Richard L. Smith is the Mark L. Reed III Distinguished Professor of statistics at the Gillings School of Global Health at UNC, and holds a joint position as professor of biostatistics. He obtained his doctoral degree from Cornell University in 1979 and has previously held academic positions at Imperial College (London), the University of Surrey (Guildford, England) and Cambridge University.
His principal areas of research are spatial statistics, time series analysis, extreme value theory and Bayesian statistics. Specific areas of expertise include spatial and time series modeling of environmental pollutants, the health effects of atmospheric pollution, the statistics of global climate change, and extreme values in insurance and finance.
Talk: Extreme Value Theory in Climate Science
Abstract: In recent years there have been a number of extreme weather events of unprecedented severity. Examples include the heatwave that hit Northwest North America in June 2021, the heatwave that affected southern England and parts of Europe in July 2021, and the very severe rainfall in Houston (Hurricane Harvey) in 2017. These events appear to be part of a pattern of extreme weather events exacerbated by global warming. Extreme value theory plays a role in quantitatively assessing the severity of the event, deciding to what extent the event can be attributed to climate change, and assessing the probability of future extreme events under various climate change scenarios. In this talk, I will discuss approaches to these problems using spatial extreme value theory and Bayesian statistics. The results confirm that all three events were very extreme events even taking climate change into account, and they also show the contrast in the likely future distribution of extremes under various climate change scenarios. I will also discuss some recent research (joint with Shaleni Kovach, graduate student) on the health effects of climate change.