hi, i have problem when i try to use forcing package to making NetCDF file for windspeed and momentum stress (i.e with forcing fields with vector data, other forcing fields with scalar data work).
error is:
DEF_VAR - unable to define variable: Uwind in NetCDF file: windspeed.nc
Invalide dimension id or name
Execution abnormally terminated in module: DEF_VAR
do u have any idea what i am doing wrong?
using forcing package
I had the same thing a couple of years ago. I think you need to be using wind stress rather than wind speed, for which the respective variable names are 'sustr' and 'svstr'. So convert your wind speeds to wind stresses, then put them into your forcing.nc file using those variable names.
Also, looking at mod_forces.F, Uwind and Vwind are on rho-points, not u- and v-points; this may be the direct cause of your problem...
I'm not sure when Uwind and Vwind are actually used; can anyone tell us?
-Evan
Also, looking at mod_forces.F, Uwind and Vwind are on rho-points, not u- and v-points; this may be the direct cause of your problem...
I'm not sure when Uwind and Vwind are actually used; can anyone tell us?
-Evan
As in many aspects of ROMS, you have choices to make. Here are some options:
1. Provide the stresses sustr and svstr directly.
2. Provide Uwind, Vwind and several other atmospheric variables and have ROMS compute the stresses from them. This is with the BULK_FLUX option, which is an atmospheric boundary layer from Fairall et al. I have been using this option lately.
I haven't used the Fortran Forcing package in a very long time now. I expect most ROMS developers are using Matlab for this instead. Also, the latest beta version of ROMS can read in a global atmospheric product and do its own interpolation to the ROMS grid internally.
1. Provide the stresses sustr and svstr directly.
2. Provide Uwind, Vwind and several other atmospheric variables and have ROMS compute the stresses from them. This is with the BULK_FLUX option, which is an atmospheric boundary layer from Fairall et al. I have been using this option lately.
I haven't used the Fortran Forcing package in a very long time now. I expect most ROMS developers are using Matlab for this instead. Also, the latest beta version of ROMS can read in a global atmospheric product and do its own interpolation to the ROMS grid internally.