Explicit incorporation of spatial variability in a biomass dynamics assessment model

Abstract

The sustainable management of fisheries has largely relied on stock assessment models that assume stocks are homogeneous throughout their domain. However, ignoring complex underlying spatial patterns can lead to increased risk of failures in management. Utilizing geostatistical approaches in conjunction with a traditional fishery biomass dynamics model, we develop a spatially-explicit modelling framework that treats the underlying population dynamics as spatial processes. Simulation experiments demonstrate that this approach reliably estimates variance parameters and accurately captures true patterns of population change. We further demonstrate the utility of our modelling framework in a real setting using data from the Canadian Maritimes Inshore Scallop Fishery. The model captures time-varying spatial patterns in both population characteristics and fishing pressure without explicit knowledge of the underlying mechanisms and retains the ability to scale up to the whole spatial domain with less associated uncertainty than for temporal models. These results lead to improved scientific advice for management, future-proofing of the assessment to shifts in stock productivity and fishing effort, and provide information that can be used to develop more effective management approaches.