Fleet-wide Emissions from Mobile CO2 Measurements

In response to regional and municipal policies, transportation agencies are increasingly integrating greenhouse gas considerations into decision making. At the local level, fuel-based methods suffer leakage, mandating a bottom-up approach based on emissions models driven by local activity data. However, high spatial and temporal resolution traffic datasets are in general scarce and subject to error. Emissions models too are based on limited data and often require inputs that are not directly measured. Here, we show that routine, on-road CO2 surface measurements can be used to improve uncertainties on both of these fronts. Using forty hours of surface concentration data collected on CA Highway 24 together with a simple atmospheric dispersion model, we simultaneously derive traffic density as a function of vehicle speed, composite vehicle parameters needed to map vehicle operation to fuel consumption, and baseline meteorological parameters such as wind speed and mixing height. We compare our results directly with traffic loop detector measurements made by California's Performance Measurement System (PeMS), with emissions predictions from EPA's MOtor Vehicle Emission Simulator (MOVES), and with weather station data included in NOAA's Meteorological Assimilation Data Ingest System (MADIS). Using both top-down and bottom-up techniques, we measure the immediate rush-hour emissions reduction associated with congestion alleviation following the opening of the Caldecott Tunnel fourth bore. We use this example to argue that routine and distributed on-road measurements of this kind could serve as a much needed policy tool for testing the impact of traffic-related emissions reduction strategies.