Earth system model learns to predict impacts of user changes on ecosystems
Earth System World Model for What-If Simulations: A Case Study for Terrestrial Ecosystems
Machine LearningArtificial Intelligence
Summary
Earth system simulations help scientists understand how ecosystems work, but they usually just follow preset rules and cannot respond to what-if questions like "What if a forest grows faster?". The authors created a new machine learning method that can learn from existing ecosystem data and respond when users make changes to parts of the system. This model can predict how ecosystems might behave when certain conditions are changed, without needing extra labeled data. Their tests show the model can make reliable long-term predictions and handle complex interactions between ecosystem components, making it useful for interactive scientific studies.
earth system simulationmachine learning emulatoraction-conditioned modelingecosystem dynamicsstate-transition learningdigital twincontrollable interventionsmasked response learninglong-horizon predictionecosystem cycles
Authors
Zhihao Wang, Ruichen Wang, Ruohan Li, Lei Ma, George Hurtt, Xiaowei Jia, Gengchen Mai, Shaowen Wang, Yiqun Xie
Abstract
Machine learning emulators have become essential for accelerating expensive Earth-system simulations, but most existing approaches remain passive forecasters: they reproduce simulator trajectories under prescribed forcings without an explicit interaction mechanism for user-specified interventions. This limits their use in interactive scientific workflows and Earth-system digital twins, where users often need to explore how a system would respond if selected state components were changed. We propose an action-conditioned world-modeling framework for Earth-system emulation that reformulates simulator trajectories as supervision for controllable state-transition learning. The key idea is transition-action pretraining: naturally observed state changes are treated as label-free action supervision, allowing the model to learn both prescribed dynamics and action-conditioned responses without manually annotated interventions. We further introduce masked response learning to infer unobserved variables under partial state edits and learn coupled system dependencies. We test this framework on ecosystem dynamics across six global regions and multiple stand ages. Experiments show that the model preserves competitive long-horizon emulation accuracy while enabling controllable structural interventions and coherent responses in coupled ecosystem-cycle variables. These results suggest a practical route from passive Earth-system emulators toward interactive, intervention-aware scientific surrogates.