7A tunnel emissions study was conducted to (partially) validate the Australian vehicle emissions software 8 COPERT Australia and PIARC emission factors. The in-tunnel fleet mix differs substantially from the average 9 on-road fleet, leading to lower emissions by factor of about 2. Simulation with the P∆P software found that in-10 tunnel air-flow roughly compensates for road gradient impacts on NO x emissions. PIARC emission factors are 11 conservative and exhibit the largest prediction errors, except for one very good agreement for LDV NO x . 12COPERT Australia is generally accurate at fleet level for CO, NO x , PM 2.5 and PM 10 , when compared with other 13 international studies, and consistently underestimates emissions by 7% to 37%, depending on the pollutant. 14 Possible contributing factors are under-representation of high/excessive emitting vehicles, inaccurate mileage 15 correction factors, and lack of empirical emissions data for Australian diesel cars. The study results demonstrate 16 a large uncertainty in speciated VOC and PAH emission factors. 17 Graphical abstract 18 19 20 Highlights 21 • Tunnel studies are useful to partially validate vehicle emissions software 22 • Air flow in tunnels can compensate the impacts of road gradients on vehicle emissions 23 • Local fleet mix is an essential factor in validation studies 24 Keywords 25 Motor vehicle; emissions; tunnel; validation; road traffic 26 1. Introduction 27 Motor vehicles are a major source of air pollution and greenhouse gas (GHG) emissions in urban areas around 28 the world. The close proximity of motor vehicles to the general population makes this a particularly relevant 29 source from an exposure and health perspective. This is illustrated by Caiazzo et al. (2013) who estimated that 30 total combustion emissions (particulates, ozone) in the U.S. account for about 210,000 premature deaths per 31year, with motor vehicles being the largest contributor, contributing to around 58,000 premature deaths per year, 32 despite the fact that road transport only contributes about 7% to total PM 2.5 emissions. 33Comprehensive measurement of vehicle emissions in urban networks is cost prohibitive due to the large number 34 of vehicles that operate on roads with different emission profiles, large spatial and temporal variability in vehicle 35 activity and many real-world factors that influence emission levels (Smit et al., 2008). The environmental 36 impacts of road traffic are therefore commonly evaluated at different scales using transport and emission models 37 and, in the case of air pollution, dispersion and exposure models. Models are also required to make projections 38 into the future. 39Vehicle-emission prediction software is well-established in Europe and the US. However, these models have 40 been found to not adequately represent Australian conditions in terms of fleet mix, vehicle technology, fuel 41 quality and climate. Large errors of up to a factor of 20 have been reported when overseas models were directly 42 applied to Australian conditions w...

No abstract