Perturbed-parameter simulations of the MJO with CAM5

 
Poster PDF

Authors

Stephen Klein — Lawrence Livermore National Laboratory

Richard Neale — NOAA - CIRES Climate Diagnostics Center
Shaocheng Xie — Lawrence Livermore National Laboratory
Kenneth R. Sperber — Lawrence Livermore National Laboratory

Category

Modeling

Description

The Madden-Julian Oscillation (MJO) is a dominant source of intraseasonal variability precipitation and is of a subject of strong interest to the ARM Climate Research Facility/Atmospheric System Research (ASR) program. In an attempt to understand and improve the typically poor simulations of the MJO by climate models, we examine perturbed-parameter simulations with the fifth version of the Community Atmosphere Model. Our ensemble consists of over 700 6-year atmosphere-only simulations with prescribed sea-surface temperatures in which 22 parameters contained in the parameterizations of shallow convection, deep convection, and large-scale clouds from a predetermined range of their acceptable values have been perturbed with a latin hypercube sampling methodology. Initial analysis suggests that the MJO could be improved by having a small value to the time-scale for adjustment in the deep convection scheme shorter, a large value to the entrainment for deep convection, and less efficient autoconversion of deep convective condensate to precipitation. The connection between the simulation of MJO in hindcast (i.e., Cloud-Associated Parameterizations Testbed [CAPT]) and free-running mode will also be discussed.