Proximity soundings for the US from global reanalysis data: 1991-1999 (Formerly P12.4)

Using soundings derived from the NCAR/NCEP global reanalysis data, we have examined environmental conditions associated with significant tornadic reports (F2 or greater), significant non-tornadic reports (65 kt winds and/or 2-inch diameter hail), nonsignificant severe reports, and other soundings for the United States from 1991 through 1999. The period was chosen because recent work suggests that it was a period of relatively consistent tornado damage assessment. The reanalysis provides information every 6 hours on a latitude-longitude grid with spacing of approximately 1.9x1.9 degrees. There are 282 locations in the reanalysis over the contiguous 48 states, leading to approximately 3.7 million soundings. 837 of the soundings are associated with significant tornadoes and 4849 are associated with non-tornadic significant storms. The size of the dataset allows us to look at regional, interannual, and intraannual variability. Although the product of CAPE and deep tropospheric wind shear is an important discriminator between severe and non-severe convection, the probability of getting severe convection, given a particular value of the product, varies strongly with region and time of year. In particular, lower values of CAPE*Shear are associated with severe convection in the cold season. This difference is likely due to the difficulties in estimating convection initiation just given large-scale environmental information of the kind associated with proximity sounding data.

Probabilities of severe convective events, given a particular combination of CAPE and deep shear show that the probabilities of significant severe storms (including tornadic and non-tornadic) are greater for high CAPE and shear. Tornadic storms, however, have a greater dependence on shear values, given a threshold value of CAPE is present.

The two results (dependence of the probability of convective initiation on time of year and location, and the increased dependence on shear of tornadic storms) have implications for our ability to estimate changes in the frequency of severe convection in a greenhouse-enhances atmosphere.