A reanalysis climatology of cool-season tornado environments over southern Australia

摘要

It is of great interest to assess the likely changes in occurrence of extreme weather events such as those associated with severe convective storms in future (changed) climates. However, while events such as tornadoes cannot be resolved by climate models, or even the higher-resolution operational numerical weather prediction (NWP) models, it has been shown that some of the environments in which they occur can be predicted. In Australia, the so-called cool-season tornado (CST) environment is defined as the area where the forecast 700 hPa Surface Lifted Index and the surface to I-km vertical wind shear exceed specified thresholds. CST threat area forecasts are issued with each operational mesoscale NWP forecast by the Australian Bureau of Meteorology. In this paper, the application of this diagnostic indicator of CST occurrence to typical climate model resolutions is tested using the NCEP/NCAR and ERA-40 reanalysis data sets. The techniques used to apply the diagnostic to these data sets are described, together with the calibration necessary to specify separate thresholds for each analysis set. Using these thresholds 40-year climatologies of CST risk over Australia are constructed for each reanalysis data set. It is shown that the two data sets indicate very similar spatial distributions of risk. Trends in frequency of occurrence are problematic, with significant discontinuities in frequency after the introduction of satellite sounding data in 1979, with an increased frequency, but little trend in numbers, of CST environments since that time. Relating inter-annual variations in the frequency of CST environments to large-scale circulation indices show little relation with regional sea surface temperature anomalies or with the Southern Oscillation Index, but showed significant negative correlation with the Southern Annular Mode Index.