Convective storms produce a wide array of natural hazards, including lightning, flash flooding, turbulence, severe wind, and tornadoes. To minimize risk associated with these hazards, early detection of developing convective storms is critical. A convective and lightning initiation nowcast algorithm is currently being researched at the University of Wisconsin/CIMSS. The algorithm is designed to identify newly developing convective storms before the storms produce lightning and/or have intense radar signatures. Observed 11-µm brightness temperatures from GOES Imagers and SEVIRI are used to compute a box-averaged cooling rate between successive observations. Areas of intense cooling are associated with vertically growing convective clouds. A cloud type (phase) algorithm developed by the GOES-R ABI Algorithm Working Group (AWG) Cloud Team is used to classify cloud top microphysical type. The cloud type information is used to determine where areas of rapidly cooling clouds are transitioning from liquid cloud to supercooled/mixed phase cloud and finally thick ice cloud. The transition to supercooled/mixed phase and thick ice cloud tops are an indication a developing convective storm will produce lightning and/or significant precipitation.

The algorithm has been validated using lightning initiation based upon ground based cloud-to-ground lightning detection networks. The validation work has been conducted using current geostationary imager data (GOES and SEVIRI, 15 minute temporal resolution). The validation results show the algorithm offers up to a 60-minute lead-time before lightning initiation. Probability of detection and false alarm statistics are being compiled and will be presented. Increased temporal resolution of future geostationary imager data (GOES-R ABI and MTG, 5 minute temporal resolution) has been shown to decrease false alarm rate. Further validation work with 5-minute rapid-scan Meteosat-8 SEVIRI data is ongoing and results will also be presented.