Relations of Thunderstorms and Cloud-to-Ground Lightning Frequencies

Temporal and spatial relationships between thunderstorms (events) and flashes were investigated using data for 1983-85 for 25 first-order stations (10 in the West and 15 along the East Coast). Thunder events were compared with flashes within three ranges: 5 km, 10 km, and 20 km, around each station. Cluster analysis revealed six geographic regions: Florida, Southeast (South Carolina, Georgia), Mid-Atlantic (Virginia, Maryland, Pennsylvania), Northeast (New York and New England), Rocky Mountains, and an intermontane area. Periods of multiple flashes not within thunder events and within 10 km of a point (most realistic for audibility), revealed that 10% to 20% (depending upon region) of all thunderstorms were missed. Also, 13% (Rockies) to 44% (Mid-Atlantic) of all thunderstorms have recorded durations too short (missed flashes before their reported start), and the average underestimated durations were from 55% (Northeast, Mid-Atlantic) to 26% ( Rockies). Flashes isolated in time and space, due to locational errors, represent 1% of all flashes in the east and 3% to 5% in the west where the data are poorer. Errors in flash data appear minimal but the errors in thunder events are sizable. Thunder events and flash frequencies related well based on major features in their average areal patterns and their between-year changes at stations. Correlations of flashes with events varied; their annual point frequencies had coefficients of +0.83 (east) and +0.67 (west). Durations of events and flash frequencies were poorly correlated with skewed distributions (often large flash frequencies in a few storms). The percent of all recorded flashes (within 10 km) in thunder events varied from 28% to 44% at western stations and from 13% to 20% at eastern stations. Thunder events with >= 1 flash varied widely, from 71% of all events at Washington, D.C. to 30% at Boston. Major east-west differences existed in the frequency of thunder events with flashes, reflecting poorer audibility of thunder in the west. Part of the difference is due to flash recording problems in the west, leaving flash frequencies that are underestimates of the true values. Latitudinal distributions were marked with north-to-south increases in thunder events and their durations, frequency of flashes, and number of flashes not in events. More missed flashes in the south suggested that atmospheric conditions in northerly U.S. latitudes enhance audibility. With a 20-km sampling radius, between 6 (Northeast) and 23 (Southeast) thunder events are not recorded yearly, but these averages drop to 1 (Northeast) and 4 (Southeast) based on flashes within 5 km. The data on thunderstorms is generally poor from two perspectives: 1) the recorded data miss sizable numbers of storm events, and 2) when recorded, 30% to 50% often underestimate durations based on nearby lightning activity.