Carbon Dioxide Storage in an Urban Environment

Urban carbon dioxide emissions may be under-estimated by observations because CO₂ storage (ΔCs) is usually poorly characterised in this environment compared to vertical fluxes. Measurements that could improve the understanding of related processes are still scarce, not least due to the complexity of the urban surface structure and the heterogeneous distribution of anthropogenic sources. Further, due to issues of cost and access it may be necessary to trade e.g. temporal resolution for spatial by using one gas analyser to measure multiple heights via a switching valve array (the profile method). Alternatively, certain areas may not be accessible (e.g. within the street canyon) and it will be necessary to estimate ΔCs using data from one height only (the flux correction or single height method). The current study addresses these methodological challenges: the effect of reduced spatial and temporal resolution on calculated ΔCs is evaluated by comparison to a ‘best practice' (BP) data set. Additionally, the observations presented allow for evaluation of the theoretical hypothesis from Finnigan (2006), that the magnitude of ΔCs underestimation varies predictably with the ratio of the measurement interval to the integral timescale of turbulence.

BP ΔCs data are compared to ΔCs calculated from a 10-height vertical profile spanning from 6.5 to 46.4 m above ground level (0.31 to 2.25 times mean building height respectively), and to ΔCs calculated from high temporal resolution CO₂ concentration data collected at one height (46.4 m) as part of an eddy covariance system. All data presented were collected in central London, during 2013. The BP system comprises two open path gas analysers measuring at 10 Hz, one within the inertial sublayer and one at half-canyon height. The profile system consists of two closed path gas analysers connected to arrays of three-port valves, allowing air to be drawn continuously from each sample location either to the gas analyser or a waste pump. Co-located inlets are used to assess the stability of each closed path gas analyser with each other and with the open path gas analysers. Non ideal measurement systems are found to substantially underestimate ΔCs. The magnitude of the under-estimation is related primarily to the response time of the sensor and secondarily to the sampling interval of the measurements. The under-estimation is predictable and empirical correction factors are presented.