Mitigation of wind-turbine clutter using range-Doppler spectral analysis

Wind turbines produce contaminating returns that can bias estimates of the spectral moments and polarimetric variables of weather signals. These biases can propagate to and negatively influence the output of automatic algorithms, such as severe weather detection and quantitative precipitation estimates. Moreover, existing ground clutter filters are ineffective at removing wind turbine clutter (WTC) contamination because the moving components of wind turbines produce clutter signals spanning a wide range of nonzero Doppler velocities. Hence, it is important to devise signal processing techniques that mitigate this special type of contamination so that weather signals can be recovered and used to estimate meaningful meteorological variables. The problem is particularly challenging because WTC signals are inherently non-stationary due to the moving wind turbine blades; thus, traditional time-domain or frequency-domain clutter-mitigation methods are ineffective. In this work, we propose a range-Doppler spectral processing technique to mitigate WTC contamination that is compatible with typical weather radar modes of operation. The proposed technique successfully exploits both spectral and spatial differences between WTC and weather signals. Following an algorithm description, simulations and real data are used to test the effectiveness of this novel WTC mitigation scheme.