Abstract:The El Niño–Southern Oscillation (ENSO) is an extremely complicated ocean–atmosphere coupling event, the development and decay of which are usually modulated by the energy interactions between multiple physical variables. In this paper, we design a multivariate air–sea coupler (ASC) based on the graph using features of multiple physical variables. On the basis of this coupler, an ENSO deep learning forecast model (named ENSO-ASC) is proposed, whose structure is adapted to the characteristics of the ENSO dynamics, including the encoder and decoder for capturing and restoring the multi-scale spatial–temporal correlations, and two attention weights for grasping the different air–sea coupling strengths on different start calendar months and varied effects of physical variables in ENSO amplitudes. In addition, two datasets modulated to the same resolutions are used to train the model. We firstly tune the model performance to optimal and compare it with the other state-of-the-art ENSO deep learning forecast models. Then, we evaluate the ENSO forecast skill from the contributions of different predictors, the effective lead time with different start calendar months, and the forecast spatial uncertainties, to further analyze the underlying ENSO mechanisms. Finally, we make ENSO predictions over the validation period from 2014 to 2020. Experiment results demonstrate that ENSO-ASC outperforms the other models. Sea surface temperature (SST) and zonal wind are two crucial predictors. The correlation skill of the Niño 3.4 index is over 0.78, 0.65, and 0.5 within the lead time of 6, 12, and 18 months respectively. From two heat map analyses, we also discover the common challenges in ENSO predictability, such as the forecasting skills declining faster when making forecasts through June–July–August and the forecast errors being more likely to show up in the western and central tropical Pacific Ocean in longer-term forecasts. ENSO-ASC can simulate ENSO with different strengths, and the forecasted SST and wind patterns reflect an obvious Bjerknes positive feedback mechanism. These results indicate the effectiveness and superiority of our model with the multivariate air–sea coupler in predicting ENSO and analyzing the underlying dynamic mechanisms in a sophisticated way.

ENSO Forecasts with Spatiotemporal Fusion Transformer Network

Spatial-temporal transformer network for multi-year ENSO prediction

Enhancing ENSO predictions with self-attention ConvLSTM and temporal embeddings

Echo State Network Based on Phase Space Reconstruction: El Niño 3 Index Forecasting

ENSONet: a data-driven ENSO forecasting model with concise spatial location learning parameters and temporal embedding

Transformer for EI Niño-Southern Oscillation Prediction

Multi-scale Spatial Features and Temporal Attention Mechanisms: Advancing the Accuracy of ENSO Prediction

Adaptive Graph Spatial-Temporal Attention Networks for Long Lead ENSO Prediction

Temporal Convolutional Networks for the Advance Prediction of ENSO

El Nino Index Prediction Based on Deep Learning with STL Decomposition

A Residual Network with Geographical and Meteorological Attention for Multi-Year ENSO Forecasts

ENSO-ASC 1.0.0: ENSO Deep Learning Forecast Model with a Multivariate Air–sea Coupler

ENSO Forecasting over Multiple Time Horizons Using ConvLSTM Network and Rolling Mechanism

Multi-Task Spatial-Temporal Transformer for Multi-Variable Meteorological Forecasting

An investigation into improving El Niño-Southern Oscillation prediction based on temporal transformer architecture

A self-attention–based neural network for three-dimensional multivariate modeling and its skillful ENSO predictions

GSTCN: Graph-Based Spatial-Temporal Convolutional Networks for Advanced ENSO Forecasting

Tendency-and-attention-informed deep learning for ENSO forecasts

Deep learning with autoencoders and LSTM for ENSO forecasting

ENSO Predictive Analytics Based on Layered Rendering

Convolutional GRU Network for Seasonal Prediction of the El Niño-Southern Oscillation