Current computer climate models can't accurately predict cloud formation, which, in turn, hinders their ability to forecast climate change from human activities. "Because of their coarse resolution, computer models produce values on large spatial scales (hundreds of kilometers) and can only represent large cloud systems," Nenes said.
Aerosol particles, however, are extremely small and measured in micrometers. This means predictive models must address processes taking place on a very broad range of scale. "Equations that describe cloud formation simply cannot be implemented in climate models," Nenes said. "We don't have enough computing power -- and probably won't for another 50 years. Yet somehow we still need to describe cloud formation accurately if we want to understand how humans are affecting climate."
To address the lack of computer power and shortcomings of existing parameterization, Nenes and his research team have developed simple, physics-based equations that link aerosol particles and cloud droplets. Then these equations can be scaled up to a global level, providing accurate predictions thousands of times faster than more detailed models.
This modeling method has proven successful in two field tests. Data was collected from aircraft flying through from cumulus clouds off the coast of Key West, Fla., in 2002, and from stratocumulus clouds near Monterey, Calif., in 2003. Compared with this real-world data, predictions from Nenes' model were accurate within 10 to 20 percent.
"We never expected to capture the physics to that degree," Nenes explained. "We were hoping for a 50 percent accuracy rate."
Another challenge in predicting climate change is to understand how aerosols' chemistry affects cloud formation. Each particle has a different potential for forming a cloud droplet, which depends on its composition, l
'"/>
14-Dec-2004