Experimental, Satellite, Microphysically-Based, Early Alerts of Severe Convective Storms, Part 1: Scientific basis, methodology and its experimental application

A new method for identification of the potential of growing convective clouds to become severe storms using their satellite retrieved microstructure was developed and tested during spring 2008 at the Storm Prediction Center in Norman, Oklahoma. The new method makes use of GOES multispectral satellite imagery for the satellite inference of developing severe convective storms. It is based on the observations that the cloud-particle effective radius (Re) in rapidly ascending cloud parcels remains small at great heights because there has not been enough time during ascent for glaciation and for the growth of large particles. The relation between Re and height can be inferred by relating the cloud-top retrieved Re to the cloud top temperature T for a field of convective clouds. The T-Re plots indicate a threat of severe storms when the inferred glaciation temperature is near the homogeneous ice nucleating temperature of -38°C, the Re at this temperature is relatively small, and the profile of T-Re is steep, indicating a slow increase in Re with height (i.e., with decreasing temperature).

The methodology requires solar illumination of the growing cloud elements for retrieving Re. Therefore, it is usable only during daylight hours. The method works best for clouds with exposed feeders. Therefore, developing convective clouds best show their severe weather potential while they are still in their early developing stages. This is typically 1 to 2 hours before the occurrence of the tornado or large hail.

The T-Re relations are analyzed by an automatic detection algorithm for the detection of “severe storm signature”. The suspected clouds are identified and flagged as “Early Alerts” (EA). ). The objective is to predict when and where a severe storm is most likely to occur one to two hours prior to the actual event. As such, the EA serves as an intermediate stage filling the large gap between severe weather watch (event up to hours away), and a warning when the event is -often about to occur within 10-20 minutes

After testing the method on historical data, the potential for its application was sufficiently encouraging to warrant the development of an operational algorithm that was implemented and tested at the Storm Prediction Center. In this presentation (Part 1) we will describe the methodology, the underlying physics and its application. The evaluation and skill of the methodology will be addressed in a companion presentation (Part 2).