Identification of NOx and Ozone Episodes and Estimation of Ozone by Statistical Analysis

Frame and daughters directives for evaluating the ambient air quality have been adopted by the EU as a part of the new strategies for pollution prevention and control and environmental management. Therefore, the prediction of ozone concentration and the identification of episodes by modeling are fundamental for protecting and preventing the population and environment against the harmful effects of this species. Under this approach, ambient air quality (immission) data in three zones: A Guarda, Corrubedo and Verin (two coastal and one interior) of Galicia (NW Spain), were collected and evaluated using statistical tools. Punctual and functional background and standard levels of ozone and NOx in the three zones have been determined for detecting abnormal situations and identifying possible emission sources. With this aim, threshold values were established by defining confidence levels. Finally, ozone concentration has been forecasted by time series modeling. Descriptive and predictive models of ozone involving different parameters depending of the area considered have been developed. Satisfactory estimation of ozone concentration was obtained in the three cases with proved efficiency, since predictive values did not exceed the 95% confidence level.