Journal of Quantitative Spectroscopy and Radiative Transfer
Modelling of atmospheric mid-infrared radiative transfer: the AMIL2DA algorithm intercomparison experiment
Introduction
The Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) [1], [2], [3] on the Envisat Earth observation satellite developed by the European Space Agency (ESA) is a mid-infrared limb emission sounder which provides vertical profiles of atmospheric species relevant to several inter-linked problems in ozone chemistry and global change. Routine data analysis under ESA responsibility covers only six the so-called key species (H2O, O3, N2O, CH4, HNO3, NO2) as well as pressure and temperature. The MIPAS data, however, contain much more information on further atmospheric trace species of very high scientific value. Thus, exploitation of MIPAS data with respect to (a) the retrieval of many more species relevant to atmospheric research such as the complete nitrogen family, chlorine source gases and reservoirs, greenhouse gases, ozone precursors, aerosols, etc., and (b) the retrieval of key species abundance profiles by methods of higher sophistication than affordable under near real time processing constraints are of high interest for better understanding of ozone chemistry and global change. In order to support the development and characterization of methods and tools for this extended analysis of MIPAS data, the project AMIL2DA (Advanced MIPAS Level 2 Data Analysis) has been initiated.
When atmospheric parameters are retrieved from atmospheric radiance spectra, accurate forward modelling of radiative transfer through the Earth's atmosphere is of particular importance, since forward modelling errors directly map on to the retrieved quantities. Intercomparison studies are a standard approach to assess the reliability of radiative transfer models [4], [5], [6]. Hence, a major component of the AMIL2DA project is the comparison of the radiative transfer (forward) models used by each group for their retrievals.
Section snippets
Theory
Disregarding scattering, which is justified in many applications of infrared remote sensing, the monochromatic radiative transfer equation [7] can be written aswhere Lν is spectral radiance at wavenumber ν at the location of the observer lobs,L0 is background radiance (assumed zero in this intercomparison experiment), l is the path coordinate, τν(l1,l2) is the spectral transmittance between two locations l1 and l2, and J is the source function,
The models
Five groups participated in this intercomparison experiment, each with their own line-by-line radiative transfer model: The Karlsruhe Optimized and Precise Radiative transfer code (KOPRA) [13], [14] of the Forschungszentrum Karlsruhe, the Reference Forward Model (RFM) of Oxford University (http://www.atm.ox.ac.uk/RFM/), a direct integration radiative transfer code by DLR (MIRART, Modular Infrared Atmospheric Radiative Transfer [15]), the Optimized Forward Model (OFM) by IFAC (formerly IROE) [16]
The intercomparison experiment
In order to detect forward model deficiencies, a cross comparison exercise was carried out to mutually validate radiative transfer codes. The idea of the setup of the intercomparison exercise was to start from simple settings proving the basic functionalities of the radiative transfer codes, and then to proceed to more complex and realistic scenarios. In order to avoid any masking of differences in the basic functionalities of the codes by additional sophistication introduced by realistic
Conclusion
The intercomparison experiment was quite enlightening for all participants. The overall interconsistency of spectra is good, except for test cases where functionalities were required which were beyond the specification of some of the codes, e.g. line coupling, non-local thermodynamic equilibrium, etc. Some differences in details are caused by different implementations of the Voigt line shape and related approximations. Different approaches to calculate the line intensity as a function of
Acknowledgements
AMIL2DA is a Shared Cost Action within the RTD generic activities of the 5th FP EESD Programme of the European Commission, Project EVG1-CT-1999-00015. M.L.-P. has been partially supported by PNE under contract PNE-017/2000-C. P. Varanasi gave access to spectroscopic cross-section data prior to their publication in HITRAN. Furthermore thanks go to J.-M. Flaud and G. Schwarz, who provided helpful comments to the manuscript.
References (53)
- et al.
Sensitivity of trace gas abundances retrievals from infrared limb emission spectra to simplifying approximations in radiative transfer modelling
JQSRT
(2002) A new implementation of the Humlicek algorithm for the calculation of the Voigt profile function
JQSRT
(1997)Optimized computation of the Voigt and complex probability functions
JQSRT
(1982)- et al.
CO2 line mixing in MIPAS limb emission spectra and its influence on retrieval of atmospheric parameters
JQSRT
(1998) - et al.
Line shape and the water vapor continuum
Atmos Res
(1989) - et al.
Efficient line-by-line calculation of absorption coefficients
JQSRT
(1999) - et al.
The HITRAN molecular spectroscopic database and HAWKS (HITRAN atmospheric workstation)1996 edition
JQSRT
(1998) - et al.
Fast computation of monochromatic infrared atmospheric transmittances using compressed look-up tables
JQSRT
(1998) - et al.
Fast monochromatic radiative transfer calculations for limb sounding
JQSRT
(2002) Rapid approximation to the Voigt/Faddeeva function and its derivatives
JQSRT
(1999)