Characterizing the coastal benthic ecology of two regions of Inuit Nunangat using machine learning

Abstract

Understanding the benthic ecology in Canada's Northern coastal areas allows for informed decision-making in their use as the Arctic undergoes rapid changes. Two regions of Inuit Nunangat (Inuit Homelands) were studied: Nunavut and Nunatsiavut. Multibeam sonar provided continuous-coverage geomorphology data, while video ground-truthing enabled us to validate sediment characteristics and benthic community distribution in the estuaries studied. Random Forest Modelling (RF) and Species Distribution Model (SDM) were employed to better understand and visualize environment-benthos relationships across the study areas. Key findings reveal that positioning within the fjord, terrain curvature, and estuary orientation to oceanographic forces are primary factors influencing benthic distribution across all study sites. The findings support the recent negotiations for the establishment of an Inuit Protected Area in Nunatsiavut and provide valuable ecological data for future marine management plans in the Canadian Arctic.