Comparing an urban metabolism model to long-term CO2 eddy-covariance measurements

Urban areas cover only 2% of the global land surface yet account for about 75% of all direct anthropogenic carbon-dioxide (CO₂) emissions. Energy conservation in urban settlements alone is crucial but insufficient to meet challenging emission reduction aspirations and targets. As much as 50% of the emissions are attributable to urban form choices - notably: density, land use mix, spatial pattern, building type and vegetation. To meet emission-reduction goals we will need appropriate decision support methods to evaluate the impact, and opportunity, of various future planning and design scenarios on all relevant sources and sinks of CO₂.

Although carbon-dioxide emissions of residential land-use at the community-level are typically dominated by fossil-fuel combustion from transportation and buildings, unaccounted in most studies are respiratory releases of CO₂ from highly managed and irrigated soils, waste decomposition, and the human metabolism as well as sequestration by urban vegetation. Our urban metabolism approach integrates all relevant net CO₂ emissions with the goal to model spatial (maps) and temporal (seasonal) output at the neighborhood-scale under current conditions and for projections of future development scenarios.

The presented urban metabolism approach uses a combination of scaling down and scaling up to quantify the net CO₂ emissions. Model input data includes automated urban object classifications based on LIDAR for building form characteristics and leaf area index estimates, optical remote sensing data for land cover characteristics. Additional data is extracted from energy utility providers, traffic counts, census data and measured radiation and climate data. Emission modeling is done by various separate model components namely a building energy model, a transportation model, and an urban vegetation and soil model to estimate GPP and ecosystem respiration.

A primarily residential area of 4 km² in Vancouver, BC has been selected in a case-study supported by Natural Resources Canada. The area is chosen to overlap with the source area of a micrometeorological flux tower. At this tower, continuous, year-long carbon-fluxes were measured as part of the CFCAS network 'Environmental Prediction in Canadian Cities'. The flux tower data is aggregated into different wind sectors that are conditionally averaged to allow a comparison between modeled and measured net emissions for different subsets of the urban surface.

Our presentation will highlight the use of aggregated carbon flux measurements to validate neighborhood carbon models, but also underline the conceptual differences between local and external emissions, and the problems associated with the modeling of mobile sources (transportation, humans) in a fixed frame of reference.