LATLON_TO_IJ error when creating ICs for regional simulation

[MattW]

Hi MPAS support folks,

I'm trying to initialize a 3 km MPAS simulation on a regional domain using HRRR output to create the ICs / LBCs. When I try to interpolate the HRRR data from the intermediate files from WPS to the regional MPAS domain I created using init_atmosphere_model, I get this error:

----------------------------------------------------------------------
Beginning MPAS-init_atmosphere Error Log File for task 0 of 1
Opened at 2024/01/04 16:24:47
----------------------------------------------------------------------

ERROR: You have not called map_set for this projection
CRITICAL ERROR: LATLON_TO_IJ
Logging complete. Closing file at 2024/01/04 16:24:47

The only places I've seen this error mentioned online are in a thread about using lat/lon data in parts of WRF where it doesn't seem to be supported (Error encounted in LATLON_TO_IJ: You have not called map_set for this projection). Does this indicate that the HRRR intermediate files I'm using have a lat/lon projection that MPAS doesn't support? I've attached my namelist.init_atmosphere and streams.init_atmosphere files below--let me know if you need any additional information!

Thanks,

Matt Wilson

 

[Ming Chen]

Matt,
Would you please upload your log file for me to take a look? Also, can you tell me where you downloaded the HRRR data? if the datafile is not too large, can you upload two HRRR data to Nextcloud storage so that I can run a test myself? Thanks.

 

[MattW]

Sure! I've attached the log file. The HRRR data was downloaded from the HRRR archive on Google Cloud, and I've been able to use it to initialize WRF simulations in the past. I can upload some of the intermediate HRRR files here, or if it's easier I can also give you the path to where I'm running MPAS since I'm running it on derecho (/glade/derecho/scratch/mawilson/mpas_tests/ohio_mesh).

 

[Ming Chen]

Thanks for the information. I will be able to access your HRRR files in derecho, so there is no need to update data to Cloud.

I will take a look at this case and get back to you. It may take some time and thanks for your patience.

 

[Ming Chen]

Hi Matt,

Just want you to know that I didn't forget your case. I am still working on it. The problem is that the HRRR map projection (PROJ_LC) is not supported for MPAS initialization. Please take a look at the code mpas_init_atm_cases.F, in which you can find the following piece of code, which indicates that only data on lat-lon and gauss grids are supported:

            if (field % iproj == PROJ_LATLON) then 
               call map_set(PROJ_LATLON, proj, &
                            latinc = real(field % deltalat,RKIND), &
                            loninc = real(field % deltalon,RKIND), &
                            knowni = 1.0_RKIND, &
                            knownj = 1.0_RKIND, &
                            lat1 = real(field % startlat,RKIND), &
                            lon1 = real(field % startlon,RKIND))
            else if (field % iproj == PROJ_GAUSS) then 
               call map_set(PROJ_GAUSS, proj, &
                            nlat = nint(field % deltalat), &
                            loninc = 360.0_RKIND / real(field % nx,RKIND), &
                            lat1 = real(field % startlat,RKIND), &
                            lon1 = real(field % startlon,RKIND))
!                            nxmax = nint(360.0 / field % deltalon), &
           end if

I have modified the code to make it able to process data on Lambert projection. But the issue still exists, indicating that more changes are necessary.

I will continue to work on this case and keep you updated.

 

[MattW]

Thanks! I've switched to using ERA5 ICs for the moment, but I'd definitely still be interested in using HRRR ICs once support for the HRRR map projection is added.

 

[Ming Chen]

Matt,
I am sorry to get back to you late, but this is because I have been distracted by other emergency issues.
I am revisiting the HRRR data issue. One thing I am not sure is that, you set

 config_nfgsoillevels = 4

However, I looked at your HRRR data and it seems that there are at least 9-10 levels of soil data. Can you double check to make sure how many levels of soil data you have?

I have fixed the code and I believe it should work as expected. But I need to run the case successfully before I can be 100% sure.

 

[Ming Chen]

Matt,
Since the input HRRR soil data are 10-level data at specific soil depth layers, MPAS doesn't support this type of soil data.

 

[MattW]

Hi Ming,

Sorry for the late response, I was at the AMS conference last week. Derecho is down for maintenance today but I can double-check the number of soil levels in my input data when it's back online later this week.

 

[Ming Chen]

Hi Matt,
I have your data and I checked the levels of soil. Its soil levels are different to that of Noah and NoahMP. Currently MPAS doesn't support the HRRR soil data, which is another reason you cannot run regional MPAS using HRRR as input.
We will add more options to MPAS in the future, but it does take time. Please pay attention to our Release Note for new features.

 

[MattW]

Thanks!