An Objective Identification Method on Short-term Heavy Rainfall Based on Radar
Based on 367 echo intensity profile data,43 short-term heavy rainfall events in Liangcheng area from June to August between 2004 and 2021 were statistically analyzed.Six parameters describing the bottom echo intensity and other aspects of the echo profiles were extracted to determine the identification thresholds for short-term heavy rainfall in Liangcheng,by analyzing radar echo profiles.The results showed that the echo intensity distribution of radar echo profiles within 1 km above ground was relatively uniform,with the median as the center,evenly distributed towards both ends.The distributions of Zmax,ET,TOP,VIL,and VILD were relatively uneven,which were all skewed to the right,with VIL and VILD exhibiting severe right skewness.The Z,Zmax,VIL,and VILD showed unimodal characteristics,while ET and TOP showed bimodal characteristics.Based on the violin graphic analysis of radar echo intensity profile characteristics,the 25th percentile was used as the identification threshold for short-term heavy rainfall.The selected thresholds were maximum profile intensity＞26.1 dBz,echo intensity with 1 km above ground＞24.2 dBz,ET＞5.3 km,TOP＞2.0 km,VIL＞0.2 kg·m-2,and VILD＞0.06 g,m-3.These thresholds were generally lower than those in central and eastern China.The identification effects of these thresholds on short-term heavy rainfalls from June to August in 2017 were evaluated.It was found that the hit ratios and false alarm ratios of rainfalls with R≥10.0 mm·h-1 and 4.9 mm·h-1≤R＜10 mm·h-1 were significantly higher than those of rainfalls with R between 1.0 mm·h-1 and 4.9 mm·h-1,while the missed forecast ratios were significantly lower.The hourly TS of short-term heavy rainfall was less than 0.15,and the hourly TS of general precipitation was 0.46.The application of this method had a certain reference value for improving the forecast accuracy of short-term heavy rainfall in Liangcheng area.
Short-term heavy rainfallObjective methodIdentification threshold
万方数据
维普
