Air pollution is highly variable in both space and time, which presents many challenges to researchers when they wish to model concentrations. The modelling of air pollution is necessary for a number of reasons, which include the determination of human health effects, providing warning of health risks, and to understand general ecosystem health. In this thesis, modelling of air pollution through both space and time has been explored, with a focus on improving models that can be used to assign air pollution exposure. The techniques presented in this thesis have leveraged the ability of mobile monitoring units to collect air pollution concentration data multiple locations throughout a study period. First, we explore the use of combining mobile air pollution monitoring data with traditional fixed location monitoring. We find that the mobile data is able to provide insight into changes in spatial pattern between two temporal periods that could not be identified solely with the fixed location monitors, which demonstrates value in this monitoring approach that can be built upon with refinement of techniques. Second, we present a method to determine the amount of classical error that will be introduced when a long-term mean concentration is calculated from a discontinuous time-series dataset, which are the type of datasets collected by mobile air pollution monitoring. Third, we merge mobile and stationary air pollution monitoring data, along with meteorological, transportation, and land use information to model the hourly air pollution field using neural network models. The models developed allowed for the assignment of air pollution exposure incorporating human activity patterns. Also, they can be used to provide a spatially refined air quality health index. Lastly, we demonstrate exposure assignment that incorporates human activity patterns to calculate the dose exposure for students during their trips to school.
This work commences with a demonstration of the basic utility of mobile air pollution monitoring data, which is to increase the number of monitored locations. Building on that utility of mobile technology, a technique was developed to estimate the error when mobile units are used for long-term estimates, similar to stationary monitoring units; and we were able to provide guiding principles for mobile monitoring data collection. Furthering our objective, to better understand the value of mobile data in a fully integrated monitoring network, we utilized both mobile and stationary data collection techniques together, in a single model, to produce accurate estimates of an air pollution field on an hourly basis. Finally, the research culminates with the demonstration of how mobile monitoring can be used for activity based air pollution exposure estimates, which was shown with a case-study of students’ trips between home and school. Overall, the chapters in this thesis work toward a better understanding of how to incorporate mobile monitoring data into air pollution assessment studies. / Thesis / Doctor of Philosophy (PhD)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/18032 |
| Date | 11 1900 |
| Creators | Adams, Matthew |
| Contributors | Kanaroglou, Pavlos, School of Geography and Geology |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0022 seconds