Introduction 1. GPATS - - what it is and the network - - how gpats is detected - - strengths and limitations 2. LIS - - what it is and the network - - how lis is detected - - strengths and limitations 1. GPATS Verification with LIS - - Extract and process LIS dates and times, accounting for leap seconds - - 8 days (4 summer 4 winter), of LIS subjectively chosen for spatially large amounts of lightning activity - - gpats restricted to LIS viewbox space/time windows resulting in #X boxes total A list of determined ES events is crucial to developing an ES climatology and synoptic classification. To discriminate U.S. ES events from SS, Colman (1990a), Moore et. al. 2003, Horgan et. al. 2007 and others used National Weather Service observer reports of thunderstorms to identify thunderstorm locations. Unfortunately, Australia does not have a dense spatial and temporal surface and upper air observing network. Using thunderstorm reports at upper air sounding locations and times may result in too few ES to create an Australian ES climatology and classify synoptic environments. In fact, a search of Melbourne Airport “Thunderstorm” Present Weather Observations at 00 and 12UTC between 2008-2015 yielded only 6 reports. An effective alternative to Bureau of Meteorology observer thunderstorms reports could be the location and timing of observed lightning strokes from the Global Position and Tracking System (GPATS), as a proxy for thunderstorm locations provided the network’s stroke location accuracy is within the typical thunderstorm spatial length scale of 4 to 40km (meso-beta scale, Fujita 1981). The European ADTNET sensor network with similar density and location calculation methods to GPATS, has published location accuracies of 5-6 km (Cummins and Murphy 2009). In addition, Kumar et. al. 2013 found almost all (94%) of GPATS strokes near Darwin over two wet seasons (October 2005–April 2006 and October 2006–April 2007), were within 10km of radar-derived convective cells. These results, although pertaining to the Australian tropical wet seasons, gives confidence to GPATS stroke location accuracies being within typical thunderstorm length scales. A limitation of the GPATS ground-based network is its ability to detect all types of lightning strokes (intra-cloud strokes (IC) and cloud to ground (CG) or ground to cloud (GC) strokes constituting “total lightning”). In a comparison of GPATS stroke types to stroke types from a high-quality research lightning flash counter (CGR4) in Brisbane, Kuleshov (2012) showed GPATS grossly under reported IC strokes, with the majority of detected strokes being either CG or GC types. Given IC to CG (or IC to GC) ratios range from 2-10 in most thunderstorms (Cummins and Murphy 2009), GPATS may be under-detecting much of the total lightning spatial coverage. A spatial comparison between GPATS strokes and stroke data from reliable total lightning sensors could quantify how much spatial coverage GPATS is missing. The Lightning Image Sensor (LIS) on the TRMM polar orbiting satellite is one such total lightning sensor, considered to be very accurate in detecting total lightning (Cummins and Murphy 2009). LIS stroke data, and spatio-temporal observation window data of the TRMM satellite swath (0.5 degree, 2 second each window respectively) is readily available. The spatial coverage of GPATS restricted to the same spatio-temporal observation windows of LIS, can be compared to the spatial coverage of LIS over Australia and , if comparable, GPATS can replace observed thunderstorm reports with superior spatial and temporal coverage. Given the vast amount of daily GPATS and LIS lightning stroke data (typically 100,000 to 300,000 strokes per day), only four summer and four winter days will be analysed. Lastly a GPATS dataset covering Australian and ranging from March 2008 to December 2014 is readily available from the Bureau of Meteorology. Any ES and SS climatology using this GPATS dataset to identify thunderstorm locations will similarly range from March 2008 to December 2014.
