Skip to content

ML-model trend representation bottleneck

Home / Open Questions / ML-model trend representation bottleneck

Background: Data-driven Earth system models currently struggle with systematic biases in prognostic fields that exhibit persistent, long-term trends, such as global mean sea-level rise.

Question / Future Work: It remains an unresolved challenge to effectively train machine-learning models to represent oceanic variables that contain strong, non-stationary climate signals, such as absolute sea surface height. Investigating whether training on detrended anomalies or employing alternative normalization strategies can mitigate the tendency of such models to collapse toward a mean historical climatological state is a key area for future development.

Why It Matters: Addressing this is critical for the long-term reliability of data-driven Earth system models, as failure to represent trends correctly limits their utility for climate change studies and long-term projection.

Metadata & Links