Minor update to facilitate multi-model coupling

[arango]

• The internal CPP option ATM_COUPLING is renamed to FRC_COUPLING since it is possible to couple to an atmospheric model or a data model (DATA_COUPLING) for the same forcing fields. It is done to consolidate with the coupling research branch that uses the ESMF/NUOPC library.

• Introduced CPP option ROMS_STDOUT to redirect ROMS standard output information to file log.roms. It coupling is advantageous to separate the standard output information of all the coupled Earth System Model (ESM) components to different log files. It is easier to track the solution of each ESM component when they are separated. In inp_par.F, we now have:

  #ifdef ROMS_STDOUT
  !
  !  Change default Fortran standard out unit, so ROMS run information is
  !  directed to a file. This is advantageous in coupling applications to
  !  ROMS information separated from other models.
  !
         stdout=20                      ! overwite Fortran default unit 6
  !
         OPEN (stdout, FILE='log.roms', FORM='formatted',                 &
       &       STATUS='replace')
  #endif
  

  Changed several routines to replace PRINT statements with WRITE (stdout,*).

• Update ROMS Nonlinear model driver nl_ocean.h to provide the correct information to time stepping kernel during model coupling:

  !
  !  Local variable declarations.
  !
  #if defined MODEL_COUPLING && !defined MCT_LIB
        logical, save :: FirstPass = .TRUE.
  #endif
        integer :: ng
  #if defined MODEL_COUPLING && !defined MCT_LIB
        integer :: NstrStep, NendStep
  #endif
        real (r8) :: MyRunInterval
  !
  !-----------------------------------------------------------------------
  !  Time-step nonlinear model over all nested grids, if applicable.
  #if defined MODEL_COUPLING && !defined MCT_LIB
  !  On first pass, add a timestep to the coupling interval to account
  !  for ROMS kernel delayed delayed output until next timestep.
  #endif
  !-----------------------------------------------------------------------
  !
        MyRunInterval=RunInterval
        IF (Master) WRITE (stdout,'(1x)')
        DO ng=1,Ngrids
  #if defined MODEL_COUPLING && !defined MCT_LIB
          step_counter(ng)=0
          NstrStep=iic(ng)
          IF (FirstPass) THEN
            NendStep=NstrStep+INT((RunInterval+dt(ng))/dt(ng))
            IF (ng.eq.1) MyRunInterval=MyRunInterval+dt(ng) 
            FirstPass=.FALSE.
          ELSE
            NendStep=NstrStep+INT(MyRunInterval/dt(ng))
          END IF
          IF (Master) WRITE (stdout,10) 'NL', ng, NstrStep, NendStep
  #else
          IF (Master) WRITE (stdout,10) 'NL', ng, ntstart(ng), ntend(ng)
  #endif
        END DO
        IF (Master) WRITE (stdout,'(1x)')
  !
  !$OMP PARALLEL
  #ifdef SOLVE3D
        CALL main3d (MyRunInterval)
  #else
        CALL main2d (MyRunInterval)
  #endif
  !$OMP END PARALLEL
  

  Recall that MyRunInterval may span the full period of the simulation, a multi-model coupling interval (RunInterval > ifac*dt), or just a single step (RunInterval=0). The zero value here is valid and can be explained by ROMS design of delayed output.