Changeset 10364 for NEMO/trunk/src/OCE/TRA/trazdf.F90

Timestamp:

2018-11-30T18:42:51+01:00 (5 years ago)

Author:

acc

Message:

Introduce Adaptive-Implicit vertical advection option to the trunk. This is code merged from branches/2018/dev_r9956_ENHANCE05_ZAD_AIMP (see ticket #2042). The structure for the option is complete but is currently only successful with the flux-limited advection scheme (ln_traadv_mus). The use of this scheme with flux corrected advection schemes is not recommended until improvements to the nonoscillatory algorithm have been completed (work in progress elsewhere). The scheme is activated via a new namelist switch (ln_zad_Aimp) and is off by default.

File:

• NEMO/trunk/src/OCE/TRA/trazdf.F90 (modified) (2 diffs)

NEMO/trunk/src/OCE/TRA/trazdf.F90

@@ -136 +136 @@
       !
       INTEGER  ::  ji, jj, jk, jn   ! dummy loop indices
-      REAL(wp) ::  zrhs             ! local scalars
+      REAL(wp) ::  zrhs, zzwi, zzws ! local scalars
       REAL(wp), DIMENSION(jpi,jpj,jpk) ::  zwi, zwt, zwd, zws
       !!---------------------------------------------------------------------
@@ -177 +177 @@
             !
             ! Diagonal, lower (i), upper (s)  (including the bottom boundary condition since avt is masked)
-            DO jk = 1, jpkm1
-               DO jj = 2, jpjm1
-                  DO ji = fs_2, fs_jpim1   ! vector opt.
-!!gm BUG  I think, use e3w_a instead of e3w_n, not sure of that
-                     zwi(ji,jj,jk) = - p2dt * zwt(ji,jj,jk  ) / e3w_n(ji,jj,jk  )
-                     zws(ji,jj,jk) = - p2dt * zwt(ji,jj,jk+1) / e3w_n(ji,jj,jk+1)
-                     zwd(ji,jj,jk) = e3t_a(ji,jj,jk) - zwi(ji,jj,jk) - zws(ji,jj,jk)
-                 END DO
-               END DO
-            END DO
+            IF( ln_zad_Aimp ) THEN         ! Adaptive implicit vertical advection
+               DO jk = 1, jpkm1
+                  DO jj = 2, jpjm1
+                     DO ji = fs_2, fs_jpim1   ! vector opt. (ensure same order of calculation as below if wi=0.)
+                        zzwi = - p2dt * zwt(ji,jj,jk  ) / e3w_n(ji,jj,jk  )
+                        zzws = - p2dt * zwt(ji,jj,jk+1) / e3w_n(ji,jj,jk+1)
+                        zwd(ji,jj,jk) = e3t_a(ji,jj,jk) - zzwi - zzws   &
+                           &                 + p2dt * ( MAX( wi(ji,jj,jk  ) , 0._wp ) - MIN( wi(ji,jj,jk+1) , 0._wp ) )
+                        zwi(ji,jj,jk) = zzwi + p2dt *   MIN( wi(ji,jj,jk  ) , 0._wp )
+                        zws(ji,jj,jk) = zzws - p2dt *   MAX( wi(ji,jj,jk+1) , 0._wp )
+                    END DO
+                  END DO
+               END DO
+            ELSE
+               DO jk = 1, jpkm1
+                  DO jj = 2, jpjm1
+                     DO ji = fs_2, fs_jpim1   ! vector opt.
+                        zwi(ji,jj,jk) = - p2dt * zwt(ji,jj,jk  ) / e3w_n(ji,jj,jk)
+                        zws(ji,jj,jk) = - p2dt * zwt(ji,jj,jk+1) / e3w_n(ji,jj,jk+1)
+                        zwd(ji,jj,jk) = e3t_a(ji,jj,jk) - zwi(ji,jj,jk) - zws(ji,jj,jk)
+                    END DO
+                  END DO
+               END DO
+            ENDIF
             !
             !! Matrix inversion from the first level