Running with non-global forcing data

[leiapauline]

Hello,

I want to run hrldas with my forcing data over 85-165E, -15,35N. I want to reduce the space needed for my input data because our study area is only over a city. What I did was to subset the Zfile and LANDSfile over the same domain as my forcing data. But when I run ./create_forcing.exe, I get an error stating that the z variable can't be read. I checked their structures but I couldn't see any difference. Is there a way for me to make this run possible? I am currently using hrldas on a docker desktop for windows. Any guidance will be much appreciated.

Sincerely,
Leia Pauline

[cenlinhe]

what forcing dataset are you using? ERA5-land? GLDAS? or something else?

[leiapauline]

I am using ERA5-Land. If it can help, my input files can be found here.

[tslin2]

Hi,

You will have to revise the subroutine code for the regional files, and recompile it (create_forcing.F) due to the default code being read Zfile and LANDSfile as the global domain.

hrldas/hrldas/HRLDAS_forcing/create_forcing.F

Line 2090 in 087f1fd

subroutine read_era5_netcdf_unit(ncid, datastruct, ierr, current)

You should change these two parts at least.

hrldas/hrldas/HRLDAS_forcing/create_forcing.F

Line 2100 in 087f1fd

integer, parameter :: NLATS = 1801, NLONS = 3600

hrldas/hrldas/HRLDAS_forcing/create_forcing.F

Line 2135 in 087f1fd

call map_set(PROJ_LATLON, datastruct%proj, lat1=90.00, lon1=0.00, &

[leiapauline]

Thanks for the guidance. The recompiled version is running.

[fdesales]

Hi there.
I was wondering if you had success in running HRLDAS using non-global ERA5 forcing data. I have edited create_forcing.F following the instructions above and recompiled the code. However, my forcing data created with the non-global data looks different than the same generated from global data. Any advice? Thanks.

[cenlinhe]

Hi, have you checked the lat,lon setup in your modified script. The setup of lon of 0-360deg is different from -180 - 180deg. Also, make sure you select the correct number of nlon and nlat. What study domain are you looking at?

[fdesales]

Thanks your response. I did update nlon and nlat according to my data but that did not solve the problem. create_forcing.exe runs fine but when I compared the forcing files generated with global and non-global data, I noticed that the latter had the latitude reversed.

The problem seem to be in the grib1_gribdata (module_grib1.F). Global grib data are read by the GRIB1_SGUP_BITMAP subroutine, while for non-global files GRIB1_SGUP_NOBITMAP is used. The latter flips the latitudes of the data.

My conclusion is that global and non-global ERA5-land grib files downloaded from Copernicus CDS have different BITMAPs. Maybe this is an issue with the new CDS.