Application of discriminant analysis to air quality data of the Metropolitan Zone of Guadalajara

Abstract

The aim of this paper was to classify the eight monitoring stations of Guadalajara´s atmospheric network, by means of discriminant analysis. Measured variables were hourly means of criterion pollutants (PM10, O3, CO, NO2, and SO2), as well as meteorological variables (temperature, relative humidity, and wind) for the period 1994-2005. The analysis was performed using the software STATGRAPHICS 5.1. Results suggested three discriminant functions to explain 82.36 % of total variability in the data. The first discriminant function explains 51.19 % of variability and distinguishes PM10 as the variable that contributes the most to define this function. The second discriminant fuction explains 20.14 % of variability; within this function temperature and humidity as meteorological variables, as well as CO and NO2 as criterion pollutants contribute the most to define this function. Finally, the third discriminant function explains 11.03 % of variability, and wind speed contributes the most to define it. The general classification table, as the result of the evaluation of the three discriminant functions, suggested an overall 27.47 % of observations correctly classified. Within this percentage, the stations best classified were Center, Miravalle and Atemajac with a 32.70, 32.16 % and 31.61% respectively. Such results do not agree with local environmental reports at which Miravalle highlights as the station with the poorest air quality and consequently the highest health impacts. Up on such findings, the authorities have set air pollution management policies for this particular area, thus failing to approach the problem in global manner. According to the results of this study, it is worth reconsidering such policies towards integrated management policies that involves the metropolitan area as a whole. With the data used and the results of the discriminant analysis it is not possible to state that the worst air pollution scenario and health impacts only occur at the southwest part of the city were Miravalle is located. Even thought the overall classification of the stations was 27.47%, the stations best classified were Center, Miravalle and Atemajac. It is well identified that within this places economic and a brought range of services concentrate here, as well as industrial activities, and the valley region of the city. These stations best classification puts forward the strong relationship between meteorological-climatic conditions and pollution patterns within urban areas.