Understanding the impact of the Indian Ocean dipole (IOD) on El Niño–Southern Oscillation (ENSO) is important for climate prediction. By analyzing observational data and performing Indian and Pacific Ocean pacemaker experiments using a state-of-the-art climate model, we find that a positive IOD (pIOD) can favor both cold and warm sea surface temperature anomalies (SSTA) in the tropical Pacific, in contrast to the previously identified pIOD–El Niño connection. The diverse impacts of the pIOD on ENSO are related to SSTA in the Seychelles–Chagos thermocline ridge (SCTR; 608–858E, 78–158S) as part of the warm pole of the pIOD. Specifically, a pIOD with SCTR warming can cause warm SSTA in the southeastern Indian Ocean, which induces La Niña–like conditions in the tropical Pacific through inter-basin interaction processes associated with a recently identified climate phenomenon dubbed the ‘‘warm pool dipole.’’ This study identifies a new pIOD–ENSO relationship and examines the associated mechanisms.
The diverse effects of pIODs on ENSO are primarily associated with two types of pIOD SSTA patterns, one with and the other without thermocline ridge warming in the western pole. The pIODs with thermocline ridge warming induce anomalous interhemispheric SST gradients that drive northerly wind anomalies over the south Indian Ocean. The northerly wind anomalies weaken the prevailing southeasterlies, further causing SST warming in the thermocline ridge and over the southeast Indian Ocean off the west coast of Australia, where local air–sea interaction gives rise to cyclonic wind anomalies to the west in subtropics, which is south of the anticyclonic wind associated with the pIOD in the tropics.
Through analysis of observational datasets and numerical experiments using a state-of-the-art climate model, we investigate the impacts of the pIOD on the simultaneous development of ENSO. Our results show that approximately one-third of the historical pIOD events are forced by El Niño, and they in turn tend to amplify El Niño. This result seems consistent with some of the previous findings, but different from the studies that use observational data in more recent periods. The discrepancy could be attributable to the decadal variations in the IOD–ENSO relationship and/or data quality issue during the early twentieth century due to the sparse observations. Furthermore, although changes in the IOD–ENSO correlation may be due to decadal variations in the ENSO variance and changes in the ENSO flavors, its causes may need further investigation. In addition, the IOGA experiments also show that 45% of the pIOD events were followed by La Niña in the next year in the model (not shown), which is consistent with the previous finding that a pIOD may favor La Niña in the following year.