Simulations of dispersion through an irregular urban building array
Introduction
Densely populated urban areas, with their complex layouts of buildings and streets, pose considerable challenges in assessing potential human exposure and environmental risks following a hazardous airborne effluent release. Interconnected streets, avenues, and intersections that separate rows and groups of buildings create “street canyons” that help to channel contaminants throughout the roadway system (DePaul and Sheih, 1986). This effect can increase ground-level concentrations by restricting the entrainment of fresh air from aloft (Belcher et al., 2015; Soulhac et al., 2013), spreading effluent through the city by channeling the flow from the expected wind direction axis (Marucci and Carpentieri, 2020), altering the boundary layer wind speed profile with height (MacDonald, 2000), and generating additional mechanical turbulence due to the building structures (Britter and Hanna, 2003). Researchers have pursued urban flow and dispersion problems for several decades because a better characterization of these processes, and subsequent improvements to the models that estimate their impacts, could save lives and better inform emergency action following an unexpected release. While the current knowledge of these urban micrometeorological concepts has greatly improved, advancement is needed to simulate these effects accurately using fast-response dispersion models, given the complicated features of urban environments. Gaussian-type dispersion models may offer valuable, widespread support in operational settings due to their rapid simulation results (Philips et al., 2013) and fewer input and postprocessing requirements. Even though these types of models produce steady-state solutions and do not incorporate the physics of flow in street canyons and around individual buildings, their predictions may be helpful when limited information is known about the release scenario. This is particularly common immediately following an unexpected release event.
A number of meteorological wind tunnel studies have examined flow and dispersion in simulated urban environments, many of which have incorporated simplified, symmetrical, or rectangular shaped arrays of buildings that result in systematically oriented avenues and street canyons (e.g., Marucci and Carpentieri, 2020; Fuka et al., 2018; Hertwig et al., 2018; Kanda and Yamao, 2016; Brixey et al., 2009; Heist et al., 2009, and others). Computational Fluid Dynamics (CFD) modeling studies (e.g., Castro et al., 2017; Kumar et al., 2015; Boppana et al., 2014, and others) have offered detailed descriptions of a plume's behavior among a variety of urban-like building arrays. These types of simulations are computationally intensive and comparisons against field or laboratory measurements are desirable. Full-scale field studies such as the Mock Urban Setting Test (MUST) (Biltoft, 2001) are also invaluable for understanding urban flows, but the immense logistical challenges and financial burdens associated with these projects make them rather rare.
This project was designed to build upon previous research from the large-scale, mock urban field study called Jack Rabbit II (JRII) (Fox et al., 2021; Nicholson et al., 2017). The study simulated dispersion within an irregular urban area, which formed street canyons and complex turbulence between and in the lee of buildings. JRII was a joint field study between the US Department of Homeland Security (DHS) and several governmental and private entities. It was designed to fill emergency response-related knowledge gaps from large, accidental, or intentional releases of dense gas within an urban or industrial-like setting (Nicholson et al., 2017). JRII took place between 2015 and 2016 at the US Army Dugway Proving Ground (DPG) test facility in the Utah desert (Gant et al., 2018) where a sequence of nine controlled dense gas releases were performed both within and in the absence of an array of 80 CONEX shipping containers. In contrast to other mock urban-style studies such as MUST and Dispersion of Localised Releases in a Street Network (DIPLOS) (Castro et al., 2017), the JRII array featured a staggered arrangement of buildings in two general sizes. While most of the building geometry was rather symmetrical, the study area also contained a single taller structure that was two CONEXs wide and stacked three high. The building area fractional density was representative of low-density suburban or light industrial areas. The JRII project permitted the study of the release mechanism and source behavior, toxic inhalation hazards, and atmospheric dispersion to improve chemical hazard dispersion modeling. The field study also helped determine safe downwind concentration distances and the further development of efficient emergency response and lifesaving measures (Nicholson et al., 2017).
After the first phase of dense gas experiments, a secondary experiment was designed to study the wind flow, turbulence, and micrometeorology within the mock urban environment (Pirhalla et al., 2020). Thirty sonic anemometers were deployed at various heights and locations around the tall building and adjacent shorter CONEXs to promote an analysis of complex wind flow regimes around the obstructions. It was found that the flow within the canopy was characterized by three distinct flow regimes, which could vary strongly only a few meters apart depending on the location of the sonic anemometer relative to an obstruction. In addition, atmospheric stability tended toward neutral within the canopy compared to the approach boundary layer flow, which impacts effluent dispersion in the street canyon.
The JRII field study was described in detail by Fox et al. (2021). The reader is also referred to Mazzola et al. (2021), Gant et al. (2021), Chang et al. (2021), Pirhalla et al. (2020), Hanna (2020), and several other articles published within the recent JRII special issue for a more complete description of the field study, release trials, dispersion results, and findings from the special sonic anemometer study. Although a gas was released during the JRII experiments, this current project is also applicable to fine particle-based releases (such as the threat of a wide-area Bacillus anthracis [anthrax] event), since any effluent will disperse within the wind's streamline flow.
To complement the field study within a controlled experimental setting, we used the JRII building layout as a starting point and then critically analyzed the dispersion patterns of a neutrally buoyant release through a series of wind tunnel experiments. To promote a closer inspection of the flow and dispersion and to provide an even denser evaluation dataset for comparison to simpler models, we also ran several CFD simulations of the wind tunnel experiments. Finally, we compared the wind tunnel and CFD results to the performance of a steady-state Gaussian-type dispersion model and then offer suggestions to reproduce the plume more effectively within an urban setting.
Section snippets
Experimental design and case descriptions
The dispersion component of the JRII field study focused entirely on dense gas releases. Additionally, only a single arrangement of CONEX buildings was deployed for all release scenarios. Laboratory and computational simulations offer the flexibility to modify the building array orientation, change characteristics and locations of the effluent releases, and to specify the approach wind directions. In this project, a series of six experimental cases were simulated with an extremely fine-scale
Analysis and results
Given the excellent performance of the ELES modeling against the wind tunnel measurements for the perpendicular incoming wind cases, we elected to use the ELES simulations to examine the behavior of the plume in this mock urban area. In addition to the performance statistics, we also examined lateral and vertical concentration profiles and assessed the extent to which the plumes were Gaussian. The Gaussian fits were then used to generate other plume characteristics, including the lateral and
Gaussian dispersion modeling
We next considered the use of a Gaussian dispersion model to estimate concentrations because of the Gaussian nature of the concentration profiles observed throughout this building array. During emergency release situations, Gaussian dispersion models could be useful due to their generally available input requirements, fast runtimes, and minimal postprocessing steps (US EPA, 2020). To represent the Gaussian class of models, AERMOD (v19191) (Cimorelli et al., 2005; https://www.epa.gov/scram) was
Concluding remarks
The goal of this multipart study was to enhance the understanding of flow and dispersion in a mock urban area using a combination of wind tunnel and modeling experiments to supplement the previously completed JRII field study. A harmful accidental or intentional effluent release (from either gases or fine particles) in urban areas may have a greater impact on human exposure due to higher population densities, channeling flows between buildings and within street canyons, and complex atmospheric
Disclaimer
The U.S. Environmental Protection Agency, through its Office of Research and Development, funded and directed the research described herein. It has been subjected to the Agency's review and has been approved for publication. Note that approval does not signify that the contents necessarily reflect the views of the Agency. Mention of trade names, products, or services does not convey official EPA approval, endorsement, or recommendation.
CRediT authorship contribution statement
Michael Pirhalla: Data curation, Conceptualization, Investigation, Methodology, Software, Formal analysis, Data curation, Validation, Writing – original draft, Project administration. David Heist: Data curation, Conceptualization, Investigation, Methodology, Software, Formal analysis, Data curation, Validation, Writing – review & editing, Project administration, Supervision. Steven Perry: Conceptualization, Methodology, Formal analysis, Investigation, Writing – review & editing, Project
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This effort was supported by the US Environmental Protection Agency's Homeland Security Research Program (HSRP) and Air and Energy (A&E) Program within the Office of Research and Development (ORD). The wind tunnel data collection work was supported by Jacobs Technology, Inc. (JTI) Contract EP-C-15-008. The ELES modeling work was completed through the High-End Scientific Computing Support Services Task Order GSQ0017AJ0019. The authors would like to acknowledge Laurie Brixey, Brittany Thomas,
References (59)
- et al.
JRII Special sonic anemometer study: a first comparison of building wakes measurements with different levels of numerical modelling approaches
Atmos. Environ.
(2021) - et al.
Development of a quality-controlled chlorine gas concentration data base for the Jack Rabbit II field experiments
Atmos. Environ.
(2021) - et al.
Measurements of wind velocities in a street canyon
Atmos. Environ.
(1986) - et al.
Numerical analysis of pollutant dispersion around elongated buildings: an embedded large eddy simulation approach
Atmos. Environ.
(2018) - et al.
DRIFT dispersion model predictions for the Jack Rabbit II model inter-comparison exercise
Atmos. Environ.
(2021) - et al.
Dense gas dispersion model development and testing for the Jack Rabbit II phase 1 chlorine release experiments, Atmos
Environ. Times
(2018) Meteorological data recommendations for input to dispersion models applied to JR II trials
Atmos. Environ.
(2020)- et al.
A baseline urban dispersion model evaluated with Salt Lake City and Los Angeles tracer data
Atmos. Environ.
(2003) The design of spires for wind simulation
J. Wind Eng. Ind. Aerod.
(1981)- et al.
Passive scalar diffusion in and above urban-like roughness under weakly stable and unstable thermal stratification conditions
J. Wind Eng. Ind. Aerod.
(2016)