High-resolution historical emission inventories of crop residue burning in fields in China for the period 1990–2013

Atmospheric Environment 138 (2016) 152e161

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

High-resolution historical emission inventories of crop residue burning in fields in China for the period 1990e2013 Jing Li, Yaqi Li, Yu Bo, Shaodong Xie* College of Environmental Science and Engineering, State Key Joint Laboratory of Environmental Simulation and Pollution Control, Peking University, Beijing, People’s Republic of China

h i g h l i g h t s
Multi-year emission inventories of crop residue open burning were established.
Agriculture mechanization ratios and locally observed emission factors were introduced.
MODIS satellite data were used to allocate the annual provincial emissions into high resolutions.
Emissions form crop residues burning have strong temporal pattern.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 October 2015 Received in revised form 26 April 2016 Accepted 2 May 2016 Available online 13 May 2016

High-resolution historical emission inventories of crop residue burning in fields in China were developed for the period 1990e2013. More accurate time-varying statistical data and locally observed emission factors were utilized to estimate crop residue open burning emissions at provincial level. Then pollutants emissions were allocated to a high spatial resolution of 10 km  10 km and a high temporal resolution of 1 day based on the Moderate Resolution Imaging Spectroradiometer (MODIS) Fire Product (MOD/ MYD14A1). Results show that China’s CO emissions have increased by 5.67 times at an annual average rate of 24% from 1.06 Tg in 1990 to 7.06 Tg in 2013; the emissions of CO2, CH4, NMVOCs, N2O, NOx, NH3, SO2, PM2.5, OC, and BC have increased by 595%, 500%, 608%, 584%, 600%, 600%, 543%, 571%, 775%, and 500%, respectively, over the past 24 years. Spatially, the regions with high emissions had been notable expanding over the years, especially in the central eastern districts, the Northeastern of China, and the Sichuan Basin. Strong temporal pattern were observed with the highest emissions in June, followed by March to May and October. This work provides a better understanding of the spatiotemporal representation of agricultural fire emissions in China and can benefit both air quality modeling and management with improved accuracy. © 2016 Elsevier Ltd. All rights reserved.

Keywords: High-resolution Emission inventory Interannual variations Crop residue burning MODIS

1. Introduction Since the pioneering study by (Crutzen et al., 1979), biomass burning, which contains forest fires, savanna and grassland fires, crop residues burning, and peat combustion (van der Werf et al., 2010), has been considered an important source of atmospheric trace species and primary fine particles that has a significant impact on regional air quality, global climate change, as well as human health (Crutzen and Andreae, 1990; Akagi et al., 2012; Permadi and Oanh, 2013; Johnson et al., 2005; Li et al., 2016a).

* Corresponding author. E-mail address: [email protected] (S. Xie). http://dx.doi.org/10.1016/j.atmosenv.2016.05.002 1352-2310/© 2016 Elsevier Ltd. All rights reserved.

In China and other agriculture-based-economy countries, crop residue burning in fields is one important kind of biomass burning and poses a serious threat to human health and air quality (Streets et al., 2003; Marlier et al., 2013; Li et al., 2015), which began to receive governmental and scientific attention within China as early as the 1990s. Streets (2003) estimated that 730 Tg of biomass was burned in Asia in 2000, 250 Tg of which was the open burning of crop residues; China and India accounted for 110 Tg and 84 Tg, respectively of the residues burning. Quantifying the magnitude and trend of pollutants and greenhouse gas emissions from crop residue burning in fields is of great important for air quality modeling and management in China. However, emissions caused by crop residue burning in fields in China has not been studied in great detail and is not well characterized. It is essential to establish

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

an accurate and high resolution emission inventory of crop residue burning in fields. One approach to estimate biomass burning emissions is based on the burned area detected by satellites and calculated as the product of burned area, fuel loads, and combustion completeness. GBA2000 (Gregoire et al., 2003), GLOBSCAR (Simon et al., 2004), GLOBCARBON initiatives (Plummer, 2006) and Global Fire Emissions Database (GFED) (Randerson et al., 2012) have yielded burned area estimates for global biomass burning. However, this method has limitations, when used to quantify small fires, particularly under continued hazy sky conditions (Roy et al., 2008; Barret et al., 2010). In China, the average farming area of a farming household is only two acres (0.001334 km2) (China statistical yearbook, 2013). Moreover, beginning in 1997, Chinese government has enacted a series of regulations and laws to prohibit field burning (Yan et al., 2006); thus, it occurs on a small scale in dispersed locations, making it difficult to assess well via satellite (Streets et al., 2003). Another approach to estimate fire emissions is the use of the Seiler and Crutzen (1980) style equation to estimate emissions by multiply statistical data and the corresponding emission factors (EFs). The reliabilities of EFs, grain-to-straw ratios, dry matter content, proportion of crop residues burned in the field, and burning efficiency are the major challenges in producing an accurate emission inventory based on this method (Klimont and Streets, 2007). The spatial and temporal distribution of such burning can be presented in grids by combining fire counts product and land cover map. Since the emissions from crop residues burning in fields tend to be largely underestimated if they are based on global burned area products, and the burning of crop residues is strongly correlated with agricultural practices, we chose the second approach to build the historical emission inventory. Emissions from crop residue burning in fields in China have been estimated in several publications using this approach (Streets et al., 2003; Yan et al., 2006; Wang and Zhang, 2008; Zhang et al., 2008; Huang et al., 2012; Jain et al., 2014; Wu et al., 2016). However, the parameters used to calculate emissions inventories were imprecise, which can result in errors in simulations and mislead policy-making. For example, most early studies used EFs determined by foreign researches or applied the same factor to the burning of different crops; most previous studies assumed the proportion of crop residue burned in fields by using statistical reports or values based on experience; the grain-to-straw ratios used in previous studies were derived from outdated data based on the results of studies conducted from 1980 to 1984. Moreover, those inventories all focused on residues open burning emissions for a single year with a low temporal and spatial resolution. For most air quality simulations, a more detailed inventory with high temporal and spatial resolution is preferred. Understanding the interannual variations of residues burning in fields emissions and their spatial and temporal distribution is essential to support the analysis and modeling of air quality and climate change issues. In this study, we used the statistical data approach to develop China’s historical emission inventories of CO2, CO, CH4, NMVOCs, N2O, NOx, NH3, SO2, PM2.5, OC, and BC from crop residues burning in fields at a spatial resolution of 10 km  10 km for the period of 1990e2013. In the process of estimating China’s crop residues burning emissions, crop-specific local measured EFs and most detailed and accurate activity data were used. For the period 2001e2013, the annual provincial results were allocated to one day intervals and 10 km grid emissions by combining daily MODIS Thermal Anomalies/Fire products and Climate Change Initiative Land Cover Maps (CCI-LC Maps). For the year prior to the MODIS period (1990e2001), the annual emissions were allocated to county-level by county-specific sow area, and the county-level

153

inventories were converted to grids at a resolution of 10 km (). 2. Methodology 2.1. Emissions estimation Burning emissions of crop residues were initially calculated at the provincial level using the product of the mass of each crop type residues burned and the corresponding EF, as shown in equation (1):

Em;i ¼

j X EFi;j  Mm;j

(1)

where E is the emission from crop residues burning in fields, EF is the corresponding emission factor, Mm,j is the mass of crop residues burned in fields, m stands for each province (including 31 provinces, autonomous regions and municipalities in China mainland excluding Hong Kong and Macau Special Administrative Regions and Taiwan province), i for different emission species (including CO2, CO, CH4, NMVOCs, N2O, NOx, NH3, SO2, PM2.5, OC, and BC); j for different crop types (including rice, wheat, corn, legumes, tubers, cotton, peanut, and rapeseed). 2.1.1. Mass of residues burning in fields The provincial mass of crop residues burned of each crop type (M) was calculated on the basis of crop production using equation (2):

Mm;j ¼ Cm;j  Rm;j  4m  s  dj

(2)

where C is crop production (China Statistical Yearbook, 1991e2014); R is the grain-to-straw ratio, which refers to recent data shown in Table S1 (Wang et al., 2012); 4 is the proportion of crop residues burned in the field, s is the proportion of dry matter in the crop residue, and d is the burning efficiency. The parameter s is assumed as 80% according to the field measurements by Zhang et al. (2013), Jain et al. (2014). Burning efficiencies (d) specific to various crops were compiled from (Turn et al., 1997) and (de Zarate et al., 2005), with values obtained for rice (0.89), wheat (0.86), corn (0.92), beans (0.68), tubers (0.8), cotton (0.8), peanuts (0.8), and rapeseed (0.82). The proportion of crop residues burned in the field (4) was an important factor to be determined. Field surveys have shown that when a crop is harvested by a combine harvester, the proportion of residues burned is much greater than that when crops are harvested manually (Yang et al., 2008; Erenstein, 2011; Mishra and Shibata, 2012). In this study, 4 for crops harvested mechanically and for those harvested manually are given respectively. Eq. (2) can be transformed into Eq. (3):

   Mm;j ¼ Aj  am;j þ Bm  1  am;j  Cm;j  Rm;j  s  dj

(3)

where am,j is the proportion of the crop harvested mechanically for crop type j in province m, 1eam,j is the proportion of the crop harvested manually, A and B are f for crops harvested by combine harvester and by hand, respectively (the rest of the variables have already been defined). am,j was taken from the China Agricultural Machinery Industry Yearbook (Table S2). The coefficient A was determined from field investigations. About 82% of mechanically harvested wheat was burned in the field, whereas that of other straw crops was about 37% (Yang et al., 2008). The value of B was determined with the same method as Li et al., 2016b. In this study China’s provinces were categorized into six groups, according to the proportion of

154

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

mechanical harvesting in each province, the rural population, GDP, and the proportion of mountainous and hilly areas. Each group has a burning proportion for manual harvesting (B), determined by field investigations in areas representative of each group. The values of A and B were shown in Table 1 (Wang and Zhang, 2008; Yang et al., 2008; Li, 2014; Li et al., 2016b). aj,m were only available for the year before 2000. For the year 1990e2000, 4 was determined with the same method as Qin and Xie (2011).

2.1.2. Emission factors Most early studies used EFs measured in foreign experiments, or applied the same factor to the burning of different crops. EFs of crop residues burning were related to residues types, conditions, or burning practices (Zhang et al., 2008). Therefore, EFs measured in local experiments can better represent emissions from the burning of crop residues in China. In this study, emission inventories were established based on local experimentally determined EFs for each crop type, as shown in Table 2.

2.2. Spatial and temporal allocation 2.2.1. Spatial and temporal allocation for MODIS period (2001e2013) In order to produce an emission inventory with a higher resolution, we used fire count satellite data to locate the emissions for 2001e2013. We selected MODIS Thermal Anomalies/Fire gridded level-3 product (MOD/MYD14A1) to determine the fire counts, which provided data from both the Terra and Aqua satellites. MOD/ MYD14A1 is produced at 1-km resolution per day. Collection 5 of the MODIS fire detection data was used in this analysis, and the MODIS data were obtained through the NASA Land Process Distributed Active Archive Center (LPDAAC), USA (https://lpdaac. usgs.gov/products/modis_products_table). Croplands were identified using the European Space Agency CCI-LC Maps (http://maps.elie.ucl.ac.be/CCI/viewer/index.php). The CCI-LC Maps are 3-epoch series of global land cover maps at 300 m spatial resolution, where each epoch covers a 5-year period (1998e2002, 2003e2007, 2008e2012). The CCI-LC Maps characterizes the land cover with 24 different classes using UN-LCCS (Land Cover Classification System) classifiers. Only MODIS active fire detections in the land cover classes defined as “Cropland” or “Mosaic cropland (>50%)” in the CCI-LC Maps were identified as crops burning in fields. In this study, if a fire event detected by MODIS sensor was located in one 300 m cropland pixel in CCI-LC Maps, it could be recognized as an agricultural fire (Liu et al., 2015). In this study, annual emissions were allocated to a high temporal resolution of 1 day based on the fire counts per day; the emission inventories were gridded at a resolution of 10 km  10 km. The emissions in the i-th gird (Ei) were calculated using the following equation:

Ei ¼

FCi  Em FCm

(4)

where FCi is the cropland fire counts in the i-th gird, FCm is the total cropland fire counts in province m, and Em is the total emissions from crop residues burning in fields in province m estimated using equation (1). 2.2.2. Spatial and temporal allocation prior to the MODIS period (1990e2000) For the period before MODIS (1990e2000), provincial emissions were allocated to the county level based on crops seeded area using equation (5).

Ec;m ¼

CAc;m  Em CAm

(5)

where Ec,m is the emissions in county c of province m, Em is the emissions in province m, and CAc,m is the crops seeded area in county c of province m, CAm is the total crops seeded area in province m. After allocating the provincial emissions to counties, we further gridded the national map at a resolution of 10 km  10 km using a GIS software, MapInfo. The emission of each county was allocated to each grid cell based on the area of each grid cell belonging to that county. For the gird cells falling over various counties, the emissions from various counties within a grid cell were aggregated, which was the total emission of that grid cell. Annual emissions were allocated to a temporal resolution of 1 day based on the average fire counts per day between the years 2001e2013. 3. Results 3.1. Emission inventories of crop residue burning in fields in 2013 3.1.1. Emissions from crop residue burning in fields The annual emissions of CO2, CO, CH4, NMVOCs, N2O, NH3, SO2, NOX, PM2.5, OC, and BC were 158.54, 7.06, 0.54, 0.92, 0.01, 0.14, 0.07, 0.45, 1.14, 0.35, and 0.06 Tg, respectively, in 2013 (Eq. (1)). CO was used as an illustrative example in this paper, as it was widely studied in the open fire emission modeling (Streets et al., 2003). Emissions from each province are shown in Fig 1. The central eastern provinces of Henan, Anhui, Jiangsu, Shandong, and the northeastern province of Heilongjiang had the highest CO emissions, which contributed 13.42%, 10.01%, 8.60%, 7.34%, and 6.43% of total national CO emissions in 2013, respectively, while Tianjin, Hainan, Xizang, Shanghai, and Beijing provinces were the least contributors to national CO emissions, which in total contributed 0.94% to total CO emissions in 2013. The crop-specific CO emissions in each province are shown in Fig. 2. The most commonly burned biomasses were residues of wheat, rice, and corn straw, accounting

Table 1 The proportion of crop residues burned in different regions. Type of region

Province

Aa Wheat

Other crops

1. 2. 3. 4. 5. 6.

Non-agriculturally developed Northern plains Northeast China Southern plains Hilly and mountains Remote

Beijing, Guangdong, Hainan, Shanghai, Tianjin Hebei, Henan, Shandong, Shanxi, Shaanxi Heilongjiang, Jilin, Liaoning, Nei Monggol, Anhui, Hubei, Hunan, Jiangsu, Zhejiang Chongqing, Fujian, Guangxi, Guizhou, Jiangxi, Sichuan, Yunnan Gansu, Ningxia, Qinghai, Tibet, Xinjiang

82% 82% 82% 82% 82% 82%

37% 37% 37% 37% 37% 37%

a

A stands for burning proportions for harvest by combine harvester. B stands for burning proportions for harvest by hand.

b

Bb

8% 6% 3% 10% 20% 8.5%

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

155

Table 2 Emission factors for pollutants emitted from crop residue burning in fields (units: g kg1 dry fuel). Crop type

CO2

CO

CH4

NMVOCs

N2O

NH3

SO2

NOx

PM2.5

OC

BC

Wheat Corn Rice Cotton others

1470a 1350a 1105m 1345 1584

60a 53a 53.2m 106 102

3.36a 4.41a 5.82 5.82 5.82

7.48a 10.40a 6.05m 9.42 9.42p

0.07a 0.14a 0.07n 0.07n 0.07n

0.37a 0.68a 2.2 2.2 2.2

0.85a 0.44a 0.4n 0.4n 0.4n

3.3a 4.3a 3.83m 2.49 5.3

7.58a 11.7a 12.1m 6.3 ,6.3

2.69, 5.02b, 2.63c 3.94, 2.43i, 2.55j, 1.48k, 2.37l 3.3,3.04d, 1.97e, 1.61f, 1.38g, 1.81h 1.83 2.3

0.491, 0.4b, 0.64c 0.35, 0.28i, 2.32j, 0.31k, 0.22l 0.49d, 0.45e, 0.28f, 0.66g, 0.74h 0.82 0.8

a

(Li et al., 2007). (Cao, 2006): b Shanxi, c Hebei, d and e Changzhou and Huai’an city in Jiangsu, f Hubei, m (Zhang et al., 2013). n (Andreae and Merlet, 2001). o (Cao et al., 2008). p (Kudo et al., 2014). Other values are from (Akagi et al., 2011). b-l

Fig. 1. CO emissions from the field burning of crop residues in Chinese provinces in 2013 (units: Gg). 1 Anhui; 2 Beijing; 3 Chongqing; 4 Fujian; 5 Gansu; 6 Guangdong; 7 Guangxi; 8 Guizhou; 9 Hainan; 10 Hebei; 11 Heilongjiang; 12 Henan; 13 Hubei; 14 Hunan; 15 Jiangsu; 16 Jiangxi; 17 Jilin18; Liaoning; 19 Nei Mongol; 20 Ningxia; 21 Qinghai; 22 Shaanxi; 23 Shandong; 24 Shanghai; 25 Shanxi; 26 Sichuan; 27 Tianjin; 28 Xinjiang; 29 Xizang; 30 Yunnan; 31 Zhejiang.

for 91.2% of the total residues burned. The burning of each crop type tended to be regional. Most of the corn, wheat, and rice were burned in northeast China, the North China Plain, and South China, respectively. This is consistent with the country’s cereal crop distribution. 3.1.2. Spatial and temporal distribution of crop residue burning in fields Emission inventories of crop residues burning in fields for the year 2013 at a spatial resolution of 10 km  10 km and temporal resolution of 1 day were established. Agricultural fire counts were used to allocate the emission. A total of 40,440 fire counts were recorded on farmland in China in 2013. The spatial and seasonal distribution of fire counts are shown in Fig. 3. Each point in Fig. 3 represents a fire count occurrence in China. The fires in spring,

g

Sichuan,

h

Guangxi, i Shandong, j Jilin,

k

Henan, l Nei Mongol.

Fig. 2. Province-level crop-specific CO emissions (Gg) in China in 2013. 1 Anhui; 2 Beijing; 3 Chongqing; 4 Fujian; 5 Gansu; 6 Guangdong; 7 Guangxi; 8 Guizhou; 9 Hainan; 10 Hebei; 11 Heilongjiang; 12 Henan; 13 Hubei; 14 Hunan; 15 Jiangsu; 16 Jiangxi; 17 Jilin18; Liaoning; 19 Nei Mongol; 20 Ningxia; 21 Qinghai; 22 Shaanxi; 23 Shandong; 24 Shanghai; 25 Shanxi; 26 Sichuan; 27 Tianjin; 28 Xinjiang; 29 Xizang; 30 Yunnan; 31 Zhejiang.

summer, autumn and winter were colored green, red, orange, and blue, respectively. Fire counts were more concentrated in summer, especially concentrated in the area around Beijing, the capital of China (Fig. 3.2). Fire counts in other seasons were relatively less regionally concentrated, and burning was uniformly scattered over most of the agricultural zone. Monthly variations of CO emissions and agricultural MODIS fire counts were shown in Fig. 4. The month when most crop fires occurred was June (25.47%), followed by May (14.40%), October (12.75%), and July (9.96%). Daily variations of CO emissions in the five highest emission provinces were described in Fig. 5. Henan, Anhui, and Shandong province locate in the North China Plain, the largest agricultural zone in China. In these

156

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

Fig. 3. Spatiotemporal distribution of agricultural fires in China in 2013: 1 Spring (MarcheMay); 2 Summer (JuneeAugust); 3 Autumn (SeptembereNovember); 4 Winter (DecembereFebruary).

Fig. 4. Monthly variations of CO emissions (grey bar) and fire counts (dashed line) in 2013.

provinces, the fire occurrences are predominantly in early and middle June, with other small peaks in early Mar, May, and October. Heilongjiang province locates in another notable area with extensive burning, the Northeastern China. In this province, late April

Fig. 5. Daily variations of CO emissions in the five highest emission provinces in 2013.

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

157

and mid-May are the sowing times (Huang et al., 2012), and early October is the only harvest time all year round. Crop resides are usually burned before sowing time or after harvest time (Wang and Zhang, 2009). Two peak emission periods, in late April and October, are presented in Fig. 5 for Heilongjiang province. Jiangsu province locates in eastern and coastal China, which has two or even three harvest times each year. As shown in Fig. 5 for Jiangsu province, agricultural emissions range over almost every month from April to September with several peaks. Fig. 6 presents the spatial distribution of CO emission in 2013 in 10 km grid cell. CO emissions from crop residues burning in fields were obviously high in the central eastern districts (Zone 1), including the southern Hebei, eastern Shandong, northern Anhui, and Henan province. The high CO emission grids also contained the Sichuan Basin and the south part of Shanxi (Zone 2), the Northeast of China (Zone 3). 3.2. Historical emission inventories of crop residues burning in fields 3.2.1. Emissions from crop residues burning in fields Temporal crop-residue production and the proportion of crop residues burned in fields (4) were presented in Fig. 7, showing that residue production increased rapidly in the early 1990s, and then remained at about 580 Tg from 1996 to 1999, it decreased in 2000, and later increased gradually after 2000, to about 718 Tg in 2013. In the past decades, wheat, rice, and corn are the main crops in China. The productions of other crops are relatively small. The productions components of different crops remain stable. So contribution of various crop residues to the total production each year has not changed significantly; wheat straw, rice straw and cornstalk contributed averagely 24%, 34% and 25% to total straw production from 1990 to 2013, respectively. Historical emissions of crop residues burning in fields were estimated for the period 1990e2013, as presented in Table 3, showing that CO emissions have increased by 567% at an annual average rate of 24% from 1.06 Tg in 1990 to 7.06 Tg in 2013. The emissions of CO2, CH4, NMVOCs, N2O, NOx, NH3, SO2, PM2.5, OC, and

Fig. 7. Chinese crop residues production (bar) and the proportion of crop residues burned in fields (f, line) from 1990 to 2013.

BC have increased by 595%, 500%, 608%, 584%, 600%, 600%, 543%, 571%, 775%, and 500%, respectively, from 1990 to 2013. CO emissions indifferent provinces and years were listed in Table S3. Spatially, provinces including Henan, Anhui, Jiangsu, Shandong, and Heilongjiang were the largest contributors to CO emissions through the estimation period. Generally, these provinces altogether made up 40.83%, 43.94%, 44.28%, 45.97%, 47.87, and 45.81% of national residues production in 1990, 1995, 2000, 2005, 2010, and 2013 respectively. 3.2.2. Spatial and temporal distribution Annual and monthly agricultural MODIS fire counts from 2001 to 2013 were shown in Fig. 8. A total of 301,762 fire counts were recorded in China from the year 2001e2013 on farmland. Annual fire counts have been on the rise from 2001 to 2013, showing a similar trend to that of pollutants emissions. Generally, the monthly variations of fire counts for different years were similar. Fire counts were extraordinarily concentrated in June, followed by March to May and October, and was fewest during November to January. Emission inventories of crop residue burning in fields at a high resolution of 10 km  10 km were established by GIS methodology, as presented in Fig. 9, which shows the spatial distribution of CO emissions in 1990, 1995, 2000, 2005, 2010, and 2013. Fig. 9 illustrates that the spatial distributions in each year were similar, but the emission intensity of high emission regions had been increasing and the area of these regions had been expanding over the years. CO emission were generally concentrated in the central eastern districts, the Sichuan Basin, and the Northeast of China, and highemitting regions have expanded significantly from 1990 to 2013, in which the central eastern districts were the most expanded area. 4. Discussion 4.1. Uncertainty analysis

Fig. 6. Spatial distribution of CO emission in 2013 in 10 km grid cell (Kg grid1).

Monte Carlo simulation was performed to study the uncertainty in the emission inventories established in this study. The overall uncertainty in emissions for crop residue burning in fields in China is that of activity data and EFs. The uncertainty of activity data is from four aspects: crop production, grain-to-straw ratio, combustion efficiency, and proportion of crop residues burned in fields. We employed different estimation methods for these four types of activity data. Crop production is from Statistical Yearbook, which has no

158

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

Table 3 Annual fire emissions for each species considered in this study (Tg year1). Year

CO2

CO

CH4

NMVOCs

N2O

NH3

SO2

NOx

PM2.5

OC

EC

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

22.80 23.14 26.53 31.83 41.26 58.74 76.21 84.38 88.74 92.55 89.86 86.32 88.02 87.92 95.84 100.73 108.18 112.41 123.18 130.74 137.71 147.50 150.40 158.54

1.06 1.08 1.24 1.47 1.92 2.75 3.54 3.93 4.12 4.31 4.07 3.93 4.00 4.14 4.37 4.57 4.89 5.04 5.53 5.87 6.16 6.60 6.70 7.06

0.09 0.09 0.10 0.12 0.15 0.22 0.28 0.31 0.33 0.34 0.30 0.29 0.30 0.30 0.33 0.34 0.36 0.38 0.41 0.44 0.47 0.50 0.51 0.54

0.13 0.14 0.16 0.19 0.24 0.35 0.45 0.49 0.53 0.55 0.50 0.48 0.49 0.50 0.54 0.56 0.60 0.63 0.69 0.74 0.78 0.85 0.88 0.92

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

0.02 0.02 0.03 0.03 0.04 0.06 0.08 0.09 0.09 0.10 0.08 0.08 0.08 0.08 0.09 0.09 0.10 0.10 0.11 0.11 0.12 0.13 0.13 0.14

0.01 0.01 0.01 0.01 0.02 0.02 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.06 0.06 0.06 0.07 0.07 0.07

0.07 0.07 0.08 0.09 0.12 0.18 0.23 0.25 0.26 0.28 0.25 0.24 0.25 0.24 0.27 0.28 0.30 0.31 0.34 0.37 0.39 0.42 0.43 0.45

0.17 0.18 0.20 0.23 0.30 0.43 0.57 0.62 0.66 0.68 0.61 0.59 0.61 0.60 0.66 0.69 0.74 0.78 0.86 0.91 0.97 1.05 1.09 1.14

0.04 0.04 0.05 0.06 0.08 0.13 0.17 0.19 0.20 0.21 0.19 0.19 0.19 0.19 0.21 0.22 0.23 0.24 0.27 0.29 0.30 0.33 0.34 0.35

0.01 0.01 0.01 0.01 0.02 0.02 0.03 0.03 0.03 0.04 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.05 0.05 0.05 0.06 0.06 0.06

Fig. 8. Annual and monthly agricultural MODIS fire counts from 2001 to 2013.

corresponding comparative data. Thus we refer to TRACE-P experience value method, and assumed a normal distribution with a coefficients of variation (CV) of ±30% (Streets, 2003). Followed the assumption by Zhao et al. (2011), grain-to-straw ratio, combustion efficiency, and the proportion of crop residues burned in fields were assumed uniform distributions. The Monte Carlo simulation is repeatedly implemented with new input values randomly selected from within the respective probability distribution of activity data of four aspects. This process was repeatedly run for 100,000 times, the range of the mass of crop residue burned in fields in 2013 in China is from 54 to 259 Tg, and the propagation of the uncertainty of the mass of crop residue burned in fields at the 95% confidence interval (CI) is about (63%, 78%). Regarding the EFs of gaseous pollutants, we followed the assumption by Streets (2003), assuming that the underlying EF measurements were normally distributed. For EFs of particulate matter, we followed the assumption by (Bond, 2004), assuming that EFs for BC, OC and PM2.5 followed a lognormal distribution. For EFs based on local observations, we assumed that they had the least uncertainty with a CV of ±50%; and EFs based on foreign studies had a CV of ±150% (Qin and Xie, 2011).

Based on the probability density distribution of activity data and EFs (Table S4), we repeatedly used the Monte Carlo simulation for 100,000 times. Fig. 10 illustrates the probability distribution of CO emission inventory of crop residues burning in fields in 2013 in China, the mean value, the 2.5th percentile value and the 97.5th percentile value of which were 7087 Gg, 2138 Tg, and 14,832 Gg, respectively. The propagation of uncertainty of CO emissions from crop residue burning in fields in 2013 at the 95% CI was about (70%, 109%). For other pollutants, the propagations of the uncertainty of the total emission at the 95% CI were as follows: CO2 (70%, 111%), CH4 (70%, 111%), NMVOCs (71%, 112%), N2O (72%, 119%), NH3 (73%, 123%), SO2 (70%, 110%), NOx (70%, 111%), PM2.5 (67%, 104%), OC (67%, 109%), BC (62%, 77%). In addition, uncertainty in the methods of spatiotemporal allocation may also be a component of the uncertainty of these emission inventories. For the period 2001e2013, the annual provincial results were allocated to 10 km grid emissions by MODIS fire counts product. Small agricultural fires may be undetected by MODIS sensor. The fire count omission of the MODIS data associated with crop-dominant areas on Brazilian Amazonia was 75% (Schroeder et al., 2008). However, these probable errors couldn’t be quantified for China without an accurate satellite product (Liu et al., 2015). For the period before MODIS (1990e2000), provincial emissions were allocated based on county-level crops seeded area, which could bring about more uncertainties in the emission allocation.

4.2. Comparison with other emission inventories Emissions from field burning of crop residues in China have been estimated in several publications. Comparison of the emissions calculated in our study with previous estimates is listed in Table 4. The emissions are determined by crop productions, grain-tostraw ratios, proportions of crop residues burning in fields and EFs. The cropland fire emissions of CO2 estimated in this study were higher than the emissions in Huang et al. (2012) and lower than other studies listed in Table 4. There are minor differences in other pollutant emissions between this study and other studies. Lower proportions of crop residues burning in fields used in Huang et al.

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

159

Fig. 9. Annual changes in spatial distribution of CO emissions at a high resolution of 10 km  10 km from crop residue burning in fields in 1990, 1995, 2000, 2005, 2010, and 2013.

Fig. 10. Probability distribution of China’s national CO emissions from crop residue burning in fields in 2013, based on 100,000 Monte Carlo simulations.

Table 4 Comparison of the emissions calculated in this study with previous estimates. Reference

Year

CO2

CO

CH4

NMVOCs

N2O

NH3

SO2

NOx

PM2.5

OC

EC

This study Tian et al., 2011 This study Huang et al., 2012 Wang and Zhang, 2008 This study Qin and Xie, 2011 This study Streets et al., 2003 Yan et al., 2006

2007

112 142 108 68 155 101 e 90 160 184

5 9 5 4 7 4 6 4 10 11

0.38 0.43 0.36 0.25 0.37 0.34 e 0.3 0.28 0.32

0.63 0.76 0.60 2.20 0.87 0.56 e 0.50 1.70 1.92

0.01 e 0.01 e e 0.01 e 0.01 e 0.01

0.1 e 0.1 0.09 0.08 0.09 e 0.08 0.13 0.16

0.05 0.04 0.05 0.02 0.06 0.05 e 0.04 0.04 0.05

0.31 0.36 0.3 0.23 0.36 0.28 e 0.25 0.4 0.47

0.78 0.54 0.74 0.27 2.17 0.69 e 0.61 e 0.47

0.24 e 0.23 0.1 0.48 0.22 0.26 0.19 0.03 0.4

0.04 0.06 0.04 0.03 0.05 0.04 0.07 0.03 0.07 0.08

2006

2005 2000

160

J. Li et al. / Atmospheric Environment 138 (2016) 152e161

(2012) were the main reason for the lower CO2 emissions. The proportions of crop residues burning in fields used in Huang et al. (2012) were about 6.6% on average. The values used in this study were ranged from 11% to 44%, similar to the values used in other studies listed in Table 4. Another important factor influencing the result was the EF value. Higher emissions in most previous estimates were caused by EFs obtained from studies in other countries, such as Akagi et al. (2011), and Andreae and Merlet (2001). Besides, all the previous studies used a combined EF for different types of crops. So the emissions in previous studies were different from those in this study. 5. Conclusions Comprehensive multi-year emission inventories of crop residue burning in fields have been established in China from 1990 to 2013, based on official statistical data, field measured burning ratios, and domestic EFs, which were further allocated at a high spatial resolution of 10 km  10 km by MODIS agricultural fire counts and county-specific sow area. Results showed that CO emissions from crop residue burning in fields in China had increased by 5.67 times at a yearly average rate of 24% from 1.06 Tg in 1990 to 7.06 Tg in 2013. The emissions of CO2, CH4, NMVOCs, N2O, NOx, NH3, SO2, PM2.5, OC, and BC have increased by 595%, 500%, 608%, 584%, 600%, 600%, 543%, 571%, 775%, and 500%, respectively, from 1990 to 2013. Temporally, the monthly variation of emissions for different years were similar. Emissions were extraordinarily concentrated in June, followed by March to May and October, and was lowest during November to January. Spatially, provinces of Henan, Anhui, Jiangsu, Shandong, and Heilongjiang were the largest contributors to CO emissions through the estimation period. The spatial distribution of gridded emissions revealed that remarkably high emission areas mainly concentrated in the central eastern districts, the three northeastern provinces, and the Sichuan Basin. High emission regions have been expanding notably from 1990 to 2013, especially in the central eastern districts. Acknowledgments This work was founded by the Environmental Protection Ministry of China for Research of emission reduction and regulatory system of VOCs in key sectors (no. 20130973) and the Natural Science Foundation for Research of establishment and validation of anthropogenic VOC emission inventory in Beijing-Tianjin-Hebei region (grant no. 91544106). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2016.05.002. References Akagi, S.K., Yokelson, R.J., Wiedinmyer, C., Alvarado, M.J., Reid, J.S., Karl, T., Crounse, J.D., Wennberg, P.O., 2011. Emission factors for open and domestic biomass burning for use in atmospheric models. Atmos. Chem. Phys. 11, 4039e4072. http://dx.doi.org/10.5194/acp-11-4039-2011. Akagi, S.K., Craven, J.S., Taylor, J.W., McMeeking, G.R., Yokelson, R.J., Burling, I.R., Urbanski, S.P., Wold, C.E., Seinfeld, J.H., Coe, H., Alvarado, M.J., Weise, D.R., 2012. Evolution of trace gases and particles emitted by a chaparral fire in California. Atmos. Chem. Phys. 12, 1397e1421. http://dx.doi.org/10.5194/acp-12-13972012. Andreae, M.O., Merlet, P., 2001. Emission of trace gases and aerosols from biomass burning. Global Biogeochem. Cycles 15, 955e966. http://dx.doi.org/10.1029/ 2000gb001382. Barret, B., Williams, J.E., Bouarar, I., Yang, X., Josse, B., Law, K., Pham, M., Le Flochmoen, E., Liousse, C., Peuch, V.H., Carver, G.D., Pyle, J.A., Sauvage, B., van

Velthoven, P., Schlager, H., Mari, C., Cammas, J.P., 2010. Impact of West African Monsoon convective transport and lightning NOx production upon the upper tropospheric composition: a multi-model study. Atmos. Chem. Phys. 10, 5719e5738. http://dx.doi.org/10.5194/acp-10-5719-2010. Bond, T.C., 2004. A technology-based global inventory of black and organic carbon emissions from combustion. J. Geophys. Res. 109 http://dx.doi.org/10.1029/ 2003jd003697. Cao, G.L., 2006. Emission Inventory of carbonaceous Aerosol from China. Doctoral Thesis. Chin. Acad. Sci., Beijing, p. 112. Cao, G.L., Zhang, X.Y., Gong, S.L., Zheng, F.C., 2008. Investigation on emission factors of particulate matter and gaseous pollutants from crop residue burning. J. Environ. Sci. China 20, 50e55. http://dx.doi.org/10.1016/s1001-0742(08) 60007-8. Crutzen, P.J., Heidt, L.E., Krasnec, J.P., Pollock, W.H., Seiler, W., 1979. Biomass burning as a source of atmospheric gases CO, H2O, N2O, NO, CH3Cl and COS. Nature 282, 253e256. Crutzen, P.J., Andreae, M.O., 1990. Biomass burning in the tropics - impact on atmospheric chemistry and biogeochemical cycles. Science 250, 1669e1678. http://dx.doi.org/10.1126/science.250.4988.1669. Erenstein, O., 2011. Cropping systems and crop residue management in the TransGangetic Plains: issues and challenges for conservation agriculture from village surveys. Agric. Syst. 104, 54e62. http://dx.doi.org/10.1016/ j.agsy.2010.09.005. Gregoire, J.M., Tansey, K., Silva, J.M.N., 2003. The GBA2000 initiative: developing a global burnt area database from SPOT-VEGETATION imagery. Int. J. Remote Sens. 24, 1369e1376. http://dx.doi.org/10.1080/0143116021000044850. Huang, X., Li, M., Li, J., Song, Y., 2012. 4A high-resolution emission inventory of crop burning in fields in China based on MODIS thermal anomalies/fire products. Atmos. Environ. 50, 9e15. http://dx.doi.org/10.1016/j.atmosenv.2012.01.017. Jain, N., Bhatia, A., Pathak, H., 2014. Emission of air pollutants from crop residue burning in India. Aerosol Air Qual. Res. 14, 422e430. http://dx.doi.org/10.4209/ aaqr.2013.01.0031. Johnson, N.G., Bryden, M., Xiao, A.R., Asme, 2005. Risk analysis and safety evaluation of biomass cookstoves. In: Proceedings of IMECE 2005: Engineering/ Technology Management. Amer Soc Mechanical Engineers, New York, pp. 185e191. Klimont, Z., Streets, D.G., 2007. Emissions inventories and projections for assessing hemispheric or intercontinental transport. In: Keating, T., Zuber, A. (Eds.), Hemispheric Transport of Air Pollution 2007. United Nations, Geneva, Atmospheric Pollution Studies No. 16, ECE/EB.AIR/94. Kudo, S., Tanimoto, H., Inomata, S., Saito, S., Pan, X., Kanaya, Y., Taketani, F., Wang, Z., Chen, H., Dong, H., Zhang, M., Yamaji, K., 2014. Emissions of nonmethane volatile organic compounds from open crop residue burning in the Yangtze River Delta region, China. J. Geophys. Res. Atmos. 119, 7684e7698. http://dx.doi.org/ 10.1002/2013jd021044. Li, J., Xie, S.D., Zeng, L.M., Li, L.Y., Li, Y.Q., Wu, R.R., 2015. Characterization of ambient volatile organic compounds and their sources in Beijing, before, during, and after Asia-Pacific Economic Cooperation China 2014. Atmos. Chem. Phys. 15, 7945e7959. http://dx.doi.org/10.5194/acp-15-7945-2015. Li, J., Wu, R., Li, Y., Hao, Y., Xie, S., Zeng, L., 2016a. Effects of rigorous emission controls on reducing ambient volatile organic compounds in Beijing, China. Sci. Total Environ. 557e558, 531e541. http://dx.doi.org/10.1016/ j.scitotenv.2016.03.140. Li, J., Bo, Y., Xie, S., 2016b. Estimating emissions from crop residue open burning in China based on statistics and MODIS fire products. J. Environ. Sci. http:// dx.doi.org/10.1016/j.jes.2015.08.024. Li, X.G., Wang, S.X., Duan, L., Hao, J., Li, C., Chen, Y.S., Yang, L., 2007. Particulate and trace gas emissions from open burning of wheat straw and corn stover in China. Environ. Sci. Technol. 41, 6052e6058. http://dx.doi.org/10.1021/es0705137. Li, Y.F., 2014. Investigation of Agricultural Burning in the Field. Master Thesis. Peking University, Beijing, China. Liu, M., Song, Y., Yao, H., Kang, Y., Li, M., Huang, X., Hu, M., 2015. Estimating emissions from agricultural fires in the North China Plain based on MODIS fire radiative power. Atmos. Environ. 112, 326e334. http://dx.doi.org/10.1016/ j.atmosenv.2015.04.058. Marlier, M.E., DeFries, R.S., Voulgarakis, A., Kinney, P.L., Randerson, J.T., Shindell, D.T., Chen, Y., Faluvegi, G., 2013. El Nino and health risks from landscape fire emissions in southeast Asia. Nat. Clim. Change 3, 131e136. http://dx.doi.org/10.1038/ nclimate1658. Mishra, A.K., Shibata, T., 2012. Synergistic analyses of optical and microphysical properties of agricultural crop residue burning aerosols over the Indo-Gangetic Basin (IGB). Atmos. Environ. 57, 205e218. http://dx.doi.org/10.1016/ j.atmosenv.2012.04.025. National Bureau of Statistics of China, 1991e2014. China Statistical Yearbook. China Statistics Press. Permadi, D.A., Oanh, N.T.K., 2013. Assessment of biomass open burning emissions in Indonesia and potential climate forcing impact. Atmos. Environ. 78, 250e258. http://dx.doi.org/10.1016/j.atmosenv.2012.10.016. Plummer, S., 2006. Establishing a earth observation product service for the terrestrial carbon community: the globcarbon initiative. Mitig. Adapt. Strateg. Glob. Change 11, 97e111. http://dx.doi.org/10.1007/s11027-006-1012-8. Qin, Y., Xie, S.D., 2011. Historical estimation of carbonaceous aerosol emissions from biomass open burning in China for the period 1990e2005. Environ. Pollut. 159, 3316e3323. http://dx.doi.org/10.1016/j.envpol.2011.08.042. Randerson, J.T., Chen, Y., van der Werf, G.R., Rogers, B.M., Morton, D.C., 2012. Global

J. Li et al. / Atmospheric Environment 138 (2016) 152e161 burned area and biomass burning emissions from small fires. J. Geophys. Res. 117 http://dx.doi.org/10.1029/2012jg002128. Roy, D.P., Boschetti, L., Justice, C.O., Ju, J., 2008. The collection 5 MODIS burned area product - global evaluation by comparison with the MODIS active fire product. Remote Sens. Environ. 112, 3690e3707. http://dx.doi.org/10.1016/ j.rse.2008.05.013. Schroeder, W., Prins, E., Giglio, L., Csiszar, I., Schmidt, C., Morisette, J., Morton, D., 2008. Validation of GOES and MODIS active fire detection products using ASTER and ETMþ data. Remote Sens. Environ. 112, 2711e2726. http://dx.doi.org/ 10.1016/j.rse.2008.01.005. Seiler, W., Crutzen, P.J., 1980. Estimates of gross and net fluxes of carbon between the biosphere and the atmosphere from biomass burning. Clim. Change 2, 207e247. http://dx.doi.org/10.1007/bf00137988. Simon, M., Plummer, S., Fierens, F., Hoelzemann, J.J., Arino, O., 2004. Burnt area detection at global scale using ATSR-2: the GLOBSCAR products and their qualification. J. Geophys. Res. Atmos. 109 http://dx.doi.org/10.1029/ 2003jd003622. Streets, D.G., 2003. An inventory of gaseous and primary aerosol emissions in Asia in the year 2000. J. Geophys. Res. 108 http://dx.doi.org/10.1029/2002jd003093. Streets, D.G., Yarber, K.F., Woo, J.H., Carmichael, G.R., 2003. Biomass burning in Asia: annual and seasonal estimates and atmospheric emissions. Glob. Biogeochem. Cycles 17. http://dx.doi.org/10.1029/2003gb002040 n/a-n/a. Tian, H.Z., Zhao, D., Wang, Y., 2011. Emission inventories of atmospheric pollutants discharged from biomass burning in China. J. Environ. Sci. China 31 (2), 349e357. Turn, S.Q., Jenkins, B.M., Chow, J.C., Pritchett, L.C., Campbell, D., Cahill, T., Whalen, S.A., 1997. Elemental characterization of particulate matter emitted from biomass burning: wind tunnel derived source profiles for herbaceous and wood fuels. J. Geophys. Res. Atmos. 102, 3683e3699. http://dx.doi.org/10.1029/ 96jd02979. van der Werf, G.R., Randerson, J.T., Giglio, L., Collatz, G.J., Mu, M., Kasibhatla, P.S., Morton, D.C., DeFries, R.S., Jin, Y., van Leeuwen, T.T., 2010. Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997e2009). Atmos. Chem. Phys. 10, 11707e11735. http://dx.doi.org/

161

10.5194/acp-10-11707-2010, 2010. Wang, S.X., Zhang, C.Y., 2008. Spatial and temporal distribution of air pollutant emissions from open burning of crop residues in China (in Chinese). Sci. Pap. Online 05, 329e333. Wang, S.W., Zhang, S.H., 2009. Thinking of comprehensive utilization of crop straw (in Chinese). Anhui Agri. Sci. Bull 15 (5), 88e90. Wang, X.Y., Xue, S., Xie, G.H., 2012. Value-taking for residue factor as a parameter to assess the field residue of field crops (in Chinese). J. China Agric. Univ. 17 (1), 1e8. Wu, R., Bo, Y., Li, J., Li, L., Li, Y., Xie, S., 2016. Method to establish the emission inventory of anthropogenic volatile organic compounds in China and its application in the period 2008e2012. Atmos. Environ. 127, 244e254. http:// dx.doi.org/10.1016/j.atmosenv.2015.12.015. Yan, X., Ohara, T., Akimoto, H., 2006. Bottom-up estimate of biomass burning in mainland China. Atmos. Environ. 40, 5262e5273. http://dx.doi.org/10.1016/ j.atmosenv.2006.04.040. Yang, S., He, H., Lu, S., Chen, D., Zhu, J., 2008. Quantification of crop residue burning in the field and its influence on ambient air quality in Suqian, China. Atmos. Environ. 42, 1961e1969. http://dx.doi.org/10.1016/j.atmosenv.2007.12.007. de Zarate, I.O., Ezcurra, A., Lacaux, J.P., Van Dinh, P., de Argandona, J.D., 2005. Pollution by cereal waste burning in Spain. Atmos. Res. 73, 161e170. http:// dx.doi.org/10.1016/j.atmosres.2004.07.006. Zhang, H., Ye, X., Cheng, T., Chen, J., Yang, X., Wang, L., Zhang, R., 2008. A laboratory study of agricultural crop residue combustion in China: emission factors and emission inventory. Atmos. Environ. 42, 8432e8441. http://dx.doi.org/10.1016/ j.atmosenv.2008.08.015. Zhang, Y., Shao, M., Lin, Y., Luan, S., Mao, N., Chen, W., Wang, M., 2013. Emission inventory of carbonaceous pollutants from biomass burning in the Pearl River Delta Region, China. Atmos. Environ. 76, 189e199. http://dx.doi.org/10.1016/ j.atmosenv.2012.05.055. Zhao, Y., Nielsen, C.P., Lei, Y., McElroy, M.B., Hao, J., 2011. Quantifying the uncertainties of a bottom-up emission inventory of anthropogenic atmospheric pollutants in China. Atmos. Chem. Phys. 11, 2295e2308. http://dx.doi.org/ 10.5194/acp-11-2295-2011.