Multi-type air pollutant emission inventory of non-road mobile sources in China for the period 1990–2017

1State Environmental Protection Key Laboratory of Food Chain Pollution Control, Beijing Technology and Business University, Beijing 100048, China 2Key Laboratory of Cleaner Production and Integrated Resource Utilization of China National Light Industry, Beijing Technology and Business University, Beijing 100048, China 3School of Ecology and Environment, Beijing Technology and Business University, Beijing 100048, China


Diesel consumption of construction machinery
where Y is the diesel consumption of construction machinery (10 4 ton), P is the power of construction machinery (10 4 kW), LF is the load factor (0.75) (Hou et al., 2019), hr is the average activity time (h) of construction machinery in one year (770) (MEPC, 2014), 1 kWh = 860 kcal, q is the calorific value of diesel, and η is diesel engine efficiency (0.35) (Zhang, 2007).
No statistical data were available on the total power of construction machinery of construction enterprises in various regions of China for the period 1990-1992. In this study, the proportion of the total power of construction machinery of construction enterprises to the total power of machinery of construction enterprises in various regions of China in 1993 was used to calculate the missing data. The method for calculation of the total power of machinery of construction enterprises in various regions of China during 1991-1992 is shown in Eq. (2): where P is the total power of machinery of construction enterprises, P1 is the total power of machinery and equipment owned by state-owned construction enterprises, and P2 is the total power of machinery and equipment owned by collective-owned construction enterprises.
The method for calculation of the total power of machinery and equipment owned by construction enterprises in various regions of China in 1990 is shown in Eq. (3): where P is the total power of machinery and equipment owned by construction enterprises, R is the power of machines per labourer, and N is the number of staff and workers.

Diesel consumption of agricultural machinery
There were no agricultural diesel consumption data available for the period 1990-1992. Therefore, these data were obtained by estimation through linear fitting of the data from 1993-2003, as shown in Fig. S1.
In certain years, relevant activity data for some provinces and cities in China were missing from local statistical yearbooks. The missing data were estimated based on the proportion of provincial data to total national data in neighbouring years.
Power of other agricultural machinery diesel engines.
where P1 is the power of other agricultural machinery diesel engines (10 4 kW), P is the total power of agricultural machinery diesel engines (10 4 kW), P2 is the total power of large-and medium-sized tractors (10 4 kW), P3 is the total power of small tractors (10 4 kW), P4 is the total power of harvesting machines (10 4 kW), P5 is the total power of vessels (10 4 kW), P6 is the total power of diesel engines for diesel-powered agricultural drainage and irrigation machinery (10 4 kW), and ξ is the ratio of the total power of diesel engines to the total power of agricultural machinery.
In this study, the category of agricultural machinery was subdivided into different emission standards based on the principles of division of emission standards and the simulated survival curves of agricultural machinery. Owing to the lack of sales data for both irrigation/drainage machinery diesel engines and combine harvesters, it was assumed that the average power of irrigation and drainage machinery diesel engines is approximately equal to that of small tractors, and that the average power of a combine harvester is approximately equal to that of large-and medium-sized tractors. Therefore, the divisions of emission standards for drainage and irrigation machinery diesel engines were consistent with that of small tractors, and the division of emission standards for combine harvesters was consistent with large-and medium-sized tractors.

Diesel consumption of vessels
where Y is the fuel consumption for passenger transport by inland river/coastal vessels (10 4 ton), r is the passenger-kilometres by inland river/coastal vessels (10 4 passengerkm), T is the average weight of vessel passengers and carry-on baggage (80 kg/person) (Feng et al., 2014), q is the fuel consumption per 10,000 tkm of inland river/coastal vessels (kg/10 4 tkm).
The method for calculation of the fuel consumption of cargo vessels is shown in Eq. (6): where Y is the fuel consumption for inland river/coastal cargo vessels (10 4 ton), r is the freight ton-kilometres by inland river/coastal vessels (10 4 tkm), q is the fuel consumption per 10,000 tkm of inland river/coastal vessels (kg/10 4 tkm).
The parameters of fuel consumption per 10,000 km of inland and coastal vessels were divided into two cases: different provinces and units directly under the State Maritime Administration. Regarding the total turnover data of the direct totals after calculating the fuel consumption, the annual fuel consumption was distributed to provinces and cities based on the proportion of a province's turnover to the national turnover. Vessel passenger and freight data were missing for 1994 and 2017. This study collected the total passenger and freight data of the waterway transportation industry in those years, and then calculated the proportion of inland river and coastal passenger and freight data of the provinces and cities in the neighbouring years to the passenger and freight data of the waterway transportation of each province. Finally, the missing data were estimated based on the calculated ratio.
The method for calculation of the fuel consumption of fishing vessels is shown in Eq. (7): where Y is the diesel consumption of fishing vessels (10 4 ton), P is the power of fishing vessels (10 4 kW), LF is the load factor (0.75) (Hsieh et al., 2009), hr is the average activity time (h) of construction machinery in one year (1300) (Yu et al., 2008;Song, 2015), 1 kWh = 860 kcal, q is the calorific value of diesel, and η is diesel engine efficiency (0.35) (Feng et al., 2014).
There were no statistical data for 2017 for the power of motorized fishing vessels.
The data for 2017 were obtained through linear estimation using data from 1990-2016, as shown in Fig. S2.

Diesel consumption of diesel locomotives
where F is the total tonnage of railway locomotives (ton), N is number of railway locomotives, and Fi is the average total tonnage of freight locomotives (ton).
where F is the total tonnage of railway locomotives (ton), F1 is the total tonnage of diesel locomotives (ton), F2 is the total tonnage of electric locomotives (ton), and F3 is the total tonnage of steam locomotives (ton).
where η is the proportion of diesel locomotives in the national total of railway locomotives, F1 is the total tonnage of diesel locomotives (ton), and F is the total tonnage of railway locomotives (ton). The value in 2014 was used to estimate the proportion of diesel locomotives in subsequent years.
The methods for calculation of the fuel consumption of diesel locomotives for railway passengers and freight are shown in Eqs. (11) and (12), respectively: where Y1 and Y2 represent diesel consumption for passenger and freight transport by diesel locomotives, respectively (10 4 ton), Q is the passenger/freight-kilometres by railways (100 million passenger-km), T is the average weight of locomotive passengers and carry-on baggage (80 kg/person) (Feng et al., 2014), q is the oil consumption by diesel locomotives (kg/10 4 tkm), η is the proportion of diesel locomotives in the national total of railway locomotives, t1 is the static load of freight cars (97.6 ton) (Feng 8 et al., 2014), t2 is the weight of the passenger car itself (879.4 ton) (Feng et al., 2014), and t3 is the weight of the fright car itself (23 ton) (Feng et al., 2014).
For the statistical data of cargo turnover not allocated to provinces and cities from 2003-2014, this study allocated this year's data to provinces and cities based on the proportion of cargo turnover of each province and city to the total national cargo turnover.

Calculation of agricultural machinery emissions
The engine-power-based approach was used to calculate the emissions of all agricultural machinery except agricultural transport vehicles, as shown in Eq. (13): where Ej,k is the total emission (ton) of pollutant i, Pj is the population of agricultural machinery j, k is the emission standard, n is the power segment, G is the average installed engine power (kW), LF is the load factor (0.65) (MEPC, 2014), hr is the average activity time (h) of agricultural machinery in one year, EF is the corresponding emission factor (g/kwh). The pollutants (i) comprised CO, NOx, HC, PM2.5, PM10, BC, OC, and VOCs; the activity time (hr) was obtained from the technical guidelines and literature (Fan et al., 2011;MEPC, 2014); and the population (P), average installed engine power (G), and emission factor (EF) were as described in Sect. 2.2 and 2.3.

Calculation of agricultural transport vehicle emissions
The method used for calculation of the emissions of agricultural transport vehicles was similar to that adopted for on-road vehicles, as shown in Eq. (14): where Ei represents the emission (ton) of pollutant i, Pj,k is the population of agricultural transport vehicle j in stage k, Mj,k is the average annual number of kilometres travelled (km) (three-wheeled transport vehicles and low-speed trucks: 23,000 and 30,900 km, respectively) (MEPC, 2014), and EF is the emission factor (g/km). The pollutants (i) comprised CO, NOx, HC, PM2.5, PM10, BC, OC, and VOCs, and the population (P) and emission factor (EF) were as described in Sect. 2.2 and 2.3.

Calculation of SO2 emissions from non-road equipment
Calculation of SO2 emissions was based on the mass balance algorithm, as shown in Eq. (15): where E is the SO 2 emissions of non-road equipment, Y is the annual fuel consumption (kg), which is described in Sect. 2.2, and S is the sulphur content of the fuel (Table S8 in the Supplementary material).

Calculation of emissions of pollutants other than except SO2 from non-road equipment (except agricultural machinery and agricultural transport vehicles)
The CO, NOx, HC, PM2.5, PM10, BC, OC, and VOCs emissions from non-road equipment other than agricultural machinery and agricultural transport vehicles were estimated based on fuel consumption, as shown in Eq. (16): where E is the CO, NO x, HC, PM2.5, PM10, BC, OC, and VOCs emission of the nonroad equipment, Y is the annual fuel consumption (kg), which is described in Sect. 2.2, and EF is the emission factor, which is described in Sect. 2.3.

Uncertainty analysis
The 95% CI and CV are calculated using Eqs. (17) and (18) where CIx is the 95% confidence interval of x, CVx is the confidence of variation of x, μx is the arithmetic average of x, σx is the standard deviation of x, n is the number of observations of x, x is emission factor. b Chongqing municipality was founded in 1997, and relevant activity level data of Chongqing municipality from 1990 to 1996 were obtained by searching the local statistical yearbook. c With horsepower converted to kilowatt by 1 horsepower= 0.735 kilowatt. d After accounting, the sum of the two is the total power of machinery of construction enterprises. a To avoid double counting, fishing vessel diesel consumption has been subtracted from agricultural diesel consumption. b Chongqing municipality was founded in 1997, and relevant activity level data of Chongqing municipality from 1990 to 1996 were obtained by searching the local statistical yearbook. c In some years, the statistical data of Tibet province is missing from the national statistical data, which can be obtained by searching the local statistical yearbook. b Chongqing municipality was founded in 1997, and relevant activity level data of Chongqing municipality from 1990 to 1996 were obtained by searching the local statistical yearbook. c Before 1999, the data were divided into directly under the jurisdiction of state and sub-total of provinces. a After 2014, the data will not be counted and 2014 data will be used for estimation. b After 2016, the data will not be counted and 2016 data will be used for estimation. c Chongqing municipality was founded in 1997, and relevant activity level data of Chongqing municipality from 1990 to 1996 were obtained by searching the local statistical yearbook.  (1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003).