A Refined Calibration Procedure of Two-Channel Sun Photometers to Measure Atmospheric Precipitable Water at Various Antarctic Sites

• Published in: Journal of Atmospheric and Oceanic Technology
• Main Authors: U. Bonafè, Vito Vitale, Angelo Lupi, Claudio Tomasi, Elena Benedetti, Luca Valenziano, Boyan Petkov
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2008
• Subjects: Rural Digital Europe; NEANIAS Atmospheric Research Community; Atmospheric Science; Ocean Engineering; Antarctic; The Antarctic; Antarc*
• Online Access: https://www.openaccessrepository.it/record/106817; https://doi.org/10.1175/2007jtecha952.1

Abstract Two-channel sun photometers can be easily employed at Antarctic sites, where harsh environmental conditions prevail, to carry out measurements of precipitable water W. In the very dry air conditions observed in the Antarctic atmosphere, water vapor does not produce strong absorption features along the sun path. Therefore, these instruments need to be calibrated using analytical forms different from the square root regime, which can be determined by simulating the output voltages measured at Antarctic sites, for the spectral near-IR curves of extraterrestrial solar irradiance, instrumental responsivity parameters, and atmospheric transmittance, relative to various measurement periods. For this purpose, average models of the Antarctic atmosphere from the ground level up to the 30-km altitude were considered for different solar zenith angles and relative humidity conditions. The ratios between the output voltages simulated in the band and window channels were plotted as a function of total water vapor content Cw, for each site and each period, to define the best-fit calibration curves, which were subsequently normalized to the field measurements to take into account the aging effects on the filter transmission characteristics. Each of the five calibration curves was found to present a slope coefficient decreasing gradually with Cw from values higher than 0.8 to about 0.6. Using these curves, measurements of W were obtained, which differ appreciably at both sea level and high-altitude sites from those given by the square root calibration curves, avoiding large overestimation errors of 10%–40% at the high-altitude sites and underestimation errors of 5%–15% at the sea level site.