19 September 2014

Analyzing the Sensitivity of Cloud Properties to CLUBB Parameters in the Single Column Community Atmosphere Model (SCAM5)


Large uncertainties remain in simulating clouds in global climate models, and part of these uncertainties result from multiple tunable parameters in cloud parameterizations. A multi-institutional team led by U.S. Department of Energy scientists at Pacific Northwest National Laboratory investigated the sensitivity of simulated shallow cumulus and stratocumulus clouds. The investigation selected tunable parameters in a newly-implemented cloud scheme in the single-column version of the Community Atmosphere Model version 5 (SCAM5), called Cloud Layers Unified by Binormals (CLUBB). The team found that most of the variance in simulated cloud fields can be explained by a small number of tunable parameters. They used a quasi-Monte Carlo sampling approach to explore the high-dimensional parameter space in CLUBB and adopted a generalized linear model to study the responses of simulated cloud fields to tunable parameters. They found that among 40 to 50 tunable parameters in CLUBB, only a handful of parameters are influential. The identified influential parameters are different for different types of clouds. This study improves the understanding of the parameter-dependence of this newly implemented scheme. This study helps to reduce the number of tunable parameters for ongoing sensitivity and calibration study of global simulations.

Zhun Guo
Guo, Z, M Wang, Y Qian, VE Larson, S Ghan, M Ovchinnikov, PA Bogenschutz, C Zhao, G Lin, and T Zhou.  2014.  "A Sensitivity Analysis of Cloud Properties to CLUBB Parameters in the Single-column Community Atmosphere Model (SCAM5)."  Journal of Advance in Modeling Earth Systems, doi:10.1002/2014MS000315.