High-tech bay study targets pollution Computers being used to pinpoint sources in Chesapeake region

June 03, 1996|By John M. Biers | John M. Biers,STATES NEWS SERVICE

WASHINGTON -- In a study that could have worldwide applications, scientists are using supercomputers to identify the sources of air and water pollution in the Chesapeake Bay region and to test cleanup options.

The technology has confirmed what scientists have long suspected: Much of the bay's pollution comes from an airshed that extends as far west as Ohio and Kentucky and as far north as Ontario, Canada.

But as far as they have come, scientists estimate they are at least a few years away from being able to use the technology to answer the question most often asked by policy-makers and the public: "Who is responsible?"

The modeling project, now in its fourth year, represents a high-tech marriage of convenience between the National Exposure Research Laboratory, an Environmental Protection Agency facility near Durham, N.C., and the Chesapeake Bay Program in Annapolis. The bay program comprises representatives from the EPA and jurisdictions surrounding the Chesapeake.

The collaboration resulted from scientists' increased awareness of atmospheric causes of the bay's environmental woes. Computer runs show that the bay's airshed -- a region with a common flow of air -- is about 5.5 times the size of the bay's watershed, which encompasses the land drained by the bay's )) tributaries.

Using computer modeling techniques, researchers plug in data about air pollutants to simulate air flow and how it affects dispersal of pollutants such as nitrogen.

The data indicate that 25 percent to 35 percent of the bay's nitrogen comes from the air rather than from ground-based runoff. Of the airborne nitrogen, some comes from power plants and some from vehicles and other mobile sources, although the ratio is not known. Most utility emissions are from the Midwest, while the auto pollution originates mostly along the East Coast.

Nitrogen is a nutrient that harms the bay by stimulating excessive growth of nuisance algae, which block sunlight and kill underwater grasses vital to marine life. Once algae sink to the bottom, they decay -- a process that uses oxygen. Without oxygen, fish and other marine organisms die, resulting in dead water.

Researchers say plugging in data about air pollutants for analysis by supercomputers complements the traditional scientific tools for studying the Chesapeake: observation and experimentation.

"Experiments run on a small scale, and, for observations, things need to have already happened. But simulations allow us to play a 'what if?' game. We're able to understand systems much better," says Lewis Linker, who heads the bay program's computer modeling division.

The technology can be used to show what would happen if emissions were limited, according to Melissa McCullough, an EPA scientist. In a number of unpublished studies, computer runs have shown that full implementation of the Clean Air Act and other emissions controls would reduce the bay's dead water.

The technology could affect numerous policy decisions, such as whether to regulate Midwestern smokestacks and other pollution sources outside the Eastern seaboard.

Already, Midwestern environmental officials and industry groups have grumbled about the possibility of increased federal oversight because of the simulations.

Bay program officials say they need time to improve the technology and give industry and policy-makers time to familiarize themselves with the results. Any regulatory adjustments, they predict, are at least 10 years away.

During computer runs, information traverses a complex path that touches four states. Using a high-speed transmission line roughly 10,000 times the speed of the average computer modem, the test makes six journeys through laboratories in Annapolis, Durham, the Waterways Experiment Station of the Army Corps of Engineers in Vicksburg, Miss., and the EPA's supercomputer center in Bay City, Mich.

The technology's most obvious limitation is the length of computer runs, which take about 360 hours, or 15 days. Much of that involves the highly complex atmospheric model, which requires 300 hours.

The upshot is that researchers are not able to test as many scenarios as they would like. McCullough identifies an "unlimited" number of areas for investigation, such as whether -- the EPA should order early phase-out of old cars or regulations on utilities.

According to Robin Dennis, senior program manager at the Durham lab, the current technology also thrusts too much guesswork on researchers, largely because of the intense challenge of simulating the airshed -- a vast territory of chemical complexity.

For example, Dennis would like to evaluate the impact of RTC unusually wet or dry seasons on the airshed. Now he is forced to use data representing three "average" years because the computers cannot process the more comprehensive data.

Researchers also would like to be able to include information about chemicals in the air other than nitrogen.

Dennis estimates that many of these advances are a few years away. Rather than the single mechanism in use now, more than 100 computers would be linked in a giant system. "Right now we're working in gigabytes; we're talking tens of gigabytes," he says.

Further off is the "electronic aquarium," which could simulate not just chemicals but fish and sea grasses as well. Dennis envisions supercomputer applications that would allow scientists to enter information about pollution counts and see the real-world consequences on the screen.

"What you're trying to do is have end points that people understand. People don't understand dissolved oxygen. To say that doesn't mean as much as having twice as many shad or saying that the crab population will come back to normal," Dennis says. " 'I can go fishing' -- that's what people what to know."

Pub Date: 6/03/96

Baltimore Sun Articles
|
|
|
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.