Influence of Straw Burning on Urban Air Pollutant Concentrations in Northeast China

Home , Baicheng, Hegang, Jiamusi, Qitaihe, Shuangyashan, Suihua

International Journal of Environmental Research and Public Health

Article Inﬂuence of Straw Burning on Urban Air Pollutant Concentrations in Northeast China

Zhenzhen Wang 1, Jianjun Zhao 1,2,3,* , Jiawen Xu 1, Mingrui Jia 1, Han Li 1 and Shijun Wang 1,2,*

1 School of Geographical Sciences, Northeast Normal University, Changchun 130024, China; [email protected] (Z.W.); [email protected] (J.X.); [email protected] (M.J.); [email protected] (H.L.) 2 Key Laboratory of Geographical Processes and Ecological Security in Changbai Mountains, Ministry of Education, Northeast Normal University, Changchun 130024, China 3 Urban Remote Sensing Application Innovation Center, Northeast Normal University, Changchun 130024, China * Correspondence: [email protected] (J.Z); [email protected] (S.W.); Tel.: +86-0431-8509-9550 (J.Z.)

Received: 3 April 2019; Accepted: 16 April 2019; Published: 17 April 2019

Abstract: Northeast China is China’s primary grain production base. A large amount of crop straw is incinerated every spring and autumn, which greatly impacts air quality. To study the degree of influence of straw burning on urban pollutant concentrations, this study used The Moderate-Resolution Imaging Spectroradiometer/Terra Thermal Anomalies & Fire Daily L3 Global 1 km V006 (MOD14A1) and The Moderate-Resolution Imaging Spectroradiometer/Aqua Thermal Anomalies and Fire Daily L3 Global 1 km V006 (MYD14A1) data from 2015 to 2017 to extract fire spot data on arable land burning and to study the spatial distribution characteristics of straw burning on urban pollutant concentrations, temporal variation characteristics and impact thresholds. The results show that straw burning in Northeast China is concentrated in spring and autumn; the seasonal spatial distributions of PM2.5, PM10 andAir Quality Index (AQI) in 41 cities or regions in Northeast China correspond to the seasonal variation of fire spots; and pollutants appear in the peak periods of fire spots. In areas where the concentration coefficient of rice or corn is greater than 1, the number of fire spots has a strong correlation with the urban pollution index. The correlation coefficient R between the number of burned fire spots and the pollutant concentration has a certain relationship with the urban distribution. Cities are aggregated in geospatial space with different R values.

Keywords: pollutant concentration; straw burning; ﬁre spots; PM2.5; PM10; AQI

1. Introduction As a large agricultural country, China’s straw crop output reached 700 million tons in 2013 [1]. Northeast China is China’s main grain production area. In 2015, the straw production in Northeast China was approximately 159 million tons, accounting for 19.2% of the country’s total straw production [2]. Pollutants produced by biomass burning have become a major source of air pollution [3–5], and straw open burning is an important contributor. Straw burning has seasonal and cyclical changes [6]. Increasing evidence shows that straw burning has become an important factor affecting urban air quality [7]. Straw burning strongly affects the atmospheric environment quality, resulting in a significant increase in the total amount of suspended particulate matter in the air, and this burning produces a large amount of harmful gases, such as PM2.5 and PM10 [7,8]. The pollutants produced by straw open burning have a long atmospheric residence time and are one of the major pollution sources in this region and the world [9]. In addition, open-air straw burning causes a series of traffic and safety

Int. J. Environ. Res. Public Health 2019, 16, 1379; doi:10.3390/ijerph16081379 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2019, 16, 1379 2 of 21 problems [6]. In recent years, the pollutant weather in the spring and autumn in the northeastern area has been frequent, with a wide range of influences, which has caused widespread concern [10]. In recent years, scholars have conducted studies on the weather caused by straw burning and have obtained a series of research results. Some scholars found that open-air biomass burning significantly impacts urban air quality by monitoring biomass produced by forests and farmland biomass burning in southeastern Queensland, Australia [7]. Under high relative humidity and south winds, the pollutant emissions from wheat straw burning in northern China and the high pollutant emissions from cities and industries have led to increased pollution in the North Plain [11]. Open burning of crop straw has important environmental impacts including reducing visibility, worsening air quality (sometimes severe smog pollution) and endangering public health [12]. A severe atmospheric haze event in Changchun city was caused by strong source emissions (such as biomass burning) and unfavourable air diffusion conditions. Among the five emission factors in Changchun, biomass burning accounted for 20% [13]. Satellite technology has also been used in national straw burning monitoring [14]. The results show that the change trend of fire spots in the buffer zone of some cities in the 700–800 km range is in good agreement with the air pollution index, and the correlation coefficient is 0.54 [14]. Traditional straw burning monitoring methods (such as point-by-point manual inspection) have the disadvantages of low efficiency, low coverage, and high costs. Satellite remote sensing has enabled quick and extensive burning monitoring due to its timeliness, wide coverage and high resolution [15]. The Moderate-Resolution Imaging Spectroradiometer (MODIS) is an advanced multi-spectral remote sensing sensor with 36 observation channels covering the main observations of the current major remote sensing satellites. Among these channels, MODIS/Terra Thermal Anomalies & Fire Daily L3 Global 1 km V006 (MOD14A1) and MODIS/Aqua Thermal Anomalies and Fire Daily L3 Global 1 km V006 (MYD14A1) thermal anomaly product data can be used directly, which is an ideal data source for monitoring straw burning. Based on MOD14A1 and MYD14A1 thermal anomaly product data, surface cover change data and pollutant monitoring site data in Northeast China, this study explores the distribution characteristics and relationship of straw burning with urban pollutant concentrations in time and space, conducts in-depth research on rice and corn concentrated planting areas, and performs a quantitative analysis of typical cities. The research results can be used to implement policies prohibiting straw burning by government departments at all levels, to establish a limited burning zone for straw burning zones, to generate scientific support for controlling straw burning, and to aid in monitoring and early warning of pollutant disasters. The factors that affect the serious pollution in the spring and autumn seasons in Northeast China have been determined to aid the development of straw harmless treatment technology at the source, to contribute to maintaining people’s health and to improve the ecological environment. The research objectives of this paper are as follows: (1) Explore the temporal and spatial distribution of straw burning fire spots and 41 city or regional pollutant concentrations in Northeast China. The relationship between straw burning and pollutant concentrations has been explored on temporal and spatial scales; (2) study the degree of influence of straw burning on pollutants in urban centres at different radius buffers and explore the maximum threshold value of straw burning on urban pollutant concentrations; (3) explore the relationship between rice- and corn-concentrated planting areas and fire spots and pollutants, and quantitatively study the extent of the impact of fire spots around typical cities in Northeast China on the city pollutants.

2. Materials and Methods

2.1. Study Area The northeastern region includes Heilongjiang Province, Jilin Province, Liaoning Province and Hulunbeier, Wulanhaote, Tongliao and Chifeng cities in the eastern part of the Inner Mongolia Autonomous Region, located at 115◦050–135◦020E, 38◦400–53◦340N[16] (Figure1). The northeastern region has a continental monsoon climate, with mild and humid summers and long cold winters [16]. Int. J. Environ. Res. Public Health 2019, 16, 1379 3 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 3 of 21

TheThe annual annual average average temperature temperature is 2.75–5.72 is 2.75–5.72 °C,◦ thC,e theaverage average annual annual precipitation precipitation is 250–700 is 250–700 mm, the mm, accumulatedthe accumulated temperature temperature above above 10 °C is 10 1500–3700◦C is 1500–3700 °C, and◦ C,the and frost-free the frost-free period is period 90–160 is days. 90–160 Corn, days. soybean,Corn, soybean, rice and rice spring andspring wheat wheat are grown are grown [16]. [Th16].e Theyield yield of crop of crop straw straw and and the the density density of of straw straw resourcesresources are are high high [17]. [17]. The The main main straw-burning straw-burning products products in in Northeast Northeast China China are are corn corn and and rice. rice.

Figure 1. Map of the study area. Figure 1. Map of the study area. 2.2. Data 2.2. DataThis study used MOD14A1 and MYD14A1 data for 2015–2017 with a spatial resolution of 1 km to extractThis study straw-burned used MOD14A1 fire spots. and The MYD14A1 land cover data type for data2015–2017 were derivedwith a spatial from the resolution surface coverageof 1 km tochange extract data straw-burned of MCD12Q1 fire (Land spots. Cover The land Type cover Yearly type L3 Globaldata were 500 mderived SIN Grid) from with the asurface spatial coverage resolution changeof 500 m,data which of MCD12Q1 was used to(Land extract Cover agricultural Type Yearly land L3 in theGlobal study 500 area. m SIN MOD14A1 Grid) with and MYD14A1a spatial resolutionproducts areof 500 based m, onwhich the 3.9wasµ mused and to 11 extractµm bands agricultural of MODIS land sensors in the on study the Terra area. satellite MOD14A1 and Aquaand MYD14A1satellite, respectively. products are By based setting on the the 3.9 thresholds μm and for11 μ firem bands spots andof MODIS background sensors temperature on the Terra di satellitefference, andthe Aqua MODIS satellite, research respectively. team extracts By thesetting temperature the thresholds anomaly for fire area spots and fireand spot background data for daytemperature and night. difference,The following the MODIS table is theresearch metadata team description extracts the of temperature the FireMask anomaly data set ofarea MOD14A1 and fire andspot MYD14A1 data for dayproducts. and night. Table The1 shows following that thetable pixel is the values metadata 7, 8 and description 9 represent of the the fire FireMask spot data data with set low of MOD14A1 confidence, andnominal MYD14A1 confidence products. and Table high confidence, 1 shows that respectively, the pixel values and these 7, 8 valuesand 9 represent are all identified the fire asspot fire data spot withdata low in this confidence, study. nominal confidence and high confidence, respectively, and these values are all identified as fire spot data in this study.

Int. J. Environ. Res. Public Health 2019, 16, 1379 4 of 21

Table 1. MOD14A1 and MYD14A1 FireMask data value description.

Value Description 0 Not processed (missing input data) 1 Not processed (obsolete; not used since Collection 1) 2 Not processed (other reason) 3 Non-fire water pixel 4 Cloud (land or water) 5 Non-fire land pixel 6 Unknown (land or water) 7 Fire (low confidence, land or water) 8 Fire (nominal confidence, land or water) 9 Fire (high confidence, land or water)

The pollutant data were derived from the China Air Quality Online Monitoring and Analysis Platform [18]. This research mainly used Air Quality Index (AQI), PM2.5 and PM10 data. Pollutant data for 41 cities in the study area were selected from March 2015 to March 2017.

2.3. Methods

2.3.1. Extraction and Treatment of Burning Straw Data In the MODIS thermal anomaly and fire product data, the values of “7, 8, 9” are fire spots; they were reassigned as 1, and the other values are non-fire spots, which were assigned a value of 0. The MOD14A1 and MYD14A1 data from the same period were combined, and the obtained data only contains the fire spot and non-fire spot data. The crop area in the MCD12Q1 land cover type data at 500-m resolution was used as a mask to extract the fire spots in the agricultural land to determine the number of fire spots of crop straw burning in the northeastern region for each day. The spring, summer, autumn and winter seasons in this study were divided according to “March, April, May”, “June, July, August”, “September, October, November”, and “December, January (the next year), February (the next year)”. According to the seasonal superposition of the fire spots in 2015 and 2016, fire spot spatial distribution maps of the four seasons of spring, summer, autumn and winter were obtained, and the annual data were superimposed to obtain the annual fire spot distribution map.

2.3.2. Contaminant Data Acquisition and Processing

The daily monitoring data of pollutants, such as PM2.5, PM10 and AQI, in 41 cities or regions in Northeast China were processed to obtain the AQI, PM2.5 and PM10 in the four seasons and the whole year for all cities in 2015 and 2016. The average value was interpolated in the Kriging space to obtain a spatial distribution map of the seasonal and annual air pollution concentrations in Northeast China.

2.3.3. Correlation Analysis We counted the number of fire spots in different buffer zones in different cities, summed the fire spots and urban pollutant data in the buffer every 8 days, and then conducted correlation analysis to obtain the correlation coefficient between the fire spots in different buffers and the pollutants in the central city. A two-dimensional map of the correlation coefficients between the fire spots in different buffers and the pollutants in the central city was obtained.

2.3.4. Selecting Typical Cities for Quantitative Analysis To analyse the quantitative relationship between fire spots and pollutants, some typical cities were selected for analysis. The selection of typical cities considered three conditions: (1) From March 2015 to March 2017, the number of straw burning points in a 150-km radius was ranked as the top 16 (number of fire points greater than 1700); (2) PM2.5, PM10, AQI and the correlation coefficient of the number of burned fire spots in different buffer zones were evaluated, and cities with R values greater than 0.5 for Int. J. Environ. Res. Public Health 2019, 16, 1379 5 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 5 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 5 of 21 twolarge or morepopulation, pollutants industrial were selected; development (3) considering and many that thevehicles, provincial the capitalprovincial has a capitals large population, were also large population, industrial development and many vehicles, the provincial capitals were also industrialconsidered development typical cities. and manyFinally, vehicles, eight thetypical provincial cities capitalsincluding were Harbin, also considered Changchun, typical Shenyang, cities. considered typical cities. Finally, eight typical cities including Harbin, Changchun, Shenyang, Finally,Songyuan, eight typicalShuangyashan, cities including Jiamusi, Harbin, Qitaihe Changchun, and Suihua Shenyang, were selected. Songyuan, Shuangyashan, Jiamusi, Songyuan, Shuangyashan, Jiamusi, Qitaihe and Suihua were selected. Qitaihe and Suihua were selected. 3. Results 3. Results Results 3.1. Analysis of Straw Fire Spots 3.1. Analysis Analysis of of Straw Straw Fire Fire Spots Spots 3.1.1. Analysis of the Temporal Distribution of Straw Fire Spots 3.1.1. Analysis Analysis of of the the Temporal Temporal Distribution of Straw Fire Spots According to the situation of farmers burning straw in the northeastern region, the straw obtainedAccording in the toto autumn thethe situation situation harvest of farmersofwas farmers burned burning burningin the straw autumn straw in the ofin northeastern that the year northeastern and region, the spring theregion, straw of the the obtained following straw obtainedinyear. the autumn Here, in the a harvest autumnstraw burning was harvest burned year was in is burned the defined autumn in from the of autumn thatSeptember year of and that 1 theof year the spring and current the of the spring year following to of August the year.following 31 Here, of the year.a strawfollowing Here, burning a year. straw year According burning is defined year to our fromis defined statistics, September from the September 1 tota of thel number current 1 of theof year fire current to points August year in to 31the August of burning the following 31 yearsof the of followingyear.2013, According 2014, year. 2015 According to and our statistics,2016 to was our the71,599, statistics, total 122,137, number the tota123,131 of firel number points and 101,588, inof thefire burning pointsrespectively. in years the of burningFrom 2013, 2013 2014, years to 2015 2016, of 2013,andthe 2016 graph2014, was 2015 of 71,599,the and number 2016 122,137, wasof fire 123,13171,599, spots 122,137,andin Northeast 101,588, 123,131 respectively.China and (Figure 101,588, From 2) respectively.shows 2013 that to 2016, within From the a 2013 graphstraw to burning of2016, the thenumberyear, graph the of of firenumber the spots number of in fire Northeast of spots fire spots was China inhighest Northeast (Figure in 2spri) showsChinang and that(Figure autumn. within 2) shows aThe straw number that burning within of fire year, a straw spots the numberburning generally year,ofincreased fire the spots number first was highestand of fire then spots in decreased, spring was andhighest which autumn. in was spri The consisteng numberand autumn.nt with of fire the The spots change number generally of ofthe fire increasedcereal spots planting generally first and area increasedthenin Northeast decreased, first and China. which then From was decreased, consistent 2013 to which 2016, with wasthe the cerealconsiste change planting ofnt thewith cereal areathe changein planting Northeast of areathe Chinacereal in Northeast plantingpeaked China. in area 2015 inFrom andNortheast 2013then togradually China. 2016, theFrom decreased cereal 2013 planting to(Figure 2016, area3).the cereal in Northeast planting China area peakedin Northeast in 2015 China and peaked then gradually in 2015 anddecreased then gradually (Figure3). decreased (Figure 3).

Figure 2. Straw fire spots in Northeast China from 2013 to 2016. FigureFigure 2. Straw 2. Straw fire ﬁre spots spots in in Northeast Northeast ChinaChina from from 2013 2013 to 2016.to 2016.

Figure 3. Change in the total number of ﬁre spots and cereal planting area in Northeast China from 2013Figure to 2016. 3. Change in the total number of fire spots and cereal planting area in Northeast FigureChina 3. from Change 2013 in to the2016. total number of fire spots and cereal planting area in Northeast China from 2013 to 2016.

Int. J. Environ. Res. Public Health 2019, 16, 1379 6 of 21

3.1.2.Int. Spatial J. Environ. Distribution Res. Public Health of 2019 Straw, 16, x Fire Spots 6 of 21 The fire spots in 2015 and 2016 were summed to obtain a fire spot distribution map for 2 years. 3.1.2. Spatial Distribution of Straw Fire Spots We recorded the frequency of fire occurrences in 2 years and used nuclear density analysis to obtain a frequencyThe map fire of spots the firein 2015 spots. and The2016 combination were summed of to the obtain two a figuresfire spot resulted distribution in Figure map for4. 2 In years. Figure 4, the fireWe spotsrecorded of strawthe frequency burning of arefire distributedoccurrences in along 2 years a largeand used area nuclear of cultivated density analysis land. The to obtain straw fire a frequency map of the fire spots. The combination of the two figures resulted in Figure 4. In Figure spots are concentrated in the southwestern (Harbin, Daqing, Qiqihar, and Suihua) and northeastern 4, the fire spots of straw burning are distributed along a large area of cultivated land. The straw fire (Shuangyashan, Jiamusi, and Hegang) parts of Heilongjiang Province, northwestern (Baicheng) and spots are concentrated in the southwestern (Harbin, Daqing, Qiqihar, and Suihua) and northeastern central(Shuangyashan, (Changchun Jiamusi, and Songyuan) and Hegang) Jilin parts Province, of Heilongjiang and central Province, Liaoning northwestern Province (Shenyang, (Baicheng) and Anshan, Liaoyang,central and (Changchun Panjin). and Among Songyuan) these Jilin areas, Province, the frequency and central of Liaoning fire spots Province in the (Shenyang, southwestern Anshan, part of HeilongjiangLiaoyang, Province and Panjin). and Among the northeastern these areas, part the of freque Liaoningncy of Province fire spots is in higher the southwestern than that of part other of areas, followedHeilongjiang by the central Province and and western the northeastern parts of Jilin part Province. of Liaoning Province is higher than that of other Theareas, fire followed points by in the central spring, and summer, western autumn parts of and Jilin winterProvince. seasons of 2015 and 2016 were summed to obtain theThe fire fire spot points distribution in the spring, maps summer, in different autumn seasons and (Figure winter5 ).seasons We recorded of 2015 theand frequency 2016 were of fire occurrencessummed in to di obtainfferent the seasons fire spot for distribution 2 years and maps used in nuclear different density seasons analysis (Figure 5). to obtainWe recorded the frequency the frequency of fire occurrences in different seasons for 2 years and used nuclear density analysis to of occurrence of fire spots (Figure5). In Figure5, the number of fire points is the highest in spring obtain the frequency of occurrence of fire spots (Figure 5). In Figure 5, the number of fire points is the and autumn and the lowest in summer and winter. The fire spots in spring are mainly distributed in highest in spring and autumn and the lowest in summer and winter. The fire spots in spring are Harbin,mainly Daqing, distributed Qiqihar in andHarbin, Suihua Daqing, in southwestern Qiqihar and Suihua Heilongjiang in southwestern Province. Heilongjiang The fire spots Province. in autumn are mainlyThe fire distributed spots in inautumn Shuangyashan, are mainly Jiamusi distribu andted Hegangin Shuangyashan, in northeastern Jiamusi Heilongjiang and Hegang Province; in Baichengnortheastern in northwestern; Heilongjiang Changchun Province; Baicheng and Songyuan in northwestern; in central Changchun Jilin Province; and andSongyuan Shenyang, in central Anshan, LiaoyangJilin Province; and Panjin and inShenyang, central LiaoningAnshan, Liaoyang Province. and The Panjin fire in spots central in winterLiaoning are Province. mainly The distributed fire in Liaoningspots in Province,winter are whichmainly isdistributed related toin theLiaoning low latitudeProvince, of which Liaoning is related Province. to the low This latitude low latitude of causesLiaoning the snow Province. cover timeThis low in Liaoning latitude causes Province the tosnow be shortercover time in winter,in Liaoning which Province leads to be straw shorter burning in winter.in winter, which leads to straw burning in winter.

Figure 4. SpatialSpatial distribution distribution map of fire ﬁre points from 2015 to 2016.

Int. J. Environ. Res. Public Health 2019, 16, 1379 7 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 7 of 21

FigureFigure 5. DistributionDistribution of of fire ﬁre spots spots in in di fferentﬀerent seasons for 2015 and 2016.

3.2. Analysis3.2. Analysis of Pollutants of Pollutants

3.2.1.3.2.1. Analysis Analysis of Pollutantsof Pollutants on on a a Temporal Temporal Scale Scale

ThreeThree parameters parameters of of pollutants, pollutants, AQI, AQI, PMPM2.52.5 andand PM PM1010, from, from March March 2015 2015 to March to March 2017 2017were were selected. Figure 6 shows that, in the 41 cities or regions in Northeast China, PM2.5, PM10 and AQI selected. Figure6 shows that, in the 41 cities or regions in Northeast China, PM 2.5, PM10 and AQI peaked in autumn and spring in 2015 and 2016. The 2015 and 2016 air quality parameters (AQI, PM2.5 peaked in autumn and spring in 2015 and 2016. The 2015 and 2016 air quality parameters (AQI, PM2.5 and PM10) show that the air pollution in most cities in 2015 was the most serious in early November, and PM ) show that the air pollution in most cities in 2015 was the most serious in early November, followed10 by December; the air pollution in most cities in 2016 was the most serious in December, followed by December; the air pollution in most cities in 2016 was the most serious in December, followed by the first half of January in the following year. There are obvious time series differences followedbetween by the2015 first and half 2016. of Figure January 2 shows in the that following the number year. of Therefire spots are was obvious highest time in autumn series di 2015.fferences between 2015 and 2016. Figure2 shows that the number of fire spots was highest in autumn 2015. The obvious pollutant concentration caused by straw burning in autumn may cause a serious air quality decline in November. Int. J. Environ. Res. Public Health 2019, 16, x 8 of 21

TheInt. J.obvious Environ. Res. pollutant Public Health concentration2019, 16, 1379 caused by straw burning in autumn may cause a serious 8 ofair 21 quality decline in November.

FigureFigure 6.6. PollutantPollutant concentration concentration changes changes in 41 in cities 41 cities or regions or regions in Northeast in Northeast China (x-axis China is the(x-axis time axis; y-axis is the city code; and z-axis is the PM , PM , and AQI levels). is the time axis; y-axis is the city code; and z-axis2.5 is10 the PM2.5, PM10, and AQI levels). The 41 cities or regions in Northeast China are coded in the order of north to south and west to The 41 cities or regions in Northeast China are coded in the order of north to south and west to east (Table2). east (Table 2). Table 2. Codes for 41 cities or regions in Northeast China. Table 2. Codes for 41 cities or regions in Northeast China. City Code City Code City Code City Code City Code City Code Heihe 1 Baicheng 15 Fuxin 29 Heihe 1 Baicheng 15 Fuxin 29 Daxinganlingdiqu 2 Songyuan # 16 Chifeng 30 DaxinganlingdiquHulunbeier 2 3 Songyuan Mudanjiang # 16 17 Chaoyang Chifeng 31 30 HulunbeierYIchun 3 4Mudanjiang Yanji 17 18 BenxiChaoyang 32 31 YIchun Hegang4 5Yanji Jilin 18 19 LiaoyangBenxi 33 32 Jiamusi # 6 Changchun # 20 Anshan 34 Hegang Shuangyashan5 # 7Jilin Tongliao 19 21 PanjinLiaoyang 35 33 Jiamusi # Qitaihe # 6 8 Changchun Siping # # 2022 Jinzhou Anshan 36 34 ShuangyashanJixi # 7 9 Tongliao Liaoyuan 21 23 Huludao Panjin 37 35 Qitaihe # 8 Siping # 22 Jinzhou 36

Int. J. Environ. Res. Public Health 2019, 16, 1379 9 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 9 of 21

Table 2. Cont. Jixi 9 Liaoyuan 23 Huludao 37 Suihua City10 CodeBaishan City24 Code CityYingkou Code38 Harbin # Suihua11 10 Tonghua Baishan 25 24 YingkouDandong 38 39 Harbin # 11 Tonghua 25 Dandong 39 Daqing 12 Tieling 26 Wafangdian 40 Daqing 12 Tieling 26 Wafangdian 40 Qiqihaer Qiqihaer13 13Fushun Fushun27 27 DalianDalian 41 41 WulanhaotWulanhaotee 14 14 Shenyang Shenyang# is# a typical # city.28 28 # is a typical city. 3.2.2. Analysis of Pollutants on a Spatial Scale 3.2.2. Analysis of Pollutants on a Spatial Scale The spatial distribution map of PM2.5, PM10 and AQI (Figure 7) shows that the air pollution concentrationThe spatial in distributionthe four seasons map in of Northeast PM2.5, PM China10 and is AQIquite (Figure different;7) shows winter that and the spring air pollution are more serious,concentration summer in is the good, four and seasons autumn in Northeast shows differ Chinaences. is quiteThe pollutant different; concentration winter and spring in autumn are more 2015 wasserious, significantly summer is higher good, andthan autumn that in shows 2016. diThefferences. spatial The distribution pollutant concentrationmap of the annual in autumn average 2015 waspollutants significantly shows higher high than pollutant that in 2016.concentrations The spatial distributionin Harbin, mapChangchun of the annual and average Shenyang. pollutants The concentrationshows high pollutant of air pollutants concentrations in urban in areas Harbin, with Changchun the three capital and Shenyang. cities of Shenyang–Changchun– The concentration of Harbinair pollutants as the axis in urban is significantly areas with higher the three than capitalthat in citiesother cities. of Shenyang–Changchun–Harbin Harbin, Changchun and Shenyang as the showaxis is a significantly clear diffusion higher trend. than that in other cities. Harbin, Changchun and Shenyang show a clear diffusionThe fire trend. spot distribution map (Figure 2) and the spatial distribution maps of PM2.5, PM10 and AQI Thein different fire spot distributionseasons and map in (Figure 2015–20162) and (Figur the spatiale 7) distributionshow that mapsthe pollutant of PM 2.5, PMconcentration10 and AQI correspondingin different seasons to spring and in and 2015–2016 autumn (Figure concentrations7) show that in the the fire pollutant spots is concentration significantly corresponding higher than that to inspring summer and autumnbut lower concentrations than the pollutant in the fireconcentration spots is significantly in winter. higher The PM than2.5, thatPM10 in and summer AQI butpollutant lower concentrationsthan the pollutant in the concentration autumn of 2016 in winter. were Theall lowe PM2.5r than, PM that10 and in AQIthe autumn pollutant of concentrations2015 and lower in than the thatautumn in the of 2016spring were of 2016. all lower The thannumber that of in fire the autumnspots in of2016 2015 was and significantly lower than thatlower in thethan spring that in of 2015, 2016. Theindicating number that of firethe spots number in 2016 of wasfire significantlyspots in the lower autumn than of that 2015 in 2015,led to indicating an increased that the pollutant number concentrationof fire spots in in the the autumn autumn of 2015of 2015, led and to an the increased decreased pollutant number concentration of straw-burning in the fire autumn spots of in 2015, the autumnand the decreasedof 2016 made number the ofconcentrations straw-burning of firePM spots2.5, PM in10 theand autumn AQI in of the 2016 autumn made of the that concentrations year lower thanof PM the2.5, pollutant PM10 and concentrations AQI in the autumn in the of same that yearperiod lower in 2015. than the pollutant concentrations in the same period in 2015.

(a)

Figure 7. Cont. Int. J. Environ. Res. Public Health 2019, 16, 1379 10 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 10 of 21

(b)

(c)

Figure 7.7. ((aa)) PMPM2.52.5,(, (bb)) PMPM1010 andand ((cc)) AQI AQI spatial spatial interpolation interpolation maps maps for for diﬀ erentdifferent seasons seasons in 2015 in and2015 2016. and 2016.

3.3. Relationship between the Number of Fire Spots and The Pollutant Concentration in Different City 3.3. Relationship between the Number of Fire Spots and The Pollutant Concentration in Different City Buffer Buffer Zones Zones In the range of 10–300 km in 41 cities or regions in Northeast China, the cumulative radius of In the range of 10–300 km in 41 cities or regions in Northeast China, the cumulative radius of 10 10 km was used to study the correlation between the number of fire spots in different radius buffers km was used to study the correlation between the number of fire spots in different radius buffers and and the PM2.5, PM10 and AQI in the central city to determine the impact of straw burning on urban the PM2.5, PM10 and AQI in the central city to determine the impact of straw burning on urban pollutants. We found that the impact of straw burning on urban pollutants has much to do with the pollutants. We found that the impact of straw burning on urban pollutants has much to do with the geographical distribution of cities. geographical distribution of cities.

Int. J. Environ. Res. Public Health 2019, 16, 1379 11 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 11 of 21

(a)

(b)

Figure 8. Cont.

Int. J. Environ. Res. Public Health 2019, 16, 1379 12 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 12 of 21

(c)

FigureFigure 8. Two-dimensional8. Two-dimensional map map of the of correlationthe correlation between between the number the number of ﬁre of spots fire andspots the and (a) PMthe 2.5, (b)(a PM) PM10 2.5and, (b ()c )PM AQI10 and in the (c) 10–300 AQI in km the radius 10–300 bu kmﬀer radius zone of buffer 41 cities zone or of regions 41 cities in Northeast or regions China in (theNortheast abscissa isChina the radius; (the abscissa the ordinate is the is radi theus; city the number). ordinate is the city number).

WithinWithin the the results results of of the the radius radius cluster cluster analysis analysis (Figure(Figure8 8),), the the results results ofof thethe clustercluster analysisanalysis of the radiithe ofradii PM of2.5 PMand2.5 PMand10 PMare10 completelyare completely consistent. consistent. Among Among them, them, the the radii radii of of 10 10 km, km, 20 20 km km and and 30 km are30 each km are of theeach same of the type, same the type, radius the radius from 40from km 40 to km 60 to km 60 is km one is one type, type, the the radius radius of 70of 70 km–150 km– km is another150 km is type, another and type, the radiusand the of radius 160 km–300 of 160 km–300 km is anotherkm is another type. type. AQI clustersAQI clusters of radii of radii are dividedare intodivided five categories. into five categories. Among them,Among the them, radii the of radi 10 kmi of and10 km 20 and km 20 are km one are class one class each, each, the radiusthe from 30radius km to from 60 km 30 is km one to class, 60 km the is radiusone class, of 70the km–200 radius of km 70is km–200 one class, km andis one the class, radius and of the210 radius km–300 of km is210 one km–300 class. Thekm is e ffoneect class. of straw The effect burning of straw on pollutants burning on has pollutants a certain has relationship a certain relationship with the radius. with the radius. From the results of the city cluster analysis (Figure 8), PM2.5 and AQI have been From the results of the city cluster analysis (Figure8), PM 2.5 and AQI have been divided into four divided into four categories, and PM10 has been divided into three categories. The first category of R categories, and PM has been divided into three categories. The first category of R values in the three values in the three10 pollutant clustering results is large, and the second category has lower R values. pollutant clustering results is large, and the second category has lower R values. The third and fourth The third and fourth categories of PM2.5 and AQI and the third category of PM10 have the lowest R categories of PM2.5 and AQI and the third category of PM10 have the lowest R values. We selected values. We selected all cities that meet at least two of the three largest R values of PM2.5, PM10 and allAQI. cities These that meetcities atare least numbered two of from the three 7 to 11 largest and fromR values 17 to 20. of PMWe2.5 defined, PM10 themand AQI.as city These type 1. cities For are numbered from 7 to 11 and from 17 to 20. We defined them as city type 1. For example, city number 7 example, city number 7 is located in the category in which PM2.5 and AQI have the largest R value is locatedin the city in clustering the category result. in which Similarly, PM2.5 weand select AQIed haveall cities the that largest meetR atvalue least in two the of city the clustering lower R result. Similarly,values of we PM selected2.5, PM10, all and cities AQI. that These meet city at numbers least two are of 4, the 6, 12, lower 13, 15,R values 16, 22, of23, PM 25 2.5and, PM 26.10 We, and AQI. Thesedefined city these numbers numbers are 4, as 6, city 12, type 13, 15, 2. The 16, 22,remaining 23, 25 and cities 26. are We located defined in the these third numbers and fourth as city type 2. Thecategories remaining of PM cities2.5 and are AQI located and in inthe the thirdthird andcategory fourth of categoriesPM10. We defined of PM2.5 theseand cities AQI as and city in type the 3. third categoryDifferent of types PM10 of. We cities defined are geographically these cities asclustered city type (Figure 3. Di ff9).erent types of cities are geographically clustered (Figure9).

Int. J. Environ. Res. Public Health 2019, 16, 1379 13 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 13 of 21

FigureFigure 9. Distribution 9. Distribution of different of different types types of of cities. cities. 3.4. Analysis of the Relationship between Fire Spots and Pollutants in Concentrated Planting Areas of Corn 3.4.and Analysis Rice of the Relationship between Fire Spots and Pollutants in Concentrated Planting Areas of Corn and Rice The concentration coefficient indicates the ratio of the per capita output value (or output) of The concentration coefficient indicates the ratio of the per capita output value (or output) of a a certain regional industry to the corresponding per capita output value (or output) of the country. certain regional industry to the corresponding per capita output value (or output) of the country. If If the concentration coefficient is greater than 1, that crop type is concentrated in the region and the concentration coefficient is greater than 1, that crop type is concentrated in the region and the the degree of specialization is high [19]. Previous studies have indicated that the cities with a rice degree of specialization is high [19]. Previous studies have indicated that the cities with a rice concentration coefficient greater than 1 in the three northeastern provinces are: Jiamusi, Shuangyashan, concentration coefficient greater than 1 in the three northeastern provinces are: Jiamusi, Jixi, Hegang, Yichun, Harbin, and Baicheng; cities with a corn concentration coefficient greater than Shuangyashan, Jixi, Hegang, Yichun, Harbin, and Baicheng; cities with a corn concentration 1 include Jiamusi, Shuangyashan, Hegang, Harbin, Suihua, Qiqihar, Daqing, Baicheng, Songyuan, coefficient greater than 1 include Jiamusi, Shuangyashan, Hegang, Harbin, Suihua, Qiqihar, Daqing, Changchun, Siping, Liaoyuan, and Fuxin [19]. Using the 150-km radius as the threshold, a city is Baicheng, Songyuan, Changchun, Siping, Liaoyuan, and Fuxin [19]. Using the 150-km radius as the the centre within the radius of 150 km. The cities with more than 1700 fire spots from March 2015 to threshold, a city is the centre within the radius of 150 km. The cities with more than 1700 fire spots March 2016 were Daqing, Qiqihar, Harbin, Suihua, Songyuan, Shuangyashan, Shenyang, Liaoyang, from March 2015 to March 2016 were Daqing, Qiqihar, Harbin, Suihua, Songyuan, Shuangyashan, Anshan, Fushun, Tieling, Benxi, Jiamusi, and Qitaihe for a total of 16. Among these cities, eight have Shenyang, Liaoyang, Anshan, Fushun, Tieling, Benxi, Jiamusi, and Qitaihe for a total of 16. Among corn or rice concentration coefficient greater than 1. As a result of the correlation coefficient between these cities, eight have corn or rice concentration coefficient greater than 1. As a result of the the cities’ PM2.5, PM10, AQI and the number of fire points in different radii, there are 14 cities that meet correlation coefficient between the cities’ PM2.5, PM10, AQI and the number of fire points in different the condition that the correlation coefficient between at least two types of pollutants and the number radii, there are 14 cities that meet the condition that the correlation coefficient between at least two of fire spots is greater than 0.5. These cities are Qitaihe, Jixi, Mudanjiang, Shuangyashan, Baicheng types of pollutants and the number of fire spots is greater than 0.5. These cities are Qitaihe, Jixi, Suihua, Jilin, Changchun, Songyuan Yanji, Harbin, Jiamusi, Liaoyuan, Siping, Hegang Songyuan and Mudanjiang, Shuangyashan, Baicheng Suihua, Jilin, Changchun, Songyuan Yanji, Harbin, Jiamusi, Yichun. Among them, 10 have a rice or corn concentration coefficient greater than 1. This finding Liaoyuan, Siping, Hegang Songyuan and Yichun. Among them, 10 have a rice or corn concentration shows that, with more concentrated rice and corn planting around a city in the northeastern region, coefficient greater than 1. This finding shows that, with more concentrated rice and corn planting the more fire spots are generated by straw burning and the greater the impact on the central city. around a city in the northeastern region, the more fire spots are generated by straw burning and the greater the impact on the central city.

Int. J. Environ. Res. Public Health 2019, 16, 1379 14 of 21

Int. J. Environ. Res. Public Health 2019, 16, x 14 of 21 3.5. Typical City Analysis 3.5. Typical City Analysis 3.5.1. Changes in the Number of Fire Spots in Different Buffers of Typical Cities The3.5.1. change Changes in in the the number Number of of fire Fire spots Spots in in diDifferentfferent buBuffersffers of Typical a typical Cities city (Figure 10) shows that the numberThe of change fire spots in the in number spring of and fire spots autumn in differen is thet highest. buffers of For a typical the three city (Figure provincial 10) shows capital that cities, the numberthe number of fire of spotsfire spots is in in the spring following and autumn order: is Harbin the highest.> Changchun For the three> Shenyang;provincial capital considering cities, the typicalthe urban number distribution of fire spots latitude, is in the the following number or ofder: fires Harbin in the > Changchun north of a city> Shenyang; is significantly considering higher the than that intypical the south. urban Additionally,distribution latitude, for the the typical number cities of fir ines the in northeasternthe north of a partcity is of significantly Heilongjiang, higher such as Shuangyashan,than that in Jiamusi the south. and Additionally, Qitaihe, the for number the typica of firel cities points in the in northeastern autumn is much part of higher Heilongjiang, than that in such as Shuangyashan, Jiamusi and Qitaihe, the number of fire points in autumn is much higher than spring. The distribution of fire points over time in different radii of these cities is highly consistent. that in spring. The distribution of fire points over time in different radii of these cities is highly In a strawconsistent. burning In a year, straw except burning for year, in Suihua, except for the in number Suihua, of the typical number city of straw-burning typical city straw-burning points reached its highestpoints valuereached in its the highest autumn value of 2015.in the autumn of 2015.

FigureFigure 10. Variations 10. Variations in the in number the number of fire of spots fire spots in diff inerent different buffers buffers of a typical of a typical city (the city x-axis (the x- is the time axis,axis is the the y-axis time isaxis, the the bu ffy-axiser radius, is the and buffer the ra z-axisdius, isand the the number z-axis ofis the fire number spots). of fire spots).

Int. J. Environ. Res. Public Health 2019, 16, 1379 15 of 21

Int. J. Environ. Res. Public Health 2019, 16, x 15 of 21 3.5.2. Changes in the Typical Urban Pollution Index 3.5.2. Changes in the Typical Urban Pollution Index The 2015–2016 typical urban pollution index change chart (Figure 11) shows that, in 2015, the highestThe peak 2015–2016 of the pollution typical urban index pollution in each city index occurred change in earlychart November, (Figure 11) followed shows that, by peak in 2015, times the in earlyhighest and peak late of December. the pollution In addition index in toeach peaks city in occurred the ﬁrst in half early of JanuaryNovember, of the followed following by peak year, times there werein early also and smaller late December. peaks in theIn addition March–May to peaks period in the of first 2015; half in 2016,of January except of forthethe following peak times year, in there the Harbin,were also Songyuan smaller peaks and Suihuain the March–May pollution indexes, period whichof 2015; occurred in 2016, in except early for November, the peak thetimes peaks in the in otherHarbin, cities Songyuan were not and obvious. Suihua The pollution pollutant indexes, concentration which occurred was high in from early December November, to January,the peaks and in thereother werecities manywere not small obvious. peaks The from pollutant March–April concen 2016.tration The was peak high value from in December March 2016 to January, was slightly and higherthere were than many of the peaksmall in peaks March from 2015. March–April Notably, the 2016. pollutant The concentrationspeak value in inMarch Changchun, 2016 was Songyuan, slightly Shenyanghigher than and of Harbin the peak in March in March and April2015. 2016 Notabl werey, morethe pollutant serious than concentrations those in 2015; in the Changchun, diﬀerence fromSongyuan, December Shenyang 2015 wasand notHarbin large. in March and April 2016 were more serious than those in 2015; the difference from December 2015 was not large.

Figure 11. Cont.

Int. J. Environ. Res. Public Health 2019, 16, 1379 16 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 16 of 21

FigureFigure 11. 11.ChangesChanges in inthe the typical typical urban urban pollution index index from from 2015 2015 to 2016.to 2016. 3.5.3. Relationship between the Number of Fire Spots and The Pollutant Concentration in Different Bu3.5.3.ffer Relationship Zones of Typical between Cities the Number of Fire Spots and The Pollutant Concentration in Different Buffer Zones of Typical Cities Considering the combination of temporal and spatial factors, there are also peaks in the spring Considering the combination of temporal and spatial factors, there are also peaks in the spring and autumn corresponding to the number of typical city straw fire spots. The pollutants PM2.5, PM10, and autumn corresponding to the number of typical city straw fire spots. The pollutants PM2.5, PM10, and AQI and the number of fire spots in different radii of the typical cities, except for Shenyang, have a and AQI and the number of fire spots in different radii of the typical cities, except for Shenyang, have maximum correlation coefficient R of 0.7 or greater (p < 0.05), as shown in Figure 12. Figure 12 shows a maximum correlation coefficient R of 0.7 or greater (p < 0.05), as shown in Figure 12. Figure 12 shows the relationship between the number of fire spots and the pollutant concentration R in the different the relationship between the number of fire spots and the pollutant concentration R in the different buffer radii of a typical city as a function of the radius at p < 0.05. buffer radii of a typical city as a function of the radius at p < 0.05.

Int. J. Environ. Res. Public Health 2019, 16, 1379 17 of 21 Int. J. Environ. Res. Public Health 2019, 16, x 17 of 21

FigureFigure 12.12.Relationship Relationship between between the numberthe number of fire of spots fire andspots the and pollutant the pollutant concentration concentration in different buffer zones of typical cities. in different buffer zones of typical cities. Among these cities, from the change of the correlation coefficient R value with the radius, both HarbinAmong and Suihuathese cities, have from a trend the of change increasing of the first correlation and then decreasing,coefficient R which value haswith strong the radius, consistency. both HarbinChangchun and Suihua decreases have first a trend at 0–20 of increasing km and increases first and atthen 20 decreasing, km–50 km. which After has 50 km,strong the consistency. correlation Changchuncoefficient Rdecreasesvalue is first basically at 0–20 unchanged km and increa at approximatelyses at 20 km–50 0.7. km. The After correlation 50 km, coetheffi correlationcient R of coefficientShenyang showsR value a tendency is basically to increase unchanged slowly atwith approx theimately radius. The0.7. RThevalues correlation of Songyuan coefficient and Jiamusi R of Shenyangtend to be shows stable a after tendency a slow to increaseincrease withslowly the with radius. the radius. Shuangyashan The R values appears of Songyuan to increase and first Jiamusi and tendthen to decrease, be stable and after finally, a slow the increaseR value with remains the radius. basically Shuangyashan unchanged. appears The correlationto increase first coeffi andcient then of decrease,Qitaihe has and the finally, largest theR value, R value and remains the overall basically trend unchanged. tends to remain The unchangedcorrelation coefficient after it first of decreases. Qitaihe hasThe the correlation largest R coe value,fficient andR thevalues overall of Changchun, trend tends Songyuan, to remain Jiamusi,unchanged Shuangyashan after it first decreases. and Qitaihe The all correlationtend to remain coefficient basically R values unchanged of Changchun, as the radius Songyuan, becomes Jiamu larger.si, SinceShuangyashan some correlation and Qitaihe coeffi allcient tendR tovalues remain do basically not satisfy unchanged the condition as the of radiusp < 0.05, becomes there larger. is data Since missing some in thecorrelation fold line coefficient in Figure R12 values. do not satisfy the condition of p < 0.05, there is data missing in the fold line in Figure 12.

Int. J. Environ. Res. Public Health 2019, 16, 1379 18 of 21

3.5.4. Analysis of the Impact of Straw Burning on Urban Pollution in Typical Cities By analysing the correlation between the number of fire spots with different radii in typical cities and urban PM2.5, under the premise of p < 0.05, we fitted and analysed the curve of the correlation coefficient R with the radius to obtain the functional relationship between them. The function formula is shown in Table3. The curve of the correlation coe fficient R and the radius of Songyuan is well fitted, and there is an obvious positively correlated relationship. The fitting curve of the R values and radii of Harbin and Suihua shows that R is positively correlated with a radius in the range of 80 km and negatively correlated with those more than 80 km (Figure 12). This finding indicates that the effect of straw burning on PM2.5 has a positive effect in the range of 80 km, while in the range of more than 80 km, the effect of straw burning on PM2.5 is gradually weakened. The curve fitting of the R values and the radii of Changchun and Jiamusi shows that the R values are positively correlated with the radius in the range of 20 to 70 km and then are basically in a stable state, indicating that the level of PM2.5 in the range of 20 to 70 km is greatly affected by straw burning. The R value of Songyuan increased with the increase of the radius at 70–200 km, indicating that PM2.5 is greatly affected by straw burning in this range. The R values and radii of Shuangyashan and Qitaihe show a negative correlation with the radius, indicating that in the range of more than 20 km, the smaller the radius, the greater the impact of straw burning on PM2.5.

Table 3. Relationship between the number of fire spots in different buffers of a typical city and the

correlation coeﬃcient R of urban PM2.5 with the radii.

City Formula R-Square Haerbin y = 10 43 0.0066x2 + 0.1064x + 0.3202 0.9249 − − Changchun y = 7 10 7x5 6 10 5x4 + 0.0019x3 0.0271x2 + 0.1811x + 0.287 0.8284 × − − × − − Shengyang y = 6 10 5x3 0.0011x2 + 0.0063x + 0.2108 0.9706 × − − Songyuan y = 0.001x2 + 0.0409x + 0.1037 0.8987 − Shuangyashan y = 5 10 7x5 5 10 5x4 + 0.0015x3 0.0229x2 + 0.1435x + 0.4955 0.7987 × − − × − − Jiamusi y = 3 10 6x4 + 0.0002x3 0.0067x2 + 0.0778x + 0.3247 0.944 − × − − Qitaihe y = 10 5x3 + 0.0006x2 0.0112x + 0.9494 0.9066 − − − Suihua y = 2 10 4x3 0.0084x2 + 0.1158x + 0.265 0.9027 × − −

We selected the radius corresponding to the maximum R value obtained by the correlation analysis between the typical city PM2.5 and the number of fire spots in different radii and determined the fitting function, as shown in Table4. The number of fire spots in the seven typical cities is linear with the buffer radius and is positively correlated. Among these cities, the maximum R value was found in Qitaihe, and the lowest was observed in Songyuan. The buffer radius corresponding to the maximum R value radius of Qitaihe, Shuangyashan, Harbin and Suihua is within 100 km. The buffer radius corresponding to the maximum R value radius of Changchun, Jiamusi and Songyuan is within 100–200 km. The results of typical urban studies show that the radius of influence of straw burning on urban pollutant concentration differs in different cities. Varied policies can be adopted to control straw burning and pollutants according to the different characteristics of each city. The straw burning zone and the limited burning zone have great reference value and practical significance. Int. J. Environ. Res. Public Health 2019, 16, 1379 19 of 21

Table 4. The number of ﬁre points in the radius with the largest R value and the PM2.5 ﬁtting function.

City Maximum R Value/PM2.5 Corresponding Buﬀer Radius/km Formula Haerbin 0.844 80 y = 0.6299x + 227.92 Changchun 0.763 180 y = 0.3407x + 311.98 Shengyang 0.238 70 y = 0.3719x + 426.05 Songyuan 0.53 200 y = 0.0845x + 220.20 Shuangyashan 0.803 40 y = 0.5837x + 225.12 Jiamusi 0.627 120 y = 0.1169x + 172.79 Qitaihe 0.959 10 y = 40.266x + 179.97 Suihua 0.784 70 y = 0.8985x + 131.24

4. Discussion

4.1. Influence of Straw Burning on Urban Air Pollutant Concentrations The correlation coefficient between the number of fire spots in the 10–300 km radius buffer zone and the urban centre PM2.5, PM10 and AQI levels in 41 cities in Northeast China was calculated. We found that the results of the cluster analysis of the radii of PM2.5 and PM10 are completely consistent. Among them, the radii of 10 km, 20 km and 30 km are each of the same type; the radius of 40 km to 60 km is one type, the radius of 70 km–150 km is one type, and the radius of 160 km–300 km is one type. AQI clusters of radii are divided into five categories. Of these categories, the radii of 10 km and 20 km are each one class, the radius of 30 km to 60 km is one class, the radius of 70 km–200 km is one class, and the radius of 210 km–300 km is one class. The effect of straw burning on pollutants has a certain relationship with the radius, but there is no uniform radius threshold in different cities. This finding is inconsistent with the results of the good correlation between the change trend of fire spots in the buffer zone of 700 to 800 km in a city in previous studies and the air pollution index [14]. In the city cluster analysis, cities with large correlation coefficient R values are aggregated. Cities with smaller R values are distributed on the periphery of cities with the largest R values. The cities with the smallest R values are distributed at the outermost periphery. These cities are in a clustered distribution state with the R value.

4.2. Relationship between the Number of Fires and Pollutants in Rice and Corn Concentrated Planting Areas In cities with a concentration coefficient of rice and corn greater than 1 in the northeastern region, the number of straw burning fire spots in the same buffer zone is higher than that in other cities. In cities with a rice or corn concentration coefficient greater than 1 in Northeast China, the correlation coefficient between the number of fire points and PM2.5, PM10 and AQI in most cities is greater than 0.5 (p < 0.05). It is indicated that the more concentrated rice and corn are planted around the city in Northeast China, the greater the amount of straw, the more fire spots generated by rice or corn stover burning and the greater the impact of the generated pollutants on the urban pollutant concentration.

4.3. Relationship between the Number of Fire Spots and Pollutants in Typical Cities The quantitative analysis of the eight selected typical cities shows that the impact radius of straw burning varies in different cities. The radius of the buffer corresponding to the maximum R values in Qitaihe, Shuangyashan, Harbin and Suihua is within 100 km. The radius of the buffer corresponding to the maximum R values in Changchun, Jiamusi and Songyuan is within 100–200 km. The impact radius of straw burning on urban pollutants varies in different cities. Varied policies can be adopted to control straw burning and pollutants according to different city characteristics.

4.4. Other Factors Influencing Urban Air Pollutant Concentrations One of the main causes of urban pollutant concentrations is the pollution source. In addition to burning straw in Northeast China, there are some effects from automobile exhaust, winter heating (October–April of the following year) and factories. The effects of other sources of pollution were not Int. J. Environ. Res. Public Health 2019, 16, 1379 20 of 21 considered in this study. The capital cities in Northeast China are densely populated, and there are many heavy industries that produce a large amount of industrial exhaust and automobile exhaust. Meteorological conditions are also an important factor in changing pollutant concentrations in urban cities. In windy weather, pollutants are quickly blown away by the wind, reducing pollutant concentrations. Wang et al. studied the effects of wheat straw burning on regional aerosols in northern China. The results show that the average AOD and PM10 concentrations in the North Plain vary greatly with wind direction and humidity [11].

5. Conclusions The fire spots of straw burning in Northeast China are concentrated in spring and autumn. The fire spots in spring are mainly distributed in southwestern Heilongjiang Province. The fire spots in autumn are mainly distributed in northeastern Heilongjiang Province, northwestern and central Jilin Province, and central Liaoning Province. From 2013–2016, for the fire spots of straw burning in northeastern China in the northeastern region, the level of PM2.5, PM10 and AQI in 2015–2016 was higher than that in winter and peaked in spring and autumn. We speculate that straw burning promotes pollutant generation to some extent. We used 41 cities in Northeast China as the centres and made buffer zones with radii increasing every 10 km. The correlation coefficient R values of PM2.5, PM10 and AQI in the central cities and the number of fire spots in different buffers were calculated, respectively. According to the variation of the R values with the radii, 41 cities were clustered and analysed. We found that the cities were clustered and distributed in regions with different R values. Radius cluster analysis shows that adjacent radii are clustered into one class. In cities with a concentration factor of rice and corn planting greater than 1 in Northeast China, the correlation coefficient of the number of fire spots and pollutants is generally higher. The main innovations and contributions of this research are: A quantitative study on the temporal and spatial distribution characteristics of fire spots and pollutants in Northeast China that was carried out to provide a reference for studying the relationship between straw and pollutants. By clustering the relationship between pollutants and straw burning, the geospatial law of the effects of straw burning on pollutants has been revealed.

Author Contributions: Conceptualization, J.Z. and S.W.; methodology, Z.W., J.X. and M.J.; software, Z.W. and H.L.; validation, Z.W., M.J. and H.L.; formal analysis, J.Z.; investigation, Z.W., J.X.; resources, M.J. and H.L.; data curation, Z.W., J.X., M.J. and H.L.; writing—original draft preparation, Z.W., J.Z., J.X., M.J. and H.L.; writing—review and editing, J.Z., Z.W. and S.W.; visualization, Z.W. and J.X.; supervision, J.Z.; project administration, J.Z. and S.W.; funding acquisition, J.Z. Funding: This research was funded by the National Natural Science Foundation of China (Grant No. 41771450, 41630749, 41871330, 41571078, 41571405), the Fundamental Research Funds for the Central Universities (Grant No. 2412019BJ001), the National Key Research and Development Project (Grant No. 2016YFA0602301). the Foundation of the Education Department of Jilin Province in the 13th Five-Year project (Grant No. JJKH20190282KJ). Acknowledgments: We thank all editors and reviewers for their insightful and valuable comments, which greatly improved the quality of our study. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Li, H.; Wang, C.; Sun, H.; Yan, X.; Liang, Q. Comprehensive utilization and sustainable development of crop straw. Res. Agric. Mech. 2017, 39, 256–262. (In Chinese) 2. Wang, J.; Tang, H.; Wang, J. Current situation and development analysis of comprehensive utilization of crop straw resources in Northeast China. J. Agric. Mach. 2017, 48, 1–21. (In Chinese) 3. Permadi, D.A.; Oanh, N.T.K. Assessment of biomass open burning emissions in Indonesia and potential climate forcing impact. Atmos. Environ. 2013, 78, 250–258. [CrossRef] 4. Chen, Y.; Xie, S.-D. Characteristics and formation mechanism of a heavy air pollution episode caused by biomass burning in Chengdu, Southwest China. Sci. Total. Environ. 2014, 473, 507–517. [CrossRef][PubMed] Int. J. Environ. Res. Public Health 2019, 16, 1379 21 of 21

5. Sun, J.; Peng, H.; Chen, J.; Wang, X.; Wei, M.; Li, W.; Yang, L.; Zhang, Q.; Wang, W.; Mellouki, A. An

estimation of CO2 emission via agricultural crop residue open ﬁeld burning in China from 1996 to 2013. J. Clean. Prod. 2016, 112, 2625–2631. [CrossRef] 6. Guan, Y.; Chen, G.; Cheng, Z.; Yan, B.; Hou, L. Air pollutant emissions from straw open burning: A case study in Tianjin. Atmos. Environ. 2017, 171, 155–164. [CrossRef] 7. He, C.; Miljevic, B.; Crilley, L.R.; Surawski, N.C.; Bartsch, J.; Salimi, F.; Uhde, E.; Schnelle-Kreis, J.; Orasche, J.; Ristovski, Z.; et al. Characterisation of the impact of open biomass burning on urban air quality in Brisbane, Australia. Environ. Int. 2016, 91, 230–242. [CrossRef][PubMed] 8. Gadde, B.; Bonnet, S.; Menke, C.; Garivait, S. Air pollutant emissions from rice straw open ﬁeld burning in India, Thailand and the Philippines. Environ. Pollut. 2009, 157, 1554–1558. [CrossRef][PubMed]

9. Chen, G.; Guan, Y.; Tong, L.; Yan, B.; Hou, L. Spatial estimation of PM2.5 emissions from straw open burning in Tianjin from 2001 to 2012. Atmos. Environ. 2015, 122, 705–712. [CrossRef] 10. Satellite Monitoring by the Ministry of Environmental Protection Shows that 93 Percent of the Country’s Straw Burning Fire Spots Were in Heilongjiang Province last Week. Available online: https://www.thepaper. cn/newsDetail_forward_1399237 (accessed on 20 November 2015). 11. Wang, L.; Xin, J.; Li, X.; Wang, Y. The variability of biomass burning and its inﬂuence on regional aerosol properties during the wheat harvest season in North China. Atmos. Res. 2015, 157, 153–163. [CrossRef] 12. Zhang, H.; Hu, J.; Qi, Y.; Li, C.; Chen, J.; Wang, X.; He, J.; Wang, S.; Hao, J.; Zhang, L.; et al. Emission

characterization, environmental impact, and control measure of PM2.5 emitted from agricultural crop residue burning in China. J. Clean. Prod. 2017, 149, 629–635. [CrossRef] 13. Chen, W.; Tong, D.Q.; Dan, M.; Zhang, S.; Zhang, X.; Pan, Y. Typical atmospheric haze during crop harvest season in Northeastern China: A case in the Changchun region. J. Environ. Sci. 2017, 54, 101–113. [CrossRef] [PubMed] 14. Li, Q.; Zhang, L.; Wu, C.; Sun, Z.; Liu, X. Monitoring of straw burning based on satellite remote sensing and analysis of its impact on air quality. J. Ecol. Rural Environ. 2009, 25, 32–37. (In Chinese) 15. Wang, L.; Tian, Q.; Bao, Y. Research on the extraction model of crop straw ﬁre based on HJ satellite IRS remote sensing data. Geogr. Sci. 2011, 29, 661–667. (In Chinese) 16. Man, W.; Wang, Z.; Liu, M.; Lu, C.; Jia, M.; Mao, D.; Ren, C. Remote sensing analysis of spatial and temporal changes of cultivated land in Northeast China from 1990 to 2013. Chin. J. Agric. Eng. 2016, 34, 1–10. (In Chinese) 17. Bi, Y.; Wang, Y.; Gao, C. Quantity and regional distribution of straw resources in China. Agric. Mech. Res. 2010, 32, 1–7. (In Chinese) 18. Air Quality Historical Data Query. Available online: https://www.aqistudy.cn/historydata/ (accessed on 3 April 2019). 19. Liu, S.; Zhang, P. A study on the geographical agglomeration pattern of agricultural industry in Northeast China. Res. Agric. Mod. 2013, 34, 591–596. (In Chinese)

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).