Accuracy analysis of wind retrieval from Korean radar network: Simulation study

Wind retrieval from Multiple Doppler radar is an excellent tool to understand dynamics of precipitation systems. Variational multiple-Doppler radar synthesis is commonly used. Uncertainty in this method is due to the Laplacian smoothing term and the balance of each cost-function. In addition, the spectral width broadening effect can overcome this uncertainty when significant turbulence and variation of winds exist within a radar sampling volume. In this study, we try to quantify the relative importance of spectral width broadening and methodology-induced error sources. Three-dimensional wind fields and microphysical variables from Weather Research and Forecasting (WRF) model data are used to simulate reflectivity, wind fields, and radial velocity data. The three-dimensional wind fields are then retrieved by a variational multiple-Doppler radar synthesis method proposed by Liou and Chang (2009). The comparison between retrieved wind and WRF model wind indicates that the area with large vorticity (> 0.007 s^-1) and vertical shear (> 0.07 s^-1) shows large root mean squared error (rmse) value (> 15 ms^-1), whereas the area with small vorticity (< 0.00025 s^-1) and vertical shear (< 0.0025 s^-1) shows lower rmse value (< 2 ms^-1). On the other hand, the comparison between retrieved radial velocity and simulated radial velocity shows small rmse value (< 2 ms^-1). These results indicate that uncertainty from spectral width broadening effect is more significant than uncertainty from variational method.

Acknowledgment This research was supported by a grant (14AWMP-B079364-01) from Water Management Research Program funded by Ministry of Land, Infrastructure and Transport of Korean government.