ARM SCM Intercomparison Helps Find Cloud Parameterization Bug
| Klein, Stephen | Lawrence Livermore National Laboratory |
The ARM Cloud Parameterization and Modeling working group has carried out several intercomparisons of Single Column Models (SCM) and Cloud Resolving Models (CRMs) to observations. The most recent intercomparison involves the simulation of clouds during the March 2000 Cloud Intensive Observing Period at the Southern Great Plains. The sub-period of March 2 - 4, during which a very strong cold front passed over the region, was intensively studied. During this period, the amount of cloud condensate varied dramatically between SCMs. The SCM of the Geophysical Fluid Dynamics Laboratory (GFDL), in particular, simulated a very large amount of liquid water in the layer of the atmosphere just beneath the freezing level. Because the SCM intercomparison suggested that the GFDL SCM was an outlier, the reasons for the large amount of liquid condensate were investigated. It was found that the numerical way in which the melting of snow was handled erroneous-ly prevented accretion and riming of the cloud liquid in the layer just beneath freezing. This error is essentially the result of the coarse vertical resolutionof the SCM. SCM simulations in which this bug is fixed show significantly reduced amounts of liquid in the simulation of the storm. At the meeting, simulations of the impact of the bug fix on the climate of the GFDL Global Climate Model will be presented. This example represents a small success of the ARM program - by focussing the attention of the GCM developers on the detailed time evolution of clouds during a case study, a bug was found which would not be have found from analysis of monthly mean output of climate models.
This poster will be displayed at the ARM Science Team Meeting.