Near-Road Observations of CO, NOy, and CO2: Evidence for a Temperature Dependence of Vehicular Emissions of NOx

Carbon monoxide (CO) and nitrogen dioxide (NO₂) are emitted mainly by combustion sources, including power plants and vehicles, and are regulated by the US Environmental Protection Agency (EPA) due to their adverse impacts on human health and the environment. According to the 2014 National Emissions Inventory (NEI), published by the US EPA, 58% of total CO emissions and 60% of total NO_x (NO_x = NO + NO₂) emissions are estimated to come from highway and off-highway vehicles. Recent studies using aircraft and surface data collected during summer found that vehicular NO_x emissions in the NEI are considerably higher than those estimated from observations, while another study based upon winter aircraft observations showed good agreement with the NEI vehicular NO_x. Our research seeks to explain the overestimation of vehicular NO_x emissions reported in the literature, particularly during summer, by studying surface observations of meteorological variables, air pollutants and traffic counts collected directly adjacent to Interstate 95 (I-95), a major highway in Maryland. We examine emission ratios (ΔCO/ΔNO_x, ΔCO₂/ΔNO_x, and ΔCO₂/ΔCO) inferred from the observations for a dependence on ambient temperature and specific humidity and find that vehicular emissions of NO_x are slightly sensitive to the specific humidity and strongly dependent on ambient temperature. This temperature sensitivity appears strongest in the early morning hours, when NO_x emissions at the I-95 site are dominated by diesel-fueled vehicles. Our observations may help resolve the seasonal discrepancy between inferred and modeled vehicular emissions of NO_x reported in recent studies. We also compare the observed temperature and specific humidity sensitivities of NO_x emissions to those present in output from the MOtor Vehicle Emissions Simulator (MOVES) model used to generate vehicle emissions for the NEI and for use in air quality models. The default configuration of MOVES shows a modest specific humidity impact and a small sensitivity to temperature above 20°C (68°F) due to an indirect air conditioning effect. Although MOVES includes a multiplicative factor that directly adjusts emissions to temperature, this factor is routinely set to unity, i.e. no sensitivity to temperature. Adjusting this multiplicative factor within MOVES to better match the observed temperature dependence of vehicular emissions will provide more realistic emissions for air quality models to identify the individual impact on surface O₃ from various emission sources.