Warning, /verification/lab_sea/README.md is written in an unsupported language. File is not indexed.
view on githubraw file Latest commit 7fe51dd4 on 2025-03-14 19:25:07 UTCd8c5b89513 Ivan*0001 Labrador Sea Region with Sea-Ice
0002 =========================================
0003
7fe51dd492 Jean*0004 ### Primary test Overview:
d8c5b89513 Ivan*0005 This example sets up a small (20x16x23) Labrador Sea experiment
0006 coupled to a dynamic thermodynamic sea-ice model (MITgcm Documentation 8.6.2).
0007
0008 The domain of integration spans $`[280, 320]^\circ`$E and $`[46, 78]^\circ`$N.
0009 Horizontal grid spacing is 2 degrees.
0010 The 23 vertical levels and the bathymetry file
0011
0012 ```
0013 bathyFile = 'bathy.labsea1979'
0014 ```
0015 are obtained from the the 2$`^\circ`$ ECCO configuration.
0016
0017 Integration is initialized from annual-mean Levitus climatology
0018
0019 ```
0020 hydrogThetaFile = 'LevCli_temp.labsea1979'
0021 hydrogSaltFile = 'LevCli_salt.labsea1979'
0022 ```
0023
0024 Surface salinity relaxation is to the monthly mean Levitus climatology
0025
0026 ```
0027 saltClimFile = 'SSS.labsea1979'
0028 ```
0029
0030 Forcing files are a 1979-1999 monthly climatology computed from the
7fe51dd492 Jean*0031 NCEP reanalysis (see [`SEAICE_PARAMS.h`](https://github.com/MITgcm/MITgcm/blob/master/pkg/seaice/SEAICE_PARAMS.h) for units and signs)
d8c5b89513 Ivan*0032
0033 ```
0034 uwindFile = 'u10m.labsea1979' # 10-m zonal wind
0035 vwindFile = 'v10m.labsea1979' # 10-m meridional wind
0036 atempFile = 'tair.labsea1979' # 2-m air temperature
0037 aqhFile = 'qa.labsea1979' # 2-m specific humidity
0038 lwdownFile = 'flo.labsea1979' # downward longwave radiation
0039 swdownFile = 'fsh.labsea1979' # downward shortwave radiation
0040 precipFile = 'prate.labsea1979' # precipitation
0041 ```
0042
0043 The experiment uses `pkg/gmredi`, `pkg/kpp`, `pkg/seaice`, and `pkg/exf`.
7fe51dd492 Jean*0044 The test is a 1-cpu, 10-hour integration. Both the atmospheric
d8c5b89513 Ivan*0045 state and the open-water surface fluxes are provided by `pkg/exf`.
0046
0047 More `pkg/seaice` test experiments, configured for low and
0048 high-resolution global cube-sphere domains are described
0049 in `MITgcm_contrib/high_res_cube/README_ice`.
0050
7fe51dd492 Jean*0051 ### Lab Sea adjoint
d8c5b89513 Ivan*0052 The `code_ad` directory provides files required to compile the adjoint
0053 version of this verification experiment. This verification
0054 experiment uses the 'divided adjoint'.
0055
0056 To compile the adjoint, one must enable the divided adjoint with the
0057 compile-time flag `USE_DIVA`, the location of which is specified in
0058 the file `build/genmake_local`.
0059 To wit,
0060
0061 ```
0062 USE_DIVA=1
0063 ```
0064
0065 To compile the adjoint without the divided adjoint, the compile-time
0066 flag `ALLOW_DIVIDED_ADJOINT` in `code_ad/AUTODIFF_OPTIONS.h` should
0067 be changed from
0068
0069 ```
0070 #define ALLOW_DIVIDED_ADJOINT
0071 ```
0072 to
0073
0074 ```
0075 #undef ALLOW_DIVIDED_ADJOINT
0076 ```
0077
0078 Note: `testreport` builds in the `lab_sea/build` directory which contains
0079 this `genmake_local` file and so it knows to use the divided adjoint.
0080
0081 ## Instructions
0082 Navigate to experiment directory
0083
0084 ```
0085 cd MITgcm/verification/lab_sea
0086 ```
0087
0088 ### 1-CPU forward experiment
0089 Configure and compile the code:
0090 ```
0091 cd build
0092 ../../../tools/genmake2 -mods ../code [-of my_platform_optionFile]
7fe51dd492 Jean*0093 [make Clean]
d8c5b89513 Ivan*0094 make depend
0095 make
0096 cd ..
0097 ```
0098
0099 To run:
0100 ```
0101 cd run
0102 ln -s ../input/* .
0103 ln -s ../build/mitgcmuv .
0104 ./mitgcmuv > output.txt
0105 cd ..
0106 ```
0107
0108 There is comparison output in the directory:
0109 ```
0110 results/output.txt
0111 ```
0112
0113 Use matlab script `lookat_ice.m` to compare the output
0114 with that from `checkpoint51f` sea-ice code:
0115 ```
0116 cd ../../../verification/lab_sea/matlab
0117 matlab
0118 lookat_ice
0119 ```
0120
0121 ### 2-CPU forward experiment
0122 Configure and compile the code:
0123 ```
0124 cd build
0125 ../../../tools/genmake2 -mpi -mods ../code [-of my_platform_optionFile]
0126 ln -s ../code/SIZE.h_mpi SIZE.h
7fe51dd492 Jean*0127 [make Clean]
d8c5b89513 Ivan*0128 make depend
0129 make
0130 cd ..
0131 ```
0132
0133 To run:
0134 ```
0135 cd run
0136 ln -s ../input/* .
0137 mpirun -np 2 ../build/mitgcmuv
0138 cd ..
0139 ```
0140
0141 ### 1-CPU adjoint experiment
0142 Configure and compile the code:
0143 ```
0144 cd build
0145 ../../../tools/genmake2 -mods ../code_ad [-of my_platform_optionFile]
0146 make adall
0147 cd ..
0148 ```
0149
0150 To run:
0151 ```
0152 cd run
0153 ln -s ../input_ad/* .
7fe51dd492 Jean*0154 ../input_ad/prepare_run
d8c5b89513 Ivan*0155 ln -s ../build/mitgcmuv_ad .
0156 ./do_run.sh
0157 cd ..
0158 ```
0159
7fe51dd492 Jean*0160 **Note:** `prepare_run` shell script is also used when running `testreport` (see below)
0161 and could be replaced by these 2 commands:
0162 ```
0163 ln -s ../input/* .
0164 ln -s ../../isomip/input_ad/ones_64b.bin .
0165 ```
0166 And the overly simple shell script "do_run.sh" just executes four times
d8c5b89513 Ivan*0167 (as specified in file "run_ADM_DIVA", `add_DIVA_runs = 4`) `mitgcmuv_ad`, saving
0168 output in intermediate files "output_adm.txt.diva_0,1,2,3", plus a final time:
0169 ```
0170 mitgcmuv_ad > output_adm.txt
0171 ```
0172 where output file `output_adm.txt` can be compared with reference output:
0173 ```
0174 results/output_adm.txt
0175 ```
0176
7fe51dd492 Jean*0177 ## Secondary tests
0178 In addition to the primary tests described above, 5 secondary forward tests and
0179 and 2 secondary adjoint tests can be run using the same executable as the corresponding
0180 primary tests but with specific input parameter files (in `input.$st\` and `input_ad.$st\`).
0181 The secondary forward tests include alternative seaice model formulations:
0182 free-drift in `input.fd/`; using EVP and `useHB87stressCoupling` in `input.hb87/` ;
0183 with `pkg/salt_plume` in `input.salt_plume/`;
0184 and two other ice-free "North-Altlantic box" set-up (formerly in `verification/natl_box/`)
0185 in `input.natl_box\` and with `pkg/longstep` in `input.longstep/`.
0186 The secondary adjoint tests are simpler version of the primary adjoint test,
0187 without seaice in `input_ad.noseaice/` and without seaice dynamics in `input_ad.noseaicedyn/`.
0188
0189 ### Instruction to run secondary tests
d8c5b89513 Ivan*0190 Run the testscript _forward_ experiments:
0191
0192 ```
0193 cd MITgcm/verification
0194 ./testreport -t lab_sea [-of my_platform_optionFile]
0195 ```
0196
0197 Standard testreport output, with all secondary tests:
0198 ```
0199 default 10 ----T----- ----S----- ----U----- ----V----- --PTR 01-- --PTR 02-- --PTR 03-- --PTR 04-- --PTR 05--
0200 G D M c m s m s m s m s m s m s m s m s m s
0201 e p a R g m m e . m m e . m m e . m m e . m m e . m m e . m m e . m m e . m m e .
0202 n n k u 2 i a a d i a a d i a a d i a a d i a a d i a a d i a a d i a a d i a a d
0203 2 d e n d n x n . n x n . n x n . n x n . n x n . n x n . n x n . n x n . n x n .
0204
7fe51dd492 Jean*0205 Y Y Y Y>11<13 16 16 16 16 16 16 14 13 13 13 16 16 16 12 14 22 16 16 16 22 16 16 16 pass lab_sea
0206 Y Y Y Y>12<16 16 16 16 16 16 16 16 16 13 13 16 14 16 13 16 22 16 16 16 22 16 16 16 pass lab_sea.fd
0207 Y Y Y Y> 4< 9 10 9 9 16 13 11 8 6 8 4 5 7 8 4 6 22 7 7 7 22 6 7 7 FAIL lab_sea.hb87
0208 Y Y Y Y 11 16 16 16 14 16 16 16 16 16 13 12 14 16 13 12 14 16 16 16>16<pass lab_sea.longstep
0209 Y Y Y Y>11<16 16 16 16 16 16 16 16 13 12 12 14 14 13 12 14 pass lab_sea.natl_box
0210 Y Y Y Y>13<16 16 16 16 16 16 16 14 16 14 13 14 16 14 13 16 22 16 16 16 22 16 16 16 pass lab_sea.salt_plume
d8c5b89513 Ivan*0211 ```
0212
0213 **Note:** Some differences in accuracy occur across different platforms as seen
0214 here for secondary test "lab_sea.hb87".
0215
0216 Run the testscript _adjoint_ experiments:
0217
0218 ```
0219 cd MITgcm/verification
0220 ./testreport -t lab_sea -ad [-of my_platform_optionFile]
0221 ```
0222
0223 Standard adjoint testreport output, with all secondary tests:
0224 ```
7fe51dd492 Jean*0225 Adjoint generated by TAF Version 6.5.1
d8c5b89513 Ivan*0226
0227 default 10 ----T----- ----S----- ----U----- ----V-----
0228 G D M C A F m s m s m s m s
0229 e p a R o d D m m e . m m e . m m e . m m e .
0230 n n k u s G G i a a d i a a d i a a d i a a d
0231 2 d e n t r r n x n . n x n . n x n . n x n .
0232
0233 Y Y Y Y 16>16<16 16 14 16 16 16 16 16 16 16 16 16 14 14 16 16 16 pass lab_sea (e=0, w=0, lfd=1, dop=1, sm=1)
0234 Y Y Y Y 14>15< 6 16 16 16 13 16 13 13 11 13 13 12 13 11 11 12 12 pass lab_sea.noseaice
0235 Y Y Y Y 16>13<16 16 16 14 16 16 13 13 16 16 16 16 14 14 16 14 16 pass lab_sea.noseaicedyn
0236 ```
