fishing fleets, a closed area alone may not be enough to reduce exploitation of a population. If there is significant movement of the fish stock between the closed area and the fishery, then the closed area is unlikely to be effective (Horwood et al., 1998). Cod are capable of moving large distances (Robichaud and Rose, 2004) so it is unlikely that a closed area on its own will be sufficient to protect the stock, and any recovery plan would need to include further effort or TAC restrictions as well.
The lack of an appreciation of risk associated with a management strategy has contributed to further difficulties with the cod recovery plan. These difficulties were manifested in the frustration of fishers and managers at how long the recovery process was going to take. Other contributors to this frustration were poor communication of the inherent uncertainty and likely probability of success, and the lack of clarity in the purpose of the recovery plan. These frustrations were amplified when the stock did not appear to "follow the plan" as originally envisaged, or worse still, it could not be shown if the plan was working. This highlights the need for scientists to communicate to both managers and fishers the uncertainty and inherent levels of risk in a strategy, and for targets to be based on measurable improvements in the stock associated with this risk.
Given the above, we suggest that a better approach to the development of a recovery plan for Irish Sea cod would include the following:
- Clear purpose which effectively communicates that the instrument of recovery is the reduction in exploitation, and how this is to be achieved.
- Clear understanding that this will require a reduction in fishing opportunities, and a consideration of the fleet-specific reduction in revenue of such reduced exploitation.
- Clear means as to how this reduction will be adhered to.
- Clear, measurable performance targets, underpinned by sufficient data collection to assess performance of recovery, and an understanding of the inherent uncertainty involved.
- A multispecies harvest plan to manage the stock when (or if) recovery is achieved.
Even so, there is no guarantee that these measures if rigidly applied will lead to recovery of the stock, and this possibility should be discussed openly. However, such suggestions should allow for clear evaluation of the state of the stock in relation to both management actions and objectives.
|
Acknowledgements |
|---|
We acknowledge those who helped in the production of this manuscript, in particular Denis Rivard, Martin Pastoors, and an anonymous referee for helpful comments on an early draft. EAC was supported through a project (Grant-aid Agreement No. PDOC/01/001) funded by the Marine Institute and the Marine RTDI Measure, Productive Sector Operational Programme, National Development Plan 2000–2006.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
-
Anon. (2000) Commission Regulation (EC) No. 304/2000 of 9 February 2000 establishing measures for the recovery of the stock of cod in the Irish Sea (ICES Division VIIA). Official Journal of the European Communities L35:10–11.
Anon. (2000) Commission Regulation (EC) No. 2549/2000 of 17 November 2000 establishing additional technical measures for the recovery of the stock of cod in the Irish Sea (ICES Division VIIa). Official Journal of the European Communities L292:5–6.
Anon. (2001) Commission Regulation (EC) No. 300/2001 of 14 February 2001 establishing measures to be applied for the recovery of the stock of cod in the Irish Sea (ICES Division VIIa). Official Journal of the European Communities L44:12–14.
Anon. (2001) Commission Regulation (EC) No. 1456/2001 of 16 July 2001 amending regulation (EC) 2549/2000 establishing additional technical measures for the recovery of the stock of cod in the Irish Sea (ICES Division VIIa). Official Journal of the European Communities L194:1.
Anon. (2004) Commission Regulation (EC) No 423/2004 of 26 February 2004 establishing measures for the recovery of the cod stocks. Official Journal of the European Communities L70:8–11.
Armstrong M.J., Gerritsen H.D., Allen M., McCurdy W.J., Peel J.A.D. (2004) Variability in maturity and growth in a heavily exploited stock: cod (Gadus morhua L.) in the Irish Sea. ICES Journal of Marine Science 61:98–112.[CrossRef][Web of Science]
Baranov T.I. (1918) On the question of the biological basis of fisheries. Nauchnyi Issledovatelskii Ikhtiologicheskii Institut Isvestia 1:81–128.
Beaugrand G., Brander K.M., Lindley J.A., Souissi S., Reid P.C. (2003) Plankton effect on cod recruitment in the North Sea. Nature 426:661–664.[CrossRef][Medline]
Beverton R.J.H. and Holt S.J. (1957) On the Dynamics of Exploited Fish Populations(Chapman, and Hall, London) (1993 facsimile reprint). 544 pp.
Borges L., Rogan E., Officer R. (2005) Discarding by the demersal fishery in the waters around Ireland. Fisheries Research 76:1–13.[CrossRef][Web of Science]
Brander K. and Mohn R. (2004) Effect of the North Atlantic Oscillation on recruitment of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 61:1558–1564.
Caddy J. F. and Agnew D. (2003) A summary of global stock recovery plans for marine organisms, including indicative information on the time to recovery, and associated regime changes that may affect recruitment and recovery success. ICES Document CM 2003/U: 08. 43 pp.
Christensen V., Guenette S., Heymans J.J., Walters C.J., Watson R., Zeller D., Pauly D. (2003) Hundred-year decline of North Atlantic predatory fishes. Fish and Fisheries 4:1–24.[CrossRef][Web of Science]
Clark R.A., Fox C.J., Viner D., Livermore M. (2003) North Sea cod and climate change – modelling the effects of temperature on population dynamics. Global Change Biology 9:1669–1680.[CrossRef][Web of Science]
Cook R.M., Sinclair A., Stefansson G. (1997) Potential collapse of North Sea cod stocks. Nature 385:521–522.[CrossRef]
Essington T.E. and Hansson S. (2004) Predator-dependent functional responses and interaction strengths in a natural food web. Canadian Journal of Fisheries and Aquatic Sciences 61:2215–2226.
Fisheries Science Services. (2003) The Stock Book: Annual Review of Fish Stocks in 2003 with Management Advice for 2004(Marine Institute, Galway) 433 pp.
In Fuiman L.A. and Werner R.G. (Eds.). Fishery Science. The Unique Contributions of Early Life Stages (2002) (Blackwell Science Ltd, Oxford) 336 pp.
Grimmett G. and Stirzaker D. (2001) Probability and Random Processes(Oxford University Press, Oxford) 608 pp.
Horwood J.W., Nichols J.H., Milligan S. (1998) Evaluation of closed areas for fish stock conservation. Journal of Applied Ecology 35:893–903.[Web of Science]
Hutchings J.A. (2000) Collapse and recovery of marine fishes. Nature 406:882–885.[CrossRef][Medline]
ICES. (2001) Report of the Working Group on the assessment of Northern Shelf Demersal Stocks 2000. ICES Document CM 2001/ACFM: 01. 438 pp.
ICES. (2003a) Working Group on Ecosystem Effects of Fishing Activities (WGECO). ICES Document CM 2003/ACE: 05. 199 pp.
ICES. (2003b) Working Group on the Assessment of Northern Shelf Demersal Stocks (WGNSDS). ICES Document CM 2003/ACFM: 01. 558 pp.
ICES. (2004a) Report of the Working Group on Methods of Fish Stock Assessments (WGMG). ICES Document CM 2004/D: 03. 238 pp.
ICES. (2004b) Report of the Working Group on the Assessment of Northern Shelf Demersal Stocks (WGNSDS). ICES Document CM 2005/ACFM: 01. 722 pp.
ICES. (2005a) Report of the Study Group on Management Strategies (SGMAS). ICES Document CM 2005/ACFM: 09. 72 pp.
ICES. (2005b) Report of the Working Group on the Assessment of Northern Shelf Demersal Stocks (WGNSDS). ICES Document CM 2006/ACFM: 13. 803 pp.
Kurlansky M. (1977) Cod: A Biography of the Fish That Changed the World(Jonathan Cape, London) 304 pp.
Marshall C.T., Kjesbu O.S., Yaragina N.A., Solemdal P., Ulltang O. (1998) Is spawner biomass a sensitive measure of the reproductive and recruitment potential of Northeast Arctic cod? Canadian Journal of Fisheries and Aquatic Sciences 55:1766–1783.
Myers R.A., Barrowman N.J., Hutchings J.A., Rosenberg A.A. (1995) Population dynamics of exploited fish stocks at low population levels. Science 269:1106–1108.
Myers R.A., Hutchings J.A., Barrowman N.J. (1997) Why do fish stocks collapse? The example of cod in Atlantic Canada. Ecological Applications 7:91–106.[CrossRef][Web of Science]
Myers R.A. and Worm B. (2003) Rapid worldwide depletion of predatory fish communities. Nature 423:280–283.[CrossRef][Medline]
Needle C.L. (2002) Recruitment models: diagnosis and prognosis. Reviews in Fish Biology and Fisheries 11:95–111.[CrossRef][Web of Science]
Needle C.L., O'Brien C.M., Darby C.D., Smith M.T. (2003) Incorporating time-series structure in medium-term recruitment projections. Scientia Marina 67:201–209.
O'Brien C.M., Fox C.J., Planque B., Casey J. (2000) Climate variability and North Sea cod. Nature 404:142.[Medline]
Pastoors M.A., Rijnsdorp A.D., Van Beek F.A. (2000) Effects of a partially closed area in the North Sea (plaice box) on stock development of plaice. ICES Journal of Marine Science 57:1014–1022.
Patterson K., Cook R., Darby C., Gavaris S., Kell L., Lewy P., Mesnil B., Punt A., Restrepo V., Skagen D.W., Stefansson G. (2001) Estimating uncertainty in fish stock assessment and forecasting. Fish and Fisheries 2:125–157.[CrossRef]
Planque B. and Frédou T. (1999) Temperature and the recruitment of Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 56:2069–2077.
Planque B., Fox C.J., Saunders M.A., Rockett P. (2003) On the prediction of short term changes in the recruitment of North Sea cod (Gadus morhua) using statistical temperature forecasts. Scientia Marina 67:211–218.[Web of Science]
Powers J. (2003) Principles and realities for successful fish stock recovery – a review of some successes and failures. ICES Document CM 2003/U: 12. 14 pp.
R Development Core Team. (2003) R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, Vienna, Austria) 3-900051-00-3 http://www.R-project.org.
Rice J. C., Shelton P. A., Rivard D., Chouinard G. A., Frechet A. (2003) Recovering Canadian Atlantic cod stocks: the shape of things to come? ICES Document CM 2003/U: 06. 23 pp.
Ricker W.E. (1954) Stock and recruitment. Journal of the Fisheries Research Board of Canada 11:559–623.
Robichaud D. and Rose G.A. (2004) Migratory behaviour and range in Atlantic cod: inference from a century of tagging. Fish and Fisheries 5:185–214.[CrossRef][Web of Science]
Savenkoff C., Castonguay M., Vezina A., Despatie S.P., Chabot D., Morissette L., Hammill M.O. (2004) Inverse modelling of trophic flows through an entire ecosystem: the northern Gulf of St Lawrence in the mid-1980s. Canadian Journal of Fisheries and Aquatic Sciences 61:2194–2214.
Walters C. and Kitchell J.F. (2001) Cultivation/depensation effects on juvenile survival and recruitment: implications for the theory of fishing. Canadian Journal of Fisheries and Aquatic Sciences 58:39–50.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Miethe, C. Dytham, U. Dieckmann, and J. W. Pitchford Marine reserves and the evolutionary effects of fishing on size at maturation ICES J. Mar. Sci., November 3, 2009; (2009) fsp248v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. J. F. Le Quesne and E. A. Codling Managing mobile species with MPAs: the effects of mobility, larval dispersal, and fishing mortality on closure size ICES J. Mar. Sci., January 1, 2009; 66(1): 122 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Groger, R. A. Rountree, M. Missong, and H.-J. Ratz A stock rebuilding algorithm featuring risk assessment and an optimization strategy of single or multispecies fisheries ICES J. Mar. Sci., September 1, 2007; 64(6): 1101 - 1115. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||