Trends in ammonia emissions from light-duty gasoline vehicles in China, 1999–2017

Journal Pre-proof Trends in ammonia emissions from light-duty gasoline vehicles in China, 1999–2017

Shengyue Li, Jianlei Lang, Ying Zhou, Xiaoyu Liang, Dongsheng Chen, Peng Wei PII:

S0048-9697(19)34350-5

DOI:

https://doi.org/10.1016/j.scitotenv.2019.134359

Reference:

STOTEN 134359

To appear in:

Science of the Total Environment

Received date:

27 July 2019

Revised date:

29 August 2019

Accepted date:

6 September 2019

Please cite this article as: S. Li, J. Lang, Y. Zhou, et al., Trends in ammonia emissions from light-duty gasoline vehicles in China, 1999–2017, Science of the Total Environment (2019), https://doi.org/10.1016/j.scitotenv.2019.134359

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier.

Journal Pre-proof

Trends in ammonia emissions from light-duty gasoline vehicles in China, 1999–2017 Shengyue Li a, Jianlei Lang a, *, Ying Zhou a, Xiaoyu Liang a, Dongsheng Chen a, Peng Wei b a

Key Laboratory of Beijing on Regional Air Pollution Control, College of Environmental & Energy

Engineering, Beijing University of Technology, Beijing, 100124, China Chinese Research Academy of Environmental Sciences, Beijing 100012, China

*

Corresponding Author:

Jo ur

na

lP

re

-p

ro

of

b

J.L. Lang, Associate Professor, PhD

College of Environmental & Energy Engineering Key Laboratory of Beijing on Regional Air Pollution Control Beijing University of Technology Beijing 100124, China Email: [email protected]

1

Journal Pre-proof Abstract: Ammonia (NH3) emitted from motor vehicles is a by-product of measures taken to reduce emissions of other pollutants (e.g. NOx and CO) and has potentially important environmental impacts. NH3 levels can be impacted by various emission standards. However, there is a lack of investigations of the influences from the implementation of different vehicular emission standards on long-term changes in NH3 emissions. To fill this gap, we estimated the inter-annual NH3 emissions of light-duty gasoline vehicles (LDGVs) under different emission standards (State 0 to

of

State V) from 1999 to 2017 and investigated the emission change characteristics under the rapidly

ro

developing Chinese economy. Results showed that the NH3 emissions from LDGVs had a sharp, 42-fold increase (from 1.8 Gg to 77.9 Gg). However, NH3 emissions per capita have begun to

-p

decrease with increases in socioeconomic development, presenting an inverted U-shaped tendency

re

as a function of per capita GDP. Further exploration indicated that the decline in emission factors, as

lP

determined by upgrades in emission standards, was the decisive factor in promoting the downward trend in per capita emissions. This suggests that continuously upgrading emission standards has

na

offset the increase in NH3 emissions due to the rapid growth of motor vehicles. Quantitative

Jo ur

scenario analysis showed a two-stage impact of emission standards on NH3 emissions: emissions would decrease 77% (48%–90% for different years) if State I and State II were not implemented; while if none of standards were upgraded (State III to State V), NH3 emissions would increase 118% (13%–224% for different years), 2–6 times the impacts from the growth of vehicle population and the decline of vehicle kilometres traveled. The data and findings in this study can provide scientific support for understanding air pollution in urban areas and for formulating further vehicle pollution mitigation measures in China and other countries.

Keywords: vehicular NH3 emissions; change trends; emission standards; driving factors

2

Journal Pre-proof 1. Introduction Ammonia (NH3) emitted from motor vehicles has potentially important environmental impacts, especially in urban areas characterized by large numbers of motor vehicles and dense populations (Chang et al., 2019; Mac Kinnon et al., 2019; Zavala-Reyes et al., 2019). Air quality monitoring studies found that motor vehicles can significantly increase the ambient NH3 concentrations on urban roadsides and in tunnels by 3–15 times that in urban and near-city background sites (Elser et al., 2018; Perrino et al., 2002; Vieira-Filho et al., 2016; Wang et al., 2018). Source apportionment

of

studies indicated that motor vehicular can contribute 12%–25% of the atmospheric NH3 in urban

ro

areas (Chang et al., 2016; Wang et al., 2018). NH3 is also an important precursor of atmospheric PM2.5 and can be transformed into NH4+ through chemical reactions. In addition, as an alkaline gas

-p

in the atmosphere, NH3 can promote the formation of acidic secondary aerosols (e.g. SO42- and

re

NO3-), which are important chemical components of PM2.5 (Fu et al., 2017; Wang et al., 2011). As a

lP

result, NH3 can also contribute to haze formation and influence N and S deposition (Cape et al., 2004; Fenn et al., 2018). Although fertilization and livestock are the dominant sources of NH3

na

emissions, in urban areas where there is nearly no farmland, no livestock, and sanitary toilet

Jo ur

conditions, motor vehicle is an important NH3 emission contributor that cannot be ignored (Meng et al., 2017). Consequently, it is of great necessity to estimate vehicular NH3 emissions to provide basic data for understanding air pollution and formulating effective pollution mitigation measures. Previous studies observed that NH3 can be emitted from gasoline vehicles equipped with three-way catalysts (TWC) (Borsari and de Assuncao, 2017). TWC is a kind of after treatment for vehicle exhaust and is designed to reduce pollutant emissions (e.g. NOx and CO), as follows in Eq. (1) (Cadogan D F and J., 1991):

However, NH3 is a by-product of this reaction, generated in a fuel-rich/reduced-O2 conditions at the catalyst, as shown in Eq. (2–3) (Livingston et al., 2009):

3

Journal Pre-proof

NH3 emissions are associated with the removal of other pollutants. Increasingly strict emission standards have been or are being implemented worldwide to mitigate the NOx and CO emitted from on-road vehicles. Upgrades to emission standards may also influence vehicular NH3 emissions. In fact, some studies have indicated that NH3 emission factors varied under different emission standards (Durbin et al., 2002; Karlsson, 2004). For example, based on emission measurement results, Huang et al. (2018) found that the mean emission factor of NH3 under State I was ~1.6

of

times that under State IV. Most previous studies that investigated vehicular emissions focused on

ro

regular pollutants, such as CO, NOx, HC and PM (Li et al., 2019; Yang et al., 2018); however, much less research has been carried out focusing on the inter-annual vehicular NH3 emissions

-p

(Kang et al., 2016; Lang et al., 2016). Furthermore, to our knowledge, there no study has been

re

published investigating the effects of implementing vehicular emission standards on long-term

lP

changes in NH3 emissions.

Over the past decades, China’s economy has experienced rapid development. The country’s

na

Gross Domestic Product (GDP) has increased, with an annual average growth rate of 9.1%, from

Jo ur

9.1 trillion yuan in 1999 to 82.7 trillion yuan in 2017 (NBSC, 2000-2018). At the same time, the population of motor vehicles has also grown sharply. The vehicle number has increased by more than 1300%, from 14.5 million in 1999 to 209.1 million in 2017 (NBSC, 2000-2018). To reduce the CO, NOx, HC, and PM emissions from motor vehicles, China has implemented successive vehicular emission standards from State I to State V from 2000 to 2017 (MEEPRC, 2018). The number of vehicles has increased with economic development, followed by intensive implementation of emission control measures in China, but their effects on NH3 emissions remain unknown. It is also unknown what drives the emissions variations. Making these issues clear will help to understand both the NH3 emissions from on-road vehicles and the experiences of vehicle pollution control in China. According to previous studies, vehicular NH3 in China is mainly emitted from light-duty 4

Journal Pre-proof gasoline vehicles (LDGVs) (Lv et al., 2019; Zhou et al., 2019); as part of the implementation of State I (in 1999 in Beijing, in 2000 nationwide) LDGVs were required to have a TWC installed. The purpose of this study is to (1) estimate inter-annual NH3 emissions from LDGVs in China from 1999 to 2017, considering the implementation of different emission standards (State 0 to State V); (2) reveal how NH3 emissions have changed with economic development, and investigate the driving factors behind this; and (3) quantitatively analyse the effects of implementing emission standards and changes in population and vehicle kilometres traveled (VKT) on the variations in

ro

of

vehicular NH3 emissions.

2. Data and methodology

-p

From 1999 to 2017, five emission standards have been implemented for LDGVs in China,

re

including State I, State II, State III, State IV and State V (Table 1); besides, we defined the stage

lP

before the implementation of State I as State 0. In this study, a distance-based method was applied to calculate standard-specific emissions of NH3 for 31 provinces in China over the past decades.

na

========================

Jo ur

Place Table 1 Here

========================

2.1 Calculation of NH3 emissions

According to the definition and classification of light-duty vehicles in relevant official document published in China (SA, 2002), two types of LDGVs were considered in our study, including PC for carrying passengers and LDGT for carrying goods. The fuel type is gasoline. More details on each type are presented in Table S1. NH3 emissions from LDGVs were computed following Eq. (4):

where i is the target year (from 1999 to 2017) for emission estimation; j represents the 31 5

Journal Pre-proof provinces in mainland China (excluding Hong Kong, Macao, and Taiwan); m is the LDGV type, including passenger car (PC) and light-duty gasoline truck (LDGT); n is the emission standard (from State 0 to State V);

is the NH3 emissions of LDGVs in area j in year i;

population of type m vehicle with emission standard n in area j in year i;

is the is the annual

average vehicle kilometres traveled for type m vehicle in area j in year i, km;

is the NH3

emission factor for type m vehicle with emission standard n in area j in year i, g/(km·veh).

of

2.2 Vehicle population

ro

To obtain the numbers of LDGVs with different emission standards, a population estimation

-p

model (Eq. 5) put forward in our previous study was adopted (Lang et al., 2012), which has been widely cited and proved to be relatively reasonable and effective (Mohammadiha et al., 2018; Yang

Jo ur

na

lP

re

et al., 2019).

where y is the implementation year of emission standard n; registrations of vehicle m with emission standard n in area j in year i; and

is the number of new is the total

number of vehicle m in area j in year i. If the number of vehicles with State n (e.g. State I) is negative, it should be set to 0, and the number of vehicles with State (n + 1) (e.g. State II) should be modified accordingly. For details can see Lang et al. (2012). The total annual vehicle population and the numbers registered in different provinces were collected from official statistical yearbooks (NBSC, 2000-2018). There was a sharp growth in the vehicle population over the past decades (Fig. S1a). The total number of LDGVs increased 24-fold from 1999 to 2017 (from 7.2 million to 180.6 million), with the average annual growth rate of 19.7%. With the implementation of emission standards, LDGVs was dominated by State 0 vehicles 6

Journal Pre-proof from 1999 to 2002, State I vehicles from 2003 to 2006, State II vehicles from 2007 to 2008, State III vehicles from 2009 to 2011, and State IV vehicles from 2012 to 2017.

2.3 Annual average VKT The annual average VKT for different vehicle types in each province were estimated following

is the total freight/passenger traffic volume for vehicle m in area i in year j,

type m vehicle in area i, %;

is the corresponding volume share, %;

ro

t-km or passenger-km;

is the actual load rate of

is the average load capacity for type m vehicle in in area i,

-p

where

of

Eq. (6) (He et al., 2005; Liu et al., 2008):

re

t/seats.

lP

The total freight/passenger traffic volumes were obtained from the official statistical yearbooks (NBSC, 2000-2018), and the other data were collected from relevant studies (Li et al., 1995; Liu et

na

al., 2007). As the vehicle population increased, the annual average VKT of PCs decreased

Jo ur

continuously from 1999 to 2017, with a total reduction of 47% (from 37,174 km to 19,686 km); however, the overall trend for LDGTs increased and the annual average VKT more than doubled, from 20,829 km in 1999 to 45,594 km in 2017 (Fig. S1b). This opposite variation tendency is probably due to the growth in the number of private cars and the development of the transport industry (Lv et al., 2019).

2.4 Standard-specific emission factors COPERT model is one of the most common methods used to model the EFs of pollutants from motor vehicles meeting European standards (Quaassdorff et al., 2016; Sun et al., 2019). Since emission standards in China are similar with that in Europe, we applied the COPERT V model to compute the annual average NH3 EFs for each vehicle type with different standards in this study. 7

Journal Pre-proof The required fuel quality information (e.g. sulphur content and Reid vapour pressure) were obtained from Chinese national and local fuel standards (SA, 2016). The inter-annual sulphur contents were listed in Table S2 for detail. Monthly meteorological data for each province in each year, including relative humidity and maximum and minimum temperatures, were obtained from the National Meteorological Information Centre (http://data.cma.cn/). From Fig. 1 we can see that there is a large discrepancy in COPERT-EFs among emission standards, which initially increased and then decreased. EFs peaked at 101.9 mg/km in State II, 6–8

of

times the value in State III to V. NH3 emissions were low from State 0 vehicles, since the TWC

ro

equipment was non-mandatory for LDGVs at this stage. Compared to the results from previous studies (Heeb et al., 2006a; Heeb et al., 2008; Heeb et al., 2006b; Huang et al., 2018; Liu et al.,

-p

2014; Suarez-Bertoa and Astorga, 2016; Suarez-Bertoa et al., 2014; Tang et al., 2016), the

re

downtrend of COPERT EFs with the upgrade of emission standard is found generally consistent

lP

with that of the average measured EFs, especially after State II. A further comparison under each emission standard suggests that the standard-specific EFs calculated by COPERT are acceptable,

na

which basically falls within the range of measured EFs under each emission standard. Taking EFs

Jo ur

under State V for example, COPERT EF (12.75 mg/kg) is in the range of local EFs (2.29–73.04 mg/km) measured by Huang et al. (2018), and comparable to the median level of 18.19 mg/kg, though it seems smaller than the average level (33.81 mg/kg) in that study. All discussion above indicates that NH3 EFs modelled by COPERT can be used further to estimate the inter-annual NH3 emissions. ======================== Place Figure 1 Here ========================

3. Results and discussion 3.1 Trends and characteristics of NH3 emissions from LDGVs 8

Journal Pre-proof Fig. 2 shows the inter-annual trends of NH3 emissions from LDGVs from 1999 to 2017 in China. A significant growth can be seen in the NH3 emissions from LDGVs over the past decades, a 42-fold increase from 1.8 Gg in 1999 to 77.9 Gg in 2017. Although emissions had a significant increase, it actually peaked in 2011 at 81.1 Gg and has presented a slightly decreased tendency in recent years. The annual variation rate in emissions was 26.8%, with an average increase rate of 40.6% during 2000–2011 and average decrease rate of 0.7% during 2012–2017. The significant increase of NH3 emissions is likely due to the sharp increase in the number of LDGVs, and the

of

slight fall after 2011 may be from the synergy effect from the slower growth in vehicle population

ro

(Fig. S1a) and the lower EFs for LDGVs with State III and after standards (Fig. 1). In terms of the vehicle type, NH3 emissions from LDGVs were entirely dominated by PC,

-p

which explained more than 90% of the total emissions. For both vehicle types, NH3 emissions

re

tended to increase first and then decrease. However, the downtrend after 2011 was more significant

between –1.5% and 3.5%.

lP

for LDGTs, with an average variation rate of –8.4%, while the emissions from PC still fluctuated

na

There is an interesting phenomenon shown in Fig. 3, in which NH3 emission from LDGVs

Jo ur

were mainly from State II vehicles for about 10 of the past 19 years, though this standard has only been in place for 3 years. The major contributors to total NH3 emissions from 1999 to 2005 were State I vehicles, with an average proportion of 85.2%; State II vehicles then became the major contributors one year after the implementation of this standard, contributing 55.8% on average annually from 2006 to 2015. Until 2016, eight years after the implementation of newer standards, the dominant contributors became State IV vehicles (34.8% on average from 2016 to 2017), but the emissions from State II vehicles were still considerable. The standard-specific distribution of vehicle population (Fig. S1a) showed that LDGVs were actually dominated by State II vehicles for only two years (from 2007 to 2008). At the same time, there was no significant variation in VKT over the past decades (Fig. S1b); we thus have reasons to conclude that the large proportion of emissions from State II vehicles over a longer period of time was very likely due to the higher EFs 9

Journal Pre-proof under this standard. This finding indicated that standard-specific EFs influenced the variation in NH3 emissions, and that further research should consider the varying EFs. For the NH3 emissions from different provinces, there is a similar distribution of regional contributions in each year (Fig. S2). The largest contributors were, in order, Beijing during the State I period (1999–2004), Guangdong from State II to the late State IV period (2005–2015), and Shandong after the State IV period ended (2016–2017). Spatial distribution of annual average emissions of each province suggests that the central and eastern regions in China emitted more NH3

of

from LDGVs (Fig. S3a). Among the 31 provinces, Guangdong and Xizang had the largest and

ro

smallest emissions with 5.7 Gg and 0.1 Gg, respectively. Analysis on emission strength shows a more obvious graded distribution descending from east to west (Fig. S3b). Xizang still ranked last,

-p

but the top regions turned to be Shanghai and Beijing, with the average emission strength of 159.8

re

kg/km2 and 156.7 kg/km2, respectively; they were more than two times that of the third largest

lP

emission strength (in Tianjin, 68.8 kg/km2). This phenomenon infers a close relationship between vehicular NH3 emission and the level of socioeconomic development. A summative analysis (pie

na

charts in Fig. 2) of the proportions of emissions by province, at each standard period, found that

Jo ur

Beijing, Guangdong, Shandong, Zhejiang, Jiangsu, Hebei, Henan, and Sichuan are the major provinces that had the largest emissions, accounting for 54.6%–63.1% of total NH3 emissions of LDGVs in China. Even more remarkably, the proportion from some relatively developed areas (e.g. Beijing, Shanghai, and Tianjin) gradually decreased; while the contribution from some economically developing provinces (e.g. Shandong, Inner Mongolia, and Hubei) tended towards growth. This may have been caused by the differences in level of economic development among the regions. Some developed areas began to control motor vehicles to reduce the environmental impacts from traffic emissions, while the number of vehicles in other developing areas were still increasing as the local economy developed. This inference can be made indirectly based on the inter-annual NH3 emissions from LDGVs in each province (Fig. S4). ======================== 10

Journal Pre-proof Place Figure 2 and Figure 3 Here ========================

3.2 NH3 emissions trends under economic development According to our discussions above and previous studies (Cai et al., 2018; Liu et al., 2017), there are some relationships between motor vehicle development and socioeconomic development. Since the NH3 emissions from LDGVs mostly come from PCs, sections 3.2 and 3.3 are thus focused

of

on the variations in NH3 emissions from PCs. To reflect the level of socioeconomic development

ro

more comprehensively, two indicators were chosen here, including Gross Domestic Product (GDP) in national perspective and Gross Regional Product (GRP) in provincial perspective. Besides, the

-p

average emission, GDP and GRP weighted by population (average emission, average GDP, and

re

average GRP for short) were calculated to eliminate the impacts caused by different levels of

lP

development. Fig. 4a presents the national average emission at the corresponding average GDP over the past decades. Similarly, Fig. 4b presents the provincial values in all 31 provinces for each year

na

studied in our research. The relationship between average emission and average GDP (or GRP)

Jo ur

represented a statistically significant inverted U shape over time—from both the national (Fig. 4a) and provincial (Fig. 4b) perspectives— with an inflection point around 2013. In other words, the average NH3 emissions began to decrease with the socioeconomic development in recent years, though it increased considerably during the earlier years. To determine why the relationship had an inverted U shape, we calculated the average vehicle population and average EF for each province, respectively. The average vehicle population was computed by dividing the number of total PC by that of people. The average EF was computed by dividing per capita NH3 emission by average vehicle population times VKT of PC, to comprehensively consider the synergy effect of population and emission standard implementation. In theory, average population, VKT, and EF all have the potential to create this shape. However, average population is actually still rising with average GRP at the present stage (Fig. 5a), and it is 11

Journal Pre-proof scarcely possible to lead to the decrease in average emission. At the same time, there has been a clear decline in both average VKT (Fig. 5b) and average EF (Fig. 5c). The difference between them is that average VKT decreased more sharply during 2003–2009, and then tended to be gentle, while the overall trend in average EF began to be downward only after 2010, which is closer to the occurrence time of the inflection point of average emission. Two scenario analyses were then carried out by estimating NH3 emissions with a fixed VKT and a fixed EF, respectively. Taking the fixed VKT scenario as an example, the emission was

of

calculated depending on actual inner-annual vehicle population and EF, and a fixed VKT in 1999;

ro

this scenario was designed to quantify the impacts of EF on the changing trend of emission. In Fig. 5d the inverted U-shape still appeared during the fixed VKT scenario, but the appearance time of

-p

the inflection point was delayed for about four years and the average emission increased; in contrast,

re

there is little downside to the fixed EF scenario. This finding suggests that the inverted U-shaped

lP

trend in average emission was determined by the EFs, while the downward-trending VKT promotes the appearance of an inflection point and lowers the total emissions level. The variation in VKT was

na

driven by the natural development of society, while the changes in EF were caused by upgrades in

Jo ur

the emission standards, which is an additional kind of human intervention. Therefore, we indicate that human action is the decisive factor accounting for the decline in NH3 emissions from LDGVs as socioeconomic development occurs in China. ======================== Place Figure 4 and Figure 5 Here ========================

3.3 Impacts of upgrades in emission standards on the variation in NH3 emissions The variations in NH3 EFs are influenced by upgrades to the emission standard update. Comparing the various COPERT EFs in Fig. 1, we can roughly separate the influence into two stages. Stage I, which is the period when State I and State II were implemented, has a higher 12

Journal Pre-proof emissions intensity; Stage II involves the later period when State III and beyond were implemented. Since NH3 emissions from LDGVs are related to the reduction in other air pollutants, especially NOx, we further investigated the relationship between the variations in NOx EFs and NH3 EFs at each emission standard, both of which were determined by the COPERT V model (Fig. S5). The two stages also clearly identified that NH3 emissions increased with the reduction of NOx in Stage I, and decreased in Stage II. Therefore, it is suggested that Stage I represents the period when emission standards have just come into force, and the control technologies were imperfect; Stage II

of

reflects an improvement of these technologies, no matter for the lower emission levels or the control

ro

effects for NH3.

Investigating the inter-annual variation rates of population, VKT, EF, and emission (Fig. 6), we

-p

found that, different from the steady and one-way change of population and VKT, the average EF

re

initially increased and then decreased, peaking at 2008, when State III for LDGVs began to be

lP

implemented nationwide (Fig. 2). This variation well reflects the two stages of emission standard implementation in China. Therefore, we divided our research period into two parts to further

na

quantify the impact of upgrades in emission standards on the variations in NH3 emission. Stage I is

Jo ur

from 2000 to 2007, and Stage II is from 2008 to 2017; 1999 and 2007 were designed as the base year for Stage I and Stage II, respectively, to evaluate the changes in estimated emissions in each scenario.

Fig. 7 shows the inter-annual NH3 emissions estimated in three hypothetical scenarios (Fixed population, Fixed VKT, and Fixed EF) and the actual scenario. For example, the inter-annual emissions in the Fixed population scenario was always estimated based on the population in base year and other variables (VKT and EF) from year to year; compared to the emissions in the actual scenario, the Fixed population scenario was designed to quantify the impacts of population on total emissions and its variations. It can be seen that the effects from variables to emissions is in the order of EF, population, and VKT. On one hand, the effect from each variable increased as time went on in each stage; on the other hand, the average effect rose from Stage I to Stage II. The 13

Journal Pre-proof effects from VKT was smallest; and if it had no variation, the total emissions would increase of 18% (4%–32%) in Stage I and 21% (3%–40%) in Stage II. If there was no growth in population, however, the emissions would decrease by an average of 50% and 62% in Stage I and Stage II, respectively. At the same time, the two-way effect from EF caused by its two-way variation analysed above was clearer here. In Stage I, no implementation of State I and State II with higher EF would reduce emissions by 28%–90% (77% on average); while in Stage II, without the implementation of newer standards with improved control technology, the emissions would rise by

of

13%–224% (118% on average), about 2–6 times that of the effects from the other two variables on

ro

average. The quantitative analysis in this section verified that the variation in EF was the major factor influencing the changing tendency of emissions at all stage, and indicated that the updates to

-p

the emission standard can effectively promote the decrease in NH3 emissions from LDGVs in

re

China.

lP

======================== Place Figure 6 and Figure 7 Here

3.4 Discussion

Jo ur

na

========================

3.4.1 Comparison with other studies

Compared to most of previous studies, this work estimated inter-annual NH3 emissions from motor vehicles in consideration of the implementation of emission standards in China. To clear and verify the significance of this improvement, we compared our results with those in other studies (Huang et al., 2018; Huang et al., 2012; Kang et al., 2016; Sun et al., 2017; Tang et al., 2016; Zhang et al., 2017) in Fig. S6. Based on the changing trend of EFs with the upgrade of emission standards (Fig. 1), NH3 emissions are supposed to have a rapid increase at the beginning, but the growth rate should then gradually slow down, especially after the implementation of State III. Fig. S6 does show an obvious rise of NH3 emissions in earlier years, no matter for those considering the 14

Journal Pre-proof implementation of emission standard or not. However, after 2008 when State III began to be implemented, some differences appear. The growth rates of the emissions taking into account standards have slowed significantly as analysed above; while the emissions not considering standards still have a rapid increase. This finding verifies the rationality of the improvement considered during the estimation process in this study; and it also again emphasizes the importance on taking emission standards into account when estimating pollutant emissions from traffic source. 3.4.2 Uncertainty and limitation

of

There are uncertainties in the estimation of NH3 emissions from LDGVs because the

ro

calculation methods used simplify the actual emission process (Dey et al., 2019). For EFs, restricted by the calculation principle of COPERT V (https://www.emisia.com/utilities/copert/documentation/),

-p

NH3 emissions estimated in this study are almost from the LDGVs equipped with TWC. There are

re

also other catalysts used for reducing pollutant emissions from vehicles apart from TWC, such as

lP

oxidation catalyst, dual-bed catalysts; but they account for only a small fraction (Zhou, 2003) and have much smaller NH3 EFs (EFIG, 2004; Heeb et al., 2006a; Huang et al., 2018; Karlsson, 2004).

na

For vehicle population, the vehicles considered in this study are all those with local brand registered

Jo ur

in local departments. However, there are still considerable effects from vehicles with non-local brand in some regions (e.g. Beijing), which are very lack of official statistics. In this study, we applied a Monte Carlo simulation, which is based on the mathematical distribution and coefficients of variation (CV) of independent variables, to evaluate these uncertainties (Caserini et al., 2013). There are three variables to estimate emissions: vehicle population, VKT, and EF. Vehicle populations were obtained from official statistical yearbooks, in which the number of vehicles registered were summarized directly without detailed investigation; hence, the uncertainty of the population was small and assumed to have a normal distribution with a CV of 5%. The VKT values were computed by the formula based on fundamental data collected from previous studies; as a result, it had a higher uncertainty and a normal distribution with a CV of 30% (Lang et al., 2014). The NH3 EFs were simulated by COPERT, which was generally 15

Journal Pre-proof recognized to be lognormal distributed with a CV of 30%. To ensure the accuracy of the results, the number of trials was set to 100,000. The simulation results showed that the uncertainty range under a confidence coefficient of 95% in 2017 was (–52%, 73%). Since LDGV emissions currently contribute to the highest proportion of total NH3 emissions from motor vehicles in China, we only considered the emissions from this type of vehicle in our study. However, other vehicles can also emit NH3. For example, diesel vehicles adopting selective catalytic reduction (SCR) technology to reduce emissions are another potential source of NH3.

of

Some advanced measurements showed that NH3 EFs of heavy-duty diesel vehicles (HDDVs) with

ro

SCR technology were also considerable (Suarez-Bertoa et al., 2016; Suarez-Bertoa et al., 2017). SCR has actually become one of the most important after-treatment technologies since the

-p

implementation of State V for HDDVs in China in 2017 (MEEPRC, 2018). Therefore, more

re

comprehensive research that considers more types of vehicles and the latest policies will be

lP

necessary in the future.

To obtain standard-specific EFs, we applied COPERT to compute NH3 EFs in this study.

na

Although China has similar emission standard stages with Europe, there are some differences

Jo ur

between these two areas, such as emission limits of pollutants, measurement methods and technical maturity of control measures. Hence, localized EFs from measurements are strongly suggested to establish an NH3 emission inventory of motor vehicles in China. However, there are still some deficiencies in the current research on this area. First, previous measurements are mostly carried out by chassis dynamometers within the laboratory, which may be not sufficient to reflect the real emission conditions on the road; second, the existing results are still not adequate and robust enough to obtain the EFs of vehicles with all emission standards; finally, most published EFs are currently from foreign measurement results. Therefore, more systematic and comprehensive measurements are needed on local NH3 EFs from motor vehicles in China for a better and more accurate understanding of its emission status and the development of further researches. 3.4.3 Implications 16

Journal Pre-proof The emission inventory covering a relatively period long time established in our study can provide a systematic understanding of emission levels and the variations in NH3 emissions from motor vehicles in China; it also supports some fundamental data for further research on its impacts on air quality or human health. In addition, according to our findings on the two-stage influence from emission standard, this study suggests that new standards should pay attention to the collaborative reduction of NOx (or CO) and NH3 emissions for the continuous declination of vehicular NH3 emissions, though there is no restriction on NH3 yet.

of

Motor vehicles have almost become the most important air pollutant source worldwide, and it

ro

will likely become more severe as the economy develops (Meng et al., 2017; Robertson, 2019). As the world's second largest economy, China has maintained a very high growth rate of GDP over the

-p

past decades, reaching 9.1% from 1999 to 2017, larger than most of countries in the world (NBSC,

re

2000-2018). Meanwhile, it has almost the fastest-growing vehicle population, more than four times

lP

the global average (http://www.oica.net/category/vehicles-in-use/). In addition, the Chinese government has experience in vehicular emission control—five vehicular emission standards (State

na

I to State V) have been successively implemented from 2000 to 2017. Therefore, China’s

Jo ur

experiences in controlling vehicle emissions may be a valuable reference for other countries around the world. Since the considerable emissions and impacts of regular pollutants (e.g. NOx and CO) emitted from motor vehicles, finding ways to reduce emissions is a necessity; actually, emission standards at different stages have been successively in place in many parts of the world (Cai et al., 2017). Hence, NH3 emissions from vehicle sources are ineluctable, and it has an important significance to synchronously minimize NH3 emission when other pollutants are controlled. In this study, we found that the variation in EF is the key factor influencing the change in NH3 emissions when population and VKT vary naturally with economic development. The level of EF is directly determined by the stage of the emission standard, which is totally led by anthropogenic actions. In other words, anthropogenic control measures can counteract the influences on pollutant emissions caused by socioeconomic development. This finding may provide some experiences or references 17

Journal Pre-proof for China and other countries for policy making and implementation in the future.

4. Conclusions This study used the COPERT model to estimate the inter-annual NH3 emissions from LDGVs in China from 1999 to 2017, and explored the impacts of the implementation of emission standards on the various emission levels over the past decades. LDGV NH3 emissions in China grew more than 42-fold for the past 19 years, increasing from 1.8 Gg in 1999 to 77.9 Gg in 2017, although a

of

slight decrease (~0.7%) has occurred in recent years. Standard-specific EFs had substantial

ro

influence on these changing emissions. Additionally, the proportion of emissions from developed areas trended downward, while that from developing provinces trended upward over time.

-p

An economy-based analysis showed that per capita NH3 emissions presented an inverted

re

U-shaped variation as a function of per capita GDP, which indicated that the emissions have begun

lP

to decrease with the development of the economy. A scenario estimation suggested that the decline in EF was the decisive factor contributing to the falling of per capita emission, even though the

Jo ur

and lowered the emission level.

na

vehicle population was still rapidly rising; meanwhile, the decline of VKT accelerated this process

EF is directly determined by the stage of emission standards. The variation rate of EF showed a two-stage impacts from standards on NH3 emissions in China. Quantitative scenario analysis found that vehicle population and VKT always affected the changes in emissions in one direction during the updated period of emission standards. However, in Stage I, no implementation of emission standards (State I and State II) would reduce NH3 emissions by 77%; in Stage II, if standard upgrades (State III to State V) had not been implemented, the emissions would have increased by 118%, 2–6 times the impacts from population and VKT on average. The variation in vehicle population and VKT is naturally driven by socioeconomic development, while the variation in EF resulted exclusively from the human policy intervention. This finding indicates that anthropogenic control measures can offset the increase in pollutant emissions caused by natural 18

Journal Pre-proof socioeconomic development. The emission inventory and changing trends investigated in our study can provide a systematic understanding of NH3 emissions from motor vehicles in China. The long-term NH3 emissions can provide basic data for further air pollution studies, especially in urban areas. In addition, the experiences learned from the changing trends and their driving factors can also offer references for effective vehicular NH3 emission control in China and other countries.

of

Acknowledgement

ro

This research was supported by the National Natural Science Foundation of China (No. 51878012 & 91644110), the National Key R&D Program of China (2017YFC0212202) and the

-p

Key Projects on Heavy Air Pollution Control of China (DQGG0303). In addition, we greatly

re

appreciate the Beijing Municipal Commission of Education and the Beijing Municipal Commission

lP

of Science and Technology for supporting this work. The authors are grateful to the anonymous

Jo ur

na

reviewers for their insightful comments.

19

Journal Pre-proof References Borsari, V., de Assuncao, J.V., 2017. Ammonia emissions from a light-duty vehicle. Transportation Research Part D-Transport and Environment 51, 53-61. Cadogan D F, J., H.C., 1991. Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed. John Wiley & Sons, Inc., New York. Cai, B., Wang, X., Huang, G., Wang, J., Cao, D., Baetz, B.W., Liu, L., Zhang, H., Fenech, A., Liu, Z., 2018. Spatiotemporal Changes of China's Carbon Emissions. Geophysical Research Letters 45, 8536-8546. Cai, J.S., Hu, M.W., Huang, W.K., 2017. A study on the emission standards of light-duty vehicles in

of

China, European Union and USA. Environmental Pollution & Control 39, 1268-1273. Cape, J.N., Tang, Y.S., van Dijk, N., Love, L., Sutton, M.A., Palmer, S.C., 2004. Concentrations of

ro

ammonia and nitrogen dioxide at roadside verges, and their contribution to nitrogen deposition. Environment Pollution 132, 469-478.

-p

Caserini, S., Galante, S., Ozgen, S., Cucco, S., de Gregorio, K., Moretti, M., 2013. A methodology

re

for elemental and organic carbon emission inventory and results for Lombardy region, Italy. Science of the Total Environment 450, 22-30.

lP

Chang, Y., Zou, Z., Deng, C., Huang, K., Collett, J.L., Lin, J., Zhuang, G., 2016. The importance of vehicle emissions as a source of atmospheric ammonia in the megacity of Shanghai.

na

Atmospheric Chemistry and Physics 16, 3577-3594. Chang, Y., Zou, Z., Zhang, Y., Deng, C., Hu, J., Shi, Z., Dore, A.J., Collett, J.L., Jr., 2019. Assessing

Jo ur

contributions of agricultural and nonagricultural emissions to atmospheric ammonia in a Chinese megacity. Environmental Science & Technology 53, 1822-1833. Dey, S., Caulfield, B., Ghosh, B., 2019. Modelling uncertainty of vehicular emissions inventory: A case study of Ireland. Journal of Cleaner Production 213, 1115-1126. Durbin, T.D., Wilson, R.D., Norbeck, J.M., Miller, J.W., Huai, T., Rhee, S.H., 2002. Estimates of the emission rates of ammonia from light-duty vehicles using standard chassis dynamometer test cycles. Atmospheric Environment 36, 1475-1482. Emission Factor and Inventory Group of the Envirionmental Protection Agency of US (EFIG), 2004. Estimating ammonia emissions from anthropogenic nonagricultural sources-draft final report. 46-47. Elser, M., El-Haddad, I., Maasikmets, M., Bozzetti, C., Wolf, R., Ciarelli, G., Slowik, J.G., Richter, R., Teinemaa, E., Huglin, C., Baltensperger, U., Prevot, A.S.H., 2018. High contributions of vehicular emissions to ammonia in three European cities derived from mobile measurements. Atmospheric Environment 175, 210-220. Fenn, M.E., Bytnerowicz, A., Schilling, S.L., Vallano, D.M., Zavaleta, E.S., Weiss, S.B., Morozumi, 20

Journal Pre-proof C., Geiser, L.H., Hanks, K., 2018. On-road emissions of ammonia: An underappreciated source of atmospheric nitrogen deposition. Science of the Total Environment 625, 909-919. Fu, X., Wang, S., Xing, O., Zhang, X., Wang, T., Hao, J., 2017. Increasing ammonia concentrations reduce the effectiveness of particle pollution control achieved via SO2 and NOx emissions reduction in East China. Environmental Science & Technology Letters 4, 221-227. He, K., Huo, H., Zhang, Q., He, D., An, F., Wang, M., Walsh, M.P., 2005. Oil consumption and CO2 emissions in China's road transport: current status, future trends, and policy implications. Energy Policy 33, 1499-1507. Heeb, N.V., Forss, A.-M., Brühlmann, S., Lüscher, R., Saxer, C.J., Hug, P., 2006a. Three-way catalyst-induced formation of ammonia—velocity- and acceleration-dependent emission

of

factors. Atmospheric Environment 40, 5986-5997.

NH3-emissions

from

gasoline-fueled

Euro-3-to

Euro-4-passenger

cars.

Atmospheric

-p

Environment 42, 2543-2554.

ro

Heeb, N.V., Saxer, C.J., Forss, A.-M., Bruehlmann, S., 2008. Trends of NO-, NO2-, and

Heeb, N.V., Saxer, C.J., Forss, A.M., Bruhlmann, S., 2006b. Correlation of hydrogen, ammonia and

re

nitrogen monoxide (nitric oxide) emissions of gasoline-fueled Euro-3 passenger cars at

lP

transient driving. Atmospheric Environment 40, 3750-3763. Huang, C., Hu, Q., Lou, S., Tian, J., Wang, R., Xu, C., An, J., Ren, H., Ma, D., Quan, Y., Zhang, Y.,

na

Li, L., 2018. Ammonia emission measurements for light-duty gasoline vehicles in China and implications for emission modeling. Environmental Science & Technology 52, 11223-11231. Huang, X., Song, Y., Li, M., Li, J., Huo, Q., Cai, X., Zhu, T., Hu, M., Zhang, H., 2012. A

1-14.

Jo ur

high-resolution ammonia emission inventory in China. Global Biogeochemical Cycles 26,

Kang, Y., Liu, M., Song, Y., Huang, X., Yao, H., Cai, X., Zhang, H., Kang, L., Liu, X., Yan, X., He, H., Zhang, Q., Shao, M., Zhu, T., 2016. High-resolution ammonia emissions inventories in China from 1980 to 2012. Atmospheric Chemistry and Physics 16, 2043-2058. Karlsson, H.L., 2004. Ammonia, nitrous oxide and hydrogen cyanide emissions from five passenger vehicles. Science of the Total Environment 334-335, 125-132. Lang, J., Cheng, S., Wei, W., Zhou, Y., Wei, X., Chen, D., 2012. A study on the trends of vehicular emissions in the Beijing-Tianjin-Hebei (BTH) region, China. Atmospheric Environment 62, 605-614. Lang, J., Cheng, S., Zhou, Y., Zhang, Y., Wang, G., 2014. Air pollutant emissions from on-road vehicles in China, 1999-2011. Science of the Total Environment 496, 1-10. Lang, J., Zhou, Y., Cheng, S., Zhang, Y., Dong, M., Li, S., Wang, G., Zhang, Y., 2016. Unregulated pollutant emissions from on-road vehicles in China, 1999-2014. Science of the Total 21

Journal Pre-proof Environment 573, 974-984. Li, X., Lopes, D., Mok, K.M., Miranda, A.I., Yuen, K.V., Hoi, K.I., 2019. Development of a road traffic emission inventory with high spatial-temporal resolution in the world's most densely populated region-Macau. Environmental Monitoring and Assessment 191. Li, X., Yan, Q., Zhao, X., 1995. China’s Urban Transport Development Strategy. China Construction Industry Press, Beijing. Liu, H., He, K., He, D., Fu, L., Zhou, Y., Walsh, M., Blumberg, K., 2008. Analysis of the impacts of fuel sulfur on vehicle emissions in China. Fuel 87, 3147-3154. Liu, H., He, K., Wang, Q., Huo, H., Lents, J., Davis, N., Nikkila, N., Changhong, C., Osses, M., He, C., 2007. Comparison of vehicle activity and emission inventory between Beijing and

of

Shanghai. Journal of the Air and Waste Management Association 57, 1172-1177.

ro

Liu, T., Wang, X., Wang, B., Ding, X., Deng, W., Lü, S., Zhang, Y., 2014. Emission factor of

Environmental Research Letters 9.

-p

ammonia (NH3) from on-road vehicles in China: tunnel tests in urban Guangzhou. Liu, Y., Liao, W., Li, L., Huang, Y., Xu, W., 2017. Vehicle emission trends in China's Guangdong

re

Province from 1994 to 2014. Science of the Total Environment 586, 512-521.

lP

Livingston, C., Rieger, P., Winer, A., 2009. Ammonia emissions from a representative in-use fleet of light and medium-duty vehicles in the California South Coast Air Basin. Atmospheric

na

Environment 43, 3326-3333.

Lv, W., Hu, Y., Li, E., Liu, H., Pan, H., Ji, S., Hayat, T., Alsaedi, A., Ahmad, B., 2019. Evaluation of

814-825.

Jo ur

vehicle emission in Yunnan province from 2003 to 2015. Journal of Cleaner Production 207,

Mac Kinnon, M., Zhu, S., Carreras-Sospedra, M., Soukup, J.V., Dabdub, D., Samuelsen, G.S., Brouwer, J., 2019. Considering future regional air quality impacts of the transportation sector. Energy Policy 124, 63-80.

Ministry of Ecology and Environment of the People's Republic China (MEEPRC), 2018. China vehicle environmental management annual report, Beijing. Meng, W., Zhong, Q., Yun, X., Zhu, X., Huang, T., Shen, H., Chen, Y., Chen, H., Zhou, F., Liu, J., Wang, X., Zeng, E.Y., Tao, S., 2017. Improvement of a global high-resolution ammonia emission inventory for combustion and industrial sources with new data from the residential and transportation sectors. Environmental Science & Technology 51, 2821-2829. Mohammadiha, A., Malakooti, H., Esfahanian, V., 2018. Development of reduction scenarios for criteria air pollutants emission in Tehran Traffic Sector, Iran. Science of the Total Environment 622, 17-28. National Bureau of Statistics of China (NBSC), 2000-2018. National Statistical Yearbook. China 22

Journal Pre-proof Statistics Press, Beijing. Perrino, C., Catrambone, M., Di Bucchianico, A.D.M., Allegrini, I., 2002. Gaseous ammonia in the urban area of Rome, Italy and its relationship with traffic emissions. Atmospheric Environment 36, 5385-5394. Quaassdorff, C., Borge, R., Perez, J., Lumbreras, J., de la Paz, D., Manuel de Andres, J., 2016. Microscale traffic simulation and emission estimation in a heavily trafficked roundabout in Madrid (Spain). Science of the Total Environment 566, 416-427. Robertson, L.S., 2019. Motor vehicle CO2 emissions in the United States: potential behavioral feedback and global warming. Weather Climate and Society 11, 623-628. Standardization Administration (SA), 2002. Classification of power-driven vehicles and trailers

of

(GB/T 15089-2001). China Quality and Standards Publishing & Media Co., Ltd., Beijing.

ro

Standardization Administration (SA), 2016. Gaosolin for motor vehicles (GB 17930-2016), Beijing. Suarez-Bertoa, R., Astorga, C., 2016. Isocyanic acid and ammonia in vehicle emissions.

-p

Transportation Research Part D: Transport and Environment 49, 259-270. Suarez-Bertoa, R., Mendoza-Villafuerte, P., Bonnel, P., Lilova, V., Hill, L., Perujo, A., Astorga, C.,

re

2016. On-road measurement of NH3 and N2O emissions from a Euro V heavy-duty vehicle.

lP

Atmospheric Environment 139, 167-175.

Suarez-Bertoa, R., Mendoza-Villafuerte, P., Riccobono, F., Vojtisek, M., Pechout, M., Perujo, A.,

na

Astorga, C., 2017. On-road measurement of NH3 emissions from gasoline and diesel passenger cars during real world driving conditions. Atmospheric Environment 166, 488-497. Suarez-Bertoa, R., Zardini, A.A., Astorga, C., 2014. Ammonia exhaust emissions from spark

Jo ur

ignition vehicles over the New European Driving Cycle. Atmospheric Environment 97, 43-53. Sun, K., Tao, L., Miller, D.J., Pan, D., Golston, L.M., Zondlo, M.A., Griffin, R.J., Wallace, H.W., Leong, Y.J., Yang, M.M., Zhang, Y., Mauzerall, D.L., Zhu, T., 2017. Vehicle Emissions as an Important Urban Ammonia Source in the United States and China. Environmental Science & Technology 51, 2472-2481. Sun, S., Zhao, G., Wang, T., Jin, J., Wang, P., Lin, Y., Li, H., Ying, Q., Mao, H., 2019. Past and future trends of vehicle emissions in Tianjin, China, from 2000 to 2030. Atmospheric Environment 209, 182-191. Tang, G., Chao, N., Wang, Y., Chen, J., 2016. Vehicular emissions in China in 2006 and 2010. Journal of Environmental Sciences-China 48, 179-192. Vieira-Filho, M.S., Ito, D.T., Pedrotti, J.J., Coelho, L.H., Fornaro, A., 2016. Gas-phase ammonia and water-soluble ions in particulate matter analysis in an urban vehicular tunnel. Environmental Science and Pollution Research International 23, 19876-19886. Wang, R., Ye, X., Liu, Y., Li, H., Yang, X., Chen, J., Gao, W., Yin, Z., 2018. Characteristics of 23

Journal Pre-proof atmospheric ammonia and its relationship with vehicle emissions in a megacity in China. Atmospheric Environment 182, 97-104. Wang, S., Xing, J., Jang, C., Zhu, Y., Fu, J.S., Hao, J., 2011. Impact assessment of ammonia emissions on inorganic aerosols in east China using response surface modeling technique. Environmental Science & Technology 45, 9293-9300. Yang, D., Zhang, S., Niu, T., Wang, Y., Xu, H., Zhang, K.M., Wu, Y., 2019. High-resolution mapping of vehicle emissions of atmospheric pollutants based on large-scale, real-world traffic datasets. Atmospheric Chemistry and Physics 19, 8831-8843. Yang, W., Yu, C., Yuan, W., Wu, X., Zhang, W., Wang, X., 2018. High-resolution vehicle emission inventory and emission control policy scenario analysis, a case in the Beijing-Tianjin-Hebei

of

(BTH) region, China. Journal of Cleaner Production 203, 530-539.

ro

Zavala-Reyes, J.C., Jeanjean, A.P.R., Leigh, R.J., Hernández-Paniagua, I.Y., Rosas-Pérez, I., Jazcilevich, A., 2019. Studying human exposure to vehicular emissions using computational

-p

fluid dynamics and an urban mobility simulator: The effect of sidewalk residence time, vehicular technologies and a traffic-calming device. Science of the Total Environment 687,

re

720-731.

lP

Zhang, X., Wu, Y., Liu, X., Reis, S., Jin, J., Dragosits, U., Van Damme, M., Clarisse, L., Whitburn, S., Coheur, P.-F., Gu, B., 2017. Ammonia emissions may be substantially underestimated in

na

China. Environmental Science & Technology 51, 12089-12096. Zhou, L., 2003. The study on three-way catalytic converter and its matching with engine. Master Thesis. Hunan University, Hunan China.

Jo ur

Zhou, Z., Tan, Q., Liu, H., Deng, Y., Wu, K., Liu, C., Zhou, X., 2019. Emission characteristics and high-resolution spatial and temporal distribution of pollutants from motor vehicles in Chengdu, China. Atmospheric Pollution Research 10, 749-758.

24

Journal Pre-proof Table Caption List: Table 1. Implementation time of each emission standard in China, Beijing, and Shanghai.

Figure Captions List: Figure 1. Emission factors of NH3 emitted from light-duty gasoline vehicles. Figure 2. Inter-annual NH3 emissions from two types of light-duty gasoline vehicles and provincial

of

contributions at periods during which different emissions standards were implemented.

ro

Figure 3. Inter-annual NH3 emissions contributed by light-duty gasoline vehicles with different

-p

emission standards.

re

Figure 4. Average NH3 emissions from passenger cars as a function of GDP (GRP) weighted by

lP

population from (a) national and (b) provincial perspectives. Figure 5. (a) Average P, (b) average VKT, (c) average EF and (d) average E in actual, fixed VKT

na

and fixed EF scenarios as a function of average GDP (GRP) weighted by population.

Jo ur

Figure 6. Inter-annual variation rates of P, VKT, EF and E with the variation of NH3 emission. Figure 7. NH3 emissions estimated by fixed P, fixed VKT, fixed EF and actual values in two stages from 1999 to 2017.

25

Journal Pre-proof Table 1. Implementation time of each emission standard in China, Beijing, and Shanghai. China

Beijing

Shanghai

State I

2000

1999

2000

State II

2005

2003

2005

State III

2008

2006

2008

State IV

2011

2008

2010

State V

2017

2013

2014

Jo ur

na

lP

re

-p

ro

of

Emission standard

26

Journal Pre-proof

NH3 emission factors (mg/km)

250 200 150

Tang et al. (2016) Huang et al. (2018) Heeb et al. (2008) Suarez Bertoa et al. (2014) This study

Heeb et al. (2006a) Heeb et al. (2006b) Suarez-Bertoa and Astorga (2016) Liu et al. (2014)

100 50

0

of

-0.5 State 0 0 0.5 State 1 I 1.5 State 2 II2.5 State 3 III3.5 State 4 IV4.5 State 5 V5.5 Mixed 6 6.5

Jo ur

na

lP

re

-p

ro

Figure 1. Emission factors of NH3 emitted from light-duty gasoline vehicles. (Mixed means the mixed situation of vehicles with multi-emission standards during the measurement.)

27

Journal Pre-proof HE HA SC 7.3% 4.5%3.4%

HE HA SC 4.8% 4.2%

State5 State V

ZJ 8.5%

HE HA SC 4.8% 4.2% JS6.6%

State3 State III State2 State II

ZJ 8.7%

JS6.4% 7.0%

ZJ 8.6%

HE HA SC JS 7.3% 4.4% 4.4% 5.3% ZJ 7.3% Others SD 36.9% 7.8% GD 9.8%

State1 State I

SD LN 8.9% SX SH YN 2.9% SN 2.8% HL 2.8% FJ GD AH 2.7% HN 2.6% HB 2.2% IM11.4% 2.1% JL 2.0% TJ GX 2.0% CQ 2.0% GZ JX 1.8% XJ GS 1.7% NX QH HI XZ 1.6% 1.5% 1.4% 1.2% 1.0% 0.7% 0.3% 0.1%

Others 41.0%

LN SD SX SH YN 3.6% SN 10.4% 3.4% HL FJ 3.1% AH 2.7% HN 2.7% HB BJ 1.7% IM GD 2.7% JL TJ 2.2% GX 2.2% 10.9% 2.8% CQ 2.4% GZ JX 2.1% XJ GS 1.5% NX QH HI XZ 1.7% 1.8% 1.4% 1.4% 0.9% 0.4% 0.3% 0.5% 0.2%

Others 41.3%

SD LN 9.1% SX SH YN 3.7% SN 3.6% HL FJ GD 3.1% AH 2.7% HN 2.6% HB 1.8% IM 12.8% 2.6% JL TJ 2.1% GX 2.1% CQ 2.4% GZ JX 2.0% XJ GS 1.5% NX QH HI XZ 1.7% 1.8% 1.5% 1.4% 1.3% 0.9% 1.0% 0.3% 0.4% 0.2%

6.9%

Others 43.8%

Others ZJ 45.4% 8.2% LN SX SH YN SD 4.3% SN HL 3.5% FJ 1.7% AH 3.0% 11.6% HN 3.0% HB 1.9% IM 3.1% JL 2.4% TJ GD BJ 2.5% GX CQ 2.7% GZ JX 2.5% XJ GS 1.7% NX QH HI 10.4% 1.2% XZ 1.6% 2.2% 0.9% 1.4% 1.6% 1.2% 1.0% 0.3% 0.5% 0.2%

LN SX SH YN SN 4.9% HL 3.4% FJ 0.9% AH 3.0% HN 2.7% HB 2.0% IM 3.2% JL 2.7% TJ GX 2.6% CQ 3.0% GZ JX 2.9% XJ GS 1.9% NX QH 0.8% HI XZ 2.6% 1.2% 1.6% 1.4% 1.2% 0.4% 0.6% 0.1%

150

120

BJ 6.1%

LDGT

PC 90

BJ 7.6%

60

BJ 16.9%

NH3 emission (Gg)

Emission standard

State4 State IV

JS 8.0%

HE HA SC 7.1% 4.8% 3.9% JS 7.6%

0

2018

2017

2016

2015

2014

2013

2012

2011

ro

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

State0 State 0

of

30

Jo ur

na

lP

re

-p

Figure 2. Inter-annual NH3 emissions from two types of light-duty gasoline vehicles and provincial contributions at periods during which different emissions standards were implemented. (BJ: Beijing; GD: Guangdong; SD: Shandong; ZJ: Zhejiang; JS: Jiangsu; HE: Hebei; HA: Henan; SC: Sichuan; LN: Liaoning; SX: Shanxi; SH: Shanghai; YN: Yunnan; SN: Shaanxi; HL: Heilongjiang; FJ: Fujian; AH: Anhui; HN: Hunan; HB: Hubei; IM: Inner Mongolia; JL: Jilin; TJ: Tianjin; GX: Guangxi; CQ: Chongqing; GZ: Guizhou; JX: Jiangxi; XJ: Xinjiang; GS: Gansu; NX: Ningxia; QH: Qinghai; HI: Hainan; XZ: Xizang.)

28

Journal Pre-proof NH3 emissions (Gg)

90

State 0

State I

State II

State III

State IV

State V

60

30

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

0

Jo ur

na

lP

re

-p

ro

of

Figure 3. Inter-annual NH3 emissions contributed by light-duty gasoline vehicles with different emission standards.

29

Journal Pre-proof 2000

2003

2006

2009

2012

2017

0.25

0.06 (a)

0.05 0.04 0.03 0.02

y = -37.19 x2 + 3.38 x - 0.02 R² = 0.98

0.01 0.00 0.00

0.01 0.02 0.03 0.04 0.05 0.06 Average GDP (million yuan/capita)

0.20 0.15 0.10 0.05

0.00 0.00

0.07

Other years y = -12.67 x2 + 2.14 x - 0.01 R² = 0.54

(b)

Average E (kg/capita)

Average E (kg/capita)

2015

0.02 0.04 0.06 0.08 0.10 0.12 Average GRP (million yuan/capita)

0.14

Jo ur

na

lP

re

-p

ro

of

Figure 4. Average NH3 emissions from passenger cars as a function of GDP (GRP) weighted by population from (a) national and (b) provincial perspectives.

30

Journal Pre-proof

0.10 0.05 (a) 0.02

0.02 0.01

0.10

y = -5176.89 x2 + 468.65 x + 29.94 R² = 0.05

60 40 20

(b)

0.00 0.00

0.14

80

0 0.00

0.03

0.08

0.02

Fixed EF

0.06

Actual 0.04

y = 11.80 x2 + 0.60 x + 0.00 R² = 1.00

0.02

y = -37.19 x2 + 3.38 x - 0.02 (d) R² = 0.98

0.00 0.04 0.06 0.08 0.10 0.12 Average GRP (million yuan/capita)

0.14

y = -38.95 x2 + 4.40 x - 0.03 R² = 0.99 Fixed VKT

(c) 0.02

0.04 0.06 0.08 0.10 0.12 Average GRP (million yuan/capita)

0.14

0.00

0.01

0.02 0.03 0.04 0.05 0.06 Average GDP (million yuan/capita)

ro

Average EF (mg/km)

100

0.04 0.06 0.08 0.10 0.12 Average GRP (million yuan/capita)

0.04

of

0.15

0.00 0.00

y = (0.01)ln(x) + 0.01 R² = 0.60

0.05 Average VKT (106 km)

0.20

y = -3.96 x2 + 2.07 x - 0.01 R² = 0.79

Average E (kg/capita)

Average P (veh/capita)

0.25

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

0.07

Jo ur

na

lP

re

-p

Figure 5. (a) Average P, (b) average VKT, (c) average EF and (d) average E in actual, fixed VKT and fixed EF scenarios as a function of average GDP (GRP) weighted by population. (P: vehicle population; VKT: vehicle kilometres traveled; EF: NH3 emission factors; E: NH3 emission)

31

Journal Pre-proof Stage I (2000-2007)

Stage II (2008-2017)

120%

100 VKT rate

EF rate

E rate

NH3 emission

90%

80

60%

60

30%

40

0%

20

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

0

1999

-30%

NH3 emission (Gg)

Variation rate

P rate

Jo ur

na

lP

re

-p

ro

of

Figure 6. Inter-annual variation rates of P, VKT, EF and E with the variation of NH3 emission. Same variable designations as in Fig. 5 caption.

32

Journal Pre-proof Fixed P scenario

Fixed VKT scenario

Fixed EF scenario

Actual scenario

270 270 Stage I (2000-2007)

Stage II (2008-2017) EF: 118% (13% ~ 244%)

210 210 VKT: 21% (3% ~ 40%)

180 180 150 150

VKT: 18% (4% ~ 32%)

P: -62% (-84% ~ -18%)

120 120 P: -50% (-78% ~ -12%)

90 90

2017 2017

2016 2016

2015 2015

2014 2014

2013 2013

2011 2011

ro 2009 2009

-p

2008 2008

2007 2007

re

2006 2006

2005 2005

2004 2004

2003 2003

2002 2002

2001 2001

2000 2000

0

1999 1999

30 30

2012 2012

of

EF: -77% (-90% ~ -48%)

60 60

2010 2010

NH3 emissions (Gg)

NH3 emission of LDGV (Gg)

240 240

Jo ur

na

lP

Figure 7. NH3 emissions estimated by fixed P, fixed VKT, fixed EF and actual values in two stages from 1999 to 2017.

33

Jo ur

na

lP

re

-p

ro

of

Journal Pre-proof

34

Journal Pre-proof

Highlights • NH3emissions from LDGVsin Chinahad a42-fold increasefrom 1999 to 2017. • Inverted U-shaped trend was found forNH3 emission per capita as a function of GDP per capita. • Decline of EFs is the decisive factor drivingthe downward trend of NH3emissions.

of

• Upgrade of emission standards can offset NH3 emissionincrease caused by rapid vehicle

Jo ur

na

lP

re

-p

ro

growth.

35