Classification and analysis of VMS data in vertical line fisheries: incorporating uncertainty into spatial distributions1
2017
Ducharme-Barth, Nicholas D. | Ahrens, Robert N.M.
Commercial fishing fleets play a critical role in the population dynamics of exploited stocks. Understanding the spatial distribution of fleets allows managers to anticipate how fishing pressure on exploited stocks changes in response to fishing regulations or to large-scale perturbations. By anticipating how fishing pressure changes, managers can develop proactive responses to better protect stocks that are vulnerable to overfishing. Modern fisheries monitoring techniques, including vessel monitoring systems (VMS), have advanced this endeavor. This paper presents a framework for using VMS data to develop spatial distributions of catch, fishing effort, and catch per unit of effort (CPUE) as well as associated estimates of uncertainty in a vertical line fishery. VMS data are classified as fishing using a random forest (RF) model. Uncertainty is calculated using a two-step approach to account for uncertainty arising from the RF modeling process and the classification accuracy of the model. This framework is applied to investigate changes in the Gulf of Mexico reef fish fishery during a period of 6 years, including the 2010 Deepwater Horizon oil spill.
显示更多 [+] 显示较少 [-]