Issues in running mHM

[Mohammedhhu]

Dear mHM expert

Recently, I have set up and prepared all the input of mHM. I am running mHM across the whole Nile River Basin (3,176,541 km2). I got this attached error, and I think it is in regard to my Latlon file.
L0 = 0.00902777785
L1 = 0.25
L11 = 0.5
I have attached the error and the information of latlon.nc for your reference.
I am looking for your suggestion and recommendation to solve this issues

mhm.zip
header_1.txt
header_2.txt
header_3.txt

[BryanSouzaM]

I believe that all layers need to be multiples of each other and have the same extensions. L0, for example, doesn't follow this rule

[Mohammedhhu]

I have checked all the morphological data regarding to the grid scale and extent.
After preparing new morphological files to the same extent as the meteorological ones, I got a new error.
Indeed, my soil_classification.txt has no value such as -9999.

what dose it mean?

[StephanThober]

Dear Mohammedhhu,

the mask in mHM is derived from the dem file. This error indicates that in your soil_class.asc you have a missing value at a grid cell where the dem has a value. Did you check that the mask in the dem.asc and soil_class.asc are the same?

Best,
Stephan

[Mohammedhhu]

Dear StephanThober

Thank you so much for your observations; yes, there are some pixels with missing data in soil and geology maps while they have values in dem map. I have solved this by filling in the missing data.

Best regards,
Mohammed

[Mohammedhhu]

Dear all

Could you please help diagnose the cause of this segmentation fault? Are there specific checks or debugging steps you recommend to resolve this issue?

Thank you for your time and support. Please let me know if you need additional details or files.

[Mohammedhhu]

Hi Pallavhydro
Thank you so much for the information and steps that you provided. I have flow the second option and I create conda environment with full compilation. Indeed the test examples coming with mhm file running well. But I have same error even after this step.

[pallavhydro]

Hi @Mohammedhhu

Thanks for reporting back.
So the debug version of mhm (mhm_debug) also doesn't provide any different runtime error message? Could you please share the error this time.

P

[Mohammedhhu]

[pallavhydro]

Interesting.

I am a bit perplexed. Usually, when we compile and run the debug version, the runtime error (backtrace) should show specifics on the modules and line numbers.

Are you sure you "successfully" compiled the debug version by:
source CI-scripts/compile_debug

And then run the corresponding and newly compiled mhm_debug file as
./mhm_debug

Important: Please make sure that the executable corresponds to the compilation. You can do that my checking the time stamp of the executable which should correspond to the time of compilation. It is very common, even amongst us developers, that we overlook the completion of compilation, and wrongly use the previously compiled executable. I hope that is not your case (??).

KR
Pallav

[Mohammedhhu]

Yes, I have successfully compiled the debug version, but when I ran, I just used ./mhm without using the ./mhm_debug.

From this, I think the error from some input data, I will check them accordingly.

Any, recommendation ....

[pallavhydro]

Great to see you got more meaningful runtime error message using the debug executable.

As per the message, you have at least one occurrence of -9998 in one or more of your input soil maps. Finding the -9998/s and replacing it/them by -9999 (standard noData value) should help.

Meanwhile, also please share the soil classification file (or screenshot of) you are using for your setup for a quick check.

Cheers
Pallav

[pallavhydro]

Hi @Mohammedhhu

I checked the mHM source code for the issue. You need to revise the soil_classdefinition.txt file for the following points -

1. Revise the nSoil_Types to the number of soil types in the file. It is currently the largest ID.
2. The MU_GLOBAL needs to be in sequential order of 1, 2, 3, ... currently it is 2, 37, 39, 47, ... So you have to rename 2 to 1, 37 to 2, and so on.
3. You have soil IDs and horizons with multiple entries. Please only keep the correct entries. E.g. 16339 has two entries each for horizon 1 and 2.

Attached is the screenshot comparing your file (left) with the reference test_domain file (right):

Hope this helps
Pallav

[Mohammedhhu]

Hi @pallavhydro

I hope this message finds you well.

I followed your suggestion to renumber the MU_GLOBAL with 1, 2, 3 instead of using the numbers derived from the soil file. However, the same error persists. I am uncertain why this issue continues, despite thoroughly checking multiple times.

To ensure I haven’t missed any critical steps or made an error, I’ve outlined below the process I followed:

I downloaded the HWSD and added the hwsd.bil file to ArcMap for resampling and masking using my study area shapefile.
The masked raster file was converted to ASCII format.
The soil_classdefinition.txt file was generated from the masked raster soil file and the soil database using an R script.

I have attached the masked raster soil map and the R code for your reference. I would greatly appreciate your advice and suggestions to help resolve this issue.

Looking forward to your insights.

Best regards,
Mohammed
Soil_NRB.zip

[pallavhydro]

Hi @Mohammedhhu

The change of the IDs to 1,2,3,... in the definition/ look up file (soil_classdefinition.txt) means your soil map (soil_class.asc) should also be adapted accordingly. Only then mhm would correctly link the spatial information and the look up table information. Since the error message is still looking for an ID based on previous system (16487) I'm assuming the adaptation of the soil IDs in soil maps is still pending, right?

Pallav

[Mohammedhhu]

Hi @pallavhydro

I have changed the soil_class.asc, and my model runs well. Thank you so much for your help, I do appreciate your support.
Thank you so much to you guys @StephanThober @BryanSouzaM

[pallavhydro]

Hi @Mohammedhhu

I'm happy your setup is running.

To bring a closure to this discussion, I'd appreciate if you set the discussion as answered.

Cheers
Pallav