An Exploration of Radar-Detected Precursors to the Onset of Lightning in Convective Storms

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 5 January 2015: 1:45 PM
225AB (Phoenix Convention Center - West and North Buildings)
Michael Colbert, Colorado State University, Fort Collins, CO; and V. Chandrasekar, E. Ruzanski, and J. Hardin

This study analyzes the evolution of hydrometeors in relationship to the onset of lightning and subsequent trends in lightning flash rate within convective storms. Twenty storms were classified into four groups (ordinary cells with lightning, ordinary cells without lightning, supercells, and storms developing within a pre-existing stratiform precipitation band). The storms were tracked beginning twenty minutes (5 NEXRAD volume scans) prior to the onset of lightning, and tracking ended when lightning ceased, or when the storm moved out of range of the available data, whichever came first. For each NEXRAD scan, the total volume of each hydrometer was computed within a box 20 kilometers in length and width and 12 km in height, centered around the convective storm of interest. Cloud-to-ground (CG) and inter-cloud (IC) lightning data came from Vaisala's National Lightning Detection Network (NLDN), and hydrometeor classification was derived by running DROPS on NEXRAD data from KFWS in the Dallas/Fort Worth area of Texas, and interpolating the output onto a Cartesian grid with 500m spacing in the horizontal and vertical directions. All storms in the analysis occurred in April-May of 2014.

Results show that the identification of low-density ice crystals between 8 and 12km in altitude typically precedes the onset of lightning in supercells by an average of 10 minutes. In ordinary individual cells, low-density ice crystals form at these heights as early as 15 to 20 minutes before the onset of lightning. In the case of thunderstorms forming out of pre-existing stratiform banded precipitation, a three-layer structure typically exists with rain between the surface and 4 to 6km altitude, followed by a 1-km thick layer of high-density ice crystals, and a layer of low-density ice crystals at the top, reaching as high as 6 to 10km. As these features transition from stratiform to convective in nature, low-density ice is noted developing between 10 and 12km in height as early as 20 minutes before the onset of lightning. In all three categories of lightning-producing convective storms, the development of high-density ice, hail, and rain are also noted prior to or coinciding with the onset of lightning.

Furthermore, this study documents the correlations between lightning flash rate and the fraction of the total volume of hydrometeors classified as ice crystals, hail stones, or plain rain over a 400km2 area surrounding the centroid of lightning. The data suggests an exponential relationship between the fraction of a storm's volume containing hailstones, and lightning flash rate. Analysis also suggests that little-to-no lightning exists in storms where less than 25% of their volumes contain ice crystals, but lightning rates quickly increase when ice crystals make up 25 to 30% of the storm's volume. From 30% onward, lightning rates again decrease with increasing fraction of ice crystals. Lightning rates tend to be at a maximum in ordinary cells and supercells when rain comprises 20-40% of a storm's volume, while lightning rates in a convective cell which originates in pre-existing stratiform precipitation are maximized when rain comprises 40-70% of the volume of hydrometeors.