Simulation Study Of Wind Field Retrieval From Doppler Weather Radar Based On Linear SVR Model

The utilization of Doppler weather radar (DWR), plays a pivotal role in advancing meteorological research and forecasting providing valuable 3D volumetric reflectivity and radial velocity observations of atmospheric systems, facilitating the estimation of local wind fields. This study presents a Support Vector Regression (SVR)-based technique for retrieving 3-D wind fields using Doppler Weather Radar (DWR) radial velocity data. The technique was evaluated through a series of simulated experiments that generated wind fields with varying levels of complexity: low, moderate, and high. The performance of the SVR-based retrieval was assessed using metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the coefficient of determination ($\mathrm{R}^{2}$). The results indicate that the SVR-based method accurately retrieves wind fields, with MAE values within $4 \mathrm{~m} / \mathrm{s}$ up to 7.5 km altitude. This study highlights the effectiveness of SVR in enhancing wind field analysis and its potential for application to real-world radar data.