SWAT+ GWFlow Crash – Floating Overflow Error (gwflow_soil.f90) – Urgent Setup Help Needed

47 views
Skip to first unread message

Benthe

unread,
Jul 9, 2025, 9:11:28 AM7/9/25
to SWAT+

Hi all,

I’m currently setting up a SWAT+ GWFlow model for my Master’s thesis, using QSWAT+ and the 2024 tutorial by Bailey 2024 as a guide. The goal is to simulate groundwater flooding and nitrate dynamics in an agricultural watershed (~542 km²) in Lower Austria. I'm following the 2024 tutorial by Ryan Bailey, and while the setup was going smoothly initially, I now encounter a fatal crash with a floating overflow error when running SWAT+ with GWFlow enabled.

As I’m working under a tight thesis deadline, I would greatly appreciate any guidance from others who’ve encountered similar problems.


forrtl: error (72): floating overflow
Image              PC                Routine            Line        Source            
rev61.0.1_64debug  00007FF79A297674  GWFLOW_SOIL                69  gwflow_soil.f90
rev61.0.1_64debug  00007FF798D58639  SWR_PERCMAIN               65  swr_percmain.f90
rev61.0.1_64debug  00007FF799CCB5C0  HRU_CONTROL               323  hru_control.f90
rev61.0.1_64debug  00007FF798832B82  COMMAND                   254  command.f90
rev61.0.1_64debug  00007FF798AF772A  TIME_CONTROL              255  time_control.f90
rev61.0.1_64debug  00007FF799E6AF1A  MAIN__                    134  main.f90
rev61.0.1_64debug  00007FF79A60B0EB  Unknown               Unknown  Unknown
rev61.0.1_64debug  00007FF79A60B819  Unknown               Unknown  Unknown
rev61.0.1_64debug  00007FF79A60B732  Unknown               Unknown  Unknown
rev61.0.1_64debug  00007FF79A60B5EE  Unknown               Unknown  Unknown
rev61.0.1_64debug  00007FF79A60B88E  Unknown               Unknown  Unknown
KERNEL32.DLL       00007FF85056E8D7  Unknown               Unknown  Unknown
ntdll.dll          00007FF850B9C34C  Unknown               Unknown  Unknown 

Since there is not yet documentation on the SWAT+gwflow I am having trouble interpreting the problem. 

Based on this forum I did the following checklist

  • All raster inputs in the same CRS (EPSG:4326).

  • DEM covers area larger than the watershed.

  • All layers clipped and aligned using QGIS.

  • HRUs and subbasins created successfully.

  • GWFlow .ini and parameter files follow the 2024 tutorial format.

  • Step 2 was successful only crashes when GWFlow isrun in SWAT+ editor for step 3. 

Questions:

  1. What specific input values (e.g., aquifer thickness, initial head, permeability) are known to trigger floating overflow errors in gwflow_soil.f90?

  2. Is it valid to have 0 or near-zero values in the aquifer thickness raster? If not, what is the recommended minimum value, and should I mask or replace 0s

  3. Does having NA values in the HRU table for LSU_ID (despite a valid LSU layer) interfere with GWFlow execution? If yes, how can this be resolved or avoided?

  4. How should initial groundwater head values be estimated for a real watershed considering export txt file? 

  5. Is there a recommended workflow to validate key GWFlow input rasters (thickness, permeability, initial head) before running the model? Are there known value ranges or consistency checks?

What I Did: 

  1. Watershed Delineation and DEM

  • Used a high-resolution DEM from Copernicus (30m).

  • The DEM covers an area larger than the watershed, as recommended.

  • Delineated 40 subbasins over ~542 km² using QSWAT+.

2. Land Use, Soil, and Slope

  • Input layers derived via Google Earth Engine, following Adrián López-Ballesteros’s YouTube tutorial.

  • Verified CRS and raster alignment (EPSG 32233 - UTM zone 33N)

  • Chose no filter or threshold HRU method; generated 343 HRUs (Step2) since I wanted to maximize recharge to gridcell mapping since it is is a small study area compared to Tutorial 2024 study area and as advised by youtube tutorial. Yet, I tried to test other threshold setups as well (from tutorials), but this didn’t affect the error.

3. Aquifer Thickness and Permeability

  • Used GLHYMPS global aquifer thickness and permeability rasters.

  • Original thickness raster had values ~100x too high — likely due to European decimal commas issue. So I corrected it using QGIS Raster Calculator, then re-exported as GeoTIFF. This showed a more realisitc aquifer thickness of this floodplain. However in the SWAT+editor export of aquifer thickness  TXT the values were often 0 or very low like 0.03, which might be problematic.

4. LSU Layer and HRU Table

  • The LSU layer appears correctly in GIS and QSWAT+, but after HRU creation, the HRU attribute table shows NA for LSU_ID.

  • Not sure if this causes problems for SWAT+ or GWFlow (no documentation available).

5. Initial Groundwater Head

  • Auto-generated from SWATgwflow  in SWAT+interface model export; values go up to 350 m. I’m unsure whether these values represent depth or elevation, or whether they need manual adjustment. They do match the boundaries of the Study area.

6. Weather Generator

  • Used SWAT+ global weather generator.

7. GWFlow Parameters

  • Based on Bailey’s 2024 tutorial gwflow.ini and groundwater parameter files.

  • No manual edits to parameter files beyond those covered in the tutorial 


This issue is blocking my thesis progress, and I would be incredibly grateful for any advice, file examples, or similar experiences you can share.

Thank you in advance for your help!

Kind regards,


Benthe Thielen


MSc trainee – Institute for Hydrology and Water Management
University of Natural Resources and Life Sciences (BOKU), Vienna




Reply all
Reply to author
Forward
0 new messages