Scientific validation of the dynamical core (GEOS-FP configuration)

[FlorianDeconinck]

The dynamical core has already been vadliated for an Aquaplanet and a HeldSuarez configuration. But those validation where done on a very specific, and sometime hard coded, set of parameters for the dynamics.

We will revalidate the core once we have raised the skill of the code to match the GEOS-FP needs.

For reference, previously we used:

• Fortran 6 month spin up then 12 month analysis on fortran/ndsl
• U/V average over 12 months
• Temperature average over 12 months

Parent: #83

---

• Metrics for validation are (re)define and documented here
• Validation for at least 1 CPU and 1 GPU backend
• Validation for at least 2 CXXX configuration