3D Radiative Transfer in Cloudy Atmospheres: Diffusion Approximation and Monte Carlo Simulation for Thermal Emission

Kuo-Nan Liou

Clouds, which occupy more than 50% of the sky, are generally finite and
inhomogeneous. They are the most important element in modulating the
energy budget of the Earth-atmosphere system and hence climate. The
potential effects of cloud geometry and inhomogeneity on the transfer of
radiation must be carefully studied in order to understand their impact on the
radiative properties of the atmosphere. Moreover, incorporation of these effects
on radiative transfer in climate and general circulation models (GCMs) remains
one of the most difficult problems due to the complexity of cloud formation
treatment in the model and the associated radiative transfer calculations. All
GCMs at present consider clouds to be plane-parallel and homogeneous with
respect to radiation calculations.
We developed a solution for the 3D diffusion radiative transfer equation in
Cartesian coordinates utilizing a four-term spherical harmonics expansion for the
scattering phase function and intensity, intended for potential incorporation in
climate models, The general second-order partial differential equation derived
from the diffusion approximation has been solved by employing an efficient full
multigrid method, which can simulate the transfer of solar and thermal infrared
radiation in inhomogeneous cloudy atmospheres with different boundary
conditions, including sharp discontinuity. The correlated k-distribution method
has been used for the sorting of gaseous absorbing lines in association with
multiple-scattering atmospheres for the calculation of fluxes covering the entire
solar and thermal infrared spectra. Comparison of the broadband fluxes and
heating rates computed from this approach to those from the equivalent
plane-parallel and 3D Monte Carlo models shows excellent agreement. For
thermal emission, we developed a 3D Monte-Carlo model for specific
application to broadband infrared radiative transfer that differs from the
conventional use of point source. The model domain consists of a variable
structure based on cubic grid cells in which the emissivities for gases and cloud
particles are parameterized.
We applied the 3D radiative transfer models to the 3D cirrus cloud fields
constructed from remote sensing on the basis of a unification of satellite and
ground-based cloud profiling radar observations over the DOE ARM-SGP site.
Two cases, which illustrate substantial horizontal and vertical variability in ice
crystal size distribution, have been chosen for the analysis of 3D fluxes and
heating rates in terms of mean and variance in different spatial scales. Pertinent
results are presented and physically discussed.

Presentation (PowerPoint File)

Back to Computational Methods in Transport