Assessing Changes in the Contribution of Main Predictors to the El Niño Forecast in Recent Decades on a Simple Statistical ModelстатьяПереводИсследовательская статья
Статья опубликована в журнале из списка Web of Science и/или Scopus
Аннотация:Based on a simple statistical model, the change in the contribution of various predictors to the forecast of two types of the El Niño phenomenon is estimated. It is shown that the most significant predictor, the heat content of the active layer of the ocean, has been losing its predictive value in recent decades, which is consistent with other studies. At the same time, accounting for components of intraseasonal tropical variability,such as Madden–Julian Oscillation and Rossby waves, can improve the quality of El Niño forecast by almost a quarter. In addition to a general assessment of the statistical model quality, its ability to predict SST anomalies for specific El Niño is tested and the contribution of various predictors to the forecast of individual events is also assessed. It is shown that a decrease in the role of heat content also manifests itself for the forecast of specific El Niño events of both types—in some cases, other predictors turn out to be more important, and sometimes taking into account the heat content worsens instead of improves the forecast.