Articles | Volume 14, issue 20
Atmos. Chem. Phys., 14, 11129–11148, 2014
Atmos. Chem. Phys., 14, 11129–11148, 2014

Research article 23 Oct 2014

Research article | 23 Oct 2014

Climatology of free-tropospheric humidity: extension into the SEVIRI era, evaluation and exemplary analysis

M. Schröder1, R. Roca2, L. Picon3, A. Kniffka1, and H. Brogniez4 M. Schröder et al.
  • 1Satellite-Based Climate Monitoring, Deutscher Wetterdienst, Offenbach, Germany
  • 2OMP/LEGOS/CNRS, Toulouse, France
  • 3Sorbonne Universités, UPMC Univ. Paris 06; CNRS/INSU; ENS; Ecole Polytechnique; LMD-IPSL, UMR8539, Paris, France
  • 4Université Versailles St-Quentin; Sorbonne Universités, UPMC Univ. Paris 06; CNRS/INSU, LATMOS-IPSL, Guyancourt, France

Abstract. A new free-tropospheric humidity (FTH) data record is presented. It is based on observations from the Meteosat Visible and Infrared Imager (MVIRI) onboard Meteosat-2–Meteosat-5, as well as Meteosat-7, and the Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard Meteosat-8 and Meteosat-9 at the water absorption band near 6.3 μm. The data set is available under clear-sky and low-level cloud conditions. With the extension to SEVIRI observations, the data record covers the period 1983–2009 with a spatial resolution of 0.625° × 0.625° and a temporal resolution of 3 h.

The FTH is the mean relative humidity (RH) in a broad layer in the free troposphere. The relation between the observed brightness temperature (BT) and the FTH is well established. Previous retrievals are refined by taking into account the relative humidity Jacobians in the training process of the statistical retrieval. The temporal coverage is extended into the SEVIRI period, the homogenization of the BT record is improved, and the full archive is reprocessed using updated regression coefficients.

The FTH estimated from the Meteosat observations is compared to the FTH computed from the RH profiles of the Analyzed RadioSoundings Archive (ARSA). An average relative bias of −3.2% and a relative root-mean-square difference (RMSD) of 16.8% are observed. This relative RMSD agrees with the outcome of an analysis of the total uncertainty of the FTH product. The decadal stability of the FTH data record is 0.5 ± 0.45% per decade.

As exemplary applications, the interannual standard deviation, the differences on decadal scales, and the linear trend in the FTH data record and in the frequency of occurrence of FTH < 10% (FTHp10) are analyzed per season. Interannual standard deviation maxima and maxima in absolute decadal differences are featured in gradient areas between dry and wet regions, as well as in areas where FTH reaches minima and FTHp10 reaches maxima. An analysis of the FTH linear trends and of the associated uncertainty estimates is achieved to identify possible problems with the data record. Positive trends in FTHp10 are featured in gradient areas between wet and dry regions, in regions where the FTH is minimum, in regions where FTHp10 is maximum, and in regions where differences between FTHp10 averaged over the 2000s and 1990s are negative. However, these positive trends in FTHp10 are associated with maximum standard deviation and are thus hardly significant. This analysis and intercomparisons with other humidity data records are part of the Global Energy and Water Cycle Experiment (GEWEX) Water Vapor Assessment (G-VAP).

Final-revised paper