Interpretable Cross-Sphere Multiscale Deep Learning Predicts ENSO Skilfully Beyond 2 Years
This addresses the problem of long-term climate forecasting for meteorologists and policymakers, offering interpretable insights into ocean-atmosphere interactions, though it appears incremental in combining existing concepts like physics-encoding with neural networks.
The paper tackles the challenge of predicting El Niño-Southern Oscillation (ENSO) beyond one year by introducing PTSTnet, an interpretable deep learning model that unifies dynamical processes and cross-scale learning, achieving predictions significantly outperforming state-of-the-art benchmarks with lead times beyond 24 months.
El Niño-Southern Oscillation (ENSO) exerts global climate and societal impacts, but real-time prediction with lead times beyond one year remains challenging. Dynamical models suffer from large biases and uncertainties, while deep learning struggles with interpretability and multi-scale dynamics. Here, we introduce PTSTnet, an interpretable model that unifies dynamical processes and cross-scale spatiotemporal learning in an innovative neural-network framework with physics-encoding learning. PTSTnet produces interpretable predictions significantly outperforming state-of-the-art benchmarks with lead times beyond 24 months, providing physical insights into error propagation in ocean-atmosphere interactions. PTSTnet learns feature representations with physical consistency from sparse data to tackle inherent multi-scale and multi-physics challenges underlying ocean-atmosphere processes, thereby inherently enhancing long-term prediction skill. Our successful realizations mark substantial steps forward in interpretable insights into innovative neural ocean modelling.