CRHMcode GUI crashes when running model with ClassWQ_pbsmSnobal module

[BZ840]

Hi,

I have two proj files: One_subbasin and Basin (which has 3 sub-basins grouped). The GUI (latest version) had no problem running the One_subbasin.prj, but it kept crashing (autoexit) when running the Basin.prj. I am not sure what is causing the issue, could you please take a look?
1998-2020.txt
Basin.txt
One_subbasin.txt

successful run with the One_subbasin.prj
crhmRun.log

GUI autoexit with no error msg prompted when running the Basin.prj:
crhmRun_failed.log

[jhs507]

The crash happens during the declaration of the of the modules in group A.

In ClassWQ_pbsmSnobal.cpp line 89 the declaration of the variable BasinSnowLoss_mWQ is made.

decldiag("BasinSnowLoss_mWQ", TDim::NDEF, "transport out of basin", "(mg/l * mm*km^2/int)", &BasinSnowLoss_mWQ, &BasinSnowLoss_mWQ_lay, numsubstances);

From there execution reaches lines 444-488 ClassModule.cpp

if (dimen == TDim::NDEF) { // save for lay loop
		Var_loop_lay_table[Var_NDEFN_cnt] = newVar->layvalues;
		Var_loop_lay_value[Var_NDEFN_cnt++] = newVar->values;
}

At this point Var_loop_lay_table and Var_loop_lay_value hold null references and attempting to index into them to write causes an exception.

I will have to try to trace down when these should have been initialized and why they were not.

However this is a water quality module and I don't think they have been properly verified yet.

@DiogoCostaPT or @lawfordp2017 may know more.

[jhs507]

The Var_loop_lay_table and Var_loop_lay_value variables have memory declared at lines 88-104 in ClassModule.cpp

	if (Var_NDEFN_cnt) {

		Var_loop_lay_table = new double**[Var_NDEFN_cnt]; // define [#Var][#layers]

		for (long jj = 0; jj < Var_NDEFN_cnt; ++jj) {
			Var_loop_lay_table[jj] = new double*[nlay];
		}

		Var_loop_lay_value = new double*[Var_NDEFN_cnt]; // define [#Var]
		for (long jj = 0; jj < Var_NDEFN_cnt; ++jj)
			Var_loop_lay_value[jj] = new double[nhru];
	}
	else {

		Var_loop_lay_table = NULL;
		Var_loop_lay_value = NULL;
	}

The declaration of memory only occurs if Var_NDEFN_cnt is > 0

Var_NDEFN_cnt is only increased from its declaration to 0 on lines 396-398 in ClassModule.cpp

		if (dimen == TDim::NDEFN) // save for lay loop
			++Var_NDEFN_cnt;

Or on lines 444-448 in ClassModule.cpp

			if (dimen == TDim::NDEF) { // save for lay loop
				Var_loop_lay_table[Var_NDEFN_cnt] = newVar->layvalues;
				Var_loop_lay_value[Var_NDEFN_cnt++] = newVar->values;
			}

[DiogoCostaPT]

Hi @jhs507 @BZ840 @lawfordp2017,

Tom Brown and I developed these modules between 2019 and 2021 (paper). Since then, @lawfordp2017 continued with the development, but the last time I heard about progress was probably a year ago or so; thus, I don't know the current status.

In any case, there are a few things to bare in mind.

• CRHM's new GUI (and structural re-organization) was being developed at the same time I was developing the WQ modules. The case study I used to develop these modules was the South Tobacco Creek Basin, and it was based on the no-GUI version that was initially set up by @lawfordp2017 for compiling in Linux. When I finished this project/paper, the new CRHM-gui version was being released for testing, but without the WQ modules. I suspect that @lawfordp2017 helped with the integration of the WQ modules in the new GUI version afterwards, but I don't know.

• I did not use groups when I applied the model to South Tobacco Creek Basin during the WQ-modules' development phase. Thus, we never came across this problem. Did you look into this after I left, @lawfordp2017 ?

Thus, my suggestion is:

• If you have time and want to help with development: dive into it and help fix this issue with groups; that would be great! :)
• if you don't have time: don't use groups

I hope this helps a bit in keeping the discussion and modelling development going - I am happy to discuss further.

[lawfordp2017]

It seems this issue didn't trigger a message to my email (or it got filtered). For anyone interested in using REWs with either erosion modelling or species modelling in the future, both the 'crhm_with_erosion' and 'rew_for_wq' branches have support for species transport in REW_routing. 'rew_for_wq' was also used recently by Taufique Mahmood for WQ modelling.

[jhs507]

@BZ840 Have you been able to use the rew_for_wq branch to address your problem?

[DiogoCostaPT]

I have not used this branch. Who developed it? From: jhs507 ***@***.***> Date: Tuesday, 4 June 2024 at 16:59 To: srlabUsask/crhmcode ***@***.***> Cc: Diogo Costa ***@***.***>, Mention ***@***.***> Subject: Re: [srlabUsask/crhmcode] CRHMcode GUI crashes when running model with ClassWQ_pbsmSnobal module (Issue #428) @BZ840<https://github.com/BZ840> Have you been able to use the rew_for_wq branch to address your problem? — Reply to this email directly, view it on GitHub<#428 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AFC5FP43O2G5UHVV4XCWAWLZFXP6FAVCNFSM6AAAAAAXWQ57U6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNBXHA4DSOBRHE>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[jhs507]

@DiogoCostaPT

The branch was developed by @lawfordp2017