306 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1956 



the coast. Thus, the forecasting problem varies for different areas 

 as well as for different storms affecting the same area. 



In this paper we shall discuss some of the things we have learned 

 about hurricanes in the past few years, the various forecasting prob- 

 lems connected with storms, the research needs, and the ultimate po- 

 tential for making accurate forecasts of hurricane inception, intensity, 

 and movement. 



ADVANCES IN KNOWLEDGE OF HURRICANES DURING THE PAST DECADE 



When Col. Joseph B. Duckworth piloted a single-engine AT-6 

 aircraft into a hurricane near Galveston, Tex., on July 27, 1943, a 

 new era began in the collection of data about hurricanes [22]. Pre- 

 viously, surface ship reports were almost the sole source of data con- 

 cerning hurricanes until the storms approached land. Once advisories 

 were issued concerning an existing hurricane, surface ships tended 

 to avoid its path. Thus in previous years each improvement in the 

 hurricane warning system usually resulted in less data being available 

 to the forecasters and to research workers. Within a year after 

 Colonel Duckworth's first flight it was routine procedure for recon- 

 naissance crews of the Air Force and Navy to make regular penetra- 

 tions of the violent storms from the Tropics. The development of 

 radar during World War II made available another powerful tool 

 for the hurricane service. Both when carried aloft and when used 

 from ground stations, radar equipment made possible observa- 

 tions about hurricanes in greater detail than ever before. Rawinsonde 

 equipment for measuring the winds, pressure, temperature, and hu- 

 midity of the air from the ground to levels above 100,000 feet have 

 come into widespread use in recent years. The additional data col- 

 lected in recent years by aircraft reconnaissance, radar observations 

 of hurricanes, and from an expanded network of upper-air sounding 

 stations have made possible increased understanding and knowledge 

 of hurricanes and greater efficiency of the hurricane warning 

 service. 



We now know much more of the vertical distribution of winds in 

 hurricanes. Prof. B. Haurwitz in 1935 [9], after computing the 

 difference in air density that would be required to counterbalance the 

 difference often observed across a hurricane in air pressure at sea 

 level, deduced that intense hurricanes must extend to above 30,000 

 feet. This has been verified. In general the winds of a hurricane 

 vary relatively little from the surface up to 10,000 feet. From there, 

 up to 30,000-50,000 feet (depending upon the characteristics of the 

 individual hurricanes) the winds gradually decrease with height. 

 Data collected during hurricane lone of 1955 and hurricane Betsy 



