The Use of Visible and IR-Based Probabilistic Overshooting Cloud Top Detection in the Verification of Global Forecasts of Convective Hazards

Observation datasets that identify thunderstorm activity are valuable for verification of global
forecasts of convective hazards. Radar is a reliable and trustworthy observation, but has limited
coverage around the globe. Satellite imagery has far greater global coverage than radar and
convective objects such as overshooting tops can be identified in the images. Overshooting
cloud tops are produced by convective updrafts that penetrate the surrounding cirrus anvil and
rise through the tropopause. Therefore, there is high confidence that convective hazards are
present when overshooting tops are detected. A Probabilistic Overshooting Cloud Top Detection
product has been developed by NASA Langley. The automated algorithm is designed to mimic
the human overshooting top identification process using IR and Visible imagery and NWP data.
Several fields from this product were investigated by NOAA/ESRL/Global Systems Division to
assess their utility in verification, including the Overshooting Top Probability, which identifies the
strongest, and often short-lived, convective cores, the IR Anvil Cloud Detection, which identifies
the surrounding thunderstorm anvils, and the Visible Texture Detection Rating. This
presentation will provide an overview of the Overshooting Cloud Top and IR Anvil Cloud
Detection products, as well as how they were used in verification techniques.