Sure --

1. First, ignore Phi2 and all the other stuff there. Those things are either not ready for prime time, or they are values which don't matter (calibration values and stuff). Otherwise...

2. Phi2 Expected value range: 0 - .84. Phi2 is the efficiency of Photosystem 2 . Read it roughly as "X% of the incoming light energy is going to do photochemistry (ie do useful work for the plant". So .8 would 80% of the energy is going to do useful work.

3. NPQ Expected value range: 0 - 5 (though it can go as high as 40). NPQ is Non-Photochemical Quenching. This is actually a cutting edge estimate of NPQ which normally takes an hour to measure, we call it NPQt. But anyway, it says "how much of the incoming light is being dissipated as heat". It is a unitless value.

4. LEF Expected value range: 0 - 500ish. LEF is Linear Electron Flow. Read as "X number of micro moles of electrons per second are flowing out of Photosystem 2." LEF is calculated as *LEF = Phi2 * light intensity PAR x .42 *. The .42 is the amount of light which is hitting Photosystem 2, as compared to that which is reflected, diffused, or hits Photosystem 1. So if your LEF value looks funny - investigate Phi2 and PAR first!

5. PAR Expected value range: 0 - 2000ish (can be higher in certain types of sun/cloud conditions). PAR is Photosynthetically Active Radiation, the portion of light which is usable by plants to do photosynthesis.

6. Chlorophyll Content Expected value range: 0 - 75ish. 75 means the leaf is very green (contains lots of chlorophyll and nitrogen), 0 is not green (very little chlorophyll / nitrogen).

What's normal / what to expect

• When PAR goes up, Phi2 goes down
• When PAR goes up, NPQ goes up
• When PAR goes up, LEF goes up but levels off at some point (see attached image as example). Use this to find outliers or possibly bad data points which sit outside this range.
• Chlorophyll content changes slowly, so does not generally relate to PAR.
  
• The One protocol should look like the attached, with the typical plateau pattern. The size of the plateau will vary based on light intensity, and a dead leaf should look flat in that region.
• NaN just means the value was not calculable (ie something tried to divide by zero somewhere). Just exclude these.

Hope this helps clarify things!

Thanks, that's helpful. When I see a negative value for any of these parameters then it basically means that the datapoint needs to be discarded? Also, the number of datapoints is significantly less than I thought I was collecting. One of the days (June 16) I sampled nearly 100 leaves, yet I've only got 49 datapoints for that day. It was one of the days where I had to get your help downloading data. Is it possible the other datapoints were lost in the great beyond?

Just a few more questions-are the time stamp data accurate? I've got one sampling date where a couple of columns are headed by "get_userdef1, 2 and 5". Can these be ignored too?

sample 940, cal 940, sample 650, cal 650 what these ??? It's sample and calibration?

The sample and cal values are blank raw absorbance values. When you calculate SPAD you need to have a blank as a reference (just like when you measure absorbance). The calculation for SPAD is:

SPAD ~ log((sample 940 - blank 940) / (sample 650 - blank 650))*100

I put a ~ instead of a = because this produces a number which correlates very well to Minolta's SPAD, but is offset. So we then calibrate the devices to an actual Minolta to remove the offset

There are also a bunch of 'userdef' values, which are the calibration to the Minolta values (see attached). We've already calibrated this stuff, so I wouldn't worry about doing it your end.

Timestamps are in GMT which is why they look off.

Hmmm I hope they weren't lost... They should be on the device or on the web... there are no more cached data points on that phone? I'm not sure, we'd have to troubleshoot that.

Yes you can ignore userdef values, see post above for description of those.

In terms of negative values, the only way to really tell what's going on is to look at the raw trace which you can do by clicking on the data point in the spreadsheet view or on a graph. Some data points are errors, and some are simply dead (ie no photosynthesis). I would argue you shouldn't exclude dead leaves (flat line) whereas you should exclude clear user error (moved the leaf or device).

Here are some good examples for The One measurement:

a good measurement: http://www.photosynq.org/projects/msu-oak-tree-study-sum15/explore_single_measurement?datum_id=41220

user error (moved the device or something): http://www.photosynq.org/projects/msu-oak-tree-study-sum15/explore_single_measurement?datum_id=86440

more user error: http://www.photosynq.org/projects/msu-oak-tree-study-sum15/explore_single_measurement?datum_id=40312

a dead leaf (flat line): http://photosynq.org/projects/99/explore_single_measurement?datum_id=90944

We're working on some ways to flag this data to make it easier to identify and to help you all develop a quicker intuition on what is good and what is bad.