The emergence of sensor-based air monitoring technology brings a major paradigm shift that allows for the collection of much needed data in microenvironments far beyond what conventional air quality monitoring (AQM) networks offer. Most conventional exposure assessment methods are based on limited fixed monitoring sites. The sensor-based systems offer a more flexible and cost-effective monitoring approach, which allows gathering data with high spatial and temporal resolution. However, considerable care is needed on the quality assurance of data reported by sensor-based systems for unattended field applications. Moreover, there remained one key unsolved challenge of sensors, sensor signal drift, which may hinder the application of this approach. The overarching objective of the work is t...[ Read more ]

The emergence of sensor-based air monitoring technology brings a major paradigm shift that allows for the collection of much needed data in microenvironments far beyond what conventional air quality monitoring (AQM) networks offer. Most conventional exposure assessment methods are based on limited fixed monitoring sites. The sensor-based systems offer a more flexible and cost-effective monitoring approach, which allows gathering data with high spatial and temporal resolution. However, considerable care is needed on the quality assurance of data reported by sensor-based systems for unattended field applications. Moreover, there remained one key unsolved challenge of sensors, sensor signal drift, which may hinder the application of this approach. The overarching objective of the work is to demonstrate that science-based data adjustment, system modification and protocol design are essential to improve the sensor data quality and that sensor-based monitoring is an effective approach to evaluate microenvironments of concern.

Three projects are reported on in this thesis covering sensor applications in conditions common in Hong Kong urban built environments. These include ad-hoc deployment of sensor instruments in controlled roadway environments (AQM network along a marathon route), campaign-based monitoring in multiple environments (simultaneous indoor and outdoor AQM in schools during high pollution episodes), and campaign-based measurement in vehicular source dominated environments (AQM network in semi-confined, highly-polluted bus terminals). Application-oriented sensor systems were designed and deployed to meet specific application requirements. A module developed in this thesis to automatically address sensor drift problem was intergraded into the sensor systems. Extensive protocol refinements were also implemented to overcome challenges encountered in each project and to meet data quality objectives.

Air monitoring data with high temporal and spatial density were successfully captured in the above-mentioned three projects. Monitoring results of the marathon showed the effects of temporary traffic control during the event. Levels of traffic related air pollutant, NO₂ and CO, declined during the control periods while PM_2.5 appeared to be less affected. Mechanical ventilation was also found to be important to ensure good air quality inside a major tunnel. AQM in schools found decreased NO₂ and O₃ levels and slightly elevated particulate matter concentration during the summertime when mechanical ventilation was in use, which ultimately drove a relative healthier indoor environment compared to the polluted outdoor air. However, high pollution conditions were found in one auditorium that had natural ventilation. Insufficient ventilation was observed in classrooms as indicated by CO₂ data. Thus, refinements are needed to both reduce pollutants from outside air while also reducing CO₂ generated by the students. In the bus terminal AQM study, diesel buses were found to be important contributors to the NO and NO₂ pollution. Pollution concentration patterns for these gases clearly tracked bus activity while PM was seen to more clearly reflect levels in outdoor air. Natural and mechanical ventilation was inadequate to dilute gaseous pollutants in semi-confined venues and placement of fresh air intakes in traffic environments may even have direct adverse effect on bus terminal air quality.

Overall, the steps taken to improve data quality from sensor-based air monitoring systems have made important contributions. When care is taken it is possible to gather needed data of high quality in situations and locations where harmful pollution condition may be present, but that is quite difficult to assess with conventional monitoring methods. There is great promise for further refinement and application of sensor-based monitoring to address personal and microenvironmental pollution and exposure conditions.