An Automated System for Detecting Large-scale Convective Storms: Application to Model Evaluation

Enroute air traffic can experience major delays if large regions of the air-space are unexpectedly closed due to convective storms. Large-convective storms such as Mesoscale Convective Systems can form without warning and persist longer than expected causing reroutes that can lead to massive delays to the National Airspace (NAS). Owing to forecast errors and the inherent limits of predictability that characterize convective storms, products are needed that portray the level of forecast uncertainty. And while probabilistic thunderstorm forecast products exist (e.g., LAMP, SREF), none specifically assess the likelihood of convection achieving specific attributes (such as duration, size) that are most devastating to air traffic flow across the NAS (as well as possibly producing extended duration impacts at individual terminals). In this study, the steps taken in developing a prototype system that provides forecast guidance that indicates the likelihood of large-scale storms are given. The system post-processes model ensemble data to predict the likelihood of continuous areas of intense precipitation (indicated by Vertically Integrated Liquid Water or VIL exceeding 3.5 kg m-2) with a maximum dimension exceeding 100 km, with gaps of less than 10 km, and a duration of at least 1 hr. Sensitivity to these criteria, and more importantly, the type (e.g., time-lagged vs multi-model) of ensemble data used are explored. A key aspect of this study is the assessment of the performance of each ensemble member as a function of time of day, region, leadtime and synoptic weather regime in order to more intelligently compute the large-scale storm likelihood field.

This research is supported by the FAA. The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.