UW Aquatic & Fishery Sciences Quantitative Seminar

Felipe Hurtado Ferro

School of Aquatic and Fishery Sciences

Are spatially-explicit assessment models better, even in the absence of tagging data?

Abstract

The importance of spatial structure for fish population modeling and fisheries stock assessment has been recognized since the days of Beverton and Holt's (1957) seminal work. However, its implementation into stock assessments has been slow, and we still have limited understanding of how including spatial structure in our models may affect their results. Most stock assessments assume populations are homogeneously distributed across their range, an assumption that in most cases will be violated. Assessments usually represent spatial structure through a 'fleets-as-areas' approach, i.e. using multiple fleets, each with different selectivity patterns, as a proxy for spatial availability. Spatially-explicit models are less commonly used, because these usually require independent data to estimate movement rates (usually some form of tagging). However, recent studies have shown that movement rates can be estimated from differences in length composition across areas. This presentation focuses on two main questions. First, how do complex spatial dynamics affect spatially-aggregated models; and second, how different stock assessment model configurations (both spatially-aggregated and spatially-explicit) perform when confronted with a spatially-structured stock with localized recruitment, and adult annual migratory behavior.