On the application of Monte Carlo solutions of the radiative transfer equation for cloudy terrestrial atmospheres

When calculating radiances and radiative fluxes for cloudy terrestrial atmospheres, atmospheric scientists usually rely, with much complacency, on 1D analytic solutions of the radiative transfer equation (RTE). However, when high precision is, or should be, required, and when analytic solutions seem remote or impossible, researchers look to Monte Carlo solutions of the RTE; often as a last resort given their reputation for commandeering computer time. The primary purpose of this talk is to demonstrate that this reputation has waned somewhat as of late due to: i) recognition that atmospheric models can tolerate large amounts of radiative noise (stemming from numerics); ii) the obvious efficiency inherent to Monte Carlo algorithms on parallel computers; and iii) the development of atmospheric models that resolve clouds sufficiently well thereby making it worthwhile to step up from ID to 3D radiative transfer. Discussions will be limited to the use of Monte Carlo algorithms to compute broadband heating rates in atmospheric dynamical models (from cloud-resolving models at one extreme and global climate models at the other), and to infer cloud properties from satellite imagery.