Meetings > PI / Science Team Meeting > Posters

Understanding uncertainties in current cloud retrievals from ARM ground-based measurements

 
Poster PDF

Authors

David D. Turner — NOAA- Global Systems Laboratory
Gerald Mace — University of Utah
Michael Jensen — Brookhaven National Laboratory
Zhien Wang — University of Colorado
Shaocheng Xie — Lawrence Livermore National Laboratory
Sally A. McFarlane — U.S. Department of Energy
Renata B. McCoy — Lawrence Livermore National Laboratory
Stephen Klein — Lawrence Livermore National Laboratory
Jennifer M. Comstock — Pacific Northwest National Laboratory
Matthew Shupe — University of Colorado
Min Deng — University of Colorado, Boulder
Robin J. Hogan — University of Reading
Maureen Dunn — Brookhaven National Laboratory
Dong Huang — NASA - Goddard Space Flight Center
Chuanfeng Zhao — Beijing Normal University

Category

Cloud Properties

Description

Accurate observations of cloud microphysics are critical to improve the representation of clouds in current climate models. However, large discrepancies are found in current cloud retrieval products. This poses a severe restriction in cloud modeling studies. Understanding the discrepancies is an important step to address the uncertainties in cloud retrievals. In this study, an in-depth analysis of seven existing ARM ground-based cloud retrievals is carried out. High-level ice clouds and boundary layer stratus clouds, which are the focus of most current retrievals, are specifically studied. Differences in the algorithm complexity and assumptions of cloud retrievals, the data used, and the cloud retrieval constraints are analyzed to understand their potential impacts on the retrieved products. It shows that the large discrepancies between different cloud retrievals, both in the retrieved cloud properties and the relationships between different cloud properties, can be partly expected from the differences in the retrieval techniques, like the algorithm parameters and the assumptions of particle-size distributions and ice crystal habits. The impact of input data used in the retrievals, such as cloud boundaries, cloud phase, and hydrometer classifications, is also an important contributor to the large discrepancies of cloud properties between different retrieval products. It indicates the need of improving accuracy and consistency in the input measurements for current retrievals. The impact of cloud retrieval constraints, such as MWR liquid water path, is also briefly described. A statistical cluster analysis technique is used to further demonstrate the systematic discrepancies between different retrievals.