Seasonal variations in δ13C and δ18O of atmospheric CO2 measured in the urban boundary layer over Vancouver, Canada in relation to fuel emissions.

Abstract:

Recent advances in techniques to measure carbon dioxide (CO₂) in urban plumes show potential for validating and monitoring emission inventories at regional to urban scale. A major challenge remains the attribution of elevated CO₂ in urban plumes to different fuel and biogenic sources. Stable isotopes are a promising source of additional information.

Here, we report a full year of measurements of CO2 mixing ratios, δ¹³C and δ¹⁸O in CO₂ in the urban boundary layer over Vancouver, Canada. The goal of the work is to link seasonally changing isotopic composition to dominant fuel sources and put the urban enhancement into the context of regional background concentrations.

Atmospheric composition in the urban atmosphere was measured continuously using a tunable diode laser absorption system (TGA 200, Campbell Scientific, Logan, UT, USA). In addition, end member signatures were determined by means of bag samples from representative fuel emission sources (gasoline, diesel, natural gas). While δ¹³C depends on the fuel type and origin (for Vancouver in 2013/14: δ¹³C gasoline 27.2‰; diesel -28.8‰; natural gas -41.6‰), δ¹⁸O is fractionated in catalytic converters (d¹⁸O gasoline vehicles -12.5‰; diesel -18.6‰; natural gas -22.7‰) and exhibits higher variability between samples. Additional signatures were determined for human, soil and plant respiration.

During the study year, monthly mean mixing ratios in the urban atmosphere ranged between 410.5 (Jul) and 425.7 ppm (Dec), which was on average 18 ppm elevated above the regional background. As expected, mean monthly δ¹³C was lower in winter than summer with seasonally changing intercepts between -33.6‰ (JJF) and -27.7‰ (MJJ). Making the simple assumption that natural gas and gasoline are the only major fuel sources, natural gas would contribute ~45% to emissions in winter and ~3% in early summer, which is lower than the downscaled Local Emissions Inventory (57% in winter and 20% in summer). Mean δ¹⁸O showed a clear annual cycle with higher monthly intercepts in summer (-15.6 ‰) than in winter (-20.2 ‰). Various mixing models were explored to characterize contributions to the urban enhancement and analyzed in combination with meteorological data.