Emission inventory and control policy for non-road construction machinery in Tianjin

The establishment of a non-road construction machinery emission inventory forms the basis for the analysis of pollutant emission characteristics and for the formulation of control policy. We analyzed and investigated data on populations, emission factors, and activity levels for the construction machinery in Tianjin to estimate an emission inventory. Finally, a variety of emission reduction scenarios were used to simulate emission reductions and propose the most effective control policy. The results show that total emissions of CO, HC, NOx, PM10, and PM2.5 from non-road construction machinery in Tianjin of 2018 reached 4180.78, 951.44, 5833.85, 383.92, and 365.70 t, respectively. Forklifts, excavators, and loaders were the three most important emission sources in Tianjin. There are clear differences in the emissions of different districts. Large machinery emissions were mainly distributed across the Binhai New Area, which includes high volumes of port machinery and tractors in Tianjin Port. Based on various emission reduction scenarios, the effect of emission reductions is estimated. The IAD affected the reduction of CO and HC emissions with RR values of 17.6% and 17.3%, respectively, while EMO affected the mitigation of PM10 and PM2.5 emissions and RR values by 18.0% and 18.4%, respectively. The emission reduction control policy for non-road construction machinery is proposed, including the accelerated updating of non-road machinery emission standards; integrating diesel engine research and development institutions to accelerate the development of vehicle after-treatment technology; and establishing a cooperation mechanism for scientific research institutes, government departments, and enterprises in the control of non-road mobile machinery emissions.


Introduction
With continuous improvements in the control of motor vehicle emissions, pollution from non-road mobile machinery has become increasingly prominent. The carbon monoxide (CO), hydrocarbon (HC), nitrogen oxides (NOx), and particulate matter (PM) emitted by non-road mobile machinery account for 18-29% of mobile sources in the world (Yan et al. 2013). Due to the numerous modes of non-road mobile machinery production, the wide range of applications involved, the large number of possessions involved, and uncontrolled pollution emissions, the management and control of those emissions pose significant challenges. According to the China Motor Vehicle Environmental Management Annual Report (2019) (MEE 2019), the NOx emission of construction machinery, agricultural machinery, ships, and railway locomotives accounted for 37%, 32%, 26%, and 3%, respectively. In 2009, the European Environmental Protection Agency published the emission inventory, which showed that non-road mobile source was an important pollution source in Europe (EEA 2019). In Yan's research, non-road mobile source will become the main source of mobile pollution after 2030 and become the largest source of CO and HC pollutants in Asia after 2050 (Yan et al. 2013).
With the acceleration of urbanization and industrialization, the use of non-road construction machinery in China has grown rapidly. With this trend, the population of excavators and loaders increased from 120,000 and 230,000 vehicles in 2001 to 1.6 million and 1.55 million vehicles in 2017, reflecting 13.3-and 6.7-fold increases, respectively (CCMA 2017). As emission standards for Chinese construction machinery are still lax, emissions of NOx and PM are high. Most construction and road construction sites are distributed in densely populated urban living areas, and human health is highly vulnerable to such pollution. Further reducing people's confidence in air quality management directly affects the reputation of major cities in China.
Research on the emission control measures of non-road mobile source and their impacts on air quality must be supported with accurate source emission inventories. Since 1991, the US EPA has conducted a large number of studies on nonroad mobile source emissions. The CO, HC, NOx, and PM emission inventories of non-road construction machinery have been established (Chow et al. 2011;Diazrobles et al. 2009;Guo et al. 2020;Lin et al. 2005;Sarwar et al. 2005). At the end of the twentieth century, the European Union also developed an air pollution emission inventory of non-road mobile source. The results showed that non-road construction machinery contributions to NOx and PM were only less than those of industrial sources (Desouza et al. 2020;Winther and Nielsen 2011).
In China, the research on emission source inventory of nonroad mobile source is still in the initial stage. The Pearl River Delta (PRD), Yangtze River Delta (YRD), and Beijing-Tianjin-Hebei (BTH) regions have begun to research the emission inventory of non-road mobile sources (Guo et al. 2020;Hou et al. 2019;Lang et al. 2018;Li et al. 2019;Zheng et al. 2009;Zhong et al. 2018). Bian Yahui has constructed a data set of non-road mobile machinery based on the activity level, comprehensive emission factors, and space-time allocation factors in PRD. A bottom-up calculation method was used to establish the non-road mobile machinery emission inventory of the PRD in 2014. Construction machinery emissions in the PRD were dominated by those generated by construction vehicles and excavators, which contribute 40.1% and 33.9%, respectively . Huang Cheng has established the emission inventory of air pollution sources for non-road mobile machinery in the YRD based on a field survey of nonroad mobile machinery in these cities of the YRD. The results showed that CO and VOCs emitted from agricultural machinery accounted for 88% and 77%, respectively. NOx and PM 2.5 emissions from construction and municipal engineering machinery contributed 49% and 35%, respectively . Non-road machinery emission test research on BTH was relatively limited and has only started recently. An emission inventory of typical agricultural machinery for Beijing from 2006 to 2016 has been established based on the actual emission factors, population, and activity level (Hou et al. 2019). The results showed that CO, NOx, HC, and PM from agricultural machinery decreased by 63.11%, 72.07%, 62.93%, and 74.67% from 2006 to 2016. In Liang's study, a high temporal-spatial resolution emission inventory for agricultural machinery in China was developed (Lang et al. 2018). Whereas the above non-road mobile source emission studies on BTH focus on agricultural machinery, there have been no evaluations of the effects of construction machinery emissions and control policies in BTH.
To fill this gap, we measured the spatial distribution of nonroad construction machinery in Tianjin through surveys and compiled the non-road mobile construction machinery emission inventory of Tianjin in accordance with the methods of the emission inventory guidelines. The Monte Carlo model was used to calculate the uncertainty of the emission inventory. Finally, emission reduction scenarios were tested to assess the effects of different policies and to provide a basis from which government managers can formulate pollution control policies in the future.

Estimating emission methods
In this study, the emission of the pollutant from non-road construction machinery sources in Tianjin was estimated for the base year 2018.
Non-road construction machinery emissions were calculated with the method based on population and rated net power from the Non-road Mobile Source Emissions Inventory Compiled Technical Guidelines (Guidelines) (MEE 2014). The calculation formula is as follows: where E is annual emissions of CO, NOx, HC, PM 10 , and PM 2.5 (t); x, y, and z are the type of non-road mobile machinery, emission standard, and power segment, respectively; P is machinery population; G is the average rated power level (kw/ unit); LF is load factor; hr is annual used hours of each nonroad mobile machinery (h); and EF is emission factor of each pollutant (g/kWh).

Construction machinery population
The population of non-road construction machinery in Tianjin was obtained from government survey statistics and the China construction machinery industry yearbook. Forklifts, loaders, and excavators are the three main types of construction machinery used in Tianjin. In addition, road rollers, bulldozers, and graders are often used in road construction. In this study, construction machinery is classified as follows: forklifts, excavators, loaders, road rollers, bulldozers, tractors, cranes, graders, pavers, milling machines, generators, and other construction machinery.
According to surveys and estimates, the population of nonroad construction machinery in Tianjin was approximately 22481 by the end of 2018. The non-road construction machinery population and type distribution in each district were detailed in Figure 1 and Table S1. Forklifts, excavators, and loaders constitute the largest proportion of construction machinery, accounting for 47%, 23%, and 7%, respectively ( Figure 1). Based on the classification of emission standards, the proportion of Pre-China I, China I, China II, and China III is 8%, 26%, 51%, and 15%, respectively ( Figure 2). Table S2 shows that non-road construction machinery met Pre-China I and China I emission standards before 2010. With the extension of use time, older vehicles were scrapped and eliminated. China II emissions continued for nearly 6 years, resulting in the China II population accounting for the largest proportion.

Emission factors and activity data
Emission factor refers to the number of pollutants emitted per unit of activity level. For non-road construction machinery, the pollutant emission concentration when outputting a unit of power per unit of time was defined as the emission factor.
In this study, basic emission factors were obtained according to the Guidelines (Table S3).
The activity levels of non-road construction machinery are closely related to the different operating regions. In this study, we conducted a large number of investigations to make up for the lack of basic data on non-road machinery activity levels, which were including the average rated power, load factor, and average annual operating hours. The average rated power was obtained through dividing the total power by the population in Tianjin. The load factor was 0.65, which was found in the relevant research. The average annual operating hours of non-road construction machinery referred to the values in the Guidelines, and the activity level of each non-road construction machinery type is shown in Table S4.

Uncertainty calculation
Due to the lack of some key data and random errors in monitoring results, and random errors, certain uncertainties will be caused in establishing the emission inventory which include errors resulting from the statistical process of determining activity levels and a lack of localized emission factors. The Monte Carlo simulation model is used to quantify the range of uncertainty (Kelliher et al. 2017;Ramirez et al. 2008;Sun et al. 2019).
Monte Carlo simulation is a commonly used random sampling simulation method. Random sampling and simulation are performed to obtain the approximate solution to a problem that characterizes the probability distribution characteristics of the problem. The Monte Carlo simulation method involves three steps: (1) construct a probability process and determine the distribution probability of each parameter used in a calculation; (2) employ sampling from the known probability distribution and generate random numbers satisfying this distribution by computer for mathematical experiments; and (3) establish various estimators and investigate the estimators of random variables after mathematical experiments to characterize the distribution characteristics of the random distribution (Wu et al. 2009).
Crystal Ball software is used for Monte Carlo simulation in this study. We conduct 10,000 simulations of the uncertainty of the emission inventory of non-road construction machinery at a 95% confidence level.

Scenario analysis for emission reduction
Several policies have been introduced in many provinces and cities to control non-road construction machinery and include the elimination of old machinery (EOM) and the installation of after-treatment devices (IAD) ( Table 1). According to a preliminary investigation, it is more difficult to install exhaust after-treatment equipment onto Pre-China I and China I emission standard construction machinery due to the poor emission conditions of such machinery. It is possible to reduce emissions by eliminating old machinery. The emissions of China II non-road mobile machinery can be significantly reduced by installing after-treatment devices (DOC and DPF). The efficiency of gaseous and particulate matter pollutant emission reduction after transformation can reach roughly 70% and 90%, respectively (Guan et al. 2015;Pirjola et al. 2017;Resitoglu et al. 2020;Resitoglu et al. 2017;Zhang et al. 2021).
Reduction effects under different control scenarios were estimated by reduction rates (RRs) relative to the normal situation (NS) as given by the following equation: where ERS is the emission reduction scenario (including IAD, EOM); E is the non-road construction machinery emissions.

Quality assurance and quality control
The quality control of non-road construction machinery inventory data involved the following three steps: (1) Data sources on activity levels were checked and calibrated to Fig. 2 Emission standard distribution of non-road construction machinery ensure the accuracy of records used in the field investigation.
(2) The consistency of the survey space and of the time range of non-road construction machinery pollution sources was checked. (3) The integrity of the survey data was checked to ensure that the data covered all types of non-road construction machinery pollution sources.

Emission inventory of non-road construction machinery
The emission inventory of non-road construction machinery in Tianjin for 2018 was estimated using the methods described in the prior section. Total atmospheric pollutant emissions of CO, HC, NOx, PM 10 , and PM 2.5 were 4180.78, 951.44, 5833.85, 383.92, and 365.70 t, respectively. These values were several times higher than those of non-road mobile sources estimated in 2015 ). The average power calculation method was used, as it is more accurate than the fuel calculation methods used in previous similar studies. The population data used in this study were also more reliable and based on actual surveys. As HC and NOx from construction machinery were the important precursors of atmospheric PM 2.5 in megacities (Guo et al. 2020), emissions of non-road mobile machinery cannot be ignored. The emission control of non-road construction machinery in such cities must be improved.
The contribution rate of emissions from different construction machines is shown in Figure 3. Among the non-road construction machinery, forklifts, excavators, and loaders were the most important emission sources in Tianjin. While the emissions of individual forklifts were not significant, they generated the largest proportion of pollutants due to their frequent use. Excavators and loaders accounted for 30.8%, contributing 40-60% of all pollutants. Therefore, control policies on non-road mobile machinery should focus on large populations of construction machines. The spatial distribution of pollutants from non-road construction machinery in Tianjin is illustrated in Figure 4. There are obvious differences in the emissions of different districts. Large machinery emissions were mainly distributed across the Binhai New Area, Wuqing District, Xiqing District, Dongli District, Jinghai District, and Beichen District. Especially for the Binhai New Area, the proportion of total emissions reached 43% for all pollutants. The Heping, Hongqiao, Hebei, Hedong, and Nankai District were the regions with the lowest emissions, which was due to the high rates of urbanization and less building construction in these five regions. Emissions of non-road construction machinery were significantly negatively correlated with the level of urbanization. The contributions of nonroad construction machinery emission sources in different districts are shown in Figure 5. The results showed that forklifts, excavators, loaders, and road rollers are construction machines with the greatest contributions in each district. Among them, the contributions of emission sources in Heping District and Binhai New Area were different from those of other districts. As Heping District has the highest level of urbanization and there are no large factories in the area, emission sources here mainly included excavators, loaders, and road rollers. The Binhai New Area is the largest district in Tianjin with a large number of factories and the Tianjin Port (Chen et al. 2016;Na et al. 2017). On port terminals, a large number of tractors were operated. Therefore, in the Binhai New Area, forklifts, excavators, loaders, road rollers, and tractors generated the most emissions.

Emissions under different reduction scenarios
Emissions from non-road construction machinery and reduction rates under different reduction scenarios are shown in Figure 6 and Table 2. The emission will be controlled to different levels under different emission reduction scenarios. The IAD has an overwhelming effect on reductions of CO and HC emissions with RR values of 17.6% and 17.3%, respectively, while EOM has limited effects on the reduction of CO and HC emissions with RR values of 3.7% and 6.9%, respectively. The results indicate that the IAD and EOM have similar effects on NOx emissions control. RR values are 16.4% and 16.0% for the IAD and EOM, respectively. For PM 10 and PM 2.5 , the RR of EOM is higher than the IAD, indicating that the elevation of emission standards has a more obvious effect on PM emission reduction. At present, research on the emissions of non-road mobile machinery is in its infancy, and there are gaps in relevant non-road construction machinery policy  measures. The simulation results of the emission reduction scenarios in this study can thus guide future control policy development.

Uncertainty analysis
In this study, non-road mobile construction machinery population, activity level, and emission factor were the three forms of emission inventory uncertainty considered. At present, several cities in China have not established a registration and filing system for non-road mobile construction machinery. Non-road mobile construction machinery is widely used, and its working environments tend to be very harsh. It is difficult to determine the population, power level, and age of construction machinery. To mitigate uncertainty as much as possible, we have done our best to carry out detailed and accurate survey classifications. However, the types of sites surveyed, construction stages considered, and the representativeness of mechanical samples may introduce uncertainty.
The activity level of construction machinery was obtained through a large sample survey of more than 20,000 machines. However, some uncertainty and large relative deviations in activity level information for some machinery may remain due to the variety of machinery considered.
Emission factors are the weakest facet of research on the emission inventory of non-road construction machinery. The research results of non-road construction machinery emission factors are still very limited. It is impossible to accurately assess the emission levels of various machinery types, engine power, and emission standards. It is thus necessary to supplement actual measurement research on the localized emission factors of various non-road mobile construction machinery to mitigate emission inventory uncertainty.
For Monte Carlo simulation, the probability distributions and the coefficients of variation (CV, the standard deviation divided by the mean) of the parameters need to be determined. Generally, the activity data (average rated power, load factor, and average annual operating hours) obtained through field investigation had a lower CV of 5%. For the emission factors, the distribution was assumed to be normal with a CV of 50%. A normal distribution with CV of 5% was used for the population (Guo et al. 2020;Lang et al. 2018;Li et al. 2019;Lv et al. 2019). The average uncertainty ranges (at a 95% confidence level) for CO, HC, NOx, PM 10 , and PM 2.5 were measured as −41.12~39.47%, −38.48~38.69%, −40.55~42.35%, −37.98~37.31%, and −37.64~37.59%, respectively. While the uncertainty of the emission inventory is inevitable, the comprehensive prediction of the emissions of non-road mobile machinery can still guide the formulation of environmental policies.

Control policy
The emission standard of non-road mobile machinery must be improved. In China, most non-road mobile machinery only adheres to China II emission standard. Many vehicles can meet the China II emission standard by installing DOC, EGR, and other after-treatment devices, which has hindered the development of advanced control technologies such as DPF and SCR. Many companies cannot afford to make significant technology and capital investments to adopt advanced control technology, hindering technological research and development.
Therefore, the short-term goal of pollution control can be achieved by improving emissions management and accelerating the installation of DPF and SCR advanced exhaust treatment devices for non-road mobile machinery. The ultimate goal of the long-term control of non-road mobile machinery exhaust pollution can be achieved by providing preferential policies for research and development enterprises, accelerating the development of DPF and SCR advanced technology, and reducing costs of transformation.
In terms of the integration and development of non-road diesel engine manufacturers, due to the small scale and insufficient technical reserves of many diesel engine manufacturers in China, their manufacturing levels are limited, and the precision of key components of developed engines is low. As a result, the combustion of domestic engines cannot achieve the best efficiency or the requirements of higher emission standards. It is necessary to integrate decentralized capital, introduce technology, improve the production of non-road engines, and ultimately achieve emission reduction and quality improvements to non-road mobile machinery.
Pollution problems related to non-road mobile machinery have been attracted to the attention of the Chinese government. It is necessary to establish a cooperative mechanism for scientific research institutes, government departments, Fig. 6 the emissions from nonroad construction machinery under different reduction scenarios enterprises, and other relevant parties in the control of nonroad mobile machinery emissions. First, scientific research institutes must establish preliminary emission standards based on the data, and the government must then promulgate a detailed plan of action according to these standards. Enterprises must provide prompt feedback on the implementation of such standards and finally achieve the precise control of non-road machinery emissions.

Conclusion
This study provides comprehensive estimations of non-road construction machinery emissions of CO, HC, NOx, PM 10 , and PM 2.5 for Tianjin. Analysis results of emission reduction scenarios show that the IAD has an overwhelming effect on the reduction of CO and HC emissions with RR values of 17.6% and 17.3%, respectively. The EMO has a more obvious effect on the reduction of PM emissions. For NOx, the IAD and EMO have similar effects on NOx emissions control with RR values of 16.4% and 16.0% for IAD and EMO, respectively.
Finally, the emission reduction control policy for non-road construction machinery is proposed to provide basic support for the future management of non-road mobile machinery. The emission standard for non-road mobile machinery must be improved. The development of vehicle after-treatment technology should be accelerated to reach international standards, and cooperation between scientific research institutes, government departments, and enterprises in the control of non-road mobile machinery emissions must be established.
Funding This work was supported by the Tianjin Science and Technology Plan Project (19YFZCSF00960, 20YFZCSN01000), the Tianjin Education Commission Research Project (2017KJ053), and the Fundamental Research Funds for the Central University.
Data Availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations
Ethics approval and consent to participate Not applicable

Consent for publication Not applicable
Competing interests The authors declare no competing interests.