Complementing single-species assessment models with age- and time-varying natural mortality: Insights from holistic ecosystem models

Abstract

Integrating information on ecosystem dynamics into single-species stock assessment has been recommended for fisheries management to head towards Ecosystem-Approach to Fisheries Management (EAFM). This paper proposes a sensitivity analysis of single-species assessments in relation to the implementation of time-varying natural mortality (M). Time-varying estimates from a holistic ecosystem model were implemented in exploratory Celtic Sea gadoid stock assessments. Two species were selected to disclose different temporal patterns in M: whiting (Merlangius merlangus) displaying a trend and haddock (Melanogrammus aeglefinus) exhibiting high variability with no trend. Implementing overall higher M levels increases Spawning Stock Biomass (SSB) and decreases average fishing mortality at age (Fbar) in both cases. Here, time-varying values induce reference points modification. Introducing a trend in M improves the assessment quality for whiting, while shifting the stock status from fully exploited to overexploited for several years. Integrating high M variability for haddock induces minimal changes in assessment quality and stock status. In both cases, short term projected sustainable catch level was reduced by 14%. Despite uncertainty around estimates, results suggest that integrating M accounting for a varying predation is crucial, especially when M disclose a trend. Future research should focus on comparing M estimates across various sources, including ecosystem models, and benchmarking the single-species models’ sensitivity to time-varying values.