Reports

This is a second set of Rnet calibrations done at Oxford Tract, after we repaired some sensors. The sensors mounted on a tripod and were connected to a CR6 and AM16/32B. We borrowed a CNR4 4-band radiometer from Ameriflux to serve as a reference sensor. The Rnet and CNR4 sensors were installed facing south, over bare soil, and leveled. The sensors and datalogger were powered with a rechargeable 12V lithium battery. Sensors were scanned every 10 seconds and the average recorded every 1 minute. Back in the lab, I calculated 10-min averages of Rnet.

I was re-testing several sensors, sn 990352, sn 990353, sn 990354, and sn 041556, that I had originally calibrated in the summer. For all sensors, there was about a 5% difference between the calibration factors. I don't think there's a clear reason to choose one or the other because they have similar, high R2 values for the linear regressions. In the summer, we have a wider range of Rnet values. However, in the fall, I kept the sensors cleaner, so in the equipment history I used the fall calibration factors.

 

Figure 1. Time series of all radiometers. We ran out of battery power from 2019-09-20 through part of 2019-09-23. Sensors are pretty well matched, except the reference sensor has a small bump (from shading?) on the morning of 2019-09-29.

Figure 2. Regression of test sensors against the reference radiometer.

Figure 3. An interesting deviation in CNR4 data that I don't think affects the calibration.  Bump in early-morning LWin that was also reflected in the Rnet sensors.

Figure 4. Another interesting deviation in CNR4 data that I don't think affects the calibration. Rain or moisture on CNR4 LWin sensor on the morning of 2019-09-26 that was not seen in the Rnet sensors.