Please help me at Sensitivity Analysis!

[fatemeh babakhani]

Hello All

I am completely new in the SWATCUP field, and I just started to do Sensitivity Analysis and then Calibrate the model based on the Sediment Component. And I face with some issue, I was wondering if you could help me at solving these problems.

1) At the first step of Sensitivity Analysis, should I choose ONE AT A TIME or GLOBAL SENSITIVITY ?! Am I supposed to change one parameter while keeping other parameters constant, then performing this process (ONE AT A TIME) for other parameters and finally check some numbers like p-factor, r-factor,... to see which parameters are more sensitive?

2) How many parameters are appropriate to choose for doing Sensitivity Analysis?! How can I choose their range value? (Can I use the default value of SWATCUP in "Text View"?)

I read the SWATCUP user manual (Specially p.49- 56) several times, but unfortunately I couldn't find my answer in the user manual! Could you please introduce me any paper, textbook or any documents that help me at doing Sensitivity Analysis in the SWATCUP....

Thank you so much,

Fatemeh

[Iman Fazeli]

Dear Fatemeh,

1) It depends on your goals for choosing One At Time or Global methods but I usually use One At Time. For One At Time method you should select three runs of every parameter and see the result. If the parameter is changed, it is sensitive.

2) You could test the parameters for your purpose. for example if you would like to check flow out and sediment you should seeking for flow out and sediment parameters to identify effective parameters. you could see range value parameters in Absolute SWAT Values file.  Some parameters that you could check out for sediment are:

PRF.bsn

ADJ_PKR.bsn

Lat_Sed.hru 

USLE K & P

hope this can helps,

Iman

--
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.

[Abbaspour, Karim]

Dear Fatemeh,

If you search the literature (and the references provided with the SWAT-CUP) you will find all the sensitive parameters to flow, sediment, N, P, etc.

But for your project, at the beginning if you are not sure a parameter is sensitive or not, then you can do a one-at-a-time sensitivity analysis to find that out.

Once you chose a whole set of important parameters you can make an iteration (consisting of say 500 simulations) and then do a global sensitivity analysis to see how all parameters perform when they are all changing simultaneously. This provides you with a different level of understanding of parameter sensitivity.

Also, after each iteration, the dotty plots can be consulted to visualize how a parameter performs as its value changes across the parameter range.

But both one-at-a-time and global sensitivity methods have their own disadvantages. Pls see the manual.

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 58 856 5359
fax: +41 58 856 5375
http://www.eawag.ch/index_EN

--

[fatemeh babakhani]

Dear Dr. Karim and Mr. Iman

Thank you so much. Now it is much more clear to me!

[Jiajia Liu]

Hi Dr Abbaspour, 

First, thank you for your amazing work and answering all of the questions here.

Second, after I finish my OAT sensitivity analysis and move on to the second step global sensitivity analysis, should I include the "not very sensitive on its own" parameters? Since the goal for global sensitivity analysis to understand the interaction between parameters, in theory, a not very sensitive parameter using OAT can be sensitive to changes made for other parameters.

Third, when I conduct global sensitivity analysis using SUFI-2, after parameters have been selected, should I just set the upper and lower bound of parameters to their model feasible range? or should i limited it to a "reasonable range for the location". 

Thank you very much!

Jiajia