HELP PLEASE: SWAT CUP Sensitivity, Calibration and Validation

[AAbino]

Good day to all!

I am new in SWAT CUP, actually in SWAT per se. At present, I tried to run the SWAT2009 using 21-year weather inputs (1991-2011).

For the sensitivity analysis, calibration and validation, I decided to use the SWAT CUP upon experiencing series of errors in the built-in sensitivity analysis, calibration and validation in SWAT2009 interface.

I was successful in running the SWAT CUP. However, I have some clarifications to check if the process that I have done was correct, as to wit:

FOR THE CALIBRATION

• File.Cio
• 21       | NBYR : Number of years simulated [is it correct to input 21 years even if my target of year of calibration is only 7 years? since the observed data that I have is only for 7 years]
• 1991    | IYR : Beginning year of simulation [is this the beginning year of simulation that I used in SWAT2009? or the beginning year of simulation of my target years i.e. 1996 since my NYSKIP = 5 years?]
• SUFI2_extract_rch.def
• output.rch     : swat output file name
• 1              : number of variables to get [i only have observed values for FLOW that is why used 1 in this section]
• 7              : variable column number(s) in the swat output file (as many as the above number)
• 17              : total number of subbasins in the project
• 1              : number of subbasins to get for the first variable
• 1              : subbasin numbers for the first variable
• 1996           : beginning year of simulation not including the warm up period
• 2002           : end year of simulation [is it the end year that I have observed data?]
• Observed.txt
• 1     : number of observed variables
• 6     : Objective function type, 1=mult,2=sum,3=r2,4=chi2,5=NS,6=br2,7=ssqr
• 0.5   : min value of objective function threshold for the behavioral solutions [what is the suggested min value of objective function threshold for the behavioral solutions? When I used 0.5, because after the SUFI2_post.bat, there is a message that Sim6.Suf i2.SwatCup>SUFI2_95ppu_beh.exe Cannot open input file /SUFI2.OUT/no_beh_sims.txt There were no solution with the objective function value >= the behavioral threshold?]
• Calibration Outputs
• 95ppu-No-Observed plot
• there is an error which said: "....Sim6.Sufi2.SwatCupIterationsIter1_newparsSUFI2.Out95ppu_g_No_Obs.txt does not exist" even though I set  SED_OUT (column 11) in the No Observation section

FOR THE VALIDATION

• Do I also need observed values for the validation?

Thank you in advance and I am looking forward on your immediate response. :D

[Karim Abbaspour]

File.Cio

• 21       | NBYR : Number of years simulated [is it correct to input 21 years even if my target of year of calibration is only 7 years? since the observed data that I have is only for 7 years]      You have data for 7 years, then why run for 21 years? But you can there is not problem
  
• 1991    | IYR : Beginning year of simulation [is this the beginning year of simulation that I used in SWAT2009? or the beginning year of simulation of my target years i.e. 1996 since my NYSKIP = 5 years?]  This is in swat. If you skip 5 years, this means swat runs from 1991 but there is no output, therefore swat-cup does not care for these years...  beginning year in swat-cup is 1996
  
• SUFI2_extract_rch.def

• output.rch     : swat output file name

• 1              : number of variables to get [i only have observed values for FLOW that is why used 1 in this section] yes
  

• 7              : variable column number(s) in the swat output file (as many as the above number)
• 17              : total number of subbasins in the project
• 1              : number of subbasins to get for the first variable
• 1              : subbasin numbers for the first variable
• 1996           : beginning year of simulation not including the warm up period

• 2002           : end year of simulation [is it the end year that I have observed data?]   No, it should be the end year of swat simulation
  
• Observed.txt

• 1     : number of observed variables
• 6     : Objective function type, 1=mult,2=sum,3=r2,4=chi2,5=NS,6=br2,7=ssqr

• 0.5   : min value of objective function threshold for the behavioral solutions [what is the suggested min value of objective function threshold for the behavioral solutions? When I used 0.5, because after the SUFI2_post.bat, there is a message that Sim6.Suf i2.SwatCup>SUFI2_95ppu_beh.exe Cannot open input file /SUFI2.OUT/no_beh_sims.txt There were no solution with the objective function value >= the behavioral threshold?]     What do you think the minimum value of the objective function for a good simulation should be? This is that number..As an analyst you need to decide, look in literature there is no fixed number, NS and R2> 0.7 are suggested to be good... for bR2 if it is > 0.4 or even 0.3, then that may be quite good...
  
• Calibration Outputs
• 95ppu-No-Observed plot
• there is an error which said: "....Sim6.Sufi2.SwatCupIterationsIter1_newparsSUFI2.Out95ppu_g_No_Obs.txt does not exist" even though I set  SED_OUT (column 11) in the No Observation section     ??
  

FOR THE VALIDATION

• Do I also need observed values for the validation?  of course!
  

Thank you in advance and I am looking forward on your immediate response. :D

--
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/groups/opt_out.
 
 

[AAbino]

Dear Dr. Karim,

Your immediate response on my queries is very much appreciated.
Likewise, thank you for the clarifications.

However, I still have further questions, if I may. [see below in green font]

On Monday, April 29, 2013 4:50:27 AM UTC+9, k_abb...@yahoo.com wrote:

File.Cio

• 21       | NBYR : Number of years simulated [is it correct to input 21 years even if my target of year of calibration is only 7 years? since the observed data that I have is only for 7 years]      You have data for 7 years, then why run for 21 years? But you can there is not problem [copy then. I will change the NBYR into 5 years and will use the remaining 2 years for the validation]
  
• 1991    | IYR : Beginning year of simulation [is this the beginning year of simulation that I used in SWAT2009? or the beginning year of simulation of my target years i.e. 1996 since my NYSKIP = 5 years?]  This is in swat. If you skip 5 years, this means swat runs from 1991 but there is no output, therefore swat-cup does not care for these years...  beginning year in swat-cup is 1996 [copy]
  
• SUFI2_extract_rch.def

• output.rch     : swat output file name

• 1              : number of variables to get [i only have observed values for FLOW that is why used 1 in this section] yes [copy]
  

• 7              : variable column number(s) in the swat output file (as many as the above number)
• 17              : total number of subbasins in the project
• 1              : number of subbasins to get for the first variable
• 1              : subbasin numbers for the first variable
• 1996           : beginning year of simulation not including the warm up period

• 2002           : end year of simulation [is it the end year that I have observed data?]   No, it should be the end year of swat simulation [copy.. will change to 2011]
  
• Observed.txt

• 1     : number of observed variables
• 6     : Objective function type, 1=mult,2=sum,3=r2,4=chi2,5=NS,6=br2,7=ssqr

• 0.5   : min value of objective function threshold for the behavioral solutions [what is the suggested min value of objective function threshold for the behavioral solutions? When I used 0.5, because after the SUFI2_post.bat, there is a message that Sim6.Suf i2.SwatCup>SUFI2_95ppu_beh.exe Cannot open input file /SUFI2.OUT/no_beh_sims.txt There were no solution with the objective function value >= the behavioral threshold?]     What do you think the minimum value of the objective function for a good simulation should be? This is that number..As an analyst you need to decide, look in literature there is no fixed number, NS and R2> 0.7 are suggested to be good... for bR2 if it is > 0.4 or even 0.3, then that may be quite good...  [copy]
  
• Calibration Outputs
• 95ppu-No-Observed plot
• there is an error which said: "....Sim6.Sufi2.SwatCupIterationsIter1_newparsSUFI2.Out95ppu_g_No_Obs.txt does not exist" even though I set  SED_OUT (column 11) in the No Observation section     ?? [below is the screencap of the message I am referring to]

 

                       In the No Observation section i.e. Extract_rch_No_Obs.def, I specified:

SUFI2          : SWAT-CUP program: SUFI2, GLUE, Para_Sol, PSO, MCMC

output.rch     : swat output file name

1              : number of variables to get (such as: discharge, sediment, ET etc.)

11             : variable column number(s) in the swat output file (as many as the number above)

17             : total number of subbasins in the project

1              : number of subbasins to get for the first variable

All            : HRU numbers for the first variable. Write "All" if equal to the total number of HRUs

1996           : beginning year of simulation not including warm up period

2002           : end year of simulation

and in the Var_file_rch_No_Obs.txt, I inputted 

SED_OUT_1.txt

SED_OUT_2.txt

SED_OUT_3.txt

SED_OUT_4.txt

SED_OUT_5.txt

SED_OUT_6.txt

SED_OUT_7.txt

SED_OUT_8.txt

SED_OUT_9.txt

SED_OUT_10.txt

SED_OUT_11.txt

SED_OUT_12.txt

SED_OUT_13.txt

SED_OUT_14.txt

SED_OUT_15.txt

SED_OUT_16.txt

SED_OUT_17.txt

Is there something wrong with what I have done that resulted to the error in 95ppu-No-Observed plot?

FOR THE VALIDATION

• Do I also need observed values for the validation?  of course! [copy.. is it okay to use 2 years of observed data for validation?]
  

THANK YOU VERY MUCH! 

 
 

[zq_cathy]

My answers is as follows, and my work is done sucessfully. and i want to know whether k_abb is right or not?
在 2013年4月28日星期日UTC+8下午8时16分50秒，AAbino写道：

Good day to all!

I am new in SWAT CUP, actually in SWAT per se. At present, I tried to run the SWAT2009 using 21-year weather inputs (1991-2011).

For the sensitivity analysis, calibration and validation, I decided to use the SWAT CUP upon experiencing series of errors in the built-in sensitivity analysis, calibration and validation in SWAT2009 interface.

I was successful in running the SWAT CUP. However, I have some clarifications to check if the process that I have done was correct, as to wit:

FOR THE CALIBRATION

• File.Cio

• 21       | NBYR : Number of years simulated [is it correct to input 21 years even if my target of year of calibration is only 7 years? since the observed data that I have is only for 7 years]

• 1991    | IYR : Beginning year of simulation [is this the beginning year of simulation that I used in SWAT2009? or the beginning year of simulation of my target years i.e. 1996 since my NYSKIP = 5 years?] my answer is 1991.
• SUFI2_extract_rch.def

• output.rch     : swat output file name
• 1              : number of variables to get [i only have observed values for FLOW that is why used 1 in this section]
• 7              : variable column number(s) in the swat output file (as many as the above number)
• 17              : total number of subbasins in the project
• 1              : number of subbasins to get for the first variable
• 1              : subbasin numbers for the first variable
• 1996           : beginning year of simulation not including the warm up period

• 2002           : end year of simulation [is it the end year that I have observed data?] Yes

• Observed.txt
• 1     : number of observed variables
• 6     : Objective function type, 1=mult,2=sum,3=r2,4=chi2,5=NS,6=br2,7=ssqr
• 0.5   : min value of objective function threshold for the behavioral solutions [what is the suggested min value of objective function threshold for the behavioral solutions? When I used 0.5, because after the SUFI2_post.bat, there is a message that Sim6.Suf i2.SwatCup>SUFI2_95ppu_beh.exe Cannot open input file /SUFI2.OUT/no_beh_sims.txt There were no solution with the objective function value >= the behavioral threshold?]
• Calibration Outputs
• 95ppu-No-Observed plot
• there is an error which said: "....Sim6.Sufi2.SwatCupIterationsIter1_newparsSUFI2.Out95ppu_g_No_Obs.txt does not exist" even though I set  SED_OUT (column 11) in the No Observation section