Need guidance on poor SWAT-CUP calibration results (FLOW_OUT – Monthly)

[Malik Waseem Iqbal khokhar]

Dear All,

I am working on hydrological modeling of the Neelum Watershed (Upper Indus Basin, Pakistan) using SWAT 2012 with SWAT-CUP (SUFI-2) for calibration.

I am calibrating monthly streamflow (FLOW_OUT) at reach 22 (outlet) for the period 1987–2004. However, my calibration results are not satisfactory.

Current statistics (Best Simulation):

• NS = -0.44

• R² = 0.32

• P-factor = 0.11

• R-factor = 0.00

• PBIAS = 68.2

• KGE = -0.23

• RSR = 1.20

Mean simulated flow = 102.68
Mean observed flow = 322.75

The model is severely underestimating discharge.

I have already:

• Checked observed data format (monthly, no missing values)

• Confirmed correct outlet reach number

• Adjusted basic flow-related parameters (CN2, ALPHA_BF, GW_DELAY, GWQMN, CH_N2, CH_K2)

• Increased number of simulations

Still, results are not improving significantly.

Could anyone please guide:

1. Which parameters should be prioritized for mountainous snow-fed basins?

2. Should I first perform sensitivity analysis separately?

3. Any specific recommendations for snowmelt-related parameters (SMFMX, SMFMN, SFTMP, SMTMP, etc.)?

Any guidance would be highly appreciated.

Thank you in advance.