(Mis)Identification of Arctic Mixed Phase Clouds By Polarization Lidar

The measurement of liquid only (LO) and mixed phase (MP) clouds in the Arctic is a critical building block in understanding surface energy and mass balances of the region. Due to the dearth of accurate measurements, Arctic LO and MP clouds are poorly represented in weather and climate models at all scales making their regional impact difficult to assess. Furthermore, measurements of LO and MP clouds are challenging because the clouds are predominantly low-lying and optically thick. These physical characteristics place stringent demands on the dynamic range of the observations and introduce large biases in cloud phase identification in current lidar based observational methodologies. This work assesses the biases in cloud micro- and macro-physical property retrievals by the Clouds Aerosols Polarization and Backscatter Lidar (CAPABL) located at Summit Camp, Greenland (72^o35’46.4”N, 38^o25’19.1”W, 3200 m asl). We show that phase identification and cloud base height are tightly linked due to signal intensity.  In observations from CAPABL taken over several months, we found that biases of up to 2 km in the median liquid cloud height estimate can occur between analog and photon counting signal analyses. This relates to false phase identification between the two signal approaches where photon counting analysis is more likely to misidentify liquid clouds as ice clouds.  Methods to improve cloud phase identification based on non-orthogonal polarization retrievals are introduced and demonstrated. In addition to improving the accuracy of cloud phase identification, these new methods increase effective sampling of the height range of interest by as much as 18%.