Spectral Element science updates/bug fixes/refactor (aka CESM2.2-updates)

[jtruesdal]

These mods are necessary to fix science bugs, improve algorithms and numerical stability.

Detailed list of mods:

• remove Rayleigh friction in SE dycore (remove namelist and associated code; not used) (src/dynamics/se/dyn_comp.F90, src/dynamics/se/dycore/prim_advance_mod.F90)
• change various namelist changes for time-stepping
  • move outfld of PMID so it is output for dry configurations (src/physics/cam/cam_diagnostics.F90)
• bug fixes for WACCM-x in dp_coupling for species dependent thermodynamics
• remove vertical diffusion code in SE dycore (not used) - (src/dynamics/se/dycore/prim_advance_mod.F90)
• change some idealized tracer fields for dynamical core test suite (cam/physics/tracer.F90)
• PGF (Pressure gradient force) discretization change (use of reference profile) to alleviate noise over steep orography (src/dynamics/se/dycore/prim_advance_mod.F90)
• changes to hyperviscosity on temperature to alleviate noise over steep orography (src/dynamics/se/dycore/viscosity_mod.F90)
• sponge layer changes for stability (increase del4 damping in sponge, i.e. introduce level dependent damping); (src/dynamics/se/dycore/global_norms_mod.F90; src/dynamics/se/dycore/dimensions_mod.F90)
• bug fixes in physconst.F90 for WACCM-x and SE (mid pressure calculation consistency)
• change namelist defaults so that hi-res and var-res support configurations (ne120, ARCTIC, ARCTCIGRIS, CONUS) aren't running with unnecessarily small CLUBB time-steps
• fix initialization bug in angular momentum diagnostics (src/physics/cam/check_energy.F90)
• print more info to log file (src/dynamics/se/dyn_comp.F90)

All regression tests using SE are expected to fail against baseline.

[jtruesdal]

@PeterHjortLauritzen Regression tests for SMS_D_Ln9.ne16_ne16_mg17.QPX2000.cheyenne_intel.cam-outfrq9s SMS_D_Ln9.ne16_ne16_mg17.FX2000.cheyenne_intel.cam-outfrq9s failing when /glade/p/cgd/amp/jet/collections/CAM-1.CESM2.2_cam6_3_024/src/dynamics/se/dycore/prim_advance_mod.F90:770 is passing negative temperature to get_molecular_diff_coef (/glade/p/cgd/amp/jet/collections/CAM-1.CESM2.2_cam6_3_024/src/utils/physconst.F90)

blowing up at 1842 in physconst.F90 when trying to raise a negative number to a power.

[PeterHjortLauritzen]

I talked to Jim and Marianna and we propose to have the bug fixes and updates be part of the CESM2.3 release rather than updating the CESM2.2 release branch (which would be a lot of work). It will be important to point users to the CESM2.3 release for any spectral-element applications.

[gold2718]

This issue was not automatically closed by #401 because the closing syntax was incorrect (did not follow the GitHub rules).