326 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1956 



empirical in nature and it is only in the last two years that promising 

 semiobjective methods have been developed. It is reasonable to be- 

 lieve that if we can first fully understand the mechanism of how hurri- 

 canes move, it will be possible to make more rapid progress in develop- 

 ing improved methods in the future. At the present time most meth- 

 ods used for forecasting either inception or intensification are very 

 hazy and show relatively little skill when used under operational 

 conditions. 



While the possibility for eventual control of hurricanes is probably 

 many years in the future, any rational approach to research on this 

 problem is dependent on first acquiring much more knowledge of 

 their structure and dynamics. The forces released in hurricanes are 

 so much stronger than any of the forces that man can bring to bear 

 against the hurricanes, it is not practicable for us to attempt control 

 by any direct method. The only possibility for controlling either the 

 movement of the hurricane or its formation is first to find some mecha- 

 nism in the hurricane itself that man can use to make the hurricane 

 control or deviate itself. That is, man must search for power brakes 

 or power steering mechanisms in the hurricane structure that man can 

 activate and thus eventually cause the hurricane to move along a cer- 

 tain path or to weaken and dissipate. 



ULTIMATE POTENTIAL FOR FORECASTING 



The development of high-speed computers has made possible the 

 use of numerical techniques in forecasting. Experimentation during 

 the last 2 or 3 years leads one to believe that ultimately it will be 

 possible to make forecasts of the movement of hurricanes for 24 to 

 48 hours in advance with such accuracy it will enable the forecasters 

 to issue hurricane warnings for coastal areas with more precision. 

 This should make it possible to warn people in relatively restricted 

 areas if they are to be affected by hurricane force winds within the 

 next 24 hours and to omit warnings for all people who are very far 

 removed from the threatened areas. It should make possible the 

 alerting of specified coastal areas 48 hours in advance of the possi- 

 bility of hurricane winds within the specified period without having 

 to alert practically the entire United States coast. One of the pur- 

 poses of the present research program is to determine precisely what 

 data are needed for making successful hurricane forecasts. Once 

 this is determined it should be possible to collect the required data 

 with less effort than is being expended at the present time in gather- 

 ing reconnaissance data and other information for use in hurricane 

 forecasting. 



