WRF data with EMEP model

[kuzo1985]

Good afternoon,dear colleagues! I'm trying to calculate wet deposition of sulfur dioxide by using EMEP MSC-W model with meteo files made by WRF model. I run the rv4.10 and I have several problem 1) If I try to use final meteofiles from WRF model without changes ( for emample, wrfout_d03_2013-10-02_00:00:00), the program writes me an error massage:

/home/kuza/EMEP_folder/EMEP_MSC-W_model.rv4.10.OpenSource/wrf_2013_2014/d3/wrfout_d03_2013-10-02_00: 
No such file or directory 

 STOP-ALL ERROR: Error in netcdf routine

application called MPI_Abort(MPI_COMM_WORLD, 9) - process 0

[unset]: write_line error; fd=-1 buf=:cmd=abort exitcode=9:system msg for write_line failure : Bad file descriptor 

2)Then I change the name of my file from "wrfout_d03_2013-10-02_00:00:00" to "wrfout_d03_2013-10-02.nc". After changing EMEP model runs the simulation, but also it Writes me such messages as

2D precipitations sum of large_scale and convective precipitations
 release profile for 3D precipitations derived from QRAIN
variable does not exist: v10NetCDF: Variable not found
  found v component of 10m wind 

and after several days of simulations :

variable does not exist: sigma_dotNetCDF: Variable not found
 variable does not exist: precipitationNetCDF: Variable not found
variable  does not exist: eddy_diffusion_coefficientNetCDF: Variable not found
 variable does not exist: relative_humidity_2mNetCDF: Variable not found
 variable does not exist: surface_stressNetCDF: Variable not found
 Warning: not reading all levels            4           1 SMOIS
 Warning: not reading all levels            4           3 SMOIS
 variable does not exist: v10NetCDF: Variable not found

Tell me please How should I you correctly use WRF data? Thank you!

[gitpeterwind]

Hi @kuzo1985,

Just for the general background: we (at the Norwegian Meteorological Institute) do not routinely use wrf to drive the model. However most of the external users do use wrf and we therefore aim at maintaining the compatibility. This means that the model will work fine with wrf metdata, but the setup maybe a bit confusing.

The colon , : , in "wrfout_d03_2013-10-02_00:00:00" is interpreted as a special character by the gfortran compiler (not for the Intel compiler). You can solve this by putting a backslash in the definition of the name in the config file (not the name of the file in your filesystem) as in: "wrfout_d03_2013-10-02_00\:00\:00" This is however not compatible with the Intel compiler. We will try to find a solution that is robust for both.

The warnings are confusing I admit, but they are only warnings, meaning that it is an informative message, but the model can do without it (possibly with a small degradation of the quality of the results). If a met variable is really needed but not found, the model will stop. (the 10 meter wind speed is actually found, but with other names).

If I understand correctly, the model does not really stop or give meaningless results, you are just worried by the messages?

It is recommended to use the latest version of the model.

[kuzo1985]

Hi, gitpeterwind ! I apologize for not responding to your message - all the time I tried to compile last version of EMEP MSC-W. Now I have solved one problem : the model runs with files named as "wrfout_d03_2013-10-02_00:00:00". But the other problems are also relevant: I am worried of Warnings, you are right. I don't know if it is need to open new issue? I try also to run file with Lambert projection.The program writes :

MapFile /home/kuza/EMEP_2018/downloads/rv4_17input/EMEP_MSC-W_model.rv4.17.OpenSource/i nput/Landuse_PS_5km_LC.nc MapFile /home/kuza/EMEP_2018/downloads/rv4_17input/EMEP_MSC-W_model.rv4.17.OpenSource/i nput/glc2000xCLMf18.nc RdLanduseCDF:LANDUSE: found 1 .../Landuse_PS_5km_LC.nc RdLanduseCDF:LANDUSE: found 2 .../glc2000xCLMf18.nc NetCDF: Index exceeds dimension bound ERRMSG: RRvalues STOP-ALL ERROR: Error in netcdf routine application called MPI_Abort(MPI_COMM_WORLD, 9) - process 0 [unset]: write_line error; fd=-1 buf=:cmd=abort exitcode=9 : system msg for write_line failure : Bad file descriptor How can I solve this problem? Thank you!

[gitpeterwind]

Hi Kuzo, It is difficult to say what goes wrong from this short output. It seems the reading of landuse is failing. Do you have the possibility to send me one day of metdata (wrfout_d03_2013-10-02_00:00:00) and your config_emep.nml file, so that I can try to reproduce the problem? Perhaps it is easier if you contact me directly at my e-mail: peter.wind@met.no If you do not know how to send a large file, I can send you a link that allows you to upload large files. Best wishes, Peter

[gitpeterwind]

I got your wrf metdata and and there was actually an error in the emep code. This is triggered when:

a) WRF metdata is used

b) Lambert projection is used

c) TRUELAT1 = TRUELAT2

This can be corrected by setting lat0_lambert = lat_stand1_lambert under line 714 in file GridValues_ml.f90:

if(USE_WRF_MET_NAMES)then call check(nf90_get_att(ncFileID, nf90_global, "TRUELAT1",lat_stand1_lambert )) call check(nf90_get_att(ncFileID, nf90_global, "TRUELAT2",lat_stand2_lambert )) lat0_lambert = lat_stand1_lambert ...

We will include this also in the next version. Thank you for the headup!

[martinsaliba]

Hi,

I would like to integrate meteo data produced by WRF with EMEP in order to run the model for the whole area of the Mediterranean Sea. Can somebody provide a copy of the config_emep.nml file regarding the required configurations?

[gitpeterwind]

you only have to set the link to the wrf metdata for the meteo, something like: meteo = '/My/PATH/wrfout_d01_YYYY-MM-DD.nc', The details will be presented at the course.

[avaldebe]

Closing down solved issues after rv4_32 and rv4_33. Please reopen if ncecesary.