The fundamental role of air-sea interactions for dynamical seasonal predictions of Indian Monsoon rain

A convergence of remarkable advances in climate observing systems and computing capacity have propelled dynamical methods for prediction to the forefront. Yet, these emergent approaches have yielded inexplicably low skill when applied to seasonal Indian monsoon rains. We show that the practice of using sea surface temperatures as specified forcing for atmospheric general circulation models is flawed for the Indian monsoon, and artificially suppresses skill. We demonstrate significant improvements in the skill of Indian monsoon predictions for atmospheric models coupled to, and fully interactive with the underlying ocean. Of particular importance is the proper representation of air-sea energy exhange over the Indian Ocean warm pool.