Two cases illustrating limitations in forecasting tropical cyclones

The first case illustrates the numerical weather products limitations in forecasting cyclogenesis.

Despite progress made over the last few years by the numerical weather products (NWP) in terms of cyclogenesis forecasting, unforecasted cyclogenesis still remains possible, as illustrated by ERNEST, a tropical cyclone which developed rapidly in the northern part of the Mozambique Channel during January 2005. ERNEST's development was not forecasted by any NWP, which analysed a low but did not forecast any deepening of the system. Without any help of the NWP, forecasters lack any form of guidance in predicting the development of tropical cyclones. Moreover, such an exceptional and unforecast cyclogenesis which develops in the vicinity of inhabited areas is an example where an alert system failed to work properly.

The second case deals with the wreck of the Samson in the northern region of Mozambique Channel in March 2004, during the tropical cyclone Gafilo.

24 hours after Gafilo's landfall on the northeastern malagasy coast, hurricane force winds continued to blow, hurricane force winds continued to blow very locally in the vicinity of the northwestern malagasy coast and likely off coast, in connection with a very active outer cloudy band of the broad clockwise circulation. This band had swept through the region between Mayotte and the northwestern Malagasy coast, far from the tracking overland residual and almost filled circulation centre. The associated squall lines seem to have resulted in the wrecking of the Samson. The continuation of such strong winds following the landfall of a tropical cyclone is truly exceptional. If ever a hurricane force winds warning could not be forecasted, detailed analyses of QuikSCAT data, available at the time of the shipwreck, clearly shows that a violent storm force warning should have been issued.