The prediction skill of Indian Ocean dipole mode in DCPP-CMIP6 decadal hindcast models

Abstract The Indian Ocean dipole (IOD) mode events are one of the most fascinating interannual ocean-atmosphere phenomenon in the tropical Indian Ocean. The zonal contrast of tropical ocean conditions foists enormous unfavorable impacts on the regional weather and climate. Until recently, the predictability of IOD events was limited to only a season in advance which poses a huge socioeconomic loss in the affected regions. We explore the DCPP-CMIP6 decadal hindcast models for their predictability of IOD events at different years after their initialization. Two CMIP6 models viz., CanESM5 and NorCPM1 (to some extent) have quite significant prediction skills for IOD events even after eight to ten years of initialization. We study these two models to probe the features that impact the IOD predictability. Customized version of model components, refined ocean-atmosphere coupler and sophisticated data assimilation system are the major attributes to their skill for IOD predictability. It is evident from CanESM5 simulations that Southern Ocean provides notable signals through Antarctic circumpolar currents that propagate to tropical Indian Ocean and influence the IOD events.