Workflow to calibrate from up to downstream?

[Afonso Augusto Magalhães de Araujo]

Below, I describe the procedures I'm adopting in an attempt to establish a consistent workflow for calibrating this project, from upstream to downstream and regionalizing the parameters. However, a series of doubts and questions arise along with seemingly inconsistent and uninspiring results. The project and the first three stations/channels to be calibrated are (Figure 1). 

After several tests in an attempt to discover how to “fix” the best parameters and advance the calibration downstream, and having tested that, when manually transferring these best parameters to a new project and running the manual calibration, the NSE was reproduced.

Thus, I adopted this as a valid procedure and calibrated the regions of channels 108 and 122 in different projects and transferred the parameters (“Current Best”) from each project to this new project to calibrate channel 89. In the individual projects I obtained the following Nash: channel 108 (NSE = 0.547) and channel 122 (NSE = 0.725).

Even though I was frustrated with the results, particularly for channel 108, I decided to move on to channel 89. However, after carefully entering the parameters for each channel and adding the initial parameters for channel 89, when running the manual calibration only the result for channel 108 seemed to be consistent, but that for channel 122 dropped from 0.725 to 0.630 (perhaps some error in the parameters), as shown in the following figure2.

But, what catches my attention the most is the fact that, despite only calibrating channel 89, the NSE for channels 108 and 122 also vary, probably impacting the calibration of channel 89. And, at the end of the calibration cycle, the result was the following (figure3).

From this "brief context," I followed up with some of the many open questions:
1. What is the correct workflow for this upstream-downstream calibration, regionalizing the parameters (by HRU and AQUs) and determining the best parameters?
2. What is the real purpose of being able to "export" the parameters for a given project? I've already tested this possibility and haven't identified any impact? How does it actually work?
3. Is there a more efficient way to edit the parameter list? For example, if we want to change the "Change Type" or the variation range?

Could you help me with these questions? Thank you.

[Celray James]

Dear Alfonso,

To have regionalised parameters, you need to set them as applying to specific objects. The interface helps with adding for HRUs and shallow aquifers but you can specify channels (for channel parameters) manually.

For each region, add the parameters selecting the area by map or manually adding the parameters...

When done, you can calibrate that region (manually or automatically) and then freeze the parameters by shrinking the parameter range (you can do this in the automatic calibration page as shown below)



If you prefer, you can manually shrink the parameter ranges by clicking on min and max in the parameters section.

Then you can move on to the next region > add parameters to region > calibrate > freeze making sure you go from upstream regions towards the watershed outlet. 

Note that the map is best used by highlighting the area around LSUs with HRUs you want and then clicking on the ones you want want to exclude one by one. You can also select one LSU and add it to list one at a time. If you have HRU List for each region, you can add them as a comma separated list too... (same applies for Channels and Aquifers)

Another Note, there is a SWAT+ Toolbox group for questions about calibration using SWAT+ Toolbox

C. James

---

From: qswa...@googlegroups.com <qswa...@googlegroups.com> on behalf of Afonso Augusto Magalhães de Araujo <afonso...@poli.ufrj.br>
Sent: Wednesday, October 8, 2025 15:14
To: QSWAT+ <qswa...@googlegroups.com>
Subject: [QSWAT+ User Group] Workflow to calibrate from up to downstream?

 

--
You received this message because you are subscribed to the Google Groups "QSWAT+" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qswatplus+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/qswatplus/4e1d0693-8cc9-4bd4-ab8e-385e7bbbbaden%40googlegroups.com.