WRF & MPAS-A Support Forum

I think aer_opt = 2 was used by the contributor applying daily aerosol data, not climatology, although if such a climatology exists, it probably can be used. And yes, the climatology can be different. Our aer_opt = 1 uses aerosol climatology from the 90s. Thompson's is obtained from global model sim...

H_DIABATIC is latent heating from microphysics. It can be positive and negative, depending on the physics that is going on in microphysics scheme.

I am not sure which version of WPS you are using. It is possible that older version of WPS cannot process new GFS data.
Please update your WPS to the latest version (WPSv4.2), then try again and let me know if you still have the same issue.

Yes they are always on their native model grids both horizontally and vertically.

Please stay with debug_level = 0 because this option doesn't always provide helpful information. You are right that you need to find the first time and grid point where NaN occurs. Form there you can trace back to other variables related to this NaN and possibly figure out what is wrong. Another opt...

Hana,
What are the error message you get when ungribbing ERA5?
Please also tell me the following information:
(1) list of ERA files you download from CISL
(2) Vtaable you use
(3) Please attach your nameless.input for me to take a look.

Would you please double check the file gfs.t00z.pgrb2.0p50.f000, and find what time it is for?
Please try the command
wgrib -s gfs.t00z.pgrb2.0p50.f000
or if you are familiar with NCL,
ncl_convert2nc gfs.t00z.pgrb2.0p50.f000
Then look at the netCDF file.

for__b_fmt_table looks like a Fortran run-time-library problem. Please make sure your LD_LIBRARY_PATH is set correctly. In addition, note that libraries and include-files for different MPI versions (OpenMPI, MPICH, etc) and/or for different underlying compiler-set are not compatible with each other....

Wenli,
You are right that it makes no sense to specify eta_level for ideal cases.

Peter, In this case at least we know for sure that the problem is induced by extremely high model top. To detect when and where the error occurs, please save wrfrst file at the nearest time before the model crash, then restart the model and save wrfout at every time step. Please look at these wrfout...

Lin, Please see my answers below to your questions: (1) You need to run WRF with multiple nodes, i.e., mpirun -n 24 ./wrf.exe Please consult your computer manage fr correct command stye (different machines might be different) (2) You don't need to specify memory for each node (3) You can always comp...

The interpolation method available in WPS can always be used with the current version of codes.

This error clearly indicates that you don't have GFS data for the time 2020-09-13_00.
Please double check your GRIB files and enamelist.wps, make sure that ALL GFS data are available for the period specified in namelist.wps.

Please type the command:
ncdump -v Times wrfbdy_d01,
and you will find how many times of boundary forcing you have. This will also tell whether your integration time period is beyond the period your boundary forcing data are available.

All,
Please take a look at the post here, which solves the problem of ungrib compiling.
viewtopic.php?f=28&t=9367

Would you please tell the following information:
(1) How did you compile WRF? Which compiler did you use? Which WRF version did you compile?
(2) How did you compile WPS, e.g., in serial or dmpar mode? Which version of WPS did you use?

It is possibly because the specified value of truelat in the previous namelist.wps is integer, while the WPS program expects a real type. However, this still doesn't explain why the code complains of a variable that is not used based on the namelist setting.

I am not sure what physics scheme are you using? Please clarify. But I don''t think CCN has the same physical meaning as Nd.
One variable, QNC, the cloud water Number concentration, might be helpful in your study. Please take a look.

Please find the lines below in Registry.EM_CMMON: state real TLEV_URB3D i{umap6}j misc 1 Z r "TLEV_URB3D" "INDOOR TEMPERATURE" "K" state real TFLEV_URB3D i{umap9}j misc 1 Z r "TFLEV_URB3D" "FLOOR TEMPERATURE" "K" change them to: state real TLEV_URB3D i{umap6}j misc 1 Z rh "TLEV_URB3D" "INDOOR TEMPER...

Xun,
The unit of potevp is m, which is true least since WRFV3.6. I don't think the computation of this variable is changed. It is possible that the definition of its unit is wrong in WRFv3.3.1.

Peter, Those netCDF errors are not really errors, and they will not lead to model crash. In this case, apparently the model physics is wrong. (1) As I suggested in my previous answer, would you please run this case with lower model top and less vertical levels? This test will let us know whether the...

svp1, svp2 and svp3 are constants for saturation vapor pressure calculation. The first two variables are dimensionless and svp3 has the unit of K. The algorithm to compute saturate water vapor pressure is ES=svp1*exp(svp2*(T-svpt0)/(T-svp3)) Note that the value of svp1 can be specified to either 611...

I suppose the model crashed very soon after the initialization. Please let me know if I am wrong, Would you please provide more information: (1) The Merra data in pressure level or on model level? (2) Can you attach the program merra2wrf for me to take a look? (3) Please double check the rsl files f...

Terrestrial water, ecosystem, and climates are intimately coupled, coevolving earth system. We aim to improve the understanding of the coupled system dynamics by developing and implementing newly developed models of soil and plant hydraulics in regional climate models through a NASA funded project. ...

Have you compared the two namelist.wps you used? Is there any difference between the two? There must be a reason why the previous run complained of the truelat setting (I suppose you run with the same georgic.exe).