Implications of spatial population dynamics for abundance estimation and catch apportionment of Alaska sablefish (Anoplopoma fimbria)

Abstract

Spatial dynamics in fish populations are challenging to identify and incorporate into population dynamics models due to the large amount of data required and the uncertainty associated with spatial processes. However, ignoring spatial dynamics can result in biased population estimates and negatively affect population sustainability or the ability to meet fishery management goals. My research explored implications of spatial dynamics for the assessment and management of Alaska sablefish (Anoplopoma fimbria). Sablefish are highly mobile, targeted in commercial and recreational fisheries, and documented to have spatial variation in population dynamics such as abundance and biomass. However, gaps remain in understanding the drivers of these spatial dynamics and their effects on fishery management. Chapter 1 focused on sablefish population dynamics in Alaska by developing a spatial stock assessment model with movement between modeled areas. Chapters 2 and 3 detail the development (Chapter 2) and application of a simulation and estimation model framework (Chapter 3) to examine alternative management scenarios and inform fisheries management. The spatial stock assessment models developed in Chapter 1 showed that a combination of population processes (movement, recruitment) and fishing have resulted in spatial differences in sablefish spawning biomass and fishing mortality, and some management areas are likely below target biomass levels despite the population as a whole being at or near biomass targets. The implications of this apparent localized depletion are unknown, but could negatively impact recruitment success if depleted areas are crucial to population sustainability. Chapter 1 also highlights the complexity of spatial stock assessment models, which require large quantities of data and pose challenges for parameter estimation. In Chapters 2 and 3, sablefish spatial population processes were simulated based on parameters estimated in the spatial stock assessment developed for Chapter 1 then assessed using a single-area, panmictic stock assessment model. Chapter 2 focuses on the simulation framework development and tests a suite of alternative population assumptions regarding movement and recruitment, which were found to be very influential in preliminary analyses. The analyses also revealed the challenges associated with a simulation framework that relies on a relatively complex stock assessment model and cannot rely on human scientists to monitor and tweak assessment model estimation processes during the automated simulations. In Chapter 3, five alternative methods for apportioning harvest opportunity among management areas were examined through the simulation to understand whether there were unexpected consequences of how catch opportunities were apportioned across space, given the assumed underlying population dynamics. We found that recruitment was poorly estimated by the single-area stock assessment model, and recruitment and movement rates between areas were drivers of population dynamics in each modeled region. We also found that no method for apportioning sablefish resulted in long-term population declines under the simulation and management assumptions, but recommended an apportionment method that used data from an annual survey so that catch opportunities are relational to observed spatial biomass. Each apportionment method resulted in different levels of catch for management areas and different year-to-year stability in catch, and these differences may be meaningful to stakeholders. This research highlights the challenges and benefits of incorporating spatial population dynamics in assessment of a relatively data-rich species with high movement rates. For species with known processes that could create spatial dynamics but insufficient data for development of spatial stock assessment models, the development and implementation of monitoring and research plans to close data gaps could be a long-term strategy for improving understanding of the species' dynamics. However, monitoring can be logistically challenging and economically prohibitive in some situations, so simulation analyses such as those conducted here may help understand which data are most important to collect or inform the development of management procedures which are robust to the most plausible hypothetical population dynamics or future environmental conditions.