A new method for estimating moments from wind profiler spectra using Hidden Markov Model (HMM) is presented. Based on a derived relation between HMM and Kalman Filter, the variance of HMM's transition probability distribution is linked with that of Kalman Filter's noise, making it possible to distinguish the atmospheric signal, ground clutter and radio frequency interference on condition that the expected value of HMM's transition probability distribution is described by the signals' continuity along range gate. A simulation study is used to assess the performance of the algorithm with the result given by human experts to be truth and their correlation coefficient is 0.9792. All parameters needed for moment estimation are taken from spectra, saving costs of modifications with the radar's location, measuring time and weather.
展开▼
