This study examines the vertically resolved cloud measurements from the cloud-aerosol lidar with orthogonal polarization instrument on Aqua satellite from June 2006 through May 2007 to estimate the extent to which the mixed cloud-phase composition can vary according to the ambient temperature, an important concern for the uncertainty in calculating cloud radiative effects. At −20°C, the global average fraction of supercooled clouds in the total cloud population is found to be about 50% in the data period. Between −10 and −40°C, the fraction is smaller at lower temperatures. However, there are appreciable regional and temporal deviations from the global mean (> AE 20%) at the isotherm. In the analysis with coincident dust aerosol data from the same instrument, it appears that the variation in the supercooled cloud fraction is negatively correlated with the frequencies of dust aerosols at the −20°C isotherm. This result suggests a possibility that dust particles lifted to the cold cloud layer effectively glaciate supercooled clouds. Observations of radiative flux from the clouds and earth's radiant energy system instrument aboard Terra satellite, as well as radiative transfer model simulations, show that the 20% variation in the supercooled cloud fraction is quantitatively important in cloud radiative effects, especially in shortwave, which are 10 − 20 W m −2 for regions of mixed-phase clouds affected by dust. In particular, our results demonstrate that dust, by glaciating supercooled water, can decrease albedo, thus compensating for the increase in albedo due to the dust aerosols themselves. This has important implications for the determination of climate sensitivity.aerosol-cloud interaction | ice nucleation | mixed cloud phase | super cooled water | cloud radiative effect C old clouds that consist of mixed-phase particles, are ubiquitous in the Earth's upper and middle troposphere. Changes in the liquid/ice-phase composition in such clouds may significantly affect the radiative balance of the earth atmosphere system because the cloud radiative properties for both the shortwave (SW) and longwave (LW) such as cloud optical depth, single-scattering albedo, and emissivity vary according to the phase of cloud particles (1-3). It is therefore of fundamental importance to examine the variation in the cloud-phase composition for accurate calculations of cloud radiative effects.In fact, the cloud-phase composition in mixed-phase clouds is complicatedly affected by several factors other than temperature; e.g., ice nuclei (IN) aerosols. Most of climate models, however, have calculated the cloud-phase composition with limited sophistication, simply as a function of grid-mean temperature (4). For example, it has been typically assumed that clouds are composed entirely of ice and liquid particles below −40°C and above 0°C, respectively, of a mixture of ice and liquid phases between −40 and 0°C. The fraction of supercooled liquid particles within mixed-phase clouds is represented as a linear function of temperatures in genera...