Air Quality in Mexico-City: Emissions, Transport, and Chemical Transformation
- 1University of Houston, Department of Earth and Atmospheric Sciences, Houston, United States of America (brappenglueck@uh.edu)
- 2Institute of Climate and Atmospheric Science, Houston, USA
- 3FM Global Insurance, Boston, USA
- 4Independent Researcher, Ciudad de México, Mexico
- 5Secretaría del Medio Ambiente, Ciudad de México, Mexico
During the period March 12-17, 2016, Mexico-City experienced its most severe smog episode since 2007. The Metropolitan Index of Air Quality (IMECA) for Mexico-City surpassed the value of 200, indicating an extremely bad situation. Hourly peak values for both, NO2 and O3, exceeded 200 ppb, while for CO more than 2 ppm were observed. Restrictions on traffic and industrial activities, among other emergency measures, were imposed. We describe results from Positive Matrix Factorization (PMF) for source apportionment based on a commixture of gasphase compounds (VOCs, CO, NO, NO2, SO2, NH3) along with equivalent black carbon (eBC), and ions (Na+, Mg2+, Ca2+, NO3-, NH4+) in combination with an analysis of regional meteorological processes and boundary layer conditions retrieved from continuous microwave radiometer measurements. Apart from more traditional emission sources, the PMF analysis also deciphered a geogenic source. Continuous boundary layer height data was used to normalize mixing ratios of pollutants representative for each source factor. This procedure allowed the retrieval of diurnal variations of pollutants predominantly determined by emissions and removal mechanisms. The results show prolonged daytime emissions of O3 precursors beyond the typical morning rush hour, an important information to optimize O3 mitigation strategies. Propylene Equivalent and Maximum Incremental Reactivity (MIR) methods identified isoprene and ethylene as the highest oxidant and O3 forming species which indicates some interchange of individual top VOC contributors to ozone formation in that city over the last decades. This presentation concludes with results from air quality modeling including Machine Learning approaches. While the Deep Neural Network, Random Forest and Gradient Tree Boosting models are depicting diurnal O3 levels nicely, as long as O3 mixing ratios are at moderate levels (≤120 ppb) only the Deep Neural Network may capture peak ozone values (>160 ppb), which are most critical with regard to public health.
How to cite: Rappenglück, B., Akther, T., Ahmad, M., Alam, J., Osibanjo, O., Retama, A., and Rivera-Hernández, O.: Air Quality in Mexico-City: Emissions, Transport, and Chemical Transformation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14092, https://doi.org/10.5194/egusphere-egu24-14092, 2024.