Bug in trcsbc.F90

[cetlod]

Context

The passive tracer surface boundary condition trend of flux through the interface ( concentration/dilution effect plus freezing/melting of sea-ice ) is not properly computed

Analysis

• In non linear free-surface, the emp budget does not affect the passive tracer at surface (through the concentration dilution effect) since the change in passive tracer will be induced by volume change.

The only change to be consider is the flux of freezing/melting coming from the possibly presence of the passive tracer in sea-ice : fmmflx * trc_i ( trc_i is the concentration of the tracer in sea-ice )

• In linear free surface case (ln_linssh=T), the volume of the ocean does not change with the water exchanges at the (air+ice)-sea interface. Therefore another term has to be added, to mimic the concentration/dilution effect associated with water exchanges : emp * trn

Note that in case of passive tracers run offline, the water flux variable to be consider must include the freezing/melting terms to be consistent with the coupling online where emp contains fmmflx

Recommendation

In each case of presence of tracer in sea-ice ( trc_i= 0 ; trc_i = trn ; trc_i =/ trn ), the flux associated to surface boundary condition must be computed as follow

      CASE ( 0 )  ! Same concentration in sea ice and in the ocean ( trc_i = trn )
         !
         DO jn = 1, jptra
            DO jj = 2, jpj
               DO ji = fs_2, fs_jpim1   ! vector opt.
                  sbc_trc(ji,jj,jn) = - fmmflx(ji,jj) * r1_rau0 * trn(ji,jj,1,jn)
               END DO
            END DO
         END DO
         !
         IF( ln_linssh ) THEN  !* linear free surface  
            DO jn = 1, jptra
               DO jj = 2, jpj
                  DO ji = fs_2, fs_jpim1   ! vector opt.
                     sbc_trc(ji,jj,jn) = sbc_trc(ji,jj,jn) + r1_rau0 * emp(ji,jj) * trn(ji,jj,1,jn) !==>> add concentration/dilution effect due to constant volume cell
                  END DO
               END DO
            END DO
         ENDIF

[cetlod]

In 15393:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found

[cetlod]

In 15394:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found