Mobile monitoring reveals congestion penalty for vehicle emissions in London

Abstract

Mobile air pollution measurements have the potential to provide a wide range of insights into emission sources and air pollution exposure. The analysis of mobile data is, however, highly challenging. In this work we develop a new regression-based framework for the analysis of mobile data with the aim of improving the potential to draw inferences from such measurements. A quantile regression approach is adopted to provide new insight into the distribution of NO_x and CO emissions in Central and Outer London. We quantify the emissions intensity of NO_x and CO (ΔNO_x/ΔCO₂ and ΔCO/ΔCO₂) at different quantile levels (τ) to demonstrate how transient high-emission events can be examined in parallel to the average emission characteristics. We observed a clear difference in the emissions behaviour between both locations. On average, the median (τ = 0.5) ΔNO_x/ΔCO₂ in Central London was 2x higher than Outer London, despite the stringent emission standards imposed throughout the Ultra Low Emissions Zone. A comprehensive vehicle emission remote sensing data set (n ≈ 700,000) is used to put the results into context, providing evidence of vehicle behaviour which is indicative of poorly controlled emissions, equivalent to high-emitting classes of older vehicles. Our analysis suggests the coupling of a diesel-dominated fleet with persistently congested conditions, under which the operation of emissions after-treatment technology is non-optimal, leads to increased NO_x emissions.