Combined retrieval of Arctic liquid water cloud and surface snow properties using airborne spectral solar remote sensing

摘要

The passive solar remote sensing of cloud properties over highlyreflecting ground is challenging, mostly due to the low contrastbetween the cloud reflectivity and that of the underlyingsurfaces (sea ice and snow). Uncertainties in the retrievedcloud optical thickness and cloud droplet effectiveradius may arise from uncertainties in theassumed spectral surface albedo, which is mainly determined bythe generally unknown effective snow grain size .Therefore, in a first step the effects of the assumed snow grainsize are systematically quantified for the conventionalbispectral retrieval technique of and for liquid water clouds. In general, the impact of uncertaintiesof is largest for small snow grain sizes. Whilethe uncertainties of retrieved are independent of thecloud optical thickness and solar zenith angle, the bias ofretrieved increases for optically thin cloudsand high Sun. The largest deviations between the retrieved andtrue original values are found with 83 % for and 62 %for .In the second part of the paper a retrieval method is presented thatsimultaneously derives all three parameters (, , ) and therefore accounts for changesin the snow grain size. Ratios of spectral cloudreflectivity measurements at the three wavelengths = 1040 nm (sensitive to ), = 1650 nm (sensitive to ), and = 2100 nm (sensitive to ) arecombined in a trispectral retrieval algorithm. In a feasibilitystudy, spectral cloud reflectivity measurements collected by theSpectral Modular Airborne Radiation measurement sysTem (SMART)during the research campaign Vertical Distribution of Ice inArctic Mixed-Phase Clouds (VERDI, April/May 2012) were used totest the retrieval procedure. Two cases of observations abovethe Canadian Beaufort Sea, one with dense snow-covered sea iceand another with a distinct snow-covered sea ice edge areanalysed. The retrieved values of , , and show a continuous transition of cloud propertiesacross snow-covered sea ice and open water and are consistentwith estimates based on satellite data. It is shown that theuncertainties of the trispectral retrieval increase for highvalues of , and low but nevertheless allowthe effective snow grain size in cloud-covered areas to be estimated.