Stochastic vectorial flow graph analysis of meteorological cause-effect relationships for drought prediction

Abstract

Advocating the view that the drought prediction models which neglect the complex feedbeck mechanisms in the cause effect relationships between variables of meteorological, hydrological and topographic origin may give faulty predictions, a system theoretic method is proposed for taking into consideration certain dominant cause-effect reIationships. The method is based on the concept of stochastic vectorial flow graph developed earlier by the author which is used in conjunction with a characterization of the interacting stochastic variables by cross-autocorrelation. A comprehensive feedback system of importent interacting Eeteorological variables is proposed for drought prediction along with an approach for computer aided analysis of such a system.