Assessing the Reliability of Rainfall if Monthly Falls are not independent

Abstract

Glover & Robinson (1953) showed that for practical purposes the rainfall at a station in any month of the year can be assumed to be normally distributed, and that the probability of the rainfall being above a certain minimum value (selected with regard to crop requirements) can be calculated accordingly. If the rainfall in anyone month is distributed independently of the rainfall in other months of the year, or season, then the probability of a pattern of rainfall month by month can be calculated by multiplying the several probabilities of the individual months. It was suggested that, in general, this was sufficient; but a method was given for obtaining the probability of a given pattern when association existed between the different months. In this the correlated monthly rainfalls, Xi' were transformed to uncorrelated variables Y . The value of each Y, when all Xi take their minimum values, was calculated and called 'minimum Y'. The probability that each Y would exceed this 'minimum Y' was calculated and the probabilities multiplied.