Traceability of solar UV measurements using the QASUME reference spectroradiometer

摘要

One major objective of the European Joint Research Project "Traceability for surface spectral solar ultraviolet (UV) radiation" was to reduce the uncertainty of spectral UV measurements. The measurement instrument used for this work was the portable UV European reference spectroradiometer QASUME. The calibration uncertainty of this instrument was decreased and validated by a comparison of direct calibrations against a primary standard for spectral irradiance, a high temperature blackbody radiator, and against a reference detector using a spectrally tunable laser as a monochromatic source. The spectral irradiance responsivity of the reference detector is traceable to the primary standard of optical power, realized through a cryogenic radiometer, and to the SI unit of meter. The measuring technique was improved by the construction of a new reference spectroradiometer, QASUMEII. An improved input optics removes the dependences of the measured solar irradiance on the angle of incident for solar zenith angle smaller than 75 deg. Moreover, a hybrid photon detection system enables continuous tracking of the instrument's responsivity changes. For both spectroradiometer systems an uncertainty budget was calculated. The improvements have reduced the measurement uncertainties of solar spectral UV irradiance measurements from 4.8% in 2005 to 2.0% (k = 2) in the spectral region above 310 nm. The largest sources of uncertainty were the absolute spectral irradiance responsivity calibration, the angular response uncertainty, and the instrument stability using the hybrid detector, which were reduced from 3.6% to 1.1%, from 1.2% to 0.6%, and from 0.65% to 0.4%, with respect to the situation prior to the project. The new instrument was validated during a four month intercomparison relative to the QASUME reference. The mean ratio of the solar irradiance scans between the two reference spectroradiometers has an offset of +0.7% and a standard deviation of +/- 1.5% for a wavelength greater than 305 nm, which is well within the combined uncertainty of 3.7% calculated from the uncertainties of the two systems. (C) 2016 Optical Society of America