Research Document - 2014/090
Comparison of the Fishery and Conservation Performance of Fixed- and Abundance-Based Exploitation Regimes for Coho Salmon in Southern British Columbia
By Josh Korman and Arlene Tompkins
Abstract
We compared a range of fixed- and abundance-based harvest policies for Coho salmon in Southern British Columbia using a simulation model. The model consisted of a two-stage (spawner-smolt, smolt-adult recruit) population dynamics component and a management component that simulated error in recruitment forecasts and harvest implementation, and was parameterized based on a meta-analysis of spawner-to-smolt stock recruitment data and information on marine survival rates from index stocks. The model simulated the dynamics of multiple populations of differing productivity and capacity within a management unit (MU). Performance under different harvest policies was evaluated based on simulated yield, inter-annual variability in yield, as well as metrics indexing the conservation status of individual populations. We simulated fixed harvest rates ranging from 0.1 to 0.8, abundance-based policies with a range of adult recruitment floors and ceilings, and harvest rates, and an abundance-based quota policy.
The maximum sustainable yield for the aggregate of populations within a management unit, assuming an average marine survival rate of 0.04, occurred at harvest rates ranging from 0.3-0.4. The inter-annual coefficient of variation (CV) in yield was relatively stable up to harvest rates of 0.3-0.4, after which it increased substantively due to overexploitation. Conservation failure rates increased rapidly with harvest rate with median values of 40% and 60% at harvest rates of 0.3 and 0.4, respectively. Extirpation rates were 5-fold higher at a harvest rate of 0.4 (ca. 50%) compared to 0.3 (ca. 10%).
Abundance-based harvest rates generally performed the same or slightly better than fixed harvest rate policies. More aggressive abundance-based policies (lower recruitment floors or ceilings) resulted in higher average harvest rates and poorer conservation performance compared to less aggressive policies. The weakness of abundance-based policies in general was their higher cv in catch. The PSC-type abundance-based harvest rate policy (three different harvest rates for three levels of status) generally performed as well as the continuous abundance-based policy that had the best conservation performance, but had slightly higher inter-annual variation in catch. The quota-based harvest policy had very poor performance. In most scenarios, the differences in yield and conservation measures between a fixed exploitation rate of 0.3 and the abundance-based harvest rate strategies with the best performance were not large.
Model performance was very sensitive to marine survival rate. Under the low survival regime (average survival of 0.01) none of the populations within the simulated management unit could be sustained, regardless of the harvest policy. This occurred because spawner-to-spawner productivity (a*MS*(1-H)) was on average less than one over the duration of each simulation, even for productive stocks. Not surprisingly, escapement and catch were much higher, and conservation and extirpation failure rates much lower, under the high marine survival regime (average=0.06). The difference in the conservation failure rates at a fixed exploitation rate of 0.3 or lower and any of the abundance-based policies was much greater under the high marine survival regime. This occurred because realized exploitation rates under abundance-based policies at high marine survival tended to be approximately 50%, resulting in a large relative increase in conservation failure rates for less productive stocks. Under abundance-based harvest policies, populations with lower productivity received less benefit from higher marine survival compared to more productive populations because harvest rates were higher.
Model performance was very sensitive to assumptions about conservation-related dynamics (Fig. 5). The conservation failure rate increased with the absolute level of the conservation limit, and increasing the limit at which extirpation occurred increased the extirpation rate and decreased escapement and catch. These results are not surprising and highlight the sensitivity of the model to parameters that determine conservation dynamics that are highly uncertain. However, conservation and extirpation limits in general did not affect the relative performance across harvest policies.
Performance measures were generally insensitive to most assumptions about meta-population dynamics with the exception of straying. There was little effect of random vs. fixed stream length assignment to populations within an MU. The number of populations that were simulated did not influence the median response across trials, but increases did reduce the extent of inter-trial variation in response. Conservation performance improved as the extent of straying of returning spawners to non-natal populations increased. This occurred because strays from productive populations to less productive ones increased the overall escapement to the less productive populations. The extent of straying had little effect on the relative performance of alternate harvest policies within scenarios.
Performance measures were sensitive to the extent of inter-annual variation in marine survival rates, but relatively insensitive to the extent of temporal correlation or inter-population correlation in marine survival. The simulation analysis demonstrated that reductions in harvest implementation error through better in-season management can potentially lead to improvements in both conservation status of less productive populations and fisheries yields.
Performance measures were not sensitive to forecast error. This was expected for the fixed exploitation rate policies that do not depend on recruitment forecasts, but was surprising for abundance-based regimes that do. This insensitivity was caused by the fundamental flaw of using aggregate abundance to determine a harvest rate that protects less productive populations within the aggregate. An abundance-based rule that depends on the aggregate recruitment will still overexploit weak populations regardless of the error in recruitment forecasts. Overexploitation of these populations leads to reduced performance in terms of both conservation and yield.
In conclusion, abundance-based harvest rate rules only make sense if their limit reference points are based on the status of the weak populations that they are designed to protect. This analysis indicated that a fixed exploitation rate of 0.3 resulted in similar yield and conservation performance relative to abundance-based policies. Considering the higher inter-annual variation in yield associated with abundance-based policies and additional management costs required for implementation (e.g., recruitment forecasting), a fixed exploitation strategy of 0.3 was the optimal harvest policy that was examined. Although this conclusion was robust to a number of model assumptions, it is preliminary and is not a management recommendation. There is considerable uncertainty about the extent of depensation in spawner-to-smolt stock-recruitment relationships, the exchangeability of stock-recruitment parameters among MUs, and potential biases in meeting target harvest rates.
Accessibility Notice
This document is available in PDF format. If the document is not accessible to you, please contact the Secretariat to obtain another appropriate format, such as regular print, large print, Braille or audio version.
- Date modified: