Convective precipitation variability as a tool for general circulation model analysis

Precipitation variability is analyzed in two versions of the Community Atmospheric Model ( CAM), the standard model, CAM, and a "multiscale modeling framework" (MMF), in which the cumulus parameterization has been replaced with a cloud-resolving model. Probability distribution functions (PDFs) of daily mean rainfall in three geographic locations [the Amazon Basin and western Pacific in December-February (DJF) and the North American Great Plains in June-August (JJA)] indicate that the CAM produces too much light-moderate rainfall (10 similar to 20 mm day(-1)), and not enough heavy rainfall, compared to observations. The MMF underestimates rain contributions from the lightest rainfall rates but correctly simulates more intense rainfall events. These differences are not always apparent in seasonal mean rainfall totals. Analysis of 3-6-hourly rainfall and sounding data in the same locations reveals that the CAM produces moderately intense rainfall as soon as the boundary layer energizes. Precipitation is also concurrent with tropospheric relative humidity and lifted parcel buoyancy increases. In contrast, the MMF and observations are characterized by a lag of several hours between boundary layer energy buildup and precipitation, and a gradual increase in the depth of low-level relative humidity maximum prior to rainfall. The environmental entrainment rate selection in the CAM cumulus parameterization influences CAM precipitation timing and intensity, and may contribute to the midlevel dry bias in that model. The resulting low-intensity rainfall in the CAM leads to rainfall-canopy vegetation interactions that are different from those simulated by the MMF. The authors present evidence suggesting that this interaction may artificially inflate North American Great Plains summertime rainfall totals in the CAM.