NPACI Computational Science Olympians Simulate Anthrax and Smallpox Attacks

Published 06/19/2003

Judges recently named the two winners of the 2003 Computational Science Olympics, with the two top student entries from San Diego State University (SDSU) focused on a frighteningly topical subject matter: bioterrorism. One of the two student groups that tied for first place simulated the spread of anthrax through a 100-story building. The other group investigated the effectiveness of smallpox inoculation programs against a bioterror attack on a population of 100,000 people, a small city approximately the size of Bismarck, North Dakota.

The contest, sponsored by the National Partnership for Advanced Computational Infrastructure's (NPACI) Education Center on Computational Science and Engineering (ECCSE) at SDSU, invites California State University undergraduate students to submit scientific research projects that incorporate advanced computing resources and web applications.

"The Computational Science Olympics are really about learning and demonstrating how computational resources can be applied to science," said Jeff Sale, a staff scientist at the ECCSE. "The students work together as a research team, using a variety of sophisticated mathematical and computational techniques to address a scientific problem. This year's entries were compelling both for the quality of the work and the subject matter, which happens to be in the news a lot these days."

"Anthrax Attack on a Building," describes three scenarios for the spread of anthrax in a high-rise building. The model uses a computer science technique called recursion to show how Anthrax spores could travel throughout a high-rise building, depending on factors that include the number of floors in the building, how much and where the bacteria were first injected into the system, the rate of air flow, and the types of air filtration used. The winning team consisted of SDSU students John Manor, Marc Ziegler, Edward Jimenez, and Claudia Valencia.

They found that in a 100-story building, a nearly full test tube of anthrax (15 cc's) could infect just over 40,000 people within 43 minutes. The bacteria were placed after a media filter that removed 99 percent of large airborne particles and fresh air from outside slowly replaced the air inside. In another simulation, the students inserted the same amount of anthrax on one floor of a building equipped with a HEPA filter and a high flow of fresh air from the outside. In contrast to the first scenario, a maximum of 443 people was infected in about 6 hours.

The students concluded that the combination of the use of a HEPA or electrostatic filter and a strong inflow of fresh air was the most effective way to reduce potential infections and to increase time available to evacuate a contaminated building.

"Modeling of a Smallpox Epidemic" simulated the effect of smallpox inoculation programs against a bioterror attack in a population of 100,000 people. The students altered several variables related to the virus' spread including the number of people initially vaccinated, the number of people vaccinated subsequent to the first infection, and the rate the virus is transmitted from person to person. The research team included SDSU undergraduates Janet Garcia, Alecia Brown, Rhiannon Czigan, and Adam Straubinger.

They calculated that for both slow and quick rates of transmission, inoculating 5000 people helped to slightly delay the peak day of infection and to lower the percentage of the population requiring immediate medical care. Vaccinating 40,000 people significantly delayed the peak day of infection, up to a year for the slower transmission rate, and lowered the percentage of people requiring medical care into the single digits on the worst day of the infection. This was pretty much as the students expected.

However, to their surprise they found that the spread of the disease was much more sensitive to another potentially controllable parameter: the number of days that the most infectious group is exposed to the uninfected population. Smallpox shows initially flulike symptoms that abate, leaving the infected individual feeling fine for a period of about four days during which they are most contagious. This period does however have some symptoms: characteristic sores in the mouth. If information could be disseminated cautioning people who are getting over a flu to look for these sores and self quarantine, the spread of the disease drops dramatically even for a one half day shortening of the infectious period contacts.

They concluded, "It appears that although initial vaccination programs are clearly important, should a viral release be imminent (how imminent is speculative), it is the reduction in transmission rate that truly buys the time to react to a threat. Public education is critical for the success of slowing the disease barring imposed quarantines, assuming efficacy thereof on both counts."

The students received their awards in a brief ceremony held at SDSU. Both groups worked on their projects as part of a class taught by SDSU Mathematics Professor Peter Salamon. He said, "Instead of just being another classroom exercise, the Computational Science Olympics gives the students an extra reason to really do their best work."

This year marks the fourth year for the Computational Science Olympics. Previous winners have included: "Simulation of The Contractile Behavior of An Isolated Cardiac Myocyte" (2000); "Spinner the Ride" and "The Nonlinear Pendulum" (2001); and "Separating Ions via Brownian Motion and Electric Fields," and "The Juan Parrondo Experience" (2002).

More about the Computational Science Olympics can be found at the ECCSE web site, http://www.edcenter.sdsu.edu/cso/index.html. For more information about this year's winners, visit http://www.edcenter.sdsu.edu/cso/cso2003.html. The next Computational Science Olympics will be held during the spring semester of 2004. - Cassie Ferguson