Emission exposure optimum for a single-destination dynamic traffic network

Highlights

• We develop a dynamic model to describe human exposure to vehicular emissions.

• A free-flow optimal solution exists that minimizes the total emission exposure.

• The health impact of vehicular emissions can be reduced by managing traffic flow.

• There is a tradeoff between total time cost and total emission exposure.

Abstract

In this paper, an integrated dynamic framework is applied to model human exposure to vehicular emissions. The framework contains three components: a time-dependent macroscopic emission model, a dynamic traffic flow model, and a line source Gaussian plume model to capture the spatial dispersion of emissions. A human exposure-based dynamic system-optimal problem is then formulated to minimize system-wide emission exposure. An illustrative network and the Sioux Falls network are used to test the proposed framework and to examine how traffic dynamics, population characteristics, and meteorological conditions affect emission exposure at various network locations. The results provide evidence that regional emission exposure can be reduced by dynamically managing network traffic. At least under the specific network setup, the paper provides a counterexample to the common conception that traffic congestion and emissions (exposure) can be minimized simultaneously, which implies that traffic system operators must properly leverage the tradeoff between congestion and emissions.

Introduction

Recently, the problem of urban air pollution has received a significant amount of attention. According to the World Health Organization, outdoor air pollution contributes to more than 4.2 million deaths annually (WHO, 2016). Studies have shown that exposure to vehicular emissions can lead to health problems such as headaches, respiratory diseases, and cardiovascular diseases (Xing et al., 2016, Burnett et al., 2018, Burns et al., 2020), even when the concentrations of pollutants are below government-mandated thresholds (Di et al., 2017). Such evidence makes it imperative to understand the relationship between urban traffic emissions and human health.

A variety of traffic emission models have been developed in the transportation field, with particular emphasis on estimating emissions as functions of traffic conditions. From a transportation modeling perspective, efforts have been made to design eco-routing (Luo et al., 2016, Alam et al., 2018, Zeng et al., 2020, Djavadian et al., 2020) or traffic control (He et al., 2015, Long et al., 2018, Song et al., 2020) strategies to reduce system-wide vehicular emissions, and controlling/managing emissions from policy-maker perspectives (Ma et al., 2015, Rodriguez-Roman and Allahviranloo, 2019, Kamishetty et al., 2020). Despite these research efforts, a gap still exists between the understanding of traffic-related emissions and their impacts on human health. The most recent research trend is to encapsulate traffic emission production, dispersion, the impacts of emission exposure on human health, and their interactions with trip behavior using integrated frameworks (Pinto et al., 2020). For example, Zhang and Batterman (2013) combined the Bureau of Public Roads (BPR) function with the MOVES model and MOBILE 6.2 to estimate emission factors and used the California Line Source Dispersion Model (CALINE4) to calculate the emission concentrations close to roadway segments. An incremental analysis was conducted to show the impacts of traffic volume on human exposure and health risks. Alzuhairi et al. (2016) integrated traffic simulation models and CALINE4 to study the effects of traffic emissions in Chicago school zones. Sun et al. (2018) developed an integrated framework to study multiple equilibrium route-choice behavior by considering the negative impacts of vehicular emissions on human health. In particular, a portion of travelers are assumed to be environmental advocates, whose trip-making decisions are based on generalized cost functions that consider travel time and emission exposure. The results show that Pareto improving solutions could be achieved if more travelers were to become environmental advocates. Luo et al. (2020) presented a holistic approach that combined the regional travel demand model, EMFAC model and R-line dispersion model to estimate near-road air pollution concentration. Human exposure to traffic-related air pollution was then considered when planning bicycle routes.

To date, most relevant studies have focused on static cases, but few have integrated traffic dynamics, emission models, dispersion models and human exposure into a coherent time-dependent framework. Lin and Ge (2006) adopted the cell transmission model to capture the spatial–temporal dynamics of traffic flow. They then applied the Gaussian dispersion model to estimate emission concentrations close to an urban street section. Aziz and Ukkusuri (2012) integrated a CO emission model into a dynamic traffic assignment (DTA) model and compared the differences between total travel time and total emissions at different congestion levels. They then used the solutions obtained from the DTA model as inputs to the Gaussian dispersion model to calculate the resulting CO concentrations. Samaranayake et al. (2014) applied a Lighthill-Whitham-Richards (LWR)-based ensemble Kalman filter model to estimate average traffic speed based on real-time traffic state measurements. They then used the estimated speed, together with real-time weather information, to query the emission rates using the EMFAC database. A simplified version of CALINE3 was implemented to estimate emission concentrations in the San Francisco Bay Area. From a transportation network-modeling perspective, it is crucial to consider the location, population density, trip behavior and other human characteristics when evaluating the regional health impacts of traffic emissions (Shekarrizfard et al., 2016, Tayarani and Rowangould, 2020).

To bridge these research gaps, in this study, an integrated dynamic framework is developed to evaluate the regional health impacts of traffic-related emissions. A dynamic traffic flow model named “double-queue,” a time-dependent macroscopic emission model, a line source dispersion model, and regional population parameters are considered in a human-centric framework (see Fig. 1 for a schematic of the framework). A network dynamic system-optimal problem that considers human exposure to vehicular emissions (called DSO-HE) is formulated to minimize network-wide emission exposure. An illustrative network and the Sioux Falls network are used to test the proposed model. The existence of a free-flow optimal solution that minimizes the system emission exposure while maintaining the system in a free-flow state is validated. Under certain meteorological conditions, we show that the emission exposure in a residential area with a high population-sensitivity parameter can be reduced by properly managing network traffic flow. The DSO-HE results are compared with the time cost-based dynamic system optimal (DSO-TC) model. The results show that the total time cost and emission exposure cannot be minimized simultaneously, which implies that traffic system operators must properly leverage the tradeoff between congestion and emissions. The proposed modeling framework can be applied to generic transportation networks to evaluate the health impacts of traffic emissions and facilitate green transportation decisions.

The remainder of this paper is organized as follows: Section 2 describes the methodology; Section 3 presents the numerical results of the experiments and managerial implications; and Section 4 provides concluding remarks.