The TRANSIT Accountability Study: Methods to Assess the Impacts of Local Congestion Programs on Traffic Air Pollution in Texas
Perry Hystad1, Mary Willis1, Elaine Hill2, Andrew Larkin1, John Molitor1, Beate Ritz3, Dennis Perkinson4, David Schrank4
1Oregon State University; Corvallis, OR, USA; 2University of Rochester; Rochester, New York, USA; 3University of California, Los Angeles; Los Angeles, California, USA; 4Texas Transportation Institute; Bryan, Texas, USA
Background: Traffic congestion is a major contributor to local air pollution. Diverse local programs have been implemented in Texas to reduce traffic congestion, for example, through electronic tolls and managed lanes to highway expansions. Few studies have empirically evaluated the direct impact of these programs on congestion, resulting changes to traffic related air pollution (TRAP), and infant health.
Methods: We leverage a diverse population-based cohort of 7.6 million births in Texas from 1996-2016 to implement the Traffic Regulations And Neonates Study In Texas (TRANSIT) Accountability Study. In conjunction with the Texas Transportation Institute, we are integrating diverse data sources to determine the most congested roadways in Texas, select historical traffic congestion projects, and quantify congestion before, during and after projects. The key dataset for measuring congestion comes from the Texas Top 100 Most Congested Roadways Project and vehicle “telematics”, where actual travel speed can be measured from connected vehicles and devices at 15 minute intervals and compared to free-flow travel speed for road segments in Texas. Major infrastructure projects are identified and characterized from administrative databases of construction billing.
Results: We have linked traffic congestion measures for ~1,800 of the most congested roadways in Texas from 2010 to 2017. Congestion metrics include total delay per mile (integrates vehicle volume with vehicle delay), and excess CO2 (integrates the EPA MOVES model with vehicle volume and delay). On these roads, congestion metrics have both increased and decreased during this time-period. We have linked over 26,000 construction projects to these congestion measures to examine change for specific roadways and projects. In preliminary analyses, we find mixed results for the impacts of different congestion reduction programs on congestion and vehicle emissions. Results also varied by trucks compared to all vehicles.
Conclusion: By leveraging novel data sources we are able to quantify congestion changes associated with different construction projects across Texas. We will integrate these data with our population-based birth cohort, develop models to examine associations between changes in traffic delay and adverse birth outcomes, and implement causal inference methods (e.g. difference-in-difference analyses) to assess the impact of local congestion reduction programs on adverse birth outcomes.