Forecasting river basin yield using information of large-scale coupled atmospheric–oceanic circulation and local outgoing longwave radiation

Abstract

ABSTRACT Global and local climate parameters influence the distribution of precipitation over continents. The spatio-temporal distribution of rainfall and the depth of rainfall over a river basin influence the river basin yield. This study deals with the prediction of the river basin scale yield of the ‘Upper Bhima River basin’ from the Maharashtra State of India. The information on large-scale circulation patterns El Nino-Southern Oscillation (ENSO), Equatorial Indian Ocean Oscillation (EQUINOO) index, Multivariate ENSO Index (MEI), and local meteorological input viz. Outgoing longwave radiation (OLR) has been used to predict the river basin scale yield. The Artificial Intelligence (AI) tool – Genetic Programming (GP) – is used for developing prediction models. Ten different combinations of input variables are attempted for the development of monthly river basin yield models to arrive at the best input variable combination for the best predictions with varying lead times. Also, three combinations of input variables were tested for the prediction of ‘Seasonal Yield’. The findings of this research work indicate that GP-derived monthly River Basin Scale Yield forecasting models are successful in the prediction of yield with a correlation coefficient of 0.83. The seasonal yields could be predicted with a correlation coefficient of 0.75.