[Wrf-users] CCSM data into WPS

[Maples, Amy C.]

WRF users,

We are attempting to run CCSM data through WRF using a domain of the continental US.  We've set up a preprocessing script to turn our netcdf data files into appropriately formatted grib files.  When we run NNRP data in netcdf through this script followed by WPS and WRF, the output appears fine.  However, the CCSM data, after being run through the exact same script, has extremely strange precipitation patterns when it comes out of WRF (at the moment, precipitation only over bodies of water).

We believe the main issues lies with the pressures.  Both input sets of grib files are set up on pressures levels and include variables for PMSL and PSFC (all in Pa).  When we run the NNRP data through WPS, the output PSFC and PMSL values look like the input grib data.  However, when we run the CCSM values through WPS, the output PSFC and PMSL have identical values which are over an unreasonably small range.

The two data sets are run through WPS using the same Vtable and without any changes to METGRID.TBL.   The only changes to namelist.wps are in the date fields.  The only real difference between the datasets, once in grib format, is that the CCSM data has missing values.  The CCSM data is interpolated from sigma levels and so has missing values below the surface of the earth, however we could extrapolate below the surface to eliminate the missing values.  But the NNRP data does not have missing values.  We have tried setting the missing_value parameters in METGRID.TBL to be the same as the grib files, but this doesn't change the WRF output.

Does anyone know what could be causing the difference in PSFC and PMSL?  Or other possible causes of the precipitation patterns?

Thanks,

Amy Maples
Center For Sustainability
Noblis, Inc.
Falls Church, VA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20110329/a3943b78/attachment.html