Monte Carlo simulation of atmospheric radiative forcings using a path-integral formulation approach for spectro-radiative sensitivities

We present recent advances in path-integral formulations designed for unbiased Monte Carlo sensitivity estimation (in the form of partial derivatives) within a coupled physics model. We establish the theoretical foundation and illustrate the approach by estimating instantaneous atmospheric radiative forcings. In climate studies, these quantities amount for the change in top-of-atmosphere (TOA) net radiative flux induced by an isolated change in surface or atmospheric constitution. Based on a path-integral framework, our approach results in estimations consistent with well-established radiative forcings in the climate community. We highlight how physics coupling through path-integral formulations yields unbiased sensitivity estimation of a radiative quantity (integrated TOA flux) to a spectroscopic parameter (fraction change in gas concentration). Furthermore, we emphasize the method's scalability, demonstrating its compatibility with computer science acceleration techniques. These latter play a key role in rendering the computational time weakly sensitive to the system's multidimensional and multiphysics complexity.