Extension of radiative transfer code MOMO, matrix-operator model to the thermal infrared. Clear air validation by comparison to RTTOV and application to CALIPSO-IIR
Résumé
1-D radiative transfer code MOMO (Matrix-Operator Model), has been extended from View the MathML source band to the whole View the MathML source spectrum. MOMO can now be used for computation of full range radiation budgets (shortwave and longwave). This extension to the longwave part of the electromagnetic radiation required to consider radiative transfer processes that are features of the thermal infrared: the spectroscopy of the water vapor self- and foreign- continua of absorption at View the MathML source and the emission of radiation by gases, aerosol, clouds and surface. MOMO's spectroscopy module, CGASA (Coefficient of Gas Absorption), has been developed for computation of gas extinction coefficients, considering continua and spectral line absorption. The spectral dependences of gas emission/absorption coefficients and of Planck's function are treated using a k-distribution. The emission of radiation is implemented in the adding-doubling process of the matrix operator method using Schwarzschild's approach in the radiative transfer equation (a pure absorbing/emitting medium, namely without scattering). Within the layer, the Planck-function is assumed having an exponential dependence on the optical-depth. In this paper, validation tests are presented for clear air case studies: Comparisons to the analytical solution of a monochromatic Schwarzschild's case without scattering show an error of less than 0.07% for a realistic atmosphere with an optical depth and a blackbody temperature that decrease linearly with altitude. Comparisons to radiative transfer code RTTOV are presented for simulations of top of atmosphere brightness temperature for channels of the space-borne instrument MODIS. Results show an agreement varying from 0.1 K to less than 1 K depending on the channel. Finally MOMO results are compared to CALIPSO Infrared Imager Radiometer (IIR) measurements for clear air cases. A good agreement was found between computed and observed radiance: biases are smaller than 0.5 K and RMSE (Root Mean Square Error) varies between 0.4 K and 0.6 K depending on the channel. The extension of the code allows the utilization of MOMO as forward model for remote sensing algorithms in the full range spectrum. Another application is full range radiation budget computations (heating rates or forcings).