Fast radiative transfer in multiple scattering atmospheres at far and mid infrared wavelengths

Recognizing the value of Far Infrared Region (FIR) observations, in September 2019, the European Space Agency (ESA) selected FORUM (Far-infrared Outgoing Radiation Understanding and Monitoring) as the 9th Earth Explorer (EE9) (Palchetti et al., 2020) whose launch is foreseen in 2027. FORUM, dedicated to mapping Earth’s far-infrared emission globally, will produce an enormous quantity of new data, requiring the implementation of fast radiative transfer models applicable to the entire IR spectral region for an effective data exploitation and analysis. Full physics models (i.e. DISORT, Stamnes et al., 1988) rely on robust and accurate numerical methodologies to solve the radiance field in presence of multiple scattering events for specific scenarios. The complexity of the multiple scattering effects makes this class of models extremely time consuming and inappropriate for large dataset analysis. To save computational time, fast radiative transfer models adopt multiple strategies which might account for approximation of the physical problem, simplified numerical solutions, code parallelization, and the extensive use of parametrizations. In the first part of this study, we investigate the level of accuracy of the Chou’s approximation (Chou et al., 1999), a fast methodology, widely used in operative frameworks for its simplicity and easy implementation. The performance of this approximate solutions is evaluated with respect to a full-physics approach over a widespread collection of atmospheric scenarios using the goal NESR of FORUM as reference metric. The results show not negligible inaccuracies when the Far InfraRed (FIR) is considered (Martinazzo et al., 2021). In the second part the study, to reduce the bias of the Chou scaling method, a correction term is modelled and computed using the solution recently proposed by Tang (Tang et al. 2018). The Tang methodology, originally created to refine the Chou flux computations, is here adapted to simulations of radiance fields over the FIR spectral range, exploiting appropriate multiplicative coefficients. The range of validity of the new methodology is then evaluated, as already done for the Chou scheme, by comparing this fast solution against full physics simulations. The comparisons show an overall reduction of the radiance residuals overs most of the cloudy cases encountered in nature. In particular, the use of the Tang methodology with the new coefficients is accurate for the computation of radiance fields in presence of thin cirrus clouds which are one of the targets of the FORUM mission.