Status of CAM-FV^3 Adiabatic Test Cases
  • The initial conditions cami_fvcubed_c46.nc were interpolated from FV inic file cami_0000-09-01_2x2.5_L26_c021002_USGS.nc. Insufficient resolution?
  • d_p_coupling is implemented (and tested for adiabatic mode).  p_d_coupling is implemented but not tested (also: tendencies must be interpolated from A grid to D grid).
  • restart files: read and write not yet implemented. Status of PIO integration?
  • dry air adjustment: not yet implemented (see discussion). In read_inidat?
  • Latest tag: fvcubed07_cam3_6_26: see instructions https://wiki.ucar.edu/pages/viewpage.action?pageId=37432298
Grid Staggering
  • Currently the initial file contains the unstaggered (A-grid) winds. These are read in and interpolated to D-grid locations (short-term fix).
  • D-grid locations not explicitly available, must be calculated as midpoint of B-grid locations
  • Minor issues with data decompositions of B-grid points; 1-2 days.
  • Brian's suggestion:

https://wiki.ucar.edu/display/ccsm/Grid+Staggering+Suggestion+for+Initial+Conditions+%28Eaton%29

Software Engineering Discussion Points
  • Getting rid of commap
  • Removal of #ifdef STAGGERED (and #ifdef SPMD)
  • Minor revisions to dyn_grid; is it sensible to use dyn_grid methods outside of dp_coupling?
  • CAM initialization: FVCUBED stepon_init will be much simpler than FV, for example.
  • dycore_is(): Add 'UNSTRUCTURED' as superset of HOMME, FVCUBED, ...?
  • What is the path for getting changes into the SVN trunk?
Minor problems
  • How to to splice US and VS at staggered locations with other unstaggered fields. Append with ncks?
  • PHIS not found in topology file (ncid_topo=20), must reasd PHIS from initial file.
  • Should be able to change CAM-FV^3 defaults by changing values in test/system/config_files/f10t5mdm, e.g. resolution to 'c46'. This does not work: one needs to change defaults in bld/config_files/defaults_fvcubed.xml. Why?
  • No labels