Improving wind power modelling through granular spatial and temporal bias correction of reanalysis data

There is a need for efficient methods to simulate wind power output to assist the expected rapid uptake of new wind farms. Reanalysis products provide our best estimate for the previous state of the atmosphere and are popular due to their global coverage and convenience. However, these models are known to misestimate wind power output by up to ±50% due to significant spatial biases. Previous work applied bias correction methods to improve power simulations. However, there has been no assessment of the spatial and temporal resolution that these bias correction factors should be derived at for the best accuracy. In this paper, we investigate the impact of the spatial and temporal resolution by grouping turbines into a varying number of clusters and varying the frequency at which correction factors are calculated across a year. The correction factors are used to simulate the power output of 4,834 turbines across Denmark resulting in monthly capacity factors. The best correction scenario decreased the error of simulated outputs by 43%. Increasing the spatial resolution of bias correction reduced the error by up to 11%. Correction factors with a bimonthly (every two months) frequency decreased the error by 3% from the time-independent correction factors.