The Importance of Temperature-Dependent Optical Properties for Radiative Closure in Super-Cooled Water Clouds

We present a re-examination of measurements from the 2001 South Pole Atmospheric Radiation and Cloud Lidar Experiment (SPARCLE) of super-cooled water clouds. Measurements from SPARCLE include infrared spectra from a ground-based Atmospheric Emitted Radiance Interferometer (AERI), cloud boundary data from a Micropulse Lidar, and radiosonde profiles. Measurements from a case study on 2 February 2001 show a stratus cloud at -32°C comprised of super-cooled water droplets. As a closure experiment, downwelling radiative thermal emission spectra of simulated clouds were compared to measured AERI spectra. Simulations employing conventional assumptions (spherical droplets and indices of refraction of water at 0°C) resulted in significant discrepancies in key spectral regions that are both sensitive to the presence of a cloud and have high signal-to-noise. These spectral discrepancies could not be resolved by the inclusion of ice in the simulated cloud. Satisfactory comparisons were obtained only after using recently published laboratory measurements of the indices of refraction of super-cooled water over the range of measured cloud temperatures. Possible implications of this result for remote sensing and climate simulation are explored. Because the indices of refraction of supercooled water are in some spectral regions intermediate between ice and liquid at 0°C, cloud phase retrievals may mistake supercooled clouds for mixed-phase clouds. In addition, the inclusion of the temperature-dependence of the optical contstants super-cooled water may be important for both surface and TOA energy budgets.