Rapid Scale Wind Profiling With Autoregressive Modeling and L-Band Doppler Radar

Radar wind profilers (RWPs) are well-established instruments for the probing of the atmospheric boundary layer, with the immense advantage of long-range and all-weather operation capability. One of their main limitations, however, is a relatively long integration time compared with other instruments, such as lidars. In the context of L-band RWP, we show that the use of autoregressive (AR) modeling for the antenna signals combined with the maximum entropy method (MEM) allows for a correct estimation of radial wind velocity profiles even with very short time samples. A systematical analysis of performance is made with the help of synthetic data. These numerical results are further confirmed by an experimental dataset acquired near the landing runways of Paris Charles de Gaulle (CDG) Airport, France, and validated using a colocated optical lidar at the Aerological Station of Payerne, Payerne, Switzerland. It is found that the AR-MEM approach can successfully derive wind estimates using integration times as short as 2.5 s where the classical spectral approach can barely provide any measurement.