A land use regression model for long-term ground-level Ozone: An application of mobile air pollution monitoring.

Authors: Anna Shadrova*, University of Toronto, Matthew Adams, University of Toronto
Topics: Spatial Analysis & Modeling, Geographic Information Science and Systems
Keywords: Air pollution, Land use regression, Ozone, Hamilton
Session Type: Paper
Day: 4/3/2019
Start / End Time: 4:30 PM / 6:10 PM
Room: Marshall South, Marriott, Mezzanine Level
Presentation File: No File Uploaded

In this paper, we investigate the spatial variation of ground-level Ozone concentrations. Air pollution observations were obtained between 2012 and 2015 with mobile sampling in Hamilton, ON, Canada. The City of Hamilton is classified as an Air Zone 3, which is the most severe air quality category in Ontario and is characterized by a concentration of large industrial sources. Mobile monitored Ozone concentrations ranged from 0 to 237 ppb, with a mean concentration of 21 ppb. Four fixed location air pollution monitors indicated an average concentration of 26 ppb across the entire period. The mobile Ozone observations were adjusted to the long-term mean using fixed location monitors in the study area. Land use characteristics were calculate within buffers of the Ozone observation locations, which included road length, road type, traffic intensity, population density, and elevation. A linear regression model and a neural network model were used in a land use regression framework to interpolate air pollution concentrations in the study area. The model associated the air pollution observations to the surrounding land use characteristics. Model cross-validation included a traditional cross-validation approach and a spatial-blocking cross validation. The statistical model was applied to estimate the 2012-2015 average ground-level Ozone concentrations on a 100 m grid.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login