Researchers: John Drake et al
Raw data, data analyses, and metadata are collected, organized, and made available to researchers worldwide. Data repositories at ORNL include CDIAC and ARM.
Computational models are developed to simulate the workings and interactions of the processes involved in the global environment. Challenges include upscaling and downscaling to link global- and regional- scale models, coupling different models (such as ocean and atmosphere), and incorporating additional environmental processes, such as groundwater.
- What are the best numerical algorithms for computing atmosphere and ocean circulations?
- How can DOE parallel computers be effectively utilized for climate simulation?
- How can model development be accelerated while maintaining model quality and integrity?
- How can collaborative technologies be extended to support an open, multi-platform code design and implementation?
Computational models are used to construct and analyze various scenarios. High resolution runs (T85 and up) of coupled models and ensemble runs (like Warren Washington and the NCAR crew are doing) can require significant commitments of computational resources.
- What are the margins of accuracy and predictability for reliable climate predictions?
- On what time and spatial scales are climate change predictions useful for assessment of impacts?
The effects that could result from a scenario are studied. One example is the statistical analyses work by Forrest Hoffman and Bill Hargrove used to calculate ecoregions.
- What effect would an increase in CO2 saturation have on vegetation?
- What effect would an increase in average global temperature have on existing ecoregions?
What do we do about it? For info on ORNL's Carbon Management activities contact Mike Farrell.