An objective scheme for forecasting the motion of tropical cyclones (MOHATT) , under development since 1967 by the U.S. Navy Fleet Numerical Weather Central and the U.S. Naval Postgraduate School, Monterey, Calif., is described and applied to the 1967-71 North Atlantic tropical cyclones for forecast intervals up to 72 hr. The MOHATT scheme involves steering of the center of the cyclone by geostrophic winds derived from heavily smoothed isobaric height fields (both analyzed and prognostic) and a

more »

... statistical correction determined by the behavior of the first 12 hr of the steering forecast. The developmental sample (1967-70) used to establish the potential accuracy of MOHATT indicates 700 mb as the optimum steering level, but the fully operational test in 1971 suggests that the 850-mb level may be an improvement for forecast intervals beyond 36 hr. An analysis of the 1971 forecast data shows errors (expressed as nautical miles per hour of forecast interval) ranging from 6.1 kt a t 12 hr (252 cases) to 5.1 kt at 42 hr (199 cases) and 5.5 k t at 72 hr (158 cases). Relative to stage of development, hurricanes are forecast with the most success, errors ranging from 4.4 kt a t 18 hr (98 forecasts) to 5.5 kt at 72 hr (85 forecasts); relative to area, the eastern Atlantic yields the more accurate forecasts. Comparison of MOHATT with the National Hurricane Center's (NHC) NHC-67 technique using a homogeneous sample of 1971 forecasts indicates that MOHATT accuracy exceeds that of NHC a t intervals beyond 24 hr, with the improvement exceeding 30 percent at 72 hr. Similar tests relative to the typhoon-tracking forecast scheme (TYRACK) developed a t the U.S. Navy Fleet Weather Central, Honolulu, Hawaii, show that MOHATT errors averaged 38 percent less than TYRACK in 1971. Interpretation of the results of comparing operational Official and MOHATT forecasts is complicated by differences in the forecast intervals of the two systems. I n any case, based on the 1971 operational test, the MOHATT forecasts are more accurate than the Official ones after 36 hr, while near equivalence prevails in the earlier intervals. Forecast examples and a discussion of various facets of the MOHATT scheme are included. a "Quasi" refers to the use of a sine function, sin #, of the following form in the geostrophic wlnd equation et latitudes less than 30": sin O'=Z[O.% sin 8+0.%)*+0.26 sin 81. 81n @'ranges from 0.125 at the Equator (the sin 0 value at 7.20 1st.) to 0.63 at 30°1atltude.