Skip to Main Content U.S. Department of Energy
Science Directorate
Page 737 of 740

Atmospheric Sciences & Global Change
Research Highlights

May 2004

New Method for Evaluating Cloud Processes and Cloud Effects in Climate Models

The Multi-scale Modeling Framework (MMF) is a new, computationally intensive approach to climate modeling being used by Pacific Northwest National Laboratory scientists to produce a more realistic treatment of cloud properties and processes. It consists of a coarse resolution global climate model (horizontal scale of ~300 km) with a cloud resolving model (horizontal scale of ~4 km) nested into each climate model grid cell.

Using this approach, scientists are evaluating the reliability of cloud simulations by any climate model. The multi-year data sets available from the Atmospheric Radiation Measurement (ARM) Program provide a rich environment for statistical evaluations. In the figure, the left panel shows a probability distribution of the net solar flux measured at the ARM Nauru site, averaged for a three-hour period each day around local noon. The mean value is 767 W/m2 and the range is 40 to 992 W/m2. The state-of-the-art NCAR Community Atmosphere Model (CAM) simulates too little solar radiation reaching the surface (a mean of 701 W/m2) and too small a range of values. PNNL's research shows this deficiency results from an over-production of clouds by the CAM. The MMF model has a mean value identical to the ARM observations, which is fortuitous. The MMF does not produce a sufficient number of thick clouds, making its minimum values too high.

Our initial application of MMF shows both promise and the need for additional work. However, unlike conventional cloud parameterizations, the path to model improvement is more straightforward. Repeated testing for many variables at multiple sites will enable improved evaluation of model performance and identification and correction of deficiencies.

Page 737 of 740

Science at PNNL

Core Research Areas

User Facilities

Centers & Institutes

Additional Information

Research Highlights Home


Print this page (?)

YouTube Facebook Flickr TwitThis LinkedIn