Líffræðifélag Íslands - biologia.is
Líffræðiráðstefnan 2025
Erindi/veggspjald / Talk/poster E129
Höfundar / Authors: William E. Butler (1), Jamie Lentin (2), Jacob M. Kasper (1), Erla Sturludóttir (1), Pamela Woods (1), Bjarki Þ. Elvarsson (1)
Starfsvettvangur / Affiliations: 1. Marine and Freshwater Research Institute, Hafnarfjörður, Iceland, 2. ShuttleThread Ltd, Manchester, UK
Kynnir / Presenter: William Butler
Single-species stock assessment models ignore many biotic and abiotic interactions that prevail in the natural world. If these interactions have a strong influence on a species’ population dynamics, their inclusion in models may help produce unbiased outputs and thus accurate advice. A useful exercise to evaluate a single-species models is to fit the model to the output of a simulated ecosystem. In such cases, a model’s ability to explain historical observations and predict future ones can be quantified and used as a measure of performance. Such exercises can inform how models are constructed and parameterised. In this study, we tested the performance of an age- and length-based model for Atlantic cod (Gadus morhua) in Iceland. The ecosystem model Atlantis served as an operating model, whilst Gadget3 was used for the estimation model. Inputs for the estimation model were generated by (1) two virtual surveys that mimicked real-life Icelandic surveys, and (2) virtual sampling from commercial catches. Uncertainties were estimated using a spatial bootstrap for disparate datasets. In general, the estimation model succeeded in capturing both the levels and temporal trends in total abundance, total biomass, and recruitment. However, small disparities between the operating and estimation model could not be explained by simple, alternative formulations of natural mortality. We discuss the potential for random effects and trophic interactions to improve the estimation model.