Prediction of winter minimum temperature of Pune by analogue and regression methods

Abstract

Objective forecasting schemes to forecast the winter minimum temperature of Pune have been developed. The schemes have been based on multiple regression, analogue and a combination of the above two techniques. As many as 12 prescriptors have been screened to select predictors for developing regression equations for 24, 15 and 12- hr forecast schemes. The predictors chosen for the regression equations have been suitably used to develop forecast schemes based on the technique of analogue and analogue cum multi-regression. The schemes have been developed from a large data sample comprising of 1116 days and have been tested in an independent sample of 403 days. The schemes have given very encouraging results. The relative importance of the various predictors and of the different schemes have been discussed.