Simulations of dispersion through an irregular urban building array
Michael Pirhalla,David Heist,Steven Perry,Wei Tang,Lydia Brouwer
DOI: https://doi.org/10.1016/j.atmosenv.2021.118500
IF: 5
2021-08-01
Atmospheric Environment
Abstract:<p>Following the release of a harmful substance within an urban environment, buildings and street canyons create complex flow regimes that affect dispersion and localized effluent concentrations. While some fast-response dispersion models can capture the effects caused by individual buildings, further research is required to refine urban characterizations such as plume channeling and spreading, and initial dispersion, especially within the presence of a nonhomogeneous array of structures. Field, laboratory, and modeling experiments that simulate urban or industrial releases are critical in advancing current dispersion models. This project leverages the configuration of buildings used in a full-scale, mock urban field study to examine the concentrations of a neutrally buoyant tracer in a series of wind tunnel and Embedded Large Eddy Simulation (ELES) experiments. The behavior, propagation, and magnitude of the plumes were examined and compared to identify microscale effects.</p><p>After demonstrating excellent quantitative and qualitative comparisons between the wind tunnel and ELES via lateral and vertical concentration profiles, we show that a nonlinear least squares fit of the Gaussian plume equation well represents these profiles, even within the array of buildings and network of street canyons. The initial plume dispersion depended strongly on the structures immediately adjacent to the release, and consequently, the near-surface plume spread very rapidly in the first few street canyons downwind of the source. The ELES modeling showed that under slightly oblique incoming wind directions of 5º and 15º, an additional 5º and 14º off-axis channeling of the plume occurred at ground level, respectively. This indicates how building structures can cause considerable plume drift from the otherwise expected centerline axis, especially with greater wind obliquity. Additionally, AERMOD was used to represent the class of fast-running, Gaussian dispersion models to inform where these types of models may be usefully applied within urban areas or groups of buildings. Using an urban wind speed profile and other parameters that may be locally available after a release, AERMOD was shown to qualitatively represent the ground-level plume while somewhat underestimating peak concentrations. It also overestimated the lateral plume spread and was challenged in the very near-field to the source. Adding a turbulence profile from the ELES data into AERMOD's meteorological input improved model estimates of lateral plume spread and centerline concentrations, although peak concentration values were still underestimated in the far field. Finally, we offer some observations and suggestions for Gaussian dispersion modeling based on this mock urban modeling exercise.</p>
environmental sciences,meteorology & atmospheric sciences