AUTHOR=Hu Shuai , Wu Bo , Zhou Tianjun TITLE=Decadal variation of prediction skill for Indian Ocean dipole over the past century JOURNAL=Frontiers in Climate VOLUME=4 YEAR=2023 URL=https://www.frontiersin.org/journals/climate/articles/10.3389/fclim.2022.1082026 DOI=10.3389/fclim.2022.1082026 ISSN=2624-9553 ABSTRACT=

Indian Ocean dipole (IOD) is one of the dominant modes of interannual variability in the Indian Ocean, which has global climate impacts and thus is one of the key targets of seasonal predictions. In this study, based on a century-long seasonal hindcast experiment from the Coupled Seasonal Forecasts of the 20th century (CSF-20C), we show that the prediction skill for IOD exhibits remarkable decadal variations, with low skill in the early-to-mid 20th century but high skill in the second half of the 20th century. The decadal variations of prediction skills for IOD are caused by two factors. The first is associated with the decadal variation of the ENSO-IOD relationship. Although individual members of the predictions can simulate the variation of the ENSO-IOD relationship, with amplitude close to that in the observation, the feature is greatly suppressed in the ensemble mean due to the asynchrony of variation phases among individual members. In the ensemble mean, the IOD evolution shows an unrealistic stable and high correlation with ENSO evolution. This causes the prediction to have much higher skill for those periods during which IOD is accompanied by ENSO in the observation. The second factor is associated with the decadal variation of IOD predictability in the prediction system. In the prediction system, the decadal variation of IOD signal strength closely follows that of ENSO signal strength. Meanwhile, the IOD noise strength shows variations opposite to the IOD signal strength. As a result, the signal-to-noise ratio greatly increases in the second half of the 20th century due to the enhancement of the ENSO signal strength, which represents the increase of IOD predictability in the prediction system.