An ensemble constrained variational analysis of atmospheric forcing data and its application to evaluate clouds in CAM5. Issue 1 (5th January 2016)

• Record Type: Journal Article
• Authors: Tang, Shuaiqi; Zhang, Minghua; Xie, Shaocheng
• Is Part Of: Journal of geophysical research. Volume 121:Issue 1(2016)
• Page Start: 33
• Page End: 48
• Publication Date: 2016-01-05
• Subjects: ensemble -- variational analysis -- large‐scale forcing data -- cloud retrievals -- model bias; Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2015JD024167
• Languages: English
• ISSNs: 2169-897X

Abstract

Abstract: Large‐scale atmospheric forcing data can greatly impact the simulations of atmospheric process models (e.g., large eddy simulations, cloud‐resolving models, and single column models (SCMs)) that are used to develop physical parameterizations in global climate models. This study introduces an ensemble variationally constrained objective analysis of atmospheric large‐scale forcing data and its application to evaluate the cloud biases in the Community Atmospheric Model (CAM5). Sensitivities of the variational objective analysis to background data, error covariance matrix, and constraint variables are presented to quantify the uncertainties in the large‐scale forcing data and state variables. Application of the ensemble forcing in the CAM5 SCM during March 2000 intensive operational period at the Southern Great Plains (SGP) of the Atmospheric Radiation Measurement Program shows that the systematic biases in the model simulations (i.e., excessive high clouds and insufficient low clouds) cannot be explained by the uncertainty of large‐scale forcing data, which points to the deficiencies of physical parameterizations. These biases are found to also exist in the global simulation of CAM5 when it is compared with satellite data over the surrounding SGP site for annual and seasonal means. Key Points: An ensemble approach of 3DCVA is introduced The derived forcing data are sensitive to background data and precipitation The uncertainties in forcing data can be used to evaluate model biases