Correlation Between Total Lightning Activity and Precipitation Particle Characteristics Observed from 34 Thunderstorms

A total of 34 thunderstorms around Shanghai and Wuhan of China are analyzed in order to determine the relationship between total lightning activity and precipitation particle characteristics. Precipitation particle concentration data are obtained from the 2A12 product of TRMM/TMI (Tropical Rainfall Measuring Mission/TRMM Microwave Image) and lightning activity data are from the TRMM/LIS (Lightning Imaging Sensor) and SAFIR3000 (Surveillance et Alerte Founder par Interferometric Radioelectirque). On a spatial scale of 0.1 degrees x0.1 degrees, a weak spatial relationship is found between total lightning and the vertically integrated content (VIC) of precipitation particles (cloud water, precipitation water, cloud ice, and precipitation ice). A strong power relationship is identified between the lightning density (D-30; fl km(-2) min(-1)), relative to a rainfall intensity threshold value of 30 mm h(-1), and the maximum rainfall intensity (R-max; mm h(-1)); the obtained regression equation is R-max = 23.10D(30)(-0.18) + 11, with a correlation coefficient of 0.841. Lightning frequency shows a significant linear correlation with the contents and covering areas of precipitation particles (in which the VICs exceed threshold values). Furthermore, ice particles above the -10 degrees C level exhibit a stronger correlation with lightning activity than those above the 0 degrees C level or the integrated ice particles at all levels. The results demonstrate that the particles responsible for the most significant charging process and lightning activity are restricted by the threshold value of VIC among the particles, which reflects the demand of the charging process on dynamic characteristics. The obtained fitting equations can provide useful reference for assimilating lightning information into numerical prediction models so as to improve the reliability of forecast results. The particle products from the prediction models are also helpful in estimating the occurrence of lightning activity within 2-6-h periods.