Simultaneous estimation of radiance and its sensitivities to radiative properties in a spherical-heterogeneous atmospheric radiative transfer model by Monte Carlo method: Application to Titan

Abstract

The study of planetary atmospheres and surfaces depends on integrating radiative transfer modelling with inversion techniques. Traditional radiative transfer models, which assume simplified geometries (e.g., parallel-plan approximation), are limited in accurately capturing the heterogeneous and spherical nature of planetary atmospheres and surfaces, especially in spectral ranges where the scattering effects are important. Numerous modelling strategies are available with full accuracy in the representation of 3D features, but coupling such realistic radiative transfer model with inversion techniques raises challenges in calculation time and gradient assessment. Thanks to the recent active research at the interface between statistical physics and computer graphics, we here tackle these challenges based on the propositions of null-collision (Galtier et al., 2013) and spatial partitioning of extinction-coefficients overestimates (Villefranque et al., 2019). Straightforwardly, these propositions are implemented here for planetary atmospheres and we also provide a slightly modified version of the algorithm that simultaneously estimates the radiance and its gradient with respect to the radiative properties (e.g., absorption coefficients, scattering coefficients and surface reflectivities). Gradient estimation preserves the insensitivity of calculation time to the geometric complexity. The method is applied to Titan. Its performance and limitations are discussed together with the perspectives of being integrated into the inversion routine of planetary data.