Accession Number : AD0674478

Title :   UTILIZATION OF PAST DATA IN NUMERICAL FORECASTING,

Corporate Author : EMMANUEL COLL BOSTON MASS ORIENTAL SCIENCE RESEARCH LIBRARY

Personal Author(s) : Koo Chen-Chao,

Report Date : MAY 1968

Pagination or Media Count : 21

Abstract : Up to now, past data have not been utilized in numerical weather forecasting due to the particular formulations of the problem. This paper indicates that the condition for solution, formed by a whole set of available meteorological data, is not compatible with the general system of weather equations. Thus, it is necessary to study the best way of formulating the problem using the greatest possible amount of available data in the condition for solution in order to obtain the optimum way of forecasting. Two possibilities are given. First, it is shown that under certain conditions, weather forecasting may be formulated as an evolutionary problem, in which past data are utilized. Secondly, past data could be partly utilized in an initial-value problem in which the system of weather equations, with certain meteorological variables eliminated, contains higher order time derivatives. The significance of these formulations is discussed. (Author)

Descriptors :   (*WEATHER FORECASTING, *NUMERICAL ANALYSIS), ATMOSPHERE MODELS, MATHEMATICAL MODELS, BAROMETRIC PRESSURE, ATMOSPHERIC MOTION, HYDRODYNAMICS, PARTIAL DIFFERENTIAL EQUATIONS, BOUNDARY VALUE PROBLEMS, CHINA

Subject Categories : Meteorology

Distribution Statement : APPROVED FOR PUBLIC RELEASE