# Advanced Long-term Earth System Forecasting

arXiv: https://arxiv.org/abs/2505.19432 · HTML: https://mlsanity.com/paper/2505.19432/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 May 2025.

## Built on by (most cited first, 9 of 9)

- [Spatiotemporal Forecasting as Planning: A Model-Based Reinforcement Learning Approach with Generative World Models](https://mlsanity.com/paper/2510.04020/) (2025)
- [Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting](https://mlsanity.com/paper/2605.06916/) (2026)
- [OMNIFLOW: A Physics-Grounded Multimodal Agent for Generalized Scientific Reasoning](https://mlsanity.com/paper/2603.15797/) (2026)
- [Learning from History: A Retrieval-Augmented Framework for Spatiotemporal Prediction](https://mlsanity.com/paper/2510.24049/) (2025)
- [Advancing Ocean State Estimation with efficient and scalable AI](https://mlsanity.com/paper/2511.06041/) (2025)
- [DIFFUMA: High-Fidelity Spatio-Temporal Video Prediction via Dual-Path Mamba and Diffusion Enhancement](https://mlsanity.com/paper/2507.06738/) (2025)
- [VISION: Prompting Ocean Vertical Velocity Reconstruction from Incomplete Observations](https://mlsanity.com/paper/2509.21477/) (2025)
- [Unlocking Out-of-Distribution Generalization in Dynamics through Physics-Guided Augmentation](https://mlsanity.com/paper/2510.24216/) (2025)
- [Decision-Oriented Uncertainty Quantification for Risk Control in Earth System Spatiotemporal Foundation Models](https://mlsanity.com/paper/2609.14821/) (2026)

## Abstract (arXiv)

Reliable long-term forecasting of Earth system dynamics is fundamentally limited by instabilities in current artificial intelligence (AI) models during extended autoregressive simulations. These failures often originate from inherent spectral bias, leading to inadequate representation of critical high-frequency, small-scale processes and subsequent uncontrolled error amplification. Inspired by the nested grids in numerical models used to resolve small scales, we present TritonCast. At the core of its design is a dedicated latent dynamical core, which ensures the long-term stability of the macro-evolution at a coarse scale. An outer structure then fuses this stable trend with fine-grained local details. This design effectively mitigates the spectral bias caused by cross-scale interactions. In atmospheric science, it achieves state-of-the-art accuracy on the WeatherBench 2 benchmark while demonstrating exceptional long-term stability: executing year-long autoregressive global forecasts and completing multi-year climate simulations that span the entire available 2500-day test period without drift. In oceanography, it extends skillful eddy forecast to 120 days and exhibits unprecedented zero-shot cross-resolution generalization. Ablation studies reveal that this performance stems from the synergistic interplay of the architecture's core components. TritonCast thus offers a promising pathway towards a new generation of trustworthy, AI-driven simulations. This significant advance has the potential to accelerate discovery in climate and Earth system science, enabling more reliable long-term forecasting and deeper insights into complex geophysical dynamics.