Improving the Roadside Environment through Integrating Air Quality and Traffic-Related Data

• Main Author: Kendrick, Christine M.
• Format: Others
• Published: PDXScholar 2016
• Subjects: Environmental Sciences
• Online Access: http://pdxscholar.library.pdx.edu/open_access_etds/3086; http://pdxscholar.library.pdx.edu/cgi/viewcontent.cgi?article=4096&context=open_access_etds

Urban arterial corridors are landscapes that give rise to short and long-term exposures to transportation-related pollution. With high traffic volumes, congestion, and a wide mix of road users and land uses at the road edge, urban arterial environments are important targets for improved exposure assessment to traffic-related pollution. Applying transportation management strategies to reduce emissions along arterial corridors could be enhanced if the ability to quantify and evaluate such actions was improved. However, arterial roadsides are under-sampled in terms of air pollution measurements in the United States and using observational data to assess such effects has many challenges such as lack of control sites for comparisons and temporal autocorrelation. The availability of traffic-related data is also typically limited in air monitoring and health studies. The work presented here uses unique long-term roadside air quality monitoring collected at the intersection of an urban arterial in Portland, OR to characterize the roadside atmospheric environment. This air quality dataset is then integrated with traffic-related data to assess various methods for improving exposure assessment and the roadside environment. Roadside nitric oxide (NO), nitrogen dioxide (NO2), and particle number concentration (PNC) measurements all demonstrated a relationship with local traffic volumes. Seasonal and diurnal characterizations show that roadside PM2.5 (mass) measurements do not have a relationship with local traffic volumes, providing evidence that PM2.5 mass is more tied to regional sources and meteorological conditions. The relationship of roadside NO and NO2 with traffic volumes was assessed over short and long-term aggregations to assess the reliability of a commonly employed method of using traffic volumes as a proxy for traffic-related exposure. This method was shown to be insufficient for shorter-time scales. Comparisons with annual aggregations validate the use of traffic volumes to estimate annual exposure concentrations, demonstrating this method can capture chronic but not acute exposure. As epidemiology and exposure assessment aims to target health impacts and pollutant levels encountered by pedestrians, cyclists, and those waiting for transit, these results show when traffic volumes alone can be a reliable proxy for exposure and when this approach is not warranted. Next, it is demonstrated that a change in traffic flow and change in emissions can be measured through roadside pollutant concentrations suggesting roadside pollution can be affected by traffic signal timing. The effect of a reduced maximum traffic signal cycle length on measurements of degree of saturation (DS), NO, and NO2 were evaluated for the peak traffic periods in two case studies at the study intersection. In order to reduce bias from covariates and assess the effect due to the change in cycle length only, a matched sampling method based on propensity scores was used to compare treatment periods (reduced cycle length) with control periods (no change in cycle length). Significant increases in DS values of 2-8% were found along with significant increases of 5-8ppb NO and 4-5ppb NO2 across three peak periods in both case studies. Without matched sampling to address the challenges of observational data, the small DS and NOx changes for the study intersection would have been masked and matched sampling is shown to be a helpful tool for future urban air quality empirical investigations. Dispersion modeling evaluations showed the California Line Source Dispersion Model with Queuing and Hotspot Calculations (CAL3QHCR), an approved regulatory model to assess the impacts of transportation projects on PM2.5, performed both poor and well when predictions were compared with PM2.5 observations depending on season. Varying levels of detail in emissions, traffic signal, and traffic volume data for model inputs, assessed using three model scenarios, did not affect model performance for the study intersection. Model performance is heavily dependent on background concentrations and meteorology. It was also demonstrated that CAL3QHC can be used in combination with roadside PNC measurements to back calculate PNC emission factors for a mixed fleet and major arterial roadway in the U.S. The integration of roadside air quality and traffic-related data made it possible to perform unique empirical evaluations of exposure assessment methods and dispersion modeling methods for roadside environments. This data integration was used for the assessment of the relationship between roadside pollutants and a change in a traffic signal setting, a commonly employed method for transportation management and emissions mitigation, but rarely evaluated outside of simulation and emissions modeling. Results and methods derived from this work are being used to implement a second roadside air quality station, to design a city-wide integrated network of air quality, meteorological, and traffic data including additional spatially resolved measurements with feedback loops for improved data quality and data usefulness. Results and methods are also being used to design future evaluations of transportation projects such as freight priority signaling, improved transit signal priority, and to understand the air quality impacts of changes in fleet composition such as an increase in electric vehicles.