Articles | Volume 5, issue 1
Atmos. Chem. Phys., 5, 153–167, 2005

Special issue: Probing the atmosphere in three dimensions for SCIAMACHY

Atmos. Chem. Phys., 5, 153–167, 2005

  24 Jan 2005

24 Jan 2005

Subtropical trace gas profiles determined by ground-based FTIR spectroscopy at Izaña (28° N, 16° W): Five-year record, error analysis, and comparison with 3-D CTMs

M. Schneider1, T. Blumenstock1, M. P. Chipperfield2, F. Hase1, W. Kouker1, T. Reddmann1, R. Ruhnke1, E. Cuevas3, and H. Fischer1 M. Schneider et al.
  • 1IMK-ASF, Forschungszentrum Karlsruhe, Karlsruhe, Germany
  • 2Institute for Atmospheric Science, University of Leeds, Leeds, UK
  • 3Observatorio Atmosférico de Izãna, Instituto Nacional de Meteorología, Spain

Abstract. Within the framework of the NDSC (Network for the Detection of Stratospheric Change) ground-based FTIR solar absorption spectra have been routinely recorded at Izaña Observatory (28° N, 16° W) on Tenerife Island since March 1999. By analyzing the shape of the absorption lines, and their different temperature sensitivities, the vertical distribution of the absorbers can be retrieved. Unique time series of subtropical profiles of O3, HCl, HF, N2O, and CH4 are presented. The effects of both dynamical and chemical annually varying trace gas cycles can be seen in the retrieved profiles. These include enhanced upwelling and photochemistry in summer and a more disturbed atmosphere in winter, which are typical of the subtropical stratosphere. A detailed error analysis has been performed for each profile. The output from two different three-dimensional (3-D) chemical transport models (CTMs), which are forced by ECMWF analyses, are compared to the measured profiles. Both models agree well with the measurements in tracking abrupt variations in the atmospheric structure, e.g. due to tropical streamers, in particular for the lower stratosphere. Simulated and measured profiles also reflect similar dynamical and chemical annual cycles. However, the differences between their mixing ratios clearly exceed the error bars estimated for the measured profiles. Possible reasons for this are discussed.

Final-revised paper