Reports

Contents

Title: Bouldin Corn - Radiometer Calibration 2021
Date:2021-06-09 - 2021-06-09
Data File: BC_calkit_20210609.csv
BC_calNDVI_20210609.csv
Refers to:BC,121038,060831,1049,1050,990181,140453,140441,140456,1051,1052,882103020,2046503357

We installed a radiometer calibration kit at Bouldin Corn from 2021-05-27 to 2021-06-09. The calibration kit included a 4-way radiometer, a PAR sensor, and a pair of Apogee NDVI sensors that we removed from BA. See the table below for sensor serial numbers and coefficients.

During the calibration period, the corn grew more than half a meter from 30 cm to 87 cm.

Radiometer cal kit on radiometer boom at Bouldin Corn.

 

Reference Variable Reference Sensor Serial Number Factory coefficient
SWin Kipp & Zonen CNR4 sn 121038 14.56 uV/(W/m2)
SWout Kipp & Zonen CNR4 sn 121038 14.17 uV/(W/m2)
LWin Kipp & Zonen CNR4 sn 121038 6.83 uV/(W/m2)
LWout Kipp & Zonen CNR4 sn 121038 5.48 uV/(W/m2)
PAR Kipp & Zonen PQS 1 sn 060831 6.69?
NDVI Incoming Apogee S2-411 sn 1049 n/a
NDVI Outgoing Apogee S2-411 sn 1050 n/a

 

Tower Variable Tower Sensor Serial Number Factory coefficient New coefficient Summary
SWin Kipp & Zonen NR01 sn 990181 9.94 uV/(W/m2)

10.03 (+0.9%)

R2=99%

  
SWout Kipp & Zonen NR01 sn 990181 9.96 uV/(W/m2)

9.91 (-0.5%)

R2=99%

  
LWin Kipp & Zonen NR01 sn 990181 9.66 uV/(W/m2)

8.82 (-8.9%)

R2=97%

  
LWout Kipp & Zonen NR01 sn 990181 9.58 uV/(W/m2)

10.78 (+12.5%)

R2=99%

  
PAR Incoming Kipp & Zonen PQS 1 sn 140453 8.12 uV/(umol/m2/s)

7.97? (-1.8%)

R2=99%

  
PAR Outgoing Kipp & Zonen PQS 1 sn 140441 8.36 uV/(umol/m2/s)

7.73? (-7.5%)

R2=99%

  
PAR Diffuse Kipp & Zonen PQS 1 sn 140456 8.26 uV/(umol/m2/s)

7.75? (-6.2%)

R2=99%

  

 

Tower Variable Tower Sensor Serial Number Calibration Equation Summary
NDVI Incoming Apogee S2-411 1051    
NDVI Outgoing Apogee S2-411 1052    
NDVI Incoming Decagon-METER SRS 882103020 n/a Do not use the SRS-based NDVI here
NDVI Outgoing Decagon-METER SRS 2046503357 n/a Do not use the SRS-based NDVI here

Shortwave Incoming

Figure 1. Incoming shortwave radiation.

Regression Data

Residuals

Figure 2. Linear regression between tower and reference sensors. Fit is very tight with R2=99.9%.

 

Shortwave Outgoing

Figure 3. Outgoing shortwave radiation. Tower's SWout has higher peaks than the reference sensor.

Regression Data

Residuals

Figure 4. Linear regression between tower and reference sensors. Fit is very tight with R2=99.6%.

 

Longwave Incoming

Figure 5. Incoming longwave radiation. It's not as synchronized as I expected given that the 2 radiometers were less than 1m away from each other.

Regression Data

Residuals

Figure 6. Linear regression between tower and reference sensors. Fit is ok, with more scatter on the lower end (nighttime). R2=97%.

 

Longwave Outgoing

Figure 7. Outgoing longwave radiation. Tower LWout had consistently higher peaks.

Regression Data

Residuals

Figure 8. Linear regression between tower and reference sensors. Fit is good with R2=99.4%.

 

Incoming PAR

Figure 9. Incoming PAR from the reference sensor and 2 tower sensors (PARin sensor and rotating shadowband (RSB) sensor used to measure diffuse PAR). The RSB sensor has slightly lower daily peaks.

Regression Data

Residuals

Figure 10. Linear regression between tower and reference sensors. Measurements <1mV were discarded to avoid too much weight on nighttime values. Fit is tight with both R2>99%.

 

Outgoing PAR

Figure 11. Outgoing PAR. The reference sensor has higher daily peaks.

Regression Data

Residuals

Figure 12. Linear regression between tower and reference sensors. Measurements <0.2mV were discarded to avoid too much weight on nighttime values. Fit is tight with R2=99.

 

NDVI

Figure 13. NDVI from a pair of reference sensors (Apogee) and 2 pairs of tower sensors (one SRS pair, one Apogee pair). The SRS NDVI is obviously degraded. The Apogee NDVI values are very similar, although one may be drifting slightly. At the beginning, the 2 Apogee sensors had a 0.2 difference, but by the end the 2 Apogee sensors had a 0.3 difference. Super small difference, though.

Regression Data

Residuals

Figure 14. Linear regression between tower and reference sensors. For a better correction for the SRS NDVI sensors, see the RT report devoted to NDVI. https://nature.berkeley.edu/biometlab/bmetdata/rtreport.php?screen=view&id=51

Figure 15. Red reflectance. The SRS sensor is obviously bad, and there is a very small difference between the 2 pairs of Apogee sensors. 

Figure 16. NIR reflectance. The 2 pairs of Apogee sensors match each other almost exactly.