Do nodata values in rasterfile cause errors?

[Joachim]

Hi,

I am using WRF-Hydro in standalone mode. I have successfully run WRF Hydro on Domains where everything is covered by land mass but currently I  am having issues when running WRF Hydro when the domain is touching the sea (i.e. the rasterfile used in the preprocessing has 'nodata' values). First I thought that the errors encountered are coming from these "nodata" entries where the sea is. I replaced all "nodata" values in the rasterfile with zero values, but I still can't run wrf_hydro.exe successfully. Therefore my question is: do nodata entries in my rasterfile cause errors? What is the best way to deal with this situation if one wants to run a domain which includes such values?

Thank you for your help

Best regards, 

Joachim

[Kevin]

Joachim,

NoData values are perfectly acceptable in the input high-resolution elevatoin layer (DEM) in the GIS Pre-processing tools. We frequently build model domains that contain coastlines and ocean areas, where the input DEM will contain NoData. Wherever a valid NoData value exists in the input DEM, it is converted to the WRF-Hydro nodata value (-9999) on the model grid. Have you checked the LU_INDEX and LANDMASK values in your input geo_em.d01.nc ("geogrid") file, to ensure that these are being represented as "water" in the Land Surface Model?

Kevin

[Joachim]

Hi Kevin

Thanks for your help! I checked the LU-Index for water which is 17 when taking the Landuse file from 'modis_landuse_20class_30s_with_lakes'. In the tutorial the 'usgs_30s' file is used, where the  water is indexed with 16. I think this is where the error is occuring. I now changed in the GEOGRID:TBL to the following:

name=LANDUSEF

        priority=1

        dest_type=categorical

        z_dim_name=land_cat

        dominant = LU_INDEX

        landmask_water =        default:17         # Calculate a landmask from this field

        interp_option =        default:nearest_neighbor

        rel_path =        default:modis_landuse_20class_30s_with_lakes/

 In the geo_em.d01.nc it is now also stated that:

:ISWATER = 17 ;

:ISLAKE = 21 ;

:ISICE = 15 ;

Unfortunately, running wrf_hydro.exe still is not working. Do you know if there is more to chnage to make sure the right index is used as water?

Thank you and best regards, 

Joachim

[aubrey]

We have not tested the code with land cover classifications that use separate categories for water and lake. Try using either the standard USGS or MODIS (no lake) classification systems.

Thanks!

Aubrey

[Joachim]

Hi Aubrey, thanks for your help! I downloaded the landuse data from the https://www2.mmm.ucar.edu/wrf/users/download/get_sources_wps_geog.html website, which is recommended in the WRF Hydro documentation. This is where I got the modis_landuse_20class_30s_with_lakes file. There is another file (without the lakes) called modis_landuse_20class_15s (is this the one you mean by 'no lake'?). Unfortunately the link doesn't work anymore and I can't try it out. Do you know where I can download the landuse files (globally) which will be compatible with WRF Hydro?

Thank you and best regards

Joachim

[Kevin]

Joachim,

You can try the older datasets, at this link: https://www2.mmm.ucar.edu/wrf/users/download/get_sources_wps_geog_V3.html

The 'landuse' layers, except 'landuse_30s_with_lakes' should work.

Kevin

[Joachim]

Hello Kevin

Thank you for your hint. I tried the dataset you mentioned but unfortunately it is not working as well. If I run it over land, it works well. The land mask layer seems fine as well. Somehow I can't figure out the root of the problem but will continue searching. Thank you for your help anyways.

Best regards, Joachim

[Joachim]

Hi everyone 

I tried to go all the way back to the Lambert-tutorial from the wrf-hydro training. Now i tried to just move the domain a little bit further south so the sea is included in the domain ( I have attached the namelist.wps file). This time I din't change the landuse file from what is provided in the WPS tutorial. The rasterfiles used in the preprocessing I took from hydrosheds, as it is recommended in the documentation. The Forcing data I took from GLDAS. Everything else I strictly did according to the tutorial. But again I have the problem, that when I try to run the simulation it crashes. The same process over land works fine. Is there something else I am missing here? Is this not supposed to work with the setup from the wrf hydro training container or the files from the WPS container? I also attached the log files.

Thank you for your help

Joachim

On Thursday, April 30, 2020 at 3:58:59 PM UTC+1, Kevin wrote:

[Walter]

Hey Joachim

I had a look at the diag_hydro.00001 file you have shared.

I faced the same problem where LWDN has a value -9999.

I wrote a shell script that replaces missing values of -9999 with 0 in the forcing data.

By doing this, I solved the error and got my model to run.

Cheers

Walter

[Joachim]

Thank you Walter

I followed your advice and now it is running. Thank you very much!

Cheers 

Joachim

[Walter Samuel]

Glad I could help you. Molly and Katelyn had answered this query earlier in another thread. 

Regards

Walter

--
You received this message because you are subscribed to the Google Groups "wrf-hydro_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hydro_use...@ucar.edu.
To view this discussion on the web visit https://groups.google.com/a/ucar.edu/d/msgid/wrf-hydro_users/8042c1a0-1c10-4ce1-b52c-3f129c93bcb4%40ucar.edu.

[Orzu Odinaev]

Hey Walter,

I have the same error, could you please share that script replacing -9999 with 0?

[Orzu Odinaev]

Hey Joachim,

I have the same error, could you please share that script replacing -9999 with 0?