Accuracy of ground-based infrared irradiance measurements in the NOAA/BSRN network

The accuracy of ground-based infrared radiation measurements has improved dramatically in the past decade. The decrease in uncertainty from ~ 30 W/m2 to ~ 3 W/m2 has been achieved through better calibrations in the laboratory and outdoors. The improvements have been so significant that it is probable that ground-based infrared measurements may soon be able to discern greenhouse gas warming from other causes of warming with unprecedented certainty (e.g., Philipona et al. JGR 2004). Participants in the Baseline Surface Radiation Network (BSRN) program have made the primary contribution to this improvement.

The purpose of this paper is to present the results from recent attempts of the NOAA/BSRN group to compare blackbody and outdoor calibration methods and data reduction formulations that tie NOAA/BSRN pyrgeometers to the Interim World Infrared Reference (IWIR), which is based on the World Infrared Standard Group (WISG) of four pyrgeometers at the World Radiation Center (WRC) in Davos, Switzerland. These four pyrgeometers are traceable to the Absolute Sky Scanning Radiometer (Philipona, Applied Optics, 2001).

Eighteen Eppley Laboratory, Inc. PIRs and one Kipp and Zonen CG-4 were included in this investigation. Three PIRs are modified with three additional dome thermistors attached; one at 45 degrees elevation in the north, and the other two separated by 120 degrees azimuth at 45 degrees elevation. These were calibrated by the WRC just before this comparison. Three unmodified PIRs, i.e., the dome thermistor was in the north near the base as installed by Eppley, were calibrated by the WRC just before the comparison. The other 13 PIRs were calibrated in the NOAA/Cox blackbody at various times over the last four years with the most recent calibrations just before these outdoor comparisons.

Instruments were mounted both unshaded and shaded, as well as unventilated and ventilated, although most were unshaded and unventilated to emphasize potential operational problems. The downwelling infrared irradiance varied between ~ 190 and ~ 320 W/m2 during the data collection period. The irradiances from these instruments are compared to examine the differences in calibration methods, the differences in dome temperature measurements, stability of calibrations over time, and the differences between equations used to calculate the irradiance. Estimates of uncertainty from this study compare favorably with earlier published estimates, but extend the sample of instruments, conditions, and methods used to obtain these results.