An application of a non-homogeneous Poisson model to study PM_2.5 exceedances in Mexico City and Bogota

Abstract

It is very important to study the occurrence of high levels of particulate matter due to the potential harm to people's health and to the environment. In the present work we use a non-homogeneous Poisson model to analyse the rate of exceedances of particulate matter with diameter smaller that 2.5 microns (PM${}_{2.5}$). Models with and without change-points are considered and they are applied to data from Bogota, Colombia, and Mexico City, Mexico. Results show that whereas in Bogota larger particles pose a more serious problem, in Mexico City, even though nowadays levels are more controlled, in the recent past PM${}_{2.5}$ were the ones causing serious problems.

Keywords:

• Bayesian inference
• Markov chain Monte Carlo algorithms
• air pollution
• change-points
• Weibull rate function

Acknowledgments

The authors would like to thank the anonymous reviewers as well as the Associate Editor for the careful reading of this work and their helpful comments and suggestions that allowed to considerably improve the presentation of the results. The authors also thank the Secretaría de Medio Ambiente of Bogota, Colombia for making the pollution data available for this study. This work is part of BMSS PhD Thesis and she thanks the Centro de Estudios Interdisciplinarios Básicos y Aplicados of the Fundación CEIBA, Colombia, for the PhD Scholarship. She also thanks the Instituto de Matemáticas of the Universidad Nacional Autónoma de México (UNAM), Mexico, for hospitality and for allowing the use of their facilities. BMSS and ERR were partially funded by the project PAPIIT-IN102416 of the Dirección General de Apoyo al Personal Académico-UNAM (DGAPA-UNAM). Part of this work was developed while ERR was in a sabbatical visit at the Department of Statistics of the University of Oxford, United Kingdom, with a grant from DGAPA-UNAM (PASPA - Sept./2018 - Feb./2019). She thanks the department for their support and hospitality.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

BMSS and ERR were partially funded by the project PAPIIT-IN102416 of the Dirección General de Apoyo al Personal Académico-UNAM (DGAPA-UNAM). Part of this work was developed while ERR was in a sabbatical visit at the Department of Statistics of the University of Oxford, United Kingdom, with a grant from DGAPA-UNAM (PASPA - Sept./2018 - Feb./2019).