Sustainable exploitation of social species: a test and comparison of models

Summary 1. Overexploitation is a major threat to the persistence of many species. A wide variety of approaches to setting ‘sustainable’ quotas for exploitation exist but there are major discrepancies between theory and practice, and only limited integration between different branches of exploitation literature. Here, we bring together and compare the efficacy of a range of different approaches to estimating sustainability. 2. A simulated population of a social mammal was used to provide data for, and compare the recommendations of, 10 widely used estimators for setting levels of sustainable exploitation. Estimators tested included four methods for setting sustainable levels of constant-effort harvesting, two approaches to assessing the sustainability of constantyield harvesting, and four systems for setting thresholds below which harvesting should cease. 3. The method used to fit catch per unit effort data to a stock dynamic model had an important influence on the variance of recommendations. Recommendations were also affected by the length of data set available and the frequency of changes in exploitation effort. Observation-error estimators were more consistent and more conservative than equilibrium, effort-averaging and process-error approaches. Harvesting at the point of maximum productivity was found to be unstable in a noisy system, suggesting the need for considerable caution when using any of these estimators. 4. Constant-yield indices, developed for use in the bush meat trade, overestimated the point at which exploitation was likely to become highly unsustainable. Sociality was an important factor underlying this finding and the assessment of sustainability in constantyield systems should give consideration to the effects of different social systems. 5. Overall, threshold-harvesting systems provided the highest mean yields in relation to extinction risk. However, the introduction of error into these systems, particularly in the form of less frequent censuses, greatly increased both variance in yields and risk of extinction.