195 |
mkdir exp3b |
mkdir exp3b |
196 |
mv *tave.0000000001.data exp3b |
mv *tave.0000000001.data exp3b |
197 |
|
|
198 |
Use matlab script lookat_exp3.m to compare |
Use matlab script lookat_exp3.m to look at |
199 |
the output of exp2 to that of exp1: |
results from exp3a and exp3b: |
200 |
cd ../../../verification/lab_sea/matlab |
cd ../../../verification/lab_sea/matlab |
201 |
matlab |
matlab |
202 |
lookat_exp3 |
lookat_exp3 |
203 |
|
|
204 |
If the dynamic ice solvers are correct, they should converge to a |
If the dynamic ice solvers are correct, they should converge to a |
205 |
constant solution. The tests above show that both the ADI and the LSR |
solution that is constant at each latitude. |
|
solver are unable to handle periodic domains and therefore that they |
|
|
cannot be "correctly" parallelized. |
|
206 |
|
|
207 |
|
|
208 |
Instructions for running Experiment 4 |
Instructions for running Experiment 4 |
230 |
Also as sea-ice ages (thickens) the difference at the tiles becomes |
Also as sea-ice ages (thickens) the difference at the tiles becomes |
231 |
increasingly small. For solver pkg/seaice/adi.F a value of NPSEUDO=10 |
increasingly small. For solver pkg/seaice/adi.F a value of NPSEUDO=10 |
232 |
appears adequate for forward integrations but cannot be used for sea-ice |
appears adequate for forward integrations but cannot be used for sea-ice |
233 |
adjoint model. Work is underway to fix this problem. |
adjoint model. With NPSEUDO=300 or greater, the tile disappears, but the |
234 |
|
computational cost is very large (11 times that of teh forward model). |
235 |
|
|
236 |
|
|
237 |
Instructions for running Experiment 5 |
Instructions for running Experiment 5 |