Future projections of offshore wind energy resources in China using CMIP6 simulations and a deep learning-based downscaling method.

Good knowledge of future wind energy resources is crucial for sitting and planning studies of wind farms. The simulation results from the Coupled Model Intercomparison Project Phase 6 (CMIP6) and a proposed new downscaling method based on the bidirectional gated recurrent unit (BiGRU) are both used in this paper to study future offshore wind energy resources in China. The proposed new downscaling method is validated and compared to two traditional methods. It is found that the spatial patterns of downscaled wind speed are highly consistent with the reference data, and biases are significantly reduced by the new method, especially in coastal and shallow water areas. Using the new method, we downscale the CMIP6 future projected simulation results and generate a new dataset of offshore wind speeds in China for the period of 2021–2100 with a resolution of 0.25°. Multi-model ensemble (MME) results project small decreases in the offshore wind speed and wind power density over the East China Sea and increases in those parameters over the South China Sea, for the middle and end of the 21st Century (2041–2060 and 2081–2100) under two representative scenarios (SSP2-4.5 and SSP5-8.5). • The first work on projection of future offshore wind energy resources in China. • A new downscaling method based on deep learning network was developed. • Global climate models' simulations were downscaled to high-resolution grids of 0.25°. • Small changes in future offshore wind speed and power density were projected in China. [ABSTRACT FROM AUTHOR]