Strong relationship ER and K600.

[demelo-limno]

Thanks for submitting your issue or question. We can address issues fastest if they are:

• New - search past issues first to see if your question has already been answered. Reopen an old issue if your question was almost but not quite answered.

• Complete - share all relevant code and console output/errors/warnings, in the order they were run and produced on your computer. Also include the output from a call to devtools::session_info() to tell us about your computer's configuration.

• Reproducible - include all the data and code necessary for us/others to recreate the problem locally. It's fine to make up data if you can't share yours, as long as the problem still comes through.

• Minimal - what is the smallest amount of data and code you can use to demonstrate the problem? This is less essential than the others but improves communication and our response time.

Delete these instructions before submitting your issue.

---

Brief problem description

What you saw on your computer

Include all code you ran (a minimal example) and all console output, errors, and warnings. Include a data file if needed.

Session information

Run the following code line (or sessionInfo() if that doesn't work) and paste in your output.

> devtools::session_info()

[demelo-limno]

Hello all!
I'm trying to model river metabolism using a long-term miniDOT dataset (between June and October). We installed our loggers in sections close to river mouth (flat reaches and low elevation). We also have hydrometric stations at the same location, which give us coupled estimates of discharge and oxygen+temp.
Since we have good discharge data, I chose to run Bayesian partial-pooling models ("binned", day_start=3, day_end=27, default options).
I found good positive relations between K600 and discharge for most of my sites. However, because I didn't have a good pre-knowledge about K600 in these rivers, I also estimated K600 using equations 1 and 7 from Raymond (2012), and I found that K600s generated by streamMetabolizer are pretty similar to those estimated by Raymond's equations. So, I believe they are fine! Shouldn't be the problem...

The thing is, I'm getting strong relations between ER and K600 generated by these models (R2=0.6-0.9) and I saw in the literature that this relationship should be low!

Any thoughts why I'm getting these results? What else could I test (I also tested MLE and bayesian no-pooling)?
I'm a beginner in this field and I appreciate any help!

I'm sharing data from one of my rivers, which you can see this clear pattern I've described.

Thanks,
Michaela

Example_discharge_data.csv
Example_raw_data.csv