| 9 |
or they will not be read correctly by MITgcm: |
or they will not be read correctly by MITgcm: |
| 10 |
|
|
| 11 |
cd run |
cd run |
| 12 |
for i in 1 2 3 4 5 ; do |
for i in 1 2 3 4 5 6; do |
| 13 |
ln -s ../input/llc_p90_${i}.mitgrid tile00${i}.mitgrid |
ln -s ../input/llc_p90_${i}.mitgrid tile00${i}.mitgrid |
| 14 |
done |
done |
| 15 |
|
|
| 16 |
|
Change topology: |
| 17 |
|
|
| 18 |
|
cd MITgcm/utils/exch2/matlab-topology-generator/ |
| 19 |
|
|
| 20 |
|
modify matlab script driver_llc.m |
| 21 |
|
line |
| 22 |
|
nr = 360; ng = 90; nb = 90; tnx=90;tny=90; |
| 23 |
|
to (e.g.) |
| 24 |
|
nr = 360; ng = 90; nb = 90; tnx=30;tny=30; |
| 25 |
|
run script |
| 26 |
|
|
| 27 |
|
This generates routine and header |
| 28 |
|
w2_e2setup.F |
| 29 |
|
W2_EXCH2_TOPOLOGY.h |
| 30 |
|
which need to go in code/ |
| 31 |
|
|
| 32 |
|
Accordingly, bathy.0??.0??.data need changing. |
| 33 |
|
|
| 34 |
|
N.B. Domain decomposition nPy not active with exch2, i.e. |
| 35 |
|
possible decomps are (for orig. with 17 tiles): |
| 36 |
|
sNx = 90 30 90 30 |
| 37 |
|
sNy = 90 30 90 30 |
| 38 |
|
nSx = 17 17 1 1 |
| 39 |
|
nSy = 1 1 1 1 |
| 40 |
|
nPx = 1 9 17 153 |
| 41 |
|
nPy = 1 1 1 1 |
| 42 |
|
|
| 43 |
|
I've generated artificial tile #18 with all zeros |
| 44 |
|
blanklist.txt needs to be generated by hand. |
| 45 |
|
Not clear how many procs to invoke in presence of blank list. |
| 46 |
|
|
| 47 |
Setup code_15lev/, input_15lev/ |
Setup code_15lev/, input_15lev/ |
| 48 |
------------------------------- |
------------------------------- |
| 49 |
|
|
| 50 |
Real bathymetry, 15 vertical levels, |
* Real bathymetry, 15 vertical levels, |
| 51 |
monthly climatological forcing via exf package. |
* monthly climatological forcing via exf package, uses |
| 52 |
|
4x4 deg. fields with on-the-fly interpolation |
| 53 |
* Bathymetry is generated via script ... in ... |
* Bathymetry is generated via script ... in ... |
| 54 |
* Initial conditions for T, S are generated via script ... in ... |
* Initial conditions for T, S are generated via script ... in ... |
| 55 |
|
|
| 57 |
(which also contains link pattern to standard tile00${i}.mitgrid fields), |
(which also contains link pattern to standard tile00${i}.mitgrid fields), |
| 58 |
then link to run directory. |
then link to run directory. |
| 59 |
|
|
| 60 |
|
Setup code_50lev/, input_50lev/ |
| 61 |
|
------------------------------- |
| 62 |
|
|
| 63 |
|
* Based on ETOPO2 bathy, |
| 64 |
|
* 50 vertical level exactly matching high_res_cube (ECCO2), incl. partial cell params. |
| 65 |
|
* forced by 6-hourly NCEP atmos. state, uses original reduced Gaussian grid |
| 66 |
|
and exf interpolation on-the-fly |
| 67 |
|
* initial conditions generated via scripts in |
| 68 |
|
MITgcm_contrib/eh3/llc/ecco-godae/climatology/ |
| 69 |
|
resulting fields are |
| 70 |
|
llc90x90x360_50lev_ptemp.bin, llc90x90x360_50lev_salt.bin |
| 71 |
|
these fields are run are run with a diffusion/conv.adj. setup of MITgcm |
| 72 |
|
for 10 days to remove some instabilities |
| 73 |
|
* input fields can be obtained at |
| 74 |
|
http://channel.mit.edu/datasets/MITgcm_input_files/MITgcm_contrib/eh3/llc/ecco-godae/data_50lev/ |
| 75 |
|
|