An evaluation is made of the performance of the methods available for forecasting tropical cyclone positions in the Australian region. The measure of skill used is the accuracy relative to a climatological-persistence (CLIPER) scheme. The CLIPER technique used in this study is described in some detail and is similar to the classical scheme of Neumann. Itis intended to be used both as a forecasting aid and a no-skill benchmark for other forecasting techniques. Of the techniques presently available to the Australian Bureau of Meteorology, only the statistical-synoptic regression scheme developed by Keenan has significant skill, measured relative to CLIPER. All the other techniques have no significant skill relative to CLIPER. Significant gains in skill may be obtained simply by combining the present forecasting techniques in an optimal linear manner. In particular, when forecasts from CLIPER and the Australian region numerical weather prediction model are optimally combined, an improvement over CLIPER of 15per cent and 17 per cent is obtained at 24 and 36 hours respectively. This latter result suggests that large gains in forecasting skill for tropical cyclone motion should prove to be readily achievable in the Australian region.
LESLIE et al. (Fri,) studied this question.