Occupancy and species interactions in biodiversity assessment of Antarctic plankton: the missing link (Manuscript submitted Ecological Processes Journal)

Plankton, as an essential ecological category, occupies the lower levels of aquatic trophic networks, playing a pivotal role in energy transfers and serving as a reliable environmental indicator for altered ecosystems.  There is a critical need for studies that consider the potential biological interactions among planktonic taxa, as they can possibly capture and quantify cascading changes in species. In this pioneering study, we addressed this gap by evaluating possible correlations through a statistical analysis of co-occurrence among different Antarctic marine phytoplankton, meso- and microzooplankton taxa. Information from over 50 environmental descriptors obtained from the Polar Macroscopic Layers, stored in the Quantartica package (open source GIS platform QGIS) was incorporated and the distribution of the selected co-occurring taxa i.e. six copepod species and six classes of phytoplanktonic groups, was modelled. The predictive maps were obtained by applying artificial intelligence techniques. Specifically, we used four different types of machine learning algorithms (TreeNet, RandomForest, CART and Ensemble) known within the Species Distribution Modelling framework, applied to presence/absence data of planktonic taxa occurring within the depth range of 0-200 m.The sampling stations were located in the Eastern portion of the Ross Sea and were sampled in the framework of the “Plankton biodiversity and functioning of the ROSS SEA ecosystems in a changing Southern Ocean” (P-ROSE) project of the Italian National Antarctic Program (PNRA), that took place in 2016-2017. The obtained models quantitatively predicted the presence of each species of copepod and phytoplankton class in terms of Relative Occurrence Index (RIO) in 3-dimensions maps for the Ross Sea. This information holds particular relevance in light of ongoing climate change  especially on  such a large scale, enabling quantitative baseline assessments, crucial in enhancing and refining long-term forecasts. Therefore our models contribute to the provision of essential ecological information through a measurable process, in one of the most pristine ecosystems globally. We argue the crucial role of the Ross Sea Region Marine Protected Area marine protected area in achieving these results.