Modelling climatically and ecologically suitable environment for benthic foraminiferal species

Abstract

Benthic foraminifera are unicellular eukaryotic organisms that play significant roles in marine food webs. Monitoring changes in their distributions and abundances are crucial to understand their roles and functions and to predict the effects of environmental impacts and climatic changes. Numerous biotic and abiotic factors influence foraminiferal distribution and abundance, but the ecology of individual taxa (i.e., autoecology) is difficult to define and this commonly results in a generalization of functions in their environment. In this study, we explore methods for inferring foraminiferal species ecology and niche requirements using Species Distribution Models (SDM). We modelled the distributions of four benthic species (i.e., Peneroplis planatus, P. pertusus, P. arietinus and Coscinospira hemprichii) that host algal endosymbionts against a large collection of predictors (ecologically meaningful environmental variables, EMEV) variables in the Arabian Gulf. To identify combinations of effective predictor variables, we compiled several models and then selected three with different variable set combinations to compare their predictive performance. Mean iron concentration, diffusion attenuation, and dissolved oxygen were identified as the most important variables determining the distribution of these species. Our model successfully predicts their current habitat suitability in the Arabian Gulf (AUROC = 92%). The model also identified areas along the western coastline of the Gulf as highly suitable habitats for the four species (Habitat Suitability Index > 0.8). Then, the model was spatially extended over the world’s oceans and marginal seas, reliably identifying areas with known distributions of the four species (AUROC = 89%). Here we demonstrate how a SDM model can be a useful tool in capturing complex habitat features for benthic organisms. Such models, when correctly applied using several remotely sensed environmental and geological variables, can be a very useful ecological tool for spatial and temporal predictions even in the current context of global climate change.