The representation of urbanized surfaces in a mesoscale numerical weather prediction (NWP) model poses a formidable challenge in dynamics and thermodynamics. Inhibited from explicit resolution of the urban boundary layer (UBL) flow by scale, the model must parameterize physical processes, requiring detailed information about the form and design (i.e., morphology) of the site. The morphology consists of obstacle (building) dimensions and distribution, street width (building density) and fractional coverage by anthropogenic materials and natural vegetation.

Uncertainty in the initial and boundary conditions of an NWP model may derive from sub-grid scale variation, incomplete model physics, measurement error or other sources. The morphology of the urban surface is typically heterogeneous and asymmetrical, thus contributing to sub-grid scale variation in a model description of the urban surface. The spatial distribution of obstacles of varying scale and orientation with respect to the large-scale wind direction can have potentially significant implications on the adjustment of the wind to the presence of the urban canopy. The distribution can also affect the local radiative budget by influencing the position and duration of shadows and the magnitude of turbulent heat fluxes. A common approach to mesoscale modeling of the urban surface involves assignment into surface aggregates according to a dominant structural form such as “high-intensity residential” or “commercial” to match the grid scale of the accompanying natural land surface geography data. This method constrains all urban surfaces to fit a universal set of urban land surface types, each with static physical characteristics (morphology parameters). This approach does not account for sub-grid scale variation within the model grid cell, however, where obstacles of varying dimension and distribution may be present. Such variation can contribute to change in the wind profile, momentum transfer and energy balance in and around an urban area. This study seeks to diagnose the sensitivity of mesoscale UBL meteorology to uncertainty in the urban morphology using the Weather Research and Forecasting (WRF) model and its urban canopy model (UCM).

The study consists of two parts: (1) Testing the model response to perturbations in the urban morphology parameters and (2) Testing the model response to increasing resolution of the urban surface from the mesoscale to the neighborhood scale. The latter tests demonstrate how increasing the spatial variability of the urban morphology (i.e., improved representation of the true urban surface) manifests in the local meteorology. The study examines the effects of morphological perturbations and spatial variability on the urban heat island circulation, turbulence kinetic energy, surface energy balance and meteorological variables in the urban boundary layer. The sensitivity of the local meteorology to urban morphology has applications to regional scale pollutant dispersion research using mesoscale NWP models. Although the UCM approach to UBL modeling does not capture urban canyon sweeps and ejections, nor adequately resolves the canyon wind profile, certain chemical species' transport rates and concentrations may be subject to this meteorological sensitivity depending on the scale of urban surface representation.