We conducted a spatial extreme value analysis of \(\hbox {NO}_2\) maxima in the Metropolitan Area of the Valley of Mexico. We used a decision tree to estimate the generalized extreme value distribution parameters. In order to validate the model, a quantile-quantile plot was constructed, and a well-fitting relationship was observed between the theoretical quantiles of the generalized extreme distribution (GEV) and the observed data. The spatial trend of each of the three GEV distribution parameters was drawn on spatial maps, in which regions with Weibull, Gumbel and Fréchet distributions of the \(\hbox {NO}_2\) maxima were observed. The return levels from a period of 25 years were mapped in order to observe future trends of extreme \(\hbox {NO}_2\) levels. We observed an increasing trend in extreme levels of \(\hbox {NO}_2\) air pollution in the northwestern part as well as in the north-central part of the Metropolitan Area of the Valley of Mexico. In contrast, the \(\hbox {NO}_2\) levels tend to decrease in the eastern zone and part of the south-central zone of the study area.
