Abstract
AbstractThe quantification of factors leading to harmfully high levels of particulate matter (PM) remains challenging. This study presents a novel approach using a statistical model that is trained to predict hourly concentrations of particles smaller than 10 m (PM10) by combining satellite‐borne aerosol optical depth (AOD) with meteorological and land‐use parameters. The model is shown to accurately predict PM10 (overall R = 0.77, RMSE = 7.44 g/m ) for measurement sites in Germany. The capability of satellite observations to map and monitor surface air pollution is assessed by investigating the relationship between AOD and PM10 in the same modeling setup. Sensitivity analyses show that important drivers of modeled PM10 include multiday mean wind flow, boundary layer height (BLH), day of year (DOY), and temperature. Different mechanisms associated with elevated PM10 concentrations are identified in winter and summer. In winter, mean predictions of PM10 concentrations >35 g/m occur when BLH is below 500 m. Paired with multiday easterly wind flow, mean model predictions surpass 40 g/m of PM10. In summer, PM10 concentrations seemingly are less driven by meteorology, but by emission or chemical particle formation processes, which are not included in the model. The relationship between AOD and predicted PM10 concentrations depends to a large extent on ambient meteorological conditions. Results suggest that AOD can be used to assess air quality at ground level in a machine learning approach linking it with meteorological conditions.
Highlights
particles smaller than 10 μm (PM10) concentrations seemingly are less driven by meteorology, but by emission or chemical particle formation processes, which are not included in the model
Results suggest that aerosol optical depth (AOD) can be used to assess air quality at ground level in a machine learning approach linking it with meteorological conditions
The underestimation is due to processes not captured by the input features, that is, street-scale processes not covered by AOD observations but still influencing PM10 observations, such as increased PM10 emissions due to traffic jams or localized dust resuspension
Summary
Extensive research has been conducted in recent years on the adverse health effects of particulate matter (PM) on the human cardiovascular system and the lungs. With increasing data availability and computational power, machine learning methods, for example, artificial neural networks (Gupta & Christopher, 2009b; Di et al, 2016) and random forests (RF) (Brokamp et al, 2017; Chen et al, 2018; Grange et al, 2018) have been applied frequently in recent years These machine learning models are beneficial as they efficiently reproduce nonlinear relationships and interactions of input features (Brokamp et al, 2017; Elith et al, 2008). The present study builds upon the approaches applied in these studies but provides a more in-depth analysis of model-inherent relationships To this end, gradient boosted regression trees (GBRT) are used to understand and quantify the conditions driving air quality, as well as determinants of the relationship between AOD and PM10. A basis is set for targeted satellite-based analyses of spatial patterns of air quality
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
