Characteristics and Uncertainty of Industrial VOCs Emissions in China

An industrial sector-based source volatile organic compounds (VOCs) emission inventory was developed for the year 2011 in China with the use of ‘emission factor method’. Based on the concept of source-tracing and a ‘bottom-up’ approach, it is estimated that the total emissions of industrial VOCs were 15.8 Tg for the year 2011, 17.7% of which came from production of VOCs, 8.4% came from storage and transport, 14.2% came from industrial processes using VOCs as raw materials, and 59.7% came from use of VOCs-containing products. Industrial VOCs emissions in East China were the largest contributors, comprising 40% of total. Four kinds of sources including furniture manufacturing, petroleum refining, machinery equipment manufacturing and printing industry took up 53.8% of total emissions. Therefore, more efficient VOCs control equipment should be paid attention to and promoted. A quantitative approach, the bootstrap simulation and Monte Carlo sampling, was used to analyze uncertainties, there is relatively medium uncertainty with an error of –42% to +71%, ranging from 9.2 to 27.2 Tg on the 95% confidence interval. To reduce VOCs emission inventory uncertainties requires giving priority to research on conducting more measurements of key sources. It’s imperative to establish domestic emission factor database.


INTRODUCTION
With the development of the economy, the complex and regional air pollution issues such as high ozone concentration, photochemical smog and haze caused by fine particles have become the most crucial topics for atmospheric environment research all over the world (Chameides et al., 1999;Chan and Yaom, 2008;Shao et al., 2009;Zhao et al., 2009;Von Schneidemesser et al., 2010;Li et al., 2011;Ling et al., 2011;Wang and Hao, 2012;Xue et al., 2013;Huang et al., 2014).As an important pollutant, volatile organic compounds (VOCs) have an adverse effect on ambient air quality and human health.Photochemical reactions occur when VOCs meet nitrogen oxides under the condition of strong light, low wind speed, low humidity, resulting in the formation of photochemical smog.Besides, VOCs are also precursors of secondary organic aerosol (SOA) (Yuan et al., 2013).What's worse, VOCs have harmful effects on the organs such as heart, lungs, liver and nervous system, causing acute and chronic poisoning, even carcinogenic and mutagenic (Zhou et al., 2011;Zhu and Liu, 2014).
Since 2010, China has become the world's second largest economy after United States, with industry, as a part of Secondary Industry in China, playing an essential role in Gross Domestic Products (GDP).Industry consumed approximately 70% of energy in 2012 and accounted for about 38% of total GDP in the meantime.Pollutants emissions generated in industry were substantial.In particular, industrial VOCs emissions (stack emissions and fugitive emissions) which are abundant, with a variety of compounds and properties, presenting complex and scattered, contributes a lot to VOCs emissions (Klimont et al., 2001;Ling et al., 2011;Fu et al., 2013).Emission inventory is a foundation of emission estimates (Chen et al., 2014).Several VOCs emission inventories have been established in China at different scales since recent decades.These VOCs inventories covered areas as large as the national scales (Wang et al., 2005;Liu et al., 2008;Wei et al., 2008;Chen et al., 2012;Cao et al., 2010;Yang, 2012;Streets et al., 2006) and down to regional (Fu et al., 2013;Huang et al., 2011;Liu et al., 2008;Lu et al., 2013;Streets et al., 2006;Zhao et al., 2012;Zheng et al., 2009a;Zheng et al., 2009b;Zheng et al., 2010;Yu et al., 2011;Zheng et al., 2009c) and urban levels (Cai et al., 2010;He et al., 2013).The accuracy of emission inventory is also essential for policy-makers and researchers.However, uncertainty in these inventories was rarely discussed or assessed in a qualitative way except few researchers (Wei et al., 2008;Zheng, Zhang et al., 2009;Zheng et al., 2010;Fu et al., 2013).
Understanding the characteristics of the industrial VOCs helps diagnose the impact of emission sources upon the real atmosphere and plan effective emission control strategies (Shin et al., 2013).Besides, industrial VOCs emissions are prone to be centrally controlled effectively among other sources.It is important and urgent that develop the industrial VOCs inventory in detail and timely.
Therefore, this paper aims to develop an industrial VOCs emission inventory based on the year 2011 in China.Furthermore, to quantify uncertainties in sector-based VOCs emission estimates by the use of bootstrap simulation and Monte Carlo sampling.

Methodology
A bottom-up approach was adopted, and the 'emission factor method' was used to calculate the VOCs emissions in this study.The annual emissions were calculated by using the following Eq.( 1): , , ,

E
( 1 ) where, j represents the province; k represents specific sector; E is annual total emission of VOCs; A is the activity data; EF denotes the emission factor; η is the removal efficiency.

Study Domain and Source Categorization
The study domain ranged over 31 provinces in China, which was exclusive of Hong Kong, Macau and Taiwan because of the limited data.
Source categorization is the foundation of an accurate emission inventory.Based on the concept of source-tracing (Chen et al., 2012;Yang, 2012), which represents VOCs' application properties and material flow process, industrial VOCs emissions mainly generated from four links: Production of VOCs, storage and transport, industrial processes using VOCs as raw materials and use of VOCs-containing products.Combining with <Industrial Classification for National Economic Activities>, source categorization of industrial VOCs emission inventory were established.Compared to other studies (Streets et al., 2006;Liu et al., 2008;Wei et al., 2008;Cao et al., 2010;Chen et al., 2012;Yang, 2012), this source categorization systematically covered the whole process of industrial VOCs emissions in more detail.84 particular industrial sources were established in total.

Compilation of Activity Data
Activity data of VOCs emission inventory refers to a variety of human activities information affecting pollutant emissions, including product yields, raw material consumption, population, etc., it's an important factor which affect the accuracy of the emission inventory.In this study, we referred much on official statistics for most activity data (e.g., fuel consumption, product output), some data were from Industry Association statistics (e.g., methanol, ink, adhesives), as shown in Table 1.63 activity data used in this study were directly collected at the provincial level, yet the rest 21 data in certain sectors, such as adhesive on decoration, wooding, etc., were available only at national level, and then were distributed to provinces by using GDP.

Determination of Emission Factors
Emission factor associates pollutant emissions with activity data, reflecting the level of pollution emission under per unit of activity intensity.In order to accurately reflect the industrial VOCs emissions, we preferred to follow basic principles for the choice of emission factors: emission characteristics measured in China were taken into consideration firstly (64 emission factors in total); when Chinese emission factors were not available, emission factors from other countries or regions were adopted for reference and correction according to domestic realities of situation.

VOCs Emissions Estimates in Regions
Table 2 gives the provincial industrial VOCs emissions in China.Industrial VOCs emissions for the year 2011 in China were 15.8 Tg, 17.7% of which came from production of VOCs, 8.4% came from storage and transport, 14.2% came from industrial processes using VOCs as raw materials, and 59.7% came from use of VOCs-containing products.The use of VOCs-containing products took up more than 50% in 71% of provinces, nevertheless, production of VOCs contributed 30%-40% in 19% of provinces.Therefore, to control the industrial VOCs emissions, the use of VOCscontaining products should be focused on primarily.The three largest contributors were Guangdong (2.18 Tg), Shandong (1.69 Tg), and Jiangsu (1.46 Tg), as shown in Fig. 1, Fig. 2, which accounted for 33.7% of the national total emissions.This phenomenon was mainly due to the more developed industries in three provinces, especially in furniture manufacturing and printing industry, resulting in greater proportion of coatings and adhesives applications, thus causing large amounts of VOCs emissions.The industrial VOCs emissions in East China were 6.5 Tg, accounting for 41% of national total emission, which were far more than other regions.This was because the provincial emissions in East China such as Shanghai, Jiangsu, Zhejiang, and Shandong occupied the front rank in China.

Regional VOCs Emissions Distribution
Fig. 3 presents industrial VOCs emissions per square kilometer and per capita of different provinces.Heavy emission densities were shown in East China, the top three average industrial VOCs emissions per square kilometer were Shanghai (186.1 ton/km 2 ), Tianjin (46.4 ton/km 2 ) and Beijing (17.7 ton/km 2 ), which were 11-113 times of the average level of China.It is the provincial different level of development, discrepant area and various industrial structure that mainly lead to huge variation range.The top three average industrial VOCs emissions per capita were Shanghai (0.05ton/capita), Tianjin (0.04 ton/capita), Zhejiang (0.02 ton/capita), respectively, which were 2-5 times as much as the average level of China.Generally speaking, there existed uneven in development among regions.

Emission Contributions by Major Sources
Among the four links of VOCs emissions, use of VOCscontaining products was the biggest contributor, comprising 59.7% of total emissions for the year 2011.What's more, furniture manufacturing accounted for 30% of the use of VOCs-containing products.28 kinds of sources including petroleum refining, basic chemical raw materials manufacturing, printing Industry, surface coating, paint, ink or pigments manufacturing, etc. took up 91.4% of total emissions.The emissions of major sources were shown as Table 3.

Assessment of Uncertainty in the Emission Inventory
A quantitative approach, the bootstrap simulation and Monte Carlo sampling, was utilized to analyze uncertainties in sector -based Industrial VOCs emissions estimates.According to previous studies, we have assumed that the activity data   and emission factor are made for a lognormal distribution (Zhao et al., 2011), besides, we estimate the error for each emitting subsector by combining the coefficients of variation (CV, or the standard deviation divided by the mean), which is based on the reliability of the data sources and the scope of application (Wei et al., 2011).The results show that there exist relatively medium uncertainties, with a 95% confidence interval of-42 to +71%, ranging from 9.2 to 27.2 Tg for this inventory.The 95% confidence interval of sectors are as shown in Table 4 and Fig. 4. The 95% confidence interval of production of VOCs, storage and transport, industrial processes using VOCs as raw materials and use of VOCs-containing products range from -58% to +136%, -61% to +135%, -8% to +83%, -62% to +115%, respectively, which is much higher than that of the total inventory.Considering the uncertainty of the emissions dates back to the transfer from activity data and emission factors, the uncertainty of sectors differs.The way to reducing uncertainty is to establish emission factor database in China, especially in the use of VOCs-containing products.
According to the sensitivity analysis, which were conducted by Monte Carlo with 10 000 simulation, as shown in Table 5, industrial VOCs emissions were assumed to be most sensitive to emission factor of Wood paintings (furniture manufacturing), contributing 37.7%, so a more specific investigation about paintings should be undertaken in China.

Comparison with Emission Inventories Comparison with Emission Inventories in China
A comparison was made with other studies in China to validate the inventory developed in this study.Table 6 presents the results and associated information for different studies.We chose some research and compare VOCs emissions in different years and scales, shown as Figs. 5  and 6.The national industrial VOCs emissions, which play an essential role in anthropogenic emissions, have an increase of     (Wei et al., 2008), 4.4 times to bio-fuel combustion source (Wei et al., 2008), 16.5 times to agricultural residues open burning (Wei et al., 2008) and 0.4 times to biogenic volatile organic compounds emissions (Li et al., 2013).
Besides, the regional industrial VOCs emissions have gone up 1 to 6 times from 2003 to 2011, because currently there are few control measures or standards for VOCs emissions in China.

Comparison with Emission Inventories in world
Industrial VOCs emissions in different countries were summarized in Table 7.It shows the industrial VOCs emissions in China are 0.6 times larger than that in European, and 1.6 times larger than that in American.In terms of the results from other countries, it is time to properly balance the development of high emission industries and VOCs emissions.

CONCLUSIONS AND RECOMMENDATIONS
Industrial sector-based source VOCs emissions inventory was developed for the year 2011 in China with the use of 'emission factor method', adopting a 'bottom-up' approach based on local source information.It is estimated from the inventory that the total emissions of industrial VOCs were 15.8 Tg for the year 2011, 59.7% of which came from the use of VOCs-containing products.
In general, the industrial VOCs emissions in East China accounted for 40% of total, ranking first in regions.The emission from furniture manufacturing, petroleum refining, chinery equipment manufacturing and printing industry were dominant based on the emission inventory, therefore, more efficient VOCs control technologies should be paid attention to and promoted, especially in mayor sources.Heavy emission densities were shown in East China, the average emission intensity in coastal provinces was much higher than that in inland provinces, part of that was because uneven in development existed among regions.
Relatively medium uncertainties were quantified with a 95% confidence interval of-42 to +71%, ranging from 9.2 to 27.2 Tg for VOCs estimates.It is highly necessary to strengthen the test of VOCs emission factors in China, whatever industrial or anthropogenic.
VOCs, as the key precursors of atmosphere pollutant, required to be reduced in order to improve the regional ozone and secondary aerosol pollution.However, the relevant regulations of VOCs have been neither comprehensive nor definite currently in China.As a consequence, more targeted and efficient control measures on reducing VOCs emissions are needed to be promoted in the national or regional planning.
For instance, as to emission factor for synthetic, <Volatile Organic Compounds Emissions Announcement Coefficient Applied for Stationary Sources Air Pollution Control Fees Application in Public or Private Places of Taiwan EPA (2009)> was used for reference.

Fig. 1 .
Fig. 1.Regional industrial VOCs emissions in China for 2011(N is short for North China; NE is short for Northeast China; E is short for East China; C & E is short for Central South China; SW is short for Southwest China; NW is short for Northwest China).

Fig. 2 .
Fig. 2. Provincial industrial VOCs emissions of total sources (a: production of VOCs, b: storage and transport, c: industrial processes using VOCs as raw materials, d: use of VOCs-containing products).

Fig. 4 .
Fig. 4. Industrial sector-based source VOCs emissions and uncertainty levels (Columns represent emissions and lines represent 95% confidence interval).

Fig. 5 .
Fig. 5. Comparison of national VOCs emissions inventories established by different studies (Columns represent emissions and lines represent 95% confidence interval).

Fig. 6 .
Fig. 6.Comparison of regional VOCs emissions inventories established by different studies (The industrial VOCs emissions of PRD in this study refer to Guangdong province, the industrial VOCs emissions of YRD in this study refer to Zhejiang province, Jiangsu province and Shanghai city).

Table 1 .
Source categorization, activity data and data sources of industrial VOCs emission inventory.
Storage & transportation of crude oil/gasoline/other oil Production/Import/Export Storage & transportation of solvent Production/Import/Export ACD
m Coke production Coke NBSC Paper production Pulp/Paper products ACD Garbage pollution treatment Sanitary landfill/Composting/MSW incineration NBSC Thermal power/Heat supply Coal/Fuel Oil/Liquefied petroleum gas/Natural gas Industrial consumption Coal/Fuel Oil/Coal gas/Liquefied/petroleum gas/Natural gas a NBSC represents National Bureau of Statistics of China.b NFIAC represents Nitrogen Fertilizer Industry Association of China.c ACD represents askci Corporation database.d PPEIAC represents Printing and Printing Equipment Industries Association of China.e CATIA represents China Adhesives and Tape Industry Association.f CDIA represents China Dyestuff Industry Association.g represents domestic PU pulp production capacity forecast from 2010 to 2012.h represents 2011 China coating industry market research report.i CBMF represents China Building Material Federation.j represents Tetrachlorethylene production status and production forecast.k DRCSCD represents Development Research Center of the State Council database.l represents Printed circuit board industry analysis report.m CEMIA represents China Electronics Materials Industry Association.

Table 2 .
Regional industrial VOCs emissions in China for 2011.

Table 3 .
Industrial VOCs emissions of major sources.

Table 4 .
Uncertainty assessment of sector-based source VOCs emission estimates.

Table 5 .
The parameters contributing most to the industrial VOCs emissions uncertainties.

Table 6 .
Comparison with other studies.
a Year refers when the inventories were established.b Huabei region includes administratively the Beijing and Tianjin Municipalities, Hebei and Shanxi Provinces, and Inner-Mongolia Autonomous Region.

Table 7 .
Comparison with other countries.