Advice on Daily Streamflow Timing in SWAT Calibration

[Rozita pm]

Dear SWAT Group,

I am calibrating a SWAT model at a daily time step and am having difficulty reproducing the timing and recession of storm-flow peaks.

The model generally produces the storm response, but some simulated peaks occur about 1–2 days too early, and the recession is too fast compared with observed discharge.

I have tested several routing parameters one at a time:

• SURLAG: little to no effect
• CH_K2: little to no effect
• CH_N2: some improvement in delaying and attenuating the peak, but not enough to resolve the problem

For example, increasing CH_N2 reduced the early peak and slightly increased flow on the following days, but the simulated hydrograph still drops much faster than the observed hydrograph.

The model is already calibrated reasonably well at the monthly scale, so I am trying to improve the daily hydrograph without degrading the overall calibration.

Has anyone experienced a similar issue? Which parameters or processes would you recommend investigating next for delayed peak timing and a slower recession? I am considering ALPHA_BF, GW_DELAY, GWQMN, or other storage/routing-related parameters.

Any suggestions would be greatly appreciated.

Best,
Rozita