MODELLING THE DAILY RAINFALL AMOUNTS OF NORTH-WEST PAKISTAN FOR AGRICULTURAL PLANNING

In this study, Gamma distribution models were used for modeling rainfall amounts on seasonal basis to derive results of potential value to agriculture. The daily rainfall data of two sites from Khyber Pukhtunkhwa Province, Pakistan viz. Peshawar and Swat (having significantly different mean annual rainfall) were analyzed. The seasonal pattern is summarized by fitting smoothed regression curves using generalized linear models. Thus for each day of the season, the mean amount of rain per rainy day could be computed using the estimates of parameters of the model. The adequacy of the fitted models for both the sites was checked by the analysis of deviance residuals and found satisfactory. The results show that Fourier series with three harmonics is suitable (for both the sites) to fit the Gamma distribution for modeling mean rain per rainy day. In addition, for both the sites, the pattern of rainfall amount (throughout the year) is found to be bimodal.