Time Series Analysis and Forecasting of Precipitation (Rainfall) Data

Abstract

According to Mattew Rodell, chief of the Hydrological Sciences Laboratory at NASA's Goddard Space Flight Center, "Groundwater in northern India is being depleted at a rate of 19.2 gigatons per year." Furthermore, a report by NITI (National Institution for Transforming India) Aayog claimed that 40 percent of India's population will have no access to drinking water by 2030. It is an undisputed fact that the water crisis is a grave issue that we face today and facts like these, forces us to experiment with new plans and measures to carefully sustain our resources. Predicting rainfall is a significant step to mitigate plans related to water crisis and helps us to effectively use water resources, manage crop productivity and plan water structures. This work also aims to compare among efficient models (such as ML, and DL) to determine the best one for predicting the rainfall with higher accuracy. The raison d’etre of this paper is to study and develop efficient models and compare them to determine an approach that can predict rainfall with higher accuracy.