A Los Alamos National Laboratory team has developed a computer program that can mitigate the dangers of hazardous chemical airborne releases by predicting where the substances will go and how fast they will get there.

Previous to this work, computer codes for emergency response were diagnostic, not prognostic, said Ted Yamada, a physicist in the Lab's Geoanalysis Group, who handles the computer modeling for the emergency response project.The diagnostic models could analyze the atmosphere at the time of an accident, but if contaminants lasted for more than a few hours the models were hampered because they couldn't predict changes in the winds, he said.

Our computer model now can predict, for the first 12 hours after the release of a gas, what will happen within about 60 miles of the site, Mr. Yamada said.

There's an obvious need for local emergency response officials to know what the chemicals will do, so appropriate action, such as evacuation, can be taken. That's why our program is important, Mr. Yamada said.

Greg Stone, a meteorologist in the Lab's Atmospheric Sciences Group, who does the projects' field experiments, said, We are taking sophisticated hydrodynamics theory and putting it to practical use.

Los Alamos is working on one particularly interesting problem for the U.S. Army, which has been directed by Congress to dispose of all existing chemical munitions by 1994.

The Los Alamos work involves supporting the Army in assuring public safety during the handling and disposal of the munitions.

Field experiment for the Los Alamos study were conducted last summer at one of the Army's storage depots, 40 miles from Salt Lake City, near Tooele, Utah. The experiments, conducted in cooperation with the National Oceanic & Atmospheric Administration, tested the Los Alamos computer model's assumptions - and gathered more information to add to the model.

Experiments included releasing harmless gas to track how it dispersed in a variety of wind, temperature and terrain conditions. Meteorological instruments, both on the ground and up to six miles high attached to balloons, collected data for a month from a 3,000-square-mile area.

The information was then taken back to Los Alamos and fed into one of the Lab's Cray supercomputers, capable of doing more than 40 million calculations a second.

One of the keys to the success of the Los Alamos program has been the ability to take work that previously required a supercomputer and make it available on small personal computers for use at an accident site.

The advancement of the microcomputer has made our job easier, Mr. Yamada said. We have attained somewhat of a computational milestone by being able to transfer the capability of our model from a supercomputer to a personal computer.

Los Alamos researchers want to broaden their model's capabilities.

We hope to do further studies to see how our model works under other conditions, he said.