Enhancing Freight-Related Emission Dispersion Estimation by Integrating Line-Source and Point-Source Models

Abstract

This study introduces the Enhanced Truck Dispersion Framework, an innovative approach to modeling emissions from idling and moving trucks in urban environments. The framework integrates the Behavior, Energy, Autonomy, and Mobility (BEAM) model for the regional transportation model; the Emission Factors (EMFAC) model and the Motor Vehicle Emission Simulator (MOVES) for emission calculations; and the Research LINE (RLINE) source model with a grid-based point-source model for dispersion analysis. Field data collected at a distribution center in Ontario, California, U.S., is used to calibrate and validate the model, particularly for idling trucks. The study simulates traffic patterns in Inland Southern California, focusing on moving vehicles and idling trucks and based on real-world data. Results reveal significant pollutant concentrations near major highways and warehouse districts, highlighting the impact of both moving and idling truck emissions on urban air quality. This comprehensive approach provides valuable insights for policymakers and urban planners in developing targeted strategies to mitigate the environmental and health impacts of truck emissions in urban areas.

The environmental impact of vehicular emissions has been a topic of significant concern in recent years ( 1 ). As urban areas continue to grow, the number of vehicles on the road increases, leading to heightened levels of pollutants in the atmosphere ( 2 ). These pollutants, primarily originating from vehicle exhaust, have been linked to a range of environmental and health issues, including air quality degradation, respiratory diseases, and global climate change ( 3 ). Among the different types of vehicle, trucks, especially those idling or moving slowly in congested areas, are notable contributors because of their prolonged operational periods and specific emission characteristics ( 4 ).

While there is substantial research on emissions dispersion, existing models often emphasize highway emissions, potentially overlooking the significant contributions of trucks operating in warehouse areas. This focus on highways may lead to an underestimation of the overall impact of truck emissions in urban environments. Addressing this gap by integrating and comparing highway and warehouse emissions is crucial for a more comprehensive understanding of urban air pollution.

The objective of this study is to develop and validate a comprehensive framework for accurately predicting emission patterns from both moving and idling trucks in urban scenarios. By integrating advanced traffic simulation models with emission and dispersion models, the research seeks to provide a robust tool for analyzing and addressing truck-related air pollution in urban environments. The Enhanced Truck Dispersion Framework introduced in this study offers a novel approach to modeling truck emissions, combining macroscopic traffic simulators with emission calculation tools and multi-scale dispersion models. By using real-world data to calibrate and validate the framework, the study provides critical insights into the spatial distribution of pollutants.

The rest of this paper is organized as follows: The next section provides background and related work on macroscopic traffic simulators, truck emission models, and truck dispersion models, highlighting their significance in understanding vehicular emissions. This is followed by a detailed description of the Enhanced Truck Dispersion Framework development, including the integration of various models and the experimental setup for data collection. Subsequently, we present a case study applying the framework to evaluate the environmental impact of urban traffic on local air quality in Inland Southern California, U.S. The Results section details the model calibration and validation process, followed by an in-depth analysis of emission dispersion patterns for both moving and idling vehicles. The final section offers conclusions, recommendations for targeted mitigation strategies, and potential directions for future research.

Background and Related Work

Macroscopic Traffic Simulators

Traffic simulation tools replicate real-world traffic situations and are employed to analyze and enhance vehicular movement patterns. These simulators are generally divided into two main categories: microscopic, focusing on the actions of individual vehicles, and macroscopic, capturing the overall traffic flow and dynamics. In this research, we predominantly concentrate on macroscopic simulators. POLARIS, renowned for its in-depth traffic flow analytics, has become indispensable in the spheres of urban planning and infrastructure decision-making ( 5 ). SIGOP-II, another significant macroscopic model, boasts algorithms and frameworks adept at addressing complex traffic issues ( 6 ). The behavior, energy, autonomy, and mobility (BEAM) model stands out for its agent-centric design, which ensures a nuanced capture of the intricate dynamics of transportation systems by emulating individual agent behaviors ( 7 ). Furthermore, BEAM’s assimilation of discrete choice models facilitates the simulation of evolving human preferences in the face of emerging technologies and policies. Its capabilities extend to energy analysis, predicting the potential repercussions of electric vehicle integration and comprehending the challenges posed by fully autonomous vehicles on our streets. Additionally, BEAM’s proficiency in evaluating urban mobility and its synergy with other modeling tools positions it as a robust option in transportation analysis.

Truck Emission Models

Vehicular emission models are designed to estimate the emission rates and factors of motorized vehicles, considering various traffic conditions and driving cycles. These models play a crucial role in understanding the environmental impact of vehicular emissions, especially from heavy-duty trucks which are significant contributors to air pollution. Depending on the granularity of data and the level of detail, these models can be broadly classified into microscopic and macroscopic categories.

Microscopic models delve into the intricacies of individual vehicle movements, capturing the nuances of their behavior on a second-by-second basis. Examples of microscopic models include the Comprehensive Modal Emissions Model (CMEM) which was initially designed for light-duty vehicles but has been adapted for heavy-duty diesel vehicles, and the Passenger Car and Heavy-Duty Emission Model (PHEM) which provides insights into emissions from both passenger cars and heavy-duty vehicles ( 8 – 10 ). Additionally, VT-Micro is another microscopic model that estimates emissions based on detailed vehicle operation data. It uses second-by-second vehicle speed and acceleration inputs to calculate emissions and fuel consumption, making it highly precise for capturing the impact of various driving behaviors and conditions ( 11 ).

On the other hand, Macroscopic models provide a broader perspective, focusing on aggregate data and average behaviors. The Emission Factors (EMFAC) model, developed by the California Air Resources Board, is a prominent example. Tailored for California’s unique vehicular and environmental landscape, EMFAC combines emission factors from various vehicle classes, using data such as average speeds and vehicle miles traveled ( 12 ). COPERT, widely used in Europe, is another significant macroscopic model. Developed by the European Environment Agency and managed scientifically by the European Commission’s Joint Research Centre, COPERT calculates emissions for different vehicle categories and operational modes. It is recognized for its detailed and comprehensive data, supporting policy decisions, air quality modeling, and environmental assessments by providing emission factors for a wide range of pollutants ( 13 ).

The Motor Vehicle Emission Simulator (MOVES), developed by the U.S. EPA, is a versatile emission modeling system designed for both macroscopic and mesoscopic analysis which can also be applied to microscopic analysis under certain conditions. MOVES is widely used for national, state, and county-level emissions inventories. It estimates emissions for mobile sources, including cars, trucks, and buses, considering various pollutants such as greenhouse gases, criteria air pollutants, and air toxins. This versatility makes MOVES a powerful tool for broad, regional, and statewide emissions assessments, as well as detailed project-level analyses when integrated with other models ( 14 , 15 ).

In this study, we focus on the MOVES and EMFAC models, for several reasons. MOVES is chosen for its comprehensive and flexible approach to emissions modeling. Its ability to operate at multiple scales allows it to provide detailed emissions estimates for various scenarios, from regional to project-level analyses. MOVES incorporates real-world driving conditions and modal activity-based assessments, making it highly accurate and reliable for diverse applications. On the other hand, EMFAC is specifically designed to address California’s vehicular emissions, considering the state’s unique environmental policies and vehicle mix. EMFAC’s detailed regional insights and robust dataset make it an essential tool for analyzing emissions within California. By leveraging the strengths of both models, we can ensure a holistic and accurate analysis of heavy-duty vehicle emissions, making MOVES and EMFAC the preferred choices for this study.

Truck Dispersion Models

Understanding the dispersion of truck-related emissions is crucial for accurately assessing their impact on urban air quality. Various models have been developed to estimate the dispersion of emissions from trucks, each with specific applications and varying levels of precision. These models can be categorized based on their source type—point or line source—and their temporal behavior—steady-state or transient.

Within the realm of point-source models, the steady-state category includes models such as AERMOD, known for its prowess in handling complex terrains by harnessing the intricacies of atmospheric boundary layer turbulence ( 16 ). However, its effectiveness diminishes for durations of less than 1 h. The Buoyant Line and Point (BLP) source model focuses on industrial emissions, especially plume rise and downwash effects, but its broader applicability remains questionable ( 17 ).

The transient class of point-source models brings forward tools such as CALPUFF ( 18 ). It is versatile, accounting for various emission sources and aligning with changing meteorological conditions. However, its precision falters in dense urban environments or short-interval analyses. A model proposed by Zegeye offers commendable computational efficiency, ideal for real-time microscopic evaluations ( 19 ). However, its widespread application potential remains under-explored.

In the arena of line source models, CALINE3 stands out in the steady-state domain, adeptly gauging emissions from less complex highway terrains ( 20 ). However, when urban traffic intricacies are introduced, its reliability diminishes. Models such as CAL3QHC and CAL3QHCR aim to bridge this gap, adapting to intricate traffic behaviors, notably at intersections ( 21 ). However, their adaptability across diverse urban settings needs further validation. R-LINE maintains a steady-state approach, focusing on near-surface releases in urban areas ( 22 ). The transient line source model, Line Source Gaussian Puff (LSGP), while effective in assessing freeway emissions, has limitations when delving into the microscopic specifics of individual vehicles ( 23 ).

In addition, computational fluid dynamics (CFD) models represent a unique category. They simulate gas flows and interactions, offering detailed insights into pollutant dispersion, especially in complex urban environments. Shi et al. used CFD to analyze vehicle-induced turbulence in street canyons, demonstrating its value for understanding urban air quality ( 24 ). However, CFD’s computational intensity can limit its use in real-time or large-scale applications.

In our study, we integrate both line-source and point-source models to enhance the accuracy and comprehensiveness of emission dispersion estimation. This integrated approach allows for a more detailed and holistic analysis of truck emissions, considering both moving and idling trucks in urban environments. By combining these models, we aim to provide a robust tool for analyzing and addressing truck-related air pollution, ultimately aiding in the development of targeted mitigation strategies.

Methods

This section outlines the methodologies employed in the study, divided into two main subsections: Enhanced Truck Dispersion Framework Development and Experiment Setup for Data Collection.

Enhanced Truck Dispersion Framework Development

To accurately model the dispersion of emissions from trucks in an urban environment, we developed the Enhanced Truck Dispersion Framework. The framework employs a bifurcated methodology to accurately model the dispersion of emissions from trucks in an urban environment, as shown in Figure 1. The first step involves utilizing BEAM, a macroscopic traffic simulator, to generate realistic traffic patterns within the study area. BEAM relies on real origin-destination (OD) data to ensure an accurate representation of daily vehicle movements. To represent the real-world traffic patterns in an area, an integrated activity-based model for passenger and freight transportation, as shown in Figure 2, is employed. This innovative model integrates both activity-based and trip-based methodologies for a comprehensive portrayal of traffic dynamics.

Figure 1.

The enhanced truck dispersion framework architecture.

Figure 2.

An integrated activity-based framework for passenger car and freight activity generation.

The activity-based segment of the BEAM model synthesizes a demographic population based on detailed census information, enabling the simulation of individual travel behavior. This procedure involves disaggregating population data from an aggregate level down to a high-resolution level, such as the census-block-group level. Passenger vehicle activities are generated from this detailed, synthesized population data, based on the principle that travel is essentially a derived demand arising from individual needs to undertake various activities. Truck activities, on the other hand, are generated separately using regional truck OD demand data derived from trip-based methods. The model also simulates local activities, inbound/outbound trips, and through traffic, aligning with traffic monitoring data to present a holistic reflection of real-world traffic conditions.

Once the BEAM traffic flow is generated, the framework differentiates trucks into two categories: idling trucks and moving trucks. Idling trucks, which are either idling or slow-moving within warehouse areas, are considered as point sources. Moving trucks, along with other vehicles, are treated as line sources.

For moving vehicles, including trucks, the EMFAC model calculates link emissions based on traffic flow characteristics, considering factors such as vehicle type, speed, and road conditions. The Research LINE (RLINE) model then simulates the dispersion of emissions from these moving vehicles, designed specifically for modeling roadway emissions in complex urban environments.

For idling trucks, the framework employs the MOVES model to calculate agent-specific emissions, allowing for a detailed assessment of emissions from individual vehicles, particularly in high-truck-traffic areas such as warehouses and distribution centers. The framework determines the number of trucks idling in warehouses and assigns each a specific idling time. A specialized low-speed movement profile represents truck activities inside warehouses, and is utilized in MOVES to compute emission factors for idling trucks. Emissions data from MOVES are processed through a 3D grid-based point-source dispersion model based on the mass conservation equation which simulates pollutant transport by incorporating advection, diffusion, and deposition processes ( 25 ). This model creates mesh areas around each warehouse to capture emission dispersion, with overlapping meshes combined to form larger areas for accurate representation. The novel dispersion model is specifically calibrated for idling trucks through extensive data collection and experimental setup, ensuring accuracy in representing unique emission patterns in warehouse settings. By integrating meteorological data such as wind speed and direction into its calculations, the model effectively captures how emissions disperse under varying atmospheric conditions. The final stage strategically places receptors to aggregate dispersion results from both on-road emissions and warehouse emissions. This integration allows for a comprehensive assessment of pollution levels, accounting for multiple phases of truck activity.

Experiment Setup for Data Collection

To calibrate the point-source dispersion model specifically for idling trucks, we conduct a field study at a distribution center. Our focus is on examining truck emissions, using CO2 as the tracer for tailpipe emissions because of its significantly higher concentrations near the truck’s tailpipe compared with the background levels. To gather comprehensive data, we use three CO2 receptors to measure CO2 concentrations at different heights and distances from the rear of an idling truck. Complementing these concentration readings, we employ a sonic anemometer to capture 3D wind patterns, ensuring a holistic understanding of the emission dispersion.

Vehicle Setup

Central to the experiment is a Volvo VNR 300 truck fitted with a D11 engine. This engine is a direct injection diesel type with a displacement of 10.8 L. Its six in-line cylinders have a bore and stroke of 4.84 × 5.98 in., and it operates with a compression ratio of 17:1. Impressively, the D11 engine can deliver up to 425 Hp and deliver a maximum torque of 1,550 lb-ft. Notably, this engine is designed to be fuel-efficient and lightweight, characteristics that influence emissions profiles and the subsequent dispersion in the environment. This truck model was selected because it represents one of the most common Class 8 truck models used in regional distribution operations, making it highly representative of the typical vehicles operating in warehouse districts and urban freight corridors ( 26 ). The truck is parked in an open area without any overhead obstructions, allowing for unrestricted vertical dispersion of emissions.

Receptors Setup

To ensure precise and comprehensive data collection for our study, we utilize a suite of advanced receptor equipment, all generously provided by the California Air Resources Board (CARB): the Portable Emissions Acquisition System (PEAQS), the LI-840A CO2 Analyzer, and the LI-820 CO2 Analyzer, as depicted in Figure 3. Before initiating the data collection process, all three receptors are calibrated to guarantee accuracy in their readings.

Figure 3.

Receptors: Portable Emissions Acquisition System (PEAQS) (left), LI-840A CO2 Analyzer (middle), and LI-820 CO2 Analyzer (right).

PEAQS stands as a state-of-the-art system tailored for real-time measurements of emissions. It excels in gauging CO2 concentrations at various distances and elevations from the emission source, with an integral feature being its compatibility with specific CO2 gas analyzers, ensuring meticulous control and data acquisition. The LI-840A CO2 Analyzer is a high-performance instrument renowned for its precision and reliability. It employs a dual-cell infrared gas analyzer, enabling simultaneous measurements of CO2. Its rapid response time and consistent results under fluctuating conditions make it a cornerstone in the data collection toolkit. Lastly, the LI-820 CO2 Analyzer, another part of the LI-COR series, is designed for scenarios demanding high precision. Its streamlined approach to CO2 analysis ensures quick and direct measurements.

Before the sampling, the background CO2 level was determined by averaging measurements taken over a period of time using three calibrated CO2 analyzers (LI-840A, LI-820, and PEAQS) placed together in an open area within the distribution center. These analyzers were positioned away from any nearby emission sources to ensure that the readings reflect true ambient conditions. This approach provides a stable and representative baseline concentration of ambient CO2, which is then used as the background level for isolating truck-specific emissions during subsequent analysis.

Building on the equipment’s capabilities, our experimental setup captures a broad spectrum of data. In accordance with Figure 4, a polar coordinate system is adopted for the measurements. Concentrations of emissions are recorded at varying distances from the rear of the truck: 1 m, 2 m, 3 m, 4 m, 5 m, and 6 m. These measurements are taken at specific angular orientations of 60°, 90°, and 120°. Furthermore, vertical data is also captured at heights of 1 m, 2 m, 3 m, and 4 m above the ground level, providing a comprehensive spatial understanding of the emissions from the truck. Each receptor independently measures CO2 concentrations at different angular orientations: LI-840A at 60°, PEAQS at 90°, and LI-820 at 120° relative to the truck’s rear. The three receptors work simultaneously to provide comprehensive spatial coverage of emission patterns.

Figure 4.

Experiment setup map.

In our experiment, each receptor inlet is affixed to a movable pillar, allowing us to adjust the distance and height for different sampling groups, as shown in Figure 5. The three receptors are deployed at the same distance behind the truck and at an equal height, capturing CO2 concentration data as one “sampling group.” During each sampling group, data is collected for 4–5 min to obtain a stable average plume signal. The CO2 measurements from each receptor are processed individually by subtracting background concentration levels, and averaged over the sampling period. A pause of 1–2 min is incorporated between consecutive sampling groups to ensure a stable plume condition for the next set of readings.

Figure 5.

Real-world experiment setup.

Meteorological Measurements

Meteorological measurements are meticulously conducted using a CSAT3 3D sonic anemometer, which is mounted at a height of 2 m. Operating at a sampling frequency of 20 Hz, this instrument is positioned to the north of the truck, ensuring it is downwind and thus optimally placed to capture the prevailing wind patterns. The sonic anemometer is instrumental in recording both the wind direction and speed.

Emission Estimation

For emission estimation, we use the OBDLink MX Bluetooth device (Figure 6) to collect fuel consumption data from the truck. Using this data, we estimate emission rates assuming complete fuel combustion.

Figure 6.

OBDLink MX bluetooth.

The data collected from this field experiment provides the necessary inputs for calibrating our point-source dispersion model, allowing for a detailed examination of emission patterns from idling trucks under real-world conditions. The data, including CO2 concentrations at various distances and heights, and 3D wind patterns, is thoroughly analyzed and discussed in the Results section.

Case Study

In this study, the Enhanced Truck Dispersion Framework is applied to evaluate the environmental impact of urban traffic on local air quality within Inland Southern California.

Traffic Data and Volume Validation

The study places particular emphasis on dynamic traffic flow scenarios, simulating 15,329 trucks and a total of 153,046 trips over a 24 h period, accounting for approximately 10% of the total traffic in the target area, as shown in Figure 7. These scenarios are crucial for understanding the impact of both truck and passenger vehicle movements on local traffic dynamics and congestion patterns.

Figure 7.

Spatial distribution of 10% sampled truck origin–destination (OD) flows, where curved arcs show OD connections and color intensity indicates truck density.

Validating the traffic volumes against real-world data is essential to establish an accurate traffic model for this area. To achieve this, we utilize the Freeway Performance Measurement System (PeMS) database, which collects traffic statistics across all California highways ( 27 ). This data is used to compare the simulated traffic in the BEAM model, as shown in Figure 8, with actual traffic observations. Figure 9 visualizes the simulated traffic within BEAM, which is used for validation against the real-world traffic data provided by PeMS. Figure 9a and b displays the comparisons between the simulated traffic volumes and the actual traffic volumes recorded by PeMS. Given that the target area is a hub of the prosperous logistics industry, with numerous warehouses particularly around highways I-10, validating traffic volumes on these routes is crucial for an accurate representation of the area’s traffic dynamics. The comparison reveals a noticeable discrepancy in traffic volumes before 5:00 a.m., indicating a disparity in traffic patterns depending on the time of day. Although the calibration and validation process are demanding, the simulated traffic volumes post-5:00 a.m. closely align with the real-world data, reflecting high-quality simulation outcomes.

Figure 8.

The visualization of simulated traffic.

Figure 9.

Comparison of simulated and observed hourly traffic volumes on I–10: (a) eastbound and (b) westbound. Blue lines show PeMS observations, and orange lines show simulations.

Warehouses and Idling Trucks

In our simulation, trucks staying within warehouses typically engage in activities such as loading, unloading, or simply idling with the engine running. To accurately represent the duration of these activities in a realistic manner, the idling time of each truck is modeled using a Gaussian distribution, as shown in Figure 10. This distribution is characterized by a mean of 1,800 s, representing the average expected idling time, and a standard deviation of 300 s. The choice of a Gaussian distribution is deliberate as it effectively captures the normal variations in operational times resulting from factors such as differences in loading or unloading speeds, the efficiency of warehouse operations, and truck-specific factors.

Figure 10.

Trucks idling time distribution.

To simulate truck idling or slow-moving activities, we incorporate a low-speed movement profile for trucks within the warehouse environment, as shown in Figure 11. In our study, idling is defined as truck operation at speeds below 3.5 mph while located within warehouse areas. It encompasses a range of activities, including slow movements as trucks navigate to and from loading docks, periods of idling as they wait for cargo to be loaded or unloaded, and other similar low-speed maneuvers that are characteristic of warehouse operations. This profile is vital for the simulation, as it provides a more nuanced representation of truck behavior as opposed to just considering them as stationary point sources of emissions.

Figure 11.

Truck idling profile.

The 3D grid-based point-source dispersion model spatially represents emissions around each warehouse. The model generates a 50 × 50 grid, with each cell measuring 10 × 10 m, ensuring detailed spatial representation of emission dispersion. This structured approach captures the nuances of emission spread, especially in varying warehouse layouts and operational scales.

In instances where the dispersion areas of nearby warehouses intersect, the simulation merges the overlapping grid sections. This involves combining overlapping cells into a larger composite grid, ensuring each 10 × 10 m cell accurately reflects cumulative emission dispersion from the overlapping warehouses. The final grid mesh is shown in Figure 12.

Figure 12.

Warehouse locations (red dots) and grid meshes (purple boxes) used for emission dispersion analysis.

Results

This section presents the findings of our comprehensive study on truck emission dispersion in urban environments. We begin by examining the calibration and validation of our model, followed by an in-depth analysis of emission dispersion patterns for both moving and idling vehicles. These results provide crucial insights into the spatial distribution of pollutants and the effectiveness of our integrated modeling approach.

Model Calibration and Validation

To ensure the accuracy and reliability of our model, we conducted a thorough calibration and validation process based on field measurements. Data collected over 5 min intervals is averaged and analyzed, with Figure 13 providing a detailed overview of wind direction and speed from the sonic anemometer.

Figure 13.

Wind direction and velocities from the sonic anemometer averaged every 5 min.

Our analytical approach involved several meticulous steps to maintain result accuracy:

• 1) Data cleaning: We addressed challenges in field data collection by removing anomalies and outliers. A lower boundary of 400 parts per million (ppm) and an upper limit of 2,000 ppm were established for CO2 concentrations, considering the background level of around 430 ppm.

• 2) Data pre-processing: This phase included normalizing emission rates and addressing potential biases or systematic discrepancies. The concentration of a substance/flow rate (C/Q) ratio is used to normalize emission rates. This normalization accounts for variations in flow, allowing for consistent comparisons of emissions across different conditions or time periods. By standardizing the data, the C/Q ratio helps reveal underlying patterns or trends in emissions that might be masked by changes in flow. Furthermore, addressing potential biases or systematic discrepancies during this phase ensures the accuracy and reliability of the normalized data, providing a solid foundation for further analysis.

• 3) Data synchronization: We aligned data from various sources (receptors, anemometers, and on-board diagnostics) by matching timestamps, ensuring accurate correspondence between model outputs and actual measurements.

Figure 14 compares model predictions with actual measurements taken at the receptors. The close match between estimated and actual data, after our rigorous data preparation, demonstrates the model’s accuracy in predicting real-world truck emission patterns.

Figure 14.

Comparison between measured and modeled pollutant concentrations; the blue line is the 1:1 reference, and red dashed lines indicate the factor-of-two bounds.

Figure 15 depicts that the relationship between plume rise and distance to the source aligns well with our predictions. It clearly demonstrates a rapid initial rise in the plume close to the emission source. However, as the plume progresses further from the source, its ascent begins to decrease, eventually stabilizing. This observed behavior is consistent with our understanding of plume dynamics and reinforces the accuracy of our predictive models for vehicular emission dispersion.

Figure 15.

Plume rise versus distance from source.

These results validate our model’s robustness and its potential for accurately simulating truck emission dispersion in urban environments.

The subsequent sections will delve into the detailed analysis of emission dispersion patterns for both moving and idling vehicles, providing crucial insights into the spatial distribution of pollutants and the effectiveness of our integrated modeling approach.

Dispersion Results and Analysis

Having validated our model, we now turn our attention to the core findings of this study. The following subsections present a detailed analysis of emission dispersion, first for moving vehicles across the urban scenario, and then for idling trucks in warehouse environments. This comprehensive examination allows us to understand the full scope of truck-related emissions and their impact on urban air quality.

Moving Vehicles Emission Dispersion

This subsection examines the spatial distribution of pollutants emitted by moving vehicles throughout the study area. Utilizing data from the BEAM traffic simulator and EMFAC emission model, and processed through the RLINE dispersion model, we present high-resolution maps of NOx and PM_2.5 (particulate matter with a diameter of 2.5 micrometers or less) concentrations. The receptors, spaced 500 m apart, capture NOx levels at various points across the map, as shown in Figure 16.

Figure 16.

NOx concentration estimated from moving vehicles ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

In the study region, major roads and highways are clearly marked, and the color-coded dots represent the concentration levels at each receptor location. The higher concentrations of NOx are primarily observed along major highways and urban areas, reflecting significant emission sources and their dispersion patterns. This detailed spatial representation helps in understanding the distribution of NOx pollution and its potential impact on the area.

Similarly, Figure 17 for $\mathrm{P}{\mathrm{M}}_{2.5}$ concentration provides insights into particulate matter pollution, with higher concentrations likely observed along major highways and urban centers.

Figure 17.

$\mathrm{P}{\mathrm{M}}_{2.5}$ concentration estimated from moving vehicles ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

Idling Vehicles Emission Dispersion

Here, we focus on the often overlooked but significant contribution of idling trucks to urban air pollution, particularly in warehouse districts. Figures 18 and 19, depicting warehouse $\mathrm{P}{\mathrm{M}}_{2.5}$ and NOx concentrations, provide a detailed spatial representation of these pollutants’ dispersion around warehouse facilities in the study area. These visualizations clearly show how $\mathrm{P}{\mathrm{M}}_{2.5}$ and NOx concentrations decrease as distance from the warehouses increases, illustrating the dispersion patterns of these pollutants in the surrounding area. Areas of high concentration are visible around the warehouses, revealing intense pollution hotspots that may not be captured by traditional traffic-focused models.

Figure 18.

NOx concentration estimated from warehouses ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

Figure 19.

$\mathrm{P}{\mathrm{M}}_{2.5}$ concentration estimated from warehouses ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

Notably, we can also observe the emission dispersion patterns influenced by wind direction. The plumes of higher concentration extend downwind from the warehouse locations, clearly illustrating the impact of meteorological conditions on pollutant dispersion. This wind-driven dispersion is particularly evident in the asymmetrical spread of pollutants around the warehouses, with higher concentrations typically found on the leeward side of the emission sources.

Combined Emission Dispersion

Our integrated framework utilizes EMFAC for calculating link emissions from moving vehicles and MOVES for calculating agent-specific emissions from idling trucks. The framework then employs RLINE for on-road emissions and our point-source model for warehouse emissions. The outputs from both dispersion models are combined in the final stage to produce grid-wise concentration estimates. This synthesis allows for a nuanced understanding of pollutant distribution and intensity across various urban scenarios, enabling accurate quantification and visualization of overall pollutant dispersion from both moving and idling trucks. Figures 20 and 21 present the combined concentration levels of NOx and $\mathrm{P}{\mathrm{M}}_{2.5}$ , respectively, offering a holistic representation of these key pollutants spatial distribution.

Figure 20.

Combined NOx concentration ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

Figure 21.

Combined $\mathrm{P}{\mathrm{M}}_{2.5}$ concentration ( $\mathrm{\mu }\mathrm{g}/{\mathrm{m}}^3$ ).

In Figure 20, we observe the spatial concentration distribution of NOx, which illustrates a higher concentration of this pollutant along major traffic routes and around warehouse facilities. This pattern indicates the significant contribution of truck activities, both in transit and during idling or low-speed movements in warehouse areas, to the overall NOx levels in the urban environment. The concentrated NOx emissions around warehouses particularly highlight the environmental impact of idling truck activities, underscoring the need for more efficient operational practices within these areas.

Similarly, Figure 21 focuses on the distribution of $\mathrm{P}{\mathrm{M}}_{2.5}$ , revealing a pattern that aligns closely with the traffic and warehouse operational dynamics. The higher concentrations of $\mathrm{P}{\mathrm{M}}_{2.5}$ in proximity to warehouses and along busy routes emphasize the compounded effect of truck emissions in these zones. The dispersion of $\mathrm{P}{\mathrm{M}}_{2.5}$ , a pollutant known for its adverse health impacts, further stresses the importance of integrated strategies that address both traffic flow and warehouse operations to improve air quality.

The combined analysis presented in these figures not only demonstrates the effectiveness of our integrated modeling approach but also serves as a crucial tool for urban planners and policymakers. By understanding the areas with heightened levels of NOx and $\mathrm{P}{\mathrm{M}}_{2.5}$ , targeted interventions can be implemented to mitigate the health and environmental impacts of truck emissions. Additionally, these results provide a foundation for further research, particularly in exploring the long-term effects of these pollutants and the effectiveness of various mitigation strategies.

To provide a clearer understanding of the concentration patterns, we have included two subfigures as examples on the top of Figure 20 that zoom in on specific areas of interest. The left subfigure illustrates the concentration distribution in areas near warehouse locations. The top end of the color-code scale captures these significant emission sources impacts, illustrating how pollutant levels remain elevated in proximity to the warehouses because of their operations, and gradually transition to lower levels further away.

The right subfigure depicts the concentration distribution in areas farther from warehouse locations, where pollutant levels are predominantly influenced by highway traffic. By mainly using the lower end of the color-code scale, this subfigure captures a more uniform distribution of lower concentrations, illustrating broader regional patterns. This perspective complements the right subfigure by emphasizing how the high concentrations near warehouses, represented by the top end of the color-code scale, differ from the lower, more regionally distributed concentrations farther away, which are primarily influenced by highway traffic. The color code effectively highlights the localized impact of warehouses while capturing broader regional air quality patterns.

Conclusions and Recommendations

This study represents a significant advancement in modeling and understanding emission dispersion from idling and slow-moving trucks within urban environments. The Enhanced Truck Dispersion Framework integrates multiple modeling frameworks, including BEAM for traffic simulation, EMFAC for emission estimation, and RLINE for dispersion modeling. Additionally, MOVES was employed to calculate emissions from idling trucks because of its detailed modeling capabilities for low-speed and stationary operations, complementing EMFAC’s California-specific strengths in estimating emissions from moving vehicles. This dual-model approach leverages the strengths of both tools to provide a comprehensive representation of truck emissions in urban areas. While EMFAC ensures compliance with California-specific requirements, MOVES allows for precise modeling of idling emissions, particularly around warehouses. This framework is adaptable for use beyond California by substituting EMFAC with local or regional emission models while retaining MOVES for idling scenarios. Field tests conducted at a distribution center provided valuable real-world data that validated our models. The results revealed high concentrations of pollutants, particularly in areas with concentrated truck activities and warehouse operations, underscoring the critical impact of truck emissions on urban air quality. The simulation of emissions from trucks idling in warehouse environments highlighted significant pollutant concentrations around these facilities, emphasizing the need for effective emission control measures. Importantly, this study demonstrates how integrating line-source and point-source models can provide insights into pollution hotspots that might not be fully captured by traditional traffic-volume-based approaches. For example, our framework accounts for emissions from idling trucks at warehouses—an area often overlooked in conventional methods—and provides a more nuanced understanding of spatial pollutant distribution.

These findings highlight the need for targeted mitigation strategies to reduce the environmental and health impacts of truck emissions, both on the road and within warehouse environments. To address the significant impact of truck emissions on urban air quality, particularly around warehouses, several targeted strategies are recommended. These include adopting advanced emission control technologies, implementing stricter zoning policies, enhancing infrastructure to support truck operations, and creating designated areas to reduce idling. More specifically:

• Encourage the adoption of advanced emission control technologies in trucks to reduce emissions from idling and low-speed operations.

• Implement zoning policies that restrict the location of warehouses and distribution centers away from densely populated areas.

• Establish designated truck parking areas with electrification options at warehouses to reduce the need for engine idling during loading and unloading.

• Invest in dedicated truck lanes and optimized signal timings to reduce pollutant hotspots caused by frequent stop-and-go conditions from high truck volumes mixing with general traffic.

These measures should be able to mitigate the environmental and health impacts of truck emissions effectively.