Lindley Graham: Inverse Sensitivity Analysis for Storm Surge Models
A single hurricane can result in many tragic deaths and cause billions of dollars in damage. Flooding from hurricane storm surge is the primary source of this destruction. Storm surge is dominated by the balance between wind forcing and bottom friction. Thus it is important to characterize uncertainty in Manning's n (a bottom friction parameter) for forecasting storm surge to improve forecasting of storm surge. We describe a framework for quantifying uncertainty in this spatially varying parameter for the ADCIRC storm surge model.
The inverse sensitivity analysis algorithm we use requires the sampling of a high dimensional probability space (~30 dimensions). To resolve this space we must run the storm surge model on a sufficiently fine mesh at each sample point. High performance computing systems provide the necessary computational power to accomplish this. Furthermore, this inverse sensitivity analysis algorithm is embarrassingly parallel making it well suited to take advantage of HPC systems.
More about this speaker:
Originally hailing from California, Lindley Graham is a second year Computational Science, Engineering, and Mathematics PhD Student at the University of Texas at Austin. Recently, he was selected to receive a 2011 National Science Foundation Graduate Research Fellowship. Working under ICES Professor Clint Dawson, Graham's research will seek to improve modeling processes in the coastal ocean. His work focuses on quantifying uncertainty in the shallow water equations in order to increase the ability to provide timely and relevant data to emergency responders. To achieve this, he will investigate polynomial chaos expansions and Karhunen-Loève expansions within the context of solving the shallow water equations with uncertain inputs such as bathymetry, salinity, bottom friction, and eddy viscosity. He is also interested in inverse problems. Graham earned his bachelor's degree in aerospace engineering at Massachusetts Institute of Technology before arriving at UT Austin last year. In his free time he enjoys fencing with the University of Texas at Austin Fencing Club.