[Mitgcm-support] Adams-Bashforth initialization

[mitgcm-support at dev.mitgcm.org]

I have a little patch that I use for the initial balanced
states (ie. solid body rotation and Rossby-Hauritz) where the
Adams-Bashforth time-stepping is initialized with a forward step.

When I ran the test-script with this patch I was surprised to
see that some expts still passed! I can understand the T,S
fields being only mildly affected but the solver and flow should
be on a different trajectory in all the expts?

              T     S     U     V
C D M    c                        
n p a R  g  m  m  m  m  m  m  m  m
f n k u  2  i  a  i  a  i  a  i  a
g d e n  d  n  x  n  x  n  x  n  x

Y Y Y Y 16 16 16 16 16 16 16 16  4 pass  adjustment.128x64x1
Y Y Y Y 16 16 16 16 16 16 16 16 16 pass  adjustment.cs-32x32x1
Y Y Y Y 13 16 16 16 16 16 16 13 16 pass  aim.5l_LatLon
Y Y Y Y  0 16 16 16 16  0  0  0  0 FAIL  exp0
Y Y Y Y  1  5  4 16 16  0  0  0  0 FAIL  exp1
Y Y Y Y  0  1  1  3  4  0  0  0  0 FAIL  exp2
Y Y Y Y  0  2  1 16 16  0  2  2  3 FAIL  exp4
Y Y Y Y  0  2 16 16 16  0  0  0  0 FAIL  exp5
Y Y Y Y  0  5  5  2  4  0  0  0  0 FAIL  front_relax
Y Y Y Y  5  9  8 16 16  3  4  7  7 FAIL  hs94.128x64x5
Y Y Y Y 13 16 16 16 16 11 16 16 13 pass  hs94.1x64x5
Y Y Y Y 14 16 16 16 16 16 13 16 13 pass  hs94.cs-32x32x5
Y Y Y Y  6 16 16  0 16 16 16 16 16 FAIL  internal_wave
Y Y Y Y  0  5  3  6  5  0  0  0  0 FAIL  natl_box

Isn't it strange that hs94.cs-32x32x5 is unaffected by
an initial Euler forward step but hs94.128x64x5 is?

To use a Campin expression, "it is curious, mmm?"

A.


The changes:

RCS file: /u/gcmpack/models/MITgcmUV/model/src/timestep.F,v
retrieving revision 1.24
diff -r1.24 timestep.F
40,43c40,43
< Caja  IF (myIter .EQ. 0) THEN
< Caja   ab15=1.0
< Caja   ab05=0.0
< Caja  ELSE
---
>       IF (myIter .EQ. 0) THEN
>        ab15=1.0
>        ab05=0.0
>       ELSE
46c46
< Caja  ENDIF
---
>       ENDIF

RCS file: /u/gcmpack/models/MITgcmUV/model/src/timestep_tracer.F,v
retrieving revision 1.2
diff -r1.2 timestep_tracer.F
34,35c34,40
<       ab15 = 1.+ tauAB
<       ab05 = -tauAB
---
>       IF (myIter .EQ. 0) THEN
>        ab15=1.0
>        ab05=0.0
>       ELSE
>        ab15=1.+tauAB
>        ab05=-tauAB
>       ENDIF

RCS file: /u/gcmpack/models/MITgcmUV/pkg/generic_advdiff/timestep_tracer.F,v
retrieving revision 1.2
diff -r1.2 timestep_tracer.F
35,36c35,41
<       ab15 = 1.+ tauAB
<       ab05 = -tauAB
---
>       IF (myIter .EQ. 0) THEN
>        ab15=1.0
>        ab05=0.0
>       ELSE
>        ab15=1.+tauAB
>        ab05=-tauAB
>       ENDIF