This research focuses on enhancing understanding of the inter-relationships between the Geostationary Operational Environmental Satellite (GOES) and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite observations, total lightning information, and polarimetric radar fields related to the mechanisms that initiate lightning within convective clouds (so-called first lightning). A hypothesis to be tested is that one can estimate the onset, intensity and character of lightning events (i.e. the relative amount of lightning expected, in terms of flash rates over time) in convective clouds across a region. Lightning within convective clouds is far from ubiquitous, and therefore considerable uncertainty exists as to why certain storms are prolific lightning producers, while others nearby may not be. The project goals are: (1) Perform basic research to relate GOES satellite-observed (infrared) convective cloud growth properties to the character (intra-cloud versus cloud-to-ground) and intensity of lightning in these clouds, and to the relationship to in-cloud dual-polarimetric Doppler radar-observed microphysical distributions; (2) Enhance understanding of convective regimes by describing the general characteristics of lightning and convective cloud development as observed by satellite and dual-polarimetric Doppler radar in various environments; (3) Extend current ability to nowcast (0-90 minutes) lightning events using GOES and MODIS.
Intellectual Merit: The research will lead to an improved understanding of the interrelationships between lightning, cloud microphysics, and of satellite observations of growing cumulus. This research will take advantage of an exiting meteorological testbed that includes dual-polarimetric radar, lightning mapping array, GOES, MODIS and Tropical Rainfall Measuring Mission (TRMM) observations, and a mobile radar. Density of lightning will be correlated in space and time to graupel, freezing altitudes and ice distributions (all inferred from polarimetric radar), and compared to GOES cloud-top cooling rates as an estimate of in-cloud vertical mass flux, cumulus cloud growth, and cloud-top microphysics. Insights into cumulus cloud electrification, recharge, and lightning types will be subsequently developed from this research.
Broader Impact: A current CI forecasting technique developed by the Principal Investigators is already being evaluated within the Federal Aviation Administration's Aviation Weather Research Program and the National Weather Service. The mechanisms, therefore, are in place for this research to immediately impact a broad operational community. In addition, collaboration with other university scientists and graduate students will further extend this research to the larger academic and educational community.
In this collaborative research, co-investigator Eugene McCaul, USRA/STI, Huntsville, AL, supplied detailed lightning analyses for selected storm cells to lead investigator John Mecikalski, University of Alabama Huntsville, and his students and post-doctoral assistants. These data were compiled to serve as ground-truth total lightning behavior for the examined storms, so that the UAH investigators could determine the exact times and locations of first lightning events in the storms. Such knowledge allowed the UAH team to study the signatures of the storms in satellite and radar data, in order to determine whether their evolving satellite and radar signatures were sufficiently distinctive and repeatable to allow prediction of lightning initiation (LI) in storms. Improved prediction of LI has the potential to provide significant safety and cost-reduction benefits to the military, aerospace, airline travel, entertainment, sporting and public recreation sectors of society. The lightning data provided consisted of comprehensive lists of the times and 3-D spatial locations of "sources" of VHF radiation emitted by the selected storms' lightning flash channels as they build stepwise through space during their development. Such sources can be detected and logged by a network of surface sensors known as a Lightning Mapping Array (LMA). Each sensor logs the time and amplitude of impulsive VHF signals associated with lightning in storms, allowing sophisticated software analyses to infer the latitude, longitude, altitude and starting time of the radiating channel segment most likely to have generated the observed signals. The resulting chronological list of "sources" is then passed to an algorithm that clusters them in time and space to produce lists of all the lightning flashes seen in the storms. LMA networks are presently operational in Oklahoma, Alabama, Florida, and several other areas. Only storms that occurred within the zones of high data quality coverage were examined in this research. For this work, co-I McCaul developed specialized software that uses LMA-derived source maps to define the time-varying centroidal latitude and longitude of a storm of interest, as well as the evolving size and shape of the storm's lightning footprint. With this information and the flash clustering algorithm, a complete list of all the sources and flashes produced by an individual network-observable storm can be compiled. A 5-min time granularity is used, not only because it happens to correspond to the approximate interval of radar volume scans, but also because it is long enough for statistically meaningful sampling of lightning in most storms, while also being short enough to discriminate the time variations of fluctuating lightning activity in the storms. Prior studies by the UAH workers have shown that GOES-IR satellite signatures of growing storms can be related to their radar signatures, allowing predictions of convective initiation, as defined by storm reflectivity thresholds. In this project, the goal is to further relate the observed patterns and trends seen in LMA data to the corresponding trend signals seen in IR satellite data and ground-based radar data, in order to develop predictions of the time of first lightning in storms with lead times as long as possible. Seven LI interest fields from GOES IR satellites were examined in this project. The interest parameters usually exhibited clear trends, especially within about 30 min of LI, and thus were confirmed to be useful in anticipating first lightning. Radar parameters have also been shown to be capable of detecting the development of conditions in storms conducive to lightning flashes. Sufficient lightning, radar and satellite data were available to allow the research team to perform regional studies of lightning initiation in two important but contrasting convective regimes: Florida and Oklahoma. The Florida environments were typically characterized by moderate convective available potential energy (CAPE), weak vertical shear, weak or nonexistent capping inversions, and very high moisture contents, while the Oklahoma environments featured larger CAPE, stronger vertical shear, stronger capping inversions and somewhat drier atmospheres. Given these environmental differences, it was expected that the storms in the two environmental regimes would generally display distinctive, detectable differences in storm intensity, growth rate, and microphysics signatures. Analysis of numerous storms in the Florida and Oklahoma regimes revealed that, for the Florida storms, many of the satellite-based trend indicators, including cloud top cooling rate and anvil area, were useful with up to 60 min lead time before LI. For the Oklahoma storms, however, lead times were only around 30 min, because of the more rapid growth of those storms dictated by their particular environmental conditions. One important implication is that earlier attempts by others to specify monolithic lead times for satellite interest fields are likely to be problematic, if they fail to account for the varying environmental conditions that regulate storm growth rates.
