Supervisor: Senior researcher Mark C. Kelly, DTU Wind Energy
Titel: Conditional Error Statistics for Offshore Winds, from WRF Modelling and ASCAT Satellite Data
“Following the findings of Hahmann et al. (2015) we observe a bias in WRF simulations when compared to anemometer or Lidar measurements in the North Sea. This bias is expressed as the difference between average WRF model output and average measurements. This thesis’s objective is to define and analyze WRF error metrics against ASCAT satellite data over the North sea and at 10 m height.
Initially we’ll look at how the data is handled and converted through interpolation to enable the comparison. This will allow for the computation of the statistics that will help us determine how well the WRF model represents the measurements.
The statistical analysis begins with the analysis of marginal pdf for the variables of wind speed, wind direction and inverse Obukhov length for the WRF data. The subsequent conditional statistics will be performed conditioning over these variables.
Two different error metrics are defined, with the raw error representing the correspondence between individual simulations and measurements while the time averaged error represents the correspondence between average simulations and measurements.
Initially the presence and effects of phase shift for WRF simulations over the error metrics is investigated. The conditional analysis continues by conditioning the error metrics over WRF wind speed, wind direction and inverse Obukhov length.
A correction for the ASCAT wind speed is presented and new conditional statistics are obtained with the comparison against the corrected data.”