SWAT+ Delineation and HPC running error

[Saba Shaghaghi]

Hello,

I am developing a SWAT+ model for the Peace River Basin in Florida using:

• SWAT+ rev. 61
• QSWAT+ 3.1
• SWAT+ Editor 3.2
• SWAT+ Toolbox 3.2
• Linux executable: swatplus-61.0.2.61-gnu-lin_x86_64-Rel

My watershed contains many lakes and reservoirs, so I tried both:

1. Burning in my own stream shapefile with lake shapefiles.
2. Letting TauDEM generate streams automatically from the DEM.

However, in both cases I encounter problems either during delineation or when running the model on HPC for calibration.

During delineation, QSWAT+ gives this error during “Constructing topology”:

“ID value 22 from inlets/outlets file ... not found as DNOIDID in channels file ... Will be ignored.”

It seems the model cannot correctly connect streams, lakes/reservoirs, and outlets.

When running the model on Linux/HPC, it crashes with:

“Floating-point exception - erroneous arithmetic operation”
in:
time_conc_init_ / proc_cha_

After checking the generated files, I found:

• all channel elevations in chandeg.con are zero,
• many channel slopes are extremely small (0.00001),
• some channel lengths are unrealistically short.

Could you please advise:

1. Why channel elevations become zero?
2. What is the recommended workflow for flat watersheds with many lakes/reservoirs?
3. How should external stream and lake shapefiles be handled correctly in SWAT+?
4. Could these topology/routing issues be causing the floating point exception?

Thank you very much for your help.

Best regards,

Saba Shaghaghi

[Chris George]

I am sorry you have problems. It is not clear what exactly you are
doing, but my advice for flat watersheds with many lakes is to burn in a
channel shapefile if you have one, and to use a lake shapefile.
Burning-in is not an alternative to delineation, but a means of
assisting it.

When you complain of short channels I think you misunderstand that a
channel in QSWAT+ is actually a channel reach, a portion of a channel
between junctions or ends, and hence linear. If junctions happen to be
close then the reach will be short. You can eliminate short channels
(by mergers) in QSWAT+.

Short channels in flat watersheds can have very small slopes.

My experience is that it is pointless seeking the source of a later
error until the earlier one(s) are fixed.

I don't understand the problem with your inlets/outlets file, but I
wonder if you checked that all its points were snapped to the channel
network.

If you would like me to look at your model, send me a zip of the
complete project directory.

Chris

On 5/26/2026 2:09 PM, Saba Shaghaghi wrote:
> Hello, I am developing a SWAT+ model for the Peace River Basin in
> Florida using: SWAT+ rev. 61 QSWAT+ 3. 1 SWAT+ Editor 3. 2 SWAT+ Toolbox

> 3. 2 Linux executable: swatplus-61. 0. 2. 61-gnu-lin_x86_64-Rel My

> watershed contains many lakes and reservoirs,

> ZjQcmQRYFpfptBannerStart
> This Message Is From an External Sender
> This message came from outside your organization.
> ZjQcmQRYFpfptBannerEnd

>
> Hello,
>
> I am developing a SWAT+ model for the Peace River Basin in Florida using:
>

> * SWAT+ rev. 61
> * QSWAT+ 3.1
> * SWAT+ Editor 3.2
> * SWAT+ Toolbox 3.2
> * Linux executable: swatplus-61.0.2.61-gnu-lin_x86_64-Rel

>
> My watershed contains many lakes and reservoirs, so I tried both:
>

> 1. Burning in my own stream shapefile with lake shapefiles.
> 2. Letting TauDEM generate streams automatically from the DEM.

>
> However, in both cases I encounter problems either during delineation or
> when running the model on HPC for calibration.
>
> During delineation, QSWAT+ gives this error during “Constructing topology”:
>
> “ID value 22 from inlets/outlets file ... not found as DNOIDID in
> channels file ... Will be ignored.”
>
> It seems the model cannot correctly connect streams, lakes/reservoirs,
> and outlets.
>
> When running the model on Linux/HPC, it crashes with:
>
> “Floating-point exception - erroneous arithmetic operation”
> in:
> time_conc_init_ / proc_cha_
>
> After checking the generated files, I found:
>

> * all channel elevations in chandeg.con are zero,
> * many channel slopes are extremely small (0.00001),
> * some channel lengths are unrealistically short.
>
> Could you please advise:
>
> 1. Why channel elevations become zero?
> 2. What is the recommended workflow for flat watersheds with many
> lakes/reservoirs?
> 3. How should external stream and lake shapefiles be handled correctly
> in SWAT+?
> 4. Could these topology/routing issues be causing the floating point

> exception?
>
> Thank you very much for your help.
>
> Best regards,
>
> Saba Shaghaghi
>

> --
> You received this message because you are subscribed to the Google
> Groups "QSWAT+" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to qswatplus+...@googlegroups.com
> <mailto:qswatplus+...@googlegroups.com>.
> To view this discussion visit https://groups.google.com/d/msgid/
> qswatplus/8365c29c-449d-4137-92f1-38ff89b79619n%40googlegroups.com
> <https://urldefense.com/v3/__https://groups.google.com/d/msgid/
> qswatplus/8365c29c-449d-4137-92f1-38ff89b79619n*40googlegroups.com?
> utm_medium=email&utm_source=footer__;JQ!!KwNVnqRv!
> HngmxOnmTCbwYOpGgrX9S_XDnKWGEtJnzqNuQEKfcMjvyifjpv_Aqp0Ra36WHLeRksilBmQfgPUb6y5zZX4GK8s$>.

[Fazl Akhtar]

This message indicates that QSWAT+ could not match the outlet point with ID 22 from your inlets/outlets shapefile to any stream channel in the generated channel network (demmchannel.shp). As a result, the outlet will be ignored during watershed delineation. This usually happens when the outlet is not located exactly on a stream, when stream threshold settings have changed after the outlet was created, or due to a temporary delineation error. 

First try clicking "Create Watershed" again as suggested by QSWAT+. If the error persists, verify that the snapped outlet point lies directly on a stream in the channel shapefile and re-snap or recreate the outlet if necessary.

I hope this helps.

Regards

--

You received this message because you are subscribed to the Google Groups "QSWAT+" group.

To unsubscribe from this group and stop receiving emails from it, send an email to qswatplus+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/qswatplus/8365c29c-449d-4137-92f1-38ff89b79619n%40googlegroups.com.

--

Best Regards
________________________________

Dr.-Ing. Fazlullah Akhtar

Senior Researcher

Center for Development Research (ZEF)

Genscherallee 3, 53113

Bonn

[Saba Shaghaghi]

Thank you for your guidance. The issue was related to the routing. I modified my river and lake maps, and now I no longer get that error either during the delineation step or later when running the model on the HPC.

However, I am now noticing another issue. During the watershed delineation process, SWAT changes the watershed boundary and removes a considerable portion of my watershed area. After the watershed creation step, some parts of the watershed are being ignored and are not included within the generated subbasins.

I am not sure why this is happening, and I would greatly appreciate your advice on what might be causing this issue.

I have attached:

1. The new version of the watershed generated using the corrected river and lake shapefiles (without routing errors),
2. The previous version where I had the routing error,
3. The original watershed area that I provided to the model as the DEM boundary in the first step.

Thank you very much for your help and support.

Saba Shaghaghi

[Chris George]

If watershed delineation comes up with a different boundary than the one
you think is correct, you can do one of three things:

- live with it
- if you have maps of the subbasin boundaries and the channels, use them
as an existing watershed model
- try a different delineation tool, eg GRASS

Chris

On 5/29/2026 6:52 PM, Saba Shaghaghi wrote:
> Thank you for your guidance. The issue was related to the routing. I
> modified my river and lake maps, and now I no longer get that error
> either during the delineation step or later when running the model on
> the HPC. However, I am now noticing

> ZjQcmQRYFpfptBannerStart
> This Message Is From an External Sender
> This message came from outside your organization.
> ZjQcmQRYFpfptBannerEnd
>

> Thank you for your guidance. The issue was related to the routing. I
> modified my river and lake maps, and now I no longer get that error
> either during the delineation step or later when running the model on
> the HPC.
>
> However, I am now noticing another issue. During the watershed
> delineation process, SWAT changes the watershed boundary and removes a
> considerable portion of my watershed area. After the watershed creation
> step, some parts of the watershed are being ignored and are not included
> within the generated subbasins.
>
> I am not sure why this is happening, and I would greatly appreciate your
> advice on what might be causing this issue.
>
> I have attached:
>

> 1. The new version of the watershed generated using the corrected river

> and lake shapefiles (without routing errors),

> 2. The previous version where I had the routing error,
> 3. The original watershed area that I provided to the model as the DEM

> boundary in the first step.
>
> Thank you very much for your help and support.
>
> Saba Shaghaghi
>
> On Friday, May 29, 2026 at 5:18:06 AM UTC-4 Khybeer wrote:
>
> This message indicates that QSWAT+ could not match the outlet point
> with ID 22 from your inlets/outlets shapefile to any stream channel
> in the generated channel network (demmchannel.shp). As a result, the
> outlet will be ignored during watershed delineation. This usually
> happens when the outlet is not located exactly on a stream, when
> stream threshold settings have changed after the outlet was created,
> or due to a temporary delineation error.
> First try clicking "Create Watershed" again as suggested by QSWAT+.

> If the error persists, *verify that the snapped outlet point lies

> directly on a stream in the channel shapefile and re-snap or

> recreate the outlet if necessary.*
> I hope this helps.
> Regards**

>
> On Tue, May 26, 2026 at 3:09 PM Saba Shaghaghi <sabash...@gmail.com>
> wrote:
>
> Hello,
>
> I am developing a SWAT+ model for the Peace River Basin in
> Florida using:
>

> * SWAT+ rev. 61
> * QSWAT+ 3.1
> * SWAT+ Editor 3.2
> * SWAT+ Toolbox 3.2

> * Linux executable: swatplus-61.0.2.61-gnu-lin_x86_64-Rel

>
> My watershed contains many lakes and reservoirs, so I tried both:
>

> 1. Burning in my own stream shapefile with lake shapefiles.
> 2. Letting TauDEM generate streams automatically from the DEM.

>
> However, in both cases I encounter problems either during
> delineation or when running the model on HPC for calibration.
>
> During delineation, QSWAT+ gives this error during “Constructing
> topology”:
>
> “ID value 22 from inlets/outlets file ... not found as DNOIDID
> in channels file ... Will be ignored.”
>
> It seems the model cannot correctly connect streams, lakes/
> reservoirs, and outlets.
>
> When running the model on Linux/HPC, it crashes with:
>
> “Floating-point exception - erroneous arithmetic operation”
> in:
> time_conc_init_ / proc_cha_
>
> After checking the generated files, I found:
>

> * all channel elevations in chandeg.con are zero,
> * many channel slopes are extremely small (0.00001),
> * some channel lengths are unrealistically short.
>
> Could you please advise:
>
> 1. Why channel elevations become zero?
> 2. What is the recommended workflow for flat watersheds with
> many lakes/reservoirs?
> 3. How should external stream and lake shapefiles be handled
> correctly in SWAT+?
> 4. Could these topology/routing issues be causing the floating

> point exception?
>
> Thank you very much for your help.
>
> Best regards,
>
> Saba Shaghaghi
>
> --
>
> You received this message because you are subscribed to the
> Google Groups "QSWAT+" group.
>
> To unsubscribe from this group and stop receiving emails from
> it, send an email to qswatplus+...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/

> qswatplus/8365c29c-449d-4137-92f1-38ff89b79619n%40googlegroups.com <https://urldefense.com/v3/__https://groups.google.com/d/msgid/qswatplus/8365c29c-449d-4137-92f1-38ff89b79619n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!KwNVnqRv!EKMb0pIwhNU4Py18nVy9TRR_HYa5VXC543Lr7fTkYjZ_ZpGcn6SPgg80P_tqSf_ZzSKbf53stbIWUTKKeaMSgGs$>.

>
>
>
> --
> Best Regards
> ________________________________
> Dr.-Ing. Fazlullah Akhtar
> Senior Researcher
> Center for Development Research (ZEF)
> Genscherallee 3, 53113
> Bonn
>

> --
> You received this message because you are subscribed to the Google
> Groups "QSWAT+" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to qswatplus+...@googlegroups.com

> <mailto:qswatplus+...@googlegroups.com>.

> To view this discussion visit https://groups.google.com/d/msgid/

> qswatplus/68a1761f-5fdb-4559-94ba-629974f2dc75n%40googlegroups.com
> <https://urldefense.com/v3/__https://groups.google.com/d/msgid/
> qswatplus/68a1761f-5fdb-4559-94ba-629974f2dc75n*40googlegroups.com?
> utm_medium=email&utm_source=footer__;JQ!!KwNVnqRv!
> EKMb0pIwhNU4Py18nVy9TRR_HYa5VXC543Lr7fTkYjZ_ZpGcn6SPgg80P_tqSf_ZzSKbf53stbIWUTKKuM_6zjc$>.