ABSTRACT

Wind profiling radars are very useful instruments in meteorological forecasting for evaluating the three dimensional wind velocities. The challenging part of this job is to identify the true Doppler peaks of atmospheric origin from a noisy Doppler profile. This paper presents a novel algorithm for the Doppler peak estimation in windprofilers. The normalized Doppler spectrum of each range bin is divided into multiple clusters and low signal to noise ratio (SNR) clusters are discarded. Then, clusters having no continuity along height axis are filtered out and finally the remaining clusters are interpolated to get the wind profile. The algorithm performs very well even when the signal SNR is consistently below −3 dB over successive range bins. The proposed algorithm is found to have improved performance compared to existing algorithms with a reduced computational time.

摘要

风廓线雷达是气象预报中用于评估三维风速的非常有用的工具。这项工作的挑战性部分是从嘈杂的多普勒剖面中识别出真正的大气多普勒峰。本文提出了一种用于风廓线仪中多普勒峰值估计的新算法。将每个距离仓的归一化多普勒频谱划分为多个簇，并丢弃低信噪比 (SNR) 簇。然后，过滤掉沿高度轴不连续的簇，最后对剩余的簇进行插值得到风廓线。即使信号 SNR 在连续范围内始终低于 -3 dB，该算法也能很好地执行。