In the mid-20^th century, transportation agencies began constructing sound walls around major highways as suggested by Title 23 of the U.S. Code of Federal Regulations, Part 772, "Procedures for Abatement of Highway Traffic Noise and Construction Noise”. Although the primary purpose of sound walls is to reduce the noise levels in residential areas next to highways, they can have a significant impact on the dispersion of traffic related pollutants. Therefore, since early 2000, U.S. Environmental Protection Agency (U.S. EPA) initiated several projects to address the impact of sound walls and surrounding vegetation on the dispersion of vehicular emissions. The results from these projects unanimously showed that incorporation of roadside structures can cause 50% reduction in the near road concentrations as compared to an unobstructed roadway⁴. Although these studies provide a thorough insight on the air quality impact of sound barriers, the question that remained unanswered is how to reflect their impact into currently available dispersion models?  Over the past 5 years, there have been some efforts to model the dispersion affected by roadway structures. Bowker et al.⁵ used Quick Urban and Industrial Complex model⁶ (QUIC) developed by Los Alamos National Laboratory to describe the results from I-440 field study. Another modeling practice was the work done by Steffens et al.⁷ where they introduced the Comprehensive Turbulent Aerosol dynamics and Gas chemistry model (CTAG) to estimate the concentrations in Idaho Falls field study⁴. Although all these studies showed some abilities in describing the impact of noise barriers, none of them provides a generic solution. Both QUIC and CTAG are numerical based models and necessitate computational resources that can become impractical for exposure analysis. This article is presenting another effort conducted by researchers at University of California Riverside (UCR) to develop a simple dispersion model that can reflect the impact of roadside structures on the near roadway concentrations, and at the same time is computationally efficient.