Sensitivity of summer ensembles of super-parameterized US mesoscale convective systems to cloud resolving model microphysics and resolution

Friday, 19 December 2014
Elizabeth Elliott1, Sungduk Yu1, Gabriel J Kooperman2, Hugh Morrison3, Minghuai Wang4 and Michael S Pritchard1, (1)University of California Irvine, Irvine, CA, United States, (2)University of California San Diego, La Jolla, CA, United States, (3)National Center for Atmospheric Research, Boulder, CO, United States, (4)Pacific Northwest National Lab, Richland, WA, United States
Microphysical and resolution sensitivities of explicitly resolved convection within mesoscale convective systems (MCSs) in the central United States are well documented in the context of single case studies simulated by cloud resolving models (CRMs) under tight boundary and initial condition constraints. While such an experimental design allows researchers to causatively isolate the effects of CRM microphysical and resolution parameterizations on modeled MCSs, it is still challenging to produce conclusions generalizable to multiple storms. The uncertainty associated with the results of such experiments comes both from the necessary physical constraints imposed by the limited CRM domain as well as the inability to evaluate or control model internal variability. A computationally practical method to minimize these uncertainties is the use of super-parameterized (SP) global climate models (GCMs), in which CRMs are embedded within GCMs to allow their free interaction with one another as orchestrated by large-scale global dynamics. This study uses NCAR’s SP Community Atmosphere Model 5 (SP-CAM5) to evaluate microphysical and horizontal resolution sensitivities in summer ensembles of nocturnal MCSs in the central United States. Storm events within each run were identified using an objective empirical orthogonal function (EOF) algorithm, then further calibrated to harmonize individual storm signals and account for the temporal and spatial heterogeneity between them. Three summers of control data from a baseline simulation are used to assess model internal interannual variability to measure its magnitude relative to sensitivities in a number of distinct experimental runs with varying CRM parameters. Results comparing sensitivities of convective intensity to changes in fall speed assumptions about dense rimed species, one- vs. two-moment microphysics, and CRM horizontal resolution will be discussed.