2015/04/30 by Brook T. Russell, Russell, Brook T., Daniel Cooley +7
Environmental Science · #Air Quality Monitoring and Forecasting #Air Quality and Health Impacts
paper · pdf · doi:10.48550/arxiv.1504.08080
This project aims to explore which combinations of meteorological conditions\nare associated with extreme ground level ozone conditions. Our approach focuses\nonly on the tail by optimizing the tail dependence between the ozone response\nand functions of meteorological covariates. Since there is a long list of\npossible meteorological covariates, the space of possible models cannot be\nexplored completely. Consequently, we perform data mining within the model\nselection context, employing an automated model search procedure. Our study is\nunique among extremes applications as optimizing tail dependence has not\npreviously been attempted, and it presents new challenges, such as requiring a\nsmooth threshold. We present a simulation study which shows that the method can\ndetect complicated conditions leading to extreme responses and resists\noverfitting. We apply the method to ozone data for Atlanta and Charlotte and\nfind similar meteorological drivers for these two Southeastern US cities. We\nidentify several covariates which help to differentiate the meteorological\nconditions which lead to extreme ozone levels from those which lead to merely\nhigh levels.\n