Bug in ENT (energy conserving at T-point) scheme

[jchanut]

Context

Newly implemented ENT scheme.

Analysis

2d computation of coriolis force in barotropic mode stepping is wrong.
One has:

         CASE( np_ENT )             ! energy conserving scheme (t-point)
         DO jj = 2, jpjm1
            DO ji = 2, jpim1   ! vector opt.

               z1_hu = ssumask(ji,jj) / ( hu_0(ji,jj) + zhup2_e(ji,jj) + 1._wp - ssumask(ji,jj) )
               z1_hv = ssvmask(ji,jj) / ( hv_0(ji,jj) + zhvp2_e(ji,jj) + 1._wp - ssvmask(ji,jj) )

               zu_trd(ji,jj) = + r1_4 * r1_e1e2u(ji,jj) * z1_hu                   &
                  &               * (  e1e2t(ji+1,jj)*zhtp2_e(ji+1,jj)*ff_t(ji+1,jj) * ( va_e(ji+1,jj) + va_e(ji+1,jj-1) )   &
                  &                  + e1e2t(ji  ,jj)*zhtp2_e(ji  ,jj)*ff_t(ji  ,jj) * ( va_e(ji  ,jj) + va_e(ji  ,jj-1) )   )
                  !
               zv_trd(ji,jj) = - r1_4 * r1_e1e2v(ji,jj) * z1_hv                    &
                  &               * (  e1e2t(ji,jj+1)*zhtp2_e(ji,jj+1)*ff_t(ji,jj+1) * ( ua_e(ji,jj+1) + ua_e(ji-1,jj+1) )   &
                  &                  + e1e2t(ji,jj  )*zhtp2_e(ji,jj  )*ff_t(ji,jj  ) * ( ua_e(ji,jj  ) + ua_e(ji-1,jj  ) )   )
            END DO
         END DO

However, inverse depths z1_hu, z1_hv hold roughly twice their correct values. One has certainly thought that zhup2_e and zhvp2_e are depths anomalies but these are total depths.

Fix

One should therefore have:

               z1_hu = ssumask(ji,jj) / ( zhup2_e(ji,jj) + 1._wp - ssumask(ji,jj) )
               z1_hv = ssvmask(ji,jj) / ( zhvp2_e(ji,jj) + 1._wp - ssvmask(ji,jj) )

[jchanut]

In 10741:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found

[jchanut]

In 10742:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found