A Machine-Learning-only GCMΒΆ

Near surface temperature, sea-level pressure, and wind. As simulated by a machine learning system trained on 20CRv2c data.

The GCM ML model produces fields of near surface temperature, sea-level pressure, and wind, at 6-hourly intervals (as well as the associated latent-space states. So we can make a video analagous to the compressed 20CRv2c version but using the ML GCM output.