SWR model instability

[Althea Walker-Hallam]

Hi,

I was hoping I could get some advice about improving the stability of my SWR channel model. 

The grid and reach sizes seem comparative to other model examples, and the discrepancy is too high but not terrible. As soon as I introduced control structures into my model, it looks like the attached image. 

Removing or grouping very small reaches had no impact.

The instability occurs whether I use a constant flux or my real data.

I have spent considerable time experimenting with the available solver variables, which improved the discrepancy, but did not change this instability.

I would be very grateful if anyone has suggestions for what else I might try.

Regards,

Althea