Automated and Objective Thunderstorm Identification and Tracking Using Operational Geostationary Lightning Mapper (GLM) Data

Schultz et al. (2016) designed a new method for objectively identifying and tracking thunderstorms derived from both radar and lightning data. This product, aptly named VILFRD, uses a combination of vertically integrated liquid (VIL) and flash rate density (FRD) from the Geostationary Lightning Mapper (GLM) to identify storm features. The product is tracked using k-means clustering within w2segmotionII in WDSS-II, resulting in a fully automated thunderstorm tracking method. Prior to Geostationary Operational Environmental Satellite (GOES) East’s launch in 2016, the VILFRD tracking algorithm utilized GLM proxy flashes derived from lightning mapping arrays to produce a best guess dataset at what GLM data would look like once operational. Now, with GOES GLM data readily available, work has begun to transition the feature identification (ID) and tracking algorithm to the GLM data stream.

This study introduces modifications made to the current VILFRD technique with the goal to efficiently incorporate GLM data into the feature ID and tracking algorithm. Ensuing impacts to lightning based algorithms (e.g., lightning jump, lightning safety) are explored to understand how the method of feature ID and tracking can potentially influence algorithm performance. Individual examples of thunderstorms and large sample analyses over the CONUS domain will be presented, assessing the performance of the newly modified VILFRD techniques. Modifications include the incorporation of group rate density (GRD) in place of FRD, changes to w2segmotionll parameters within WDSS-II, and using a lightning-only tracking method to serve areas where radar data is not available. Thunderstorm tracking using GLM alone highlights potential for storms to be tracked anywhere within the GOES East and GOES West fields of view.