w2w problem

[Foxcrash2021]

Describe your issue

Hello, I encountered a problem while using w2w. I used the underlying surface updated with land use data from Liaoning Province in 2020 in China (which involves the reclassification of USGS by geographical data), and then used it/ Geogrid.exe generated a geographic file, and then I used w2w to generate an LCZ file, but this error occurred

I don't know if it's because I updated the land use data

w2w --version

w2w -0.5.0

nc-config --all

netcdf-fortran-4.5.3
netcdf-c-4.7.4

Installed Packages

numpy==1.23.1

Traceback

https://user-images.githubusercontent.com/94523993/226560059-15b435bd-d4a9-4cda-9c60-33a2b51bcd5e.png

[matthiasdemuzere]

I do think this error occurs because of the updated land use data, that is not recognised by W2W. Yet hard to say without knowing how exactly you have updated the land use using "geographical data".

@dargueso perhaps you can follow-up?

[dargueso]

@matthiasdemuzere I can follow up, but I need information on the new land use data used. It looks like there is a problem with variables LU_INDEX and LANDUSEF, which may not be consistent. But hard to tell without the files. @Foxcrash2021, can you send us more info? Thanks

[Foxcrash2021]

Okay, I'll provide more information, just a moment

[Foxcrash2021]

Let me briefly describe how I update the data. First, I used the high-precision tif data from the Chinese Academy of Sciences to perform a reclassification based on USGS categories for the provinces I want to simulate. Some categories are missing, but the maximum number of categories 24 corresponds to USGS. Then

, I used the gdal tool to convert tif data into bil data. Based on the generated header file. hdr, I created an index file and binary geographic data, If LCZ partitioning is not used, just use/ I can continue to run the geogrid.exe process until a wrfout is generated. My wrf version is 4.3

[dargueso]

@Foxcrash2021, using alternative data sources is not supported by W2W in principle. However, I can have a look and see if I can help you. To that purpose, I need to see the actual geo_em file in order to understand where the error is coming from.

[Foxcrash2021]

ok,thanks,I will upload later

[Foxcrash2021]

here. 2020_geo_em.zip

[Foxcrash2021]

"I did not update the d01 data and used the default ones, while I updated the d02 and d03 data."

[dargueso]

Hi @Foxcrash2021, I'm checking your files and there are a few things that puzzle me. Assuming you're using a USGS 24-cat system, in geo_em.d01.nc , it looks like there is no Urban areas (category 1). In geo_em.d02.nc and geo_em.d02.nc, the entire domain is category 16 (which is water in USGS), so it is impossible to replace any urban. Furthermore, if the domain would be entirely urban, the code isn't able to replace by the surroundings, because that information is not provided. It seems that you have to double check the conversion of your data. WRF works fine but I'm not sure the landscape is described as you expected.

[Foxcrash2021]

Hi @Foxcrash2021, I'm checking your files and there are a few things that puzzle me. Assuming you're using a USGS 24-cat system, in geo_em.d01.nc , it looks like there is no Urban areas (category 1). In geo_em.d02.nc and geo_em.d02.nc, the entire domain is category 16 (which is water in USGS), so it is impossible to replace any urban. Furthermore, if the domain would be entirely urban, the code isn't able to replace by the surroundings, because that information is not provided. It seems that you have to double check the conversion of your data. WRF works fine but I'm not sure the landscape is described as you expected.

Thank you for your reply. I discovered what you said later, and it is true. It may be that something went wrong when I converted the data. I will carefully review it immediately. Thank you again for pointing out this key issue to me