CHEMISTRY

ITHACA, NY --With a death toll of over 2000, the tsunami that struck Papua New Guinea on July 17, 1998 was one of the most devastating on record. Many of the victims drowned in the Sissano Lagoon behind the seaside Arop villages. While researchers were able to show the public a general picture of the event shortly after it occurred, time and powerful computing resources were needed to explain the unique power of this devastating event. Cornell civil and environmental engineers in the lab of Professor Philip Liu used the Dell/Intel/Windows Velocity high-performance cluster at the Cornell Theory Center (CTC) to show that the collapse of the sea floor caused by the associated earthquake intensified the wave's destructive power. Papua New Guinea is a seismically active region, where many tectonic plates converge and slide past each other. The earthquake that caused the giant 1998 wave shocked the Sandaun province of northwestern New Guinea, about 65 km northwest of the port city of Aitape. Its magnitude was estimated at 7.0 on the Richter scale. Approximately 20 minutes after the first shock, Warapo and Arop villages were totally destroyed by tsunamis, which produced waves of 7 to 10 meters in height at the coast. Working with Liu, graduate student Patrick Lynett, simulated the track and size of the wave and its runup, based on details provided by an international team of tsunami researchers. This information included the bathymetry (undersea landscape) of the region, in addition to the forces generated by the shaking and dynamics of the quake and records of the runup gathered in New Guinea. The greatest wave height reported was more than 30 feet and can be explained by the unusual nature of the event. A huge underwater landslide, of a magnitude rarely seen, occurred near the shore. The ground movement disturbed the ocean above it and created the devastating tsunami. Lynett's simulation captures this complexity and shows the path of the wave as it surges through the lagoon where it destroyed the coastal villages. Shortly after the event, Liu said, "In order to design the coastal structure properly and to reduce the damages which might be caused by a future tsunami, it is essential to be able to estimate the impact force acting on the structure." His team's simulations serve this purpose by going to great lengths to recreate the event in detail. Such simulations make an important contribution to risk assessment for regional planners and contribute to our basic understanding of the behavior of these devastating waves, no longer thought to be limited to remote regions of the Pacific. Planners like Liu recommend construction of sea walls and implementation of zoning policies that ban building in high-risk areas as ways to minimize the destruction caused by such natural disasters. However, Liu is quoted by CNN to say, "It is questionable that in a country like Papua New Guinea such policies could be implemented." In a country with limited resources, a flat coastal landscape, and sub-standard construction, planners have to make realistic recommendations. For example, instead of a storm shelter, families are encouraged to pick a sturdy Casuarina tree for quick climbing and to fit their tree out as a temporary refuge. Tsunami destruction is caused both by the water and the debris that is picked up by the wave and hurled at buildings. Thus Tsunami experts recently recommended such straightforward, but important, plans as preventing relocation to vulnerable backwater regions, placing simple memorials to remind inhabitants of the danger and discourage resettlement of high risk locales, and planting sturdy trees instead of fragile palms along the coast. CTC's Windows-based Velocity cluster consists of 64 Dell PowerEdge 6250 Quad Servers with Intel Pentium III Xeon 500 Mhz processors. CTC is a high-performance computing and interdisciplinary computational research center located at Cornell University. CTC receives funding from Cornell University, New York State, a number of federal agencies, and Corporate Program members. For further information visit www.tc.cornell.edu