基于天气雷达的长江三峡区间降雨定量估测方法

Radar-based quantitative precipitation estimate in the Three Gorges region of Yangtze River

天气雷达能够提供高时空分辨率的降雨观测,但其观测误差来源较多,影响到降雨估测精度。本研究针对长江三峡区间开发了基于单部地面天气雷达的定量降雨估测方法,包括波束遮挡分析、地物杂波抑制、降雨类型自动识别及多种Z-R关系转换等关键环节,并结合地面雨量观测对雷达估测结果进行了检验。结果显示,雷达的降雨估测精度随探测距离增加呈现明显的衰减趋势,较好估测降雨的范围是100公里以内。通过比较不同算法的结果,发现波束遮挡是万县雷达估测误差的主要来源之一。地物杂波是另一类主要误差来源,导致50公里范围内雷达探测降雨的误报率较高。通过识别降雨类型,针对不同降雨类型采用不同Z-R转换关系,可以降低降雨估测的偏差。2010年夏季降雨的观测结果表明,该算法对典型层状(对流)降雨会有低估(高估),总体上对逐时降雨量的估测略微偏低。

Weather radar can detect spatiotemporal variability of precipitation in high resolution, but it suffers several sources of error. In this study, over the Three Gorges region （TGR） in Yangtze River, we have developed a first version of single radar based quantitative precipitation estimation （QPE） algorithms including various critical procedures, such as beam blockage analysis, ground clutter filter, rain type identification, and multiple Z-R relations. Current radar based QPE shows obvious range dependent performance, which suggests it could only reliably retrieve rainfall within the range of 100 km. Further, comparison among three datasets processed with different algorithms reveals that beam blockage is one of major error sources; within the range of 50km, affected by ground clutters, false alarm is another typical error. Multiple Z-R relations have effectively reduced bias at various ranges. The overall performance indicates that radar QPE generally underestimates （overestimates） typical stratiform （convective） storm rainfall events, and as a whole, it tends to slightly underestimate hourly rainfall during the 2010 summer.