!5.2A TOTAL LIGHTNING AS AN INDICATOR OF MESOCYCLONE BEHAVIOR Sarah M. Stough* 1, Lawrence D. Carey! 1, and Christopher J. Schultz1,2 1Department of Atmospheric Science, University of Alabama in Huntsville, Huntsville, AL, USA 2Earth Science Office, NASA Marshall Space Flight Center, Huntsville, AL, USA ! 1. INTRODUCTION the particle scale is followed by storm scale ! charge separation due to differences in particle fall With the increasing availability of total (i.e., in- speeds and the action of a vigorous updraft. The cloud and cloud-to-ground) lightning data, many low-to-mid-level updraft that is responsible for studies have documented the relationship cloud electrification via NIC and ultimately flash between total lightning activity and severe weather production could also contribute to the tilting of (e.g., Williams et al. 1999; Goodman et al. 2005; environmental horizontal vorticity into the vertical, Steiger et al. 2007; Montanya et al. 2009; Schultz the subsequent stretching of this vertical vorticity, et al. 2009; Darden et al. 2010; Gatlin and and the development and intensification of a Goodman 2010; Pineda et al. 2011; Schultz et al. mesocyclone (Rakov and Uman 2003; Lemon and 2011). This relationship has recently been Doswell 1979). explored further through the framework of the A quasi-steady, rotating updraft, or lightning jump, which is characterized by a mesocyclone, extending through the depth of a definable rapid increase in lightning flash rate storm is often the primary indicator in the initial (Schultz et al. 2009; Gatlin and Goodman 2010; diagnosis of a severe supercell storm. Although Schultz et al. 2011). From related research, it has only roughly 26% of mesocyclones have been been found that rapid increases in total lightning found to be associated with tornadoes, flash rate, or lightning jumps, often precede approximately 90% of all mesocyclones are instances of severe phenomena at the ground. associated with severe phenomena (Stumpf et al. These results indicate that lightning data may 1998; Trapp et al. 2005). Despite this knowledge, possess some operational utility in providing challenges remain in nowcasting severe weather increased confidence in warning decisions given that include correctly identifying and diagnosing added awareness of storm characteristics, the first severe storm of the day as well as resulting in increased warning lead time. providing advanced warning on the first tornado of In order to maximize the capabilities of total the day (Brotzge and Ericksen 2009; Brotzge and lightning data for nowcasting severe storms, its Donner 2013). fusion with proven tools has become a major goal The work here explores the temporal in the research and operational communities. This relationship between enhanced supercell rotation is also driven by anticipated widespread total and intensification of lightning activity objectively lightning detection capabilities afforded by the identified by the Schultz et al. (2009; 2011) two- Geostationary Operational Environmental Satellite sigma lightning jump. Using National Weather Series R (GOES-R) Geostationary Lightning Service (NWS) Weather Surveillance Radar – Mapper (Goodman et al. 2013). As such, this 1988 Doppler (WSR-88D) and Lightning Mapping study lays some of the conceptual groundwork for Array (LMA) data, an initial investigation of fusing radar with total lightning on a national level supercell thunderstorms is conducted to determine into a multi-sensor algorithm for severe weather how lightning coupled with radar may give earlier detection and forecasting. indication of updraft strength to improve situational The premise for such an algorithm is based awareness, increase warning lead time, or upon the microphysical and kinematic connection potentially “tip the scales” between severe versus between storm electrification and dynamics. In tornado warnings given a priori environmental particular, the updraft plays a pivotal role in both knowledge. Section 2 provides additional charge separation leading to flash initiation and information about the datasets used, while Section mesocyclogenesis. The primary means for cloud 3 details analysis methods. Results and electrification is thought to be the rebounding interpretation are provided in Section 4 with collisions between graupel and ice crystals in the concluding remarks and future work outlined in presence of supercooled water, or so-called non- Section 5. inductive charging (NIC) (Takahashi 1978). NIC at ! !* Corresponding author address: Sarah M. Stough, Univ. of Alabama in Huntsville, Earth System Science Center, Huntsville, AL 35805-1912; e-mail: [email protected] !1 2. DATA factor characteristics and Doppler velocity ! calculations from several S-band radars in the The primary data sources for this study WSR-88D network. Crum and Alberty (1993) include total lightning data from local LMA explain the benefits and limitations of the original networks as well as archived Level II and Level III NEXRAD network, which has since been data from several S-band WSR-88D radars in the upgraded to dual-polarization capabilities. Next Generation Weather Radar (NEXRAD) Reflectivity was generally used in each case to network available from the National Oceanic and a s s e s s s t o r m s t r u c t u r e f o r s u p e r c e l l Atmospheric Administration (NOAA) National characteristics. Meanwhile, Doppler velocity data Climatic Data Center (NCDC). These data were were analyzed for qualities of storm-scale rotation collected from the North Alabama and Central and the presence of a mesocyclone. Oklahoma regions for three separate storm case In addition to the Level II digital level base studies. data of reflectivity and Doppler velocity, two of the ! output product datasets from the Level III Radar 2.1 Lightning Data Product Generator (RPG) were also used. The ! NOAA National Severe Storms Laboratory (NSSL) Data from the North Alabama Lightning Digital Mesocyclone Detection Algorithm (MDA) Mapping Array (NALMA) and the Oklahoma and Tornado Detection Algorithm (TDA), referred Lightning Mapping Array (OKLMA) were used in to in the data by product codes “NMD” and “TVS”, this study (Goodman et al. 2005; MacGorman et respectively, were chosen so that an objective al. 2008). LMAs are small-scale networks that definition and time history of storm rotation would measure the time of arrival (TOA) of very high be available for analysis (Mitchell et al. 1998; frequency (VHF) radio waves emitted by lightning Stumpf et al. 1998). While both algorithms rely discharges. These networks are capable of upon spatial Doppler velocity constraints for detecting individual point VHF radiation sources identification, the MDA additionally requires associated with electrical breakdown, such that persistent identification through time of rotation they can be mapped in two or three dimensions to defined by different horizontal spatial requirements represent individual stepped leaders of a flash. To and more detailed vertical spatial constraints than reconstruct a lightning flash, an algorithm is used the TDA. The TDA product is described in detail in to cluster the sources by time and location Mitchell et al. (1998), while the MDA is proximity factors into groups comprising the flash. documented by Stumpf et al. (1998). Only the Here, a flash clustering algorithm similar to that presence of a tornado vortex signature (TVS) from described in McCaul et al. (2005) is used. In the TDA output, or lack thereof, was considered in general, LMA stations typically record the time and this study; however, the mesocyclone strength magnitude of the peak radiation emitted from index (MSI) attribute from the NMD output was lightning in intervals of 80 "s, in a local unused chosen as an analysis parameter. This attribute television channel (e.g., at about 80 MHz for the takes into account vertically integrated strength NALMA). These measurements result in tens to ranks of rotation computed and thresholded based hundreds of recorded source points detected per on gate-to-gate Doppler velocity difference and flash from multiple stations, at least six for better shear (Stumpf et al. 1998). spatiotemporal resolution, whose time recordings ! and stationary positions are used to locate the 3. METHODS emittance time and location of each source in a ! flash. For each source point, a chi2 statistic is Three supercell storms were analyzed for calculated revealing a goodness of fit and quality lightning flash rate and storm-scale rotation of the data. Together, all of these points can characteristics. While the NMD and TVS data were provide mapped sources with horizontal (vertical) already in the form of post-processed algorithm location errors of less than 500 m (1000 m) within output, associating lightning flashes with storms, a range of 100 km of the network (Koshak et al. calculating lightning jumps, and determining a 2004). Outside of this network range, however, proxy for mesocyclone presence, persistence, and measurements have been determined to decrease strength in Doppler velocity data required data in location accuracy, particularly for height manipulation before combined analysis. For the calculations (Koshak et al. 2004). process of associating lightning flashes with ! particular storms and conversion of Doppler 2.2 Radar Data velocity data into layered azimuthal shear fields, the Warning Decision Support System - Integrated Storms selected for this study were analyzed Information (WDSS-II) was used (Lakshmanan et and interrogated based