Global flash clustering and infrared radiance characteristics: Analysis of TRMM/LIS data

Abstract

Lightning activities are usually related to climate change and meteorological disasters. We investigated the spatiotemporal distribution of global lightning activities and infrared radiance characteristics of lightning flashes. Based on the Lightning Imaging Sensor (LIS) data product from 1998 to 2014, the spatiotemporal distributions of flash duration, length, footprint, and radiance were analyzed. K-means was used to perform cluster analysis on global lightning data. Finally, based on LIS event data, the lightning radiance in a single detection was processed statistically, and the probability density of the flash radiance was fitted. The results indicated that the flash density had significant spatiotemporal distribution characteristics, which were primarily affected by the climates of different geographical locations and seasons. According to the K-means clustering algorithm, lightning was geographically divided into six categories. The flash density in the area with the strongest lightning activities was 62.9280 fl·km−2·a-1 and the mean annual radiance was 1.1371 J·m−2·sr-1·μm−1·a-1, primarily distributed in North America, southern South America, central and southern Africa, southern Asia, and northwest and southeast Australia. The strength of lightning activities was mainly affected by the climate. Offshore waters, low latitude areas, and the barrier of mountain ranges cause strong convective activity in the atmosphere, which leads to strong lightning activity. Additionally, human activities, such as high-temperature exhaust gas and aerosols emitted by factories, will also affect atmospheric convection activities and cloud formation, consequently affecting lightning activities. The probability density functions of lightning flash radiance obtained by single detection had almost the same form in the radiance range of 0–0.6 J·m−2·sr-1·μm−1 regardless of the geographic location. This may be related to the inherent discharge properties of lightning at low radiance. When the radiance was greater than 0.6 J·m−2·sr-1·μm−1, the difference between the probability density functions of the radiance of the six categories was subtle. The lightning clustering result and the probability density function radiance obtained by single detection are helpful for the identification and prediction of lightning flash radiation characteristics.