Accession Number : AD0775848

Title :   Numerical Procedures for Analyzing and Predicting Mesocale Tropical Weather Patterns.

Descriptive Note : Final rept. 26 Feb 71-31 Dec 73,

Corporate Author : STANFORD RESEARCH INST MENLO PARK CALIF

Personal Author(s) : Mancuso,Robert L. ; Wolf,Daniel E.

Report Date : JAN 1974

Pagination or Media Count : 59

Abstract : A multi-layer numerical model was developed for forecasting sybsynoptic weather over tropical regions. The model is based principally on winds, vorticity, and divergence that are potentially observable from geosynchronous satellites. It has been developed for use on a mini-computer system and is designed to facilitate testing of different formulations. Experiments were carried out to test various advection techniques since it is particularly important that mesoscale features be preserved in the forecast. In these experiments, the different techniques were compared with regard to their accuracy and computational efficiency. A quasi-Lagrangian scheme was found to give significantly better results than the more conventional methods, such as the leap frog. Investigations were carried out in regard to using cloud motions as wind input to the numerical forecasting model. A method for determining winds from satellite data in clear areas was proposed: tracking of humidity patterns on isentropic surfaces. (Modified author abstract)

Descriptors :   *Weather forecasting, *Tropical regions, Numerical analysis, Meteorological satellites, Data reduction, Wind, Cloud cover, Atmospheric temperature, Water vapor, Interpolation

Subject Categories : Meteorology

Distribution Statement : APPROVED FOR PUBLIC RELEASE