Scalar Dispersion from Point and Area Sources in a Realistic Urban Environment

Accurate modeling of pollutant dispersion within and above urban canopies is critical to properly predict air quality in urban environments. Using large eddy simulation (LES), we investigate scalar dispersion from point and area sources in a typical North American neighborhood using topography and foliage density derived from airborne LIDAR scans with 1 m resolution in Vancouver, BC, Canada. The added drag force due to trees is parameterized in the LES as a function of the leaf area density (LAD) profile. Conversely, drag from buildings is accounted for using a direct forcing approach immersed-boundary method. The scalar advection-diffusion equation is discretized in a finite-volume framework, and accurate mass conservation is enforced through a recently developed Cartesian cut cell method. Simulations are performed with trees for different values of LAD, representative of summer and winter conditions, as well as a case without trees. The effects of varying mean wind direction (derived from observed wind climatologies) on dispersion patterns are also considered. Scalar release locations in the LES are informed by spatially distributed measurements of carbon dioxide concentration; CO₂ is used as a tracer for fossil fuel emissions, since source strengths are well-known and the contribution from biological processes in this setting is small (<10%). A detailed emissions inventory allows us to consider the separate and combined effects of emissions from vehicle exhaust and heating of buildings on air quality. The effects of leaf area density, source height, and wind direction on scalar statistics including the growth of the mean concentration plume and the fraction that escapes the urban canopy layer will be considered. Previous studies have demonstrated that the presence of trees strongly modifies sweep and ejection patterns for the momentum flux; here we consider the related issue of how vegetation influences coherent structures responsible for scalar transport.