Abstract. We introduce a new multispectral method for the retrieval of optical thickness and effective radius from cloud transmittance, which is less sensitive to effective radius than cloud reflectance. Based on data from the moderate spectral resolution observations of the Solar Spectral Flux Radiometer (SSFR) and Shortwave Spectroradiometer (SWS), we use the spectral shape of transmitted radiance as a means of retrieving effective radius from cloud transmittance. The observations were taken during the International Chemistry Experiment in the Arctic Lower Troposphere and at the Southern Great Plains (SGP) site of the Atmospheric Radiation Measurement (ARM) Climate Research Facility. The spectral shape was quantified by fitting a slope to the normalized transmittance between 1565 nm and 1634 nm. The retrieval was performed by comparing the observed slope at 1565 nm and the transmittance at 515 nm with a pre-calculated library (lookup table). An estimate of the retrieval uncertainty was provided by propagating the uncertainty of the observations through the best-fit algorithm. We compare the new retrieval with an algorithm that uses transmittance at two wavelengths, a method often used with cloud reflectance. The liquid water path (LWP) is derived from the retrieved optical thickness and effective radius, assuming a cloud with effective radius varying linearly with altitude above cloud base, and compared to the retrieved liquid water path from a microwave radiometer. Retrievals from two MODIS overpasses of the SGP were also compared. The data taken from the SGP was under thicker cloud than the case used from ICEALOT, with average optical thickness of 44 and 22, respectively. For the time period with the thicker clouds, the dual-wavelength method and the slope method retrieved nearly indistinguishable results. The dual-wavelength method, however, resulted in slightly higher average relative effective radius uncertainty of 12.9 μm±12.8%, as compared to 12.8 μm±8.9% from the slope method. The thinner cloud case resulted in a significant difference between the dual-wavelength and slope algorithms with average retrieved effective radius and uncertainties of 12.5 μm±8.4% and 17.0 μm±21.0% for the slope and dual-wavelength methods, respectively. The retrieved optical thickness values for this case were nearly identical. The average derived LWP was within 12.5% and 20% of the MWR LWP for the ARM and ICEALOT data. For a homogeneous cloud case, the MODIS retrievals (optical depth, effective radius, and LWP) were within the uncertainty of the SWS retrievals. Inhomogeneous clouds resulted in lesser agreement between the MODIS and SWS retrievals.;
