A Weighted Multiple Regression Model to Predict Rainfall Patterns: Principal Component Analysis approach

In this study, a multiple regression models developed to explain and predict mean annual rainfall in Zimbabwe. Principal component analysis is used to construct orthogonal climatic factors which influence rainfall patterns in Zimbabwe. The aim of the study is to develop a simple but reliable tool to predict annual rainfall one year in advance using Darwin Sea Level Pressure (Darwin SLP) value of a particular month and a component of Southern Oscillation Index (SOI) which is not explained by Darwin SLP. A weighted multiple regression approach is used to control for heteroscedasticity in the error terms. The model developed has a reasonable fit at the 5%statistical significance level can easily be used to predict mean annual rainfall at least a year in advance. DOI: 10.5901/mjss.2014.v5n7p3