Predicting intraurban airborne PM1.0-trace elements in a port city : Land use regression by ordinary least squares and a machine learning algorithm
Zhang, Joyce J. Y., Sun, Liu, Rainham, Daniel, Dummer, Trevor J. B., Wheeler, Amanda J., Anastasopoulos, Angelos, Gibson, Mark and Johnson, Markey. (2022). Predicting intraurban airborne PM1.0-trace elements in a port city : Land use regression by ordinary least squares and a machine learning algorithm. Science of the Total Environment. 806(1), p. Article 150149. https://doi.org/10.1016/j.scitotenv.2021.150149
|Authors||Zhang, Joyce J. Y., Sun, Liu, Rainham, Daniel, Dummer, Trevor J. B., Wheeler, Amanda J., Anastasopoulos, Angelos, Gibson, Mark and Johnson, Markey|
Airborne particulate matter (PM) has been associated with cardiovascular and respiratory morbidity and mortality, and there is some evidence that spatially varying metals found in PM may contribute to adverse health effects. We developed spatially refined models for PM trace elements using ordinary least squares land use regression (OLS-LUR) and machine leaning random forest land-use regression (RF-LUR).
Two-week integrated measurements of PM1.0 (median aerodiameter < 1.0 μm) were collected at 50 sampling sites during fall (2010), winter (2011), and summer (2011) in the Halifax Regional Municipality, Nova Scotia, Canada. PM1.0 filters were analyzed for metals and trace elements using inductively coupled plasma-mass spectrometry. OLS- and RF-LUR models were developed for approximately 30 PM1.0 trace elements in each season. Model predictors included industrial, commercial, and institutional/ government/ military land use, roadways, shipping, other transportation sources, and wind rose information.
RF generated more accurate models than OLS for most trace elements based on 5-fold cross validation. On average, summer models had the highest cross validation R2 (OLS-LUR = 0.40, RF-LUR = 0.46), while fall had the lowest (OLS-LUR = 0.27, RF-LUR = 0.31). Many OLS-LUR models displayed overprediction in the final exposure surface. In contrast, RF-LUR models did not exhibit overpredictions. Taking overpredictions and cross validation performances into account, OLS-LUR performed better than RF-LUR in roughly 20% of the seasonal trace element models. RF-LUR models provided more interpretable predictors in most cases. Seasonal predictors varied, likely due to differences in seasonal distribution of trace elements related to source activity, and meteorology.
|Keywords||outdoor air pollution; pollution exposure; land-use regression (LUR); particulate matter (PM); PM trace elements; machine learning|
|Journal||Science of the Total Environment|
|Journal citation||806 (1), p. Article 150149|
|Digital Object Identifier (DOI)||https://doi.org/10.1016/j.scitotenv.2021.150149|
|Open access||Published as ‘gold’ (paid) open access|
|Research or scholarly||Research|
File Access Level
|Online||06 Sep 2021|
|Publication process dates|
|Accepted||01 Sep 2021|
|Deposited||19 Jan 2022|
|License: CC BY-NC-ND 4.0|
|File access level: Open|
4views this month
2downloads this month