Abstract Species distribution models (SDMs) are essential for mapping natural resources and supporting spatial planning. Most SDMs focus on total species abundance and overlook intraspecific variation such as size-structure. This can influence inferred species–environment relationships since habitat preferences differ among individuals of different sizes. This study explores how size-specific habitat preferences affect spatial distribution and its environmental drivers, using Atlantic cod (Gadus morhua) in the Baltic Sea as a model species. We examined whether (i) predictor importance varies among size classes, (ii) distribution–environment relationships differ among size classes, and (iii) habitat overlap varies over time; for the latter, we identified the factors driving these changes. Additionally, we assessed how prediction accuracy compares between size-specific and combined SDMs. Using Baltic International Trawl Surveys data from 1993 to 2021, we defined cod size classes based on percentiles of the overall length distribution. For each class, we modeled density (individuals/km²) as a function of selected environmental variables using random forest models. The results revealed consistent differences in predictor importance and density–environment relationships among size classes, as shown by size-specific partial dependence functions. We found no significant temporal changes in habitat overlap among size classes. Although size-specific SDMs showed moderate accuracy loss compared to combined models, they provided valuable insights into size-structured spatial patterns and environmental responses that are not captured by combined models. This framework offers a flexible and informative alternative to conventional SDMs. By accounting for intraspecific variability, it enhances ecological understanding and can inform spatial management strategies by highlighting size-specific patterns of habitat use and overlap. The approach is broadly applicable to other size-structured species and ecological contexts.
Building similarity graph...
Analyzing shared references across papers
Loading...
Szymon Smoliński
National Marine Fisheries Research Institute
Krzysztof Radtke
National Marine Fisheries Research Institute
Max Lindmark
Swedish University of Agricultural Sciences
ICES Journal of Marine Science
Swedish University of Agricultural Sciences
NOAA National Marine Fisheries Service Pacific Islands Fisheries Science Center
National Marine Fisheries Research Institute
Building similarity graph...
Analyzing shared references across papers
Loading...
Smoliński et al. (Fri,) studied this question.
synapsesocial.com/papers/69fd7fcdbfa21ec5bbf085b2 — DOI: https://doi.org/10.1093/icesjms/fsag065