Further insights on the influence of the Indian Ocean dipole on the following year's ENSO from observations and CMIP5 models

Abstract

Recent observational studies have suggested that negative and positive Indian Ocean dipole (IOD) events (nIOD and pIOD, respectively) favor a transition toward, respectively, El Niño and LaNiña events one year later. These statistical inferences are however limited by the length and uncertainties in the observational records. This paper compares observational datasets with twenty-one 155-yr historical simulations from phase 5 of CMIP (CMIP5) to assess IOD and El Niño-Southern Oscillation (ENSO) properties along with their synchronous and delayed relationships. In the observations and most CMIP5models, it is shown that El Niños tend to be followed by La Niñas but not the opposite, that pIODs co-occur more frequently with El Niños than nIODs with La Niñas, that nIODs tend to be followed by El Niños one year later less frequently than pIODs by La Niñas, and that including an IOD index in a linear prediction based on the Pacific warm water volume improves ENSO peak hindcasts at 14 months lead.TheIOD-ENSOdelayed relationship partly results froma combination ofENSOintrinsic properties (e.g., the tendency forElNiños to be followed by LaNiñas) and from the synchronous IOD-ENSOrelationship. The results, however, reveal that this is not sufficient to explain the high prevalence of pIOD-Niña transitions in the observations and 75%of theCMIP5models, and of nIOD-Niño transitions in 60% of CMIP5 models. This suggests that the tendency of IOD to lead ENSO by one year should be explained by a physical mechanism that, however, remains elusive in theCMIP5 models. The ability of many CMIP5models to reproduce the delayed influence of the IOD on ENSO is nonetheless a strong incentive to explore extended-range dynamical forecasts of ENSO.