Determine which features to port from current implementation places

[sgreenbury]

Currently we have implementations in a few different locations relating to spatiotemporal data:

• Explore wrapping the well models, metrics, dataset and utils autoemulate#828
• SpatioTemporalSimulation class autoemulate#827
• Improve API for SpatioTemporal emulators autoemulate#816
• Integrate the well training approaches with the current FNO example for reaction/advection-diffusion autoemulate#834
• Support warm-starting for optimization-based emulator models autoemulate#306

Current places we've begun implementations:

• PRs:
  • Add initial emulator base class for The Well autoemulate#840
  • Explore extending TheWellEmulator API autoemulate#863
• https://github.com/alan-turing-institute/spatio_temporal_forecasting
• physicsnemo:
  • https://github.com/alan-turing-institute/autoemulate/tree/855-856-physicsnemo-explore
  • https://github.com/alan-turing-institute/spatio_temporal_forecasting/tree/physicsnemo