SWAT and SWAT CUP calibration using SUFI2

[azreen azman]

Hello all.

I have some questions regarding my analysis using Arc Swat and Swat Cup. At my study area, there are 2 dams. The discharge from the first dam (6979ha) flows into the second dam (36900ha). The second dam operated since 1985 while the first dam operated starting from 2015. I only have one river flow observed data point (1988-present) to use in calibration process. I watched your Youtube about Swat and Swat Cup. My questions are:

1. When I want to key in the reservoir data in Swat such as RES_ESA, RES_EVOL, RES_PSA, RES_PVOL and RES_VOL, there are the maximum value that I can key in which are very small than the exact value of my study area information. Is it okay if I just put the maximum values? I am afraid that the values will affect my simulation because they are too small than the exact values.

2. For the calibration part, I used SUFI 2. In Youtube, the values that have to look into are p factor (>0.7) and r factor (close to 1). But, almost all journals that I read, refer to r squared and NSE values to indicate that the calibration is good enough before I can proceed to the validation part. My calibration only has the r squared value around 0.3. I repeated the analysis several times by adjusting the minimum and maximum values of the parameters but the r squared still remains around 0.3.

I really hope you can help me to figure these out. All your help and answers are much appreciated. Thank you.

[2w2e]

- If you have discharge from the dam, then you don't really need all those dam information. SWAT cannot really do dam  management for you based on those information. 

- Your model is uncertain. You need to look at p-factor and r-factor. R2 and NS are only for reference. Using parameters associated with the best R2 or NS could be very misleading. 

I see that a lot of people use only R2 or NS, etc. Because of that, I wrote a short article, which will be published soon in STOSEN. I will send a copy when its out.

Karim 

[azreen azman]

Thank you so much for the feedback. Really appreciate it Prof. Can I get the clear picture in choosing the parameters in streamflow calibration? From my reading, the most sensitive parameters in streamflow calibration are CN2, SURLAG, SOL_AWC,SOL_K, CH_N2, ESCO, EPCO, ALPHA_BF,GW_DELAY, GW_REVAP, OV_N, GW_QMN, REVAPMN, RCHARG_DP AND REVAPMN. Based on your Youtube video, you mention that each iteration does not have to perform with many parameters.

1. Does it mean that I have to separate the parameters based on the file name for each iteration? I get confused because the values of p-value and t-stat will change for iterations.

2. I still can't get the clear picture in using the r__,v__ or a__ in adjusting the min and max values of the parameter. Can someone explain more about this?

Thank you so much.

--
You received this message because you are subscribed to a topic in the Google Groups "SWAT-CUP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/swat-cup/bYKoqK_2Qeg/unsubscribe.
To unsubscribe from this group and all its topics, send an email to swat-cup+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/swat-cup/39b4b886-9420-46ca-9318-d8ae5f881629n%40googlegroups.com.