Sciences in Cold and Arid Regions  2015, 7 (6): 722-729 PDF

#### Article Information

JinChang Li, HaiXia Liu, Yong Liu, ZhiZhu Su, ZiQiang Du. 2015.
Land use and land cover change processes in China's eastern Loess Plateau
Sciences in Cold and Arid Regions, 7(6): 722-729
http://dx.doi.org/10.3724/SP.J.1226.2015.00722

### Article History

Accepted: February 12, 2015
Land use and land cover change processes in China's eastern Loess Plateau
JinChang Li1 , HaiXia Liu1, Yong Liu1, ZhiZhu Su2, ZiQiang Du1
1. Institute of Loess Plateau, Shanxi University, Taiyuan, Shanxi 030006, China;
2. Historical Culture School, Shanxi University, Taiyuan, Shanxi 030006, China
Abstract: Using Landsat remote sensing images, we analyzed changes in each land use type and transitions among different land use types during land use and land cover change (LUCC) in Ningwu County, located in the eastern Loess Plateau of China, from 1990 to 2010. We found that grassland, woodland, and farmland were the main land use types in the study area, and the area of each type changed slightly from 1990 to 2010, whereas the area of water, construction land, and unused land increased greatly. For the whole area, the net change and total change were insignificant due to weak human activity intensity in most of the study area, and the LUCC was dominated by quasi-balanced two-way transitions from 1990 to 2010. The insignificant overall amount of LUCC appears to have resulted from offsetting of rapid increases in population, economic growth, and the implementation of a program to return farmland to woodland and grassland in 2000. This program converted more farmland into woodland and grassland from 2000 to 2010 than from 1990 to 2000, but reclamation of woodland and grassland for use as farmland continued from 2000 to 2010, and is a cause for concern to the local government.
Key words: land use and land cover change     remote sensing     forest restoration     Ningwu County     Loess Plateau

1 Introduction

In the last few decades, dramatic conversions among various Land use patterns have occurred due to population growth, food scarcity, urbanization, and shifts in regional Land use policy(Luo et al.,2008; Wang et al.,2008; Bandeira et al.,2011; Huang et al.,2014). These changes have had major impacts on regional water cycles, biogeochemical cycles, biological diversity, terrestrial ecosystem productivity, and climate change(Foley et al.,2005; Sun et al.,2009). As a result, Land use and Land cover change(LUCC)has become a major focus of global change research since the 1990s(Lambin et al.,2001; Reynolds et al.,2007), particularly in the developing country of China(Zhou et al.,2011; Liu et al.,2014). Numerous studies have been performed to underst and the current status, trends, driving forces, and environmental impacts of LUCC(Geist and Lambin, 2002; Munroe and Müller, 2007; Overmars et al.,2007; Hasselmann et al.,2010; Skaloš et al., 2011). The goal is to ultimately achieve sustainable Land utilization based on the insights provided by these studies(Turner et al.,1995; Lambin, 2002).

Since the Chinese government initiated a program to return farmLand to woodLand and grassLand in the Loess Plateau in 2000, tremendous LUCC has occurred in the regions such as northern Shaanxi Province and eastern Gansu Province(Chen et al.,2003; Zhang et al.,2007; Fu et al.,2010). The increase in grassLand and woodLand cover has been considered an important beneficial result of the project(Ma et al.,2002; Liu and Deng, 2010; Chen et al.,2011). However, accelerated industrialization and urbanization as a result of economic growth and population increases have also greatly affected LUCC through the conversion of Land to construction Land (Fu et al.,1989; Chen et al.,2001; Qiu and Cao, 2011). Shanxi Province lies in the eastern part of the Loess Plateau, in the middle reaches of the Yellow River and the upper reaches of the Haihe River. The woodLand and grassLand restoration project in this region has been supported financially by the Chinese government, but little research about the resulting LUCC processes has been conducted.

In the present research, we used Ningwu County of China's Shanxi Province as a case study to explore the LUCC process during the last two decades by integrating remote sensing and GIS techniques. Our overall objectives were to underst and the trends, conversion processes, and driving forces responsible for LUCC and to evaluate the effects of woodLand and grassLand restoration since 2000 in the area.

2 Materials and methods 2.1 Study area

Ningwu County is located in northern Shanxi Province, from 111°50′E to 112°37′E and from 38°31′N to 39°09′N. The county covers about 1, 941 km2, with average elevation of 2, 000 m(Figure 1). The elevation increases from the middle of the study area to the southeast and northwest, and the lowest elevations extend from southwest to northeast through the middle of the study area. The county is bounded by the Yunzhong Mountains on the east, by the Guanchen and Luya mountains on the west, and by the Lvliang and Hongtao mountains on the southwest and north, respectively.

 Figure 1 Location of the study area and topographic map

The study area is located in a semiarid region of northern China, and is dominated by a continental monsoon climate. Average annual temperature is about 6.2 °C, with a minimum mean monthly temperature of −9.9 °C in January and a maximum of 20.1 °C in July. Average annual precipitation is about 600 mm, about 80% of which occurs during the growing season from June to September. Average annual pan evaporation rate is about 1, 700 mm. There are two growing seasons annually, and the main crops are wheat, millet, broom corn millet, and potatoes.

2.2 Data sources and processing

We used three sets of remote sensing images in this study, including Land sat ETM+ images acquired on July 24, 2000 and Land sat TM images acquired on August 22, 1990 and July 12, 2010, all with a 30-m spatial resolution. In addition to these data, we used topographic maps at a scale of 1:50, 000 surveyed by the Chinese Mapping Agency in the early 1980s, as well as vegetation and soil maps provided by several Chinese government agencies for ancillary materials to assist our classification of the Land use types during the visual interpretation process.

We performed geometric corrections using version 9.1 of the ERDAS Imagine software(ERDAS, Norcross, GA). The ETM+ images acquired in 2000 were georeferenced using ground control points derived from the topographic maps. The mean positional errors for georectification of the TM and ETM+ images were controlled to less than 1.5 pixels in mountainous regions and less than 1 pixel in flat regions, which are acceptable levels of precision for large-scale surveys. The TM images from 1990 and 2010 were matched with the geometrically corrected ETM+ images from 2000 by means of an image-to-image matching method provided by the ERDAS Imagine software.

To accurately classify the Land cover, we investigated the study area at the beginning of our research by taking a series of photographs of each type of Land cover and recording each photograph's geographic coordinates using a global positioning system(GPS)receiver. We then found the corresponding Land cover in the remote sensing images at those geographic coordinates to build an interpretation symbol database for use in subsequent image interpretation. In addition, we digitized the topographic maps at a scale of 1:50, 000 to provide vector data, then transformed the vegetation and soil maps to the same projection mode and coordinate system so that they could be overlaid on the other images to provide assistance with image interpretation.

The polygons with different Land use types in the images acquired in 2000 were labeled according to their cover class. Once the digital map of the Land coverage in 2000 was complete, the polygons were copied and the segments that needed modification were changed based on the 2010 and 1990 images; they were updated by adding, deleting, or modifying lines in order to reflect LUCC from 1990 to 2000 and from 2000 to 2010. The visual interpretation process was completed using version 9.1 of the ArcMap software(ESRI, RedLand s, CA).

Using the resource and environment database established by the Chinese Academy of Sciences, we classified the Land use types in Ningwu County into six classes: farmLand , woodLand , grassLand , water body, construction Land , and unused Land .Descriptions of these Land use types are presented in Table 1, and their distribution in each of the three years is shown Figure 2. The interpretation results for the study area were validated in the field in August 2011. Subsequent corrections were made after field validation to ensure the classification accuracy was better than 95%.

Table 1 Land use classification scheme used in this study
 No. Classification Description 1 FarmLand Areas cultivated with dense annual crops and vegetables, including dry Land and irrigated Land 2 WoodLand Areas of dense forest, open forest, orchards, and nurseries 3 GrassLand Land with natural grassLand cover, including steppes and grazing Land s 4 Water body All areas of open water, including rivers, streams, lakes, and reservoirs or ponds 5 Construction Land All Land used to construct human structures, including residential, commercial, and industrial buildings as well as transportation facilities, highways, railways, and single- and multiple-family houses 6 Unused Land Uncultivated areas with sparse plant cover, including saline or alkali Land along river lake beaches, barren rocky or s and y Land in sloping fields, bare Land , and cultivated Land without crops at the time the image was acquired(e.g., fallow or ab and oned Land )
 Figure 2 Land use and Land cover maps for Ningwu County in 1990, 2000, and 2010

To define the LUCC that occurred in Ningwu County from 1990 to 2000, from 2000 to 2010, and from 1990 to 2010, we used the Overlay Tool in the ArcInfo 8.3 GIS software(ESRI)to compute the geometric intersection of areas of each Land use type in the three periods based on the Land use maps from each year. The output coverage file was converted into shapefile format. The cross-tabulation table was calculated using the PivotTable Wizard in Microsoft Excel, and was output as a transition matrix. The magnitudes and directions of LUCC were determined based on the transition matrix.

2.3 Methodology for identifying the overall status of LUCC

In the present study, we applied the model developed by Luo et al.(2008)to characterize the net changes and total changes in Land use types, and the resulting overall LUCC status in Ningwu County. Net change was defined as the maximum gain or loss of area minus the minimum gain or loss of area for each Land use type. Total change equals the sum of the gains and losses of area for each Land use type. The net change(Snet) and total change(Stot)are defined as follows:

 $\begin{array}{l} {S_{{\rm{net}}}} = \frac{{\sum\nolimits_{i = 1}^n {\left| {{U_{bi}} - {U_{ai}}} \right|} }}{{2\sum\nolimits_{i = 1}^n {{U_{ai}}} }} \times 100\% \\ \quad \;\,= \frac{{\sum\nolimits_{i = 1}^n {\left| {\Delta {U_{{\rm{in}} - i}} - \Delta {U_{{\rm{out}} - i}}} \right|} }}{{2\sum\nolimits_{i = 1}^n {{U_{ai}}} }} \times 100\% \end{array}$ (1)

 $\begin{array}{l} {S_{{\rm{tot}}}} = \frac{{\sum\nolimits_{i = 1}^n {(\Delta {U_{{\rm{in}} - i}} + \Delta {U_{{\rm{out}} - i}})} }}{{2\sum\nolimits_{i = 1}^n {{U_{ai}}} }} \times 100\% \\ \quad \;\,= \frac{{\sum\nolimits_{i = 1}^n {\Delta {U_{{\rm{out}} - i}}} }}{{\sum\nolimits_{i = 1}^n {{U_{ai}}} }} \times 100\% \\ \quad \;\,= \frac{{\sum\nolimits_{i = 1}^n {\Delta {U_{{\rm{in}} - i}}} }}{{\sum\nolimits_{i = 1}^n {{U_{ai}}} }} \times 100\% \end{array}$ (2)

where Uai and Ubi denote the area of Land use type i in the initiaLand final stages of a period, respectively; ΔUouti is the sum of the areas of Land use type i converted into other Land use types; ΔUini is the sum of the areas of other Land use types converted into Land use type i; and n is the number of Land use types(i.e., n = 6 in the present study). Using Equations (1) and (2), the index Ptot can be defined to characterize the total LUCC status for the study area:

 ${P_{{\rm{tot}}}} = \frac{{{S_{{\rm{net}}}}}}{{{S_{{\rm{tot}}}}}} = \frac{{\sum\nolimits_{i = 1}^n {\left| {\Delta {U_{{\rm{in}} - i}} - \Delta {U_{{\rm{out}} - i}}} \right|} }}{{\sum\nolimits_{i = 1}^n {(\Delta {U_{{\rm{in}} - i}} + \Delta {U_{{\rm{out}} - i}})} }}$(Stot ≠ 0, 0 ≤ Ptot ≤ 1) (3)

For the convenience of discussion and application of this index system, Luo et al.(2008)divided Ptot arbitrarily into four categories:

1)0≤ Ptot ≤0.25: LUCC exhibits balanced two-way transitions, and for most Land use types in the study area, the area gained is equal to or differs only slightly from the area lost.

2)0.25< Ptot ≤0.50: LUCC is quasi-balanced, and for most Land use types, the area gained is clearly larger than or less than the area lost.

3)0.50< Ptot ≤0.75: LUCC is dominated by one-way transitions, with an unbalanced status, and for most Land use types, the area gained is much greater or much less than the area lost.

4)0.75< Ptot ≤1: LUCC is characterized by extremely unbalanced one-way transitions, in which some Land use types increase in area with little or no loss, whereas others decrease in area, with little or no gain.

3 Results 3.1 The characteristics of LUCC

Table 2 shows thatgrassLand , woodLand , and farmLand were the main Land use types in Ningwu County in all three years; moreover, they accounted for about 97% of the total Land , and their total areas change only slightly(by less than 8%)during both periods(1990-2000 and 2000-2010). However, the total areas of bodies of water, construction Land , and unused Land all increased greatly(by 3% to 120%)during the two periods. From 1990 to 2000, the areas of grassLand and unused Land decreased by 2.6% and 5.4%, respectively. In contrast, the area of farmLand increased by 2.1%, versus increases of 1.6% for woodLand , 10.1% for bodies of water, and 3.2% for construction Land during the same period. From 2000 to 2010, the areas of farmLand decreased greatly(by 7.8%), whereas woodLand and grassLand changed little(by less than 1%), and the areas of bodies of water, construction Land , and unused Land increased greatly(each by more than 69.0%). From 1990 to 2010, the area of farmLand and grassLand had decreased by 5.9% and 2.8%, respectively, whereas the area of woodLand increased by 1.3%. The largest increases were for bodies of water(119.8%), construction Land (74.4%), and unused Land (71.4%). Figure 2 shows that the LUCC of the study area occurred mainly in the northeast and southwest.

Table 2 The areas of each Land use type and their proportional change during three periods in Ningwu County
 Land use type Area(ha) Change(%) 1990 2000 2010 1990-2000 2000-2010 1990-2010 FarmLand 39, 446.83 40, 282.42 37, 127.25 2.12 −7.83 −5.88 WoodLand 70, 652.14 71, 762.39 71, 595.51 1.57 −0.23 1.34 GrassLand 80, 014.30 77, 905.78 77, 765.24 −2.64 −0.18 −2.81 Water body 1, 425.48 1, 569.73 3, 133.69 10.12 99.63 119.83 Construction Land 1, 847.50 1, 906.58 3, 222.21 3.20 69.00 74.41 Unused Land 760.14 719.47 1, 302.49 −5.35 81.03 71.35 Total 194, 146.38 194, 146.38 194, 146.38 0.00 0.00 0.00

Table 3 shows the transition matrix for LUCC in Ningwu County during the study period. To a large extent, LUCC from 1990 to 2010 was dominated by mutual conversions between grassLand , woodLand and farmLand , which accounted for 68.1% of the total conversions.

Table 3 Transition matrix for LUCC in Ningwu County from 1990 to 2000, 2000 to 2010, and 1990 to 2010(unit: ha)
 (a)1990-2000 FarmLand WoodLand GrassLand Water body Construction Land Unused Land FarmLand 38, 138.30 204.25 949.74 125.65 18.96 9.93 WoodLand 121.93 69, 077.35 1, 241.44 178.45 32.98 0.00 GrassLand 1, 911.96 2, 469.39 75, 395.88 204.43 32.64 0.00 Water Body 104.28 0.83 268.53 1, 051.83 0.00 0.00 Construction Land 5.96 0.00 10.77 8.76 1, 822.01 0.00 Unused Land 0.00 10.58 39.41 0.61 0.00 709.54 Net change from 1990 to 2000 835.59 1, 110.25 −2, 108.52 144.25 59.08 −40.67 (b)2000−2010 FarmLand WoodLand GrassLand Water body Construction Land Unused Land FarmLand 35, 822.56 407.63 2, 985.84 137.08 770.83 158.50 WoodLand 102.58 68, 053.06 2, 825.46 606.98 63.02 111.29 GrassLand 929.71 3, 083.87 71, 715.07 1, 351.50 418.53 407.10 Water Body 241.77 2.91 185.69 1, 035.02 104.34 0.00 Construction Land 30.64 27.35 44.01 3.11 1, 801.48 0.00 Unused Land 0.00 20.70 9.16 0.00 64.01 625.60 Net change from 2000 to 2010 −3, 155.17 −166.88 −140.54 1, 563.96 1, 315.63 583.02 (c)1990−2010 FarmLand WoodLand GrassLand Water body Construction Land Unused Land FarmLand 34, 949.51 439.96 2, 863.89 222.70 815.57 155.20 WoodLand 157.22 67, 023.56 2, 460.00 854.07 62.55 94.74 GrassLand 1, 682.28 4, 095.23 71, 988.55 1, 423.02 411.74 413.48 Water Body 305.87 2.85 367.41 625.47 100.47 23.40 Construction Land 32.36 2.64 36.82 7.81 1, 767.87 0.00 Unused Land 0.00 31.27 48.57 0.61 64.01 615.67 Net change from 1990 to 2010 −2, 319.58 943.37 −2, 249.06 1, 708.21 1, 374.71 542.35 Transitions represent transfers of area from the Land use type in the first column to the Land use types across the top of the table. Net positive and negative values represent an increase or decrease(respectively)in the area of the Land use type during the specified period.
3.2 The overall status of LUCC

Based on the change in Table 3 and Equation (3), we analyzed the net and total changes for each Land use type, and calculated the overall status of LUCC(Table 4). The net and total changes were insignificant(less than 2.5% and 8.9%, respectively)during the entire study period and during both sub-periods. The second period had a greater net change and total change than the first period. The values of Snet and Stot for the whole study period(from 1990 to 2010)were higher than the corresponding values in either of the two sub-periods(1990 to 2000 and 2000 to 2010)because longer duration of period resulted in higher sum of the changes.On the other h and , the values of Snet and Stot for the entire period were lower than the corresponding sums for the two sub-periods, probably because the LUCC processes returned some of the altered Land use types to their original Land use types(Luo et al.,2008).

Table 4 Net and total changes in Land use, and overall status of LUCC(based on the Ptot index)in Ningwu County from 1990 to 2010
 Item 1990-2000 2000-2010 1990-2010 Net change(Snet, %) 1.11 1.78 2.35 Total change(Stot, %) 4.10 7.77 8.85 Ptot 0.27 0.23 0.27

The Ptot index was 0.27 from 1990 to 2000 and from 1990 to 2010, suggesting that for most Land use types, the area gained was clearly larger or less than the area of lost, and that the overall status of LUCC was dominated by quasi-balanced two-way transitions. From 2000 to 2010, the Ptot index decreased slightly(to 0.23), suggesting that during the most recent period, most Land use types showed little or no difference in the areas gained and lost; the overall status of LUCC was therefore dominated by balanced two-way transitions.

4 Discussion

Population and economic growth, combined with shifts in Land use policy, have long been considered the major factors responsible for LUCC(e.g., Lin and Ho, 2003; Tanrivermis, 2003; Long et al.,2007; Luo et al.,2008; Liu et al.,2009). The population of Ningwu County was 135.73 thous and in 1990, 152.94 thous and in 2000, and 161.16 thous and in 2010, representing an increase of 19% over the course of the study period, with a slower increase during the second period(5%)than during the first period(13%). Simultaneously, the GDP increased from RMB 0.13 billion in 1990 to RMB 0.32 billion in 2000(an increase of 146%), and then increased dramatically to RMB 2.26 billion in 2010(an increase of 606%). The increase in the area of farmLand from 1990 to 2000 may have mainly resulted from the rapid population increase, which led to a requirement for more farmLand to feed the larger population. However, the decrease in farmLand from 2000 to 2010 may have resulted mainly from implementing of the woodLand and grassLand restoration project(began in 2000), which converted more farmLand into woodLand and grassLand than occurred from 1990 to 2000(Table 3). A substantial area of woodLand and grassLand were nonetheless converted into farmLand from 2000 to 2010(Table 3), which indicated that woodLand and grassLand were still being reclaimed for agriculture in parts of the study area during this period.

WoodLand and grassLand decreased slightly from 2000 to 2010 despite the woodLand and grassLand restoration project implemented in 2000. This was mainly caused by the expansion of construction Land , bodies of water, and unused Land . With the population and economy increases, the construction of residential, commercial, and industrial buildings as well as transportation facilities, highways, and railways during the study period resulted in continuous expansion of construction Land at the expense of farmLand and grassLand . Large amounts of farmLand , grassLand , and woodLand were turned into man-made lowLand s as a result of soil excavation combined with rapid expansion of construction Land . Because precipitation is highest during the summer(from June to August), when the Land sat images were acquired, some of these lowLand s might have been filled temporarily with water and therefore classified as bodies of water in the Land use map. This is likely to be the major factor that contributed to the increased area of bodies of water(Sun et al.,2009). The loss of bodies of water to farmLand , construction Land , and unused Land was higher than the gain in bodies of water from these competing Land use during the study period, and this can be attributed to the encroachment of farmLand and construction Land on bodies of water, accompanied by increased consumption of industriaLand domestic water as a result of socioeconomic development. GrassLand s that degenerated into bare Land as a result of overgrazing and farmLand that was ab and oned may be the main reasons for the increased area of unused Land during the study period.

The population density(82 per km2) and per capita GDP(￥14, 023 CNY)in the study area were lower than the corresponding values for Shanxi Province as a whole(228 per km2 and ￥26, 000 CNY, respectively) and for China as a whole(140 per km2 and ￥29, 000 CNY, respectively)in 2010, which suggests that the intensity of human activity was weaker in the study area than in the province and China as a whole. The most intense human activity was concentrated within the northeast and southwest which have lower elevations and flatter terrain than other areas. Despite the rapid increase in GDP(more than 600%)between 2000 and 2010, most of this economic activity was provided by activities with high value-added(e.g., industrial production)within an existing area or new increased small area of construction Land , thereby leading to relatively minor impacts on most of the study area despite the large increase in GDP. The insignificant net and total changes for the total Land use types in the study area(Table 4)may have been caused mainly by the weak intensity of human activity in most of the study area. Although woodLand and grassLand restoration was only implemented in part of the area, these changes combined with rapid expansion of construction Land , bodies of water, and unused Land since 2000 to produce more significant net and total changes from 2000 to 2010 than from 1990 to 2000(Table 4).

5 Conclusions

In this study, we described the LUCC processes in Ningwu County, in the eastern part of China's Loess Plateau, using multi-temporal remote sensing images. We found that the net and total changes were insignificant in the study area due to the weak intensity of human activity and the concentration of the most intense human socioeconomic activity within a relatively small portion of the overall study area. The LUCC patterns mainly included mutual conversions between farmLand , woodLand , and grassLand , and gains in the areas of bodies of water, construction Land , and unused Land as a result of transitions from farmLand , woodLand , and grassLand .

The rapid population increase from 1990 to 2000 resulted in a large reclamation of grassLand and unused Land to produce the farmLand that was needed to feed this population. The combination of rapid population and economic increases resulted in large proportional increases in the area of construction Land (including residential, commercial, and industrial buildings as well as transportation facilities, highways, and railways)at the expense offarmLand and grassLand throughout the study period; However, the actual increase in area was relatively small, so the area of construction Land still represented a small proportion of the county's total area. The insignificant overall amount of LUCC appears to have resulted from offsetting of rapid increases in population, economic growth, and the implementation of a program to return farmLand to woodLand and grassLand in 2000. This program converted more farmLand into woodLand and grassLand from 2000 to 2010 than from 1990 to 2000, but reclamation of woodLand and grassLand for use as farmLand continued from 2000 to 2010, and is a cause for concern to the local government.

Acknowledgments:

This study was supported by the Open Fund Project of the Key Laboratory of Desert and Desertification, Chinese Academy of Sciences(No. KLDD-2014-001), the Important Specialized Science and Technology Item of Shanxi Province, China(No. 20121101011), and the Natural Science Foundation of China(Nos. 41271513, 41271030).

References
 Bandeira JM, Coelho MC, Sá ME, et al., 2011. Impact of land use on urban mobility patterns, emissions and air quality in a Por-tuguese medium-sized city. Science of the Total Environment, 409(6): 1154–1163. DOI: 10.1016/j.scitotenv.2010.12.008. Chen FJ, Zhang XP, Fu YL, et al., 2011. Land use change of Majiagou Catchment in Ansai County, Northern Shaanxi Province in recent 20 years. Bulletin of Soil and Water Conservation, 31(1): 244–248. Chen LD, Messing I, Zhang SR, et al., 2003. Land use evaluation and scenario analysis towards sustainable planning on the Loess Plateau in China—case study in a small catchment. Catena, 54(1–2): 303–316. DOI: 10.1016/S0341-8162(03)00071-7. Chen LD, Wang J, Fu BJ, et al., 2001. Land-use change in a small catchment of northern Loess Plateau, China. Agriculture, Ecosystems & Environment, 86(2): 163–172. DOI: 10.1016/S0167-8809(00)00271-1. Foley JA, DeFries R, Asner GP, et al., 2005. Global consequences of land use. Science, 309(5734): 570–574. DOI: 10.1126/science.1111772. Fu BJ, Chen LD, Ma KM, 1999. The effect of land use change on the regional environment in the Yangjuangou Catchment in the Loess Plateau of China. Acta Geographica Sinica, 54(3): 241–246. Fu YL, Zhang XP, Chen FJ, et al., 2010. Land use/cover pattern survey under the background of grain for green policy implementation in the loess hilly-gully region—A case study of Majiagou Catchment, Ansai County. Research of Soil and Water Conservation, 17(6): 81–85. Geist HJ, Lambin EF, 2002. Proximate causes and underlying driving forces of tropical deforestation. Bioscience, 52: 143–150. DOI: 10.1641/0006-3568(2002)052[0143:PCAUDF]2.0. Hasselmann F, Csaplovics E, Falconer I, et al., 2010. Technological driving forces of LUCC: Conceptualization, quantification, and the example of urban power distribution networks. Land Use Policy, 27(2): 628–637. DOI: 10.1016/j.landusepol.2009.08.016. Huang XJ, Li Y, Yu R, et al., 2014. Reconsidering the controversial land use policy of "linking the decrease in rural construction land with the increase in urban construction land": a local government perspective. China Review—An Interdisciplinary Journal on Greater China, 14: 175–198. Lambin EF, 2002. Linking causes, drivers and pathways with rates and patterns of land change. In: LUCC International Project Office. LUCC Newsletter No. 8. Institut Cartográfic de Cata-lunya, Spain. Lambin EF, Turnerb BL, Geista HJ, et al., 2001. The causes of land-use and land-cover change: moving beyond the myths. Global Environment Change, 11: 261–269. DOI: 10.1016/S0959-3780(01)00007-3. Lin GCS, Ho SPS, 2003. China’s land resources and land-use change: insights from the 1996 land survey. Land Use Policy, 20(2): 87–107. DOI: 10.1016/S0264-8377(03)00007-3. Liu H, Zhao WZ, Chang XL, et al., 2009. Characterizing landscape dynamics of a small catchment under ecological rehabilitation interventions in Northwestern China. Landscape and Urban Planning, 93(3–4): 201–209. DOI: 10.1016/j.landurbplan.2009.07.007. Liu JY, Deng XZ, 2010. Progress of the research methodologies on the temporal and spatial process of LUCC. Chinese Science Bulletin, 55(14): 1354–1362. DOI: 10.1007/s11434-009-0733-y. Liu Y, Huang XJ, Yang H, et al., 2014. Environmental effects of land-use/cover change caused by urbanization and policies in Southwest China Karst area—A case study of Guiyang. Habitat International, 44: 339–348. DOI: 10.1016/j.habitatint.2014.07.009. Long HL, Tang GP, Li XB, et al., 2007. Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China. Journal of Environmental Manage-ment, 83(3): 351–364. DOI: 10.1016/j.jenvman.2006.04.003. Luo GP, Zhou CH, Chen X, et al., 2008. A methodology of characterizing status and trend of land changes in oases: A case study of Sangong River watershed, Xinjiang, China. Journal of Environmental Management, 88(4): 775–783. DOI: 10.1016/j.jenvman.2007.04.003. Ma AQ, Chen DJ, Wang JH, et al., 2002. Landscape pattern and differentiation in Longdong Loess Plateau Based on RS and GIS. Journal of Soil and Water Conservation, 16(3): 56–59. Munroe DK, Müller D, 2007. Issues in spatially explicit statistical land-use/cover change (LUCC) models: Examples from western Honduras and the Central Highlands of Vietnam. Land Use Policy, 24(3): 521–530. DOI: 10.1016/j.landusepol.2005.09.007. Overmars KP, Verburg PH, Veldkamp TA, 2007. Comparison of a deductive and an inductive approach to specify land suitability in a spatially explicit land use model. Land Use Policy, 24(3): 584–599. DOI: 10.1016/j.landusepol.2005.09.008. Qiu HJ, Cao MM, 2011. Land-use change in gully region of Loess Plateau—A case study of Luochuan County, Shaanxi Province. Bulletin of Soil and Water Conservation, 31(2): 207–210 . Reynolds JF, Smith DMS, Lambin EF, et al., 2007. Global deser-tification: building a science for dryland development. Science, 316(5826): 847–851. DOI: 10.1126/science.1131634. Skalo J, Weber M, Lipsky Z, et al., 2011. Using old military survey maps and orthophotograph maps to analyse long-term land cover changes—Case study (Czech Republic). Applied Geo-graphy, 31(2): 426–438. DOI: 10.1016/j.apgeog.2010.10.004. Sun ZY, Ma R, Wang YX, 2009. Using Landsat data to determine land use changes in Datong basin, China. Environmental Ge-ology, 57(8): 1825–1837. DOI: 10.1007/s00254-008-1470-2. Tanrivermis H, 2003. Agricultural land use change and sustainable use of land resources in the mediterranean region of Turkey. Journal of Arid Environments, 54: 553–564. DOI: 10.1006/jare.2002.1078. Turner BL II, Skole D, Sanderson S, et al., 1995. Land use/land cover and land cover change science/research plan. IGBP Re-port No. 35, HDP Report 7, IGBP of the ICSU and HDP of the ISSC, Stockholm and Geneva. Wang XH, Zheng D, Shen YC, 2008. Land use change and its driving forces on the Tibetan Plateau during 1990–2000. Ca-tena, 72(1): 56–66. Zhang XM, Yu XX, Wu SH, et al., 2007. Response of land use/coverage change to hydrological dynamics at watershed scale in the Loess Plateau of China. Acta Ecologica Sinica, 27(2): 414–421. DOI: 10.1016/S1872-2032(07)60013-4. Zhou DY, Zhang LJ, Zhang L, et al., 2011. The effect of landscape park on urban heat island: a case study of Harbin city. Areal Research and Development, 30: 73–78.