SUFI2 cannot calibrate the SWAT model, but I can do it Manually!

[Hadi]

I have a SWAT model that its primary performance is pretty good (with NSE=0.3 and R^2=0.45) and manually changing the CN and Groundwater delay parameters can calibrate/validate the Model. But When I am using the SUFI2 at SWAT-CUP it just deteriorates the results (NSE= -0.06 and R^2=0.03).
- I have given it several tries, but none gave a better outcome.
- Even when I define the parameter's band (value) in the Par_inf.txt and tie them to the manually calibrated parameters values, it still cannot give a satisfactory result!
- I am quite sure that the Sufi2 setup is right according to the instructions.
However,
Could you please advise me on the possible wrong doings at running the SUFI2! or what could be the problem here?

Thanks.
Hadi

[Abbaspour, Karim]

Perhaps two different swat.exe versions are being use..

--
You received this message because you are subscribed to the Google Groups "SWAT-CUP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swat-cup+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Hadi]

Dear Dr. Abbaspour,

Thanks. I followed your comment, if I have understood you correctly, and the results is as follows;

I replaced the "swat.exe" of SWAT-CUP in a SUFI2 project with the previously used SWAT.exe of my ArcSWAT. I renamed the swat2012.exe file according to SWAT-CUP naming format [swat.exe]. This way I tried to make sure the same swat.exe is used for both ArcSWAT and SWAT-CUP.

- However, the results were the same as before (NSE= -0.16 and R^2=0.02). It is actually not meeting the threshold level of 0.5 for NSE and not producing the "beh" file!
- I have also attached the 95ppu plot of SWAT-CUP calibration and my manual calibration results [Blue= Observation, Red=Simulation].
- I have even tried this on three different computers for dozens of times, but results were the same.

I hope you and the other researchers can help me with some other ideas on this case.

Thanks,
Hadi

[Hadi]

I need to add that the swat.exe version of the both ArcSWAT and SWAT-CUP has been 64-bit. However, as we are able to choose between swat.exe Debug or Release versions in ArcSWAT, I just made sure that both programs use Release version.

- Still the result is bad and nowhere near calibration!
- I have also run the SWAT-CUP program with initial swat parameter values at iteration one (simulation no = 1, parameters no =1, r__CN2 [0, 0]) and yet the simulation is far different than ArcSWAT simulations (Flow_out)!
- Both the SWAT-CUP and ArcSWAT are now running swat.exe version 627, the only unknown is if I should use the debug or release version at ArcSWAT to match with SWAT-CUP.

[Karim Abbaspour]

well if you did make any runs with swatcup, then probably the parameters are different from the original textinout. Just copy everything from the backup directory into the main project directory. If you are running the same swat version, then you should get the same answers. It does not matter if you run debug version or not.

Best, Karim

 

-------------------------------------------------
Dr. K.C. Abbaspour
Eawag, Swiss Federal Institute for Aquatic Science and Technology
Ueberlandstr. 133, P.O. Box 611, 8600 Duebendorf, Switzerland
email: abba...@eawag.ch
phone: +41 44 823 5359
fax: +41 44 823 5375
http://www.eawag.ch/index_EN

[Hadi]

Dear Dr. Abbaspour,

I am sincerely grateful for your help and guidance on users problems with the program.

Report on my case:
After having followed your comments I found that there is a problem with my SWAT-CUP project set up which is explained below and I hope it would be helpful to others as well.

My SWAT Project simulation years: 1980 to 2012
Warm up: 10 Years
Calibration period: 2000-2007
Validation period:  2008-2012

in SWAT-CUP (SUFI2) I should have followed the Blue format which instead I had used the Red one:
1990  2000         : beginning year of simulation not including the warm up period
2007  2007         : end year of simulation
Observation format: First column is a sequential number from beginning of the simulation (1990 in my case)
121    FLOW_OUT_1_2000    10.2
122    FLOW_OUT_2_2000    9.1
124    FLOW_OUT_4_2000    15.9
1    FLOW_OUT_1_2000    10.2
2    FLOW_OUT_2_2000    9.1
4    FLOW_OUT_4_2000    15.9
(Users: Please remember that missing data should be considered in the numbering order as above)

and I could benefit from the program to improve my SWAT model calibration.

Thank you very much.
Hadi

[Diva Bhatt]

Thank you Hadi...Your Report on your SWAT-CUP run has been really a good and useful guide to me...Thank you very much!

Diva

[Mamush Tekle]

Dear Kari, I calibrated and validated the SWAT model with SWAT-CUP , then I try to run the SWAT-CUP for climate change scenario , The swat-cup stops at first run with forrtl: error(72). when I run the SWAT-CUP folder with the SWAT2012_debug.exe it runs for well. Now my question is that possible to run the calibrated model  with swat2012_debug.exe run instead of SWAT2012? can it gives the same result as the swat-cup run for the climate change scenario?

[Atiqullah Azad]

Dear SWAT-CUP Community,

I’m facing a puzzling issue and would appreciate your insights.

A few days ago, I calibrated the USGS watershed 06820500 using SWAT-CUP (SUFI-2) and achieved excellent results (e.g., NSE = 0.931, R² = 0.94). Today, I re-ran the exact same calibration — without changing any files, parameters, bounds, or settings — yet the results are significantly worse (e.g., NSE dropped to ~0.52, high PBIAS).

Could you help me?

It is worth noting that I currently have no previous parameter sets.