Next Article in Journal
Can Government Budget Management Reconcile Environmental Governance with Sustainable Economic Development?
Previous Article in Journal
Investigating Antecedents of Intention to Use Green Agri-Food Delivery Apps: Merging TPB with Trust and Electronic Word of Mouth
Previous Article in Special Issue
Increasing Electric Vehicle Charger Availability with a Mobile, Self-Contained Charging Station
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 17
Issue 8
10.3390/su17083718
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Impact of Last-Mile Delivery Fleet Electrification on Emissions, Dispersion, and Health: An Environmental Justice Analysis Based on Dallas County, Texas

by
Jaesik Choi
and
Kate Hyun
*
Department of Civil Engineering, University of Texas at Arlington, 416 Yates St., Arlington, TX 76010-1539, USA
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(8), 3718; https://doi.org/10.3390/su17083718
Submission received: 7 February 2025 / Revised: 28 March 2025 / Accepted: 15 April 2025 / Published: 20 April 2025
(This article belongs to the Special Issue Effects of CO2 Emissions Control on Transportation and Its Energy Use)
Browse Figures
Review Reports Versions Notes

Abstract

:
The increasing popularity of online shopping leads to more last-mile deliveries and higher emissions from last-mile delivery trucks. Despite small amounts being emitted from the tailpipe of delivery trucks, there can be a significant amount of fine particulate matter that degrades the health quality of residents from aggregated delivery trucks at the community level. Addressing the environmental impact of last-mile deliveries is needed to achieve the sustainability goals because air pollution mitigation can reduce mortalities. This study employs a comprehensive methodology to assess the health impact of fine particulate matter from last-mile delivery trucks and the benefits of their electrification. It uses a three-tiered modeling approach, incorporating emissions measurement, exposure level evaluation, and health impact assessment. In addition, this paper shows the changes in health impacts at the various levels of the fleet electrification rate using a case study in Dallas County, Texas. The results indicate that higher fine particulate matters are concentrated near or on the major roadways and transportation facilities. This study also shows the relationships between last-mile delivery emissions and socio-economic variables and found that younger, racial minorities and low-income communities are exposed to higher last-mile emissions due to the proximity of their residences to major truck corridors. To evaluate the health impacts of fleet electrification, this study uses 2%, 10%, and 30% market shares of fleet electrification. The results indicate that for 2% of the market share of fleet electrification to last-mile delivery trucks, up to 1 death per year can be prevented, while 70 deaths per year can be prevented with 30% of electrification for the whole traffic.

1. Introduction

The share of electronic shopping (or e-commerce) has been increasing since 1999 in the United States, and online shopping has been the dominant shopping method since 2008 [1]. For example, online shopping accounted for roughly 12.3% of annual sales in the United States in 1999; in 2008, it accounted for over 50.0% of annual sales. This trend has not gone below 50% since 2008. The World Economic Forum [2] has noted that the international popularity of e-commerce is based on increased e-commerce sales, increased urbanized areas, and advanced technology. However, the COVID-19 pandemic has amplified and accelerated the trend of increasing last-mile deliveries, estimating a 25% increase in business-to-customer delivery globally in 2020 [2]. Consequently, the last-mile delivery market was evaluated at USD 131.5 billion in 2021, and the market value is expected to reach USD 288.9 billion by 2031 [3].
A growing number of trucks used for last-mile delivery consequently adds to the concerns about community health outcomes. Tailpipe emissions of trucks are known to emit pollutants such as nitrogen oxide, volatile organic compounds, and various sizes of particulate matter, all of which undermine community health [4]. Amongst these pollutants, fine particulate matter (PM2.5) is a highly concerning material associated with severe health issues, such as premature death, heart disease, and abnormal pregnancy [5,6,7]. Lawton [8] showed that last-mile delivery emissions account for 4.5% of total emissions generated in an overall supply chain process.
An obvious solution to improve air quality from the tailpipe can be to use electric vehicles for last-mile deliveries; even plug-in electric vehicles that use both electricity and fossil fuels can reduce 54% of carbon footprint emissions [9]. Peng et al. [10] found that emissions such as carbon dioxide, nitrogen oxide, PM2.5, and other pollutants are reduced by up to 80% of volatile organic compounds when alternative energy vehicles are applied. Despite the clear benefits of using electric vehicles for last-mile deliveries, scant research exists that aims to understand the health impact of last-mile delivery trucks’ emissions and the benefits of fleet electrification at a regional level. Regional-level analyses are useful to understand the broader level impacts of emissions, such as health impacts for communities, which will be essential to designing environmental policies [11]. However, most of the previous research [12,13,14,15,16,17,18] estimates emissions at the corridor level to capture microscopic traffic movements as the key elements of emission estimations.
Recent studies on measuring health impacts have used a structured three-tiered modeling approach: emission assessment, exposure level measurement from dispersion modeling, and health impact evaluation. For example, Tomar et al. [19] estimate health impacts to assess the effects of emissions in Saharanpur, India, analyzing sectors such as transportation, industry, residential sources, and burning from various fields (e.g., cooking oil from households, burning waste, and residue burning in agriculture) that generate PM2.5 emissions and found that their estimated premature mortalities are 6372 cases and various diseases like chronic obstructive pulmonary disease or circulation problems. Chi et al. [20] also used the structured approach to compare the effects of advanced technologies and stricter emission control policies in 31 regions of China between 2020 (a baseline scenario) and 2030 (a scenario with stronger environmentally friendly policies). They compare the nitrogen oxide emissions and sulfur dioxide emissions in their study area under these two scenarios to assess improvements. They found that implementing stronger air pollutant control policies could reduce 452,551 fatalities.
On the other hand, recent research trends in last-mile emissions in last-mile delivery fields suggest that the current last-mile system can reduce the carbon footprint of last-mile emissions. This can be achieved by turning internal combustion engine trucks into electrified vehicles [21,22,23,24] or by using parcel lockers [25,26,27].
A few studies have investigated the health and economic impacts of various transportation strategies, including vehicle replacement or fleet electrification [28,29,30,31,32,33]. For example, Lee et al. [14] compared the economic impact of health quality comparison in avoided mortalities from the Clean Truck Program, which is a program to improve air quality from the Port of Long Beach and Los Angeles in California by replacing old heavy-duty diesel vehicles with zero-emission and eco-friendly fleets. Lee et al. [14] observed that the Clean Truck Program can save the equivalent of USD 440 million in health outcomes, including premature deaths. Pan et al. [34] compared the health benefits of fleet electrification in premature deaths at the metropolitan level in the United States and identified that when the total sale shares of battery electric vehicles and plug-in hybrid electric vehicles reach a 75% level in 2050, the Greater Los Angeles Area will be able to save 1163 premature deaths in a year. When new diesel trucks are replaced with electric trucks, up to USD 8.5 billion per year will be able to be saved from these climate and health outcomes [35]. Similarly, the health and economic impacts of transportation policies are assessed in other countries. In China, for example, a stronger policy to reduce PM2.5 emissions can decrease annual fatalities from 74,799 in 2025 to 38,774 in 2030, reducing 36,025 deaths over this period. This policy could also save the Chinese government CNY 263.1 billion [36].
Further investigations into community-level health impacts, particularly to understand the most vulnerable and impacted communities from last-mile delivery emissions, are limited, whereas a few researchers [37,38] show the relationship between online shopping delivery frequencies and socio-economic variables. Low-income and racial minorities are particularly vulnerable because their residential places can often be concentrated along heavy truck corridors [39,40]. Demetillo et al. [41] measured nitrogen dioxide emissions at the census tract level using a satellite-based monitoring instrument in the Houston metropolitan area. They found that higher emissions of nitrogen dioxide are captured in low-income non-white communities because of more houses being located near the roadways compared to higher-income or white communities. Kane [42] identified that Hispanic and African Americans are more exposed to PM2.5 levels as high as 15% and 18%, respectively, compared to the average communities in California.
This study overcomes the abovementioned research gaps by developing three-tiered modeling approaches consisting of emissions estimation, dispersion modeling, and community health impact analysis at a regional level. This comprehensive modeling will estimate (i) the amount of emissions with diverse fleet electrification for last-mile delivery using sensitivity analysis to capture the uncertainties present in the market penetration of electric vehicles and (ii) health impacts and associated disparities in communities using a case study of Dallas County, Texas.

2. Methods

Figure 1 shows the overall process of health impact modeling from last-mile deliveries. First, this study adopts truck demand estimation models created by Arabi et al. [43] to identify last-mile delivery demands in 2024. Second, this study uses the improved motor vehicle emission simulator (MOVES-Matrix 2.0) to estimate PM2.5 emissions during peak operating hours, 6AM to 9AM and 4PM to 7PM, from 1 January to 31 December 2024, and third, the Research Line source model (RLINE v1.2) is used to assess exposure levels. Lastly, this paper uses the Benefits Mapping and Analysis Program—Community Edition (BenMAP-CE v 1.5.8) to evaluate health impacts based on the community’s exposure to PM2.5 concentration levels.
The whole process can be split into two major pieces: one is the conversion from link-level analysis (i.e., emission modeling) to point-level analysis (i.e., dispersion modeling), which may downscale the output (i.e., step 2 to step 3), and the other conversion is from the point-level analysis to the regional-level analysis (i.e., measuring health impact) (i.e., step 3 to step 4). To calculate health impact, the average air quality in the regional level is required, and conversion to downscaled exposure levels will be solved by another conversion at the regional-level analysis. To estimate emissions and exposure levels, this paper assumes that geometry has not changed since 2019.

2.1. Step 1: Truck Demands Estimation Model

This study adopts Arabi et al.’s model [43] to estimate hourly last-mile truck demands. Arabi et al. [43] used scenario assessment methods to estimate future last-mile truck demands for Dallas, Texas, using pre- and during-pandemic daily last-mile truck data between 2019 and 2020. Their research reflects a potential increase in online shopping activities and the consequent last-mile deliveries post-pandemic based on the sample last-mile trips gathered pre- and during the pandemic from smartphone location data using Streetlight.
Note that the estimated truck demands by Arabi et al. [43] represent the sample truck demands between freight warehouses (e.g., Amazon freight centers) and residential areas collected by Streetlight. Streetlight sampling rates vary from 1% to 35% depending on locations [44]. Considering various sampling sizes by regions, this paper uses 10%, indicating that last-mile truck samples gathered by Streetlight represent 10% of total trucks on the network in the study area [30]. Given that the sampling rates from Streetlight are not stable by region, and that the data can be recorded incorrectly, we believe that 10% well represents the area to conduct a regional analysis of the last-mile electrification of delivery trucks.
In addition, this study separately estimates the rest of the traffic, including general passenger and commercial traffic using the 2019 traffic data from the Texas Department of Transportation (TxDOT) Roadway database [45].

2.2. Step 2: MOVES-Matrix 2.0

The motor vehicle emission simulator (MOVES) is the emission modeling tool. MOVES can calculate emissions with various pollutants, such as carbon dioxide, PM2.5, and other byproducts, while vehicles are operated. One of the drawbacks of MOVES is the computational burden [46,47]. Therefore, some studies achieve efficient computation by combining MOVES with other traffic simulators or by developing algorithms based on MOVES [46,48].
Table 1 shows the list of input data in measuring emissions through MOVES or MOVES-Matrix 2.0. The input data of MOVES or MOVES-Matrix 2.0 are categorized into four sections, including vehicle characteristics, driving cycle, roadway geometry, and meteorology. Driving behaviors can vary depending on whether vehicles are in urban or rural areas and whether they are on interstates or arterials, even if the average speed is the same. For example, vehicles drive differently on urban interstates compared to rural interstates, despite having the same average speed. Emissions vary based on the fuel type vehicles use. Temperature from meteorology must be shown in Fahrenheit, and temperature and relative humidity must be shown as integers [49].
MOVES-Matrix 2.0 developed by Guensler et al. [48] uses a query system to estimate the emissions rate on a given corridor based on the original emission estimation tool by the United States Environmental Protection Agency, the motor vehicle emission simulator (MOVES). The input data include the age distribution of vehicles, meteorological data, the speed distribution of the vehicles, and vehicle composition rates on the network as the same data for MOVES. Compared to MOVES, MOVES-Matrix 2.0 enhances calculation performance 200 times faster than MOVES [48] because of the pre-estimated emission rates in the system. MOVES-Matrix 2.0 also analyzes the emissions of the network that the user imports by links and by pollutants like MOVES.

2.3. Step 3: Dispersion Model

Dispersion modeling calculates pollutant levels at specific receptors, allowing users to capture pollutant exposure levels. The closer spacing of receptors generally results in more accurate exposure levels. However, this increased accuracy comes at the cost of computational efficiency. Because our ultimate goal is to assess health impact from the electrification of last-mile trucks at a regional level, we ensure that the spacing between receptors is granular enough to measure exposure levels for communities at a block group level while maintaining computational efficiency.
This study uses RLINE v 1.2 to measure exposure levels. RLINE v 1.2 employs virtual receptors to measure exposure levels at specific points. Table 2 shows the list of input data in RLINE. The input data for RLINE v 1.2 include receptor coordinates in the Universal Transverse Mercator system, meteorological data from a meteorological preprocessor, emission rates for each link, and the geometry information of the links [50,51].
Kim et al. [52] highlight the importance of computational efficiency in running dispersion models for regional analysis and recommend simplifying receptor grids to solve computational challenges by applying a spacing of as granular as 200 m at the county or municipal levels [50,53]. Because this step is an intermediate stage between emissions and health impact, the spacing between receptors can be minor as long as every block group contains at least one receptor. This study uses a spacing of 500 m to measure exposure levels to keep the granularity in the analysis while facilitating computational efficiency [50,54].

2.4. Step 4: BenMAP-CE v 1.5.8

BenMAP-CE v 1.5.8 estimates the health impact of the study area and economic factors. If the health impact is calculated based on mortality, BenMAP-CE v 1.5.8 calculates the economic impact of mortality. BenMAP-CE v 1.5.8 converts point exposure levels from the dispersion model output into block group-level exposure levels using grid cell interpolation methods [55].
Table 3 lists input data for assessing health impact. The input data for BenMAP-CE v 1.5.8 are the incidence rates (i.e., morbidity rates or mortality rates of a disease relevant to the pollutant) in the study area, census data in the study area, incidence functions by age range and by diseases, and air quality for each scenario [55]. Users may remove block groups from the study area when the block groups do not have any population information.
Note that BenMAP-CE v 1.5.8 does not show the number of people who will be deceased or ill in each scenario. Instead, it shows differences in health impact between two scenarios (e.g., base vs. case study) in the study area. The output of BenMAP-CE v 1.5.8 is the health impact from the air quality discrepancy between two different scenarios and the economic impact from the health impact by multiplying the social loss cost from the unit with the health impact.

2.5. Step 5: Community Impact Assessment

The last step of the modeling is to compare the health impacts on communities to ultimately understand any health disparity in emissions caused by last-mile delivery trucks [56,57,58] particularly for disadvantaged groups. We use the K-means clustering method to identify vulnerable socio-economic groups in each block group in terms of the following socio-economic variables [59].
  • Age: seniors (aged 65+) and children (aged 17 years old or less) represent social minorities.
  • Race: non-whites represent racial minorities.
  • Economic measures using median household income, the percentage of no vehicles, unemployed, and households under poverty levels characterize economic minorities.

3. Data

3.1. Study Area

The study area is Dallas County, Texas, one of the most populated counties in the Dallas-Fort Worth metropolitan area with 2,613,539 residents. This study uses the American Community Survey data of 2020 at the block group level to gather population demography. There are 1570 block groups in Dallas County, Texas, but the American Community Survey shows that there are 19 block groups that do not have any populations.

3.2. Truck Demands Data and Geometry Data

This study estimates general daily traffic using the Texas Department of Transportation database [45]. This study applies 2% as the average annual growth rate from 2008 to 2035 [60], based on the estimated vehicle miles of travel growth rates from 2008 to 2035. This study also uses the TxDOT Roadway inventory to gather roadway geometries, including the length of links and the number of lanes of the roadways in Dallas County, Texas.

3.3. Meteorological Data

Meteorology data are obtained from the Texas Commission on Environmental Quality [61]. The measurement of emissions requires meteorological conditions (e.g., temperature and relative humidity) and concentration levels [51]. Table 4 shows the meteorological data observed in 2020 that are used to estimate exposure levels.
Sensible heat flux shows energy exchange because of the temperature difference between the surface and the atmosphere [62]. When sensible heat flux is positive, heat moves from the air to the land [63]. Surface friction velocity is the effect of wind force on the surface. The speed of wind can be changed close to the surface, and friction velocity differs with surface roughness, such as whether the surface is paved or if the surface is forested [64]. Station pressure is atmospheric pressure as the force by the air above a specific location due to gravity. Because it is sensitive at a particular elevation and the weight of the air column [65], standard pressure is different by location.

4. Results

4.1. Last-Mile Delivery Daily Truck Demands and Daily General Traffic in Dallas County

Figure 2a,b show the daily traffic volumes of the last-mile delivery trucks and general traffic, which includes private cars, and all commercial trucks combined. When thinner lines are used, then there are less traffic demands than other links. Similarly, when thicker lines are used, then there are more traffic demands than other links.
Note that the last-mile traffic data used by Arabi et al. [43] were collected through Streetlight as a sample of truck traffic to forecast truck demand in 2024. The general traffic data in 2024 are based on the reported traffic in 2019 from the TxDOT roadway database by applying the annual traffic growth rate of 2% from 2019 to 2024.
Significant last-mile delivery trucks are shown near the transportation facilities, A (DFW International Airport), B (Addison Airport), C (Dallas Love Field), D (Dallas Executive Airport), and E (Union Pacific Mesquite Station), due to loading/unloading activities at warehouses located near major freeways and airports.

4.2. Last-Mile Truck Traffic and General Traffic Emissions

Figure 3a,b show the hourly PM2.5 emissions of last-mile delivery traffic and general traffic in Dallas County, respectively, for peak operating hours of general traffic between 6AM and 9AM and 4PM and 7PM. In Figure 3a,b, when thicker lines are used, denser PM2.5 emissions are generated. As expected, major highways show higher emission rates, and links near the transportation infrastructure facilities, such as airports (locations A, B, and C) and the intermodal terminal, show higher emissions. Specifically, the maximum emissions (343.38 g/hour) are captured near the intermodal terminal (location E). However, the last-mile delivery truck emissions account for a small portion of the total emissions as the maximum emissions are 10.28 g/hour from as high as 2594 vehicles last-mile delivery truck/day.
Given that the general traffic results in considerable amounts of PM2.5 emissions, we decided to use the total traffic and the corresponding emission rates for further analysis. The next dispersion modeling calculates exposure levels at the given receptors. Dispersion outputs are sensitive to meteorological conditions, such as the amount of precipitation, wind direction, and wind speed, considering the dispersion of pollutants. Figure 4 shows the dispersion modeling outputs with various color usage: silver (minimal exposure) to yellow (moderate) and red (significant). Note that moderate-to-critical exposures may pose health risks and contribute to adverse health outcomes. Those risk areas appear along with the major roadways, such as I-30, I-635, I-20, and the George Bush Turnpike near the Addison Airport (location B), the Dallas Love Field Airport (location C), the Dallas Executive Airport (location D), and the Union Pacific-Mesquite Intermodal Terminal (location E).
Yellow, orange, or red receptors are shown along major roadways, such as I-30, I-635, I-20, and the George Bush Turnpike (toll roadway). They are near the Addison Airport (location B), the Dallas Love Field Airport (location C), the Dallas Executive Airport (location D), and the Union Pacific-Mesquite Intermodal Terminal (location E).

4.3. Emission Exposure Assessments

This section compares the exposure rates from last-mile deliveries, particularly for susceptible socio-economic groups. We first define susceptible and vulnerable socio-demographic groups using a K-means clustering approach. The elbow method gives the optimal number of clusters based on all the selected socio-demographic and economic variables. For example, Figure 5 shows the optimal cluster number determination for the percentage of children. The within-cluster sum of squares dynamically changes at a number of clusters of three [59,66], so the optimal number of clusters was chosen to be three. This pattern is consistent over all socio-economic variables; therefore, this study uses three groups, referred to as low, medium, and high in the corresponding socio-economic variables, as shown in Table 5.
For each variable, Table 5 shows groups clustered as low, medium, and high with the corresponding range. For instance, the cluster of ‘low’ senior community represents the block groups with 0% to 11.62% senior populations and the cluster of ‘medium’ shows the senior proportions in the block groups between 11.65% and 26.2%. The results show that block groups with lower senior populations are exposed to a higher level of last-mile emissions, which aligns with previous studies that show that seniors are less likely to use online shopping [58,67]. Similarly, block groups with higher portions of children are exposed to more emissions. Additionally, racial minorities and low-income communities are more associated with higher last-mile emissions.

4.4. Health Impacts

Figure 6 shows expected yearly community-level health impacts compared to the base traffic level (i.e., no traffic loaded on the network) in 2024. Figure 6a shows the health impacts when last-mile delivery trucks are added to the network, where the daily average PM2.5 background concentration level in Dallas County is given to 8 μg/m3. Figure 6b shows the health impacts when the general traffic, including last-mile trucks, is added to the network where the daily average PM2.5 background concentration level in Dallas County is 10.1 μg/m3. The healthy air quality of the PM2.5 exposure levels as the background is 5.9 μg/m3 by assuming that no critical diseases are found to endanger human lives [7]. This indicates that the output map shows the difference in mortalities between these two results: the base and control scenarios.
Green areas show fewer ‘more mortalities’, and red areas show higher ‘more mortalities’ in Figure 6, meaning that when the color of the block groups is closer to red, their residents are likely to suffer from the symptoms of the higher PM2.5 exposure levels. This study identifies that the adverse health impact of ‘more mortalities’ is found in the block groups that are located near the airports and the terminals due to the frequent trips of last-mile delivery trucks. However, up to 1.3 mortality/year is observed in this scenario.
Figure 6a shows the health impact when only last-mile trucks are considered, and Figure 6b shows the health outcomes when the general vehicles are added to the study area. Compared to the last-mile only scenario, a much higher number of mortality rates (up to 247 mortalities/year) is shown with the general traffic. Notably, the most critical block group is located near E (Union Pacific Mesquite Station), where last-mile routes are on I-30 and adjacent to US-80 in Figure 6b.

4.5. Electrification Impact on Health and Sensitivity Analysis

This section compares the health impact of fleet electrification to measure how much fleet electrification improves air quality at the community level using both last-mile and general traffic electrification. This study develops nine different scenarios, including the electrification of (a) last-mile delivery trucks only (the first column in Figure 7), (b) all trucks (the second column in Figure 7), and (c) whole traffic (the third column at Figure 7), with three different market penetration scenarios at 2%, 10%, and 30%. This study first uses 2% market penetration for last-mile delivery trucks as a realistic scenario [68] to showcase the most probable health impact of fleet electrification, specifically avoiding mortalities with the fleet electrification scenario compared to non-fleet electrification, as shown in Figure 7a.
The expected maximum avoided mortality is up to 0.301 deaths per year at 2% market penetration of fleet electrification for last-mile delivery trucks. However, replacing only last-mile delivery trucks with electrified trucks does not significantly improve health impacts. In contrast, adopting 2% of electric vehicles across all vehicle types can save up to five deaths per year. The difference in the avoided mortalities between the electric truck scenario and the non-electric truck scenario improvement is not significant in Dallas County as a whole. For example, there is nearly no discernible difference between the scenarios of non-electrification and electrification in the southern area, especially at the areas at the boundary of Dallas County (i.e., 0 to 0.0530 deaths/year). The smaller avoided mortalities (i.e., 0.0530 deaths/year) are because the differences in air quality between before and after fleet electrification are insignificant when 2% conversion rates are applied.
Therefore, this paper investigates electrification impact at 10% and 30% market penetrations. Figure 7b,c show the avoided mortalities from different fleet electrification adoptions. Specifically, compared to the 2% scenario (the first row in Figure 7), higher market penetration rates of electric vehicles show significantly more avoided mortalities. For instance, when the fleet electrification adoption ratio reaches 30% for all trucks, the maximum avoided mortalities can be 19 deaths per year. Furthermore, when the fleet electrification adoption reaches 30% for general traffic, up to 70 deaths per year can be avoided. At 10% fleet electrification adoption for all truck scenarios, the reduction impact of avoided mortalities is less than 30% fleet electrification; this value decreases to four people/year. In summary, higher penetration rates of electric vehicles can result in more residents being safe from diseases or fatalities caused by PM2.5 emissions.
According to the United States Department of Transportation, the value of a statistical life was USD 13.2 million in 2023 per person [69]. In the study area, at a 2% penetration rate of electric vehicles, Dallas County could have saved USD 4768.88 million (USD 4.77 billion) in 2023. At a 10% penetration rate of electric vehicles, Dallas County could have saved USD 23,370.80 million (USD 23.38 billion) in 2023. At a 30% penetration rate of electric vehicles, Dallas County could have saved USD 66,293.80 million (USD 66.30 billion) in 2023.
Table 6 shows the specific numbers in Figure 7 of the block groups that expect better health outcomes due to electrification. Out of 1551 block groups in Dallas County, 152 and 1549 block groups are expected to experience better health outcomes when 30% of last-mile trucks and general traffic are electrified.

5. Discussions and Conclusions

5.1. Discussions

This study introduces an approach to estimate health impacts from last-mile delivery scenarios at different truck estimations and various fleet electrification scenarios at the community level. We use comprehensive modeling steps: (i) emission models, (ii) dispersion modeling, and (iii) measuring health impacts to more accurately capture emissions from last-mile trucks and estimate health impacts at the regional level. Additionally, we quantify the level of PM2.5 exposures by the socio-economic variables of block groups using the block group data of Dallas County.
This study finds up to 247 fatalities/year when all trucks are used with internal combustion engine vehicles compared to healthy air quality. However, when 30% of last-mile delivery trucks are electrified, fatality rates are reduced by 24%. This study also shows the relationships between last-mile delivery emissions and socio-economic variables and identifies that age and race factors are mostly associated with last-mile delivery emissions. Notably, households with lower median household incomes are likely to be exposed to more emissions than those with higher incomes, which may create disproportional impacts because wealthier households engage in more online shopping, increasing last-mile deliveries on the network [70,71,72], and consequently, higher incomes may contribute to greater last-mile emissions.
This study presents how health quality is improved at the block groups in communities when fleet electrification penetration rates increase regionally. These findings illustrate that non-white, younger, and high-income communities are more likely to be exposed to PM2.5 emissions compared to their counterparts (i.e., white, senior, and low-income communities) for various reasons. Younger communities have less technology hurdles compared to senior groups, which may lead to higher online shopping usage resulting in greater exposure to emissions. However, the reasons for exposure disparities between high-income and low-income groups and between white groups and non-white groups are more complex.
The reason why non-white households or lower-income households are exposed to higher PM2.5 emissions from last-mile deliveries can be attributed to several factors, including the fact that residential areas are more closely related to major highways and arterials. Logistic service providers (LSPs) tend to select highways and major arterials as truck corridors to maximize operational efficiency by choosing the shortest path from warehouses and delivery locations.
However, LSP decision-making extends beyond route optimization. Factors such as warehouse locations, demand concentration, and operational costs also influence route planning, which may inadvertently lead to disproportionate emissions in certain socio-economic areas. Furthermore, infrastructure plays a crucial role in the distribution of emissions. Areas with limited access to distribution hubs may experience higher emissions due to longer travel distances and fewer optimized delivery routes.

5.2. Future Study

Reducing unnecessary emissions in the transportation network is essential for improving community health. There are some potential solutions to minimize tailpipe emissions, such as (i) optimizing last-mile routes, (ii) utilizing delivery pick-up points to reduce last-mile trucks’ travel, and (iii) replacing environmentally friendly vehicles.
LSPs can refine delivery routes by considering parcel locker locations and customer preferences [73,74]. Optimized routes could help prevent excessive exposure to emissions in financially disadvantaged block groups using advanced routing algorithms.
Establishing temporary terminals or micro-hubs could reduce emissions by shortening last-mile delivery distances. To complete customers’ orders or to change vehicles in the middle of last-mile deliveries, LSPs can use temporary terminals [25,75]. They can prepare various types of temporary terminals (i.e., parcel lockers or interim terminals to change vehicles).
As this paper shows the fleet electrification impact in Figure 7, electric vehicles reduce tailpipe emissions and enhance health impacts. Some electric vehicles, such as electric assistance bikes or drones, help LSPs to reach “hard-to-reach” places [76,77]. Rather than looking for a better place to complete deliveries to the refused areas, LSPs can use alternative vehicles to complete deliveries to customers’ preferred places. However, ensuring that electrification benefits are distributed equitably across communities remains a key policy challenge.
Policymakers or LSPs can develop strategies to reduce PM2.5 emissions for better community health. Solutions may include incentivizing the use of micro-distribution hubs, improving parcel locker infrastructure, and encouraging alternative last-mile delivery modes, such as electric bikes and drones. Additionally, integrating environmental justice metrics into logistics planning could help mitigate emission disparities across different socio-economic groups.

5.3. Limitation

This paper includes several limitations due to data availability issues. For example, the assumptions made for meteorological data and traffic data in 2024, as well as the sampling rates of 10%, may require validation. This study assumed 10% to represent a regional sampling rate of last-mile trucks from Streetlight due to the lack of availability at a link level [30]. However, a more accurate and granular sampling rate may provide a more accurate estimation of delivery truck demands as well as health impacts.
This study estimated general traffic in 2024 by applying an annual growth factor from 2019 to 2024 with the traffic demands in 2019. This approach may over- or under-estimate the traffic demand in 2024; therefore, a future study should be conducted using real-time link data to produce more accurate link-level volumes on the network.
Similarly, due to limited sources and limited access, this paper cannot evaluate potential morbidities from PM2.5 emissions. This study could find mortality rates publicly [78]; however, morbidity rates are not available publicly, and therefore this study excluded the impact from morbidity. When morbidity rates with economic burden are available, a more accurate health impact with economic burden can be expected.
There can be more economic variables that can generate last-mile deliveries, which leads to more PM2.5 emissions. Although this paper uses three economic variables (no vehicles, no employment, and median household incomes), this paper expects more economic variables that may show the strong relationship between economic variables and PM2.5 emissions.
Lastly, this study assumed 500 m for spacing between receptors by assuming the minimum distance to contain one or more receptors in every block group in the study area from the unit, which is how the American Community Survey collects the demographic information. While this study uses 500 m spacing, this study may have skipped the small size of block groups unintentionally.

Author Contributions

The authors confirm their contribution to the paper as follows: study conception and design: K.H. and J.C.; data collection: J.C.; analysis and interpretation of results: K.H. and J.C.; draft manuscript preparation: K.H. and J.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Average Daily Traffic data in 2019: https://www.txdot.gov/data-maps/roadway-inventory.html (accessed on 5 February 2025). American Community Survey (2018–2022): https://data2.nhgis.org/main (accessed on 5 February 2025).

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. United States Census Bureau. Annual Retail Trade Survey: 2022. Available online: https://www.census.gov/data/tables/2022/econ/arts/annual-report.html (accessed on 23 April 2024).
  2. World Economic Forum. Pandemic, Parcels and Public Vaccination Envisioning the Next Normal for the Last-Mile Ecosystem; World Economic Forum: Cologny/Geneva, Switzerland, 2021. [Google Scholar]
  3. Allied Market Research. Last Mile Delivery Market by Service Type, by Vehicle Type, by Mode of Operation, by Delivery Mode, by Application, by Destination: Global Opportunity Analysis and Industry Forecast, 2021–2031; Allied Market Research: Pune, India, 2022. [Google Scholar]
  4. United States Environmental Protection Agency. Smog, Soot, and Other Air Pollution from Transportation. Available online: https://www.epa.gov/transportation-air-pollution-and-climate-change/smog-soot-and-other-air-pollution-transportation#:~:text=Air%20pollution%20emitted%20from%20transportation,volatile%20organic%20compounds%20(VOCs) (accessed on 3 September 2023).
  5. Li, C.; Managi, S. Contribution of On-Road Transportation to PM2.5. Sci. Rep. 2021, 11, 21320. [Google Scholar] [CrossRef] [PubMed]
  6. National Toxicology Program. NTP Monograph: Systematic Review of Traffic-Related Air Pollution and Hypertensive Disorders of Pregnancy; National Toxicology Program: Research Triangle Park, NC, USA, 2019. [Google Scholar]
  7. World Health Organization. WHO Global Air Quality Guidelines. Particulate Matter (PM2.5 and PM10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide; World Health Organization: Geneva, Switzerland, 2021. [Google Scholar]
  8. Lawton, G. The Environmental Challenges of Last-Mile Delivery. Available online: https://www.techtarget.com/searcherp/feature/The-environmental-challenges-of-last-mile-delivery#:~:text=Supply%20chain%20emissions%20make%20up%20typically%2090%25%20of,Neutral%2C%20a%20nonprofit%20organization%20located%20in%20San%20Francisco (accessed on 16 April 2023).
  9. Tonachel, L. Study: Electric Vehicles Can Dramatically Reduce Carbon Pollution from Transportation, and Improve Air Quality. Available online: https://www.nrdc.org/bio/luke-tonachel/study-electric-vehicles-can-dramatically-reduce-carbon-pollution-transportation#:~:text=As%20shown%20below%2C%20a%20PEV,the%20electricity%20grid%20carbon%20intensity (accessed on 7 June 2024).
  10. Peng, L.; Liu, F.; Zhou, M.; Li, M.; Zhang, Q.; Mauzerall, D.L. Alternative-Energy-Vehicles Deployment Delivers Climate, Air Quality, and Health Co-Benefits When Coupled with Decarbonizing Power Generation in China. One Earth 2021, 4, 1127–1140. [Google Scholar] [CrossRef]
  11. Grant, M.; D’ignazio, J.; Ang-Olson, J.; Chavis, A.; Gallivan, F.; Harris, M.; Rooney, K.; Silla, T.; Wallis, E.; Siwek, S. Assessing Mechanisms for Integrating Transportation-Related Greenhouse Gas Reduction Objectives into Transportation Decision Making; The National Academies Press: Washington, DC, USA, 2010. [Google Scholar]
  12. Yavari, A.; Bagha, H.; Korala, H.; Mirza, I.; Dia, H.; Scifleet, P.; Sargent, J.; Shafiei, M. ParcEMon: IoT Platform for Real-Time Parcel Level Last-Mile Delivery Greenhouse Gas Emissions Reporting and Management. Sensors 2022, 22, 7380. [Google Scholar] [CrossRef]
  13. Shahmohammadi, S.; Steinmann, Z.J.N.; Tambjerg, L.; Van Loon, P.; King, J.M.H.; Huijbregts, M.A.J. Comparative Greenhouse Gas Footprinting of Online versus Traditional Shopping for Fast-Moving Consumer Goods: A Stochastic Approach. Environ. Sci. Technol. 2020, 54, 3499–3509. [Google Scholar] [CrossRef]
  14. Lee, G.; You, S.I.; Ritchie, S.G.; Saphores, J.D.; Jayakrishnan, R.; Ogunseitan, O. Assessing Air Quality and Health Benefits of the Clean Truck Program in the Alameda Corridor, CA. Transp. Res. Part A Policy Pract. 2012, 46, 1177–1193. [Google Scholar] [CrossRef]
  15. Liu, H.; Xu, X.; Rodgers, M.O.; Xu, Y.A.; Guensler, R.L. MOVES-Matrix and Distributed Computing for Microscale Line Source Dispersion Analysis. J. Air Waste Manage Assoc. 2017, 67, 763–775. [Google Scholar] [CrossRef] [PubMed]
  16. Xu, Y.A.; Liu, H.; Rodgers, M.O.; Guin, A.; Hunter, M.; Sheikh, A.; Guensler, R. Understanding the Emission Impacts of High-Occupancy Vehicle (HOV) to High-Occupancy Toll (HOT) Lane Conversions: Experience from Atlanta, Georgia. J. Air Waste Manag. Assoc. 2017, 67, 910–922. [Google Scholar] [CrossRef] [PubMed]
  17. Li, L. Life Cycle Greenhouse Gas Emissions for Last-Mile Parcel Delivery by Automated Vehicles and Robots. Master’s Thesis, University of Michigan, Ann Arbor, MI, USA, 2021. [Google Scholar]
  18. Xu, X.; Liu, H.; Anderson, J.M.; Xu, Y.; Hunter, M.P.; Rodgers, M.O.; Guensler, R.L. Estimating Project-Level Vehicle Emissions with Vissim and MOVES-Matrix. Transp. Res. Rec. 2016, 2570, 107–117. [Google Scholar] [CrossRef]
  19. Tomar, G.; Nagpure, A.S.; Jain, Y.; Kumar, V. High-Resolution PM2.5 Emissions and Associated Health Impact Inequalities in an Indian District. Environ. Sci. Technol. 2023, 57, 2310–2321. [Google Scholar] [CrossRef]
  20. Chi, Y.; Zhang, Y.; Zhang, X.; Liu, B. Potential and Health Impact Assessment of Air Pollutant Emission Reduction: A Case Study of China. Clean Technol. Environ. Policy 2024, 27, 1053–1066. [Google Scholar] [CrossRef]
  21. Wangsa, I.D.; Wee, H.M.; Hsiao, Y.L.; Rizky, N. Identifying an Effective Last-Mile Customer Delivery Option with an Integrated Eco-Friendly Inventory Model. INFOR Inf. Syst. Oper. Res. 2022, 60, 165–200. [Google Scholar] [CrossRef]
  22. Fajardo-Calderin, J.; Masegosa, A.D.; Elejoste, M.P.; Moreno-Emborujo, A.; Cantero-Lopez, X.; Angulo, I. An Artificial Intelligence-Based Software Module for the Optimization of Collaborative Delivery in Last-Mile Logistics. In Proceedings of the 9th International Physical Internet Conference, Athens, Greece, 13–15 June 2023. [Google Scholar]
  23. Garus, A.; Alonso, B.; Raposo, M.A.; Grosso, M.; Krause, J.; Mourtzouchou, A.; Ciuffo, B. Last-Mile Delivery by Automated Droids. Sustainability Assessment on a Real-World Case Study. Sustain. Cities Soc. 2022, 79, 103728. [Google Scholar] [CrossRef]
  24. McLeod, F.N.; Cherrett, T.J.; Bektas, T.; Allen, J.; Martinez-Sykora, A.; Lamas-Fernandez, C.; Bates, O.; Cheliotis, K.; Friday, A.; Piecyk, M.; et al. Quantifying Environmental and Financial Benefits of Using Porters and Cycle Couriers for Last-Mile Parcel Delivery. Transp. Res. D Transp. Environ. 2020, 82, 102311. [Google Scholar] [CrossRef]
  25. Bienzeisler, L.; Wage, O.; Friedrich, B. Cutting Redundancy, Slashing Emissions: The Role of Consolidation for a Sustainable Last-Mile Delivery. In Proceedings of the FISTS 2024: 4th Forum for Innovative Sustainable Transportation Systems, Riverside, CA, USA, 26–28 February 2024; Institute of Electrical and Electronics Engineers Inc.: New York, NY, USA, 2024. [Google Scholar]
  26. Dong, B.; Hovi, I.B.; Pinchasik, D.R. Analysis of Service Efficiency of Parcel Locker in Last-Mile: A Case Study in Norway. Transp. Res. Procedia 2023, 69, 918–925. [Google Scholar] [CrossRef]
  27. Peppel, M.; Spinler, S.; Winkenbach, M. Integrating Mobile Parcel Lockers into Last-Mile Delivery Networks: An Operational Design for Home Delivery, Stationary, and Mobile Parcel Lockers. Int. J. Phys. Distrib. Logist. Manag. 2024, 54, 418–447. [Google Scholar] [CrossRef]
  28. Pennington, A.F.; Cornwell, C.R.; Sircar, K.D.; Mirabelli, M.C. Electric Vehicles and Health: A Scoping Review. Environ. Res. 2024, 251, 118697. [Google Scholar] [CrossRef]
  29. Calatayud, V.; Diéguez, J.J.; Agathokleous, E.; Sicard, P. Machine Learning Model to Predict Vehicle Electrification Impacts on Urban Air Quality and Related Human Health Effects. Environ. Res. 2023, 228, 115835. [Google Scholar] [CrossRef]
  30. Cárdenas, I.; Beckers, J.; Vanelslander, T. E-Commerce Last-Mile in Belgium: Developing an External Cost Delivery Index. Res. Transp. Bus. Manag. 2017, 24, 123–129. [Google Scholar] [CrossRef]
  31. Choma, E.F.; Evans, J.S.; Hammitt, J.K.; Gómez-Ibáñez, J.A.; Spengler, J.D. Assessing the Health Impacts of Electric Vehicles through Air Pollution in the United States. Environ. Int. 2020, 144, 106015. [Google Scholar] [CrossRef]
  32. Filigrana, P.; Levy, J.I.; Gauthier, J.; Batterman, S.; Adar, S.D. Health Benefits from Cleaner Vehicles and Increased Active Transportation in Seattle, Washington. J. Expo. Sci. Environ. Epidemiol. 2022, 32, 538–544. [Google Scholar] [CrossRef]
  33. Ma, C.; Madaniyazi, L.; Xie, Y. Impact of the Electric Vehicle Policies on Environment and Health in the Beijing–Tianjin–Hebei Region. Int. J. Environ. Res. Public Health 2021, 18, 623. [Google Scholar] [CrossRef] [PubMed]
  34. Pan, S.; Yu, W.; Fulton, L.M.; Jung, J.; Choi, Y.; Gao, H.O. Impacts of the Large-Scale Use of Passenger Electric Vehicles on Public Health in 30 US. Metropolitan Areas. Renew. Sustain. Energy Rev. 2023, 173, 113100. [Google Scholar] [CrossRef]
  35. Hennessy, E.M.; Scown, C.D.; Azevedo, I.M.L. The Health, Climate, and Equity Benefits of Freight Truck Electrification in the United States. Environ. Res. Lett. 2024, 19, 104069. [Google Scholar] [CrossRef]
  36. Cai, X.; Zheng, S.; Zhang, X.; Ye, Z.; Liu, C.; Tan, Z. The Impact of CO2 Emission Synergy on PM2.5 Emissions and a Dynamic Analysis of Health and Economic Benefits: A Case Study of China’s Transportation Industry. J. Clean. Prod. 2024, 471, 143405. [Google Scholar] [CrossRef]
  37. Figliozzi, M.; Unnikrishnan, A. Home-Deliveries before-during COVID-19 Lockdown: Accessibility, Environmental Justice, Equity, and Policy Implications. Transp. Res. D Transp. Environ. 2021, 93, 102760. [Google Scholar] [CrossRef]
  38. Wang, X.C.; Zhou, Y. Deliveries to Residential Units: A Rising Form of Freight Transportation in the U.S. Transp. Res. Part C Emerg. Technol. 2015, 58, 46–55. [Google Scholar] [CrossRef]
  39. Pratt, G.C.; Vadali, M.L.; Kvale, D.L.; Ellickson, K.M. Traffic, Air Pollution, Minority and Socio-Economic Status: Addressing Inequities in Exposure and Risk. Int. J. Environ. Res. Public Health 2015, 12, 5355–5372. [Google Scholar] [CrossRef]
  40. Houston, D.; Li, W.; Wu, J. Disparities in Exposure to Automobile and Truck Traffic and Vehicle Emissions near the Los Angeles-Long Beach Port Complex. Am. J. Public Health 2014, 104, 156–164. [Google Scholar] [CrossRef]
  41. Demetillo, M.A.G.; Navarro, A.; Knowles, K.K.; Fields, K.P.; Geddes, J.A.; Nowlan, C.R.; Janz, S.J.; Judd, L.M.; Al-Saadi, J.; Sun, K.; et al. Observing Nitrogen Dioxide Air Pollution Inequality Using High-Spatial-Resolution Remote Sensing Measurements in Houston, Texas. Environ. Sci. Technol. 2020, 54, 9882–9895. [Google Scholar] [CrossRef]
  42. Kane, E.J. Evaluating Equity in Transportation Greenhouse Gas Emission Reduction Strategies in Local Climate Action Planning in Oakland, California. Master’s Thesis, University of San Francisco, San Francisco, CA, USA, 2022. [Google Scholar]
  43. Arabi, M.; Hyun, K.; Mattingly, S.P.; Imran, M.A. Post-Pandemic Last-Mile and Shopping Center Trip Estimations Using Scenario Analysis. In Proceedings of the Transportation Research Board Annual Meeting, Washington, DC, USA, 8–12 January 2023. [Google Scholar]
  44. StreetLight Data, Inc. StreetLight InSight® Metrics: Methodology and Data Sources; StreetLight Data, Inc.: San Francisco, CA, USA, 2018. [Google Scholar]
  45. Texas Department of Transportation 2019 Roadway Inventory. Available online: https://www.txdot.gov/inside-txdot/division/transportation-planning/roadway-inventory.html (accessed on 28 July 2023).
  46. Ramirez-Sanchez, E.; Tang, C.; Xu, Y.; Renganathan, N.; Jayawardana, V.; He, Z.; Wu, C. NeuralMOVES: A Lightweight and Microscopic Vehicle Emission Estimation Model Based on Reverse Engineering and Surrogate Learning. arXiv 2025, arXiv:2502.04417. [Google Scholar]
  47. Zhao, Y.; Sadeka, A.W. Computationally-Efficient Approaches to Integrating the MOVES Emissions Model with Traffic Simulators. In Proceedings of the Procedia Computer Science, Barcelona, Spain, 5–7 June 2013; Elsevier, B.V.: Amsterdam, The Netherlands, 2013; Volume 19, pp. 882–887. [Google Scholar]
  48. Guensler, A.; Liu, R.; Xu, H. MOVES-Matrix for High-Performance Emission Rate Model Applications; National Center for Sustainable Transportation: Davis, CA, USA, 2018. [Google Scholar]
  49. United States Environmental Protection Agency. MOVES2014a User Guide; United States Environmental Protection Agency: Washington, DC, USA, 2015.
  50. Texas Commission on Environmental Quality. Air Quality Modeling Guidelines; Texas Commission on Environmental Quality: Austin, TX, USA, 2019.
  51. United States Environmental Protection Agency. User’s Guide for the AERMOD Meteorological Preprocessor (AERMET); United States Environmental Protection Agency: Washington, DC, USA, 2022.
  52. Kim, D.; Liu, H.; Rodgers, M.O.; Guensler, R. Development of Roadway Link Screening Model for Regional-Level near-Road Air Quality Analysis: A Case Study for Particulate Matter. Atmos. Environ. 2020, 237, 117677. [Google Scholar] [CrossRef]
  53. Kim, D.; Liu, H.; Xu, X.; Lu, H.; Wayson, R.; Rodgers, M.O.; Guensler, R. Distributed Computing for Region-Wide Line Source Dispersion Modeling. Comput.-Aided Civ. Infrastruct. Eng. 2021, 36, 331–345. [Google Scholar] [CrossRef]
  54. Zou, B.; Wilson, J.G.; Zhan, F.B.; Zeng, Y. An Emission-Weighted Proximity Model for Air Pollution Exposure Assessment. Sci. Total Environ. 2009, 407, 4939–4945. [Google Scholar] [CrossRef]
  55. United States Environmental Protection Agency. Environmental Benefits Mapping and Analysis Program-Community Edition User’s Manual; United States Environmental Protection Agency: Washington, DC, USA, 2022.
  56. Barua, L.; Zou, B.; Zhou, Y.; Liu, Y. Modeling Household Online Shopping Demand in the U.S.: A Machine Learning Approach and Comparative Investigation between 2009 and 2017. Transportation 2023, 50, 437–476. [Google Scholar] [CrossRef]
  57. Giuliano, G. The Challenges of Freight Transport in Cities. In Handbook on City Logistics and Urban Freight; Edward Elgar Publishing: Cheltenham, UK, 2023. [Google Scholar]
  58. Krishnamurthy, S.; Suppiah, S. A Study to Explore Behavioral Dynamics and Expectations of the Elderly People from Last Mile Delivery Services. In HCI International 2021—Posters, Proceedings of the 23rd HCI International Conference, Virtual, 24–29 July 2021; Filipe, J., Ghosh, A., Prates, R.O., Zhou, L., Eds.; Springer International Publishing: Cham, Switzerland, 2021. [Google Scholar]
  59. Syakur, M.A.; Khotimah, B.K.; Rochman, E.M.S.; Satoto, B.D. Integration K-Means Clustering Method and Elbow Method for Identification of the Best Customer Profile Cluster. In Proceedings of the IOP Conference Series: Materials Science and Engineering, Bali, Indonesia, 16–17 April 2018; Institute of Physics Publishing: Bristol, UK, 2018; Volume 336. [Google Scholar]
  60. Texas Department of Transportation. The Statewide Long-Range Transportation Plan 2035; Texas Department of Transportation: Austin, TX, USA, 2010.
  61. Texas Commission on Environmental Quality. Meteorological Data for Refined Screening with AERMOD. Available online: https://www.tceq.texas.gov/permitting/air/modeling/aermod-datasets.html (accessed on 16 April 2023).
  62. Ledley, T.S. Energy Balance Model, Surface. Encycl. Atmos. Sci. 2003, 747–754. [Google Scholar]
  63. Center for Multiscale Modeling of Atmospheric Processes. Climate. Available online: https://hogback.atmos.colostate.edu/cmmap/learn/climate/energy4.html#:~:text=So%2C%20a%20positive%20latent%20heat,the%20atmosphere%20to%20the%20surface (accessed on 24 February 2024).
  64. Garimella, S. Condensation in Minichannels and Microchannels. In Heat Transfer and Fluid Flow in Minichannels and Microchannels; Elsevier Science: Amsterdam, The Netherlands, 2014. [Google Scholar]
  65. National Weather Service. Pressure Definitions; National Weather Service: Silver Spring, MD, USA, 2023. [Google Scholar]
  66. Cui, M. Introduction to the K-Means Clustering Algorithm Based on the Elbow Method. Account. Audit. Financ. 2020, 1, 5–8. [Google Scholar] [CrossRef]
  67. Reiffer, A.; Kübler, J.; Briem, L.; Kagerbauer, M.; Vortisch, P. Analysing Long-Term Effects of the COVID-19 Pandemic on Last-Mile Delivery Traffic Using an Agent-Based Travel Demand Model. In Proceedings of the 12th International Scientific Conference on Mobility and Transport, Singapore, 5–7 April 2022; Antoniou, C., Rau, A., Busch, F., Hariharan, M., Eds.; Springer: Berlin/Heidelberg, Germany, 2023. [Google Scholar]
  68. Miller, A.; Wang, M.; Fulton, Q. Spatial Scenarios for Market Penetration of Plug-in Battery Electric Trucks; National Renewable Energy Laboratory: Golden, CO, USA, 2022.
  69. United States Department of Transportation. Departmental Guidance on Valuation of a Statistical Life in Economic Analysis. Available online: https://www.transportation.gov/office-policy/transportation-policy/revised-departmental-guidance-on-valuation-of-a-statistical-life-in-economic-analysis (accessed on 27 March 2025).
  70. Giuliano, G.; Fang, J.; Binder, R.B.; Ha, J.; Holmes, A. Travel Behavior in E-Commerce: Shopping, Purchasing, and Receiving; METRANS Transportation Center (Calif.): Los Angeles, CA, USA; National Center for Sustainable Transportation (NCST) (UTC): Davis, CA, USA, 2022. [Google Scholar] [CrossRef]
  71. PYMNTS. High-Income Consumers Shop Online 40% More Than Low-Income. Available online: https://www.pymnts.com/news/ecommerce/2024/high-income-consumers-shop-online-40percent-more-than-low/#:~:text=High%2DIncome%20Consumers%20Shop%20Online%2040%%20More%20Than%20Low%2DIncome.&text=Additionally%2C%20the%20study%20%E2%80%9CThe%20Online%20Features%20Driving,to%20want%20to%20buy%20directly%20from%20brands (accessed on 25 March 2025).
  72. Feger, A. Discover In-Store, Buy Online: How High-Income Consumers Shop. Available online: https://www.emarketer.com/content/discover-in-store--buy-online--how-high-income-consumers-shop (accessed on 24 March 2025).
  73. Rincon-Garcia, N.; Waterson, B.; Cherrett, T.J.; Salazar-Arrieta, F. A Metaheuristic for the Time-Dependent Vehicle Routing Problem Considering Driving Hours Regulations—An Application in City Logistics. Transp. Res. Part A Policy Pract. 2020, 137, 429–446. [Google Scholar] [CrossRef]
  74. Bonomi, V.; Mansini, R.; Zanotti, R. Last Mile Delivery with Parcel Lockers: Evaluating the Environmental Impact of Eco-Conscious Consumer Behavior. IFAC-PapersOnLine 2022, 55, 72–77. [Google Scholar] [CrossRef]
  75. Trott, M.; Baur, N.F.; Auf der Landwehr, M.; Rieck, J.; von Viebahn, C. Evaluating the Role of Commercial Parking Bays for Urban Stakeholders on Last-Mile Deliveries—A Consideration of Various Sustainability Aspects. J. Clean. Prod. 2021, 312, 127462. [Google Scholar] [CrossRef]
  76. Bányai, T. Impact of the Integration of First-Mile and Last-Mile Drone-Based Operations from Trucks on Energy Efficiency and the Environment. Drones 2022, 6, 249. [Google Scholar] [CrossRef]
  77. Nguyen, Q.T.; Patrix, J.; Oulmakki, O.; Ramirez-Rios, D.G.; Verny, J. Estimating Costs and CO2 Emissions for Last Mile Delivery of Groceries in Paris Using Alternative Modes and Routing Strategies. Int. J. Logist. Res. Appl. 2023, 27, 2325–2350. [Google Scholar] [CrossRef]
  78. Institute for Health Metrics and Evaluation. US Health Map. Available online: https://vizhub.healthdata.org/subnational/usa (accessed on 24 April 2023).
Sustainability 17 03718 g001
Figure 1. Flowchart of methodology.
Figure 1. Flowchart of methodology.
Sustainability 17 03718 g001
Sustainability 17 03718 g002
Figure 2. (a) Last-mile delivery trucks and (b) general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Figure 2. (a) Last-mile delivery trucks and (b) general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Sustainability 17 03718 g002
Sustainability 17 03718 g003
Figure 3. PM2.5 emissions of (a) last-mile trucks and (b) general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Figure 3. PM2.5 emissions of (a) last-mile trucks and (b) general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Sustainability 17 03718 g003
Sustainability 17 03718 g004
Figure 4. PM2.5 exposure levels of general traffic (unit: μg/m3). A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Figure 4. PM2.5 exposure levels of general traffic (unit: μg/m3). A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Sustainability 17 03718 g004
Sustainability 17 03718 g005
Figure 5. Optimal number of clusters for the percentage of nonadult.
Figure 5. Optimal number of clusters for the percentage of nonadult.
Sustainability 17 03718 g005
Sustainability 17 03718 g006
Figure 6. Mortality outcomes compared to healthy air quality of (a) last-mile delivery trucks and (b) the general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Figure 6. Mortality outcomes compared to healthy air quality of (a) last-mile delivery trucks and (b) the general traffic. A: Dallas-Fort Worth International Airport, B: Addison Airport, C: Dallas Love Field, D: Dallas Executive Airport, and E: Union Pacific Mesquite Station.
Sustainability 17 03718 g006
Sustainability 17 03718 g007
Figure 7. Health impact changes when trucks are electrified under several scenarios.
Figure 7. Health impact changes when trucks are electrified under several scenarios.
Sustainability 17 03718 g007
Table 1. The list of input data for emission modeling.
Table 1. The list of input data for emission modeling.
TypeInput DataDescription
Vehicle CharacteristicsVehicle classification (vehicle composition)Passenger cars, trucks, buses
Rolling resistance
Fuel typeGasoline, diesel, electric, alternative fuels
Vehicle age distributionEngine wear, technology types
Driving Behavior (operating mode distribution)Speed, accelerationSecond-by-second speed, acceleration profiles
Regional typeUrban, rural
Restricted or unrestrictedStop-and-go (unrestricted), highway cruising (restricted)
Geometry InformationSlopeRoad gradient
Meteorology and TimeTemperature and relative humidity, hour and monthAir conditioning usage
Table 2. The list of input data for dispersion modeling.
Table 2. The list of input data for dispersion modeling.
TypeInput DataDescription
Link sourceEmissions rates on a linkEmissions
Coordinates of starting and ending points, number of lanes, geometry information, average vertical dispersionLink information (location, roadway geometry, average height of vehicles’ tailpipes)
Meteorological sourceWind speed, wind direction, temperature, precipitation amount, precipitation type, relative humidity, station pressureOutput from the AERMOD Meteorological Preprocessor (AERMET)
Receptor sourceReceptor coordinatesX coordinate, Y coordinate, Z coordinate
Table 3. The list of input data for BenMAP-CE v 1.5.8.
Table 3. The list of input data for BenMAP-CE v 1.5.8.
TypeInput DataDescription
Incidence/prevalence rateIncidence ratesMorbidity rates or mortality rates
PopulationPopulationNumber of racial populations
Health impact functionIncidence functionIncidence (i.e., diseases) function
Air qualityExposure levelsExposure levels
Table 4. Descriptive statistics of the meteorology data at the peak operating hours in Dallas County, Texas.
Table 4. Descriptive statistics of the meteorology data at the peak operating hours in Dallas County, Texas.
DescriptionMinMaxMeanStandard Deviation
Hourly temperature (K)269.2313.1292.719.19
Hourly relative humidity (%)1310065.6521.21
Sensible heat flux (W/m2)−6461.7−11.0038.09
Surface friction velocity (m/s)0.092.070.700.30
Wind speed (m/s)0.6914.084.992.14
Wind direction (degrees)1360169.0896.67
Station pressure (mb)9771019994.666.06
Note: 273.15 K = 0 °C = 32 °F where water is frozen.
Table 5. Average emissions by clusters at block groups.
Table 5. Average emissions by clusters at block groups.
VariableSocio-Economic Variable RangeAverage Emissions
(g/hour/Block Group)		Sample Size
MinimumMaximum
Seniors (%)Low011.6243.66408
Medium11.6526.243.28620
High26.3162.5739.79457
Nonadults (%)Low0.4617.8539.27851
Medium17.8731.1144.60485
High31.1359.4344.69149
Non-white (%)Low035.9546.18408
Medium36.0865.3260.09620
High65.4210068.84457
No vehicles (%)Low06.8843.001025
Medium6.9420.5743.05344
High20.6670.3846.21116
No employment (%)Low017.4461.18577
Medium17.4831.6859.32664
High31.7976.6957.05244
Median household income (USD)Low075,86544.69851
Medium76,153149,43243.11485
High150,408250,00135.58149
Table 6. Number of block groups which show better health impact when using electrified vehicles.
Table 6. Number of block groups which show better health impact when using electrified vehicles.
Electrified ScenarioLast-Mile Delivery Trucks (Percent)All TrucksGeneral Traffic
2%11 (0.71%)229 (14.76%)1236 (79.69%)
10%44 (2.84%)812 (52.35%)1533 (98.84%)
30%152 (9.8%)1504 (96.97%)1549 (99.87%)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Choi, J.; Hyun, K. The Impact of Last-Mile Delivery Fleet Electrification on Emissions, Dispersion, and Health: An Environmental Justice Analysis Based on Dallas County, Texas. Sustainability 2025, 17, 3718. https://doi.org/10.3390/su17083718

AMA Style

Choi J, Hyun K. The Impact of Last-Mile Delivery Fleet Electrification on Emissions, Dispersion, and Health: An Environmental Justice Analysis Based on Dallas County, Texas. Sustainability. 2025; 17(8):3718. https://doi.org/10.3390/su17083718

Chicago/Turabian Style

Choi, Jaesik, and Kate Hyun. 2025. "The Impact of Last-Mile Delivery Fleet Electrification on Emissions, Dispersion, and Health: An Environmental Justice Analysis Based on Dallas County, Texas" Sustainability 17, no. 8: 3718. https://doi.org/10.3390/su17083718

APA Style

Choi, J., & Hyun, K. (2025). The Impact of Last-Mile Delivery Fleet Electrification on Emissions, Dispersion, and Health: An Environmental Justice Analysis Based on Dallas County, Texas. Sustainability, 17(8), 3718. https://doi.org/10.3390/su17083718

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop