An empirical study on rainfall patterns of monsoon season in the north-west India using time series models

摘要

Rainfall is an important natural source of water, which is the most essential component of living organisms. The study of rainfall modeling and forecasting are critically important for an agrarian based country like India. Whereas, the fluctuations of economy and security of food depends on the timely availability of rainfall. The current study adopts time series modeling schemes to model the rainfall pattern of the Southwest monsoon using data from 1901 to 2015 in the North-West India. The univariate time series multiplicative SARIMA models were implemented for forecasting. The diagnostic checks showed that the seasonal model SARIMA (1, 0, 1) x (1, 1, 1)([4]), with AIC score of 4373.97, fits the series well, and realistic forecasts were obtained. The accuracy level of each model is computed by various performance measures. The identified SARIMA model was found to be an appropriate tool for forecasting of the Southwest monsoon rainfall patterns.