EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Water Quality Mixing Model (WQMM) for Environmental Flow Release Monitoring

Gabriel Sentlinger
Gabriel Sentlinger
  • Fathom Scientific Ltd., R&D, Bowen Island, Canada (

Environmental Flow Release monitoring can be an expensive undertaking in active watercourses normally suitable for run-of-river hydropower projects.  In order to attain acceptable (<10%) uncertainty in the derived flow series, it is necessary for a Qualified Professional (QP) to make several site visits to measure a range of flows in order to calibrate a stage-discharge (rating) curve.  With climate change, the need to measure drought conditions and respond appropriately is crucial for habitat health and to prevent fish stranding.  The current study employs a Water Quality Mixing Model (WQMM) to estimate flows at a downstream site from an existing hydropower plant using a modified constant rate mixing model.  This is an independent estimate of flow entirely distinct from the stage-discharge curve.  The method can be employed anywhere there is a sufficient mixing length and sufficiently distinct WQ traits.  The method can reduce both maintenance costs and flow uncertainty where Environmental Flow Release Monitoring is required.

How to cite: Sentlinger, G.: Water Quality Mixing Model (WQMM) for Environmental Flow Release Monitoring, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11759,, 2020

Display materials

Display file

Comments on the display material

AC: Author Comment | CC: Community Comment | Report abuse

Display material version 2 – uploaded on 30 Apr 2020
Further clarification of abbreviations.
  • CC1: Comment on EGU2020-11759, Jérôme Le Coz, 03 May 2020

    Gabe, great work, thanks for sharing.

    About your slide 11: How did you retrain the rating curve, manually or using a probabilistic approach? The quantification of the impact of temperature difference (DT) on dilution gauging errors is very interesting. I understand the plot shows the discharge residuals between the automated tracing system (TMM) and the rating curve (RC), correct? Then, it would be interesting to consider the RC uncertainty in the comparison.

    Also, is there a relation between DT and discharge Q? (eg DT increases when Q decreases due to lower mixing?) We can expect that RC uncertainty increases for lower Q, that's why I'm asking.

    Still on that slide 11, is it reasonable to assume that TMM errors due to DT are independent? I'd rather think that a given DT would produce a systematic error in the conductance-concentration rating, hence a systematic discharge error.

  • AC1: Comment on EGU2020-11759, Gabriel Sentlinger, 04 May 2020

    Hi Jerome,

    I believe I used a maximum likelihood fit, but it may have been a simple Euclidean fit.  I know what you're asking though, and they may not be significant difference.  What we're showing here is actually variance about the mean RC fit.  

    I did plot the DT vs Discharge and there is a negative correlation, as you suggest.  This is shown in Figure 17 and Figure 18.  This is only showing data from Jun 17 to Sep 4, 2017.  I've confirmed that a similar relationship exists in other periods.

    I think I know what you mean about DT vs Q error.  If the US Temp error is always cooler, and the DS always hotter than the true temperature, for example, then a bias would be introduced.  However, there are independent factors that affect the error, time of day, heat flux, random instrument error, Q conditions (ie pooling, subsurface flow, backeddys).  I would like to investigate further because the technique shows great promise.


Display material version 1 – uploaded on 29 Apr 2020, no comments