Conditions and trends of air pollution levels in the metropolitan Zone in the city of Toluca

摘要

The city of Toluca is located 60 km. to the west of Mexico City, and due to its accelerated demographic and industrial growth, one of the main concerns of the State Environmental Department is the potential increment in air pollution indexes. Records about important pollutants have been collected every hour since April 1993 for seven locations in the zone. These data have been analyzed statistically looking for trend changes and potential differences among locations. Due to the great amount of missing data, only two locations provided enough weekly averages to allow for statistical analysis. The daily data were deseasonalized and weekly averages used to estimate missing data via ARIMA models. To construct forecasts and to detect trend changes, ARIMA models were adjusted to the available series. To identify the corresponding models bootstrap samples from the actual series were obtained and the ACF and PACF functions constructed from these samples. The estimation phase was performed by using actual data and at the validation stage, bootstrap principles were also applied, this time the bootstrap samples were generated from the series residuals. Forecasts for the last 10 weeks of 1994 were computed, and the forecasting ability of the models finally tested by computing performance measures from the actual data corresponding to these weeks. Discussion of main findings and the validated models for each series are presented.