A probabilistic approach to determine the thermodynamic cloud phase using passive satellites

The cloud thermodynamic phase (ice / mixed-phase / liquid) is a crucial parameter to understand the earth radiation budget, hydrological cycle and atmospheric thermodynamic processes. The phase partitioning of clouds and their parameterization in global climate models have therefore become of particular interest.

To improve our understanding of the frequency of occurrence and temporal evolution of cloud phase, geostationary passive sensors can be very useful due to their wide field of regard and high temporal resolution. However, the retrieval of cloud phase using passive instruments is challenging since the spectral signature of the phase is weak compared to other parameters of the clouds and atmosphere. Especially the distinction between ice and mixed-phase clouds is difficult and previous efforts to retrieve cloud phase often only distinguished between ice and liquid phase.

We present a new method to detect clouds and retrieve their phase using the passive instrument SEVIRI aboard the geostationary satellite Meteosat Second Generation. The method uses probabilities derived from active observations (the Lidar-Radar product DARDAR) of cloud top phase. Combining these probabilities for different SEVIRI channels gives probabilities for the presence of a cloud and for its cloud top phase. Our probabilistic approach includes a measure of uncertainty and allows us to distinguish between ice, mixed-phase, supercooled liquid, and warm liquid clouds. The method is tested against active satellite measurements and shows good agreement. Finally, we discuss its advantages and limitations. In the future, we plan to use our method to study the microphysical (such as optical thickness and effective radii) and macrophysical (such as temporal evolution and extent) properties of ice and mixed-phase clouds.