New possibilities for atmospheric analysis with Monte Carlo

In recent years, significant advancements have been made in scientific computing utilizing the Monte Carlo method, particularly at the intersection of two key communities: the heat transfer community and the computer graphics community. The first breakthrough is that the computation time required to solve radiative transfer becomes invariant, whether the ground is described by a single building or thousands of them, whether the 3D clouds are poorly resolved or refined. We recently published in the journal PNAS a climate-analysis-oriented application showing this computation time invariance, extending to the spectral, spatial, and temporal domains. The second breakthrough of the paper lies in the nonlinear coupling of radiative transfer and gas spectroscopy, allowing estimation of unbiased line-by-line radiative quantities, even when considering scattering. This opens up new analysis prospects, initially applied to Earth and now expanding to Titan.