The ECHAM4 stand-alone simulation considerably overestimates the equatorial Indian Ocean (IO) rainfall while underestimates the monsoon rainfall near 15oN over the Indian subcontinent. This feature is commonly seen in many other AGCMs (Gadgil and Sajani 1998). Upon coupling ECHAM4 with the UH ocean model (without heat flux correction), the equatorial IO rainband in the stand alone experiment splits into two zonal rain belts similar to the observed rainfall distributions. The monsoon rain belt near 15oN is stronger than the equatorial IO rain belt. These two rain belts are connected by the northward propagating intraseasonal oscillation that was significantly enhanced by the air-sea interaction. The atmosphere-ocean coupling not only improved the general rainfall pattern and the timing of onset, peak, and withdrawal of mean Asian-Pacific summer monsoon, but also improved the CISO of the Asian-Pacific summer monsoon. In the stand-alone simulation, the CISO were too strong in the equatorial IO. The air-sea coupling, on the one hand, produces more realistic magnitude of CISO variability at the equator, on the other hand, significantly enhances the CISO over the Arabian Sea and the Bay of Bengal. The northward propagation of CISO in the above regions is more robust in the coupled system than in the stand-alone run.
Both the local and remote air-sea interactions in the tropical Indian and Pacific Oceans contribute to the improved simulations of the mean Asian summer monsoon and its CISO. The local impact is through changes of shortwave radiation and latent heat flux at the equatorial IO. In the coupled experiments, the increased cloud covers (surface winds) associated with the excessive rainfall near the equatorial IO reduce (increase) the downward solar radiation (upward latent heat flux). Both effects cool the sea surface, thus reducing the local rainfall. The reduced equatorial IO SST further enhances the local Hadley cell and favors the development of rainband around 15oN. This local negative feedback mechanism also diminishes the equatorial IO CISO appeared in the stand-alone atmospheric run. The remote impact is through the change of surface winds in the western-central Pacific. The latent heat released in the rainfall near the equatorial IO and South China Sea excites an atmospheric Kevin wave response in the equatorial western-central Pacific. The resulting surface easterly wind cools the sea surface in the coupled system, reducing the rainfall over the equatorial western-central Pacific. The reduced equatorial rainfall further enhances the Indian monsoon rainfall through strengthening Walker circulation.