SWAT+ Toolbox's results significantly differs from SWAT+ Editor. What is wrong with it?

Ehsan Qasemipour

unread,

Mar 1, 2021, 10:52:44 AM3/1/21

to SWAT+

Hello everyone,

I just modelled a catchment in SWAT+ Editor and the results showed good agreement with the observed data. But when I tried to calibrate the model using the SWAT+ Toolbox, there was a very big difference between simulated and observed data with NSE=-0.5!

It should be noted that I am using the latest versions of both SWAT+ Editor (2.0.0, and SWAT+ model rev. 60.5.2) and SWAT+ Toolbox (v0.7.5).

Have you any idea what is wrong with my this?

Thanks,

Ehsan

ch51_calibration.png

ch51.png

celray.chawanda

unread,

Mar 1, 2021, 10:58:55 AM3/1/21

to Ehsan Qasemipour, SWAT+

Can you send me your entire project on we transfer?

use celray[at]openwater.network

James

--
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/97f9f9d0-48b7-4899-8ea6-39f127907055n%40googlegroups.com.

Celray James CHAWANDA

unread,

Mar 1, 2021, 2:33:20 PM3/1/21

to SWAT+

Dear All,

Here is what the problem is. The results of channel flows can be found in either channel_sdmorph files or channel_sd files. For daily flows, the channel flows are exactly the same, but for monthly, and yearly flows, they are different.

channel_sdmorph flows are averaged while channel_sd flows are added up, so the monthly and yearly flows in channel_sd are incorrect. Keeping this in mind, QSWAT+ gets its flows from channel_sd while SWAT+ Toolbox gets its flows from channel_sdmorp. Thus the graph you get for flows in QSWAT+ is NOT correct if it is for monthly and yearly flows. I have attached an excel file if you are interested to see why the monthly flows from channel_sd are not correct.

James

Why_SD_MonthlyFlowsAreWrong.xlsx

Ehsan Qasemipour

unread,

Mar 2, 2021, 12:32:13 PM3/2/21

to Celray James CHAWANDA, SWAT+

Hello James,

I followed your descriptions and I understood the differences between the sd and sd_morph values. I tried to validate your explanations in three different projects but I wasn't able to find any differences. Here are attached the images. Is there anything else I should consider?

Best Regards,

Ehsan

--

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/70a1adb3-47e0-4c28-aca8-f35062924472n%40googlegroups.com.

ch1_flo_out_Toenepi.png

ch46_flo_out_Dashahe.png

ch71_flo_out_Mohaka.png

Kean Michael Cabigao

unread,

Mar 26, 2022, 6:59:20 AM3/26/22

to SWAT+

Hello,

Ive also observed a difference in NSE values between the SWAT+ toolbox graph and Visualize plot in QSWAT+. However, in my case, I have been using the daily flows. Wherein, as stated in this conversation, both channel_sd and channel_sdmorph for the daily time step have the same values. Or is the negative values in my observed data (considered to be missing values) affect the computation in the NSE values on either SWAT+ Toolbox/QSWAT+?

Thanks.