Role of the Western Pacific Heat Buildup for the Timing, Magnitude and Prediction of ENSO

The theoretical predictability of El Niño-Southern Oscillation (ENSO) is on the order of years, but despite the continued efforts to improve its prediction, long-term forecasts are still limited. One reason is that many ENSO models have not incorporated the subsurface ocean dynamics in the most efficient way targeted at long-lead forecasting. At the same time, the ongoing provision of high-resolution data by the TAO-TRITON array since the end of the TOGA Program in 1994 has allowed to track in detail the buildup of heat in the western tropical Pacific subsurface, which typically occurs about two years before the peak of warm events. We incorporate this dynamically-relevant information in the form of predictor variables of temperature at different depths and regions of the equatorial ocean, as well as wind stress in the central Pacific in a flexible statistical dynamic components model to retrospectively forecast the Niño3.4 Index in the period 1970-2016. The use of specific regions and depths, rather than the traditional integration of heat content, has allowed us to significantly extend in time the predictive capacity of our model. It successfully predicts all the major EN episodes, including the recent extreme 2015/16 El Niño up to two and a half years in advance. We also discover that the events are predicted much more accurately after the completion of the observational array in the tropical Pacific in 1994.
Furthermore, through observational records and numerical experiments we explore the sensitivity of El Niño to the magnitude of the heat buildup in the ocean subsurface 21 months in advance. Within the framework of the experiments, the coupled ocean-atmosphere system compensates any initial decrease in heat content and evolves towards a new recharge and a delay of one year in the occurrence of a warm event. The results point to a non-linear dependence between the intensity of the subsurface heat buildup and the magnitude and timing of the subsequent El Niño.