Need new subroutine for time variables

[jgipsl]

The time related variables in ORCHIDEE needs to be put together in a new subroutine. The calculation of EndOfYear, FirstTsOfYear, LastTsDay should be done at the same time as sec and month are updated, done in call intsurf_time. The following part in slowproc should be put elsewhere :

    IF ( sec == dt_sechiba .AND. month==1 .AND. day==1 ) THEN
       ! The current time step is the first sechiba timestep of a new year
       IF (printlev_loc>=4) WRITE(numout,*) "This is a new day and a new year: month, day, sec=", month, day, sec
       FirstTsYear=.TRUE.
    ELSE
       FirstTsYear=.FALSE.
    END IF
    
    IF ( sec == 0 ) THEN
       ! The current time step is the last sechiba time step on a day
       LastTsDay=.TRUE.
       
       IF ( month == 1 .AND. day == 1 ) THEN
          ! The current time step is the last sechiba time step on a year
          ! JG : note that month=1, day=1, sec=O is the last day of the year. 
          !      This is due to a problem before the first call to slowproc. 
          !      slowproc_main enters the first time on the 2nd time step (1800s)
          EndOfYear = .TRUE.
          IF (printlev_loc>=4) WRITE(numout,*) "This is the last sechiba time step of a year, EndOfYear is activated"
       ELSE
          EndOfYear = .FALSE.
       END IF
    ELSE
       LastTsDay = .FALSE.
       EndOfYear = .FALSE.
    END IF

There is a time shift introduced in with the call to intsurf_time(from intersurf). The first calculated time step is sec=1800s. The last time step of the day is sec=0 and the date of the day already turned into the new date. This behaviour needs also to be revised at the same time. See comment JG above.

[jgipsl]

The new time module has been introduced in the trunk rev [4646]. No changes for dim2_driver or coupled mode with LMDZ. Each time-step is now described by the time interval(variables are indexed _start or _end).

Done by J Polcher, J Ghattas