The Warning Decision Support System – Integrated Information (WDSS-II, Lakshmanan et al. 2007) software allows users to simultaneously view and manipulate multiple data sources. WDSS-II contains a suite of algorithms that combine Near-Storm Environment (NSE) information from the Rapid Update Cycle (RUC) model with single or multiple WSR-88D radar data to compute any number of discernible parameters (e.g., reflectivity at specific isotherm levels). These derived parameters then can be superimposed onto CG lightning information from the NLDN and/or total lightning data from LDAR/LMA networks to examine lightning patterns within individual storm cells. This paper will describe our recent algorithm modifications and illustrate the benefits of automating the development of datasets used to investigate lightning producing storms.

WDSS-II contains an algorithm that uses Hierarchical K-Means clustering to identify and track mesoscale features as small as 10 km². The original algorithm identifies regions of composite reflectivity greater than 30 dBz and determines the type of storm (i.e., isolated supercell, line, pulse, or non-severe). We modified the algorithm to cluster several additional gridded fields based on user defined thresholds and to compute higher order parameters along the storm's path. We also revised the original storm type identification algorithm to include lightning parameters, created a new algorithm to track clusters of LDAR densities providing 1-min temporal resolution, and used the algorithm's advection feature to produce short-term forecasts of total lightning. Our paper will briefly describe each of these modifications to provide background for our newest set of algorithms (described below) that are being used to create a robust dataset that allows for a statistical investigation of lightning patterns within severe and non-severe storms.

Total lightning networks are regional in coverage, and few NWS WFOs have access to their data in real-time. Therefore, we are developing our dataset to include storms from various LDAR/LMA networks to investigate regional influences. Preliminary NLDN-Radar relationships applicable to all areas of the country will be discussed, as will the advantages of including IC lightning in the dataset. The GOES-R Global Lightning Mapper (GLM) will provide total lightning data at a resolution of 10 km. Therefore, we are examining GLM proxy parameters obtained from total lightning data to maximize the utility of our database and provide risk reduction for the GLM. The paper also will evaluate methods for determining the convective core and anvil regions of a storm as well as a technique to locate the source of flashes between these regions. Finally, the paper will describe a statistical approach to compare severe and non-severe storms, as well as the perceived benefits of employing these newly developed WDSS-II tools in an operational setting.