Data-driven large-scale spatial planning framework for determining size and location of offshore wind energy development: A case study of China

Offshore wind energy (OWE) is now considered as a vital renewable energy source in numerous coastal countries due to high potential, vast ocean areas, and the scarcity of onshore land resources. Therefore, solving spatial siting decision issues are extremely important for identifying the potential suitable regions and enacting developing strategies for future offshore wind energy facilities. In this study, a data-driven methodological framework was proposed for addressing complex planning problems of offshore wind energy sources by taking into account a set of suitability criteria clusters: resources endowments, oceanographic features, Land–Sea interactions, and potential conflict with current maritime activities. Through linking 6084 samples of existing offshore wind turbines and 12 spatial explaining factors, a high-efficient machine learning regression and classification model, Random Forest algorithm, was conducted to predict ideal size and location of offshore wind energy facilities at the grid cell level. The methodology was applied to the related offshore areas of China with about 18,400 km of coastline. Results highlight that location suitability of OWE farms is greatly determined by water depth and distances from fishing areas, whereas capacity size of wind turbines depended on the synergistic effect of multiple factors. Suitability and size maps were developed to determine the most suitable locations and optimal wind installation capacity in the sea space. Under existing wind power technical conditions, almost 15.6% of marine areas within 50 m water depth (approximately 251 GW) was classified as having above moderate wind energy suitability. The proposed ensemble learning framework provides efficient and useful practical tools for addressing the spatial planning issues of offshore wind farms, and informing policy development from the perspectives of both siting and size.