An analysis of the long-term trend in ozone data from two Houston, Texas monitoring sites

Abstract This paper uses the theory of extreme values to estimate the long-term trend in the probability that ozone measurements made at two monitoring sites in Houston, TX exceed a threshold level specified by the U.S. Environmental Protection Agency (EPA). In addition, a bootstrap testing procedure is used to test if the trend is statistically significant. The analysis takes into account the correlation in the series of daily ozone readings, seasonality in the data, and missing values in the ozone series. The empirical results indicate a downward trend in the probability of exceeding the EPA's specified threshold level at both sites. However, the magnitude of the downward trend is much greater at one site than at the other, even though both sites are in Houston. In fact, hypothesis testing results suggest that only one site has a statistically significant trend in the long-term probability of an exceedance.