Retrieving atmospheric temperature profiles by microwave radiometry using a priori information on atmospheric spatial-temporal evolution

• Published in: IEEE transactions on geoscience and remote sensing Vol. 39; no. 9; pp. 1896 - 1905
• Main Authors: Basili, P., Bonafoni, S., Ciotti, P., Marzano, F.S., d'Auria, G., Pierdicca, N.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.09.2001; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Applied geophysics; Atmospheric measurements; Atmospheric modeling; Atmospheric temperature; Atmospherics; Brightness temperature; Covariance matrix; Data assimilation; Earth sciences; Earth, ocean, space; Evolution; Exact sciences and technology; Geostatistics; Internal geophysics; Kalman filters; Kriging; Mathematical models; Meteorology; Microwave radiometry; Microwave theory and techniques; Principal components analysis; Satellite broadcasting; Satellites; Studies; atmosphere; geostatistics; microwaves; kriging; Space remote sensing; accuracy; covariance; principal components analysis; assimilation; brightness temperature; new methods; radiometry; Kalman filters; filtering
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/36.951080

A new approach is presented to determine atmospheric temperature profiles by combining measurements coming from different sources and taking into account evolution models derived by conventional meteorological observations. Using a historical database of atmospheric parameters and related microwave brightness temperatures, the authors have developed a data assimilation procedure based on the geostatistical Kriging method and the Kalman filtering suitable for processing satellite radiometric measurements available at each satellite pass, data of a ground-based radiometer, and temperature profiles from radiosondes released at specific times and locations. The Kalman filter technique and the geostatistical Kriging method as well as the principal component analysis have proved very powerful in exploiting climatological a priori information to build spatial and temporal evolution models of the atmospheric temperature field. The use of both historical radiosoundings (RAOBs) and a radiative transfer code allowed the estimation of the statistical parameters that appears in the models themselves (covariance and cross-covariance matrices, observation matrix, etc.). The authors have developed an algorithm, based on a Kalman filter supplemented with a Kriging geostatistical interpolator, that shows a significant improvement of accuracy in vertical profile estimations with respect to the results of a standard Kalman filter when applied to real satellite radiometric data.