AbstractObservations of carbon dioxide (CO₂) mixing ratios in the urban boundary layer (UBL) are rare, even though there is potential for such measurements to be used to monitor city-scale net CO₂ emissions. This work presents a unique dataset of CO₂ mixing ratios observed in the UBL above Vancouver, British Columbia, Canada, by means of a tethered balloon system over a continuous 24-h summertime period. Vertical profiles of CO₂ mixing ratios are found to vary according to UBL thermal structure and mechanical dynamics (development of convective and nocturnal boundary layers, vertical mixing from mechanical turbulence, horizontal advection from land–sea thermal breezes, and vertical entrainment). A box model is applied to quantify net city-scale surface emissions to the UBL volume using the measured rate of change of UBL CO₂ mixing ratios and estimated CO₂ advection and entrainment fluxes. The diurnal course of city-scale net emissions predicted by the model is similar to simultaneous local-scale eddy-covariance CO₂ flux measurements, although there are relatively large uncertainties in hourly model calculations of horizontal advection and vertical entrainment fluxes due to inputs of regional background CO₂ mixing ratios. Daily city-scale emissions totals predicted by the model (20.2 gC m⁻² day⁻¹) are 35% larger than those measured simultaneously on an urban local-scale eddy-covariance flux tower and are within 32% of a spatially scaled municipal greenhouse gas inventory. However, these methods are not expected to agree exactly because they represent different spatial source areas and include different CO₂ source and sink processes.