Characteristics of mixed-phase clouds. II: A climatology from ground-based lidar

In part I it was demonstrated that supercooled liquid-water clouds can occur in the form of thin but radiatively significant layers that are distinctive in lidar imagery due to their high backscatter coefficient. In this paper, 18 months of near-continuous lidar data from two midlatitude locations are analysed to estimate the frequency of occurrence of such clouds as a function of temperature. An algorithm is developed that uses the integrated backscatter to identify liquid-water clouds with a visible optical depth of greater than 0.7 (i.e. those that scatter more than half of the incident radiation), and is found to compare favourably with microwave-radiometer measurements of liquid-water path. From data taken with a lidar pointing at 5degrees from zenith, the frequency of supercooled liquid-water layers over Chilbolton in southern England is found to fall steadily with temperature; 27% of clouds between -5 degreesC and -10 degreesC are found to contain significant liquid water, failing to only 6% of clouds observed between -25 degreesC and -30 degreesC. The horizontal extent of the layers typically ranges between 20 and 70 km. When the lidar is pointed directly at zenith, specular reflection by horizontally aligned plate crystals is found to bias the statistics between -10 degreesC and -20 degreesC. The importance of supercooled liquid-water clouds in the radiation budget is reduced when thick ice clouds are present above them, so we then use simultaneous cloud radar data to estimate the optical depth of any cloud above. It is found that around 30% of supercooled liquid-water clouds with temperatures between 0 degreesC and -20 degreesC have ice above with a visible optical depth in excess of 0.5, falling to 10% between -20 degreesC and -30 degreesC. Given the substantial optical depth of the supercooled water itself, we conclude that in the majority of cases when supercooled water is present, it will dominate the radiative properties of the cloud profile. Finally, we compare the occurrence of supercooled-liquid clouds with the amounts found in the models of the UK Met Office and the European Centre for Medium-Range Weather Forecasts (ECMWF). Both models are found to produce too much supercooled liquid at warmer temperatures and too little at colder temperatures (with virtually none being simulated below -20 degreesC), although the occurrence of supercooled cloud is far higher in the ECMWF model than the Met Office model. The observations in this paper are limited to one climatic zone but the forthcoming spaceborne lidars will be able to extend these comparisons to the whole globe.