Evaluating the risks of red tide mortality misspecification when modeling stock dynamics

Abstract

Extensions of single-species stock assessments to include ecosystem considerations are one step towards achieving an ecosystem approach to fisheries management. While red tide mortality has been estimated within base assessment models for both gag grouper and red grouper in the Gulf of Mexico since the late 2000 s, considerable uncertainty remains regarding how best to incorporate this source of mortality in stock assessment and the potential risks of misspecifying its timing and its effect across age-classes. We conduct simulation testing of an approach to estimating red tide mortality, which treats red tide as a bycatch fleet where all encountered fish are discarded dead. Accounting for episodic natural mortality events by correctly specifying the years in which these events occurred was the most unbiased and precise approach for estimating population quantities, although the estimated red tide mortalities were highly variable in magnitude. While not estimating red tide mortality provided reasonably unbiased estimates of terminal biomass and fishing mortality rate, which are used to determine stock status, caution must be exercised when red tide events occur in the terminal year(s), as ignoring episodic mortality can lead to these events being misinterpreted in model fitting and inflating estimates of fishing mortality (and potentially overfishing status). Our simulations highlight the importance of continued data collection on how red tides affect marine resources, and continued community and stakeholder engagements to gain an understanding of the historical timing and impact of red tide events.