Development of a global multi-layered cloud retrieval system

The accurate quantification of ice cloud properties is essential for the characterization of global hydrological and radiation budgets. This quantification is often complicated by the occurrence of multi-layer clouds and cloud overlap. Current satellite cloud retrievals are usually based on the assumption that all clouds are a homogenous single-layer, despite the frequent occurrence of overlapped cloud systems. Cloud overlap can contribute large errors to the retrieval of many cloud properties, including ice water path (IWP), cloud height, optical depth, thermodynamic phase, and particle size. For multi-layered cloud systems with ice clouds above water clouds, one of the greatest impediments to accurately determining cloud ice mass is the influence of underlying liquid water clouds and precipitation on the radiances observed at top-of-atmosphere (TOA). The optical depth derived from the reflected visible and infrared radiance represents the combined effects of all cloud layers. This effect causes the optical depth to be overestimated by approximately 40% due to the influence of the underlying cloud. Radiative transfer calculations also suggest that the underlying clouds must be properly characterized and that the radiative transfer of the overlapped cloud system must be taken into account.

To overcome this problem, we have developed a global multi-layered cloud retrieval system (MCRS). A two-layer cloud model was used to simulated multi-layered cloud radiative characteristics and to develop the visible reflectance parameterization. Based on the results from the two-layer cloud model, two algorithms were developed for the retrieval of multi-layer cloud properties over ocean (MCRS-Ocean) and over land (MCRS-Land). Over ocean, microwave radiances are used to estimate cloud liquid water path (LWP) and cloud water temperature of the lower level water cloud. The cloud properties of the ice cloud are then derived using the MCRS-Ocean algorithm. Over land, the variability in surface emissivity renders the microwave approach nearly useless. Therefore, a parameterization scheme using surface observations and environmental conditions is used to estimate the cloud liquid water path of the lower level cloud instead of the microwave retrievals. The cloud LWP derived from microwave radiometers over four ARM SGP sites is used to validate the parameterization scheme. The MCRS-Ocean and MCRS-Land were applied to TRMM VIRS/TMI and ARM SGP data, respectively. . The preliminary results suggest that the MCRS can significantly improve the accuracy and reduce overestimation of optical depth and IWP. For example, for the 2000 calendar year study period over four ARM SGP sites, the optical depth and IWP were reduced, from original overestimates, by approximately 30.0% and 29.0%, respectively. These results suggest that the combination of different instruments and perspectives should lead to a much improved characterization of ice clouds and their impact on the radiation budget.