Gaussian Mixture Method for Kdp estimation
2019
Estimating Kdp is a quite difficult task in radar meteorology, being estimation based on computin a derivative of a range profile affected by noise and various measurement errors.Different methods exist, but are difficult to compare each other, since all comes with advantages and disadvantages.The manuscript by Wen et al. describes a novel approach, called Gaussian Mixture Method (GMM) is proposed and compare its performance with the classical method based on linear regression (LR).Results in terms of QPE computed with respect to 4 raingauges at different distance from the radar shows a slightly better (but not for all the raingauges, only for the two farther from the radar) performance of the new method.It is not easy to provide a convincing validation of Kdp estimation, but some suggestions to improve it are provided below (note than pages 2-14 describes the method, but only pages 14-16 are related to validation).