Mapping Seafloor Sediment Distributions Using Public Geospatial Data and Machine Learning to Support Regional Offshore Renewable Energy Development

Abstract

Given the rapid expansion of offshore wind development in the United States (US), the accurate mapping of benthic habitats, specifically surficial sediments, is essential for mitigating potential impacts on these valuable ecosystems. However, offshore wind development has outpaced results from environmental monitoring efforts, compelling stakeholders to rely on a limited set of public geospatial data for conducting impact assessments. The present study therefore sought to develop and evaluate a systematic workflow for generating regional-scale sediment maps using public geospatial data that may pose integration and modeling challenges. To demonstrate this approach, sediment distributions were characterized on the northeastern US continental shelf where offshore wind development has occurred since 2016. Publicly available sediment and bathymetric data in the region were processed using national classification standards and spatial tools, respectively, and integrated using a machine learning algorithm to predict sediment occurrence. Overall, this approach and the generated sediment composite effectively predicted sediment distributions in coastal areas but underperformed in offshore areas where data were either scarce or of poor quality. Despite these shortcomings, this study builds on benthic habitat mapping efforts and highlights the need for regional collaboration to standardize seafloor data collection and sharing activities for supporting offshore wind energy decisions.