Studying the impacts of wildfire emissions on Ozone in the Las Vegas Valley using cluster analysis and a statistical model

In recent years, the observed ozone concentrations have approached, and occasionally exceeded, the federal ozone standard in the Las Vegas Valley (LV Valley), located approximately 270 miles downwind from Los Angles, California at the southern tip of Clark County, Nevada. In addition to the contributions from local and regional transport, such ozone exceedances may have been associated with wildfire emissions in the western United States where wildfire activities have increased in recent decades. In this study, the U.S. EPA's “Omnibus Meteorological Data Set” (OMD) and daily peak 8-hour O3 of local and upwind areas of LV Valley for five summer months during 2004-2008 were used to develop a statistical model to identify wildfire events and study their relationships with high ozone episodes. Using the cluster analysis of the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model, the back-trajectories in the LV Valley were calculated for seven clusters of meteorological conditions and observed peak O3 mixing ratios. A statistical model was then developed for each cluster based on the cluster analysis using polynomial regression equations. Compared with the model simulation without cluster analysis and without accounting for wildfire emissions, the statistical model with cluster analysis significantly improved the coefficient of determination (R2) from 0.59 to 0.71-0.87 and reduced the standard errors from 6.16 to 3.33-4.87 ppb. The predicted peak O3 mixing ratios for suspected wildfire days were calculated with the statistical model based on their assigned clusters to examine the impact of wildfire events. Several wildfire events were identified. Based on the differences between observations and predictions with the statistical models without accounting for wildfire emissions, the estimated enhancement in the peak O3 mixing ratios during wildfires range from 0.49 ± 4.87 to 24.94 ± 4.87 ppb. For example, the observed O3 mixing ratios are 100 and 105 ppb on June 29 and 30, 2005, respectively. The predicted O3 from the statistical model are 76.69 ± 3.33 and 89.82 ± 3.33 ppb, respectively. These results provide evidence on the impacts of wildfires on O3 that warrant further examination.