terrorist detonates a dirty bomb in Chicago, releasing toxic chemicals, radiation and radioactive elements into the atmosphere. When will the toxins reach Milwaukee and points north? In what concentrations? If they don’t move toward Milwaukee, where will they go and how will they disperse into the atmosphere? With no similar past experiences in the Great Lakes to learn from, atmospheric scientists at UWM are trying to answer these questions by simulating such events with mathematical models. Collecting current conditions— such as wind speed and direction, atmospheric pressure, and temperature—from numerous monitoring sites throughout the Great Lakes basin, computers are assimilating the data to help scientists predict how airborne hazardous materials from various possible introduction points will be concentrated and dispersed. Other modeling projects are simulating movement of toxins introduced into Lake Michigan. The projects, now based at the UWM Great Lakes WATER Institute’s Center for Water Security, began as informal discussions in the summer of 2001 between atmospheric scientist Paul Roebber and the WATER Institute about a possible collaboration. “We came up with the idea that water security would be an idea that people would be interested in because Lake Michigan is the drinking supply for a major part of the population,” he said. “Then September 11th happened.”   From www.uwm.edu/~kahl/Air This sequence of maps projects the concentration of a simulated pollutant release from Minneapolis at 6:00 a.m. on November 14, 2003 throughout a 48-hour period (Click on picture for larger image). Roebber and his fellow professors from the atmospheric sciences program in UWM’s math department soon found themselves working with greater urgency. Roebber’s computer system forecasts the weather over and near Lake Michigan up to two-and-a-half days ahead (with updates four times a day) and has a resolution of four miles. It surpasses any other forecasting configuration in the country for this job, he says. Jonathan Kahl’s model predicts where airborne pollutants will travel though the atmosphere and land on the Earth’s surface. Using current weather data, the model simulates four releases per day from 50 locations within the Great Lakes basin—from Minneapolis in the west to New York City in the east; from Thunder Bay, Ontario in the north to Cincinnati in the south. Projected airborne concentrations, as well as pollutant deposition to the Earth’s surface, are graphically depicted on Kahl’s Web site (www.uwm.edu/~kahl/Air) Meanwhile, Kyle Swanson and Anastasios Tsonis have modified a widely used ocean model to simulate Lake Michigan’s circulation. Their adaptation of the Princeton Ocean Model operates at a resolution of 4 kilometers horizontally and 100 meters vertically (Lake Michigan averages approximately 85 meters deep, with a maximum depth of 292 meters). They have already used it to model the spread of chemical or biological contaminants within the lake and to predict the paths that pollutants could take to reach vital locations such as the Chicago city water intake. From www.uwm.edu/Dept/GLWI/cws/projects/swanson.html Left: Lake Michigan circulation as simulated by the Princeton Ocean Model for a recent July. Note the strong near-shore currents. Right: Evolution of an initial contaminant field (circle) over a two-month period. Note the complex distribution due to current structure. Such modeling could not only predict the path of a hazardous material from a known source, but also work backwards to help pinpoint the origin of an unknown release. It took a year for Swanson and Roebber to develop their respective models. Although both models are providing interesting results, to be truly effective the two have to talk to each other. “If you want to know where the water is going in the lake,” Roebber says, “you have to look at the air [above the lake], because the air is driving the water.” But if you want to know what’s happening with the air above the lake, you have to look at the water. The lake’s surface temperature can drop by as much as 10 degrees centigrade in a mere 12 hours, completely altering the water’s effect on the atmosphere above it. The lake also creates complex patterns in local winds. “We don’t have that in the modeling we’re doing right now,” Roebber says. Coupling the two models, each with roughly 10,000 lines of computer code, will be a challenge, Swanson says. “When you marry them they’re not necessarily compatible.” After a year of joint work, Swanson estimates that the project is still a year or two from completion. When it’s done, the researchers expect the UWM model to be the most detailed air-water model ever developed.