Reports
Contents
| Title: | Delta SW PAR Comparison 10-2016 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Date: | 2017-05-08 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Data File: | DeltaRadComp_PAR.csv DeltaRadComp_SWLW.csv |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Refers to: | BA,EE,MB,SW,TA,TW,WP,140456,140457,Q100421,140451,PQS110340,10162,140454,PQS110338,050597,PQS110341 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
We compared Delta radiometers to calculate new calibration coefficients for PARin and SWin. The reference sensors were deployed at our Twitchell Rice site as part of the “Golden Radiometer†setup. The SWin sensor (Huskeflux NR01 sn 2420) was deployed from 2016-09-22 to 2016-10-11. The PARin sensor (K&Z PQS1 sn 140456) was deployed from 2016-09-22 to 2016-10-06. The raw output of the SWin sensors is in mV (data were stored every 10 seconds and stored as 30-minute averages). The factory calibration coefficient [uV/(W/m2)] converts the mV reading to W/m2. Similarly, the raw output of the PAR_in sensors is in mV, the factory calibration coefficient [µV/(µmol/m2·s)] converts the mV reading to µmol/m2·s. PARin
Â
Figure 1. Linear regression between three PARin reference sensors. All three sensors were similar, so we used K&Z PQS1 (sn 140456) as the reference sensor for this analysis.
Figure 2. Linear regression between raw [mV] values of the Delta sensors and umol/m2/s values of the reference sensor. The slope of this regression line gives us the updated calibration coefficient [(µmol/(m2/s))/mV] for each Delta sensor. I forced the regression line through 0 (not shown here). Figure 3. Time series of uncorrected PARin data. The max difference between sensors is ~250 umol/m2/s. Figure 4. Time series of corrected PARin data. The max difference between sensors is smaller, ~100 umol/m2/s. SW data had the smallest difference after correction ~5 µmol/(m2·s), while TA had the largest ~250 µmol/(m2·s). The other sites had differences of ~120 µmol/(m2·s) or less.  SWin
Â
Figure 5. Linear regression between two SWin reference sensors. The two sensors were similar, so we used Hukseflux NR01 sn 2420 as the reference sensor for this analysis.
Figure 6. Linear regression between raw [mV] values of the Delta sensors and W/m2 values of the reference sensor. The slope of this regression line gives us the updated calibration coefficient [(W/m2)/mV] each Delta sensor. I forced the regression line through 0 (not shown here). TW has a different regression line because it is a K&Z CM11 pyranometer, and the other sites all have Huskeflux NR01 radiometers. Figure 7. Time series of uncorrected SWin data. The max difference between sensors is ~30 W/m2. Figure 8. Time series of corrected SWin data. The max difference between sensors is now smaller, ~15 W. SW had the largest difference after correction, ~6 W, while the other sites had differences of ~1 W or less. Â LWin We compared LWin sensors with the reference LWin sensor at our Twitchell Rice site. We determined it did not make sense to calibrate incoming longwave sensors with reference sensors if the sensors were at separate sites, since incoming longwave radiation is much more sensitive to sky conditions such as cloud cover, the amount of water vapor in the air, and air temperature. Figure 9. Time series of uncorrected LWin data. The TA data was messy. Â Temperature Figure 10. Time series of PRT. Â |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| |