Converting CERRA (lambert conformal) to rotated latlon grid
Added by Michał Jędrzejczyk 5 months ago
Hi,
I want to regrid CERRA data to rotated latlon so that it is compatible with JRodos software and so that the grid is coarser.
My questions:
1. Do u,v vectors need any special treatment like rotating from grid-relative form to another or is cdo -f grb2 remapbil,mylcc_regrid14rgll.txt uv_lambert.grib2 uv_latlon.grib2 command enough to interpolate the vectors correctly?
2. When interpolating to another hybrid model level structure, does cdo automatically detect that it is dealing with velocities or do u,v need to have specific variable names or parameter ids?
The content of mylcc_regrid14rgll.txt:
gridtype = projection gridsize = 31073 xsize = 193 ysize = 161 xname = rlon xlongname = "longitude in rotated pole grid" xunits = "degrees" yname = rlat ylongname = "latitude in rotated pole grid" yunits = "degrees" xfirst = -10.000 xinc = 0.125 yfirst = -16.500 yinc = 0.125 scanningMode = 64 grid_mapping = rotated_pole grid_mapping_name = rotated_latitude_longitude grid_north_pole_latitude = 32.5 grid_north_pole_longitude = -170.0 north_pole_grid_longitude = 0.0
Thank you,
Michał
Replies (1)
RE: Converting CERRA (lambert conformal) to rotated latlon grid - Added by Rostislav Kouznetsov 2 months ago
Hi Michał,
We have the same issue with SILAM. So far it cannot run in LCC, so we approximate it with rotta.
For small areas it is quite accurate and avoids remapping.
You can try to craft something out of the attached script to the netcdf file obtained with
$ cdo -f nc copy uv_lambert.grib2 uv_lambert.nc
Once you make the corresponding rotated-grid spec, you can use
$ cdo setgrid,yourrll.grid uv_lambert.grib2 uv_rll.grib2
just to replace the grid.
There will be slight mismatch between the grid, but in return you will have no artifacts of remapping and resampling GRIB. The latter are arguably more evil...
BR
Rostislav
| lcc2silamgrid.py (4.82 KB) lcc2silamgrid.py |