Estimating the Effects of Fishing on Ecosystem Stability Using Time-Series Methods

Understanding how fisheries affect ecosystems and interactions within them remains an important challenge. Here, we use time-series approaches and output from an ecosystem model of the California Current Large Marine Ecosystem to estimate how fishing changes ecosystem stability.  The Atlantis Ecosystem Model of the California Current is a spatially explicit, multi-sector model that includes oceanographic forcing, a food web model, and fishing.  We modeled fishing on low-trophic level species to understand how removing biomass from the prey base of the food web changes ecological interactions and community stability.  We fit multivariate autoregressive state-space models to 25 year time-series of Atlantis output, incorporating spatial variation in biomass for each time series.  We estimated the interaction matrix between functional groups, and used that matrix to estimate properties of stability such as return time and resilience. We found that stability changes with increased fishing pressure, but these changes are not related to total biomass removed, or trophic level or body size of the target species.  This work demonstrates a new method for detecting the effects of perturbations to ecosystems on the stability of communities. These stability metrics can be used to predict approaching thresholds as a function of fishing and environmental changes.