Characterizing changes in extreme ozone levels under 2050s climate conditions: An extreme-value analysis in California

Abstract

Ground-level ozone has a well established climate penalty associated with changing meteorological conditions under future climate change, but most existing research focuses exclusively on ozone-temperature relationships defined at specific locations. Using an extreme-value theory approach, we model relationships between temperature, humidity, and vapor pressure deficit and maximum daily 8 h ozone concentrations (MDA8) across California using a combination of ozone station network data and Community Multiscale Air Quality Modeling System (CMAQ) model output from 2008 to 2017. We use a spatial regression with time varying bias coefficients to fuse station observations and modeled output into a spatially explicit gridded ozone dataset at 12 km resolution, then fit independent Point Process models to every grid cell to estimate 5, 20, and 50 year ozone return-levels across California. We evaluate the impact of climate change on ozone return levels and episode days under RCP4.5 conditions using an ensemble of 18 downscaled CMIP5 climate models, finding changes of -6t to 8 ppb in effective return levels and −8 to 16 episode days, Our results further strengthen the evidence for anozone climate penalty in California and show the advantages of considering multiple climate variables when modeling ground-level ozone. • Extreme ozone levels are affected by multiple meteorological variables, including temperature, relative humidity, and vapor pressure deficit. • Climate change increases extreme ozone levels across most of California. • Extreme value theory model fits are improved with meteorological covariates.