*On the construction of porous geometry*

For all particles we utilize the experimentally-determined complex index of refraction by Patsalas et. al. \cite{Patsalas_2002}.

*on the DDA* The discrete dipole approximation (DDA) was used to solve Maxwell’s equations. Details of the numerical method are summarized in \cite{Randrianalisoa_2014} as well as in the review by Draine and Flatau \cite{Draine_1994}

*on the chosen dipole spacing and numerical validation*

A summary of the simulation parameters can be found in Table \ref{tab:DDA_summary}

We calculate the radiative properties of clouds made up of monodisperse, cold, uniformly distributed particles of three different types. Representations and descriptions of the three particle types can be found in Figure \ref{fig:problem_setup}(a)-(b). The radiative transfer equation (RTE) has been solved in one dimension using the Monte Carlo method to recover the cloud reflectivity, absorbtivity, and spatially dependent divergence of the radiative flux. Using the notation of \cite{Modest_2013} the RTE is given by,

The computational domain is represented in \ref{fig:problem_setup}(d). On the left-hand side, radiative flux is incident on the computational domain with incoming angles between -15 and 15 degrees and a spectral distribution corresponding to the emission from Xenon arc lamps. This situation was chosen to match the irradiance of the inner lamps of the recently constructed solar simulator described in \cite{Bader_2014a}, \cite{Leveque} resulting in a uniform power input of \(\approx 1000\) suns; future experimental studies are anticipated with this physical setup.

Cloud scattering calculations were validated against single scattering measurements and theoretical calculations of bidirectional reflectivity and transmissivity of polystyrene spheres as a function incidence angle\cite{Hottel_1970}.

Talk about Lorenz-Mie theory and VAT.

Here we employ the correlations from José found in \cite{Rhodes_2008}. We assume the suspension is dilute such that the particles are entrained and fluid–particle interactions dominate such that each particle may be treated as an isolated, suspended particle We choose a density of particles corresponding to just over the choking velocity, such that we can assume the particles are staying within the irradiation zone for sufficient time. This estimate is based on the correlation recommended in \cite{Rhodes_2008} for vertical pneumatic transport. In the this limit of dilute transport fluidization, it is expected that there will be no variations in axial concentration of the particles \cite{Jakobsen_2014}.