Tests C grid - Githubissues

JFLemieux73 commented 2 years ago

ncview.uvel.pdf ncview.uvelE.pdf ncview.vvel.pdf ncview.vvelN.pdf

JFLemieux73 commented 2 years ago

Components of rheology term strintx.pdf strintxE.pdf strinty.pdf strintyN.pdf

apcraig commented 2 years ago

To setup the channel case, do something like the following.

./cice.setup -m cheyenne -e intel -g gbox12 -p 1x1x12x12x1 -s boxsyme,diag1,run1day,gridb --test smoke --testid gD01

in ice_in, set

  histfreq    = ‘d’,‘1’,‘x’,‘x’,‘x’
  close_boundaries = .false.   ! not required
  ice_data_type  = ‘channel’
  ew_boundary_type = ‘cyclic’
  f_aice     = ‘d1’

Results in table form as follows. This is output at i=5 for all j (1:12). Results at each i location are identical. The channel case here is 12x12 with land at T points j=1,2,11,12 and aice=0.5 at j=4,5,6,7,8. No ice at j=3,9,10. Column 1 = B grid uvel Column 2 = C grid uvelE Column 3 = CD grid uvelE Column 4 = CD grid uvelN Column 5 = CD grid uvelE with pstar=0. Column 6 = CD grid uvelN with pstar=0.

            UVEL_1   UVELE_1  UVELE_1   UVELN_1  UVELE_1  UVELN_1
1    /  1:     ....     ....      ....     ....     ....     ....
2    /  2:     ....     ....      ....     ....     ....     ....
3    /  3:  0.08421  0.00000  0.000000  0.08421  0.00000  0.08421
4    /  4:  0.08421  0.08420  0.001599  0.08421  0.08421  0.08421
5    /  5:  0.08421  0.08420  0.002078  0.08421  0.08421  0.08421
6    /  6:  0.08421  0.08420  0.002238  0.08421  0.08421  0.08421
7    /  7:  0.08421  0.08420  0.002078  0.08421  0.08421  0.08421
8    /  8:  0.08421  0.08420  0.001599  0.08421  0.08421  0.08421
9    /  9:  0.00000  0.00000  0.000000  0.00000  0.00000  0.00000
10   / 10:  0.00000  0.00000  0.000000  0.00000  0.00000  0.00000
11   / 11:     ....     ....      ....     ....     ....     ....
12   / 12:     ....     ....      ....     ....     ....     ....

Same for v velocities,

              VVEL_1    VVELN_1    VVELE_1    VVELN_1  VVELE_1 VVELN_1
1    /  1:       ....       ....       ....       ....    ....    ....
2    /  2:       ....       ....       ....       ....    ....    ....
3    /  3: -1.420E-17 -6.481E-04  0.000E+00 -2.033E-19  0.0000  0.0000
4    /  4: -8.501E-18  1.780E-07  1.677E-04 -4.794E-19  0.0000  0.0000
5    /  5: -2.831E-18  9.961E-08  1.198E-04 -6.700E-20  0.0000  0.0000
6    /  6:  2.831E-18 -1.093E-07 -1.290E-18  1.572E-19  0.0000  0.0000
7    /  7:  8.501E-18 -1.574E-07 -1.198E-04  9.004E-19  0.0000  0.0000
8    /  8:  1.420E-17  6.481E-04 -1.677E-04  4.571E-19  0.0000  0.0000
9    /  9:  0.000E+00  0.000E+00  0.000E+00  0.000E+00  0.0000  0.0000
10   / 10:  0.000E+00  0.000E+00  0.000E+00  0.000E+00  0.0000  0.0000
11   / 11:       ....       ....       ....       ....    ....    ....
12   / 12:       ....       ....       ....       ....    ....    ....

The obvious issue is the decoupled CD solution associated with Column 3 in U. The other thing to note is the relatively poor performance of the V velocity in the C solution, Column 2 in V. Pstar=0 fixes the problems with the CD solution.

apcraig commented 2 years ago

Here are the latest tables for the same channel cases as above with ndte=1200 and a new column for VP results. Results are unchanged. VP "v" is not quite as zero as some other cases.

Column 1 = B grid uvel Column 2 = C grid uvelE Column 3 = CD grid uvelE Column 4 = CD grid uvelN Column 5 = CD grid uvelE with pstar=0. Column 6 = CD grid uvelN with pstar=0. Column 7 = B grid uvel with kdyn=3 (vp)

            UVEL_1   UVELE_1  UVELE_1   UVELN_1  UVELE_1  UVELN_1  UVEL_1
1    /  1:     ....     ....      ....     ....     ....     ....     ....
2    /  2:     ....     ....      ....     ....     ....     ....     ....
3    /  3:  0.08421  0.00000  0.000000  0.08421  0.00000  0.08421  0.08421
4    /  4:  0.08421  0.08420  0.001599  0.08421  0.08421  0.08421  0.08421
5    /  5:  0.08421  0.08420  0.002078  0.08421  0.08421  0.08421  0.08421
6    /  6:  0.08421  0.08420  0.002238  0.08421  0.08421  0.08421  0.08421
7    /  7:  0.08421  0.08420  0.002078  0.08421  0.08421  0.08421  0.08421
8    /  8:  0.08421  0.08420  0.001599  0.08421  0.08421  0.08421  0.08421
9    /  9:  0.00000  0.00000  0.000000  0.00000  0.00000  0.00000  0.00000
10   / 10:  0.00000  0.00000  0.000000  0.00000  0.00000  0.00000  0.00000
11   / 11:     ....     ....      ....     ....     ....     ....     ....
12   / 12:     ....     ....      ....     ....     ....     ....     ....

              VVEL_1    VVELE_1    VVELE_1    VVELN_1  VVELE_1 VVELN_1   VVEL_1
1    /  1:       ....       ....       ....       ....    ....    ....       ....
2    /  2:       ....       ....       ....       ....    ....    ....       ....
3    /  3: -1.999E-17 -3.240E-04  0.000E+00 -4.366E-19  0.0000  0.0000 -1.007E-11
4    /  4: -1.196E-17 -3.240E-04  1.677E-04 -1.903E-19  0.0000  0.0000 -1.117E-12
5    /  5: -3.983E-18  3.929E-08  1.198E-04 -1.449E-19  0.0000  0.0000 -1.134E-13
6    /  6:  3.983E-18 -4.227E-09  4.461E-18  1.590E-21  0.0000  0.0000  1.134E-13
7    /  7:  1.196E-17 -4.222E-08 -1.198E-04  2.518E-19  0.0000  0.0000  1.117E-12
8    /  8:  1.999E-17  3.240E-04 -1.677E-04  5.885E-19  0.0000  0.0000  1.007E-11
9    /  9:  0.000E+00  3.240E-04  0.000E+00  0.000E+00  0.0000  0.0000  0.000E+00
10   / 10:  0.000E+00  0.000E+00  0.000E+00  0.000E+00  0.0000  0.0000  0.000E+00
11   / 11:       ....       ....       ....       ....    ....    ....       ....
12   / 12:       ....       ....       ....       ....    ....    ....       ....

apcraig commented 2 years ago

If I force v=0 is step_vel for the channel CD case, v==0 in output and the u solutions are unchanged.

Baseline uE, vE, uN, vN:

            UVELE_1    VVELE_1   UVELN_1   VVELN_1
1    /  1:      ....       ....     ....       ....
2    /  2:      ....       ....     ....       ....
3    /  3:  0.000000  0.000E+00  0.08421 -4.366E-19
4    /  4:  0.001599  1.677E-04  0.08421 -1.903E-19
5    /  5:  0.002078  1.198E-04  0.08421 -1.449E-19
6    /  6:  0.002238  4.461E-18  0.08421  1.590E-21
7    /  7:  0.002078 -1.198E-04  0.08421  2.518E-19
8    /  8:  0.001599 -1.677E-04  0.08421  5.885E-19
9    /  9:  0.000000  0.000E+00  0.00000  0.000E+00
10   / 10:  0.000000  0.000E+00  0.00000  0.000E+00
11   / 11:      ....       ....     ....       ....
12   / 12:      ....       ....     ....       ....

v set to zero:

            UVELE_1  VVELE_1  UVELN_1 VVELN_1
1    /  1:      ....    ....     ....    ....
2    /  2:      ....    ....     ....    ....
3    /  3:  0.000000  0.0000  0.08421  0.0000
4    /  4:  0.001599  0.0000  0.08421  0.0000
5    /  5:  0.002078  0.0000  0.08421  0.0000
6    /  6:  0.002238  0.0000  0.08421  0.0000
7    /  7:  0.002078  0.0000  0.08421  0.0000
8    /  8:  0.001599  0.0000  0.08421  0.0000
9    /  9:  0.000000  0.0000  0.00000  0.0000
10   / 10:  0.000000  0.0000  0.00000  0.0000
11   / 11:      ....    ....     ....    ....
12   / 12:      ....    ....     ....    ....

apcraig commented 2 years ago

A few other results before I forget as I continue testing,

increasing the width of the ice in the channel by 1 gridcell does not change the CD results.
setting aice to 0.1 (was 0.5) brings the E solution much closer to the ideal solution. I think this is not surprising. We know setting pstar=0 results in correct answers, as p goes to zero, the E results gets closer to the correct answer.
adding low concentration ice in the channel problem where there was zero ice before results in correct E and N results! Concentrations as low as 0.005 work, leaving aice=0.5 in the main channel ice. This suggests the problem really is somehow related to how some of the equations are discretized at the ice edge.
setting stresspT=stressmT=stresspU=stressmU=c0 results in same wrong solution.

dabail10 commented 2 years ago

Here is the first subcycle: stress12T 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 stress12U 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 3.078947368421053E-004 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000

Second subcycle: stress12T 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 stress12U 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.863973461411587 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 -0.863973461411587 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 -5.399834133822417E-005 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 -5.399834133822417E-005 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 6.149286482003538E-004 6.157812421928050E-004 6.157812421928050E-004 6.157812421928050E-004 6.149286482003538E-004 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 6.157812421928050E-004 6.157812421928050E-004 6.157812421928050E-004 6.157812421928050E-004 6.157812421928050E-004 6.157812421928050E-004 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000

End of run: stress12T 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 stress12U 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 779.587470870799 467.699878980278
155.889843963600 -155.889843963600 -467.699878980278
-779.587470870798 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 -1.949297449315755E-002 -1.948812718854237E-002 -1.948623049545003E-002 -1.948812718854232E-002 -1.949297449315756E-002 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 strintxN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelE 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 1.598515885989748E-003 2.078016700778288E-003 2.237839921854726E-003 2.078016700778291E-003 1.598515885989755E-003 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 uvelN 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 8.421243560722592E-002 8.421243560722592E-002 8.421243560722592E-002 8.421243560722592E-002 8.421243560722592E-002 8.421243560722592E-002 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000

dabail10 commented 2 years ago

Here are the zeta/eta

ksub = 1zetax2T 0.000000000000000E+000 etax2T 0.000000000000000E+000 zetax2U 243.847278997721
etax2U 60.9618197494304
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 243.847278997721
etax2U 60.9618197494304

ksub=2 zetax2T 0.000000000000000E+000 etax2T 0.000000000000000E+000 zetax2U 243.847278997721
etax2U 60.9618197494304
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 243.847278997721
etax2U 60.9618197494304

end of run: zetax2T 0.000000000000000E+000 etax2T 0.000000000000000E+000 zetax2U 243.847278997721
etax2U 60.9618197494304
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 243.847278997721
etax2U 60.9618197494304

dabail10 commented 2 years ago

Fixed values.

zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861
zetax2T 487.694557995443
etax2T 121.923639498861
zetax2U 487.694557995443
etax2U 121.923639498861

apcraig commented 2 years ago

With fixes to viscous coeff U calc using capping, CD grid

           UVELE_1    VVELE_1   UVELN_1   VVELN_1
1    /  1:     ....       ....     ....       ....
2    /  2:     ....       ....     ....       ....
3    /  3:  0.00000  0.000E+00  0.08421 -1.879E-17
4    /  4:  0.08420  5.362E-08  0.08421 -1.699E-17
5    /  5:  0.08420  3.234E-08  0.08421 -7.977E-18
6    /  6:  0.08420 -4.004E-22  0.08421  7.977E-18
7    /  7:  0.08420 -3.234E-08  0.08421  1.699E-17
8    /  8:  0.08420 -5.362E-08  0.08421  1.879E-17
9    /  9:  0.00000  0.000E+00  0.00000  0.000E+00
10   / 10:  0.00000  0.000E+00  0.00000  0.000E+00
11   / 11:     ....       ....     ....       ....
12   / 12:     ....       ....     ....       ....

apcraig commented 2 years ago

For the C grid with update viscous coeffs,

           UVELE_1    VVELN_1
1    /  1:     ....       ....
2    /  2:     ....       ....
3    /  3:  0.00000 -6.481E-04
4    /  4:  0.08419 -1.158E-07
5    /  5:  0.08421  3.232E-08
6    /  6:  0.08421 -3.232E-08
7    /  7:  0.08421  1.158E-07
8    /  8:  0.08419  6.481E-04
9    /  9:  0.00000  0.000E+00
10   / 10:  0.00000  0.000E+00
11   / 11:     ....       ....
12   / 12:     ....       ....

apcraig commented 2 years ago

I have started to look at the boxsyme test for the B/C/CD cases. The default setup has advection and thermodynamics off just like our channel flow. It also has aice=1.0. The ice is basically "stuck". This produces a "stuck" solution and while the velocities in the three cases have similar properties, they are also quite different. I believe we are just looking at different numerics. If I change aice to 0.5, then I get free drift in the interior, zero velocity against the wall, and the solutions are a lot more similar except in the corners. I don't see how we can ever have identical results near the boundaries given the different staggers. I am running for 6 days and have increased ndte to 1200. I want to turn on advection next.

Should the boxsym tests be using aice=1.0 in general. Is that really the test we want? Should I continue to test aice=1.0 or is aice=0.5 better? Or do we really need to do both?

apcraig commented 2 years ago

I attach images of uvel and vvel for the B/C/CD results for the boxsyme test with aice=0.5 after 5 days with ndte=1200. The scales are the consistent. I am not plotting u=v=0 at the walls. Outside of the corners, you can see that the "change in color" is slightly different for each solution, but those differences are small. I think these look pretty good. Looking at some raw numbers, the E and N solutions seems to be quite consistent for the CD solution which is good. The one thing that does jump out is the solution on the left and right edge for the C grid especially noticeable in the v solution in the corners. The maximum magnitude is not right at the edge as it is for other solutions.

vvel_B vvel_C vvel_CD vvelN_C vvelN_CD vvelE_CD uvel_B uvel_C uvel_CD uvelE_C uvelE_CD uvelN_CD

eclare108213 commented 2 years ago

Should the boxsym tests be using aice=1.0 in general. Is that really the test we want? Should I continue to test aice=1.0 or is aice=0.5 better? Or do we really need to do both?

aice=1.0 is too stiff of a problem. aice=0.5 doesn't really test the stress terms, although it is useful for checking symmetry and similar kinds of debugging. aice=0.9 or a little higher could be okay. This is why the box2001 test is set up with a ramp in aice from 0 to 1 -- you get all the various kinds of behavior.

apcraig commented 2 years ago

Some other test results from today

the boxslotcyl is bit-for-bit for B/C/CD (as expected). dynamics is off, advection is on.
boxsyme tests with advection turned on. The B and CD look very close after 5 days, the C is a little different. But after 3 months, hi is still very similar between the 3. You can see differences, but I can't say which is better or worse.

hi_box_3mon_B hi_box_3mon_C hi_box_3mon_CD

-boxwallblockp5 (an aice=0.5 box on the east wall with east forcing) with advection after 3 months is as follows. Now we have an interesting case that is starting to show noise in both the C and CD cases. The block is being push against the wall and spreading. Maybe this is the next thing to debug?

hi_block_3mon_B hi_block_3mon_C hi_block_3mon_CD

eclare108213 commented 2 years ago

Wow, very interesting. How small are the values left behind (in purple)? Wondering if I can use this to debug residual ice issues.

apcraig commented 2 years ago

I believe the purple values are less than 0.01. I tried to enhance the shading there too, but it got smaller than I thought we cared about. That purple basically represents where the box started. We have eastward forcing and it mostly piles up and spreads on the east wall, but a tiny bit is left behind apparently.

dabail10 commented 2 years ago

I redid the kmtislands (CD) case and printed some values. On the second sub-cycle, I get:

uvelE 2.472165410902706E-004 (i=50,j=33) uvelE 2.472165410902706E-004 (i=49,j=33) uvelE 2.472165410902706E-004 (i=51,j=33) uvelE 0.000000000000000E+000 (i=50,j=34) uvelE 0.000000000000000E+000 (i=50,j=32) stresspU -397.468827643619
stresspU -397.468827643619
stresspU -397.468827643619
stresspU 0.000000000000000E+000 stresspU -397.468827643619
stress12U -99.3672069109048
stress12U -99.3672069109048
stress12U -99.3672069109048
stress12U 0.000000000000000E+000 stress12U 99.3672069109048

So, doing some back of the envelope calculations, assuming vvel = 0.

`` divU = 0. tensionU = 0. shearU = 16000. (0. - 2.47x10^-4) = 3.952 deltaU = 3.952 tmpcalc =~ 20000./3.952 =~ 5060. zetax2U = 5060. etax2U = 0.255060. = 1265. rep_prsU =~ 20000.

stresspU =~ -20000. stress12U =~ 0.51265.3.952 `` This assumes a strength of 20000. It is probably less than this. Anyhow, do these numbers make sense for a one-gridpoint wide channel? Should there be a stresspU / stress12U value at the land gridcell below, but not above?

JFLemieux73 commented 2 years ago

Hi Dave,

I think we should get nonzero values for stresspU and stress12U both above and below.

I am working on a special test case for one-gridpoint wide channels. I think we can get an analytical solution. It will be easier to understand what is going on.

JFLemieux73 commented 2 years ago

There are analytical solutions for one grid cell channels with cyclic BCs. Have a look at the pdf below. For strong winds there is a plastic solution. For weak winds we can also find a viscous solution. It also allows us to calculate the shear stresses at the walls.

1cell_channel.pdf

JFLemieux73 commented 2 years ago

Hi Tony,

About the figures above...Have you tried to reduce the time step to see if you still get the noise in the C (and CD) solutions?

apcraig commented 2 years ago

I have not tried reduced timestep. I assume that means ndte=1200? I can give that a try.

JFLemieux73 commented 2 years ago

Sorry I meant the advective time step. It is probably set to one hour. I would try 15min.

JFLemieux73 commented 2 years ago

Hi Tony,

As you can see we have decided to create one issue per test case. You could go ahead and create issues for the other test cases you are testing (e.g. boxwallblockp5). It would be nice if the first comment of each issue was the command for the setup (and possibly things that need to be modified in ice_in). Finally maybe we could close this issue. Thanks!

apcraig commented 2 years ago

Closing, see #684 #686 #687 #688 #689

CICE-Consortium / CICE

Tests C grid #682