Cirrus clouds have a profound impact on the radiation balance of the Earth--atmosphere system. Accurate representation of their radiative properties is critical to understanding climate and predicting climate change. This paper casts the split-window cirrus cloud retrieval technique in an optimal estimation framework facilitating direct inclusion of explicit cloud boundary information from complementary sensors as well as providing a suite of diagnostic tools for evaluating the dominant sources of uncertainty in all retrieved quantities. Errors in retrieved microphysical properties are used to determine the resulting errors in the calculation of global-scale radiative budgets. Uncertainties in optical depth and effective radius are found to diminish from ~45% and ~80%, respectively, in the absence of explicit cloud boundary information to ~15% and ~60% when accurate radar-based or lidar-based estimates are included. It is demonstrated that the improvements to cirrus cloud optical properties afforded by accurate cloud boundary information may lead to as much as a factor of 3 increase in the accuracy to which their impact on the Earth's radiative balance can be modeled. Colocated infrared radiances from the Moderate-Resolution Imaging Spectroradiometer (MODIS) instrument aboard the Earth Observing System (EOS) Aqua satellite and cloud radar observations from the CloudSat satellite will soon allow the retrieval presented here to be integrated into an operational retrieval of the vertical distribution of cloud properties on a global scale.