Improved Predictability of the Indian Ocean Dipole Using Seasonally Modulated ENSO Forcing Forecasts
Despite recent progress in seasonal forecast systems, the predictive skill for Indian Ocean Dipole (IOD) remains typically limited to a lead time of one season or less in both dynamical and empirical models. We developed a new simple stochastic‐dynamical model (SDM) to predict the IOD using seasonally modulated El Niño–Southern Oscillation (ENSO) forcing together with a seasonally modulated Indian Ocean coupled ocean‐atmosphere feedback. The SDM demonstrates considerably improved IOD forecast skill compared to current operational models.
Since IOD events have large socioeconomic and environmental impacts globally, predicting them has become a scientific challenge of considerable importance. Our findings provide a novel powerful perspective for improving IOD predictions. Importantly, the proposed new model framework can be easily adopted operationally and will allow seasonal forecast centers to translate their skillful ENSO forecasts into much improved predictions of the IOD at significantly longer lead times. These longer lead times are critical for policymakers to be able to implement mitigation measures. Furthermore, our results imply that potential future ENSO improvements in models should also directly translate into more skillful IOD predictions.
In the two related papers on the subject, we developed and validated a simple stochastic-dynamical model (SDM) to predict the IOD using the forecasted ENSO conditions. The forecast skill of the SDM is compared to operational North American Multi-model Ensemble (NMME) models. We found that the new SDM constitutes a large improvement compared to current operational IOD predictions. The forecasted IOD–ENSO relationship deteriorates away from the observed relationship with increasing lead time, which might be one reason that limits the IOD predictive skill in coupled models. Furthermore, our findings clearly demonstrate that operational IOD predictability beyond persistence is largely controlled by ENSO predictability and the signal-to-noise ratio of the system.