A research program and initial results from a tropical short-term forecasting experiment in Darwin, Australia, are described. The program, involving both real-time forecasting phases combined with ongoing post-analysis phases, is directed at improving our understanding and short-term forecasting capability of tropical weather phenomena. Based on Doppler radar observations, the key forecasts are convective initiation, wind-shift changes and the occurrence of strong winds. Initial results show a systematic relationship between the occurrence of tropical convection and boundary-layer convergence lines visible in the optically-clear air by Doppler radar. This provides the rationale for making probabilistic forecasts of the initiation of convection and wind-shift changes. The results also indicate the utility of Doppler radar in providing timely warnings of the occurrence of strong winds in the zero to three-hour time frame and in identification of microbursts.
Keenan et al. (Sun,) studied this question.