Problem with running obsgrid.exe for PX-LSM Nudging

WRF-CMAQ
1

Reference: PXLSM Soil Nudging - #12 by rgilliam

I was not able to post replies to the above discussion. So creating a new topic to continue discussion.

I have 4 wrf cases: 2021 Run, 2022 Run, 2023 Run, 2024 Run. All runs completed successfully, ie., both WPS and WRF ran successfully without obsgrid.

Now I want to add PX-LSM Nudging to all cases: 2021 Run and 2024 Run completed successfully with obsgrid.

However 2022 Run and 2023 Run gave error while executing ./obsgrid.exe.

I am attaching the log files: Obsgrid - Google Drive

2022 Run Error

Time Loop Processing, date = 2022-01-21_18:00:00

Using ./empty.obs:2022-01-21_18 as obs input file

Number of observations successfully ingested: 0.
Number of empty observations discarded: 0.
Number of observations discarded outside of domain: 0.

There were no valid observations found within the domain for this time period.

ERROR EXIT
For sea level pressure, request pressure level 1001 mb.
ERROR_EXIT in error_handler

2023 Run Error

Time Loop Processing for SFC FDDA, date = 2023-01-06_09:00:00

Using ./empty.obs:2023-01-06_09 as obs input file

Number of observations successfully ingested: 0.
Number of empty observations discarded: 0.
Number of observations discarded outside of domain: 0.

There were no valid observations found within the domain for this time period.

Doing Cressman for surface data
ROI for the surface, Cressman scan number 1 is 260.0 grid points or 520.00km ( 2.0 km spacing).
ROI for the surface, Cressman scan number 2 is 182.0 grid points or 364.00km ( 2.0 km spacing).
ROI for the surface, Cressman scan number 3 is 128.0 grid points or 256.00km ( 2.0 km spacing).
ROI for the surface, Cressman scan number 4 is 90.0 grid points or 180.00km ( 2.0 km spacing).

forrtl: severe (153): allocatable array or pointer is not allocated
Image PC Routine Line Source
obsgrid.exe 00000000004D4FF4 for_dealloc_alloc Unknown Unknown
obsgrid.exe 000000000040FC33 Unknown Unknown Unknown
obsgrid.exe 000000000041EC05 Unknown Unknown Unknown
obsgrid.exe 000000000040CD32 Unknown Unknown Unknown
libc-2.17.so 00002B87E3335555 __libc_start_main Unknown Unknown
obsgrid.exe 000000000040CC29 Unknown Unknown Unknown

I followed the same steps for all cases. What could be the reason for errors in 2022 run and 2023 run.

Looking forward to any kind of help.

Thank you
Anie

2

Let me know if your metgrid met_em file for 2023-01-06_09:00:00 is the same size as the prior times.

met_em.d01.2023-01-06_09:00:00.nc

I see you are using an empty obs file. That rules out an error in the Little R obs file for that hour. All I can figure is your met_em file for this hour does not have some legit fields because the Ungrib file was incomplete.

Otherwise, I would need your met_em file for this day so I could test myself.
met_em.d01.2023-01-06_*.nc

You could load these on your same google drive for me to download.

3

Thank you for your reply.

Ungrib was successfully completed in both 2022 and 2023 runs (./wrf.exe also worked fine without obsgrid).

I have attached met_em.d01.2023-01-06_*.nc as well to the same google drive.

Looking forward to your guidance.

Thank you

4

I cannot say exactly why, but almost 100% sure obsgrid is crashing because the met_em.d01.2023-01-06_12:00:00.nc and met_em.d01.2023-01-06_18:00:00.nc files only have the variable VV as a 2D variable. Never seen this before. This would surely cause problems in Obsgrid.

It may not cause WRF to crash because this is not an initial condition time. Unsure if you are creating a wrffdda_d01 file for this domain and now real.exe handles this. I’d scan the wrffdda_d01 file for the same time as these odd met_em files to see if the 3D V wind component has missing data. Since the model is nudged to these values it could still run through this period.

For fixing, you need to go back to the intermediate files produced by ungrib.exe and look at the size of the files for Jan 6, 2023 (12 & 18 UTC). I’m sure the size of these files is slightly different than the 00 and 06 UTC files. In this case you have to go back further and look at the GRIB files of the analysis. See if these are different as well. You might need to use a different analysis for these hours or if NAM 12 km I could look in our archive for replacements.

Standing by.