Meteorology-driven variability of air pollution (PM1) revealed with explainable machine learning

doi:10.5194/acp-2020-469

GSTDTAP > 地球科学

DOI	10.5194/acp-2020-469
	Meteorology-driven variability of air pollution (PM1) revealed with explainable machine learning
	Roland Stirnberg, Jan Cermak, Simone Kotthaus, Martial Haeffelin, Hendrik Andersen, Julia Fuchs, Miae Kim, Jean-Eudes Petit, and Olivier Favez
	2020-07-27
发表期刊	Atmospheric Chemistry and Physics
出版年	2020
英文摘要	Air pollution, in particular high concentrations of particulate matter smaller than 1 µm in diameter (PM1), continues to be a major health problem, and meteorology is known to substantially contribute to atmospheric PM concentrations. However, the scientific understanding of the complex mechanisms leading to high pollution episodes is inconclusive, as the effects of meteorological variables are not easy to separate and quantify. In this study, a novel, data-driven approach based on empirical relationships is used to characterise the role of meteorology on atmospheric concentrations of PM1. A tree-based machine learning model is set up to reproduce concentrations of speciated PM1 at a suburban site southwest of Paris, France, using meteorological variables as input features. The contributions of each meteorological feature to modeled PM1 concentrations are quantified using SHapley Additive exPlanation (SHAP) regression values. Meteorological contributions to PM1 concentrations are analysed in selected high-resolution case studies, contrasting season-specific processes. Model results suggest that winter pollution episodes are often driven by a combination of shallow mixed layer heights (MLH), low temperatures, low wind speeds or inflow from northeastern wind directions. Contributions of MLHs to the winter pollution episodes are quantified to be on average ~ 5 µg/m³ for MLHs below 500 m agl. Temperatures below freezing initiate formation processes and increase local emissions related to residential heating, amounting to a contribution of as much as ~ 9 µg/m³. Northeasterly winds are found to contribute ~ 5 µg/m³ to total PM1 concentrations (combined effects of u- and v-wind components), by advecting particles from source regions, e.g. central Europe or the Paris region. However, in calm conditions (i.e. wind speeds < ~ 2 m/s), the lack of dispersion leads to increased PM1 concentrations by ~ 3 µg/m³. Unusually high PM1 concentrations in summer are generally lower compared to winter peak concentrations, and are characterised by a higher content of organics. Meteorological drivers of summer peak PM1 concentrations are temperatures above ~ 25 °C (contributions of up to ~ 2.5 µg/m³), dry spells of several days (maximum contributions of ~ 1.5 µg/m³) and wind speeds below ~ 2 m/s (maximum contributions of ~ 3 µg/m³). High-resolution case studies show a large variability of processes, which together lead to high PM1 concentrations. Processes vary even within seasons. A high pollution episode in January 2016 is shown to be driven by a drop in temperature (maximum contributions of 11 µg/m³), which enhances formation of secondary inorganic aerosols (SIA) and likely causes an increase in local wood-burning emissions. In contrast, during December 2016, high PM1 concentrations are caused mainly by a shallow MLH and low wind speeds. It is shown that an observed decrease in pollution levels is linked to a change in wind direction, advecting cleaner, maritime air to the PM measurement site (combined contributions of u- and v-wind-components of ~ −4 µg/m³). The application of SHAP regression values within a machine learning framework presents a novel and promising way of analysing observational data sets in environmental sciences.
领域	地球科学
URL	查看原文
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.173/C666/handle/2XK7JSWQ/286647
专题	地球科学
推荐引用方式 GB/T 7714	Roland Stirnberg, Jan Cermak, Simone Kotthaus, Martial Haeffelin, Hendrik Andersen, Julia Fuchs, Miae Kim, Jean-Eudes Petit, and Olivier Favez. Meteorology-driven variability of air pollution (PM1) revealed with explainable machine learning[J]. Atmospheric Chemistry and Physics,2020.
APA	Roland Stirnberg, Jan Cermak, Simone Kotthaus, Martial Haeffelin, Hendrik Andersen, Julia Fuchs, Miae Kim, Jean-Eudes Petit, and Olivier Favez.(2020).Meteorology-driven variability of air pollution (PM1) revealed with explainable machine learning.Atmospheric Chemistry and Physics.
MLA	Roland Stirnberg, Jan Cermak, Simone Kotthaus, Martial Haeffelin, Hendrik Andersen, Julia Fuchs, Miae Kim, Jean-Eudes Petit, and Olivier Favez."Meteorology-driven variability of air pollution (PM1) revealed with explainable machine learning".Atmospheric Chemistry and Physics (2020).