Sensitivity analysis of observation nudging in WRF model for wind flow simulation over complex terrain

Accurate simulation of wind flow in complex terrain is a challenge for mesoscale numerical simulation methodology. In order to study the WRF setup suitable for the simulation of wind flow over complex terrain to improve the real-time observational nudging capability of the WRF model in simulating complex terrain wind flow, we chose a real complex terrain wind farm in China to study the sensitivity of the WRF configurations, including grid space, nesting method, updating precision terrain, changing terrain categories, and observational nudging time frequency, etc. and compared the simulation results of the wind speed, wind direction time series, and wind speed profiles with the lidar measurement data. The results show that the RSME and MAE using rough grid space with nudging do not improve when updating high-precision topography, there is no significant difference between the observation data nudging results of the one-way and two way nested methods, whereas combining high-precision topography, and observation data nudging method with high grid resolution can make the simulated MAE and RMSE is the lowest and reduced by 28% compared with no using these methods. The method of this paper can be further used for the numerical simulation of wind flow and wind turbine wake in complex terrain with meso-micro sacale coupling.