Deep Learning Atmospheric Models Reliably Simulate Out-of-Sample Land Heat and Cold Wave Frequencies
This work addresses the reliability of fast deep learning climate simulators for extreme events, which is crucial for climate scientists and policymakers, but it is incremental as it builds on existing DL-GCM methods.
The study evaluated deep learning-based general circulation models (NGCM and DLESyM) against a conventional model (HiRAM) for simulating land heatwaves and coldwaves under out-of-sample early-20th-century conditions, finding that both DL models generalized successfully with skill comparable to HiRAM, though DLESyM overestimated frequencies due to excessive temperature autocorrelation.
Deep learning (DL)-based general circulation models (GCMs) are emerging as fast simulators, yet their ability to replicate extreme events outside their training range remains unknown. Here, we evaluate two such models -- the hybrid Neural General Circulation Model (NGCM) and purely data-driven Deep Learning Earth System Model (DL\textit{ESy}M) -- against a conventional high-resolution land-atmosphere model (HiRAM) in simulating land heatwaves and coldwaves. All models are forced with observed sea surface temperatures and sea ice over 1900-2020, focusing on the out-of-sample early-20th-century period (1900-1960). Both DL models generalize successfully to unseen climate conditions, broadly reproducing the frequency and spatial patterns of heatwave and cold wave events during 1900-1960 with skill comparable to HiRAM. An exception is over portions of North Asia and North America, where all models perform poorly during 1940-1960. Due to excessive temperature autocorrelation, DL\textit{ESy}M tends to overestimate heatwave and cold wave frequencies, whereas the physics-DL hybrid NGCM exhibits persistence more similar to HiRAM.