Hydrological calibration issue

11 views
Skip to first unread message

Anish Katwal

unread,
Jul 17, 2026, 11:53:25 AM (6 days ago) Jul 17
to ArcSWAT

Hi Everyone,

I am working on a SWAT model for the Cupsaw Lake watershed in Ringwood, New Jersey. It is a small forested lake watershed, about 10.7 km², and my final research goal is to estimate phosphorus loading to the lake, especially the possible contribution from septic/OWTS systems.

The main challenge is that the watershed does not have a direct stream gauge or measured lake inflow data. Because of this, I have been using a nearby donor watershed and drainage-area-ratio scaling to develop proxy inflow targets. However, since the donor watershed is larger than Cupsaw, I am not fully confident that direct daily or monthly flow matching is appropriate. I am now considering using the donor information more as seasonal hydrologic signatures rather than the exact observed flow.

After structural cleanup and several calibration tests, the model can improve overall NSE/KGE, but it still strongly overpredicts summer/late-summer inflow, especially the July–September period. For example, one recent run improved overall KGE to about 0.5, but JAS seasonal bias remained very high.

I am considering a multi-criteria calibration approach where I screen runs by overall PBIAS, KGE/NSE, seasonal PBIAS, JAS low-flow behavior, parameter plausibility, and pathway behavior such as SURQ, LATQ, GW_Q, PERC, and ET.

I would be very grateful for your advice on a few points:

  1. For a small ungauged lake watershed, is it reasonable to calibrate using donor-derived seasonal signatures rather than direct DAR-scaled daily/monthly flow?
  2. When summer low-flow is the main problem, would you recommend NSE, KGE, log-NSE, PBIAS, or a custom seasonal objective?
  3. Is it acceptable to retain a behavioral hydrology ensemble instead of choosing only one best NSE/KGE run?
  4. Are there any SWAT parameters or structural settings you would prioritize for persistent summer overprediction in a small forested watershed?

Any suggestions would be greatly appreciated.

Suraj Shah

unread,
Jul 17, 2026, 7:12:28 PM (6 days ago) Jul 17
to Anish Katwal, ArcSWAT
Hi 
I suggest you to use ET and soil moisture as an calibration target, as they are major moisture drivers in case of ungauged basin. 


Best Regards 
Er. Suraj Shah
Hydrological remote sensing Enthusiast
researchgatelinkedinfacebook

येत् भावो -तत् भवोति!! 


--
You received this message because you are subscribed to the Google Groups "ArcSWAT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arcswat+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/arcswat/51d7485d-df8a-433d-ba7c-8c56c1c1b544n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages