1 | WARNING: For now, the test case ICE_ADV2D only works if the logical "ll_neg" is set to FALSE in the routine icedyn_adv_umx.F90 |
---|
2 | (it is still unclear why) |
---|
3 | ------- |
---|
4 | Purpose |
---|
5 | ------- |
---|
6 | This demonstration case can serve different purposes: |
---|
7 | |
---|
8 | 1) The main one: study of the advection of a patch of ice in a bi-periodic channel with a slab ocean (i.e. 1 ocean layer) |
---|
9 | ==> set ln_icethd=false in namelist_ice_cfg |
---|
10 | |
---|
11 | 2) Secondary: study of ice thermodynamics in the same basin |
---|
12 | ==> set ln_icedyn=false in namelist_ice_cfg |
---|
13 | ==> eventually change usrdef_sbc.F90 in MY_SRC to fit your needs |
---|
14 | |
---|
15 | ----------- |
---|
16 | Experiments |
---|
17 | ----------- |
---|
18 | Two experiments can be configured: |
---|
19 | |
---|
20 | a) A simple channel at 3km horizontal resolution (slightly variable => +- 10%) |
---|
21 | ==> in MY_SRC/usrdef_hgr.F90, uncomment the part "variable scale factors" and comment part "constant scale factors" |
---|
22 | |
---|
23 | b) A simple channel at 3km horizontal resolution (constant scale factors) |
---|
24 | ==> in MY_SRC/usrdef_hgr.F90, comment the part "variable scale factors" and uncomment part "constant scale factors" |
---|
25 | |
---|
26 | ---------- |
---|
27 | How to run |
---|
28 | ---------- |
---|
29 | |
---|
30 | a) Compile and run the model once to get a mesh_mask.nc file with the following command: |
---|
31 | ../../../makenemo -r ICE_ADV2D -n ICE_ADV2D -m X64_ADA -j 4 |
---|
32 | mpirun ./nemo -np 1 |
---|
33 | |
---|
34 | b) Create the initial condition file for sea-ice (initice.nc) by running this python script: |
---|
35 | python ./make_INITICE.py |
---|
36 | |
---|
37 | c) Run the model a second time |
---|
38 | mpirun ./nemo -np 4 |
---|
39 | |
---|
40 | --------------- |
---|
41 | What to look at |
---|
42 | --------------- |
---|
43 | In case of purpose 1, One can test |
---|
44 | --- |
---|
45 | a) the advection scheme: Ultimate-Macho (ln_adv_UMx=T) versus Prather (ln_adv_Pra=T) |
---|
46 | for a square (ice concentration) or a gaussian (ice volume) |
---|
47 | with either a constant velocity (ln_dynADV2D=T) |
---|
48 | or a constant ice-atm. stress, thus velocity is calculated by rheology (ln_dynRHGADV=T) |
---|
49 | with 1 or 5 ice categories (jpl=1 or 5). |
---|
50 | (note that ln_dynADV2D=T only works with jpl=1) |
---|
51 | |
---|
52 | In case of purpose 2, one can test conservation of properties: |
---|
53 | --- |
---|
54 | b) ice should not change at all if surface fluxes = 0 and SST = freezing temperature |
---|
55 | |
---|
56 | -------------------------------------- |
---|
57 | Interpretation of the results (remarks) |
---|
58 | -------------------------------------- |
---|
59 | - Prather conserves the max values but also creates side lobes |
---|
60 | - UM does not conserve the max but does not create side lobes |
---|
61 | - The "unmoving" ice (if any) is due to ice rheology which states that ice mass below a certain thresold (1kg/m2) is considered to move at the ocean velocity (thus 0 m/s) |
---|