Sensitivity to morphology in urban boundary layer modeling at the mesoscale

The representation of urbanized surfaces in a mesoscale numerical weather prediction (NWP) model poses a formidable challenge in dynamics and thermodynamics. The heterogeneity of surface type generally precludes explicit solution, requiring a parameterization of physical processes to understand the meteorology of the urban boundary layer. These processes include radiative transfer, drag, anthropogenic heat sources as well as the transport of turbulent quantities of heat and momentum and their exchange with the inertial layer above. Mesoscale modeling of urban morphology often requires information about roughness element (building) dimensions and street width (building density) and is especially pertinent to the parameterization of these physical processes. This study seeks to evaluate the sensitivity of key atmospheric vector and scalar quantities in the urban boundary layer to small perturbations in urban morphological parameterizations using the NCEP/NCAR Weather Research and Forecasting Model (WRF) and its newly coupled urban canopy model (UCM) for mesoscale forecasting.

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 typically heterogeneous urban morphology contributes substantially to sub-grid scale variation in the description of the model surface. A common response to this problem is to classify surface aggregates according to a dominant structural form such as “high-intensity residential” or “commercial.” This allows for the creation of a set of urban land surface types, each with distinct physical and thermal characteristics. However, sub-grid scale variations in morphology remain. Many relevant meteorological quantities, such as urban canopy wind speed, urban boundary layer depth, turbulent kinetic energy strength, vertical sensible heat flux, street canyon temperature, etc., all derive from parameterizations influenced by the urban morphology. Of crucial interest to urban environmental modeling is pollutant transport, of which a viable short-term forecast depends on accurate model estimates of these quantities. Thus, urban air quality studies are inevitably subject to the uncertainty caused by sub-grid scale variations in urban morphology representations. Current remote sensing and GIS techniques can provide a very accurate, high-resolution depiction of the static urban surface. However, the integration of such data into an operational NWP model may be costly and spatially limited. It is important to understand the sensitivity of urban meteorology to uncertainties in the mesoscale urban morphology to demonstrate the limitations in deterministic NWP forecasts of the urban boundary layer for meteorological and air quality concerns.

In this experiment, the principal morphology parameters driving WRF-UCM: mean structural height, street width, the fraction of surface urbanized and the mean structural drag coefficient are perturbed in a controlled urban environment. The perturbations reflect common sub-grid scale variations, drawn to represent a broad array of urban surface arrangements. Replicates combine to form an ensemble of solutions demonstrating the non-linear response of the urban boundary layer meteorology to simple sub-grid scale morphology variations in WRF-UCM. Further investigations probe the response by time-of-day and the divergence of solutions in time versus control runs. Results provide a preliminary quantitative estimate of the variation that can be expected in the boundary-layer meteorology manifested through sub-grid scale variation in the urban morphology. Such quantification of uncertainty may be of particular interest to urban air quality modelers interested in point and line source dispersion when representing a city with mesoscale urban canopy models or other models lacking precise, high-resolution morphological analyses.