New Computer Model Concept Could Solve Big, Real-World Problems On A Small, Porous Scale

The Department of Energy’s Pacific Northwest National Laboratory today was awarded a Scientific Discovery through Advanced Computing, or SciDAC, grant to develop a computer model that can simulate biogeochemical processes on multiple scales. This computational advancement would enable researchers to make more accurate predictions of the movement and fate of contaminants in groundwater so that appropriate cleanup and human safety measures can be applied to the problem.

DOE today announced $60 million in grants for 30 SciDAC projects over the next three to five years.

Specific to the PNNL-led project, the SciDAC program provides the framework within which computational and earth scientists will work together to bring the most powerful computers and latest software developments to bear on the challenging groundwater problem.

Organizations collaborating with PNNL on the modeling project will be the University of California, San Diego; Idaho National Laboratory; and Oak Ridge National Laboratory.

Subsurface transport processes play a central role in issues that are of concern globally, as in DOE missions, such as cleanup of Cold War legacy contamination, development of energy resources, reduction of global climate impacts through subsurface sequestration of carbon dioxide, and safe storage of subsurface waste for the long-term.

Computer models that simulate groundwater’s behavior already exist, but with limited ability to share information between models at different scales. Researchers hope that a new, sophisticated model will be able to combine individual, small-scale simulations and this will create a larger picture where real-world problems can be analyzed and, hopefully, solved with a single model instead of developing them on a case-by-case basis.

Big problems, such as cleaning up contaminated groundwater at old plutonium production sites, are often complex on the surface, and even more complicated below the surface – that is, at the “pore” scale.

Consider a five-foot pathway of water that’s flowing underground. It isn’t a “clean” tunnel; rather, it’s a mixture of water, minerals, contaminants, and dirt, making it a stew that researchers must examine closely in order to determine how and where the water will flow.

If researchers simply examined the stew on the large scale, they might neglect small-scale (e.g., microscopic- or pore-scale) processes that ultimately control the large-scale behavior. But with a sophisticated computer model that marries various simulations representing different processes on a variety of scales, ranging from microbial growth in pores to migration of contaminants to rivers or wells, researchers get a clearer picture of what’s happening underground now and what’s likely to happen in the future.

A similar project that received a SciDAC grant will seek to enhance groundwater monitoring so that researchers can better understand the migration of radionuclides. The knowledge gained from detailed modeling could help with remediation efforts in locations where cleaning up the groundwater is necessary to avoid contamination of areas it contacts, such as rivers. PNNL will collaborate with Los Alamos National Laboratory on the project.

Other SciDAC projects that PNNL will collaborate on, are: A Scalable and Extensible Earth System Model for Climate Change Science, led by Oak Ridge National Laboratory; Design and Testing of a Global Cloud-Resolving Model, led by Colorado State University; Modeling Multiscale-Multiphase-Multicomponent Subsurface Reactive Flows using Advanced Computing, led by Los Alamos National Laboratory; Petascale Data Storage Institute, led by Carnegie Mellon University; Center for Interoperable Technologies for Advanced Petascale Simulations, led by Lawrence Livermore National Laboratory; Center for Technology for Advanced Scientific Component Software, led by ORNL; and Scientific Data Management Center for Enabling Technologies, led by Lawrence Berkeley National Laboratory.

The SciDAC program was launched in 2001 to develop new tools and techniques for advancing scientific research through computer simulation in all mission areas of DOE’s Office of Science.

Article Written by