Fernandez, M., Alves, F., Ferreira, R., Fischer, J., Thake, P., Nunes, N., Caldeira, R. & Dinis, A. (2021) Modeling fine-scale cetaceans’ distributions in oceanic islands: Madeira archipelago as a case study.Frontiers in Marine Science, Online early, . DOI:10.3389/fmars.2021.688248 (IF2020 4,912; Q1 Marine & Freshwater Biology)
Species distributional estimates are an essential tool to improve and implement effective conservation and management measures. Nevertheless, obtaining accurate distributional estimates remains a challenge in many cases, especially when looking at the marine environment, mainly due to the species mobility and habitat dynamism. Ecosystems surrounding oceanic islands are highly dynamic and constitute a key actor on pelagic habitats, congregating biodiversity in their vicinity. The main objective of this study was to obtain accurate fine-scale spatio-temporal distributional estimates of cetaceans in oceanic islands, such as the Madeira archipelago, using a long-term opportunistically collected dataset. Ecological Niche Models (ENM) were built using cetacean occurrence data collected on-board commercial whale watching activities and environmental data from 2003 to 2018 for 10 species with a diverse range of habitat associations. Models were built using two different datasets of environmental variables with different temporal and spatial resolutions for comparison purposes. State-of-the-art techniques were used to iterate, build and evaluate the MAXENT models constructed. Models built using the long-term opportunistic dataset successfully described distribution patterns throughout the study area for the species considered. Final models were used to produce spatial grids of species average and standard deviation suitability monthly estimates. Results provide the first fine-scale (both in the temporal and spatial dimension) cetacean distributional estimates for the Madeira archipelago and reveal seasonal/annual distributional patterns, thus providing novel insights on species ecology and quantitative data to implement better dynamic management actions.