How do NEWA and ERA5 compare for assessing offshore wind resources and wind farm siting conditions?

Abstract

In advance to the construction of a wind farm, a wind resource assessment is performed, often by the use of wind atlas data. Two widely used and publicly available numerical datasets, ERA5 and the recently created NEWA, are compared and evaluated regarding wind climatology, variability and extreme wind speeds using a comprehensive database with high quality offshore measurements. NEWA shows to be more accurate but less precise in terms of predicting mean wind speed and correlation coefficients. However, the higher temporal resolution of NEWA causes a better representation of the wind variability for the offshore sites. Extreme wind speed predictions based on both numerical datasets are underestimated on average. The spectral correction method is applied and found to be appropriate for the correction of the error in predicting the 50-year return wind speed from the numerical datasets for some sites while it fails for others.