Quantitative Clairvoyance or Deductive Delusion? Do Multi-Species Models Improve Fisheries Management Advice?

Multi-species stock-assessment models (MSCAAs) can quantify the direct and indirect effects of climate variability and fisheries harvest on species populations. MSCAAs are designed to address long recognized limitations of prevailing single species management, notably non-stationarity in mortality and maximum sustainable yield (MSY) and as such may help increase forecast accuracy. That said, interpreting emergent MSCAA biological reference points, especially in terms of long-term projections for climate change analyses, is challenging. Here we discuss retrospective and projection results of CEATTLE, a MSCAA developed for three groundfish species from the Alaskan Bering Sea (USA). For two of the three species, the multi-species model achieved a slightly higher over-all fit to the data than the single species model. Inclusion of temperature in projections did have a strong effect on BRPs values, although relative influence of climate and trophic effects depended on recruitment assumptions. In general, inclusion of trophic interactions amplified climate-driven declines or increases in fish abundances. Our results suggest that fisheries models that do not include trophic interactions or climate effects might over-estimate sustainable harvest rates for fish species negatively impacted by climate change and under-estimate harvest rates for predator species positively impacted by climate change.