A Review on Indian Summer Monsoon Rainfall Prediction Using Machine Learning Techniques

A significant part of the Indian economy is still reliant on agriculture, which is directly linked to rainfall. Southwest monsoon precipitation over the nation is mostly consistent. The country's average seasonal rainfall is 89 cm (1941-1990 period mean value), with a variance coefficient of about 10%. A year above(less than) 110% (90%) rainfall is known as excess (drought) year. Currently, Indian meteorological Department (IMD) is using ensemble multiple regression (EMR) and projection pursuit regression (PPR) models for 1st stage and 2nd prediction of Indian summer monsoon rainfall (ISMR) respectively. In the paper, we study different potential climate parameters relationships with ISMR and various machine learning models. An analysis of percentage deviation of ISMR data from seasonal mean rainfall for the period (1991-2019) is also provided which shows deficit, normal, and excess monsoon rainfall years.