I have calibrated my SWAT model with monthly observed flow data, and the NSE coefficient is as high as 0.85 at monthly scale, while when I plot a FDC curve with daily flow data, it does not match well. Generally, the model overestimate the peak flow for a flood process, and the variability of simulated flow is much higher than that of observed, therefore, I want the simulated flow to be more smoother.
I think the SURLAG parameter from the basins.bsn file should be responsible for this.
I have changed the SURLAG from the default value of 4 to 1, 8,22, but it does not result in any difference in FLOW_OUTcms. I wonder why?
Is there any other responsible for this ?
Dear Hsang Jaknr Hieng ,
As per my view and perception since you are working on continuous simulations, peak flows and daily variations of flow are not much important. For continuous simulations check with N.S.E and total streamflow volume dear. If it was an event scale study (flood modeling) then peak flow and daily variations are of major importance.
Not only with SURLAG check with OV_N as well.
Hope my answer would help you
Thanks and with regards
Miyuru
In fact, I am aiming to analysis hydrological extremes, so the peakflow is important for me. I have also noticed the parameter OV_N, CH_N2, and CH_N1, but I think SURLAG is directly responsible for the peak flow, but it seems not works for my project, I wonder why? Is there other options wrongly closed by me?
Dear Hsang Jaknr Hieng ,
Usually CH_N2 and CH_N1 (Manning's roughness) are not used for calibration. Even many papers use CH_N2 and CH_N1 as calibration parameters, These values should be estimated with the recommendations of observed field investigations or else from recommended ranges of literature.
Yes you check with SURLAG.
Thanks and with regards
Miyuru
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India