Urban Expansion and Agricultural Land Loss in China: A ... - MDPI

sustainability Article

Urban Expansion and Agricultural Land Loss in China: A Multiscale Perspective Kaifang Shi 1,2 , Yun Chen 2, *, Bailang Yu 1, *, Tingbao Xu 3 , Linyi Li 4 , Chang Huang 5 , Rui Liu 1 , Zuoqi Chen 1 and Jianping Wu 1 1

2 3 4 5

*

Key Laboratory of Geographic Information Science, Ministry of Education, East China Normal University, Shanghai 200241, China; [email protected] (K.S.); [email protected] (R.L.); [email protected] (Z.C.); [email protected] (J.W.) CSIRO Land and Water, Canberra 2601, Australia Fenner School of Environment and Society, The Australian National University, Linnaeus Way, Canberra 2601, Australia; [email protected] School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China; [email protected] College of Urban and Environmental Sciences, Northwest University, Xi’an 710127, China; [email protected] Correspondence: [email protected] (Y.C.); [email protected] (B.Y.); Tel.: +61-2-6246-5729 (Y.C.); +86-21-5434-1172 (B.Y.)

Academic Editor: Marc A. Rosen Received: 25 May 2016; Accepted: 8 August 2016; Published: 11 August 2016

Abstract: China’s rapid urbanization has contributed to a massive agricultural land loss that could threaten its food security. Timely and accurate mapping of urban expansion and urbanization-related agricultural land loss can provide viable measures to be taken for urban planning and agricultural land protection. In this study, urban expansion in China from 2001 to 2013 was mapped using the nighttime stable light (NSL), normalized difference vegetation index (NDVI), and water body data. Urbanization-related agricultural land loss during this time period was then evaluated at national, regional, and metropolitan scales by integrating multiple sources of geographic data. The results revealed that China’s total urban area increased from 31,076 km2 in 2001 to 80,887 km2 in 2013, with an average annual growth rate of 13.36%. This widespread urban expansion consumed 33,080 km2 of agricultural land during this period. At a regional scale, the eastern region lost 18,542 km2 or 1.2% of its total agricultural land area. At a metropolitan scale, the Shanghai–Nanjing–Hangzhou (SNH) and Pearl River Delta (PRD) areas underwent high levels of agricultural land loss with a decrease of 6.12% (4728 km2 ) and 6.05% (2702 km2 ) of their total agricultural land areas, respectively. Special attention should be paid to the PRD, with a decline of 13.30% (1843 km2 ) of its cropland. Effective policies and strategies should be implemented to mitigate urbanization-related agricultural land loss in the context of China’s rapid urbanization. Keywords: urban expansion; agricultural land loss; nighttime light data; China

1. Introduction Agricultural land loss, as a type of land use/cover change (LUCC), is one of the most important factors that affect food security [1–4]. Human activities, especially urbanization, have resulted in a significant loss of agricultural land during the past decades around the world [5–8]. Substantial areas of agricultural land, including cropland, forest, and grassland, have been converted into artificial or impervious surfaces [3,9,10]. Therefore, mapping and quantifying agricultural land loss following urban expansion are essential to understanding its impact on food security.

Sustainability 2016, 8, 790; doi:10.3390/su8080790

www.mdpi.com/journal/sustainability

Sustainability 2016, 8, 790

2 of 16

China has been experiencing unprecedented urban expansion since its “reform and opening up” starting from the late 1970s [11,12]. The urban area in China has increased about four-fold [13]. China’s rapid urbanization has contributed to a massive agricultural land loss that could threaten its food security [14]. For example, urbanization in the more developed eastern China led to a loss of 7% from 1995 to 2000 [15]. In addition, the expansion of urban area has reduced agricultural land by 34% in some major cities of the Beijing–Tianjin–Hebei region in China from 1990 to 2000 [16]. Approximately 24% of agricultural land was converted to urban land in Changzhou, Jiangsu province, China from 2004 to 2011 [17]. Hence, the study of agricultural land loss due to urbanization is particularly crucial in China. Previous studies have estimated agricultural land loss caused by urban expansion in two major ways. Firstly, socioeconomic statistical data based on administrative units were used to obtain information on loss [15,18]. In spite of their authoritativeness, the socioeconomic statistical data only provide numeric records of loss for an entire administrative unit without showing internal spatial patterns [6]. Secondly, medium to high spatial resolution remotely sensed images have been widely employed for loss estimation [6,16,19]. These data mainly include Landsat Thematic Mapper (TM)/Operational Land Imager and Thermal Infrared Sensor (OLI–TIRS) images [20,21]. However, since they are limited by their geographic coverage [22,23], large amounts of computation and labor are required to obtain one-time information on loss at a national scale. As a consequence, few studies have accurately mapped urbanization-related agricultural land loss in a timely manner, especially at the national scale. To deal with the above existing issues, this study aims to map urban expansion and agricultural land loss in China from 2001 to 2013. The objectives are (1) to map urban expansion at 1 km resolution in China using the nighttime stable light (NSL), normalized difference vegetation index (NDVI), and water body data; (2) to monitor urbanization-related agricultural land loss at 1 km resolution in China using the above geographic data and land use/cover data; (3) to analyze urbanization-related agricultural land loss from national scale to regional and metropolitan scales. The remainder of this study is organized as follows. Section 2 describes the study areas and data source. Section 3 introduces the methodology used. Section 4 analyzes the results of urban expansion and agriculture land loss. Section 5 presents the discussion, and the conclusions are given in Section 6. 2. Study Areas and Data Sources 2.1. Study Areas Study areas were selected from three different levels for multiple-scales analysis. Following the first national level, the second is the regional level. Due to China’s uneven socioeconomic development, different regions with great disparities of urban expansion have been formed. In this study, China was divided into three regions (eastern, central, and western) based on their socioeconomic development and geographical position (Figure 1). It is noted that these divided regions have been commonly accepted and widely used to analyze China’s socioeconomic development [24,25]. The metropolitan scale forms the third level. Because China’s urban expansion is concentrated in some metropolitan areas which contribute the most to agricultural land loss [16], six typical metropolitan areas—Shanghai–Nanjing–Hangzhou (SNH), Beijing–Tianjin–Tangshan (BTT), Pearl River Delta (PRD), Chengdu–Deyang–Mianyang (CDM), Zhengzhou–Luoyang–Jiaozuo (ZLJ), and Wuhan–Ezhou–Huangshi (WEH) (Figure 1)—were selected to represent this level.

Sustainability 2016, 8, 790 Sustainability 2016, 8, 790

3 of 16 3 of 15

Figure 1. 1. The The location location of of study study areas. areas. Figure

2.2. Data Data Source Source 2.2.

The The Defense Defense Meteorological Meteorological Satellite Satellite Program’s Program’s Operational Operational Linescan Linescan System System (DMSP–OLS) (DMSP–OLS) NSL NSL data National Oceanic Oceanic and and Atmospheric Atmospheric Administration Administration (NOAA)/National (NOAA)/National data were obtained from the National Geophysical Geophysical Data Data Center Center (NGDC). (NGDC). These These data data include include the the lights lights from from cities, cities, towns, towns, and and other other sites sites with persistent lighting and removed removed ephemeral events, such as as fire, fire, and present present the annual annual average average brightness units of of six-bit six-bitdigital digitalnumbers numbers(DN) (DN)ranging rangingfrom from 0 to They cover area of −180 brightness in units 0 to 63.63. They cover anan area of ´180 to to degrees longitude and −65toto75 75degrees degreesin inlatitude, latitude,at at aa spatial spatial resolution resolution of 30 arc-seconds 180180 degrees in in longitude and ´65 arc-seconds (about NSL data were collected by three different DMSP satellites (F15,(F15, F16, and (about 11km). km).Since Sincethe the NSL data were collected by three different DMSP satellites F16,F18), and they not be usedused to map urban expansion due to tothe F18), could they could notdirectly be directly to map urban expansion due thelack lackof of continuity continuity and and comparability [26–29]. In Inthis thisstudy, study, assumed would continuously grow in comparability [26–29]. wewe assumed thatthat the the NSLNSL datadata would continuously grow in China, China, DN values in an year earlier year would betoequal to or smaller thaninthose in year. a later year. and theand DNthe values in an earlier would be equal or smaller than those a later Spatial Spatial data were converted intoformat rasterand format and resampled to aresolution spatial resolution of 1projected km and data were converted into raster resampled to a spatial of 1 km and projected into theAzimuthal Lambert Azimuthal Equal Area Projection with to reference WGS84 datum. into the Lambert Equal Area Projection with reference WGS84 to datum. The The Moderate Moderate Resolution Resolution Imaging Imaging Spectroradiometer Spectroradiometer (MODIS) (MODIS) monthly monthly NDVI NDVI and and water water body body data (NLC) data were used to data were were also also used usedto tomap mapurban urbanexpansion. expansion.The Thenational nationalland landcover/use cover/use (NLC) data were used estimate urbanization-related agricultural land loss in China. Landsat TM/OLI–TIRS images and to estimate urbanization-related agricultural land loss in China. Landsat TM/OLI–TIRS images and provincial gross domestic product (GDP) were applied to assess the accuracy of urban area extraction. The statistical statistical data dataon ongrain grain production were employed to quantify the impact of urbanizationproduction were employed to quantify the impact of urbanization-related related agricultural land a local scale. A summary of used the data used in this studyinisTable given agricultural land loss at aloss localatscale. A summary of the data in this study is given 1. in Table 1.


4 of 15

Table 1. Description of the data used in this study.


4 of 16

Source Year 2001, 2005, Table 1. stable Description of the data used in this study.NOAA/NGDC Global nighttime NSL 2009, and light composite data (http://www.ngdc.noaa.gov/eog/dmsp.html) 2013 Data Data Description Year Source Monthly NDVI composite 2001, 2005, Global nighttime stable light 2001, 2005, 2009, NOAA/NGDC Geospatial Data Cloud NSL MODIS–NDVI datacomposite of China data at 1 km 2009, and and 2013 (http://www.ngdc.noaa.gov/eog/dmsp.html) (http://www.gscloud.cn) resolution 2010 Monthly NDVI composite 2001,2013 2005, 2009, Geospatial Cloud Water body Shapefile ESRI BaruchData Geoportal MODIS–NDVI dataofofwater China in at China and 2010 (http://www.gscloud.cn) 1 km resolution Ten images covering six Water body Shapefile of waterTianjin, in China 2013 ESRI Baruch Geoportal cities (Beijing, 2001 and Geospatial Data Cloud TM/OLI–TIRS Shanghai, Chengdu, Ten images covering six 2013 (http://www.gscloud.cn) citiesand (Beijing, Tianjin, Geospatial Data Cloud Wuhan Hefei) at 30 m TM/OLI–TIRS 2001 and 2013 Shanghai, Chengdu, Wuhan (http://www.gscloud.cn) resolution and Hefei) at 30 m resolution National land cover/use National Data Sharing Infrastructure of National land cover/use National Data Sharing Infrastructure of Earth NLC data at 1 km resolution in 2000 Earth System Science NLC data at 1 km resolution 2000 System Science China (http://www2.geodata.cn/index.html) in China (http://www2.geodata.cn/index.html) Annual dataofof The Database of Economic and Annualstatistical statistical data TheStatistical Statistical Database of Economic and Social 2001, 2005, Socioeconomic (1088Yuan) Yuan) 2001, 2005, 2009, Social Development by the Knowledge National twotypes—GDP types—GDP (10 Socioeconomic two Development by the National 2009, and 4 t) 4 t) and 2013 andgrain grainproduction production (10 censusdata data Infrastructure of China census and (10 Knowledge Infrastructure of China 2013 in China China (http://tongji.cnki.net) in (http://tongji.cnki.net) Administrative Shapefiles files of of provinces, National GeomaticsCenter Center of Administrativ Shape provinces, National Geomatics ofChina China 2008 2008 boundaries cities in China (http://ngcc.sbsm.gov.cn/article/en/or/an/) e boundaries cities in China (http://ngcc.sbsm.gov.cn/article/en/or/an/) Data

Data Description

3. Methodology Methodology 3. Three main were undertaken to map land loss caused urban expansion Three mainprocedures procedures were undertaken to agricultural map agricultural land lossby caused by urban in China: urban expansion was extracted from 2001 to 2013; agricultural land was mapped in 2000; expansion in China: urban expansion was extracted from 2001 to 2013; agricultural land was mapped and agricultural land lossland caused urban by expansion was estimated at national,atregional, urban in 2000; and agricultural lossbycaused urban expansion was estimated national,and regional, metropolitan scales (Figure 2). and urban metropolitan scales (Figure 2).

NSL

NDVI

Water body

Urban area extraction using the NFS method 1) Identifying the transition zones 2) Quantifying central urban areas 3) Defining marginal urban areas 4) Mapping urban expansion

NLC

Mapping agricultural land

Procedure I

Procedure II TM/OLI– TIRS

Urban expansion in China from 2001 to 2013

Input

(Validation) GDP

Agricultural land of China in 2000

Estimating agricultural land loss caused by urban expansion Procedure III

Method

Agricultural land loss at a national scale

Agricultural land loss at a regional scale

Output

Figure Figure 2. 2. Flowchart Flowchart of of methodology. methodology.

Agricultural land loss at a metropolitan scale


5 of 16

3.1. Extracting Urban Expansion Accurately extracting urban expansion is a prerequisite for quantifying agricultural land loss caused by urbanization. The NSL data have been demonstrated as an accurate and valuable data source for mapping urban expansion at regional, national, continental, and global scales [30–34] due to their appropriate temporal and spatial resolution. Among the published methods for urban area extraction from NSL data, the neighborhood focal statistic (NFS) method has been effectively used due to its simplicity [35,36]. However, due to the “blooming” natural features in the NSL data, the NFS method might confuse some non-urban pixels located in the marginal urban areas, and misclassify many natural features situated within urban regions. In order to improve the accuracy of urban extraction, we mapped urban expansion in China by applying the NFS method to the NSL, NDVI, and water body data from 2001 to 2013. The specific steps for urban extraction are as follows: (1)

(2)

(3)

(4)

Identifying the transition zones. Because the NSL data could be treated as Digital Elevation Model (DEM) images, identifying the dividing lines between ridges and valleys is equivalent to distinguishing urban and non-urban areas. However, due to the “blooming effect” of the NSL data, these lines are actually transition zones, including both urban and non-urban areas. In this study, the transition zones were identified by using maximum and minimum NFS calculations. Quantifying central urban areas. According to the differences of pixel values in the NSL data, the regions with relatively high values on the side of transition zones were classified as central urban areas. Defining marginal urban areas. Since the transition zones are more heterogeneous landscapes, the marginal urban areas hidden in the transition zones were extracted using the minimum NFS calculation. Mapping urban expansion. The primary urban maps were obtained by overlaying central and marginal urban areas. To deal with any misclassifications in these primary urban maps, we applied NDVI and water body data to remove natural features located within urban areas from the primary urban maps. Moreover, urban areas were assumed to continuously grow outward, and an urban pixel of reclassified urban maps in an earlier year would remain as urban in a later year. Additional information can be found in the supplementary materials.

Two methods were used to validate the mapping of urban area estimation from the NSL, NDVI, and water body data. Since Yang et al. [37] and He et al. [13] demonstrated that GDP correlated well with urban expansion, we firstly evaluated the estimated results against GDP at the provincial level in China. Secondly, Landsat TM/OLI–TIRS images were employed to validate the spatial accuracy of urban areas. Considering Landsat TM/OLI–TIRS images have much higher spatial resolution (30 m) than the NSL data (1 km), they are efficient to represent the real pattern of urban areas [20,30,38,39]. The urban areas extracted from Landsat TM/OLI–TIRS images were produced by the maximum likelihood classifiers based on training data extracted from typical urban areas. These results at 30 m resolution were then aggregated to 1 km urban maps to facilitate the intercomparison [20]. Finally, the Kappa, overall accuracy (OA), commission error (CE), and omission error (OE) were calculated and used for the accuracy estimation. 3.2. Mapping Agricultural Land in 2000 The land use/cover classification scheme (2006–2020) developed by the Outline of the National Overall Planning on Land Use [40] was adopted in this study and employed to extract agricultural land from the 2000 NLC data. According to this classification scheme, any farmland, including cropland, grassland, and forest, can be regarded as agricultural land. Several land cover types in the 2000 NLC data were combined to map three agricultural land classes—cropland, grassland, and forest. Specifically, paddy and dry land were merged into cropland; forest, shrub wood, open and other forest were reclassified as forest; dense, moderate, and sparse grass were all considered as grass


6 of 16

cover in this study. The detailed descriptions of each land use/cover type in the NLC data can be found in Ran et al. [41]. Sustainability 2016, 8, 790

6 of 15

3.3. Estimating Agricultural Land Loss Caused by Urban Expansion 3.3. Estimating Agricultural Land Loss Caused by Urban Expansion Agricultural land loss from urban expansion refers to the conversion of the corresponding land landuse. lossInfrom refers the conversion of the corresponding land coverAgricultural into urban land this urban study, expansion we calculated thetoareas of agricultural land loss in the urban cover into urban land use. In this study, we calculated the areas of agricultural land loss in the urban expansion process for each pixel using Formula (1): expansion process for each pixel using Formula´ (1): ¯ 2001 ALpi, jq “ AOp2000 ˆ Urban2013 ´ i − Urbani ),, (, ) = ( i,, )jq × (

(1) (1)

whereAL AL (i,j)isis the area j type agricultural land at pixel the ith pixel 2001 AO(i,j) where the area of jof type agricultural land lost at lost the ith from 2001from to 2013; AOto represents (i,j) (i,j)2013; 2013 2001 2013 2001 represents area of j type agricultural land at the ith pixel in i2000.and Urban i Urban i are class the area of jthe type agricultural land at the ith pixel in 2000. Urban Urban class values at iand are values at the ith pixel in the urban land map in 2001 and 2013, respectively. The class value of the ith pixel in the urban land map in 2001 and 2013, respectively. The class value of 1 represents urban1 represents urban while 0 is non-urban. The total area agricultural landwas losscalculated in each pixel was while 0 is non-urban. The total area of agricultural landofloss in each pixel through calculated through summation as presented in Formula (2): summation as presented in Formula (2): ÿm (2) ALi “= ∑ AL(p,i,j)q, (2) j“1

where ALi is the total area of agricultural land loss within the ith pixel from 2001 to 2013; m represents where ALi is the total area of agricultural land loss within the ith pixel from 2001 to 2013; m represents the original number of agricultural land types in 2000. the original number of agricultural land types in 2000. 4. Results 4. Results 4.1. Validation Validation of of Urban Urban Area Area Extraction Extraction 4.1. The regression regression model model was was firstly firstly used used to to compare compare GDP GDP against against the the mapped mapped urban urban areas areas at at the the The provinciallevel levelininChina. China.Based Basedononcurrent current data availability, 31 provinces were to implement provincial data availability, 31 provinces were usedused to implement the the assessment. The results showed the estimated urban expansion hadcorrelation a strong correlation assessment. The results showed that thethat estimated urban expansion had a strong (R ě 0.75) (R ≥ the 0.75) with the GDP fromat2001 to 2013 at of the 0.001 level(Figure of significance (Figurecapital 3). Six with GDP growth fromgrowth 2001 to 2013 the 0.001 level significance 3). Six important important capital cities (Beijing, Tianjin, Shanghai, Chengdu, Wuhan, and Hefei) with various cities (Beijing, Tianjin, Shanghai, Chengdu, Wuhan, and Hefei) with various levels of urbanizationlevels were of urbanization chosenthe asspatial samples to validate theareas. spatial accuracy of urban areas. Beijing, chosen as sampleswere to validate accuracy of urban Beijing, Tianjin, and Shanghai are the Tianjin, and Shanghai areathe most developed cities with a large-scale urban expansion in are China. most developed cities with large-scale urban expansion in China. Chengdu, Wuhan, and Hefei the Chengdu, Wuhan, andinHefei are the(Chengdu) socioeconomic centers(Wuhan in the western (Chengdu) and central socioeconomic centers the western and central and Hefei) regions, respectively. (Wuhanthese and cities Hefei)are regions, respectively. Hence, these cities are the most representative samples for Hence, the most representative samples for urban expansion validation. The results urban expansion validation. The results showed that the urban areas of these cities extracted using showed that the urban areas of these cities extracted using the NSL, NDVI, and water body data in the NSL, NDVI, and water data in 2001 andof 2013 presented an OA average valuean ofaverage 0.63, an 2001 and 2013 presented an body average Kappa value 0.63, an average valueKappa of 93.58%, average value and of 93.58%, an average 2.67%,4and average OE value of 3.75% CE valueOA of 2.67%, an average OE valueCE of value 3.75% of (Figures and an 5, Table 2). The relatively high (Figures 4 and 5, Table 2). The relatively high accuracy suggested that the proposed method could be accuracy suggested that the proposed method could be effectively and accurately mapping urban effectivelyin and accurately mapping urban expansion in China. expansion China.

12,000 9,000 6,000 3,000

(a)

0

GDP growth (108 Yuan)


2005–2009 R = 0.81, P < 0.001

20,000

25,000

16,000 12,000 8,000 4,000

(b)


2001–2005 R = 0.91, P < 0.001

15,000

0 0

400

800

1,200

1,600

Urban expansion (km2)

2,000

2009–2013 R = 0.75, P < 0.001

20,000 15,000 10,000 5,000

(c)

0 0

500

1,000

1,500

2,000


2,500

0

500

1,000

1,500

2,000

2,500


Figure3. 3. Correlation Correlationbetween betweenurban urbanexpansion expansionand andGDP GDPgrowth growthatatthe theprovincial provincial level. 2001– Figure level. (a)(a) 2001–2005; 2005; (b) 2005–2009; (c) 2009–2010. (b) 2005–2009; (c) 2009–2010.


7 of 16 7 of 15

Figure 4. Accuracy assessment of urban areas extracted from the NSL, NDVI, and water body data in images in in 2013. 2013. comparison with those from Landsat OLI–TIRS images

Table 2. 2. Accuracy Accuracy assessment of urban areas extracted from the NSL, NDVI, and water body data in Table TM/OLI–TIRS images. comparison with those from Landsat TM/OLI–TIRS images.

Year

Index Kappa Kappa OA (%) OA (%) 2001 2001 CE(%) (%) CE OE(%) (%) OE Kappa Kappa OA OA(%) (%) 2013 2013 CE (%) CE (%) OE (%) OE (%) Kappa Kappa OA (%) Average OA (%) Average CE (%) CE (%) OE (%) OE (%) Year

Index

Beijing Beijing 0.62 0.62 95.22 95.22 1.62 1.62 3.17 3.17 0.61 0.61 90.46 90.46 5.33 5.33 4.21 4.21 0.61 0.61 92.84 92.84 3.47 3.47 3.69 3.69

Tianjin Shanghai Chengdu Wuhan Tianjin Shanghai Chengdu Wuhan 0.60 0.62 0.56 0.63 0.60 0.62 0.56 0.63 95.34 89.78 98.51 96.73 95.34 89.78 98.51 96.73 1.38 3.22 0.25 0.55 1.38 3.22 0.25 0.55 3.27 7.00 1.24 2.72 3.27 7.00 1.24 2.72 0.62 0.62 0.67 0.65 0.62 0.62 0.67 0.65 89.75 82.96 94.89 93.43 89.75 82.96 94.89 93.43 4.55 9.24 2.36 2.54 4.55 9.24 2.36 2.54 5.70 7.80 2.75 4.02 5.70 7.80 2.75 4.02 0.61 0.62 0.61 0.64 0.61 0.62 0.61 0.64 92.55 86.37 96.70 95.08 92.55 86.37 96.70 95.08 2.96 6.23 1.30 1.55 2.96 6.23 1.30 1.55 4.49 7.40 2.00 3.37 4.49 7.40 2.00 3.37

Hefei Hefei 0.66 0.66 98.75 98.75 0.11 0.11 1.15 1.15 0.74 0.74 97.15 97.15 0.85 0.85 2.00 2.00 0.70 0.70 97.95 97.95 0.48 0.48 1.57 1.57


8 of 16 8 of 15


8 of 15

Figure5.5.Accuracy Accuracyassessment assessment of of urban urban areas areas extracted extracted from Figure from the the NSL, NSL, NDVI, NDVI, and andwater waterbody bodydata datainin Figure 5. Accuracy assessment of urban areas extracted from the NSL, NDVI, and water body data in comparison with those from Landsat TM images in 2001. comparison with those from Landsat TM images in 2001. comparison with those from Landsat TM images in 2001.

4.2.Urban UrbanExpansion Expansionfrom from2001 2001 to to 2013 2013 4.2. 4.2. Urban Expansion from 2001 to 2013 Thespatial spatialdistribution distribution of of urban urban expansion expansion in The in China China from from 2001 2001 to to2013 2013mapped mappedusing usingthe theNSL, NSL, The spatial distribution of is urban expansion in China fromthe 2001past to 2013 mapped using the NSL, NDVI, and water body data shown in Figure 6. Over decade, China experienced NDVI, and water body data is shown in Figure 6. Over the past decade, China experienced accelerated NDVI, andurban water body data shown in Figure 6. Over the past decade, China experienced accelerated at is a rate much higher rate than the world average. China’s total urban area urban expansion, atexpansion, a much higher than the world average. China’s total urban area increased from accelerated urban expansion, at a much higher rate than the world average. China’s total urban area 2 2 2 2 increased from 31,076 km in 2001 to 80,887 km in 2013, with an average annual growth rate of 13.36%, 31,076 km in 2001 to 80,8872 km in 2013, with an2 average annual growth rate of 13.36%, while the world increased from average 31,076 km in 2001 toto80,887 km only in 2013, with[42]. an average annual rate of 13.36%, while the world 2000[42]. was 3.20% Consistent withgrowth China’s urbanization average from 1990average to 2000from was 1990 only 3.20% Consistent withConsistent China’s urbanization trend, the urban while the world from 1990 to 2000 was only 3.20% [42]. with China’s urbanization trend, thewithin urbanthe expansion within the three regions increased continuously from 2001 to 2013. expansion three regions increased continuously from 2001 to 2013. However, our trend, the urban expansion within the three regions increased continuously from 2001 tomapping 2013. However, our mapping showed a great difference in between urban expansion speed with between these regions, showed a great difference in urban expansion speed these regions, the eastern region However, our mapping showed a great difference in urban expansion speed between these regions, with the eastern region having the largest percentage of total landareas area(1.69%; to become urban areas 2 ), the(1.69%; having the largest percentage of total land area to become urban 28,463 km central with the eastern region having the largest percentage of total land area to become urban areas (1.69%; 28,463experiencing km2),2 the central region experiencing medium levels of growth urbanization with a netkm growth of 0.75% 2 ) of the region medium levels of urbanization with a net of 0.75% total 28,463 km ), the central region experiencing medium levels of urbanization with(7736 a net growth of 0.75% 2 (7736 km )2 of the total land area, while the western region showing the lowest percentage of its total2 land area, thetotal western region showing the lowest percentage of itslowest total area (0.20%;of 13,612 km ) (7736 kmwhile ) of the land area, while the western region showing the percentage its total area (0.20%; 13,612 km2)2 to become urban areas (Table 3). to area become urban areas 3). (0.20%; 13,612 km(Table ) to become urban areas (Table 3).

Figure6. Urban expansion expansion in Figure Urban in China China from from2001 2001 to 2013. Figure 6.6.Urban 2001to to2013. 2013.

Sixmetropolitan metropolitanareas areas accounted accounted for for 2.88% 2.88% of China’s total land area but contained 33.21% ofof Six China’s total land area contained 33.21% Six metropolitan areas accounted for 2.88% ofof China’s total land area butbut contained 33.21% of the the country’s new urbanized areas from 2001 to 2013 (Table 3). The SNH, BTT, and PRD experienced the country’s urbanized to 2013 (Table 3). The SNH, experienced country’s new new urbanized areasareas fromfrom 20012001 to 2013 (Table 3). The SNH, BTT,BTT, and and PRDPRD experienced high high levels of urbanization with more than 2900 km22 of newly urban areas, accounting for more than 2 high levels of urbanization with more than 2900 km of newly urban areas, accounting for more than levels of urbanization with more than 2900 km of newly urban areas, accounting for more than 6.5% 6.5% of their total land areas. For the CDM, ZLJ and WEH, more than 2% of their land areas were theirland total landFor areas. For the CDM, andmore WEH, more 2%land of their areas were of6.5% theiroftotal areas. CDM, andZLJ WEH, than 2% than of their areasland were converted converted into urban areasthe from 2001ZLJ to 2013 (Table 3). converted into urban areas from 2001 to 2013 (Table 3). into urban areas from 2001 to 2013 (Table 3).


9 of 16 9 of 15

Table 3. Urban expansion and agricultural land loss at regional and urban agglomeration scales.

Table 3. Urban expansion and agricultural land loss at regional and urban agglomeration scales. Agricultural Land in Urban Expansion Agricultural Land Loss Urban Expansion Agricultural 2000 from 2001 to 2013 from 2001 Land to 2013 Total Land Agricultural Administrative Land in 2000 2 from 2001 to 2013 Loss from 2001 to 2013 Administrative Total Land Area (km ) Unit Area Percentage Area Percentage Area Percentage Unit Area (km2) Area Percentage Area Percentage Area Percentage 2 2 2 (km ) (%) (km ) (%) (km ) (%) (km2) (%) (km2) (%) (km2) (%) Eastern region 1,684,856 1,547,427 91.84 28,463 1.69 18,542 1.10 Eastern region 1,684,856 1,547,427 91.84 28,463 1.69 18,542 1.10 Central region 1,027,217 951,166 92.60 7736 0.75 5276 0.51 Central region 1,027,217 951,166 92.60 7736 0.75 5276 0.51 Western region 6,725,648 4,574,035 68.01 13,612 0.20 9262 0.14 Western 6,725,648 4,574,035 68.01 13,612 0.20 9262 0.144.97 SNH region 95,097 77,261 81.24 6787 7.14 4728 SNH 95,097 77,261 81.24 6787 7.14 4728 4.973.90 BTT 4048 32,718 76.00 2907 6.75 1680 BTT 4048 32,718 76.00 2907 6.75 1680 3.905.14 PRD 52,556 44,636 84.93 4387 8.35 2702 PRD 52,556 44,636 84.93 4387 8.35 2702666 5.141.67 CDM 39,833 38,112 95.68 939 2.36 ZLJ 26,759 24,177 90.35 928 3.47 CDM 39,833 38,112 95.68 939 2.36 666580 1.672.17 WEH 14,731 11,166 75.80 592 4.02 ZLJ 26,759 24,177 90.35 928 3.47 580325 2.172.21 WEH 14,731 11,166 75.80 592 4.02 325 2.21

4.3. Agricultural Land in 2000 4.3. Agricultural Land in 2000 The spatial distribution of agricultural land in China is shown in Figure 7. In 2000, agricultural The spatial distribution of agricultural land in China is shown in Figure 7. In 2000, agricultural land covered 74.94% (7,072,628 km2 ) of the total area, with cropland, forest, and grassland accounting land covered 74.94% (7,072,628 km2) of the total area, with cropland, forest, and grassland accounting 2 2 ) of the for 19.10% (1,802,821 km ),2 23.93% (2,257,876 km22 ), and 31.91% (3,011,931 km total area, for 19.10% (1,802,821 km ), 23.93% (2,257,876 km ), and 31.91% (3,011,931 km2) of the total area, 2 respectively. The eastern region had high proportions of cropland and forest: 41.88% (705,574 km ) respectively. The eastern region had high proportions of cropland and forest: 41.88% (705,574 km2) and 42.16% (710,365 km2 )2 of their total area, respectively. Of the total area, cropland and forest made and 42.16% (710,365 km ) of their total area, respectively. Of the total area, cropland and forest made 2 ) in the central region while grassland covered up 41.68% (428,093 km22) and 42.59% (437,499 km 2) in the central region while grassland covered 41.56% up 41.68% (428,093 km ) and 42.59% (437,499 km 2 ) in the western region. In addition, six metropolitan areas also had different 41.56% (2,794,869 (2,794,869 km2) km in the western region. In addition, six metropolitan areas also had different proportions ofof agricultural oftheir theirtotal totalareas, areas,cropland cropland was 54.97% proportions agriculturalland. land. In In terms terms of of proportions proportions of was 54.97% 2 2 2 (14,710 kmkm ) 2in the WEH and and 54% 54%(21,511 (21,511km km2) in ) inthe the CDM. Forest was (14,710 ) in theZLJ, ZLJ,54.25% 54.25%(7992 (7992km km2)) in in the the WEH CDM. Forest was 2 )2 in the PRD. It should be pointed out that grassland in particular accounted for 56.52% (29,705 km 56.52% (29,705 km ) in the PRD. It should be pointed out that grassland in particular accounted for 2 11.21% (4465 km 11.21% (4465 km) 2of ) oftotal totalarea areaininthe theCDM CDM (Figure (Figure 8). 8).

Figure 7. The spatial distribution of agricultural land in China in 2000. Figure 7. The spatial distribution of agricultural land in China in 2000.


10 of 16


10 of 15

Cropland

Forest

Grassland

(a)

2,000,000 1,500,000

16.50%

1,000,000

41.88%

42.16%

9.95% 41.68% 42.59%

500,000 7.80%

60,000

41.56%

2,500,000

Agricultural land (km2)

Agricultural land (km2)

3,000,000

50,000

30,000

Central region

Western region

Grassland

(b) 56.52% 27.23% 54.00%

48.50%

20,000

26.36%

30.47%

21.47%

10,000 0

Eastern region

Forest

40,000

8.36%

0

Cropland 52.96%

1.06%

SNH

6.03%

BTT

54.97%

11.21% 2.05%

PRD

CDM

26.88%

54.25%

8.49%

ZLJ

19.39% 2.15%

WEH

Figure Figure 8. 8. Agricultural Agricultural land land of of 2000 2000 at at (a) (a) regional regional and and (b) (b) metropolitan metropolitan scales. scales.

4.4. Urbanization-Related Urbanization-Related Agricultural Land Loss from 2001 to 2013 at National, Regional, and Metropolitan Scales Widespread urban urban expansion expansion led led to to aa total loss of agricultural land of 33,080 km22 accounting for 0.47% of the total area of agricultural land of China in 2000 (Figure 9). Cropland experienced severe 2 or 2 or 1.37% ofof thethe total area of cropland. Forest decreased by aby rate loss with with aa high highdecline declineofof24,783 24,783km km 1.37% total area of cropland. Forest decreased a 2 2 of 0.17% of the area area of forest, which is equal to about 3750 km China. Grassland was least rate of 0.17% oftotal the total of forest, which is equal to about 3750 in km in China. Grassland was 2) of affected, with with a lowapercentage 0.15%0.15% (4547(4547 km2 ) km of its total area being lost tolost urban expansion. least affected, low percentage its total area being to urban expansion.

Figure land loss loss in Figure 9. 9. Spatial Spatial distribution distribution of of urbanization-related urbanization-related agricultural agricultural land in China China from from 2001 2001 to to 2013. 2013.

The impact of of urbanization urbanizationon onthe thearea areaofofagricultural agriculturalland landdiffered differedinin the three regions (Figure The impact the three regions (Figure 9). 2 or 1.2% of its total agricultural land area, whereas the central 9). The eastern region lost 18,542 km 2 The eastern region lost 18,542 km or 1.2% of its total agricultural land area, whereas the central region 5276 km km22 (0.55%), 9262 km km22 (0.20%). region lost lost 5276 (0.55%), and and the the western western region region lost lost 9262 (0.20%). Specifically, Specifically, the the eastern eastern 2 region more than 15,000 kmkm accounting for 2.24% of itsof total 2 accounting region experienced experiencedsevere severecropland croplandlosses losseswith with more than 15,000 for 2.24% its cropland area. Although less severe, the central and western regions also lost a considerable total cropland area. Although less severe, the central and western regions also lost a considerable proportion study period. period. A proportion (0.7%–1%) (0.7%–1%) of of their their cropland cropland during during the the study A relatively relatively significant significant forest forest loss loss 2). The central and western region had low levels also occurred in the eastern region (0.31%; 2180 km 2 also occurred in the eastern region (0.31%; 2180 km ). The central and western region had low levels of forest loss, accounting for 0%–0.15% of their total forest areas. Additionally, grassland loss was


11 of 16


11 of 15

of forest loss, accounting for 0%–0.15% of their total forest areas. Additionally, grassland loss was concentrated in the central region, which accounted for 0.49% (420 km22)) of of the the total total area area of of grassland. grassland. 2 Contrarily, Contrarily, less less than than 1.5% 1.5% (3561 (3561 km km2) of grassland loss occurred occurred in in the the western western region region (Figure (Figure 10). 10).

16,000

Cropland

Forest

5,000

Grassland

2.24%

(a)

12,000 8,000 0.70%

1.00%

4,000

0.13%

0.31% 0.43%

0.13% 0.49%

0.09%

Agricultural land loss (km2)


20,000

Eastern region

Central region

Western region

Forest

Grassland

4,000

(b)

3,000 13.30%

2,000

7.45%

1,000

2.68% 0.69%

0

0

Cropland

8.97%

3.38%

SNH

0.79% 1.96%

BTT

5.77%

PRD

3.00%

3.78% 3.65% 0.12% 0.13% 0.14% 0.62% 0.81% 3.15%

CDM

ZLJ

WEH

Figure 10. 10. Agricultural Figure Agricultural land land loss loss following following urban urban expansion expansion from from 2001 2001 to to 2013 2013 at at (a) (a) regional regional and and (b) (b) urban metropolitan scales. urban metropolitan scales.

A significant loss of of agricultural land has has been A significant loss agricultural land been driven driven by by urban urban expansion expansion in in the the six six metropolitan metropolitan 2 areas. Around 10,681 km of agricultural land was converted into urban land use from 2001 2013 2 areas. Around 10,681 km of agricultural land was converted into urban land use from 2001 to to 2013 in in these metropolitan areas. The metropolitan areas with the largest percentages of their total these metropolitan areas. The metropolitan areas with the largest percentages of their total agricultural agricultural landhave areas thaturbanized have beenwere urbanized the SNH 4728 (6.05%; km2), PRD land areas that been in the were SNH in (6.12%; 4728(6.12%; km2 ), PRD 2702(6.05%; km2 ), 2 2 2702 km ), and BTT (5.13%; 1680 km ). The WEH and ZLJ underwent medium levels of agricultural 2 and BTT (5.13%; 1680 km ). The WEH and ZLJ underwent medium levels of agricultural land loss 2 of the total agricultural land area, land aloss with aofdecrease of 2.91% (3252.40% km2) (580 and 2.40% with decrease 2.91% (325 km2 ) and km2 ) of(580 the km total) agricultural land area, respectively. 2 respectively. In addition, agricultural land loss of 1.75% (666 km ) was detected the CDM. It should 2 In addition, agricultural land loss of 1.75% (666 km ) was detected in the CDM. Itinshould be noted that be noted that a significant percentage (>8.5%) of cropland was lost in the SNH and PRD. Special a significant percentage (>8.5%) of cropland was lost in the SNH and PRD. Special attention should be attention be also paidabout to the6% PRD, where about grassland also paid should to the PRD, where of grassland was6% lostof(Figure 10). was lost (Figure 10). 5. Discussion 5. Discussion 5.1. Correlation between Urbanization-Related Agricultural Land Loss and Grain Grain Production Production Decrease Decrease During therapid rapid urban expansion last decade, considerable urbanization-related During the urban expansion of theof lastthe decade, considerable urbanization-related agricultural agricultural has been found in China,inespecially in the six metropolitan areas.and ThePRD SNHwere and land loss hasland beenloss found in China, especially the six metropolitan areas. The SNH PRD were particularly having experienced an alarming in agricultural land coverage; particularly notable fornotable havingforexperienced an alarming drop in drop agricultural land coverage; more more than of cropland in these was converted into urban landThe use.evaluated The evaluated than 8% of8% cropland in these areas areas was converted into urban land use. resultsresults were were reliable because they were similar to those in Tan et al. [16] and Liang et al. [17]. In addition, reliable because they were similar to those in Tan et al. [16] and Liang et al. [17]. In addition, compared compared other studies [43–45], we also the urbanization-related agricultural loss in with otherwith studies [43–45], we also found the found urbanization-related agricultural land loss land in China to China to be much morethan serious that countries. of other countries. be much more serious thatthan of other Normally, there is is aa positive positive correlation correlation between between agricultural agricultural land loss and grain production decrease [46]. [46]. However, However,due duetoto technological progress heavy of chemical fertilizer thethe technological progress andand heavy use use of chemical fertilizer and and pesticides, China’s grain production presented a growth trend duringthe thepast pastdecades decades [47–49]. [47–49]. pesticides, China’s grain production presented a growth trend during Nevertheless, agricultural land lossloss has has a positive correlation with Nevertheless, the thesignificant significanturbanization-related urbanization-related agricultural land a positive correlation grain production decrease and partly threatens foodfood security in some developed areas. Figure 11 with grain production decrease and partly threatens security in some developed areas. Figure presents the relationship between urbanization-related agricultural land loss and grain production 11 presents the relationship between urbanization-related agricultural land loss and grain decrease in 16 typical cities (Figure 1) within the SNH and PRD, showing a positive correlation with R-value of 0.65. Specifically, the substantial area (>500 km22)) of of agricultural agricultural land land lost lost in Shanghai, Guangzhou, Suzhou, Suzhou, Ningbo, Ningbo, Foshan, Foshan, and and Wuxi correlated correlated to a decrease of 2.01 million ton of grain production. Moreover, waswas a significant heterogeneity in the quantity qualityand of agricultural Moreover,there there a significant heterogeneity in the and quantity quality of land with each city,with and therefore has beeneach experienced grain production decrease. agricultural land each city,each andcity therefore city has different been experienced different grain However, relativelyHowever, small agricultural land loss could a disproportionate productioneven decrease. even relatively small cause agricultural land loss grain couldproduction cause a decrease. For example, Zhaoqing had a lowFor level of agricultural land (137level km2of ), yet showed disproportionate grain production decrease. example, Zhaoqing hadloss a low agricultural 2 aland 0.23loss million ton grain production decrease during the study period (Figure 11). Hence, it very (137 km ), yet showed a 0.23 million ton grain production decrease during the study is period important toHence, implement urban planning and agricultural land protection policies in these cities. (Figure 11). it is viable very important to implement viable urban planning and agricultural land protection policies in these cities.


R = 0.65

1,000

(a)

800 600 400 200 0 0

10 20 30 40 Grain production decrease (104 ton)

50

50 40

Grain production decrease

1,200

Agricultural land loss

1,000

(b)

30

800 600

20

400

10

200 0

0 Guangzhou Shanghai Suzhou Ningbo Foushan Wuxi Dongguan Changzhou Nanjin Huizhou Hangzhou Jiangmeng Jiaxing Shenzhen Zhaoqing Zhuhai

Grain production decrease (104 ton)


1,200


12 of 16 12 of 15

Figure 11. agricultural land loss and grain production 11. Relationship between urbanization-related urbanization-related agricultural production decrease in in the the typical typical cities. cities. (a) Correlation Correlation relationship; relationship; (b) (b) trend trend relationship. relationship.

5.2. Strategies for Mitigating Urbanization-Related Agricultural Agricultural Land Land Loss Loss There is a pressing need for China to relieve urbanization-related agricultural land loss, but not hinder the growth of the the economy. economy. Key measures to solve current current land land problems problems include include rational rational urban planning and strict agricultural land protection. Specific strategies were suggested as follows: (1) (1) rationally rationally control control the the direction direction and and speed speed of urbanization urbanization to avoid blind urban sprawl; (2) increase efficiency of urban land use and adjust land use structure to optimize optimize the allocation of urban land resources; (3) strictly implement the protection policies on the quantity quantity and quality quality of of agricultural agricultural land; and (4) effectively effectively conduct conduct land land development, development, consolidation, consolidation, and and reclamation. reclamation. Due to different areas areas with with great greatdisparities disparitiesofofurban urbanexpansion expansionand andeconomic economic development, development, it it is imperative to implement differentiated strategies to alleviate urbanization-related agricultural is imperative to implement differentiated strategies to alleviate urbanization-related agricultural land land For eastern the eastern region, more attention should paidtotothe the protection protection of of high-quality loss. loss. For the region, more attention should bebe paid agricultural land and improvement efficiency of urban urban land land use. use. Because agricultural land loss was mainly from high-quality agricultural land in the eastern region, it is necessary to strictly protect the remaining remaining high-quality high-quality agricultural agricultural land, land, and and rationally rationally implement implement land land development, development, consolidation, consolidation, and reclamation to supplement the shortage of agricultural land. In addition, the increased efficiency of urban land use could use its limited land resources more effectively, and slow down the rapid urban expansion. expansion.As Asthe the central and western regions are important the important producing areas in central and western regions are the food food producing areas in China, China, more attention should be paid to the strict protection policies on agricultural land, especially more attention should be paid to the strict protection policies on agricultural land, cropland. However, However, because becausethe thecentral centraland andwestern westernregions regionsare areatatananearly earlystage stage urbanization, ofof urbanization, it it would impossible to completely relieve agricultural inshort the short It is probably would bebe impossible to completely relieve agricultural landland loss loss in the term.term. It is probably more more feasible in a short period (1) increase the efficiency of urban landon use onbasis the basis of intensive feasible in a short period to: (1)to: increase the efficiency of urban land use the of intensive land land use;harmonize (2) harmonize agricultural land demands among variousurban urbanorganizations, organizations;and and (3) (3) solve use, (2) agricultural land demands among various conflicts of agricultural agricultural land demand demand among among various various urban urban departments. departments. For the six six metropolitan metropolitan areas already at aa relatively relatively high high development development stage of urbanization, urbanization, rational urban land use is an urgent task to relieve agricultural land loss. For example, example, high-tech high-tech manufacturing, manufacturing, financial, financial, service, and Internet industries with low urban land demands should be vigorously supported, while highly land-demanding land-demanding industries industries should should be be shut shut down down or relocated relocated to other places. Corresponding policies should be made in time to benefit agricultural land protection. For instance, tax and loan preference and financial subsidies subsidies could could be be given given to to companies companies with with low low land landdemands. demands. Meanwhile, Meanwhile, extra taxations efficiency, and poor land taxations should should be imposed on industries with land waste, low land output efficiency, use structure. structure. The six metropolitan areas also need to rationally guide the flow of population to avoid the blind expansion of of urban urban living living space space due due to to aa high high intensity intensity of of population. population. 5.3. Limitations and 5.3. Limitations and Future Future Perspectives Perspectives There are aafew fewlimitations limitations that need to improved be improved in future studies. For example, some There are that need to be in future studies. For example, some natural natural be excluded fromareas, urbanbut areas, NFS method fallsofshort of eliminating featuresfeatures could becould excluded from urban the but NFSthe method falls short eliminating all the

“blooming” natural features because the NSL data have a relatively coarse spatial resolution. The


13 of 16

all the “blooming” natural features because the NSL data have a relatively coarse spatial resolution. The accuracy of mapping agricultural land loss also largely depends on the NLC data. Although the accuracy of the NLC data has been proven, mismatches between the estimated urban areas and NLC data still exist, resulting in some errors when evaluating agricultural land loss caused by urban expansion. We will analyze the relationship between urbanization and agricultural land using time series nighttime light data and NLC data to investigate the direction and trend of urbanization-related agricultural land loss in China. 6. Conclusions Most studies on the spatial pattern of urbanization-related agricultural land loss in China have been conducted at the local scale. Few have considered it in a timely manner, especially at the national scale. Based on the integration of the NSL, NDVI, water body, and NLC data, this study examined urban expansion and agricultural land loss in China at multiple scales from 2001 to 2013. Urban expansion was firstly mapped using the NSL, NDVI, and water body data. The urbanization-related agricultural land loss was then evaluated from a national scale to regional and metropolitan scales. The validation results demonstrated that the NFS method could be effective and accurate for mapping urban expansion. The urban expansion was mainly identified in the eastern region, SNH, BTT, and PRD. The widespread urban expansion has caused a total loss of agricultural land of 33,080 km2 in China from 2001 to 2013. At a regional scale, the high level of agricultural land loss was concentrated in the eastern region. At a metropolitan scale, agricultural land loss was mainly found in the SNH, PRD, and BTT. Because significant urbanization-related agricultural land loss has a positive correlation with grain production decrease in some developed areas, there is a pressing need for China to relieve urbanization-related agricultural land loss. The relief strategies for China should mainly focus on rational urban planning and strict agricultural land protection. For the eastern region, more attention should be paid to improving the efficiency of urban land use and enhancing the protection of high quality cultivated land, while the central and western regions should incorporate and harmonize agricultural land demands among various urban departments. Rationalizing urban land use is also an urgent task to relieve agricultural land loss in the six metropolitan areas. Supplementary Materials: The following are available online at www.mdpi.com/2071-1050/8/8/790/s1, Figure S1: Identifying the transition zones and central urban areas from the topographic maps, Figure S2: Extracting the marginal urban areas from mixed maps, Figure S3: Mapping primary and reclassified urban maps. Acknowledgments: This work is supported by the National Natural Science Foundation of China (No. 41471449), the Natural Science Foundation of Shanghai (No. 14ZR1412200), the Innovation Program of Shanghai Municipal Education Commission (No. 15ZZ026), the Fundamental Research Funds for the Central Universities of China, and the China Scholarship Council (No. 201406140007). The authors also wish to thank their colleague Susan Cuddy for her helpful suggestions. Author Contributions: Yun Chen, Bailang Yu, and Jianping Wu conceived and supervised the research topic. Kaifang Shi and Bailang Yu proposed the methods. Kaifang Shi, Linyi Li, Chang Huang, Rui Liu, and Zuoqi Chen processed the data. Kaifang Shi, Yun Chen, and Tingbao Xu analyzed the results and wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2.

Lambin, E.F.; Turner, B.L.; Geist, H.J.; Agbola, S.B.; Angelsen, A.; Bruce, J.W.; Coomes, O.T.; Dirzo, R.; Fischer, G.; Folke, C. The causes of land-use and land-cover change: Moving beyond the myths. Glob. Environ. Chang. 2001, 11, 261–269. [CrossRef] Kraemer, R.; Prishchepov, A.V.; Müller, D.; Kuemmerle, T.; Radeloff, V.C.; Dara, A.; Terekhov, A.; Frühauf, M. Long-term agricultural land-cover change and potential for cropland expansion in the former Virgin Lands area of Kazakhstan. Environ. Res. Lett. 2015. [CrossRef]


3.

4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17.

18.

19. 20. 21.

22.

23.

24.

14 of 16

Gutzler, C.; Helming, K.; Balla, D.; Dannowski, R.; Deumlich, D.; Glemnitz, M.; Knierim, A.; Mirschel, W.; Nendel, C.; Paul, C. Agricultural land use changes—A scenario-based sustainability impact assessment for Brandenburg, Germany. Ecol. Indic. 2015, 48, 505–517. [CrossRef] Barati, A.; Asadi, A.; Kalantari, K.; Azadi, H.; Witlox, F. Agricultural land conversion in Northwest Iran. Int. J. Environ. Res. 2015, 9, 281–290. Del Mar López, T.; Aide, T.M.; Thomlinson, J.R. Urban expansion and the loss of prime agricultural lands in Puerto Rico. Ambio J. Hum. Environ. 2001, 30, 49–54. [CrossRef] Pandey, B.; Seto, K.C. Urbanization and agricultural land loss in India: Comparing satellite estimates with census data. J. Environ. Manag. 2015, 148, 53–66. [CrossRef] [PubMed] Pribadi, D.O.; Pauleit, S. The dynamics of peri-urban agriculture during rapid urbanization of Jabodetabek Metropolitan Area. Land Use Pol. 2015, 48, 13–24. [CrossRef] Thapa, R.B.; Murayama, Y. Examining spatiotemporal urbanization patterns in Kathmandu Valley, Nepal: Remote sensing and spatial metrics approaches. Remote Sens. 2009, 1, 534–556. [CrossRef] Zhong, T.; Huang, X.; Zhang, X.; Wang, K. Temporal and spatial variability of agricultural land loss in relation to policy and accessibility in a low hilly region of southeast China. Land Use Pol. 2011, 28, 762–769. Richards, P.D.; Walker, R.T.; Arima, E.Y. Spatially complex land change: The Indirect effect of Brazil’s agricultural sector on land use in Amazonia. Glob. Environ. Chang. 2014, 29, 1–9. [CrossRef] [PubMed] Bai, X.; Shi, P.; Liu, Y. Society: Realizing China’s urban dream. Nature 2014, 509, 158–160. [CrossRef] [PubMed] Zhou, L.; Dickinson, R.E.; Tian, Y.; Fang, J.; Li, Q.; Kaufmann, R.K.; Tucker, C.J.; Myneni, R.B. Evidence for a significant urbanization effect on climate in China. Proc. Natl. Acad. Sci. USA 2004, 101, 9540–9544. [CrossRef] [PubMed] He, C.; Liu, Z.; Tian, J.; Ma, Q. Urban expansion dynamics and natural habitat loss in China: A multiscale landscape perspective. Glob. Chang. Biol. 2014, 20, 2886–2902. [CrossRef] [PubMed] Jiang, L.; Deng, X.; Seto, K.C. The impact of urban expansion on agricultural land use intensity in China. Land Use Pol. 2013, 35, 33–39. [CrossRef] Deng, X.; Huang, J.; Rozelle, S.; Zhang, J.; Li, Z. Impact of urbanization on cultivated land changes in China. Land Use Pol. 2015, 45, 1–7. [CrossRef] Tan, M.; Li, X.; Xie, H.; Lu, C. Urban land expansion and arable land loss in China—A case study of Beijing-Tianjin-Hebei region. Land Use Pol. 2005, 22, 187–196. [CrossRef] Liang, C.; Penghui, J.; Manchun, L.; Liyan, W.; Yuan, G.; Yuzhe, P.; Nan, X.; Yuewei, D.; Qiuhao, H. Farmland protection policies and rapid urbanization in China: A case study for Changzhou City. Land Use Pol. 2015, 48, 552–566. [CrossRef] Ma, L.; Guo, J.; Velthof, G.L.; Li, Y.; Chen, Q.; Ma, W.; Oenema, O.; Zhang, F. Impacts of urban expansion on nitrogen and phosphorus flows in the food system of Beijing from 1978 to 2008. Glob. Environ. Chang. 2014, 28, 192–204. [CrossRef] Fazal, S. Urban expansion and loss of agricultural land—A GIS based study of Saharanpur City, India. Environ. Urban. 2000, 12, 133–149. [CrossRef] Cao, X.; Chen, J.; Imura, H.; Higashi, O. A SVM-based method to extract urban areas from DMSP-OLS and SPOT VGT data. Remote Sens. Environ. 2009, 113, 2205–2209. [CrossRef] Ma, Q.; He, C.; Wu, J.; Liu, Z.; Zhang, Q.; Sun, Z. Quantifying spatiotemporal patterns of urban impervious surfaces in China: An improved assessment using nighttime light data. Landsc. Urban Plan. 2014, 130, 36–49. [CrossRef] Liu, J.; Liu, M.; Tian, H.; Zhuang, D.; Zhang, Z.; Zhang, W.; Tang, X.; Deng, X. Spatial and temporal patterns of China’s cropland during 1990–2000: An analysis based on Landsat TM data. Remote Sens. Environ. 2005, 98, 442–456. [CrossRef] Liu, J.; Zhang, Z.; Xu, X.; Kuang, W.; Zhou, W.; Zhang, S.; Li, R.; Yan, C.; Yu, D.; Wu, S. Spatial patterns and driving forces of land use change in China during the early 21st century. J. Geogr. Sci. 2010, 20, 483–494. [CrossRef] Shi, K.; Chen, Y.; Yu, B.; Xu, T.; Chen, Z.; Liu, R.; Li, L.; Wu, J. Modeling spatiotemporal CO2 (carbon dioxide) emission dynamics in China from DMSP-OLS nighttime stable light data using panel data analysis. Appl. Energy 2016, 168, 523–533. [CrossRef]


25.

26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

36.

37.

38.

39. 40.

41. 42.

43. 44. 45. 46. 47.

15 of 16

Xie, Y.; Weng, Q. Detecting urban-scale dynamics of electricity consumption at Chinese cities using time-series DMSP-OLS (Defense Meteorological Satellite Program-Operational Linescan System) nighttime light imageries. Energy 2016, 100, 177–189. [CrossRef] Elvidge, C.D.; Ziskin, D.; Baugh, K.E.; Tuttle, B.T.; Ghosh, T.; Pack, D.W.; Erwin, E.H.; Zhizhin, M. A fifteen year record of global natural gas flaring derived from satellite data. Energies 2009, 2, 595–622. [CrossRef] Wu, J.; He, S.; Peng, J.; Li, W.; Zhong, X. Intercalibration of DMSP-OLS night-time light data by the invariant region method. Int. J. Remote Sens. 2013, 34, 7356–7368. [CrossRef] Liu, Z.; He, C.; Zhang, Q.; Huang, Q.; Yang, Y. Extracting the dynamics of urban expansion in China using DMSP-OLS nighttime light data from 1992 to 2008. Landsc. Urban Plan. 2012, 106, 62–72. [CrossRef] Pandey, B.; Joshi, P.; Seto, K.C. Monitoring urbanization dynamics in India using DMSP/OLS night time lights and SPOT-VGT data. Int. J. Appl. Earth Observ. Geoinf. 2013, 23, 49–61. [CrossRef] Shi, K.; Huang, C.; Yu, B.; Yin, B.; Huang, Y.; Wu, J. Evaluation of NPP-VIIRS night-time light composite data for extracting built-up urban areas. Remote Sens. Lett. 2014, 5, 358–366. [CrossRef] Elvidge, C.D.; Baugh, K.E.; Dietz, J.B.; Bland, T.; Sutton, P.C.; Kroehl, H.W. Radiance calibration of DMSP-OLS low-light imaging data of human settlements. Remote Sens. Environ. 1999, 68, 77–88. [CrossRef] Small, C.; Pozzi, F.; Elvidge, C.D. Spatial analysis of global urban extent from DMSP-OLS night lights. Remote Sens. Environ. 2005, 96, 277–291. [CrossRef] Elvidge, C.D.; Tuttle, B.T.; Sutton, P.C.; Baugh, K.E.; Howard, A.T.; Milesi, C.; Bhaduri, B.; Nemani, R. Global distribution and density of constructed impervious surfaces. Sensors 2007, 7, 1962–1979. [CrossRef] Sutton, P.C. A scale-adjusted measure of “urban sprawl” using nighttime satellite imagery. Remote Sens. Environ. 2003, 86, 353–369. [CrossRef] Su, Y.; Chen, X.; Wang, C.; Zhang, H.; Liao, J.; Ye, Y.; Wang, C. A new method for extracting built-up urban areas using DMSP-OLS nighttime stable lights: A case study in the Pearl River Delta, southern China. GISci. Remote Sens. 2015, 52, 218–238. [CrossRef] Su, Y.; Chen, X.; Ye, Y.; Wu, Q.; Zhang, H.O.; Huang, N.; Kuang, Y. The characteristics and mechanisms of carbon emissions from energy consumption in China using DMSP/OLS night light imageries. Acta Geogr. Sin. 2013, 68, 1513–1526. Yang, Y.; He, C.; Zhang, Q.; Han, L.; Du, S. Timely and accurate national-scale mapping of urban land in China using Defense Meteorological Satellite Program’s Operational Linescan System nighttime stable light data. J. Appl. Remote Sens. 2013, 7, 1–18. [CrossRef] Huang, Q.; He, C.; Gao, B.; Yang, Y.; Liu, Z.; Zhao, Y.; Dou, Y. Detecting the 20 year city-size dynamics in China with a rank clock approach and DMSP/OLS nighttime data. Landsc. Urban Plan. 2015, 137, 138–148. [CrossRef] Zhang, Q.; Seto, K.C. Mapping urbanization dynamics at regional and global scales using multi-temporal DMSP/OLS nighttime light data. Remote Sens. Environ. 2011, 115, 2320–2329. [CrossRef] The Land Use/Cover Classification Scheme (2006–2020) Developed by the Outline of the National Overall Planning on Land Use. Available online: http://www.mlr.gov.cn/xwdt/jrxw/200810/t20081024_111040. htm (accessed on 10 August 2016). (In Chinese) Ran, Y.; Li, X.; Lu, L. Evaluation of four remote sensing based land cover products over China. Int. J. Remote Sens. 2010, 31, 391–401. [CrossRef] Angel, S.; Sheppard, S.; Civco, D.L.; Buckley, R.; Chabaeva, A.; Gitlin, L.; Kraley, A.; Parent, J.; Perlin, M. The Dynamics of Global Urban Expansion; World Bank, Transport and Urban Development Department: Washington, DC, USA, 2005. Dadi, D.; Azadi, H.; Senbeta, F.; Abebe, K.; Taheri, F.; Stellmacher, T. Urban sprawl and its impacts on land use change in Central Ethiopia. Urban For. Urban Green. 2016, 16, 132–141. [CrossRef] Sharma, R.; Joshi, P. Mapping environmental impacts of rapid urbanization in the National Capital Region of India using remote sensing inputs. Urban Clim. 2016, 15, 70–82. [CrossRef] Dewan, A.M.; Yamaguchi, Y. Land use and land cover change in Greater Dhaka, Bangladesh: Using remote sensing to promote sustainable urbanization. Appl. Geogr. 2009, 29, 390–401. [CrossRef] Chen, J. Rapid urbanization in China: A real challenge to soil protection and food security. Catena 2007, 69, 1–15. [CrossRef] Qin, X.; Zhang, F.; Liu, C.; Yu, H.; Cao, B.; Tian, S.; Liao, Y.; Siddique, K.H. Wheat yield improvements in China: Past trends and future directions. Field Crop. Res. 2015, 177, 117–124. [CrossRef]


48. 49.

16 of 16

Yan, T.; Wang, J.; Huang, J. Urbanization, agricultural water use, and regional and national crop production in China. Ecol. Model. 2015, 318, 226–235. [CrossRef] Yi, F.; Sun, D.; Zhou, Y. Grain subsidy, liquidity constraints and food security—Impact of the grain subsidy program on the grain-sown areas in China. Food Policy 2015, 50, 114–124. [CrossRef] © 2016 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/).