Factors associated with NO2 and NOX concentration gradients near a highway

Abstract

The objective of this research is to learn how the near-road gradient, in which NO₂ and NO_X (NO + NO₂) concentrations are elevated, varies with changes in meteorological and traffic variables. Measurements of NO₂ and NO_X were obtained east of I-15 in Las Vegas and fit to functions whose slopes (dC_NO2 /dx and dC_NOX /dx, respectively) characterize the size of the near-road zone where NO₂ and NO_X concentrations from mobile sources on the highway are elevated. These metrics were used to learn about the near-road gradient by modeling dC_NO2 /dx and dC_NOX /dx as functions of meteorological variables (e.g., wind direction, wind speed), traffic (vehicle count), NO_X concentration upwind of the road, and O₃ concentration at two fixed-site ambient monitors. Generalized additive models (GAM) were used to model dC_NO2 /dx and dC_NOX /dx versus the independent variables because they allowed for nonlinearity of the variables being compared. When data from all wind directions were included in the analysis, variability in O₃ concentration comprised the largest proportion of variability in dC_NO2 /dx, followed by variability in wind direction. In a second analysis constrained to winds from the west, variability in O₃ concentration remained the largest contributor to variability in dC_NO2 /dx, but the relative contribution of variability in wind speed to variability in dC_NO2 /dx increased relative to its contribution for the all-wind analysis. When data from all wind directions were analyzed, variability in wind direction was by far the largest contributor to variability in dC_NOX /dx, with smaller contributions from hour of day and upwind NO_X concentration. When only winds from the west were analyzed, variability in upwind NO_X concentration, wind speed, hour of day, and traffic count all were associated with variability in dC_NOX /dx. Increases in O₃ concentration were associated with increased magnitude near-road dC_NO2 /dx, possibly shrinking the zone of elevated concentrations occurring near roads. Wind direction parallel to the highway was also related to an increased magnitude of both dC_NO2 /dx and dC_NOX /dx, again likely shrinking the zone of elevated concentrations occurring near roads. Wind direction perpendicular to the road decreased the magnitude of dC_NO2 /dx and dC_NOX /dx and likely contributed to growth of the zone of elevated concentrations occurring near roads. Thus, variability in near-road concentrations is influenced by local meteorology and ambient O₃ concentration.

Keywords: Dispersion; NO2; Near-road; Nitrogen dioxide; Oxides of nitrogen.

Fig. 1

Map of the Las Vegas near-road sampling site (Kimbrough et al., 2017).

Fig. 2

Concentration (in ppb) roses for NO₂ during stable conditions (top left) and unstable conditions (top right), and concentrations roses for NO_X during stable conditions (bottom left) and unstable conditions (bottom right). Concentrations presented in this figure were measured at the 25 m site to the east of I-15.

Fig. 3

GAM models of NO₂ gradient (dC_NO2/dx) with respect to each smooth meteorological function. (first row l-r): Temperature, sensible heat flux, inverse Monin-Obukhov length scale; (second row l-r): wind speed, wind direction, traffic count; (third row, l-r): ambient O₃ concentration, upwind NO_X concentration, hour of day.

Fig. 4

GAM models of NO₂ gradient (dC_NO2/dx) with respect to each smooth meteorological function when winds are constrained to come from the west (210–330°). (first row l-r): Temperature, Sensible heat flux, inverse Monin-Obukhov length scale; (second row l-r): wind speed, traffic count; (third row, l-r): ambient O₃ concentration, upwind NO_X concentration, hour of day.

Fig. 5

GAM models of NO_X gradient (dC_NOX/dx) with respect to each smooth meteorological function. (first row l-r): Temperature, Sensible heat flux, inverse Monin-Obukhov length scale; (second row l-r): wind speed, wind direction, traffic count; (third row, l-r): ambient O₃ concentration, upwind NO_X concentration, hour of day.

Fig. 6

GAM models of NO_X gradient (dC_NOX/dx) with respect to each smooth meteorological function when winds are constrained to come from the west (210–330°). (first row l-r): Temperature, Sensible heat flux, inverse Monin-Obukhov length scale; (second row l-r): wind speed, traffic count; (third row, l-r): ambient O₃ concentration, upwind NO_X concentration, hour of day.