Statistical behavior of ozone in urban environment

This paper analyzes the statistical behavior of the ground level ozone concentrations (GLO) observed at a major traffic intersection in Delhi. Five sets of data, i.e. summer (May to July, high solar radiation data), winter (November to January, low solar radiation data), spring (March to April), autumn (September to October), and the entire year have been used to study the seasonal variation in the statistical behavior of GLO. Appropriate statistical distribution form has been identified from alternative candidate distribution models using the goodness-of-fit methods and parameters have been estimated using the method of maximum likelihood. The yearly, winters, spring, and summer datasets were found to follow the log-normal distribution model, while autumn dataset followed Weibull distribution. Analysis shows that ozone concentrations also show similar statistical behavior like other air pollutants and fit mainly to the log-normal distribution as reported for other pollutants in different studies. The seasonality of the datasets shows higher skewness during summers due to longish tail of the distribution mainly on account of higher photo–chemical activity. The probability density functions corresponding to the five datasets were used to compute the probability of exceedence of the National Ambient Air Quality Standards and return period of violation of standards. The distributions have also been used to classify the study region under various air quality descriptor categories. The region is found to violate the air quality compliance criteria 17% of the recorded times in the year. Alternative measures have been discussed to reduce the precursor emissions in order to achieve the air quality goals.