Multi-Year ENSO Forecasts Using Parallel Convolutional Neural Networks With Heterogeneous Architecture

El Ni˜no-Southern Oscillation (ENSO), as one of the main drivers of Earth’s interannual climate variability, can cause a wide range of climate anomalies, and hence the multi-year ENSO forecasts is deemed as a paramount scientific issue. However, most of the existing works relying on the conventional iterative mechanism fail to achieve an accurate long-term prediction due to the error accumulation. The methods based on deep learning (DL) apply the parallel modeling scheme for different lead times instead of a single iteration model, but this scheme leverages the same DL model for prediction, which can not fully mine the variability of different lead times, resulting in a decrease of the prediction accuracy. To solve this problem, we propose a novel parallel deep convolutional neural network employed with heterogeneous architecture. Here, by adaptively selecting network architectures for different lead times, we realize variability modeling of different tasks (lead times) and finally improve the reliability of long-term prediction. Furthermore, we propose a relationship between different prediction lead times and neural network architecture from a unique perspective, namely the receptive field originally proposed in computer vision. According to the spatial-temporal correlated area and sampling scale of lead times, the size of convolution kernel and the mesh size of sampling are adjusted as the lead time increases. Experimental results demonstrate that our proposed method outperforms the other well-known methods, especially when long-term prediction is implemented.