Time-variant productivity in biomass dynamic models on seasonal and long-term scales

• Published in: ICES Journal of Marine Science
• Main Authors: Mildenberger, Tobias K, Berg, Casper W, Pedersen, Martin W, Kokkalis, Alexandros, Nielsen, J Rasmus
• Other Authors: Zhou, Shijie, EMFF, European Maritime and Fisheries Fund, Danish Fisheries Agency
• Format: Article in Journal/Newspaper
• Language: English
• Published: Oxford University Press (OUP) 2019
• Subjects: Ecology; Aquatic Science; Ecology, Evolution, Behavior and Systematics; Oceanography; Gadus morhua
• Online Access: http://dx.doi.org/10.1093/icesjms/fsz154; http://academic.oup.com/icesjms/advance-article-pdf/doi/10.1093/icesjms/fsz154/30040850/fsz154.pdf

Abstract The productivity of fish populations varies naturally over time, dependent on integrated effects of abundance, ecological factors, and environmental conditions. These changes can be expressed as gradual or abrupt shifts in productivity as well as fluctuations on any time scale from seasonal oscillations to long-term changes. This study considers three extensions to biomass dynamic models that accommodate time-variant productivity in fish populations. Simulation results reveal that neglecting seasonal changes in productivity can bias derived stock sustainability reference levels and, thus, fisheries management advice. Results highlight the importance of biannual biomass indices and their timing relative to the peaks of the seasonal processes (i.e. recruitment, growth, mortality) for the estimation of seasonally time-variant productivity. The application to real-world data of the eastern Baltic cod (Gadus morhua) stock shows that the model is able to disentangle differences in seasonal fishing mortality as well as seasonal and long-term changes in productivity. The combined model with long-term and seasonally varying productivity performs significantly better than models that neglect time-variant productivity. The model extensions proposed here allow to account for time-variant productivity of fish populations leading to increased reliability of derived reference levels.