National empirical models of air pollution using microscale measures of the urban environment
- Author
- Tianjun Lu, Julian D. Marshall, Wenwen Zhang, Perry Hystad, Sun-Young Kim, Matthew J. Bechle, Matthias Demuzere (UGent) and Steve Hankey
- Organization
- Abstract
- National-scale empirical models of air pollution (e.g., Land Use Regression) rely on predictor variables (e.g., population density, land cover) at different geographic scales. These models typically lack microscale variables (e.g., street level), which may improve prediction with fine-spatial gradients. We developed microscale variables of the urban environment including Point of Interest (POI) data, Google Street View (GSV) imagery, and satellite-based measures of urban form. We developed United States national models for six criteria pollutants (NO2, PM2.5, O-3, CO, PM10, SO2) using various modeling approaches: Stepwise Regression + kriging (SW-K), Partial Least Squares + kriging (PLS-K), and Machine Learning + kriging (ML-K). We compared predictor variables (e.g., traditional vs microscale) and emerging modeling approaches (ML-K) to well-established approaches (i.e., traditional variables in a PLS-K or SW-K framework). We found that combined predictor variables (traditional + microscale) in the ML-K models outperformed the well-established approaches (10-fold spatial cross-validation (CV) R-2 increased 0.02-0.42 [average: 0.19] among six criteria pollutants). Comparing all model types using microscale variables to models with traditional variables, the performance is similar (average difference of 10-fold spatial CV R-2 = 0.05) suggesting microscale variables are a suitable substitute for traditional variables. ML-K and microscale variables show promise for improving national empirical models.
- Keywords
- Empirical models, street-level features, urban form, exposure assessment, machine learning, LAND-USE REGRESSION, GOOGLE STREET VIEW, EXPOSURE ASSESSMENT, ULTRAFINE PARTICLES, BLACK CARBON, PM2.5, NUMBER, CANADA, OXIDES
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-8742676
- MLA
- Lu, Tianjun, et al. “National Empirical Models of Air Pollution Using Microscale Measures of the Urban Environment.” ENVIRONMENTAL SCIENCE & TECHNOLOGY, vol. 55, no. 22, 2021, pp. 15519–30, doi:10.1021/acs.est.1c04047.
- APA
- Lu, T., Marshall, J. D., Zhang, W., Hystad, P., Kim, S.-Y., Bechle, M. J., … Hankey, S. (2021). National empirical models of air pollution using microscale measures of the urban environment. ENVIRONMENTAL SCIENCE & TECHNOLOGY, 55(22), 15519–15530. https://doi.org/10.1021/acs.est.1c04047
- Chicago author-date
- Lu, Tianjun, Julian D. Marshall, Wenwen Zhang, Perry Hystad, Sun-Young Kim, Matthew J. Bechle, Matthias Demuzere, and Steve Hankey. 2021. “National Empirical Models of Air Pollution Using Microscale Measures of the Urban Environment.” ENVIRONMENTAL SCIENCE & TECHNOLOGY 55 (22): 15519–30. https://doi.org/10.1021/acs.est.1c04047.
- Chicago author-date (all authors)
- Lu, Tianjun, Julian D. Marshall, Wenwen Zhang, Perry Hystad, Sun-Young Kim, Matthew J. Bechle, Matthias Demuzere, and Steve Hankey. 2021. “National Empirical Models of Air Pollution Using Microscale Measures of the Urban Environment.” ENVIRONMENTAL SCIENCE & TECHNOLOGY 55 (22): 15519–15530. doi:10.1021/acs.est.1c04047.
- Vancouver
- 1.Lu T, Marshall JD, Zhang W, Hystad P, Kim S-Y, Bechle MJ, et al. National empirical models of air pollution using microscale measures of the urban environment. ENVIRONMENTAL SCIENCE & TECHNOLOGY. 2021;55(22):15519–30.
- IEEE
- [1]T. Lu et al., “National empirical models of air pollution using microscale measures of the urban environment,” ENVIRONMENTAL SCIENCE & TECHNOLOGY, vol. 55, no. 22, pp. 15519–15530, 2021.
@article{8742676,
abstract = {{National-scale empirical models of air pollution (e.g., Land Use Regression) rely on predictor variables (e.g., population density, land cover) at different geographic scales. These models typically lack microscale variables (e.g., street level), which may improve prediction with fine-spatial gradients. We developed microscale variables of the urban environment including Point of Interest (POI) data, Google Street View (GSV) imagery, and satellite-based measures of urban form. We developed United States national models for six criteria pollutants (NO2, PM2.5, O-3, CO, PM10, SO2) using various modeling approaches: Stepwise Regression + kriging (SW-K), Partial Least Squares + kriging (PLS-K), and Machine Learning + kriging (ML-K). We compared predictor variables (e.g., traditional vs microscale) and emerging modeling approaches (ML-K) to well-established approaches (i.e., traditional variables in a PLS-K or SW-K framework). We found that combined predictor variables (traditional + microscale) in the ML-K models outperformed the well-established approaches (10-fold spatial cross-validation (CV) R-2 increased 0.02-0.42 [average: 0.19] among six criteria pollutants). Comparing all model types using microscale variables to models with traditional variables, the performance is similar (average difference of 10-fold spatial CV R-2 = 0.05) suggesting microscale variables are a suitable substitute for traditional variables. ML-K and microscale variables show promise for improving national empirical models.}},
author = {{Lu, Tianjun and Marshall, Julian D. and Zhang, Wenwen and Hystad, Perry and Kim, Sun-Young and Bechle, Matthew J. and Demuzere, Matthias and Hankey, Steve}},
issn = {{0013-936X}},
journal = {{ENVIRONMENTAL SCIENCE & TECHNOLOGY}},
keywords = {{Empirical models,street-level features,urban form,exposure assessment,machine learning,LAND-USE REGRESSION,GOOGLE STREET VIEW,EXPOSURE ASSESSMENT,ULTRAFINE PARTICLES,BLACK CARBON,PM2.5,NUMBER,CANADA,OXIDES}},
language = {{eng}},
number = {{22}},
pages = {{15519--15530}},
title = {{National empirical models of air pollution using microscale measures of the urban environment}},
url = {{http://doi.org/10.1021/acs.est.1c04047}},
volume = {{55}},
year = {{2021}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: