Advancing Indian monsoon rainfall predictions

Abstract

Despite great strides made in seasonal climate forecasting using dynamical models, skill in predicting the Indian monsoon is woefully poor. Our analysis of the reasons for failure exposes a flaw in the popular design of dynamical prediction systems. The approach of driving atmospheric models with a projected ocean surface temperature presupposes Indian monsoon variability to be a consequence solely of the atmosphere reacting to the ocean. We demonstrate significant improvements in the skill of Indian monsoon predictions when atmospheric models are coupled to, and fully interactive with the ocean. The additional feedback of the atmosphere onto the ocean is thus deemed critical for harvesting skilful monsoon predictions.