High-resolution mapping of vehicle emissions of atmospheric pollutants based on large-scale, real-world traffic datasets

Abstract

Abstract. On-road vehicle emissions are a major contributor to elevated air pollution levels in populous metropolitan areas. We developed a link-level emissions inventory of vehicular pollutants, called EMBEV-Link (Link-level Emission factor Model for the BEijing Vehicle fleet), based on multiple datasets extracted from the extensive road traffic monitoring network that covers the entire municipality of Beijing, China (16 400 km2). We employed the EMBEV-Link model under various traffic scenarios to capture the significant variability in vehicle emissions, temporally and spatially, due to the real-world traffic dynamics and the traffic restrictions implemented by the local government. The results revealed high carbon monoxide (CO) and total hydrocarbon (THC) emissions in the urban area (i.e., within the Fifth Ring Road) and during rush hours, both associated with the passenger vehicle traffic. By contrast, considerable fractions of nitrogen oxides (NOx), fine particulate matter (PM2.5) and black carbon (BC) emissions were present beyond the urban area, as heavy-duty trucks (HDTs) were not allowed to drive through the urban area during daytime. The EMBEV-Link model indicates that nonlocal HDTs could account for 29 % and 38 % of estimated total on-road emissions of NOx and PM2.5, which were ignored in previous conventional emission inventories. We further combined the EMBEV-Link emission inventory and a computationally efficient dispersion model, RapidAir®, to simulate vehicular NOx concentrations at fine resolutions (10 m × 10 m in the entire municipality and 1 m × 1 m in the hotspots). The simulated results indicated a close agreement with ground observations and captured sharp concentration gradients from line sources to ambient areas. During the nighttime when the HDT traffic restrictions are lifted, HDTs could be responsible for approximately 10 µg m−3 of NOx in the urban area. The uncertainties of conventional top-down allocation methods, which were widely used to enhance the spatial resolution of vehicle emissions, are also discussed by comparison with the EMBEV-Link emission inventory.

PMIF v1.0: an inversion system to estimate the potential of satellite observations to monitor fossil fuel CO<sub>2</sub> emissions over the globe

“…Even if emissions can be effectively constrained by CO2M for clumps whose emissions are larger than 2 MtC per year, the sum of annual emission budgets from these large clumps account only for 54% of the total CO 2 clump emissions and for 36% of the total global fossil fuel CO 2 emissions (accounting for diffuse emissions outside the clumps), according to the clump definition of Wang et al (2019) and the ODIAC emission map. For a specific country, clumps with emissions larger than 2 MtC per year typically represent less than 50% of the total national emissions (accounting for diffuse emissions outside the clumps).…”

“…The size of the plumes and the magnitude of XCO 2 enhancements in these plumes are tightly linked to the emissions. Wang et al (2019) developed an algorithm to extract, from gridded emission maps, a conservative set of area (cities) and point sources (power plants) with intense emissions around the globe which can generate coherent XCO 2 plumes that may be observed from space, given the precision of current satellite observations. This set was conservative because it is inferred for idealized meteorological condition without wind.…”

“…These emitting sources were called "emission clumps". Wang et al (2019) identified 11,314 individual clumps which contribute 72% of the global fossil fuel CO 2 emissions from the ODIAC (Open-source Data Inventory for Anthropogenic CO 2 version 2017, Oda et al, 2018) 1 km resolution inventory. Broquet et al (2018) showed that the part of the XCO 2 plumes exploited by the atmospheric inversion in satellite images correspond to few hours of the clump emissions before the satellite overpass.…”

“…The XCO 2 signature of the earlier clump emissions is too diluted to be filtered from the measurement errors and the signature of other CO 2 sources and sinks. Further, emissions from a given clump vary in time during the day, for instance due to the variations of traffic in cities (Yang et al, 2019), from day to day and between seasons, with more emissions associated to heating in winter over cold regions (Bré on et al, 2017). Therefore, the estimate of annual budgets of the clump emissions based on satellite observation during daytime (generally for a fixed local time since most of the missions use heliosynchronous orbits) and for low cloud coverage is a challenge, and cannot rely on the direct information from the satellite imagery.…”

“…It would be a high-resolution spectrometer, with 2 km × 2 km resolution pixels and a swath of 300 km, and it would be placed on a sun-synchronous orbit ensuring global coverage in 4 days. The PMIF inversion system relies on the list of clumps extracted by Wang et al (2019) from the ODIAC inventory, on the Gaussian plume model to simulate the XCO 2 plumes generated by the emissions from these clumps, on an analytical inverse modeling framework, and on a combination of overlapping assimilation windows to solve for the inversion problem over the globe and a full year. It also addresses the question of temporal extrapolation that is needed to generate estimates of annual emissions from the information of a limited number of time windows for which emissions are well constrained by the direct satellite images, by accounting for the temporal auto-correlation of the prior uncertainties.…”