An Enhanced Storm Warning and Nowcasting Model in Pre-Convection Environments

A storm tracking and nowcasting model was developed for the contiguous US (CONUS) by combining observations from advanced baseline imager (ABI) numerical weather prediction (NWP) short-range forecast data, along with precipitation rate CMORPH (the Climate Prediction Center morphing technique). random forest based adopted using maximum as benchmark convection intensity, location time of storms optimized optical flow (OF) continuous tracking. Comparative evaluations showed that models had higher accuracy severe areas equal to or larger than 5000 km2 over smaller samples, loweraccuracy cases 1000 km2, while sample-balancing applied possibilities detection (PODs). typical convective event August 2019 presented illustrate application on local (LSS) identification warnings in pre-convection stage; successfully provided a lead 1–2 h before heavy rainfall. Importance score analysis overall impact ABI much NWP, brightness temperature difference between 6.2 10.3 microns ranking at top terms feature importance.