SGI's Cray T3E at EPA to Help States Gain Cleaner AirMountain View 02 Sep 99"Our goal is to provide the tools states need to meet federal clean air standards," said Dr. Gary Foley, director of the National Exposure Research Laboratory, U.S. Environmental Protection Agency (EPA), headquartered in Research Triangle Park, N.C. "The models we will create with the supercomputer are as close to reality and the ability to predict the future as we can get. The ability to run these more comprehensive models and, just as important, visualize their results enables analysts to get a nearly immediate grasp of what the data are telling them about different 'what if' scenarios of air pollution control strategies." Cray's supercomputer is located at EPA's research facilities in Research Triangle Park and will support environmental research for the agency nationwide. One of EPA's primary uses of its Cray T3E system will be to model harmful atmospheric pollutants, such as carbon dioxide, sulfur dioxide, nitrogen oxide, photochemical oxidants and particulate matter. Common sources of these pollutants include vehicle emissions, coal combustion, fertilizers, ammonia, smog, dust and soot. Foley noted that during the past 20 years, the United States has successfully reduced emissions of harmful pollutants from "big, obvious sources" such as power plants, industrial operations and vehicles. As a result, more U.S. cities meet Clean Air Act standards today than did just 10 years ago. "But it takes more sophisticated computer models to address such issues as smaller, less obvious sources of air pollutants, such as outdoor grills, outboard motors and lawnmowers, to name just a few." Supercomputing is essential to this increasingly difficult mission to preserve clean air because the models run by the EPA on its Cray T3E supercomputer become the tools with which states can identity problem areas and test "what if" control scenarios. These "what if" scenarios - performed by running the air quality models over smaller, local areas - allow state officials to see how alternative control strategies targeting "hot spots" will impact the concentration of air pollutants in their specific area of concern. While states typically run these smaller models on their own, less powerful computer systems, such as PCs and workstations, the accuracy of EPA's large air quality model directly impacts the integrity of the local models, and the effectiveness of the control plans developed from them. Inefficient control plans can result in substantial costs to society - either in continued environmental damage through inadequate reductions, or in adoption of less effective control measures. This makes multimillion-dollar supercomputers, which create more accurate air quality models, a cost-effective tool. In addition to air quality simulation - the Cray T3E supercomputer's primary function - other applications to benefit from this system's increased computing power include integrated air-watershed-estuary models such as those supporting the agency's Chesapeake Bay Project. The goal of this regional model, which is one of the most complete environmental models in the world, is to give the people responsible for making decisions on pollution controls, such as the governors of three Mid-Atlantic states, the ability to quickly understand pollution problems that cross air and water boundaries. "The Cray T3E will allow us to simulate important resources such as clams, oysters and underwater grasses to better understand how nitrogen, phosphorus and sediment interact with the area's living resources above and below the surface. It also allows us to simulate in one day what used to take two weeks," said Lewis Linker, environmental scientist and coordinator of the Chesapeake Bay Project's living resource models. "This 15-fold speed up means affected states can test many more "what if" scenarios in order to develop the most cost-effective, equitable and achievable reduction strategies."
|
|