Estimating traffic contribution to particulate matter concentration in urban areas using a multilevel Bayesian meta-regression approach.
Environment International 2020 ; 141: 105800.
Heydari S, Tainio M, Woodcock J, de Nazelle A
DOI : 10.1016/j.envint.2020.105800
PubMed ID : 32474298
PMCID : PMC7273192
URL : https://linkinghub.elsevier.com/retrieve/pii/S0160412020317554
Abstract
Quantifying traffic contribution to air pollution in urban settings is required to inform traffic management strategies and environmental policies that aim at improving air quality. Assessments and comparative analyses across multiple urban areas are challenged by the lack of datasets and methods available for global applications. In this study, we quantify the traffic contribution to particulate matter concentration in multiple cities worldwide by synthesising 155 previous studies reported in the World Health Organization (WHO)'s air pollution source apportionment data for PM and PM. We employed a Bayesian multilevel meta-regression that accounts for uncertainties and captures both within- and between-study variations (in estimation methods, study protocols, etc.) through study-specific and location-specific explanatory variables. The final sample analysed in this paper covers 169 cities worldwide. Based on our analysis, traffic contribution to air pollution (particulate matter) varies from 5% to 61% in cities worldwide, with an average of 27%. We found that variability in the traffic contribution estimates reported worldwide can be explained by the region of study, publication year, PM size fraction, and population. Specifically, traffic contribution to air pollution in cities located in Europe, North America, or Oceania is on average 36% lower relative to the rest of the world. Traffic contribution is 28% lower among studies published after 2005 than those published on or before 2005. Traffic contribution is on average 24% lower among cities with less than 500,000 inhabitants and 19% higher when estimated based on PM relative to PM. This quantitative summary overcomes challenges in the data and provides useful information for health impact modellers and decision-makers to assess impacts of traffic reduction policies.