A study of the variability of thunderstorm electrical events based on very-low-frequency electromagnetic data

Abstract

The incidence of lightning discharge was estimated, the relationship of lightning-flash density to thunderstorm days was studied, and the causes of the variability in c loud-to -ground discharges were explored. The primary data were counts of sferics. The estimates of the incidence of discharge include areal and -global statistics. The global estimate was 1 .3 x 10^-5 km^-2 s^-1. Also included are monthly values over a regularly-spaced grid that includes much of the Eastern Hemisphere. The Southern Hemisphere (S .H.) tropics (0° -20°S) were higher in average incidence of discharge than the Northern Hemisphere (N .H.) tropics (0° -20°N). The extra-tropical zone of the S.H. (20°S-35°S) was dominant in incidence of discharge relative to other zones. Lightning occurred much less frequently in high vs. middle latitudes but the number of discharges per km^-2 s^-1 during stormy periods was similar in magnitude. The relationship between flash density and thunderstorm days was studied for the months of January, April, August, and November with a curvilinear regression analysis. Many useable regressions were produced. The relationship varied by season. The regressions for all the data produced estimates of flash density which were generally the same order of magnitude as that of previous studies. All estimates of flash density predicted by the regressions based on stratifications of the data were low in comparison to previous studies. The large-scale causes of the variability of cloud-to -ground discharges (dependent variable) were studied. February and August data for a selected area of the Eastern Hemisphere were chosen for study. Multiple linear regression (11 regressors in February and 12 in August) was used with and without a logarithmic transformation of the dependent variable. The regressors were available on the National Meteorological Center (NMC) grid and were averaged in various ways to conform to the resolution of the dependent variable (10 °-data blocks). The thrust was to explain the variability of cloud-to -ground discharges when significant lightning had occurred since this is the problem of physical and practical importance...