Application of the information approach to the analysis of two-year microwave observations of the atmosphere by the RPG-HATPRO radiometer at St. Petersburg University

• Published in: International journal of remote sensing 2016-07, Vol.37 (14), p.3346-3364
• Main Authors: Kostsov, V. S., Timofeyev, Yu. M., Zaitsev, N. A., Poberovsky, A. V., Osipov, S. I.
• Format: Article
• Language: English
• Subjects: Atmospherics; Channels; Humidity; Intervals; Mathematical analysis; Radiometers; Sampling; Seasons
• ISSN: 0143-1161; 1366-5901
• DOI: 10.1080/01431161.2016.1199060

Since June 2012 the ground-based microwave (MW) radiometer RPG-HATPRO (Radiometer Physics GmbH - Humidity and Temperature Profiler) G3 (generation 3) has been operating at the Faculty of Physics of St. Petersburg State University in constant mode with sampling interval of 1-2 s. The results of MW radiation brightness temperature measurements in 14 spectral channels of the RPG-HATPRO radiometer have been analysed for the time period December 2012 to November 2014. Within this time period, different data ensembles are compiled that have different duration and describe the variety of atmospheric situations at the observational site. Two different approaches are used for the assessment of the information characteristics of MW measurements. The first one is based on calculation of Shannon's information content that characterizes the number of atmospheric states distinguishable by measurements. We use the modification proposed by Kozlov - the 'information volume' approach. The second approach is based on calculation of the information quantity that characterizes the degree of correlation between measurements and atmospheric state parameter variations. The information quantities are calculated and analysed for the complete set of spectral channels and for two subsets that comprised seven so-called 'temperature' and seven so-called 'humidity' channels separately. Kozlov's 'information volume' approach is also applied to the problem of estimation of the optimal sampling interval of MW measurements. It is shown that for St. Petersburg University observation site that represents subarctic region with maritime climate: (1) the information volume of MW measurements for a cold and dry season is significantly less than for a warm and humid season due to lack of information from 'humidity channels'; (2) for stable atmospheric conditions the sampling interval of MW measurements should not be greater than 100-200 s; (3) the degree of correlation between MW measurements and atmospheric temperature and humidity does not considerably depend on season of observations.