Sensitivity Analysis and Uncertainty Quantification of Turbine-Height Wind Speeds to Parameters in Yonsei-University Planetary Boundary-Layer Scheme in the Weather Research and Forecasting Model

We conducted Sensitivity Analysis (SA) of simulated turbine-height wind speeds to 15 selected parameters applied in the Yonsei-University (YSU) planetary boundary-layer (PBL) scheme using multiple uncertainty quantification approaches. The study is based on 256 simulations using the Weather Research and Forecasting (WRF) model over the Second Wind Forecast Improvement Project (WFIP2) region with complex terrain. An efficient sampling algorithm is used to explore the multiple-dimensional parameter space and several different SA methods (Sobol, Multivariate Adaptive Regression Splines) are applied in the analysis. Results show consistent parameter sensitivity across different SA methods. We found that 3 out of 15 parameters contribute more than 95% total variance and first order effects dominate compared to interaction effects. All identified sensitive parameters significantly affect the vertical profiles of wind speed and the magnitude of wind shear. The relative contributions of individual parameters are found to be dependent on both the terrain slope and atmospheric stability. The physical mechanisms of how the turbine-height winds are affected by the parameters are also discussed. Our results could provide insights into the design of field measurements to constrain the most sensitive and uncertain parameters in the PBL parameterization.