Assessment and prediction of monthly Indian summer monsoon rainfall by using nineteen large-scale circulation indices with four lags

Authors : Rahul Verma* and Ganesh D. Kale**

*Research Scholar, Department of Civil Engineering, Sardar Vallabhbhai National Institute of Technology, Surat-395007, Gujarat, India

**Associate Professor, Department of Civil Engineering,Sardar Vallabhbhai National Institute of Technology, Surat-395007, Gujarat, India

Download the book : “Artificial Intelligence and Water Resource Development”(adv)

Abstract

Indian summer monsoon rainfall (ISMR) provides 80% of total annual precipitation and thus has a tremendous impact on water resource management, agricultural yield and, consequently, on India's gross domestic product. In this study, assessment of hydro-climatic teleconnection between monthly ISMR and large-scale atmospheric/oceanic circulation indices is performed by using monthly composite index (MCI) and machine-learning technique named support vector regression (SVR) with linear kernel. The MCIs between monthly ISMR and final selected significant indices (FSSIs) are formed by using multivariate linear regression (MLR) for development phase periods 1951-1985, and 1951-1988 and these are tested during testing phase periods 1986-2014, and 1989-2014, respectively. Similarly, SVR model also has the same training and testing periods as of MLR model. The correlation coefficient is evaluated between observed and simulated monthly ISMR corresponding to both MLR and SVR models corresponding to development/training and testing phases. The study revealed that, correlation coefficients obtained by SVR model are better than that of MLR model for testing phases.