Atmospheric longwave irradiance uncertainty: Pyrgeometers compared to an absolute sky-scanning radiometer, atmospheric emitted radiance interferometer, and radiative transfer model calculations

Because atmospheric longwave radiation is one of the most fundamental eleme nts of an expected climate change, there has been a strong interest in impr oving measurements and model calculations in recent years. Important questi ons are how reliable and consistent are atmospheric longwave radiation meas urements and calculations and what are the uncertainties? The First Interna tional Pyrgeometer and Absolute Sky-scanning Radiometer Comparison, which w as held at the Atmospheric Radiation Measurement program's Southern Great P lains site in Oklahoma, answers these questions at least for midlatitude su mmer conditions and reflects the state of the art for atmospheric longwave radiation measurements and calculations. The 15 participating pyrgeometers were all calibration-traced standard instruments chosen from a broad intern ational community. Two new chopped pyrgeometers also took part in the compa rison. An absolute sky-scanning radiometer (ASR), which includes a pyroelec tric detector and a reference blackbody source, was used for the first time as a reference standard instrument to field calibrate pyrgeometers during clear-sky nighttime measurements. Owner-provided and uniformly determined b lackbody calibration factors were compared. Remarkable improvements and hig her pyrgeometer precision were achieved with field calibration factors. Res ults of nighttime and daytime pyrgeometer precision and absolute uncertaint y are presented for eight consecutive days of measurements, during which pe riod downward longwave irradiance varied between 260 and 420 W m(-2). Compa risons between pyrgeometers and the absolute ASR, the atmospheric emitted r adiance interferometer, and radiative transfer models LBLRTM and MODTRAN sh ow a surprisingly good agreement of <2 W m(-2) for nighttime atmospheric lo ngwave irradiance measurements and calculations.