Abstract: The China Meteorological Administration’s operational wind profiler radar network fails to distinguish precipitation and turbulence signals in power spectral density (PSD) data under rainfall, causing errors in atmospheric vertical/horizontal wind speed products. To address this, we propose a PSD processing algorithm combining Savitzky-Golay filtering and sliding Z-test for noise estimation, outperforming existing methods in simulations. A detection method using statistical and physical checks identifies bimodal precipitation signals, enabling discrimination between turbulence and fall velocity via Gaussian fitting. Validation with Beijing radar data (2023-07-30) shows the algorithm accurately distinguishes atmospheric motion from precipitation fall velocity under all conditions, aligning with the fifth generation ECMWF reanalysis (ERA5) data. Current operational products misinterpret fall velocity as atmospheric motion, introducing significant errors.