Objective prediction of severe thunderstorm environments: Preliminary results linking a decision tree with an operational regional NWP model

The output from the Australian operational regional numerical weather prediction model has been used to provide input thermodynamic and kinematic fields to a decision tree designed to diagnose the likelihood of thunderstorms and whether the thunderstorm environment is conducive to the development of severe, supercell, or tornadic supercell thunderstorms. On targeted cases of observed severe weather described in this paper the system successfully diagnoses severe thunderstorms and appears to discriminate between tornadic and nontornadic cases. It is shown that the thunderstorm decision tree of Colquhoun has considerable potential, when coupled with a regional NWP model, to provide forecast guidance of areas of thunderstorms, severe thunderstorms, tornadic thunderstorms, and whether these storms are likely to be associated with flash floods, downbursts, strong winds, etc. The results from targeted severe weather case studies show a very good degree of correspondence with the location and type of severe weather observed, and if these results were to be replicated in operational practice, the system would be of great benefit to weather forecast teams. The results of a 33-day trial indicate that the results from the case studies did not occur due to excessive overprediction, with a relatively low number of false alarms and relatively high number of hits based on the admittedly fairly loose subjective criterion for a hit. These results indicate that the system may provide a useful alerting system to form the basis of the definition of thunderstorm watch areas in operational practice in Australia.