Implementation and Initial Evaluation of a Real-Time Lightning Jump Algorithm for Operational Use

A Lightning Jump Algorithm (LJA), developed and described by Schultz et al. (2009, 2011), identifies rapid increases in the rate of total lightning (intra-cloud plus cloud-to-ground) from multiple ground-based VHF Lightning Mapping networks across the continental United States. These jumps have been shown to be pre-cursors of severe weather in thunderstorms. Real-time implementation and testing of this algorithm began 2 Apr 2012 and continued through Fall 2012. The real-time system uses the WDSSII framework to integrate WSR88D reflectivity data with the RUC model and LMA lightning flashes. Storms are then tracked automatically using the K-Means and SegMotion algorithms (Lakshmanan et al., 2003, 2009). Prior to the experiment, these algorithms were tuned to optimally produce complete and continuous track histories without human intervention. The real-time, automated nature of the experiment assures that the verification statistics of the lightning jump are also inherently tied to the storm identification and tracking. For verification and evaluation of the jump, ground truth data collection via the Severe Hazards Verification Experiment (SHAVE; Ortega et al. 2008) was incorporated throughout the entire real-time data collection phase of the experiment, though Storm Data will be the primary source for verification. One goal of the project is to evaluate how this enhanced verification dataset (with a density 10 to 100 times greater than that from the reports in Storm Data) alters our understanding of the performance of the LJA as an effective warning tool, comparing specific "severe storm periods" with individual reports. For cases in which SHAVE data are unavailable, the LJA statistics may also be compared with radar proxies for severe weather such as the Maximum Expected Size of Hail (MESH). However, the primary objective of this experiment is to evaluate the broader applicability of the LJA as a severe weather warning decision assistance tool.