Past, Present and Future Perspectives of Seasonal Prediction of Indian Summer Monsoon Rainfall: A Review

This paper presents a detailed review and discussion on the long-range forecast (LRF) of Indian summer monsoon rainfall (ISMR), its present status, problems, and perspectives with an emphasis on the Indian meteorological department (IMD) operational forecasting methods and its forecasts. Verifications of IMD LRF for 1924-1987 revealed a success rate of nearly 64%. IMD operational LRF skills (correlation coefficient) during the period 1988-2020 were found as 0.25 and 0.34 for the 1st -stage (April) and 2nd -stage (June), respectively. From 1988 to 2020 (33-years), 21-years of predictions were predicted out of the confidence band (±4%LRF). Out of these 21-years, 18-years were individual or co-occurring El Niño-Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) event years. It indicates that IMD operational models have failed to capture ENSO and IOD phenomenon effects in their predictions considerably, despite the developing stage of ENSO and IOD events during the monsoon season. Moderate success and limitations of statistical models of IMD lead to the launching of the Monsoon mission (MM) in 2012 to use the dynamical prediction system by using the climate forecast system. Within a span of 7-years, MM has brought the skill of ISMR prediction to 0.71, which is above the older potentially predictable limit estimate of 0.65. Overall, in the last century, enormous progress has been made on the LRF of the ISMR. The accurate capturing of the ENSO and IOD phenomenon, incorporation of non-stationarity components in the predictors, improvements in the model microphysics and initial conditions could improve prediction skills.