Seminar- Investigating the causes of extreme ozone levels

02/01/2013 - 12:00pm
Mesa Laboratory, Directors Conference Room

Brook Russell, CSU Fort Collins

It is well known that air temperature is associated with ground level ozone. However, air temperature alone does not completely explain extreme levels of the air pollutant. The goal of this research is to identify the combination of meteorological covariates that is most strongly associated with extreme ozone levels.  The method we propose relies on bivariate extreme value theory. We also present a metric to measure the tail dependence between two variables. Two variables with strong tail dependence tend to have extremely large values occur together.  We suggest an approach that uses data mining to find the combination of meteorological covariates that has the strongest tail dependence with ozone. A simulation study will be presented in addition to a discussion of preliminary results from an Atlanta, Georgia data set.

Date: Friday, February 1, 2013 - 12:00pm
Location:  Mesa Laboratory, Directors Conference Room