Changeset 2126 for trunk/LMDZ.COMMON/libf/dyn3d/integrd.F

Timestamp:

Apr 26, 2019, 11:18:52 AM (6 years ago)

Author:

emillour

Message:

Common dynamics:
Some work to enforce total tracer mass conservation in the dynamics.
Still to be further studied and validated.
For now these changes are triggered by setting a "force_conserv_tracer"
flag to ".true." in run.def (default is ".false." to not change anything
with respect to previous versions).
When force_conserv_tracer=.true. then:

1. Rescale tracer mass in caladvtrac after tracer advection computations
2. Recompute q ratios once atmospheric mass has been updated in integrd

These steps technically ensure total tracer mass conservation but it
might be the tracer advection scheme and/or time-stepping updating
sequence of fields that should be rethought or fixed.
EM

File:

• trunk/LMDZ.COMMON/libf/dyn3d/integrd.F (modified) (3 diffs)

trunk/LMDZ.COMMON/libf/dyn3d/integrd.F

@@ -7 +7 @@
      &  )
 
-      use control_mod, only : planet_type
+      use control_mod, only : planet_type,force_conserv_tracer
       USE comvert_mod, ONLY: ap,bp
       USE comconst_mod, ONLY: pi
@@ -66 +66 @@
       REAL vscr( ip1jm ),uscr( ip1jmp1 ),hscr( ip1jmp1 ),pscr(ip1jmp1)
       REAL massescr( ip1jmp1,llm )
+      REAL :: massratio(ip1jmp1,llm)
 !      REAL finvmasse(ip1jmp1,llm)
       REAL p(ip1jmp1,llmp1)
@@ -242 +243 @@
 
       endif ! of if (planet_type.eq."earth")
+
+      if (force_conserv_tracer) then
+        ! Ehouarn: try to keep total amont of tracers fixed 
+        ! by acounting for mass change in each cell
+        massratio(1:ip1jmp1,1:llm)=massescr(1:ip1jmp1,1:llm)
+     &                             /masse(1:ip1jmp1,1:llm)
+        do iq=1,nq
+        q(1:ip1jmp1,1:llm,iq)=q(1:ip1jmp1,1:llm,iq)
+     &                        *massratio(1:ip1jmp1,1:llm)
+        enddo
+      endif ! of if (force_conserv_tracer)
 c
 c