Climate Model Improvements At present, climate-focused models are transitioning to earth system models that include a broader range of parameters (Box 4.2; Figure 4.8). Some submodules have already been linked to climate models for studies of future and past climate (e.g., dynamic vegetation changes, fire ecology, ice sheet growth and decay, and hydrology of lakes and rivers). A focused effort is required to develop new submodels that are especially applicable to the kinds of past climate experimentation envisioned in this program. Examples would include simulating plant-animal interactions and community ecology; simulating isotope fractionation in evaporation/precipitation cycles; simulating the biogeochemistry of lakes and oceans, including sedimentation; and simulating the sources, transports, and sinks of dust. Simplified versions of such submodules are already being constructed and tested for some climate models, but much additional development, refinement, and testing are required. The addition of these new submodules will provide model output that would increasingly match the environmental signals recorded in sediments and other fossil records.
Computational Resources Requirements A major increase in computational resources will be required to simulate the range of specific time intervals over the past 8 Ma needed to address the relationship between paleoclimates and hominin evolution and dispersal. Multiple simulations will be required to test the sensitivity of the results to uncertainties in the forcing variables (e.g., CO2 levels, topographic characteristics). Simulations at high spatial resolution will be required to resolve the relatively fine-scale details of climate, vegetation, and hydrology that are contained in environmental records in regions of complex local topography, such as the East African Rift System.
Support Requirements No new physical facility is necessarily required—computer resources can be located at existing facilities, and then be made available to the community via virtual networks. A small support staff of scientists and technicians would be required to facilitate the research activities of the scientists from many disciplines that would be involved in these coordinated interdisciplinary studies, and capabilities for archiving and retrieving model output and observations and graphical and statistical tools for efficient model/data comparison would also be needed. The model/data storage facilities and the small support staff could be located at an existing facility. Although increased computational resources are urgently needed for U.S. scientists working in this field, possibilities may also exist for shared computational facilities with the European Union and other nations or groups of nations. It will also be important to broaden electronic access to computational resources and data/model archives so that scientists from Africa and Eurasia can collaborate with U.S. scientists.