Deep learning for skillful long-lead ENSO forecasts

El Niño-Southern Oscillation (ENSO) is one of the fundamental drivers Earth's climate variability. Thus, its skillful prediction at least a few months to years ahead utmost importance society. Using both dynamical and statistical methods, several studies reported ENSO predictions various lead times. Predictions with long times, on other hand, remain difficult. In this study, we propose convolutional neural network (CNN)-based system heterogeneous CNN parameters for each season modified loss function predict 18–24 ahead. The developed indicates that model highly in predicting times high skills extreme events compared Scale Interaction Experiment-Frontier ver. 2 (SINTEX-F2) systems. analysis can overcome spring barrier, major hindrance systems, improvement skill partly be attributed seasonal models used study also use model. attempted identify precursors using heatmap analysis.