1 |
heimbach@mit.edu 07-Nov-2002. |
2 |
This verification experiment is almost identical to the |
3 |
global_ocean.90x40x15 experiment, except that it uses |
4 |
the exf (and cal) package instead of the |
5 |
external_fields_load routine. |
6 |
To obtain identical results between the two experiments, |
7 |
the following 2 modifications are necessary: |
8 |
1. in external_fields_load.F: |
9 |
replace the line |
10 |
Imytm=int(myTime*rdt+0.5) |
11 |
by |
12 |
Imytm=int((myTime+1296000.)*rdt+0.5) |
13 |
2. in exf_set_climsst.F: |
14 |
comment the 'quality' check, i.e. avoid the setting |
15 |
to freezing temperature if less than climtempfreeze. |
16 |
|
17 |
================================ |
18 |
|
19 |
menemenlis@jpl.nasa.gov 05-Aug-2003 |
20 |
|
21 |
On-the-fly spatial interpolation has been added to pkg/exf. |
22 |
It is enabled by defining CPP option USE_EXF_INTERPOLATION |
23 |
in EXF_CPPOPTIONS.h or in ECCO_CPPOPTIONS.h |
24 |
Both bi-linear and bi-cubic (#define BICUBIC) are supported. |
25 |
|
26 |
This package is a placeholder until more general coupler |
27 |
is made available by ESMF project. It assumes that input |
28 |
grid is Cartesian with arbitrary latitude increments in the |
29 |
y-direction but with equidistant longitude increments in |
30 |
the x-direction. The input grid must encompass the complete |
31 |
output grid in the y-direction (i.e., extrapolation is not |
32 |
supported) and it is assumed that the grid is periodic in |
33 |
the x-direction. |
34 |
|
35 |
The verification/global_with_exf experiment has been modified |
36 |
to use the USE_EXF_INTERPOLATION option, the input grids |
37 |
being defined by variables *_lon0, *_lon_inc, *_lat0, *_lat_inc |
38 |
*_nlon and *_nlat in input/dat.exf and input/data.exf_clim |
39 |
|
40 |
*_lon0, *_lat0 :: lon and lat of sw corner of global input grid |
41 |
*_lon_inc :: scalar x-grid increment |
42 |
*_lat_inc :: vector y-grid increments |
43 |
*_nlon and *_nlat :: input x-grid and y-grid size |
44 |
|
45 |
In this particular example the input and output grids are |
46 |
the same, so the results of the verification experiment remained |
47 |
unchanged from before. |