A POWER REGRESSION MODEL FOR LONG RANGE FORECAST OF SOUTHWEST MONSOON RAINFALL OVER INDIA

Abstract

A detailed analysis of southwest monsoon (June to September) rainfall over India of several decades vis-a-vis the regional and global antecedent signals in numerous permutations and combinations has led the authors to conclude that a long range forecast based on one, two, three or four parameters as attempted by several workers in the past, cannot be reliable on all occasions as indeed has proved to be the case. The parametric and power regression models utilizing 16 parameters, described in the present paper, suggest that it is a tapestry of several parameters and interactive nature of the regional and global climatic forcings that govern the quality and quantity of the monsoon. A detailed analysis of non-linear interactions among the antecedent climatic conditions and the monsoon has led the authors to introduce the concept of proportionate weightage to the signals of different parameters. This has led to the development of a power regression model, which is able to quantify the effect of each parameter. Details of the model are presented, Based on the model, the India Meteorological Department has been issuing the operational long range forecast of monsoon rainfall over India as a whole during the past 3 years, 1988 to 1990, and these forecasts have proved to be correct.