Next Article in Journal
Challenges and Disadvantages of PVA-Based Media Application in Wastewater Treatment: A Mini-Review
Previous Article in Journal
Bridging Knowledge Gaps and Charting Future Directions in Urban and Industrial Air Pollution Research
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Effects of Ambient Temperature on NOx Emissions from Heavy-Duty Diesel Vehicles Measured in Utah

by
Amber L. Gurecki Allen
*,
Darrell B. Sonntag
and
Gary A. Bishop
Department of Civil & Construction Engineering, Ira A. Fulton College of Engineering, Brigham Young University, Provo, UT 84602, USA
*
Author to whom correspondence should be addressed.
Retired, Chemistry and Biochemistry, College of Natural Sciences and Mathematics, University of Denver, Denver, CO 80208, USA.
Environments 2025, 12(9), 293; https://doi.org/10.3390/environments12090293
Submission received: 30 June 2025 / Revised: 14 August 2025 / Accepted: 21 August 2025 / Published: 26 August 2025

Abstract

This study investigates the effects of ambient temperature on NOx (NO + NO2) emissions from model year 2011 and later heavy-duty (HD) diesel vehicles. Emission measurements were collected in Perry, Utah, using the Fuel Efficiency Automobile Test (FEAT) remote sensing device. Data were limited to model year 2011 and later to focus on vehicles likely equipped with selective catalytic reduction (SCR) systems, which control tailpipe NOx emissions and are shown to be temperature sensitive. HD diesel vehicles measured in the winter of 2020 had consistently higher NOx emissions than those measured in the summer of 2023, most significantly for vehicles aged 0 to 3. A non-linear model fit to the data that accounts for age effects, predicts fleet-average NOx emissions to be two times higher at colder ambient temperatures (−4.4 °C, 24 °F) than warmer ambient temperatures (28.1 °C, 82.5 °F). The temperature effect from this study supports temperature effects observed in other studies measuring real-world emissions from HD diesel vehicles. One possible improvement to the accuracy of NOx emission inventories could be including a temperature effect for SCR-equipped HD diesel vehicles.

1. Introduction

Heavy-duty (HD) diesel vehicles contribute significantly to on-road emissions. Nitrogen oxides (NO + NO2), or NOx emissions are of particular concern, as they are precursors to the formation of ground-level ozone and fine particulate matter (PM2.5). NO2 is a critical air pollutant under the National Ambient Air Quality Standards (NAAQS) and causes direct harm to human health. NO2 is found in elevated levels near roadways [1], and those living near major roadways have an increased risk of asthma, cancer, and cardiovascular and pulmonary diseases [2]. In the United States, HD trucks only account for 5% of vehicles, but are estimated to contribute 24% of NOx emissions, as of 2023 [3,4]. To control NOx emissions from HD diesel trucks and to meet the EPA heavy-duty engine NOx emissions requirements [5], engine manufacturers began phasing in selective catalytic reduction systems (SCRs) to HD diesel vehicles beginning in engine model year 2010. SCR systems have been shown to be effective at reducing engine-out NOx emissions, with typical NOx conversion efficiencies ranging between 70% and 90% [6].
While SCR systems are an effective aftertreatment to reduce HD NOx emissions they have known limitations and temperature sensitivities. SCR systems require high temperatures for efficient NOx emission control. Several in-use studies have demonstrated that HD SCR systems are less effective at colder ambient temperatures [7,8] and during low-load driving cycles that have low exhaust temperatures [9,10,11]. Tan et al. (2019) [6] measured NOx conversion efficiency for 68 vehicles (engine model years 2012–2015) across a wide range of SCR inlet temperatures and operating conditions. At exhaust temperatures below 200 °C, NOx conversion efficiencies were on average 59%, compared to 73.7% for inlet temperature ranges of 200 °C to 250 °C, and 81.1% for temperatures over 250 °C. SCR systems will stop dosing urea, necessary for NOx reduction, below a certain exhaust temperature threshold, which is between 200 °C and 250 °C for some vehicles [12]. To account for this, the 2010 US EPA HD engine in-use emission standards are excluded when the exhaust temperature is below 250 °C [13]. HD diesel vehicles operating during colder temperatures or short distances, like short-haul drayage trucks, may experience this type of temperature-induced decreased SCR efficiency.
In addition to temperature effects, the deterioration of emissions control technology from continuous use may significantly impact NOx emissions. Several studies have investigated HD diesel NOx emission trends over time by vehicle or engine model year, but not specifically by vehicle age [14,15]. Additionally, several studies have found that tampering of SCR systems significantly increases tailpipe NOx emissions, though the prevalence of tampering in heavy-duty fleets remains uncertain [16,17,18]. Most studies have demonstrated the effects of temperature and aging on NOx emissions on only a small number of HD trucks. There is still a limited understanding of what the effects are on NOx emissions from a large fleet of SCR-equipped in-use HD diesel vehicles. The uncertainty about the effects of temperature, aging, and the frequency of tampering presents uncertainty to emission inventories from the HD freight sector. Estimates from vehicle emission models, including MOVES, developed by the US EPA [19], and EMFAC, developed by the California Air Resources Board [20], are used in statewide emission inventories to estimate pollution contributions from all types of vehicles. Currently, the vehicle emission rates for model year 2010 and later HD diesel vehicles in MOVES4 are based on manufacturer-run in-use testing of vehicles under 435,000 miles [21]. Understanding emissions outside of laboratory test settings is critical for providing insight into how vehicles are operating in the real world.
Several recent studies have evaluated the real-world effect of ambient temperature on NOx emissions from in-use vehicles. Hall et al. (2020) [22] observed that near-road NOx emissions were two times higher at −5 °C than 25 °C. Evans et al. (2019) [23] measured near-road NOx concentrations at three ranges of ambient temperature and observed approximately two times higher NOx at −5 °C than 15 °C. Grange et al. (2019) [24] measured light-duty (LD) vehicles in the UK from 2017 to 2018 using remote sensing and found NOx emissions were two times higher at 0 °C than 25 °C. Researchers in China have also investigated the effects of temperature on NOx emissions from HD diesel vehicles. He et al. (2024) [11] observed 17 HD diesel vehicles at different temperature and load conditions on a chassis dynamometer and concluded that hot-start emissions were 37% to 53% lower than the cold-start NOx emissions. Similarly, Wang et al. (2023) [25] used plume-chasing methods to observe NOx emissions from HD diesel vehicles, finding that emissions increased by 49% across the 35 °C to −3 °C temperature range.
In 2020, Bishop et al. (2022) [26] measured over 1500 HD trucks in the wintertime in Northern Utah and found they had consistently higher NOx emissions by model year than those measured in California [27]. Vehicles operating in Utah also had fuel-based NOx emission rates significantly higher compared to rates estimated by MOVES [26]. However, with a sample limited to the winter in cross-state comparisons, the elevated NOx emissions observed in Utah cannot be attributed to temperature effects, higher frequency of tampering, or larger impacts of aging in the Utah fleet.
This study provides a follow-up to the wintertime 2020 (Winter 2020) HD measurements collected by Bishop et al. (2022) [26]. HD diesel emission measurements were collected in the summer of 2023 (Summer 2023) at the same location in Perry, Utah. Both data sets are used to investigate whether NOx emissions from HD trucks in Utah differ between the winter and the summer.

2. Materials and Methods

The data were collected using Fuel Efficiency Automobile Test (FEAT), a remote sensing device designed to measure vehicle tailpipe emission across a single lane of traffic. FEAT measures the fuel-specific emission rate of carbon monoxide (CO), hydrocarbons (HC), nitrogen oxides (NO + NO2), and ammonia (NH3). This analysis focuses on NOx due to the large contribution that HD trucks make to total NOx emissions [28,29].
FEAT has been used to collect data in numerous locations and has been proven to be an effective and accurate method for remote sensing emission measurements [30,31], including for HD diesel vehicles [26,32,33,34]. For a detailed overview of FEAT, see Bishop and Stedman (1996) [35]. FEAT operates by quantifying the absorption of infrared (IR) and ultraviolet (UV) light by gases. FEAT also measures opacity, which can be used as an indication of particle emissions. A light source emitting IR and UV light is placed on one side of the roadway, directed at a light detector on the opposite side of the roadway. The detector is calibrated with reference gases to ensure measurement accuracy and is connected to a computer system. When a vehicle breaks the beam of light, the computer triggers the detector to start recording measurements. Measurements taken for 0.2 s before the vehicle passes are used as a reference. FEAT then measures every 0.01 s for 0.5 s after the vehicle has passed, collecting 50 data points. From these observations, ordinary least squares regression is used to estimate the ratio of each pollutant concentration to the concentration of CO2, which is corrected based on the roadside calibration factors. FEAT does not need to capture the full exhaust plume because each pollutant concentration is measured relative to CO2. Using an assumed 0.86 carbon density of gasoline and diesel fuel, (g-carbon/g-fuel) and the measured molar ratios of CO, HC, NO, NO2, and NH3 to CO2, an emission rate is calculated for each pollutant relative to fuel consumed (g/kg-fuel).
Pairs of infrared detectors are used to measure vehicle speed and acceleration as they pass by the FEAT sensors. At the moment the detector beam is broken, the camera is also triggered to capture an image of the vehicle’s license plate, which is saved to the corresponding vehicle measurement. License plate images are manually transcribed during data post-processing. For Utah-plated trucks, vehicle-specific (non-personal) information was obtained in cooperation with the Utah Department of Motor Vehicles. For non-Utah plated vehicles, license plate information was matched to vehicles using publicly available registration data [36].
The data were collected about an hour north of Salt Lake City, Utah, at the Port of Entry in Perry. At the Port, trucks are directed to one of two lanes, each with a different speed: 20 mph or 3 mph. Before approaching the on-ramp, all trucks merge into a single lane. A truck can also be asked to stop and pull in for an inspection before leaving the Port, but most remain off the freeway for less than 5 min. HD diesel vehicles were measured in Winter 2020 and Summer 2023 on the on-ramp as they were leaving the Port of Entry to merge back onto the interstate (I-15) heading south (see Figure A1).
HD diesel vehicles have either vertical (stack) or ground-level (turn-down) exhaust pipes. During the Winter 2020 campaign, vehicle measurements were collected in two configurations, Low FEAT and High FEAT. In the Low FEAT configuration, the FEAT detectors and light-source were positioned 0.3 m above the ground on the roadside to capture the ground-level exhaust pipes, which are either horizontal or downturned toward the ground (see Figure A2). In High FEAT, the FEAT detectors were placed on two, 4.3 m high scaffolding structures positioned on either side of the road to measure emission rates from trucks with vertical exhaust pipes (see Figure A3). Vehicles with stack exhaust pipes tended to be older than those with ground-level exhaust pipes. The majority of model year 2009 and older trucks had vertical exhaust pipes, while model years 2010 to 2014 had an equal proportion of vertical and ground-level pipes, and approximately two thirds of the model year 2015 and newer trucks had ground-level exhaust pipes (see Figure A4).
Due to the complexity and cost of setting up High FEAT, Summer 2023 measurements were only collected from the Low FEAT configuration, capturing emissions only from vehicles with ground-level exhaust pipes. The observations from the Summer 2023 campaign show that approximately 60% of trucks passing through the Perry Port of Entry had ground-level exhaust pipes. To maintain consistency, Winter 2020 data was limited to include only ground-level exhaust measurements. Additionally, this study limits measurement data to only include vehicles with chassis model years 2011 and newer for all analyses. Vehicle weight class is also limited to include only Class 6, Class 7, and Class 8 medium-heavy-duty (MHD) and heavy-heavy-duty (HHD) trucks. Though the EPA’s 2010 NOx emissions standard went into effect for engine model year 2010, previous studies have observed that engine model years tend to be 1 year older than the chassis model year on average [26]. Limiting this analysis to chassis model year 2011 and newer represents when SCR-equipped trucks are present in the fleet, and when ground-level exhaust tucks start to gain prevalence over elevated exhaust pipes (see Figure A4).
First, the sample means of each age group and accompanying 95% confidence intervals were calculated to evaluate differences between the winter and summer campaigns by age. Generalized additive models (GAMs) were then used to estimate smooth means by vehicle age without imposing a shape to the data (e.g., exponential or linear). By incorporating data from all age groups, the GAM predictions are expected to be more stable when applied across multiple samples of data, whereas the simple means will overfit the variation in each sample. Thus, the GAM should be a more robust predictor of the population means than the sample means. All model year 2011 and later data were used in the GAM fits, including older ages with limited data. The increased uncertainty in the GAM predictions for ages with limited data are reflected in larger confidence intervals. The GAMs are fit using the R package mgcv (version 1.9-3), and use generalized cross-validation to estimate optimal smoothing penalties to best predict each observation when they are is excluded from the data set [37].
The ratio between the mean emission rates in the winter and summer by vehicle age are calculated using the sample means and the GAM means. The winter-to-summer ratio can be considered an age-specific multiplicative temperature effect. Next, a non-linear regression model was used to estimate fuel-based NOx emission rates (gNOx/kg-fuel) by temperature and vehicle age. The model formulation was informed by visualizing the data by temperature and age, where there appears to be an exponential relationship between vehicle age and NOx emissions. The trends in the means by age also reflect a multiplicative relationship between age and temperature, with the impact of temperature appearing proportional to the NOx emissions at each age group. The exponential model was fit because it provides an average multiplicative temperature effect, which can be compared to temperature effects from other studies. Additionally, the non-linear model can be used to estimate the relative increase in emissions that is attributed to temperature or aging. While this model includes age as a factor, the effect of age is not fully explored in this paper. Including the age effect in the model is important for identifying the true temperature effect by accounting for the confounding age effect. An exploration of the effect of age on SCR deterioration can be found in Allen et al.
Existing methodologies for modeling the effect of temperature and deterioration on NOx emissions support the use of non-linear exponential model to the HD data in this study. Both exponential and multiplicative effects have been found to be a useful way to represent the effects of temperature and deterioration on vehicle emissions. For example, the MOVES model from the US EPA applies a multiplicative temperature adjustment for HD diesel vehicles starting in model year 2027 [7] and applies multiplicative adjustments to account for aging effects [21]. Laboratory testing of a model year 2027 prototype engine showed approximately two times higher emissions at cold ambient temperatures compared to the standard laboratory temperature [7]. The multiplicative effect of deterioration has not been well-researched for HD diesel vehicles but is well-documented for LD gasoline vehicles. MOVES applies a log-linear (exponential) approach to estimate the effects of LD deterioration, using a three-piece log-linear spline model to estimate NOx emissions across different age groups [38]. California’s EMFAC model alternatively adopted a power function fit to determine HD vehicle deterioration based on odometer mileage, which assumes that NOx emission rates continue to increase with age, but the rate of increase decreases with age [20].
The non-linear regression model takes the form shown below in Equation (1). The model was developed for this study as a way to quantify a temperature effect while also considering changes in NOx emissions as vehicles age. Applying this non-linear regression model allows an average temperature effect to be derived from the data across all vehicle ages. The temperature effect, e ( b · t e m p ) , estimates the multiplicative increase in NOx emissions relative to 0 °C. The relative temperature 0 °C, rather than the observed −4.4 °C, was chosen to serve as a simple baseline for interpreting the model coefficients. The age effect e ( c · a g e ) , estimates the multiplicative increase in NOx emissions relative to emissions at age 0.
N O x = a × e ( b · t e m p ) × e ( c · a g e )
where
  • NOx = estimated NOx emissions in g/kg-fuel,
  • a = baseline emissions at age 0 (years) and temperature 0 (°C),
  • b = the temperature coefficient (°C),
  • c = the age coefficient (years),
Though the non-linear regression model explains trends in the data, several limitations should be addressed. With only two sets of data collected in 2020 and 2023, emission data are observed only from model years 2011 to 2020 trucks at two different ages; therefore, there is a lack of emission data for all possible combinations of model year and chassis age. The exponential effect of age is sufficient to account for the age and model year effects in the data, but we do not recommend using the age effect beyond the observed locations and calendar years. Due to small sample sizes of older vehicles, the predicted NOx emissions for those ages may not be representative of the entire age fleets. The presence of few and high-emitting vehicles may skew the model-predicted age effects. The model format used explains the trends present in the data and provides an average temperature effect across all measured vehicle ages. However, applying this model format to data outside of this study is not recommended. Further limitations of the model can be found in the Section 4.

3. Results

Measurements were collected during two days of testing of the ground-level exhaust for both the Winter 2020 (~10 h) and Summer 2023 (~15 h) campaigns. The daily average temperature for both days of the Winter 2020 campaign was −4.4 °C (24 °F). The daily average temperature for both days of the Summer 2023 campaign was 28.1 °C (82.5 °F). Table 1 provides an overview of each campaign, including basic location information, vehicle counts, vehicles matched to registration data, and average emission rates of measured pollutants.
HD diesel vehicles measured during Winter of 2020 have significantly higher NOx emissions than those measured during Summer 2023. Mean NOx emissions and 95% confidence intervals (CI) were calculated for each vehicle age group. The average NOx emissions for the Winter 2020 campaign are 9.1 g/kg-fuel NOx (CI 7.8–10.4), while the Summer 2023 measured average was significantly lower at 5.7 g/kg-fuel NOx (CI 4.2–7.1) (see Table 1, Figure A5). Winter 2020 vehicles have 60% (1.6 times) higher mean NOx emissions than the vehicles measured in Summer 2023.
Measurements were also examined by model year to investigate trends in NOx emissions. While there is an overall trend of increased NOx emission for older model years, there is no significant difference or noticeable trend in NOx emission between the winter and summer measurements by model year (see Figure A6). At times, average NOx emission from Summer 2023 is higher than Winter 2020 emission, and vice versa. The reason there is no noticeable trend is likely due to the confounding effect of vehicle age, which is not illustrated by model year analysis alone, and uncertainties in the data as a result of limited measurements.
To compare the difference in NOx emissions between winter and summer, it is more appropriate to plot average emissions by vehicle age at the time it was measured. Table 2 shows the sample size for each model year with its corresponding age at the time of measurement for each campaign. This approach considers the effects of age on catalyst deterioration and NOx emissions. Figure 1 depicts mean NOx emission rates by vehicle age. For vehicles 0 to 9 years, the Winter 2020 campaign mean NOx emissions are consistently higher compared to vehicles of the same age measured during Summer 2023. There are distinct trends among groups of vehicle ages for both Winter 2020 and Summer 2023 NOx emission measurements. For vehicles 0 to 3 years, mean NOx emissions are level and keep a tight range. For Winter 2020 measurements, mean NOx emission rates for vehicles 0 to 3 years range between 4.6 and 6.0 g/kg-fuel, while Summer 2023 mean emission rates have a range between 1.1 and 3.5 g/kg-fuel for the same ages. NOx emissions then increase steadily for ages 4 to 6, but after age 6 they increase at a much steeper rate. Larger uncertainties exist for NOx emissions from older vehicles due to fewer vehicle measurements.
Two separate GAMs were fit separately to the data from the Winter 2020 and Summer 2023 campaigns to estimate mean NOx emissions as a function of age, as shown in Figure 2 and Table 3. The GAM smoothing has an impact on the estimated mean emission rates. For example, the GAM mean for age 9 from the Winter 2020 campaign is noticeably lower than the sample mean, which is only based on three observations. Additionally, the GAM mean for NOx emission rates at age 6 in the Summer 2023 campaign is more than twice as high as the sample mean.
Table 3 contains the sample means and the GAM means for each campaign and age, and the ratio of the means between winter and summer campaigns by age. In general, the ratio or multiplicative temperature effect is largest for the newest vehicles (age 0 to 3 years), while the absolute difference in mean emissions is greatest for the older vehicles. HD vehicles aged 0 to 3 had 2.2 to 4.5 times higher emissions in the winter, while HD vehicles aged 3 to 9 had only 1.3 to 2.5 times higher emissions in the winter (see Table 3). These results are consistent with Figure 1, which shows significant differences in the mean NOx emissions for ages 0, 1, and 3, but not for the other ages.
Next, a non-linear model was fit to the data using the Non-linear Least Squares (nls) function in the stats package of R [39]. All the coefficients were highly significant (p-value < 0.05), as shown in Table 2. The model included the inputs of vehicle age by year, and temperature in degrees Celsius (°C). The model parameters can be used separately to identify the fleet-average multiplicative effect that both temperature and age have on NOx emissions, using our model form shown in Equation (2).
N O x = 3.417 × e ( 0.021 · t e m p ) × e ( 0.287 · a g e )
Figure 3 shows the model adequately fits the mean HD NOx emission rates across vehicle age for each campaign. The model fit to the HD data suggests that there is a strong correlation between NOx emissions and increasing vehicle age across both campaigns. Table 4 shows the output for each model term, including the model estimate, 95% confidence interval, and p-value. The model coefficient for temperature (−0.021) has a highly significant p-value of <2 × 10−16. The model coefficient for age (0.287) has a highly significant p-value of <3 × 10−12. Figure 3 illustrates the predicted mean HD NOx emissions as a function of temperature across −4.4 °C to 28.1 °C for each vehicle age, plotted alongside the individual vehicle measurements. There is a large variation in the individual emission rates at each age and temperature. The multiplicative age effect between the two temperatures is consistent across ages but most pronounced in older vehicles (ages 7–9), where predicted mean NOx emissions are approximately 50 g/kg-fuel in winter compared to 25 g/kg-fuel in summer. Sparse data for age 9 contributes to uncertainty in predicted NOx emissions.
The model estimated the fleet-average multiplicative temperature effect to be 2.0 (CI 1.6–2.4) between average temperatures of the Summer 2023 and Winter 2020 campaigns, meaning that the model estimates that NOx emissions are more than two times higher at −4.4 °C (24 °F) than 28.1 °C (82.5 °F) when controlling for the differences in vehicle ages between the two campaigns. Figure 4 illustrates the predicted NOx emissions (g/kg-fuel) across a temperature range present in this study, with each box representing a different vehicle age. Points on the graph are individual NOx emission measurements from each campaign. Gray bands around the center line represent 95% confidence intervals of the predicted means, calculated using a bootstrap resampling method.
From the model, vehicles are predicted to have 31 (CI 22–47) times higher NOx emissions at age 12 than they did at age 0 (see Table 3). The observed deterioration in our model is much steeper than what current emission models predict. Several factors that could influence these extreme predictions include the lack of data for older ages, deterioration of the SCR emission control technology, or could possibly be an indication of SCR tampering. Further analysis of the effect of aging seen in this study is explored in Allen et al.

4. Discussion

Age by age, NOx emissions were consistently higher in winter than summer. This study also demonstrated that the multiplicative temperature effect is larger for newer vehicles. New and fully functional SCR systems should be working effectively at high exhaust temperatures. When the SCR system is not operating due to low exhaust temperatures, there should be a large increase in NOx emissions. For aged SCR systems with degraded performance, the NOx emissions should increase when the SCR is not operating at low temperatures, but the relative increase should be lower, since the emissions are not fully controlled at high exhaust temperatures.
The fleet-average temperature effect estimated from the non-linear regression model in this research compares well with what other studies have seen. Table 5 summarizes several studies that have investigated temperature effects on NOx emissions from diesel vehicles, both HD and LD. The observed NOx temperature effect was summarized for each study (Equation (A1)), and temperature ranges from other studies compare well to the temperatures observed in this study. The temperature effect was then calculated from Equation (2) across the same temperature ranges from each study’s observation. The modeled temperature in this study compares well to what other studies have observed and thus is believed to be applicable beyond the vehicle operating conditions observed at the Perry weigh station. The difference in altitudes across studies is not believed to have an impact on observed temperature effects (see Appendix A.1).
Although age was in important factor included in the non-linear regression model, a limited sample size of older, SCR-equipped vehicles limits the ability to draw solid conclusions on the effects of aging on NOx emissions from HD diesel vehicles for the oldest vehicles in this study (ages 9–12). The age effect predicted by the model presented in this study is significantly larger than age effects determined by emission models and other researchers. For example, the MOVES and EMFAC models predict that after 12 years of aging, emissions from HD diesel vehicles would increase by 1.7 and 1.9, respectively (calculated for chassis model years 2011 to 2023). Testing of three HD diesel vehicles in China showed that after driving 200,000 km (equivalent to 2.3 years of aging), NOx emissions increased by 10–30% [40]. A further investigation of the impact of aging on NOx emissions from SCR- and non-SCR-equipped vehicles will be conducted using the data presented here, along with additional data, in Allen et al.
All models are a simplification of reality, and this model has limitations that should be considered. The model formulation assumes that the baseline emission a (age 0, temperature 0 °C) is constant across model years 2011 to 2024, and that the age and temperature effect are the same for all model years. The primary motivation for these assumptions is the limitations of the data, but they are consistent with what other studies have found. Engine manufacturers are given time to certify engines to new standards, and therefore, not all 2011 engines are certified to the 0.2 g/bhp-h standard. Chassis model year was not included in this model, and it is assumed that the included age effect likely also includes a model year effect (see Figure A6). For this model, age was chosen over model year because the Summer 2023 measurements have 2 years and 7 months additional aging compared to the Winter 2020 measurements. Additionally, previous data analysis also has not detected significant differences in emissions by model year for SCR-equipped trucks. The analysis of NOx emissions conducted for the MOVES model from HD trucks certified to the 0.2 g/bhp-h standard has shown no significant differences in NOx emissions from 2010 to 2013 and 2014 to 2015 engine model year groups [21]. These trucks were measured by engine manufacturers as part of the in-use HD program with mileage below 435,000 miles [21]. Additionally, an analysis of NOx emissions from SCR-equipped HD vehicles by Tan et al. (2019) [6] showed that while emissions varied significantly depending on vehicle use and make, there was no significant difference in NOx emissions by engine model year (2012–2015).
Existing emission models do not currently include a temperature effect for NOx emissions from HD diesel vehicles. This and other studies have observed temperature dependencies for NOx emission reduction equipment from HD diesel vehicles. Failing to include a temperature effect for NOx could indicate that states are underestimating emissions during colder periods of the year when calculating their emission inventories and creating state implementation plans (SIPs) that detail mitigation efforts. Including a temperature sensitivity for NOx emissions from HD diesel vehicles could improve the accuracy of emission inventories and improve NOx reduction strategies going forward. More discussion on the impact of temperature on emission models and state emission inventories can be found in Allen et al.
Table 5. Overview of previous studies investigating the effects of ambient temperature on NOx emissions from HD diesel vehicles.
Table 5. Overview of previous studies investigating the effects of ambient temperature on NOx emissions from HD diesel vehicles.
StudySummaryTemperature RangeStudy NOx
Temperature Effect
Equation (2) Temperature Effect 2
Evans et al., 2019 [23]Measured near-road NOx concentrations at three ranges of ambient temperature. Observed significantly higher NOx during colder temperatures.−15 °C to
15 °C
~1.8 11.9
Grange et al., 2019 [24]On-road remote sensing of LD diesel vehicles across the UK from 2017 to 2018. Higher NOx at colder ambient temperatures.0 °C to
25 °C
2.01.7
Hall et al., 2020 [22]Used near-road and aircraft observations to observe CO, NOx and CO2. Found that NOx emissions increase at colder temperatures.−5 °C to
25 °C
2.01.9
Söderena et al.,
2020 [41]
Measured tailpipe emissions from one car equipped with SCR during 2018 and 2019. No significant difference in emissions across temperature range for the vehicle.7 °C to
17 °C
No significant observed effect1.2
US EPA, 2023 [21]EPA’s MOVES model temperature adjustment for HD NOx starting in model year 2027. Adjustment is based on in-lab testing of a prototype model year 2027 engine at cold ambient temperatures in 2022. Found variation in temperature effect by regulatory class.−4.4 °C to
25 °C
1.2 to 1.31.8
Wang et al., 2019 [8]Measured vehicle emissions at Fort McHenry Tunnel in Baltimore, MD, in 2015. HD vehicles were estimated separately from LD and had higher emissions at colder temperatures.7.1 °C to 29.7 °C1.71.6
Wærsted et al., 2022 [42]Measured road traffic NOx emissions for 4 years at 46 sites from 2016 to 2019 in Norway. Observed higher NOx emissions at colder temperatures.−7 °C to
14 °C
2.71.6
1 Temperature effect was estimated from graphical representation of data.

Author Contributions

Conceptualization, A.L.G.A., D.B.S. and G.A.B.; methodology A.L.G.A., D.B.S. and G.A.B.; software, G.A.B.; validation, A.L.G.A., D.B.S. and G.A.B.; formal analysis, A.L.G.A. and D.B.S.; investigation, A.L.G.A.; resources, A.L.G.A. and D.B.S.; data curation, A.L.G.A., D.B.S. and G.A.B.; writing—original draft preparation, A.L.G.A.; writing—review and editing, A.L.G.A., D.B.S. and G.A.B.; visualization, A.L.G.A. and D.B.S.; supervision, D.B.S.; project administration, D.B.S.; funding acquisition, D.B.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was performed under funding provided by a Mentored Research Grant from the BYU Ira A. Fulton College of Engineering and the Utah Governor’s Office of Economic Opportunity Strategic Innovation Grant Contract 240632114.

Data Availability Statement

The Winter 2020 campaign data can be accessed at https://digitalcommons.du.edu/feat/ (accessed on 12 May 2024). The Summer 2023 data can be accessed at https://scholarsarchive.byu.edu/data/ (accessed on 20 August 2025).

Acknowledgments

Dylan Fitt, Dave Andrada, and Andrew Coleman conducted collection and processing of the data as student research assistants. Rodney May and Clayton Harrison and staff in the Civil & Construction Engineering Department built, obtained, and repaired equipment for us to transport and operate the FEAT at BYU. Stephen Goodrich and staff from the Utah Department of Transportation enabled us to conduct the Summer 2023 campaign at the Perry Port of Entry. The staff from the Utah Department of Motor Vehicles provided vehicle information from the recorded license plates. Finally, we recognize the late Don Stedman for conceiving and developing the FEAT which has been used to measure real-world emissions for nearly 40 years.

Conflicts of Interest

The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Abbreviations

The following abbreviations are used in this manuscript:
NOxNitrogen oxides
FEATFuel Efficiency Automobile Test
HDHeavy-duty
SCRSelective catalytic reduction
COCarbon monoxide
HCHydrocarbons
NH3Ammonia
LD Light-duty

Appendix A

Figure A1. Sampling Location. Perry Port of Entry in Perry, Utah, about 1 h north of Salt Lake City.
Figure A1. Sampling Location. Perry Port of Entry in Perry, Utah, about 1 h north of Salt Lake City.
Environments 12 00293 g0a1
Figure A2. Ground-Level Exhaust. Images of ground-level or turn-down exhaust pipes underneath a truck (a) and while in operation during additional sampling in West Valley City, Utah (b).
Figure A2. Ground-Level Exhaust. Images of ground-level or turn-down exhaust pipes underneath a truck (a) and while in operation during additional sampling in West Valley City, Utah (b).
Environments 12 00293 g0a2
Figure A3. Stack Exhaust. Image of an elevated or stack exhaust pipe during a different sampling campaign in West Valley City, Utah.
Figure A3. Stack Exhaust. Image of an elevated or stack exhaust pipe during a different sampling campaign in West Valley City, Utah.
Environments 12 00293 g0a3
Figure A4. Exhaust Location Distribution. Count of elevated and ground-level exhaust pipe HD diesel trucks measured in Winter 2020. The ground-level exhaust counts have been multiplied by a factor of 2.2 to account for the differences in time spent collecting the data at different heights during the campaign (10 h at ground level [0.3 m], 22 h elevated [4.3 m]). Elevated-exhaust vehicles make up a majority of older model years, while starting with model year 2014, the majority is taken over by the ground-level exhaust vehicles.
Figure A4. Exhaust Location Distribution. Count of elevated and ground-level exhaust pipe HD diesel trucks measured in Winter 2020. The ground-level exhaust counts have been multiplied by a factor of 2.2 to account for the differences in time spent collecting the data at different heights during the campaign (10 h at ground level [0.3 m], 22 h elevated [4.3 m]). Elevated-exhaust vehicles make up a majority of older model years, while starting with model year 2014, the majority is taken over by the ground-level exhaust vehicles.
Environments 12 00293 g0a4
Figure A5. Fleet Average NOx. Mean NOx emissions (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns. Fleetwide average NOx emissions are 1.7 times higher in the winter than in the summer. Error bars are 95% confidence intervals of the mean.
Figure A5. Fleet Average NOx. Mean NOx emissions (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns. Fleetwide average NOx emissions are 1.7 times higher in the winter than in the summer. Error bars are 95% confidence intervals of the mean.
Environments 12 00293 g0a5
Figure A6. Boxplots by Chassis Model Year. Boxplot spread of NOx emissions by chassis model year for the Winter 2020 and Summer 2023 campaigns. Numbers underneath each box are the number of vehicles in each sample group (see also Table 2).
Figure A6. Boxplots by Chassis Model Year. Boxplot spread of NOx emissions by chassis model year for the Winter 2020 and Summer 2023 campaigns. Numbers underneath each box are the number of vehicles in each sample group (see also Table 2).
Environments 12 00293 g0a6
Figure A7. Non-linear Model Predictive Curve. The individual prediction curves for mean NOx emissions across the two Utah campaigns, with 95% confidence intervals represented by the gray-shaded area calculated using a resampling method with the nlsBoot() function from the nlstools() package in R.
Figure A7. Non-linear Model Predictive Curve. The individual prediction curves for mean NOx emissions across the two Utah campaigns, with 95% confidence intervals represented by the gray-shaded area calculated using a resampling method with the nlsBoot() function from the nlstools() package in R.
Environments 12 00293 g0a7
Equation (A1): How the calculated temperature effect from our model was obtained, using other studies and high and low temperature as inputs. The temperature coefficient calculated by our model is −0.021 (see Equations (1) and (2), and Table 4).
T e m p e r a t u r e   E f f e c t = e 0.020880   ×   l o w e s t   t e m p e r a t u r e   ( ° C )   e 0.020880   ×   h i g h e s t   t e m p e r a t u r e ( ° C )

Appendix A.1

Altitude can have a varied effect on HD vehicle NOx emissions [33,43], as well as SCR operation. A high altitude results in lower ambient pressure, lower air-to-fuel ratio, lower engine power, delayed ignition, decrease in ignition temperature, increase in in-cylinder temperature, changes to the combustion duration, and exhaust temperature [44,45,46,47]. These impacts can either increase or decrease engine-out NOx emissions. Engine manufacturers adjust the engine controls (such as fuel injection timing) to balance fuel economy, engine power, and engine-out emissions. For SCR-equipped engines, thermal management strategies, such as intake and exhaust throttling and post-engine fueling, are used to increase exhaust temperature and SCR NOx conversion efficiency while decreasing NOx emissions. However, these strategies have a difference impact on NOx emissions at different altitudes [48]. The impact of altitude on modern diesel engines with SCR aftertreatment is complex because it depends not only on how altitude impacts the inputs to the engine and emission control system (e.g., decreased pressure), but primarily on how the engine and emission control system has been programmed to respond to the higher altitude. These controls could vary by engine manufacturer, model, model year, and the operating mode of the vehicle.
Zheng et al. (2023) [46] tested a SCR-equipped China VI heavy-duty diesel engine at different simulated altitudes and found that NOx emissions remained well controlled up to 9842 feet/3000 m but had low SCR NOx conversion efficiency and subsequent high tailpipe NOx emissions above 13,123 feet/4000 m. Chen et al. (2023) [48] found that light-duty diesel vehicles measured at high elevations (1560 m) had greater NOx emissions by a factor of 2 compared to those measured at low elevations (300 m to 400 m) in Switzerland. After driving a SCR-equipped HD vehicle across the United States at varying altitudes, Kappanna et al. (2013) [49] observed NOx emissions to be an order of magnitude higher at 5000 feet (1524 m) compared to the 2010 certification standard, noting that the US EPA 2010 HD NOx standards do not apply at high altitudes (>5500 feet, 1676 m) and manufacturers can deactivated the SCR system at those altitudes. The elevation at our Utah location (4301 feet, 1131 m) is well below the 5500 feet exception in the regulation.

Appendix A.2

As shown in Table 1, the Winter 2020 campaign has a near-zero value for average vehicle acceleration, while the Summer 2023 campaign has a slightly more positive average acceleration. This difference in average acceleration does not impact the differences observed in NOx emissions between the Winter 2020 and Summer 2023 campaigns. Figure A8 illustrates measured vehicle acceleration against NOx emissions for individual vehicles. Both campaigns have a similarly bell-shaped distribution of emission to acceleration, with the majority of points falling between −2.5 to 2.5 mph/s. Vehicles with extreme high and extreme low measured acceleration have relatively low measured NOx, while the majority of high emitters fall in the middle of the distribution. When examining acceleration trends across each vehicle age, the distribution for each age reflected the fleet-wide trends.
Figure A8. Acceleration and NOx. Acceleration against measured NOx emissions from the Winter 2020 and Summer 2023 campaigns. Not all valid NOx measurements are present in this figure, as some vehicles may not obtain a valid acceleration reading from the FEAT system.
Figure A8. Acceleration and NOx. Acceleration against measured NOx emissions from the Winter 2020 and Summer 2023 campaigns. Not all valid NOx measurements are present in this figure, as some vehicles may not obtain a valid acceleration reading from the FEAT system.
Environments 12 00293 g0a8

References

  1. Beckerman, B.; Jerrett, M.; Brook, J.R.; Verma, D.K.; Arain, M.A.; Finkelstein, M.M. Correlation of nitrogen dioxide with other traffic pollutants near a major expressway. Atmos. Environ. 2008, 42, 275–290. [Google Scholar] [CrossRef]
  2. Brugge, D.; Durant, J.L.; Rioux, C. Near-highway pollutants in motor vehicle exhaust: A review of epidemiologic evidence of cardiac and pulmonary health risks. Environ. Health 2007, 6, 23. [Google Scholar] [CrossRef] [PubMed]
  3. US EPA. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990–2022—Main Text; 430-R-24-004; US EPA: Washington, DC, USA, 2024. Available online: https://www.epa.gov/system/files/documents/2024-04/us-ghg-inventory-2024-main-text_04-18-2024.pdf (accessed on 21 August 2025).
  4. US EPA. Air Quality Trends Show Clean Air Progress. Available online: https://gispub.epa.gov/air/trendsreport/2024/ (accessed on 30 June 2025).
  5. US EPA. Final Rule for Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles. Available online: https://www.epa.gov/regulations-emissions-vehicles-and-engines/final-rule-greenhouse-gas-emissions-standards-and-fuel (accessed on 2 April 2024).
  6. Tan, Y.; Henderick, P.; Yoon, S.; Herner, J.; Montes, T.; Boriboonsomsin, K.; Johnson, K.; Scora, G.; Sandez, D.; Durbin, T.D. On-Board Sensor-Based NOx Emissions from Heavy-Duty Diesel Vehicles. Environ. Sci. Technol. 2019, 53, 5504–5511. [Google Scholar] [CrossRef] [PubMed]
  7. US EPA. Emission Adjustments for Onroad Vehicles in MOVES4. 2023; EPA-420-R-23-021. Available online: https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P10186KX.pdf (accessed on 9 February 2024).
  8. Wang, X.; Khlystov, A.; Ho, K.-F.; Campbell, D.; Chow, J.C.; Kohl, S.D.; Watson, J.G.; Lee, S.-C.F.; Chen, L.-W.A.; Lu, M. Real-World Vehicle Emissions Characterization for the Shing Mun Tunnel in Hong Kong and Fort McHenry Tunnel in the United States; HEI Report 199; Health Effects Institute: Boston, MA, USA, 2019; Available online: https://www.healtheffects.org/publication/real-world-vehicle-emissions-characterization-shing-mun-tunnel-hong-kong-and-fort (accessed on 20 August 2025).
  9. Quiros, D.C.; Thiruvengadam, A.; Pradhan, S.; Besch, M.; Thiruvengadam, P.; Demirgok, B.; Carder, D.; Oshinuga, A.; Huai, T.; Hu, S. Real-World Emissions from Modern Heavy-Duty Diesel, Natural Gas, and Hybrid Diesel Trucks Operating Along Major California Freight Corridors. Emiss. Control Sci. Technol. 2016, 2, 156–172. [Google Scholar] [CrossRef]
  10. Quiros, D.C.; Smith, J.; Thiruvengadam, A.; Huai, T.; Hu, S. Greenhouse gas emissions from heavy-duty natural gas, hybrid, and conventional diesel on-road trucks during freight transport. Atmos. Environ. 2017, 168, 36–45. [Google Scholar] [CrossRef]
  11. He, L.; Li, G.; Wu, X.; Zhang, S.; Tian, M.; Li, Z.; Huang, C.; Hu, Q.; Wu, Y.; Hao, J. Characteristics of NOX and NH3 emissions from in-use heavy-duty diesel vehicles with various aftertreatment technologies in China. J. Hazard. Mater. 2024, 465, 133073. [Google Scholar] [CrossRef]
  12. Majewski, A.W. Urea Dosing and Injection Systems. Available online: https://dieselnet.com/tech/cat_scr_diesel_urea_dosing.php (accessed on 14 July 2024).
  13. CFR. Not-to-Exceed Test Procedures; CFR: New York, NY, USA, 2017; Volume 40, CFR 86.1370. Available online: https://www.ecfr.gov/current/title-40/section-86.1370 (accessed on 20 August 2025).
  14. Preble, C.V.; Harley, R.A.; Kirchstetter, T.W. Plume Capture Measurement of Vehicle Emissions at the Caldecott Tunnel for Heavy-Duty Emission Program Development & Verification; Final Report. 2025. Available online: https://ww2.arb.ca.gov/sites/default/files/2025-07/Research%20Contract%2020RD004%20Final%20Report.pdf (accessed on 13 August 2025).
  15. Preble, C.V.; Harley, R.A.; Kirchstetter, T.W. Control Technology-Driven Changes to In-Use Heavy-Duty Diesel Truck Emissions of Nitrogenous Species and Related Environmental Impacts. Environ. Sci. Technol. 2019, 53, 14568–14576. [Google Scholar] [CrossRef]
  16. Fulper, C.; US EPA, J.W.; Sabisch, M.; Ruminski, B.; Conference, O. Developing Tools to Quickly Find Potentially High Emitting Heavy-Duty Vehicles. 2024. Available online: https://www.cert.ucr.edu/sites/default/files/2024-03/1.2_carl_fulper_presentation.pdf (accessed on 7 July 2024).
  17. Braun, C.; Badshah, H.; Hosseini, V.; Jin, L.; Miller, J.; Rodríguez, F. Heavy-Duty Emissions Control Tampering in Canada—International Council on Clean Transportation. Available online: https://theicct.org/publication/hdv-emissions-tampering-can-mar22/ (accessed on 23 April 2024).
  18. US EPA. Tampered Diesel Pickup Trucks: A Review of Aggregated Evidence from EPA Civil Enforcement Investigations; US EPA: Washington, DC, USA, 2020. Available online: https://www.epa.gov/sites/default/files/2021-01/documents/epaaedletterreportontampereddieselpickups.pdf (accessed on 16 May 2024).
  19. US EPA. Latest Version of MOtor Vehicle Emission Simulator (MOVES). Available online: https://www.epa.gov/moves/latest-version-motor-vehicle-emission-simulator-moves (accessed on 3 June 2024).
  20. CARB. EMFAC2021 Model and Documentation | California Air Resources Board. Available online: https://ww2.arb.ca.gov/our-work/programs/msei/emfac2021-model-and-documentation (accessed on 3 June 2024).
  21. US EPA. Exhaust Emission Rates for Heavy-Duty Onroad Vehicles in MOVES4. Available online: https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P10186TB.pdf (accessed on 23 April 2024).
  22. Hall, D.L.; Anderson, D.C.; Martin, C.R.; Ren, X.; Salawitch, R.J.; He, H.; Canty, T.P.; Hains, J.C.; Dickerson, R.R. Using near-road observations of CO, NOy, and CO2 to investigate emissions from vehicles: Evidence for an impact of ambient temperature and specific humidity. Atmos. Environ. 2020, 232, 117558. [Google Scholar] [CrossRef]
  23. Evans, G. Near-Road Air Pollution Pilot Study Final Report. 2019. Available online: http://hdl.handle.net/1807/96917 (accessed on 19 June 2025).
  24. Grange, S.K.; Farren, N.J.; Vaughan, A.R.; Rose, R.A.; Carslaw, D.C. Strong Temperature Dependence for Light-Duty Diesel Vehicle NOx Emissions. Environ. Sci. Technol. 2019, 53, 6587–6596. [Google Scholar] [CrossRef]
  25. Wang, H.; Zhang, S.; Wu, X.; Wen, Y.; Li, Z.; Wu, Y. Emission Measurements on a Large Sample of Heavy-Duty Diesel Trucks in China by Using Mobile Plume Chasing. Environ. Sci. Technol. 2023, 57, 15153–15161. [Google Scholar] [CrossRef] [PubMed]
  26. Bishop, G.A.; Haugen, M.J.; McDonald, B.C.; Boies, A.M. Utah Wintertime Measurements of Heavy-Duty Vehicle Nitrogen Oxide Emission Factors. Environ. Sci. Technol. 2022, 56, 1885–1893. [Google Scholar] [CrossRef]
  27. Haugen, M.J.; Bishop, G.A.; Thiruvengadam, A.; Carder, D.K. Evaluation of Heavy- and Medium-Duty On-Road Vehicle Emissions in California’s South Coast Air Basin. Environ. Sci. Technol. 2018, 52, 13298–13305. [Google Scholar] [CrossRef]
  28. UDEQ. Statewide Emissions Inventories. Utah Department of Environmental Quality. Available online: https://deq.utah.gov/air-quality/statewide-emissions-inventories (accessed on 23 May 2024).
  29. US EPA. 2020 National Emissions Inventory (NEI) Data. Available online: https://www.epa.gov/air-emissions-inventories/2020-national-emissions-inventory-nei-data (accessed on 3 June 2024).
  30. Popp, P.J.; Bishop, G.A.; Stedman, D.H. Development of a High-Speed Ultraviolet Spectrometer for Remote Sensing of Mobile Source Nitric Oxide Emissions. J. Air Waste Manag. Assoc. 1995 1999, 49, 1463–1468. [Google Scholar] [CrossRef]
  31. Burgard, D.A.; Dalton, T.R.; Bishop, G.A.; Starkey, J.R.; Stedman, D.H. Nitrogen dioxide, sulfur dioxide, and ammonia detector for remote sensing of vehicle emissions. Rev. Sci. Instrum. 2006, 77, 014101. [Google Scholar] [CrossRef]
  32. Haugen, M.J.; Bishop, G.A. Repeat Fuel Specific Emission Measurements on Two California Heavy-Duty Truck Fleets. Environ. Sci. Technol. 2017, 51, 4100–4107. [Google Scholar] [CrossRef]
  33. Bishop, G.A.; Morris, J.A.; Stedman, D.H.; Cohen, L.H.; Countess, R.J.; Countess, S.J.; Maly, P.; Scherer, S. The Effects of Altitude on Heavy-Duty Diesel Truck On-Road Emissions. Environ. Sci. Technol. 2001, 35, 1574–1578. [Google Scholar] [CrossRef]
  34. Bishop, G.A.; Schuchmann, B.G.; Stedman, D.H. Heavy-Duty Truck Emissions in the South Coast Air Basin of California. Environ. Sci. Technol. 2013, 47, 9523–9529. [Google Scholar] [CrossRef] [PubMed]
  35. Bishop, G.A.; Stedman, D.H. Measuring the Emissions of Passing Cars. Acc. Chem. Res. 1996, 29, 489–495. [Google Scholar] [CrossRef]
  36. Oregon DOT. Trucking Online. Available online: https://www.oregontruckingonline.com/cf/MCAD/pubMetaEntry/carrierInfoByVeh/ (accessed on 22 April 2024).
  37. Wood, S.N. Generalized Additive Models: An Introduction with R, 2nd ed.; Chapman and Hall/CRC: Boca Raton, FL, USA, 2017. [Google Scholar]
  38. US EPA. Exhaust Emission Rates for Light-Duty Onroad Vehicles in MOVES4. 2023; EPA-420-R-23-028. Available online: https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P101872R.pdf (accessed on 22 April 2024).
  39. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistial Computing: Vienna, Austria, 2023; Available online: https://www.R-project.org/ (accessed on 16 June 2024).
  40. Lyu, L.; Ji, Z.; Yin, H.; Wang, J.; Yang, W.; Tan, J.; Hao, L.; Wang, X.; Wang, H.; Ge, Y.; et al. NOx emission deterioration in modern heavy-duty diesel vehicles based on long-term real driving measurements. Environ. Res. 2023, 232, 116396. [Google Scholar] [CrossRef] [PubMed]
  41. Söderena, P.; Laurikko, J.; Weber, C.; Tilli, A.; Kuikka, K.; Kousa, A.; Väkevä, O.; Venho, A.; Haaparanta, S.; Nuottimäki, J. Monitoring Euro 6 diesel passenger cars NOx emissions for one year in various ambient conditions with PEMS and NOx sensors. Sci. Total Environ. 2020, 746, 140971. [Google Scholar] [CrossRef] [PubMed]
  42. Wærsted, E.G.; Sundvor, I.; Denby, B.R.; Mu, Q. Quantification of temperature dependence of NOx emissions from road traffic in Norway using air quality modelling and monitoring data. Atmos. Environ. X 2022, 13, 100160. [Google Scholar] [CrossRef]
  43. Human, D.M.; Ullman, T.L.; Baines, T.M.; Human, D.M.; Ullman, T.L.; Baines, T.M. Simulation of High Altitude Effects on Heavy-Duty Diesel Emissions. In Earthmoving Industry Conference & Exposition; SAE International: Warrendale, PA, USA, 1990. [Google Scholar]
  44. US EPA. Control of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements. Available online: https://www.govinfo.gov/content/pkg/FR-2001-01-18/pdf/01-2.pdf (accessed on 17 June 2024).
  45. Wang, X.; Ge, Y.; Yu, L.; Wang, X.; Ge, Y.; Yu, L. Combustion and Emission Characteristics of a Heavy-Duty Diesel Engine at Idle at Various Altitudes. SAE Int. J. Engines 2013, 6, 1145–1151. [Google Scholar] [CrossRef]
  46. Zheng, Y.M.; Xie, L.B.; Liu, D.Y.; Ji, J.L.; Li, S.F.; Zhao, L.L.; Zen, X.H. Emission Characteristics of Heavy-Duty Vehicle Diesel Engines at High Altitudes. J. Appl. Fluid Mech. 2023, 16, 2329–2343. [Google Scholar] [CrossRef]
  47. Nie, X.; Bi, Y.; Liu, S.; Shen, L.; Wan, M. Impacts of different exhaust thermal management methods on diesel engine and SCR performance at different altitude levels. Fuel 2022, 324, 124747. [Google Scholar] [CrossRef]
  48. Chen, Y.; Wu, X.; Hu, K.; Borken-Kleefeld, J. Nox emissions from diesel cars increase with altitude. Transp. Res. Part Transp. Environ. 2023, 115, 103573. [Google Scholar] [CrossRef]
  49. Kappanna, H.; Besch, M.; Thiruvengadam, A.; Delgado, O.; Cozzolini, A.; Carder, D.; Gautam, M.; Hu, S.; Huai, T.; Ayala, A.; et al. Greenhouse Gas Emissions of MY 2010 Advanced Heavy Duty Diesel Engine Measured Over a Cross-Continental Trip of USA; SAE International: Warrendale, PA, USA, 2013. [Google Scholar]
Figure 1. Mean NOx emissions (g/kg-fuel) for each vehicle age. Numbers under each bar indicate the number of vehicles in each age group. The error bars represent 95% confidence intervals of the mean, calculated using a t-distribution. Wide error bars for age 9 are indicative of relatively small sample sizes for that age in the data.
Figure 1. Mean NOx emissions (g/kg-fuel) for each vehicle age. Numbers under each bar indicate the number of vehicles in each age group. The error bars represent 95% confidence intervals of the mean, calculated using a t-distribution. Wide error bars for age 9 are indicative of relatively small sample sizes for that age in the data.
Environments 12 00293 g001
Figure 2. GAM predictions. NOx emission (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns by vehicle age. The means by age and campaign are displayed with x marks. Smooth mean emission rates by age are predicted using GAMs and shown with the dark gray lines. Shaded light-gray areas are 95% confidence intervals of the mean predicted values from the GAMs.
Figure 2. GAM predictions. NOx emission (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns by vehicle age. The means by age and campaign are displayed with x marks. Smooth mean emission rates by age are predicted using GAMs and shown with the dark gray lines. Shaded light-gray areas are 95% confidence intervals of the mean predicted values from the GAMs.
Environments 12 00293 g002
Figure 3. Non-linear model-predicted curve overlain on individual measurements for each temperature (campaign) and vehicle age. Gray band represents 95% confidence intervals of the mean, which were calculated using bootstrap resampling.
Figure 3. Non-linear model-predicted curve overlain on individual measurements for each temperature (campaign) and vehicle age. Gray band represents 95% confidence intervals of the mean, which were calculated using bootstrap resampling.
Environments 12 00293 g003
Figure 4. Non-linear model-predicted curve across temperatures (−4 °C to 29 °C) overlain on individual measurements for each temperature (campaign). Each grid represents the model predicted curve for the indicated vehicle age (years). The gray shaded region around each line represents a 95% confidence interval of the predicted mean, calculated using bootstrap resampling.
Figure 4. Non-linear model-predicted curve across temperatures (−4 °C to 29 °C) overlain on individual measurements for each temperature (campaign). Each grid represents the model predicted curve for the indicated vehicle age (years). The gray shaded region around each line represents a 95% confidence interval of the predicted mean, calculated using bootstrap resampling.
Environments 12 00293 g004
Table 1. Informational overview of campaigns. The 95% confidence intervals were calculated using a t-distribution of the mean.
Table 1. Informational overview of campaigns. The 95% confidence intervals were calculated using a t-distribution of the mean.
CampaignWinter 2020Summer 2023
LocationPerry, UTPerry, UT
Dates12/10/20–12/11/207/31/23–8/01/23
Daily Average Temperature (°C/°F)−4.4, 2428.1, 82.5
Daily Average Humidity (%)61.329.4
Location Elevation
(Meters/Feet)
1131, 43011131, 4301
Number of Vehicles Matched to Records549527
Average Vehicle
Acceleration (mph/s)
0.030.45
Average Vehicle Speed (mph)26.428.6
Location Slope (degrees)
Median Vehicle Age (Years)22
Median Chassis Model Year20182021
Mean gNOx/kg-fuel
(95% CI)
9.1
(7.8–10.4)
5.7
(4.2–7.1)
Mean gNO 1/kg-fuel
(95% CI)
8.6
(7.4–9.8)
5.0
(3.7–6.2)
Mean gNO2/kg-fuel
(95% CI)
0.5
(0.4–0.6)
0.7
(0.4–1.0)
1 NO in NO2 molecular weight equivalent.
Table 2. Chassis model year and age distribution for each campaign, with vehicle counts included for each sample group.
Table 2. Chassis model year and age distribution for each campaign, with vehicle counts included for each sample group.
Chassis Model YearWinter 2020Summer 2023
AgenAgen
2024--038
2023--0118
2022--191
2021050260
2020093362
20191112443
2018282537
2017345617
2016460731
2015554811
2014624911
2013716102
2012810114
201193122
Table 3. GAM predictions. Mean NOx emissions (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns by vehicle age. The ratio of the winter/summer emission rates by age is calculated using the sample means and the GAM means.
Table 3. GAM predictions. Mean NOx emissions (g/kg-fuel) from the Winter 2020 and Summer 2023 campaigns by vehicle age. The ratio of the winter/summer emission rates by age is calculated using the sample means and the GAM means.
Winter 2020Summer 2023Ratio of Sample Means
(Winter/Summer)
Ratio of GAM Means
(Winter/Summer)
Age Sample Means GAM Means Sample Means GAM Means
04.64.71.11.04.24.5
16.05.91.31.74.53.4
25.45.53.52.51.62.2
36.06.72.13.02.92.2
411.611.28.58.11.41.4
514.714.312.411.11.21.3
616.216.93.06.95.52.5
725.225.414.413.31.71.9
839.943.428.631.61.41.4
980.470.444.839.91.81.8
10--12.439.1--
11--45.135.3--
12--14.021.5--
Table 4. Non-linear model results. The 95% CIs were calculated using bootstrap resampling and prediction packages from R (nlsBoot, nlsBootPredict) [39].
Table 4. Non-linear model results. The 95% CIs were calculated using bootstrap resampling and prediction packages from R (nlsBoot, nlsBootPredict) [39].
TermParameterEstimate95% Confidence Intervals
(Lower, Upper)
p-Value
aIntercept (gNOx/kg-fuel)3.417(2.706, 4.231)<2 × 10−16
bTemperature (°C)−0.021(−0.027, −0.015)<3 × 10−12
cAge (Years)0.287(0.259, 0.321)<2 × 10−16
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

Gurecki Allen, A.L.; Sonntag, D.B.; Bishop, G.A. Effects of Ambient Temperature on NOx Emissions from Heavy-Duty Diesel Vehicles Measured in Utah. Environments 2025, 12, 293. https://doi.org/10.3390/environments12090293

AMA Style

Gurecki Allen AL, Sonntag DB, Bishop GA. Effects of Ambient Temperature on NOx Emissions from Heavy-Duty Diesel Vehicles Measured in Utah. Environments. 2025; 12(9):293. https://doi.org/10.3390/environments12090293

Chicago/Turabian Style

Gurecki Allen, Amber L., Darrell B. Sonntag, and Gary A. Bishop. 2025. "Effects of Ambient Temperature on NOx Emissions from Heavy-Duty Diesel Vehicles Measured in Utah" Environments 12, no. 9: 293. https://doi.org/10.3390/environments12090293

APA Style

Gurecki Allen, A. L., Sonntag, D. B., & Bishop, G. A. (2025). Effects of Ambient Temperature on NOx Emissions from Heavy-Duty Diesel Vehicles Measured in Utah. Environments, 12(9), 293. https://doi.org/10.3390/environments12090293

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