Bayesian Exploration of Cloud Microphysical Sensitivities in Mesoscale Cloud Systems

Abstract:

It is well known that changes in cloud microphysical processes can have a significant effect on the structure and evolution of cloud systems. In particular, changes in water phase and the associated energy sources and sinks have a direct influence on cloud mass and precipitation, and an indirect effect on cloud system thermodynamic properties and dynamics. The details of cloud particle nucleation and growth, as well as the interactions among vapor, liquid, and ice phases, occur on scales too small to be explicitly simulated in the vast majority of numerical models. These processes are represented by approximations that introduce uncertainty into the simulation of cloud mass and spatial distribution and by extension the simulation of the cloud system itself.

This presentation demonstrates how Bayesian methodologies can be used to explore the relationships between cloud microphysics and cloud content, precipitation, dynamics, and radiative transfer. Specifically, a Markov chain Monte Carlo algorithm is used to compute the probability distribution of cloud microphysical parameters consistent with particular mesoscale environments. Two different physical systems are considered. The first example explores the multivariate functional relationships between precipitation, cloud microphysics, and the environment in a deep convective cloud system. The second examines how changes in cloud microphysical parameters may affect orographic cloud structure, precipitation, and dynamics. In each case, the Bayesian framework can be shown to provide unique information on the inter-dependencies present in the physical system.