SWAT+ NSE

[Altin Bektashi]

Hello everyone,

I have built my model for a 1500 km2 catchment and my NSE for the simulated vs observed discharge in the visualise interface is 0.99 and PBIAS is also 0.99 without any calibration. I feel this is very unlikely. Can someone help me in this direction? Moreover, when I upload the TextInOut file in the SWAT+ toolbox it gives me completely different (negative) results as compared to the plot of NSE in "Visualise". Did anyone have a similar problem or has any suggestion? Thank you in advance.

[Natalja C.]

Hello,

I do not understand the issue :)

1. Can you confirm the NSE and PBIAS is so high by yourself? Export the flows into xls (or any other tool) and calculate those yourself.

2a. If you confirm this - congratulations. --> There is o issue.

2b. If you get different values - that means that something is wrong with your interface/tools/methods that you are currently using to perform the statistics. --> you have to contact the support of the tools creators or figure it out on yourself :)

Best,

Natalja

[Altin Bektashi]

Hi Natalja,

Thanks for your response. I have prepared csv files for daily and monthly discharge data in my outlet point to look at the model performance. Then I plotted my channel's results against the observed data in step 4 (visualise). Here I get .99 for NSE and .99 for PBIAS too, before even calibrating or changing any parameters in the model. I have triple checked both the input data from the respective stations in my catchment as well as the observed discharge. However, when I go to the swat+ toolbox and plot the results I get an NSE of -3.45.

I was wondering if was there anything wrong with the swat+ editor or the toolbox. Getting almost perfect NSE results without calibrating in step 4 and the negative result from the toolbox both seem very unrealistic to me. I do not know to proceed from this point.

Thank you

--
You received this message because you are subscribed to the Google Groups "SWAT+" group.
To unsubscribe from this group and stop receiving emails from it, send an email to swatplus+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/swatplus/de99e465-9dc0-40d4-8b0e-708be2668641n%40googlegroups.com.

[Natalja C.]

Try extracting the data yourself from the swat output files, not from any of the editors.

Navigate to your swat+ project, locate the file in question (I assume it is channel_sdmorph for flow) and extract the flow_out column yourself and calculate the NSE, PBIAS, R2, whichever by yourself using the excel, R, python, any other direct tools.

Then you will know who is correct, SWAT+toolbox or the editor.

Best,

Natalja

[Altin Bektashi]

Hi Natalja,

OK, I will try that and see the results. In my case I have to use channel_sd_mon or channel_sd_day. 

Thank you so much for your help.

To view this discussion on the web visit https://groups.google.com/d/msgid/swatplus/9d92f9c9-0157-42b2-af94-b0f263f42e61n%40googlegroups.com.