Development and Validation of a Time-Lag Correction for Vaisala Radiosonde Humidity Measurements

• Published in: Journal of atmospheric and oceanic technology Vol. 21; no. 9; pp. 1305 - 1327
• Main Authors: Miloshevich, L M, Paukkunen, A, Voemel, H, Oltmans, S J
• Format: Journal Article
• Language: English
• Published: 01.09.2004
• ISSN: 0739-0572; 1520-0426
• DOI: 10.1175/1520-0426(2004)021(1305:DAVOAT)2.0.CO;2

This study presents a method of improving the accuracy of relative humidity (RH) measurements from Vaisala RS80 and RS90 radiosondes by applying sensor-based corrections for well-understood sources of measurement error. Laboratory measurements of the sensor time constant as a function of temperature are used to develop a correction for a time-lag error that results from slow sensor response at low temperatures. The time-lag correction is a numerical inversion algorithm that calculates the ambient ('true') humidity profile from the measured humidity and temperature profiles, based on the sensor time constant. Existing corrections for two sources of dry bias error in RS80 humidity measurements are also included in the correction procedure: inaccuracy in the sensor calibration at low temperatures, and chemical contamination of sensors manufactured before June 2000 by nonwater molecules from the radiosonde packaging material. The correction procedure was evaluated by comparing corrected RS80-H measurements with simultaneous measurements from the reference-quality NOAA/Climate Modeling and Diagnostics Laboratory balloon-borne cryogenic hygrometer. The time-lag correction is shown to recover vertical structure in the humidity profile that had been 'smoothed' by the slow sensor response, especially in the upper troposphere and lower stratosphere, revealing a much sharper troposphere-stratosphere transition than is apparent in the original measurements. The corrections reduced the mean dry bias in the radiosonde measurements relative to the hygrometer from 4% RH at -20 degree C and 10% RH at -70 degree C to about plus or minus 2% RH at all temperatures, and the variability at low temperatures is substantially reduced. A shortcoming of the existing contamination correction is also uncovered, and a modification is suggested. The impact of the corrections on several radiosonde datasets is shown.