Investigating Real-World Emissions of China ’ s Heavy-Duty Diesel Trucks : Can SCR Effectively Mitigate NOx Emissions for Highway Trucks ?

The Euro IV emission standards for heavy-duty diesel vehicles require a substantial reduction in NOx emissions, spurring the introduction of selective catalytic reduction (SCR) systems. However, previous studies have found unsatisfactory control of NOx emissions for SCR-equipped urban buses, which has raised concern among policy-makers and researchers regarding the capability of SCR adopted by heavy-duty diesel trucks (HDDTs) in real-world applications. This study tested sixteen HDDTs in China between 2010 and 2014, including six SCR-equipped Euro IV HDDTs using a portable emissions measurement system (PEMS). On-road emission measurement results show no significant difference in NOx emissions between Euro II and Euro III HDDTs. In contrast, we observed a substantial reduction in real-world NOx emissions, as low as 25.4 g kg-fuel, for six SCR-equipped HDDTs (Euro IV) compared with those without SCR systems (Euro II and Euro III), providing an overall reduction of ~50%. However, real-world brake-specific NOx emission factors for the SCRequipped HDDTs were higher by ~45% than the lab test limit of 3.5 g kWh due to off-cycle NOx emissions, indicating the importance of introducing real-world emission test requirements for HDDTs. Due to the introduction of SCR systems, distinctive impacts from real-world operating conditions are observed for HDDTs. For example, fuel-based NOx emission factors steadily decrease as the vehicle speed increases due to higher exhaust temperatures, which improve the efficiency of SCR systems.


INTRODUCTION
Heavy-duty diesel vehicles (HDDVs) have been identified as one of the most important contributors to air pollution, health impacts and climate change (Johnson et al., 2009;Wu et al., 2011;Silverman et al., 2012;Bond et al., 2013;Zhang et al., 2013;Wu et al., 2014) and have contributed to an ascending trend in China's total anthropogenic NO x emissions (Richter et al., 2005;Wu et al., 2012;Zhao et al., 2013).Due to the severe environmental problems related to NO x emissions, China's "Twelfth Five-Year Plan (2011)(2012)(2013)(2014)(2015)" set a clear goal of mitigating total NO x emissions by 10% over five years (NPC, 2011).More recently, China's State Council released the "Action Plan for Air Pollution Prevention and Control" in September 2013, containing the strictest control measures in recent years (State Council, 2013), which was followed by the release of local action plans at provincial or city levels (e.g., Beijing, Zhang et al., 2014a).HDDVs are key targets for NO x emission control because of their high emission factors and vehicle-specific use intensity.According to the official report released by the MEP (2012), total NO x emissions from on-road vehicles in China were 6.4 Tg in 2011, 67% of these emissions were contributed by HDDVs (including public transit buses, highway trucks and coach buses).Moreover, Wu et al. (2012) estimated that the total NO x emissions of the national HDDV fleet in 2009 exceeded the amount provided in the official report based on a large sample of HDDVs tested with on-road emission measurement by ~1.0 Tg.Even for megacities, such as Beijing and Guangzhou, which are pioneers for controlling vehicle emissions in China (Zhang et al., 2013(Zhang et al., , 2014a)), their annual reduction in vehicle emissions of NO x has been significantly lower than for other pollutants (e.g., CO, THC and PM 2.5 ) over recent years.Zhang et al. (2014a) warned that the mitigation of NO x emissions from gasoline vehicles is partially offset by increased NO x emissions from HDDVs due to their unsatisfactory real-world NO x emission controls.Therefore, it is urgently necessary to strengthen the control of real-world NO x emissions from China's HDDVs through effective technology measures and policy interventions, in particular for heavy-duty diesel trucks (HDDTs), which account for a large portion of China's HDDV fleet (MEP, 2012).
From the perspective of China's vehicle emission control history, the Euro IV (equivalent to China IV) emission standards for HDDVs were first planned for implementation in 2010.The most significant advancement originating from the Euro IV standards is that selective catalytic reduction (SCR) would become a prevailing after-treatment technology to control exhaust NO x emissions.However, the promulgation of the Euro IV emission standards for HDDVs was postponed by the MEP until July 2013 for several reasons (Zhang et al., 2014a).For example, a sufficient supply of low-sulfur fuel that complied with the Euro IV standards (e.g., 50 ppm or lower) was limited to major cities (e.g., Beijing and Shanghai) and unavailable nationwide (Zhang et al., 2010;Yue et al., 2015).Furthermore, the shortage of urea (e.g., AdBlue) filling facilities discouraged its use.In China, several previous on-road studies using portable emission measurement systems (PEMS) have reported real-world NO x emissions of Euro I to Euro III diesel trucks (Liu et al., 2009;Huo et al., 2012;Wu et al., 2012).In Beijing, recent on-road studies evaluated SCR-equipped Euro IV and Euro V urban diesel buses and reported unexpectedly high realworld NO x emissions due to their low exhaust temperatures under congested driving conditions and low loads (Wu et al., 2012;Fu et al., 2013;Zhang et al., 2014b).Therefore, there is considerable uncertainty in the control of NO x emissions for Euro IV HDDVs and in-use PEMS measurement programs are needed (Lowell and Kamakaté, 2013;Vlachos et al., 2014).Considering the very large population and different operating conditions of trucks compared urban busses and that very few studies have been conducted on SCR-equipped HDDTs in China, it is important to investigate real-world NO x emissions for diesel trucks in China equipped with SCR systems.
This paper aims to provide policy-makers with a timely investigation of real-world NO x emissions from modern diesel trucks in China, particularly for SCR-equipped Euro IV trucks.We employed a PEMS to obtain high temporal resolution emission profiles for sixteen HDDTs, covering emission standard categories ranging from Euro II to Euro IV.We calculated the NO x emission factor for each vehicle as a weighted average of various road types (e.g., urban roads, suburban roads and freeways) to explore the impacts of tightened emission standards and real-world operating conditions.With a focus on SCR-equipped HDDTs, this study provides the analysis of major factors (e.g., road type, vehicle speed, and load mass) affecting NO x emissions for Euro IV highway trucks.

Experimental Section
On-road emission tests were performed in Beijing and Guangdong between 2010 and 2014.The tested vehicles, including five Euro II HDDTs, five Euro III HDDTs and six Euro IV HDDTs (see Table 1), were recruited from local freight companies.Each vehicle tested has been properly maintained and serviced.In recent years, individual truck weights in China's truck fleet have significantly increased (Wu et al., 2016).For example, more than 82% of total truck sales had gross vehicle weights (GVWs) exceeding 19 tons in 2009.Moreover, for tractors, nearly 98% of total sales had GVWs exceeding 25 tons.Therefore, we tested three tractors (i.e., IV-1, IV-2 and IV-3) with GVWs exceeding 25 tons.Heavier Euro IV HDDT samples would result in relatively high fuel consumption and NO x emission when developing emission inventories for China's truck fleet.Euro II and Euro III HDDTs are not equipped with advanced after-treatment devices to control NO x emissions.In contrast, Euro IV trucks use urea-based SCR systems and on-board diagnostic (OBD) equipment.In particular, these SCR systems adopt vanadiumbased catalysts that have a higher tolerance to sulfur (e.g., up to 350 ppm) than copper-or iron-based catalysts.In accordance with the requirement of the regulations regarding real-world emissions measurement (BEPB&BAQTS, 2013), market fuel from the certified refueling stations was used and the fuel quality complied with China III or China IV standard (see Table S1).The difference is mostly sulfur content (< 350 ppm for China III and < 50 ppm for China IV), which should be applicable to SCR devices (Walker et al., 2004;Jiang et al., 2010).All drivers were instructed to maintain their usual driving behaviors during the test trips.
To explore the impacts of operating conditions on vehicle emissions, the test routes for each truck included various typical road types that can be categorized into urban roads, suburban roads and freeways.The measurement duration of each test for each vehicle lasted more than 2 hours with a hot start, in which the measurements for each typical road category continued for more than 30 min to ensure adequate real-world vehicle driving data and emission profiles.Average vehicle speeds on urban roads, suburban roads and freeways were approximately 20 km h -1 , 50 km h -1 and 70 km h -1 , respectively (see Table 2).In addition to driving conditions, load mass is another factor influencing realworld fuel consumption and emissions.Long-distance freight trucks in China often overload to increase their operation profit.Therefore, we set various load weights for the tested vehicles, ranging from 0.5 to 20.5 tons, which is illustrated in Table 2.These loads represent load rates ranging from less than 5% to approximately 120%.Among the tested HDDTs, five vehicle samples (i.e., II-1, II-2, III-1, III-2 and IV-1; see The load mass was estimated based on the mass of the PEMS, the weight of the driver and the mass of the artificial load, among which the sum of PEMS mass and driver weight is ~0.5 tons.temperature, relative humidity and air pressure between each comparison test pair were within 5°C, 10% and 1 mbar, respectively, to eliminate impacts from environmental conditions on vehicle emissions. A PEMS (SEMTECH-DS, Sensors Inc.) was applied to measure on-road emissions of gaseous pollutants for the diesel trucks.SEMTECH-DS employed a non-dispersive ultraviolet (NDUV) module for separate measurement of exhaust NO and NO 2 concentrations.Moreover, a heated flame ionization detector (HFID) measured the total hydrocarbon (THC) content, and a non-dispersive infrared (NDIR) analyzer was employed to measure CO and CO 2 concentrations (Liu et al., 2009;Wu et al., 2012).To ensure measurement accuracy, the analyzers were zero and span calibrated before each test.Real-time exhaust flow rates were recorded via an exhaust flow meter (SEMTECH-EFM).Instantaneous vehicle speed and location information (longitude, latitude and altitude) were recorded using a GPS receiver integrated into the SEMTECH-DS.Moreover, the vehicle interface module of the SEMTECH-DS can record real-time signals regarding the engine operation mode (e.g., engine speed, torque and power) for the six Euro IV diesel trucks equipped with OBD systems.

Data Processing
Emission rates (unit in g s -1 ) for each HDDT were collected at one-second intervals in this study.We first calculated distance-based emission factors (g km -1 ) of all gaseous pollutants (e.g., NO x , CO, THC, and CO 2 ) by road type for each individual vehicle based on the instantaneous emission rate profiles and driving conditions: where, EF dist.,i,j is the average distance-based emission factor for road type i and pollutant j, g km -1 ; t is the total test time on road type i, s; ER i,j is the instantaneous emission rate for road type i and pollutant j, g s -1 ; and v i is the instantaneous vehicle speed for road type i, km h -1 .
To eliminate effects of vehicle size (e.g., engine power rating and/or GVW) between each of the tested trucks, we also calculated the fuel-based emission factor for better comparison with a carbon balance method (Wu et al., 2012;Zhang et al., 2014c) where EF fuel,i,j is the average fuel-based emission factor for road type i and pollutant j, g kg-fuel -1 ; W C is the carbon mass ratio in diesel fuel, i.e., 0.866 (Zhang et al., 2014b, c); and EF dist.,i,CO 2 , EF dist.,i,CO and EF dist.,i,THC are the average distance-based emission factors for CO 2 , CO and THC, respectively, for road type i, g km -1 .
To eliminate the impacts of driving conditions on each HDDT sample, we further weighted the overall average emission factors for the tested vehicles based on their emission factors for various road types: where EF j is the weighted average emission factor for pollutant j, g kg-fuel -1 or g km -1 ; and EF f,j , EF sub,j and EF urb,j are the average emission factors of pollutant j for freeways, suburban roads and urban roads, respectively, g kg-fuel -1 or g km -1 .We referred to the local standard DB11/965-2013 of Beijing (BEPB&BAQTS, 2013) for the recommended weighting coefficients.

Real-World NO x Emissions Associated with Tightened Emission Standards
Fig. 1 presents the weighted NO x emission factors for the tested HDDTs.The weighted fuel-based NO x emission factors are 49.1 ± 11.4, 47.9 ± 17.7 and 25.4 ± 3.3 g kg-fuel -1 , respectively, for Euro II to Euro IV HDDTs.Little reduction in NO x emissions from Euro II to Euro III HDDTs is observed.Furthermore, a wide range in fuel-based NO x emission factors for Euro II and Euro III HDDTs, i.e., from 27 to 67 g kg-fuel -1 , are identified.This finding might be attributed to several factors, including different engine technologies, emission control strategies, load masses and maintenance performances (Almén and Erlandsson, 2011;Ligterink et al., 2012;Carrese et al., 2013).In contrast, for the six Euro IV HDDTs equipped with urea-SCR systems, their fuel-based NO x emission factors on real roads are significantly lower, by an average of approximately 50%, than those without SCR systems.To compare with the emission standard limits (units in g kWh -1 ), we also derived realworld brake-specific NO x emission factors for the six Euro IV HDDTs based on the engine operation data.For the Euro II and Euro III HDDTs, we applied engine combustion efficiency values of 209 and 206 g-fuel kWh -1 , respectively (CRAES, 2010), to convert between the different units.The real-world brake-specific NO x emission factors for Euro II, III and IV HDDTs, which are listed in Table 3, are higher than their corresponding laboratory emission limits (i.e., 7.0, 5.0 and 3.5 g kWh -1 for Euro II, III and IV HDDTs based on laboratory engine bench tests, respectively) by 47%, 97% and 45%, respectively, similar to the gaps between real-world emissions and regulatory limits previously reported by Wu et al. (2012).
Nevertheless, a recent investigation revealed that numerous counterfeit HDDTs are available in China's market, which claims to meet the Euro IV standard with SCR systems.However, no SCR systems or urea tanks can be found on these trucks (CCTV, 2014).These counterfeit Euro IV HDDTs have been determined to have emission performances similar to Euro III HDDTs or even older models.Consequently, more stringent inspections during the conformity of production and enhanced in-use emission testing compliance programs (e.g., PEMS and remote sensing regulation) for Euro IV HDDVs are urgently needed in China.
Published studies regarding real-world NO x emissions of HDDVs are summarized in Table 3 from a broad review of the literature.There have been very few studies, and only two of them measured the Euro IV trucks in China including this study.In the current study, the focus of the investigation  2. NO x is presented as the sum of NO and NO 2 , which are also shown separately.(Liu et al., 2009) ~48 (37-60) (Euro II) (mostly GVW< 5 tons) ~50 (44-58) (Euro III) Five cities in China; Trucks (Huo et al., 2012) 44 ± 25 (Euro II) 51 ± 15 (Euro III) 36 ± 11 (Euro IV) b Beijing, China; Buses (Wu et al., 2012) 46.0 ± 16.0 (Euro II) 46.6 ± 4.1 (Euro III) 41.7 ± 6.1 (Euro IV) 6.6 and 9.5 for two Euro IV buses Beijing, China; Buses (Liu et al., 2011) 7.86 and 11.05 for two Euro III buses 6.88 and 7.08 for two Euro IV buses Beijing, China; Buses (Fu et al., 2013) 9.61~12.44(Euro IV) Beijing, China; Buses (Zhang et al., 2014b) 30.0 (Euro V) 5.6 (Euro V) EU; Trucks (Bonnel et al., 2011) 4.7 (3.1-8.5)(Euro IV) Netherlands; Trucks (Ligterink et al., 2009) 8.4 (city) and 7.1 (freeway) for Euro III; 6.5 (city) and 2.6 (freeway) for Euro V; Sweden; HDDTs c (Erlandsson et al., 2008) 4.80 (Euro IV) US; HDDTs (Krishnamurthy et al., 2007) 4. 3-5.1 (MY 2002) d US; HDDTs (Johnson et al., 2009) 3.80 , average of in-use NTE events (MY 2004) d Note: a Brake-specific emission factors are converted based on an assumed engine efficiency of 209 g-fuel kWh -1 for Euro II trucks and 206 g-fuel kWh -1 for Euro III trucks; b average NO x emission factor of two Euro IV trucks tested in Jinan, Shandong, in 2011; c net vehicle weight is 15 tons, and the vehicles were equipped with exhaust gas recirculation (EGR) systems; and d MY 2002 andMY 2004 in the US are comparable to the Euro IV standards.was on their performance in terms of controlling NO x emissions.Similar to previous PEMS test results (Liu et al., 2009;Huo et al., 2012;Wu et al., 2012), we added new on-road evidence of the unsuccessful NO x emission control for pre-Euro IV HDDVs due to the absence of effective after-treatment devices (e.g., SCR systems) and a strong inuse inspection program.Furthermore, Wang et al. (2012) employed a mobile platform to measure on-road NO x emissions of 440 chased trucks in Beijing and Chongqing and found that the median fuel based NO x emission factors were 40-50 g kg-fuel -1 , highly coincident with PEMS test results.Krishnamurthy et al. (2007) identified a trade-off between higher fuel economy and controlling NO x emissions under real-world conditions.Unfortunately, fuel economy is one of the most important concerns for truck manufacturers and owners; thus, without an effective in-use compliance program, NO x emissions can easily run out of control.
For SCR-equipped trucks (Euro IV), the average NO x emission factor (i.e., 25.4 ± 3.3 g kg-fuel -1 ) is comparable to previous test results in the EU and US for trucks complying with Euro IV or similar emission standards (Erlandsson et al., 2008;Johnson et al., 2009;Bonnel et al., 2011).Previous studies have indicated only small emission reductions for Euro IV buses equipped with urea-SCR systems compared with Euro III buses (Liu et al., 2011;Wu et al., 2012;Fu et al., 2013).Low exhaust temperatures (below 200°C during most of the operating time) under congested urban driving conditions and low loads result in unsatisfactory performance of SCR systems, which has also been shown for Euro IV and Euro V HDDVs in low-speed urban driving circumstances (Ligterink et al., 2009;Carslaw et al., 2011;Velders et al., 2011).Unlike the operating conditions of urban buses, higher driving speeds and loads for the tested SCR-equipped trucks lead to significantly higher exhaust temperatures, which activate the SCR systems.This condition facilitated the effectiveness for SCR to convert NO x to N 2 .In the following section, additional analysis on the impacts of real-world operating conditions is provided.
Direct NO 2 emissions (i.e., primary NO 2 ) from modern diesel vehicles are of significant concern due to their oxidation potential to aggravate secondary pollution, e.g., ozone.We present separate primary NO and NO 2 emission factors in Fig. 1.The averaged primary NO 2 emission factors for Euro II, Euro III and Euro IV HDDTs are 1.8 ± 1.6, 1.7 ± 1.2 and 0.1 ± 0.1 g kg-fuel -1 , respectively, indicating that the average fractions of primary NO 2 in total NO x are 3.7%, 3.5% and 0.4%.This finding agrees well with previous results for Euro III (3.2%) and Euro IV (1.0%) buses (Wu et al., 2012).Lower primary NO 2 emissions from Euro IV HDDTs might be attributed to the introduction of SCR systems, without the combined use of a diesel oxidation catalyst (DOC) and a diesel particle filter (DPF).Therefore, the primary NO 2 fraction for Euro IV HDDTs is lower than those in other geographical regions (e.g., EU) because the application of DOC and DPF systems would cause concern regarding the increased primary NO 2 fraction (e.g., up to 50%) (Carslaw et al., 2011;Hu et al., 2012;Velders et al., 2011).Upcoming Euro VI emission standards for HDDVs, which are scheduled to be implemented no later than 2017 in Beijing (probably two or three years later for nationwide implementation), will set very stringent limits on mitigate diesel particle emissions and will promote broad application of DOC and DPF systems.The real-world primary NO 2 emissions of Euro VI HDDVs must be monitored when DOC and DPF systems become widely available in the market.

Impact of Driving Conditions and Load Mass on NO x Emissions
Table 4 summarizes the fuel-based NO x emission factors for the tested HDDTs associated with emission standards and road types.For HDDTs without SCR systems (Euro II and Euro III), little variation in the fuel-based NO x emission factors is found among the three different roads (average of 45-52 g kg-fuel -1 ).This finding indicates that for HDDTs without SCR systems, the impacts of driving conditions on NO x emissions are consistent with the trend in fuel consumption.However, the pattern for HDDTs with SCR (Euro IV) is very different.A significant decrease in the fuel-based NO x emission factors for HDDTs with SCR systems is found because their average speeds increase from urban roads (~20 km h -1 ) to freeways (~70 km h -1 ).Fig. 2 presents the average fuel-based NO x emission factors according to instantaneous speed bins, which clearly shows the same trend as in Table 4.For HDDTs without SCR systems, the average fuel-based NO x emission factors are not sensitive to changes in vehicle speed.However, for HDDTs with SCR systems, fuel-based NO x emissions are reduced from ~55 g kg-fuel -1 under extremely congested conditions (e.g., 0-10 km h -1 ) to approximately 20 g kg-fuel -1 on freeways.
Similar to Velders et al. (2011) and Verbeek et al. (2010), our study also identifies how driving conditions (e.g., road type and vehicle speed) are related to the real-world performance of SCR systems.If the average emission level of Euro III HDDTs is seen as the baseline, the NO x emission reductions for SCR-equipped Euro IV HDDTs are 18%, 49% and 58% on urban roads, suburban roads and freeways, respectively, which is because exhaust temperatures at low  speeds might be too low for proper functioning of SCR systems.A similar conclusion has been previously identified as the major reason for poor NO x emission control performance in Euro IV diesel buses (Wu et al., 2012;Fu et al., 2013).Alternatively, using the OBD-derived engine combustion efficiency, the brake-specific NO x emission factors for the Euro IV HDDTs are 8.5 g kWh -1 , 5.0 g kWh -1 and 3.8 g kWh -1 on urban roads, suburban roads and freeways, respectively.As a result, only the brake-specific NO x emission factors on freeways are close to the emission limit (3.5 g kWh -1 ).Similar findings have also been reported based on real-world PEMS studies in Europe (Velders et al. 2011;Lowell and Kamakaté, 2013).A more effective in-use compliance program, e.g., the real-world emission test regulation, is required to control off-cycle emissions (Lowell and Kamakaté, 2013), especially for modern HDDVs using advanced after-treatment devices (e.g., SCR systems).Consequently, in the development of a high-resolution vehicle emission inventory for future scenarios that has a considerable penetration of advanced HDDVs with SCR systems, we suggest that the impacts from driving conditions should be carefully considered in the spatial allocation process associated with different road types (Zheng et al., 2014).Five representative HDDTs were selected from the sixteen tested trucks to explore the real-world NO x emission change with load mass.Due to larger individual emissions difference in each group meeting Euro II and Euro III HDDTs (see Fig. 1), one high emission and one low emission trucks were picked up.However, the average NO x emission factors were similar for the six Euro IV HDDTs, thus only one were selected to analyze the impact of load mass.Fig. 3 presents the fuel-based NO x emission factors under high and low loads.The high load mass is equal to around twice the low load mass, but unlike four Euro II and III HDDTs, the low and high loads of the Euro IV HDDT were higher, because of its large vehicle size.In general, the fuel consumption for all five HDDTs increases by 1-9% (see Table 2) under high loads compared to low loads.For HDDTs without SCR systems (Euro II and Euro III), the fuel-based NO x emission factors increase by 1-31% for a doubling in the load mass, which is because higher engine loads generally increase the combustion temperature, thereby enhancing NO x formation (Almén and Erlandsson, 2011).
For the HDDT with SCR (i.e., IV-1) in this study, the fuel consumption also increases under higher loads.However, unlike the HDDTs without SCR systems, the NO x emission factor under high loads is approximately 22 g kg-fuel -1 , a reduction of 18% relative to low loads.Lowell and Kamakaté (2013) also reported the same effect for Euro V diesel trucks with SCR systems, suggesting a NO x emission reduction of ~20% under a 100% load compared to a 50% load.We verified the measured exhaust temperatures using a thermocouple temperature probe integrated into the SEMTECH-EFM; the average temperature was found to increase from 199°C to 258°C for a doubling in the load mass.The higher temperatures enhance the efficiency of SCR systems (Fu et al., 2013).Zhang et al. (2014b) also found that the use of on-board air-conditioning systems for SCR-equipped Euro IV diesel hybrid buses significantly increase real-world fuel consumption and resulted in better SCR performance due to higher exhaust temperatures.However, the HDDVs complying with the US 2010 standard might achieve very low NO x emissions (e.g., lower than 1 g km -1 ), although some challenges still remain under unfavorable operating conditions, which are associated with lower exhaust temperatures (e.g., low speed, idle conditions and low load mass) (Misra et al., 2013;Spreen et al., 2014).

CONCLUSIONS
We measured real-world emissions of sixteen HDDTs complying with Euro II to Euro IV emission standards using a PEMS.The fuel-based NO x emission factors for all of the tested trucks were calculated based on emission profiles measured at one-second intervals.Furthermore, we analyzed the impacts of real-world operating conditions on NO x emissions, with a special focus on six HDDTs equipped with SCR systems.
The average fuel-based NO x emission factors for Euro II, III and IV HDDTs were 49.1 ± 11.4, 47.9 ± 17.7 and 25.4 ± 3.3 g kg-fuel -1 , respectively.We found no significant reduction in real-world NO x emissions as the emission standards tightened from Euro II to Euro III.In contrast, thanks to SCR systems, the average fuel-based NO x emission factors for the six SCR-equipped HDDTs (Euro IV) decreased by ~50% compared to HDDTs without SCR systems (Euro II and Euro III), and NO 2 emissions were also considerably lower for Euro IV HDDTs.These findings suggest a better performance of SCR systems for HDDTs relative to those installed in the urban bus fleet.However, we also found that the real-world brake-specific NO x emissions were higher than the emission limits by 47%, 97% and 45% for the tested Euro II, Euro III and Euro IV HDDTs, respectively, indicating a substantial gap between real-world emissions and laboratory-derived emission limits.
We also highlighted the importance of real-world operating conditions on NO x emissions for SCR-equipped HDDTs.
For driving conditions, we found a significant decrease in the fuel-based NO x emission factors for SCR-equipped HDDTs with increased vehicle speeds, which differed from the findings for HDDTs without SCR systems.The fuelbased NO x emission factors for the six HDDTs with SCR systems on suburban roads and freeways were lower than those on urban roads by ~41% and ~55%, respectively.
Our study provides first-hand data for further improvement in local vehicle emission inventories and policy-making related to controlling vehicle emissions in China.

Fig. 1 .
Fig. 1.Weighted fuel-based NO x emission factors for each tested HDDT.The emission factors are the test results for low loads, when trucks were tested with varying load mass.Specific load mass information is given in Table2.NO x is presented as the sum of NO and NO 2 , which are also shown separately.

Fig. 2 .
Fig. 2. Fuel-based NO x emission factors for the tested HDDTs in various vehicle speed bins.
Fig.3.Fuel-based NO x emission factors under high and low loads for five HDDTs (Note: Four Euro II and III HDDTs (i.e., II-1, II-2, III-1 and III-2) emission tests were performed in Beijing and One Euro IV HDDTs (i.e., IV-1) in Guangdong).

Table 2 )
were tested under two load mass conditions for comparison.Discrepancies in the average ambient

Table 1 .
Summary of tested diesel trucks.Roman numeral and Arabic numeral refer to the emission stage and the vehicle number, respectively.For example, II-1 means the 1st tested truck complying with Euro II; b On road emission were performed in plains of Beijing and Guangdong province; c The three Euro IV HDDTs were recruited from local freight companies and were operating in a demonstration program for Euro IV HDDVs in the Guangdong province; d vanadium-based SCR catalyst, model SCRV020, manufactured by Cummins; e vanadium-based SCR catalyst, model SCRV021, manufactured by Cummins; and f vanadium-based SCR catalyst, model 1208010-43A, manufactured by First Automobile Works (FAW).

Table 2 .
Average vehicle speed, load mass and fuel consumption for the tested trucks.

Table 3 .
Summary of real-world NO x emissions for the HDDVs tested using PEMS.

Table 4 .
NO x emission factors for all tested HDDTs associated with emission standards and road types.Average vehicle speeds on urban roads, suburban roads and freeways were approximately 20 km h -1 , 50 km h -1 and 70 km h -1 , respectively.