SWAT-MODFLOW crashes during run

[YinLong Huang]

Hi all, 

I'm currently working on a SWAT-MODFLOW project and the coupled model keeps crashing at some random time. 

If I use ground elevation for my MODFLOW model, then the coupled model crashes at day 1, and if I use a constant head for each layer in MODFLOW then the model crashes at the end of second year.

I have another SWAT-MODFLOW model from someone else to compare my input data with his, and I found out that the linking files are fine, and modflow.mfn file is also good. So I wonder if the crash is related to the model convergence? I checked all solver files and they should be all good (tolerance adjusted to be maximum), I set all my recharge to be 0 for all periods, so I'm kind of confused right now about where else should I check.

Sorry for the inconvenience and hope to hear back from anyone. 

Best

Yinlong Huang

[Seonggyu Park]

Hi Yinlong Huang,

You can also check the overall water budget discrepancy in the MODFLOW *.lst, *list, or *out file whether it is over 1%. To avoid the model convergence failure, the solver must converge within limited outer and inner iterations. And the overall discrepancy needs to be less than 1%.

Best,

Seonggyu

[YinLong Huang]

Hi Seonggyu, 

Thanks for that information. I checked the .LST file and the percentage of discrepancy is below 1%. However, I did notice one warning message that says "FAILED TO MEET SOLVER CONVERGENCE CRITERIA IN TIME STEP  279 OF STRESS PERIOD  1", which might be the source of error. That confuses me because my river stage as well as recharge in MODFLOW are set to be constant, so is it because of the initial head that I set was problematic?

Thanks again for your suggestion! 

Best regards