Atmospheric Environment 39 (2005) 4729–4743 Dispersion within a model urban area Jing Yuan, Akula Venkatram à Department of Mechanical Engineering, University of California, Riverside, CA 92521, USA Received 23 November 2004; accepted 20 April 2005 Abstract This paper analyzes the data collected in a field study conducted in Dugway Proving Ground, Utah, from 17 to 26 July 2001. The experiment was designed to simulate a source in an urban area modeled at a scale of roughly 1:10. The model urban canopy was constructed with 55-gallon drums laid out in a 5 by 9 array. Propylene (C 3 H 6 ), the tracer, was released and sampled at 3 receptor arcs both within the barrel array and over flat terrain. Turbulence, velocity, and temperature were measured with sonic anemometers. A comparison of observations made with and without the obstacle array indicated that the obstacle array significantly increased lateral and vertical dispersion. The arc maximum surface concentrations observed within the array were well explained by a dispersion model based on that of Van Ulden [1978. Atmospheric Environment 12, 2125–2129] and modified by Venkatram [2004. Atmospheric Environment 38, 1337–1344] to use on-site measurements of mean wind and turbulence. The major conclusion of this study is that estimating dispersion within the urban canopy requires flow information below the canopy top. r 2005 Elsevier Ltd. All rights reserved. Keywords: Dispersion; Obstacle array; Model urban area; Field experiment; Tracer study; Plume spread; Dispersion model; Data analysis 1. Introduction We can understand features of dispersion within the urban canopy layer through controlled experiments in physical models of urban areas. Such experiments, conducted by Davidson et al. (1996) and Macdonald et al. (2000) and others have provided useful informa- tion on plume behavior within an array of obstacles that mimic an urban canopy. These studies found that plume spreads in the model canopy are generally larger than that over flat terrain. But they have not related these spreads to turbulence within the canopy. In the study reported in this paper, we examine the feasibility of estimating plume spreads and concentrations in a model canopy using flow and turbulence measured within the canopy. The dispersion model used to estimate concen- trations is similar to that applied in flat terrain. The question is whether the model still applies within an obstacle array if appropriate flow inputs are used. The data required to evaluate the model were obtained from an experiment conducted in Dugway Proving Ground, Utah, which was designed to simulate a source in an urban area modeled at a scale of roughly 1:10. 2. Experiment design and implementation The experimental design was based on the assumption that the horizontal and vertical scales of turbulence that govern dispersion within the urban canopy are propor- tional to the dimensions of the buildings (Macdonald et al., 1998). This allowed us to use obstacles to simulate ARTICLE IN PRESS www.elsevier.com/locate/atmosenv 1352-2310/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2005.04.021 à Corresponding author. E-mail address: akula.venkatram@ucr.edu (A. Venkatram).