 |
Inside this Issue |
||||
| Next Generation Tsunami Modeling | |||||
Dr. Zygmunt Kowalik,
Institute of Marine Science; |
|||||
Story by Leone Thierman
Because Alaska coastal communities are located in one of the most active seismic regions of the world, Alaska has the greatest earthquake and tsunami potential in the entire United States. To ensure reliable tsunami early detection and hazard assessment capabilities, it is essential to create a numerical model to forecast future tsunami impact and flooding limits in specific coastal areas. Tsunamis are a series of long waves often generated by large-scale earthquakes or landslides that occur along active faults near continental margins or subduction zones. Although a tsunami’s wavelength from crest to crest can extend more than hundreds of kilometers, it’s height in deep ocean is commonly less than a meter from crest to trough. A tsunami is considered a shallow-water wave, since it is noticeably affected by bottom topography, especially when the wavelength is at least 20 times the depth of the water beneath it. The speed of a shallow-water wave is determined by the ratio of the water’s depth to the wavelength of the wave – its celerity – C=√Depth x Gravity. In addition, the rate at which a wave loses its energy is inversely related to its wavelength. Since a tsunami has a very large wavelength, it will lose little energy as it propagates. Therefore, in deep water, a tsunami will travel at high speeds and over great transoceanic distances with limited energy loss. For example, when the ocean is 5 km deep, tsunamis can travel about 220 m/s (~500 miles per hour), nearly the speed of a jet airplane, and they can move from one side of the Pacific Ocean to the other side in less than one day.
As a tsunami leaves the deep water of the open sea and moves into the shallower water of the coastal areas in its path, it undergoes a transformation. As its speed decreases, the amplitude increases, and the wavelength shortens. The wave may appear as a rapidly rising or falling tide that rushes up onto the beach without breaking, or as a turbulent wall of water with higher water behind it, which is called a bore. Tsunamis may reach a maximum onshore vertical height above sea level, called a run-up height, of 30 meters (100 feet). The wave causes devastation as it moves inland, smashing ocean-floating objects into structures onshore and carrying off loose objects and people in its path as it retreats. Earthquakes generate most tsunamis, particularly those that propagate Pacific-wide. Landslides cause some tsunamis, such as the 1958 Lituya Bay, Alaska earthquake-induced rock slide, which generated a 525-meter splash-up immediately across the bay and leveled trees across LaChausse Spit before leaving the bay. When sediment and rock slump down-slope and are redistributed across the sea floor, a volume of fluid flows out and up from the sea bottom and propagates from there. A Different Perspective
A team led by Dr. Zygmunt Kowalik and his research assistant Juan J. Horrillo, both with the Institute of Marine Science (IMS); Ed Kornkven, of the Arctic Region Supercomputing Center (ARSC); and William Knight, of the West Coast and Alaska Tsunami Warning Center are working together to develop and refine a new code that pairs current tsunami modeling with the volume of fluid (VOF) method developed at Los Alamos National Laboratory. The VOF method solves the fully nonlinear Navier-Stokes equation adding three basic ingredients to solve the free surface configuration: a scheme to locate the surface, an algorithm to track the surface as a sharp interface moving through a computational grid, and a means of applying boundary conditions at the surface. Because the vertical acceleration associated with tsunami shallow-water open-ocean waves is small compared with the gravity acceleration, tsunamis are still resolved using 2D hydrostatic models. These models take into account only horizontal velocities and calculate the vertical sea level changes through the mass conservation equation. The goal is to use VOF to incorporate the vertical motion component, especially in the generation and run-up, and to couple its results with the currently used 2D shallow water models. At present, high-resolution VOF 2D models are used to calculate only tsunami generation and run-up. The IMS’s tsunami numerical modeling is focused on simulation and prediction of wave height, run-up, inundation area and wave-structure interactions. The modeling also involves tsunami generation, which encompasses sub-marine and sub-aerial landslides, sea bottom uplift and rock impacting the water surface. “This new code being developed here at UAF offers us another tool and gives us another perspective to validate other models,” says Horrillo. “With this tool we can determine the forces of tsunamis and design coastal structures accordingly.” Pushing the Computational EnvelopeTaking VOF to a whole new level in tsunami modeling requires extensive CPU time and high resolution, so at this point it is mostly suitable for small domains with complex topography, like bays and channels, especially in the subduction zones of the Aleutian Islands. Kornkven has been assisting the IMS team with ways to improve this code using tools from the Cray X1 to highlight areas that consume CPU time. Kornkven massages the code to take advantage of the unique vector and multi-streaming capabilities of the Cray X1. The idea is to change the form of the equations and order of calculations so that the code works more efficiently and produces the same results.
“Zygmunt and Juan really have a vision for the magnitude of the problem they want to solve,” says Kornkven. “Given the code is amenable to a vector mode, it can take advantage of the unique features that give the Cray X1 it’s performance capabilities.” “So far, we have made some progress using VOF in 2D. In the near future we expect to jump into 3D,” says Horrillo. |
|||||
| Page 1 | Challenges Index | |||||
