We present a new set of analytic models for the expansion of HII regions powered by UV photoionisation from massive stars and compare them to a new suite of radiative magnetohydrodynamic simulations of turbulent, self-gravitating molecular clouds. To
perform these simulations we use the Eulerian adaptive mesh magnetohydrodynamics code RAMSES-RT, including radiative transfer of UV photons. Our analytic models successfully predict the global behaviour of the HII region provided the density and velocity structure of the cloud is known. We give estimates for the HII region behaviour based on a power law fit to the density field assuming that the system is virialised. We give a radius at which the ionisation front should stop expanding (stall). If this radius is smaller than the distance to the edge of the cloud, the HII region will be trapped by the cloud. This effect is more severe in collapsing clouds than in virialised clouds, since the density in the former increases dramatically over time, with much larger photon emission rates needed for the HII region to escape a collapsing cloud. We also measure the response of Jeans unstable gas to the HII regions to predict the impact of UV radiation on star formation in the cloud. We find that the mass in unstable gas can be explained by a model in which the clouds are evaporated by UV photons, suggesting that the net feedback on star formation should be negative
The treatment of radiation transport in global circulation models (GCMs) is crucial to correctly describe Earth and exoplanet atmospheric dynamics processes. The two-stream approximation and correlated-k method are currently state-of-the-art approxim
ations applied in both Earth and hot Jupiter GCM radiation schemes to facilitate rapid calculation of fluxes and heating rates. Their accuracy have been tested extensively for Earth-like conditions, but verification of the methods applicability to hot Jupiter-like conditions is lacking in the literature. We are adapting the UK Met Office GCM, the Unified Model (UM), for the study of hot Jupiters, and present in this work the adaptation of the Edwards-Slingo radiation scheme based on the two-stream approximation and the correlated-k method. We discuss the calculation of absorption coefficients from high temperature line lists and highlight the large uncertainty in the pressure-broadened line widths. We compare fluxes and heating rates obtained with our adapted scheme to more accurate discrete ordinate (DO) line-by-line (LbL) calculations ignoring scattering effects. We find that, in most cases, errors stay below 10 % for both heating rates and fluxes using ~ 10 k-coefficients in each band and a diffusivity factor D = 1.66. The two-stream approximation and the correlated-k method both contribute non-negligibly to the total error. We also find that using band-averaged absorption coefficients, which have previously been used in radiative-hydrodynamical simulations of a hot Jupiter, may yield errors of ~ 100 %, and should thus be used with caution.
