A Scalable Framework for Modeling Near-Road Air Pollution Using Representative Road Types and Orientation-Informed Emission Profiles

DEEPSHIKHA OLA, Huyen Le, Jeffrey Bielicki, Andrew A. May, The Ohio State University

     Abstract Number: 414
     Working Group: Urban Aerosols

Abstract
Traffic emissions are a major contributor of particulate matter (PM), nitrogen oxides (NOx), and carbon monoxide (CO) in urban areas While county-level emission inventories are useful for regional-scale assessments of the impacts of these traffic emissions, they lack the spatial granularity required to evaluate air quality and exposure at the street or neighborhood level. This limitation becomes especially critical within urban microenvironments, where traffic conditions can vary dramatically across locations.

To address this gap, we propose a practical, data-driven framework that constructs a high-resolution emissions dataset using a set of representative roads. Rather than relying on statistical clustering, which yielded limited differentiation, we categorized roads based on two key parameters: average traffic volume and speed. Traffic volume was divided into three categories (low, medium, and high), and speed into five bins (≤25, 26 and ≤35, 36 and ≤45 mph, 46 and ≤55, and >55mph), forming a comprehensive matrix of road types. Using 2023 vehicle activity data from the StreetLight database, project-level emission factors will be estimated through the Motor Vehicle Emission Simulator (MOVES). These emission factors across the representative road types will be integrated with the R-LINE dispersion model to simulate pollutant concentrations across the community level. To enhance transferability, we will further develop a generalized emissions-concentration inventory by incorporating wind direction relative to road orientation across 16 directional bins. The accompanying code will demonstrate how we combine the representative roads and wind directions to quantify traffic emissions in a complete road network in Columbus, OH.

Our work enables the generation of concentration gradient profiles for each road type, allowing estimation of near-road air pollution in different communities using local road network that fall into the same representative categories. This dual approach, combing road classification with directional dispersion, strikes a balance between spatial detail and computational efficiency, offering a robust, scalable tool for community-level air quality modeling and exposure analysis across U.S. cities. By leveraging this data-driven approach, policymakers can better identify neighbourhoods that require stricter regulations and targeted mitigation strategies to improve air quality.