Eddy-covariance isofluxes of carbon dioxide measured over an urban surface in Vancouver, Canada

Stable isotope ratios of carbon dioxide (CO2), in particular δ13C, separate between combustion of gasoline (δ13C = -27.3‰), diesel (δ13C = -28.8‰) and natural gas (δ13C = -41.6‰). For example 'Keeling plots' of δ13C have been used to partition the enhanced urban CO2 using simultaneous background data. We propose an alternative approach to directly characterize and partition urban emissions at fine temporal and spatial scales by measuring isofluxes using the eddy-covariance (EC) approach. EC fluxes of 12CO2 and 13CO2 were measured continuously on a 30m-tall tower located in Vancouver, Canada (Fluxnet ID "Ca-VSu) between March and July 2016. The tower is located near an intersection of arterial roads in a residential area characterized by building emissions from natural gas furnaces. The goal is (1) to determine δ13C of the urban emission mix at any given time, and (2) to use this information to partition total measured CO2 fluxes into combustion of natural gas and the mix of gasoline+diesel (g+d, assuming an a priori ratio). Respiratory fluxes have been shown to be small and are neglected. Fluxes were measured using a sonic anemometer (CSAT-3, CSI) and a tuneable diode laser absorption spectroscopy (TDLAS) system scanning absorption lines at 2308.171 (13CO2) and 2308.225 cm-1 (12CO2) at 10 Hz (TGA200, CSI). Every 10 minutes, the TDLAS was calibrated against three tanks, referenced against NOAA-ESRL/INSTAAR, UoC standards. Isoflux ratios (F13C) were calculated as F13C = 1000 [(w'13CO2' / w'12CO2') / RVPD-CO2) - 1]. The average measured F13C was -32.67‰, which corresponds to a mixture of 32% natural gas and 68% g+d. This matches well against a top-down inventory for the City of Vancouver, which estimates 36% natural gas and 64% g+d. Fluxes were correlated with traffic counts, with higher fluxes during daytime. F13C increased with increasing flux strength because the highest emissions originated from traffic intersections (g+d). F13C was higher during daytime with -30.8‰ (23% natural gas) and lower at nighttime with -33.6‰ (43% natural gas) when traffic emissions were lower. Although the TDLAS is characterized by noise in the order of 1.5‰, the strong fluxes recorded (up to 100 µmol m-2 s-1) make it possible to determine realistic isoflux ratios that can be used to precisely constrain urban δ13C.