The impact of alternative harvesting strategies in a resource–consumer metapopulation

Summary 1.  The majority of our knowledge of harvested populations has been drawn from studies on single‐species occupying continuous landscapes. Population dynamics in spatially structured landscapes are markedly different from those in continuous habitats. We investigate the effects of several harvesting strategies on the metapopulation dynamics of the bruchid beetle Callosobruchus maculatus and its parasitoid Anisopteromalus calandrae, when locally extinction prone populations of bruchids were harvested. By harvesting the resource in this closed resource–consumer interaction, we could examine the scenario where humans and natural predators share a common resource. 2.  Using population‐level models in single and coupled‐patch systems, simulations were run to estimate sustainable harvesting levels for each harvesting strategy. These harvesting levels were then implemented in experimental metapopulation microcosms. Experiments were run for multiple generations and long‐term time series of population size and harvest yield were collected. 3.  Controversially, harvesting resulted in larger regional population sizes in harvested metapopulations than in unharvested metapopulations. Similarly, conservative harvesting strategies such as fixed escapement harvesting and harvesting with refuges resulted in smaller population sizes than fixed quota or fixed proportion harvesting strategies. Fixed proportion harvesting gave rise to the largest population sizes, fewest local extinctions and largest yields. Assuming both species are of ecological importance, fixed proportion is therefore the optimal harvesting strategy for this model system. 4.   Synthesis and applications. We demonstrate that, under certain conditions, increasing local mortality can increase population sizes. This ‘hydra effect’ may be caused by the advantage of higher rates of local population turnover for dispersive species in patchy landscapes, or due to the interruption of overcompensatory density‐dependence in the host populations. These results are of particular relevance both in the development of sustainable harvesting policies in multi‐species communities or, conversely, the control of pest species inhabiting spatially structured landscapes. We have shown that predictive models can be a useful tool in the estimation of sustainable harvesting limits where no a priori knowledge of the system exists. We also demonstrate that spatial structure, the effects of interspecific interactions and knowledge of density‐dependent population dynamics must be included in the generation of sustainable harvesting theory.