Evaluation of NMC Upper-Stratospheric Temperature Analyses Using Rocketsonde and Lidar Data

• Published in: Bulletin of the American Meteorological Society 1993-05, Vol.74 (5), p.789-799
• Main Authors: Finger, F. G., Gelman, M. E., Wild, J. D., Chanin, M. L., Hauchecorne, A., Miller, A. J.
• Format: Article
• Language: English
• Subjects: Artificial satellites; Atmosphere; Atmospheric and Oceanic Physics; Atmospheric temperature; Datasets; Earth, ocean, space; Exact sciences and technology; External geophysics; General properties of the high atmosphere; Ground truth; Lidar; Meteorology; Physics; Physics of the high neutral atmosphere; Radar; Standard error; Tempera; Temperature; Temperature gradients; Temperature measurement
• ISSN: 0003-0007; 1520-0477
• DOI: 10.1175/1520-0477(1993)074<0789:EONUST>2.0.CO;2

Daily NMC analyses, constructed from operational TOVS data since 1978, are used to monitor behavior of middle atmospheric temperature. Capability of the upper-stratospheric analyses (5, 2, 1, and 0.4 mb) to provide temporally consistent temperature fields depends on adjustments derived from ground-truth observations. These adjustments compensate for biases in the analyses caused by behavioral differences in data derived from successive operational satellite instruments and by changes in data and analysis procedures. This paper supports previous studies showing that observations from the datasonde rocket system provide ground-truth adjustments with a precision of 1°–3°C. The number of datasonde observations has diminished substantially in recent years, putting this adjustment system at risk. Falling-sphere rocket temperature data are shown to have variability in excess of that judged to be acceptable for use in the adjustment system. The capability for Rayleigh lidar to provide high-quality temperature data needed for ground truth is examined by comparing NMC analysis temperatures, adjusted by datasonde measurements, with observational values from regularly operating lidar systems in France since 1978. Agreement between the two databases is found to be good in recent years. This is further verified by comparisons between the datasonde-computed adjustments and independent analysis adjustments derived from the lidar database. It is concluded that high-quality lidar measurements, if made available from low, medium, and high latitudes, could provide the essential data for use in the analysis adjustment system.