Create NetCDF module for writing standalone CRTM outputs

[chengdang]

Create NetCDF module for writing standalone CRTM outputs

[chengdang]

Attempting to use RT solution as the only input (without atmosphere data structure) for CRTM_RTSolution_WriteFile_netCDF:

  FUNCTION CRTM_RTSolution_WriteFile_netCDF( &
    Sensor_Filename   , &  ! Input
    RTSolution        , &  ! Input
    Quiet             , &  ! Optional input
    Debug             ) &  ! Optional input (Debug output control)
  RESULT( err_stat )

Noticed CRTM_RTSolution_type%n_Layers has never been assigned in CRTM calculations. This variable is needed for layer outputs. Modified Common_RTSolution.F90 to assign CRTM_RTSolution_type%n_Layers in modules Assign_Common_Output*.

Updates Aug 13:
CRTM_RTSolution_type%n_Layers should be assigned when calling CRTM_RTSolution_Create( rts, N_LAYERS )

Once this is added in check_crtm.F90, there is no need to reassign this variable. Restored all changes in module Assign_Common_Output

[chengdang]

Initial code updates have been pushed to branch: https://github.com/JCSDA/CRTMv3/tree/feature/cd_RTSout_netCDF

This branch allows users to output CRTM RT Solution in netCDF files, for example, see

CRTMv3/src/Build/libsrc/test/check_crtm.F90

Line 374 in 1ee69fd

! 8d. **** OUTPUT THE RESULTS TO NetCDF files ****

New CRTM_RTSolution_WriteFile_netCDF module writes RT solution of the same sensor (multi channels) and same profile to one single file. For multiple profiles, it generates multiple files.

[chengdang]

Remaining work:

• 2-D array outputs need to be added:

  CRTMv3/src/RTSolution/CRTM_RTSolution_Define.f90

  Line 1395 in 1ee69fd

  ! CD: Not working, need to come back and check the size of these RT 2D arrays

• Consider adding wavelength?
• Consider adding angles as global attributes?

[chengdang]

Layered outputs have been added in module CRTM_RTSolution_WriteFile_netCDF. For this module to work properly, the users must call CRTM_RTSolution_Create to allocate n_Layers and layered outputs. Check_CRTM.f90 is updated to reflect such requirement: commit 1ee69fd

The following variables are not added to RTS solution for now, since:

• Wavelength is stored in the spc coeff file.
• Input angles should be part of the input profiles.

Closing this issue for now.