Within the current scenario of ongoing global environmental change, Marine Ecosystem Models (MEMs) are developed to analyse the past and future dynamics of life in oceans. One of such efforts is EcoOcean, a complex, mechanistic and spatiotemporal explicit MEM of the global oceans based on a trophodynamic core. To predict species distributions and dynamics under contrasting scenarios of climate change and human pressures, EcoOcean requires as inputs the species native ranges, species-specific functional responses for key environmental variables, and time-varying maps of environmental variables delivered by Earth System Models (ESMs). The different sources of uncertainty in these inputs may influence the validity and precision of EcoOcean results. In this study, we evaluate the use of global Bayesian additive regression trees (BART) as a promising new alternative to traditional Species Distribution Models (SDMs) based on classification tree methods to generate these inputs to EcoOcean. To test BART's capability as an SDM on a global scale, we performed a suitability study with the worlds’ 18 penguins as a widely distributed community in the Southern Hemisphere. Our results show that BART is a powerful approach to predict the potential distribution of penguin species, as well as their relationship with key environmental variables, on a global scale. Besides their usefulness as inputs for informing existing MEMs such as EcoOcean, our assessments provide insights on the past, present and future distribution of the whole penguin community. This information is essential for evaluating penguin responses to global change, and for guiding viable management approaches with global conservation policy objectives

11th International Penguin Conference, 4-9 September 2023, Viña del Mar, Chile

Peer reviewed