Superstorm Sandy is merely the most recent high-impact weather event to raise concerns about extreme weather events becoming more frequent or more severe. Previous examples include the western European heatwave of 2003, the Russian heatwave and the Pakistan floods of 2010, and the Texas heatwave of 2011. However, it remains an open question to what extent such events may be “attributed” to human influences such as increasing greenhouse gases. One way to answer this question is to run climate models under two scenarios, one including all the anthropogenic forcing factors (in particular, greenhouse gases) while the other is run only including the natural forcings (e.g. solar fluctuations) or control runs with no forcings at all. Based on the climate model runs, probabilities of the extreme event of interest may be computed under both scenarios, followed by the risk ratio or the “fraction of attributable risk”, which has become popular in the climatology community as a measure of the human influence on extreme events. This talk will discuss statistical approaches to these quantities, including the use of extreme value theory as a method of quantifying the risk of extreme events, and Bayesian hierarchical models for combining the results of different climate models. This is joint work with Xuan Li (UNC) and Michael Wehner (Lawrence Berkeley Lab). Event flyer.
Dr. Eunshin Byon is an Assistant Professor in the Department of Industrial and Operations Engineering at the University of Michigan, Ann Arbor, USA. She received her Ph.D. degree in Industrial and Systems Engineering from the Texas A&M University, College Station, USA, and joined the University of Michigan in 2011.
Importance sampling has been used to improve the efficiency of simulations where the simulation output is uniquely determined, given a fixed input. We extend the theory of importance sampling to estimate a system’s reliability with stochastic simulations. Thanks to the advance of computing power, stochastic computer models are employed in many applications to represent a complex system behavior. More
To quantify and minimize the uncertainties in the design and operational stage, we model and analyze the dependency of wind turbine responses (e.g., power generation, loads and condition monitoring sensor measurement) on operating conditions and the interactions among turbines. Our research entails several areas… More