Hayes, LiamStocks, MattBlakers, Andrew2023-08-160360-5442http://hdl.handle.net/1885/295604Accurate long-term wind speed data is important for understanding the role of offshore wind farms in future energy systems. Meteorological reanalyses, such as ERA5, are relied upon by the wind energy industry and researchers. Being unaffected by onshore topography and surface roughness, the historic generation of offshore wind farms can be accurately predicted using such weather reanalysis. In this work we present a new method for using ERA5 weather data to model long term (>40 year) hourly wind generation for individual offshore wind farms. The model is validated against 57 offshore wind farms in Europe, and reduces the root mean squared error in hourly and daily capacity factor predictions by 10% and 18% respectively when compared to the Renewables Ninja. Further, 40 years (from 1980 to 2019) of ERA5 hourly wind speeds within 200 km of the coast is made easily available for energy system research on our accompanying website (windtlas.xyz).This work was funded in part by the ANU Grand Challenge, Zerocarbon Energy for the Asia-Pacific and the Australian Renewable Energy Agency project, STORES-2. Responsibility for the views, information or advice expressed herein is not accepted by the Australian Government.application/pdfen-AU© 2021 Elsevier LtdERA5Offshore wind farmWind power modellingReanalysisCapacity factorHourly timeseries202110.1016/j.energy.2021.1206032022-07-24