The shortwave radiative effect of an ice cloud observed over the Atmospheric Radiation Measurement program's Southern Great Plains site in Oklahoma is investigated. Airborne microphysical data from a cloud particle imager, optical array probes, and forward scattering probes are used to construct vertical profiles of the size and shape distributions of ice crystals. Due to uncertainties associated with measuring the sizes and shapes of small ice crystals with maximum dimensions less than 120 mu m, five alternate size-shape distributions are derived and combined with existing databases of wavelength-dependent single-scattering properties of idealized ice crystals to obtain vertical profiles of optical properties. The dependence of the surface and the top-of-the-atmosphere fluxes on these uncertainties is simulated with a radiative transfer model. In addition, surface fluxes are compared against measurements at the surface. It is found that the differences between the modeled and measured fluxes are too large to be explained by uncertainties in the shape and concentrations of small ice crystals. Sensitivity tests suggest that the discrepancies occur because the real optical thickness is larger than that derived from the aircraft profiles most of the time. When the optical thickness was derived based on modeled and measured direct fluxes, the modeled total downward flux agreed well with the measurements. Slightly (less than 10%) reducing the asymmetry parameter, which is possibly associated with the presence of surface roughness, air bubble inclusions or other nonidealities in ice crystals, may further improve the agreement with observations.