On the Band-Averaged Radiative Transfer Calculation in a Mixture of Absorptive Gas and Scattering Medium

In remote sensing of cloud and aerosol, it is necessary to simulate the band-averaged radiance received by a remote sensing instrument. Even in a narrow spectral band, the atmospheric gas absorption coefficients vary substantially at different wavelengths while the optical properties of cloud and aerosol vary little.

The line-by-line radiative transfer calculation gives accurate band-averaged radiance but is too costly for operational retrieval. Instead, we can first compute band-averaged gas absorption optical thickness and band-averaged cloud and aerosol optical properties. Subsequently, the band-averaged optical thickness of absorptive gases and cloud or aerosol are added up, and we calculate the total single-scattering albedo accordingly. The total band-averaged optical thickness and single-scattering albedo are inputs in radiative transfer calculations, which is assumed to be monochromatic. The monochromatic approximation introduces biases. When the optical thickness of a cloud or an aerosol layer is much larger than gas absorption optical thickness, the biases are not significant. However, in some cases, the absorptive gases and cloud or aerosol have comparable optical thicknesses such as in oxygen A-band, where the biases may be unacceptable for remote sensing applications.

In this study, we propose a method to mix the band-averaged gas absorption and cloud or aerosol optical thicknesses in radiative transfer calculations. In monochromatic radiative transfer calculations, the path optical thickness is obtained by multiplying the optical thickness in zenith direction by the secant of zenith angle, which is invalid for band-averaged optical thickness. In the proposed method, the path band-averaged gas absorption optical thickness is directly computed by a fast gas absorption parameterization algorithm, which decreases the biases due to monochromatic approximation in band-averaged radiative transfer calculations. In this presentation, we will take oxygen A-band as an example to illustrate the application of the proposed method.