EN中文

Journal of Geographical Sciences >

2016 , Vol. 26 >Issue 3: 272 - 296

DOI: https://doi.org/10.1007/s11442-016-1268-1

Orginal Article

Scenario simulation and landscape pattern dynamic changes of land use in the PovertyBelt around Beijing and Tianjin: A case study of Zhangjiakou city, Hebei Province

  • SUN Piling , 1 ,
  • XU Yueping , 1 ,
  • YU Zhonglei 2 ,
  • LIU Qingguo 3 ,
  • XIE Baopeng 1 ,
  • LIU Jia 1
Expand

Author: Sun Piling (1984-), PhD Candidate, specialized in land use/cover change, sustainable utilization of land resources. E-mail:

*Corresponding author: Xu Yueqing (1972-), PhD and Associate Professor, E-mail:

Received date: 2015-04-15

  Accepted date: 2015-10-12

  Online published: 2016-03-20

Supported by

National Natural Science Foundation of China, No.41171088, No.41571087

Copyright

Journal of Geographical Sciences, All Rights Reserved
Fold

Abstract

Land use/cover change has been recognized as a key component in global change and has attracted increasing attention in recent decades. Scenario simulation of land use change is an important issue in the study of land use/cover change, and plays a key role in land use prediction and policy decision. Based on the remote sensing data of Landsat TM images in 1989, 2000 and 2010, scenario simulation and landscape pattern analysis of land use change driven by socio-economic development and ecological protection policies were reported in Zhangjiakou city, a representative area of the Poverty Belt around Beijing and Tianjin. Using a CLUE-S model, along with socio-economic and geographic data, the land use simulation of four scenarios-namely, land use planning scenario, natural development scenario, ecological-oriented scenario and farmland protection scenario-were explored according to the actual conditions of Zhangjiakou city, and the landscape pattern characteristics under different land use scenarios were analyzed. The results revealed the following: (1) Farmland, grassland, water body and unused land decreased significantly during 1989-2010, with a decrease of 11.09%, 2.82%, 18.20% and 31.27%, respectively, while garden land, forestland and construction land increased over the same period, with an increase of 5.71%, 20.91% and 38.54%, respectively. The change rate and intensity of land use improved in general from 1989 to 2010. The integrated dynamic degree of land use increased from 2.21% during 1989-2000 to 3.96% during 2000-2010. (2) Land use changed significantly throughout 1989-2010. The total area that underwent land use change was 4759.14 km2, accounting for 12.53% of the study area. Land use transformation was characterized by grassland to forestland, and by farmland to forestland and grassland. (3) Under the land use planning scenario, farmland, grassland, water body and unused land shrank significantly, while garden land, forestland and construction land increased. Under the natural development scenario, construction land and forestland increased in 2020 compared with 2010, while farmland and unused land decreased. Under the ecological-oriented scenario, forestland increased dramatically, which mainly derived from farmland, grassland and unused land. Under the farmland protection scenario, farmland was well protected and stable, while construction land expansion was restricted. (4) The landscape patterns of the four scenarios in 2020, compared with those in 2010, were more reasonable. Under the land use planning scenario, the landscape pattern tended to be more optimized. The landscape became less fragmented and heterogeneous with the natural development scenarios. However, under the ecological-oriented scenario and farmland protection scenario, landscape was characterized by fragmentation, and spatial heterogeneity of landscape was significant. Spatial differences in landscape patterns in Zhangjiakou city also existed. (5) The spatial distribution of land use could be explained, to a large extent, by the driving factors, and the simulation results tallied with the local situations, which provided useful information for decision-makers and planners to take appropriate land management measures in the area. The application of the combined Markov model, CLUE-S model and landscape metrics in Zhangjiakou city suggests that this methodology has the capacity to reflect the complex changes in land use at a scale of 300 m×300 m and can serve as a useful tool for analyzing complex land use driving factors.

Key words： land use change; Markov model; CLUE-S model; landscape metrics; Zhangjiakou city

Cite this article

SUN Piling , XU Yueping , YU Zhonglei , LIU Qingguo , XIE Baopeng , LIU Jia . Scenario simulation and landscape pattern dynamic changes of land use in the PovertyBelt around Beijing and Tianjin: A case study of Zhangjiakou city, Hebei Province[J]. Journal of Geographical Sciences, 2016 , 26(3) : 272 -296 . DOI: 10.1007/s11442-016-1268-1

1 Introduction

Land use changes result from the interaction between humans and nature (Vitousek et al., 1997; Liao et al., 2011) and involve complex mechanisms and processes (Sohl and Claggett, 2013). Land use and land cover change (LUCC) is a complex process subject to the interactions between natural and social systems on different temporo-spatial scales (Veldkamp and Lambin, 2001). It is a principal component, and main cause, of global environmental changes, and it has emerged recently as an important focus in the study of land use change. At present, the study of LUCC and its effects has increasingly become a core element of global environmental change and sustainable development research (Feng et al., 2013; Xu et al., 2011; Xu et al., 2013b). Increasing attention has been paid to the driving factors involved in such processes, and the effects of LUCC. The simulation of global and regional land use changes has been employed as a new method for studying LUCC (Liu et al., 2002; Couclelis, 2005). Models and scenario analysis are useful tools for understanding land use patterns and the complex mechanisms of socio-economic and physical variables that influence land use change, and have been recognized as excellent tools for simulating land use changes (Li et al., 2001; Britz et al., 2011; Deng et al., 2013). Currently, different approaches and models have been widely used in many researches on LUCC and have yielded good results (Evans et al., 2001; Tang et al., 2009; Zhang et al., 2014b). Some models, especially dynamic models, can support decision makers in simulating future scenarios. For instance, Markov chains, system dynamics (SD) models, cellular automata (CA) models, multi-agent system (MAS) models, and the Conversion of Land Use and its Effects at Small regional extent (CLUE-S) model have frequently been applied to the simulation of spatio-temporal land use changes (Hu et al., 2013; Liang et al., 2011; Ou et al., 2014; Chen et al., 2009; Tian and Ren, 2012; Zeng et al., 2014; Wu et al., 2014). The CLUE-S model has been considered as an excellent tool for simulating land use changes. It can be used to explore the competition between different types of land uses based on systems theory, to simulate different land uses simultaneously, and to produce a spatially explicit display of the simulation results (Overmars and Verburg, 2007; Hu et al., 2013). Recently, the CLUE-S model has been used to simulate land use change in certain hot spots, such as plateau cities, mining cities, Heihe River basin, Huangshui River basin, Luoyugou watershed, metropolitan fringe areas, loess hilly areas, and the Danjiangkou reservoir area (Zhang et al., 2013a; Dai and Zhang, 2013; Feng et al., 2013; Huang et al., 2012; Deng et al., 2013; Li et al., 2011b). In terms of hot spots, it is necessary to pay more attention to the ecologically fragile areas, where land use changes remarkably. Analyzing land use changes with the CLUE-S model in such areas makes a significant contribution to the research into global change.
Land use is a direct indicator of human interaction with the landscape. Land use cover may be defined as the primary landscape on the earth’s surface influenced by both human activities and natural conditions. Landscape patterns may be defined as the quantitative description of the spatial composition and configuration of land use, which have both spatial and temporal characteristics. Landscape pattern indexes are proposed to quantitatively describe landscape patterns. Currently, many researches on landscape patterns are focused on analyzing landscape pattern changes in land use. However, landscape pattern indices have not been included in land use change models (Yang et al., 2014).
Recently, remote sensing (RS) and geographical information system (GIS) have often been combined to assess land use change effectively. RS integrated with GIS provides an excellent framework for data capture, storage, processing and analysis (Wang et al., 2010). In order to analyze land use changes systematically, landscape metrics have been integrated into LUCC models such as the CLUE-S model, Markov-CA (cellular automata) model, Markov-CLUE-S model and SD (system dynamics)-CLUE-S model (Zeng et al., 2014; Wang et al., 2014; Hu and Dong, 2013; Zhang et al., 2014b; Hu et al., 2013; Luo et al., 2010; Zheng et al., 2015). These methods have proved useful in the analysis and prediction of land use changes. The Markov-CA and CLUE-S models integrate landscape pattern indices (LPIs) to accomplish the simulation of land use dynamics, which can provide scientific bases for land use planning and management.
The Poverty Belt around Beijing and Tianjin, located in the north of Hebei Province, bordering with Beijing and Tianjin city, is a typical ecologically fragile area and comprises 31 administrative districts. This belt involves the whole of Chengde city and Zhangjiakou city, and part of Baoding city. Owing to the harsh natural environment, population pressure, and unreasonable human activities such as over logging and steep-slope farming, the Poverty Belt has been experiencing serious soil erosion, desertification of land, natural disasters and environmental deterioration, resulting in a vicious cycle of poverty. With the initiation of the integrated and coordinated development of the Beijing-Tianjin-Hebei region, the society and economy in the Poverty Belt will develop rapidly. The competition for land use among different industries and departments will intensify, and the contradiction between land supply and demand will become more marked. Although there have been many researches into the causes of, and solution to, the poverty of the Poverty Belt around Beijing and Tianjin, most of these have focused on development strategies and patterns relating to the regional economy, ecological compensation mechanisms, environmental deterioration, and the effects of land use changes (Sun et al., 2014). Few researches on land use changes using LUCC models and scenario analysis have been conducted in relation to the Poverty Belt around Beijing and Tianjin. The main problems of this area are closely associated with land use changes, so the effective analysis and simulation of land use changes is critical to rectifying many of the problems relating to the eco-environmental management of the Poverty Belt. There is an urgent need to simulate land use change scenarios driven by socio-economic and ecological protection policies to ensure sustainable land use planning and management in this particularly fragile and poverty-stricken environment.
The principal aims of this study were: (1) to propose an integrated model based on the Markov-CLUE-S model and LPIs; (2) to describe the temporal and spatial characteristics of land use change and identify the primary driving factors; and (3) to predict and simulate the evolution of land use patterns and landscape characteristics under different scenarios.

2 Study area

Zhangjiakou city, located in the northwest of Hebei Province, is a main part of the Poverty Belt and plays a key role in water conservation and ecological security for Hebei and Beijing. The region’s latitude ranges from 39°30′ to 42°10′N, while its longitude ranges from 113°50′ to 116°30′E. It covers an area of 3.68×104 km2 (Figure 1), which includes 13 counties (Kangbao, Zhangbei, Guyuan, Shangyi, Huai’an, Wanquan, Xuanhua, Yuxian, Yangyuan, Zhuolu, Huailai, Chongli and Chicheng) and 4 districts (Qiaoxi, Qiaodong, Xuanhua and Xiahuayuan). Its elevation ranges from 320 m to 2841 m above sea level. The study area is located in the transition zone of the North China Plain and the Inner Mongolia Plateau, with elevations increasing from southeast to northwest. Zhangjiakou city is one of the most important areas for conserving water and providing an ecological defense for Beijing and Tianjin. Moreover, it is also a typical mountainous region with ecosystem vulnerability and sensitivity. It is characterized by a typical temperate continental monsoon climate. The mean annual precipitation is about 350 mm and the annual average temperature is 7.8°C, both of which have clear seasonal variations. However, with the growth in population and economic activities, land use in the study area has changed dramatically over the past two decades. The ongoing land use change shows obvious spatial differentiation characteristics. There is higher terrain in Bashang Plateau with a significantly undulating plateau landscape. While the landscape under the dam is characterized by hills and mountains, which has changed dramatically with complex terrain, Zhangjiakou has undergone rapid urbanization and economic growth since the economic reforms of 1995, and by 2010 the total population of the city had reached 4.35 million according to China’s sixth population census. With per-capita grain output up to 315 kg, the city’s per-capita net income for farmers was 4119 yuan RMB, and its GDP was approximately 966.42 billion yuan then. Therefore, Zhangjiakou city is a typical representative of the Poverty Belt around Beijing and Tianjin.
Figure 1 Location of the study area and its elevation

3 Data and methods

3.1 Data sources and processing

The data used in this study comprised: (1) remote sensing (RS) data, including Landsat Thematic Mapper (TM) data images for 1989, 2000 and 2010; (2) data for soil organic matter (at a resolution of 100 m), derived from the Institute of Remote Sensing Applications, Chinese Academy of Sciences; (3) terrain data, supplied by the national digital geomorphic database at a 1:1,000,000 scale and the geospatial data cloud information platform (http://www. gscloud.cn/; accessed November 2015); (4) socio-economic data, derived from annual statistics of Zhangjiakou city and rural annual statistics of Hebei Province; and (5) road data (i.e., highways, railways, state roads) and river data, derived from the present land use map of Zhangjiakou city at a scale of 1:10,000.
The processing procedure for the source data was as follows: (1) Subjects for land use were determined from the original data. The data for land use cover was derived from historical TM data images by means of a user-computer interactive interpretation method and by conducting a field survey to check selected points in the field and on maps (Li et al., 2006). Following geometrical image correction and geo-referencing, the average location errors were estimated at less than 30 m (about one pixel). Global accuracies of 86.82%, 88.92% and 90.46% were obtained for the 1989, 2000 and 2010 images, respectively. The Kappa coefficients were 0.84, 0.87 and 0.88, respectively for the same years. This accuracy met the recommended value (Ellis et al., 2010), so the data was available for further study (Zhang et al., 2013b). Based on the classification system reference to the national resources and environmental background dynamic remote sensing survey database, these land use maps of 1989, 2000 and 2010 were classified into eight land use types: farmland, garden land, forest land, grassland, residential and industrial land, transportation land, water body and unused land (Gong et al., 2009). Simultaneously, considering the requirement for the CLUE-S model, the proportion of land use type must be over 1% of the total area (Zhang et al., 2013a). Residential and industrial land and transportation land are classified as construction land. As a result, we chose farmland, garden land, forest land, grassland, construction land, water body and unused land as data sources for the CLUE-S model (Figure 2). Furthermore, ESRI ArcGIS 9.3 was used for the spatial analysis. Topographic maps were collected for establishing the digital elevation model (DEM) to obtain other data (i.e, slope, aspect) and also had a spatial resolution. (2) For the unification of the geo-reference of subjects, we used the Gauss Kruger (GK) projection coordinates with original longitude 117ºE, original latitude 0ºN, Xi’an geodetic datum and Krassovsky ellipsoid. (3) For the rasterization of the vector data, a one-sample Kolmogorov-Smirnov (KS) test was initially used to test the socio-economic data. The results showed that the asymptotic significance level was above 0.05. They indicated that the population and per-capita net income of the farmers followed a normal distribution, and that other socio-economic data approximated to a normal distribution by log transformation. Therefore, the spatial distribution of socio-economic data could be obtained by spatial interpolation using the Kriging and natural neighbor methods. According to the requirements of the CLUE-S model combined with the actual location of Zhangjiakou city, many types of data layers were transferred to a grid format with a resolution of 300 m×300 m using the same projection coordinates, on which spatial logic and algebraic computations were calculated at each subject layer. Finally, all variables were normalized to allow a uniform measurement system (Locantore et al., 2004).
Figure 2 Land use pattern in Zhangjiakou city throughout 1989-2010

3.2 Methods

3.2.1 Land use change measurement
Based on the research, the land use dynamic degree (D) is used as a base indicator to describe the rate of land use change. The land use dynamic degree is calculated as follows (Zhang et al., 2014a):
where Di is the dynamic degree of land type i in the period T; N is the number of land use type; Sai and Sbi represent the area of land that uses type i at the beginning and end of the period, respectively; and T is the study period. The land use dynamic degree is an index that can quantitatively describe the regional difference in the rate of land use change and predict the trend in land use change (Wang and Bao, 1999). Moreover, the land use dynamic degree reflects the comprehensive influence of socio-economic activities on land use changes.
The land use change intensity index (C) can be used to recognize the main types of land use change, which can be expressed as follows (Luo et al., 2014):
where Ci represents the intensity index of land use change type i; Ai is the area of land use change type i; and the value of Ci is between 0 and 1. The closer the value of Ci to 1, the more marked the land use change in a given region. Next, we can select the main types of land use changes by the rank of the value of Ci.
3.2.2 Markov model
The Markov model is used for state transition analysis and forecasting, and is suitable for predicting the conversion of land use in a given region (He et al., 2011; Hu et al., 2013). The Markov model is closely related to dynamic distributed-lag models and consists of two primary land components: the transition matrix and the transition probability matrix, which represent the number and the probability of land shifting from one land use type to other types in an observed period (Han et al., 2015). The land use distributions at the beginning (St) and at the end (St+1) of a time period, as well as a transition probability matrix (Pij) representing land use change that occurred during the defined period, are used to construct the Markov model, which is expressed as follows:
where St+1 and St are the land use statuses at the time-points t+1 and t, respectively, and Pij is a transition probability matrix satisfying the following conditions:
where i and j are the land use types at the time points t+1 and t, respectively.
3.2.3 Landscape metrics analysis
Landscape change is the visual sign of land use change (Feng et al., 2010). A large set of landscape metrics has been developed recently (Su et al., 2012). The capacity of landscape metrics to indicate an ecological process has been discussed by several researchers (Su et al., 2012; Jones et al., 2013; Fan and Myint, 2014; Tian et al., 2014). The selection of landscape metrics in this paper was based on the capacity to reflect the characteristics of landscape patterns for the study area. The landscape fragmentation index (F), landscape shape index (LSI), perimeter-area fractal dimension index (PAFRAC), Shannon’s diversity index (SHDI), interspersion and juxtaposition index (IJI), contagion index (CONTAG) and aggregation index (AI) were selected to describe the differences in landscape patterns in different scenarios. The landscape metrics above were calculated using Fragstats 4.2 software (Frazier, 2014).
3.2.4 The CLUE-S model and scenarios of land use changes
(1) The CLUE-S model
The land use simulation model CLUE-S (the Conversion of Land Use and its Effects at Small regional extent) utilizes the spatial dominant dynamic simulation model established by Wageningen University (Verburg and Veldkamp, 2004; Verburg and Overmars, 2009). The CLUE-S model includes a non-spatial demand module and spatially explicit land use allocation module, which needs to consider the following four elements: land use requirements, location characteristics and suitability, spatial policies and restrictions, and land use type specific conversion settings (Luo et al., 2010). The model can be developed to simulate land use change using empirically quantified relations between land use and its driving factors in combination with dynamic modeling of competition and interactions between the spatio-temporal dynamics of land use systems (Verburg et al., 2002; Verburg and Veldkamp, 2004). To date, the CLUE-S model has been widely used in international researches, and many scholars have conducted simulation by controlling different scenario settings and the total demand for the corresponding land patterns (Batisani and Yarnal, 2010; Verburg and Overmars, 2009; Luo et al., 2010; Wu et al., 2014; Zheng et al., 2015).
(2) Logistic regression
Generally, conversions of land use are expected to take place at locations that have the highest preference for a specific type of land use at a given moment. It can be calculated as a probability of a certain grid cell by logistic regression, as follows (Luo et al., 2010; Zeng et al., 2014; Wang et al., 2014; Zheng et al., 2015):
where Pi is the probability of a grid cell for the occurrence of the considered land use type i, the parameters X refer to the driving factors, n is the number of driving factors, and the coefficients β are estimated through logistic regression using the actual land use pattern as a dependent variable. β0 is the constant coefficient. Exp(β) values indicate the change in odds upon one unit change in the independent variable. When Exp(β)>1, the probability increases upon an increase in the value of the independent variable. When Exp(β)<1, the probability decreases. The relative operating characteristic (ROC) is used as a quantitative measure to evaluate the fit of the logistic regression model (Pontius and Schneider, 2001). The value of ROC ranges from 0.5 to 1.0. An ROC above 0.70 shows that the driving factors had a greater explanatory power for a certain land use type.
(3) Scenarios setting and land use elasticity
Based on local statistical socio-economic information from 1989 to 2010, and related land use planning and policies in Zhangjiakou city, we designed four variable settings of population increase, economic growth and urban development for the 10 years following 2010. Then, according to the combination of these settings, we defined four scenarios - namely, the land use planning scenario, natural development scenario, ecological-oriented scenario and farmland protection scenario - to predict land use demand for 2020 by applying various models, such as the Markov model, GM (1, 1) model and Grey prediction model. (1) The land use planning scenario: this scenario is the reference one, which may hasten the socio-economic development with an 11.7% annual increase in GDP rate during 2010-2020. Total population may increase by 3.82‰ annually from 2010 to 2020 and the urbanization rate will increase to 54.2% in 2020. Based on the planning targets of land use types from the Land Use Master Plan of Zhangjiakou (2006-2020), the structure of land use can be obtained by data interpolation according to the restrictions under land use planning in 2020. (2) The natural development scenario: this scenario almost keeps the present pace with socio-economic development, under which the GDP will increase by 12.0% annually and the total population will increase by 4.98‰ annually from 2010 to 2020. In addition, the tendency of land use changes is predicted by applying time series methods and the GM (1, 1) model. (3) The ecological-oriented scenario: this scenario emphasizes protecting natural reserves, water resources, and biological diversity protection areas and requires all growth to be allocated outside these designated areas. In this scenario, the GDP will increase 10.0% annually and the forest coverage rate will add up to 34.0% in 2020. (4) The farmland protection scenario: in this scenario, basic farmland will be designated as a restricted area in which the farmland cannot be converted into other land use types. The GDP will increase 9.6% annually and the total population will increase by 4.53‰ annually from 2010 to 2020 under this scenario. The Grey prediction model and multi-objective programming model were used to simulate the land use demand for Zhangjiakou city in 2020.
The last decision rule concerns elasticity, which is an indicator of the conversion cost. For a given type of land use, the elasticity ranges from 0 (easy conversion) to 1 (irreversible change), and the value is positively related to the difficulty in converting this type of land use to other types (Wassenaar et al., 2007; Batisani and Yarnal, 2010). The transition matrix of land use types is calculated to ensure that the relative elasticity is extrapolated from current data. Based on the reference data in the study area during 1989-2010, the values of conversion elasticity for different land use types were tuned so that they were suitable for the calibration of the model. According to the defined scenarios, specific conversion elasticity values of land use types were defined and implemented in the model during 2010-2020 (Table 1).
Table 1 Values of land use type conversion elasticity (ELAS) in Zhangjiakou city
Relative elasticity Farmland Garden
land		 Forestland Grassland Construction land Water
body		 Unused
land
Land use planning scenario 0.7 0.9 0.8 0.8 0.9 0.9 0.5
Natural development scenario 0.8 0.8 0.8 0.9 0.9 0.9 0.4
Ecological-oriented scenario 0.8 0.9 0.9 0.9 0.9 0.9 0.6
Farmland protection scenario 0.9 0.9 0.9 0.9 0.9 0.9 0.8

4 Results and analysis

4.1 Land use dynamics

In this section, we have described the characteristics of land use dynamics by changes in land use quantity structure and land use types.
4.1.1 Change in land use quantity structure
Land use has changed significantly over the whole period from 1989 to 2010 in Zhangjiakou city (Table 2). The main types of land use in Zhangjiakou city are farmland, forestland and grassland, which accounted for about 85% of the total area, while garden land and other land use types occupied a relatively small proportion of the total area. Apparently, the area of farmland, grassland, water body and unused land shrank significantly throughout 1989-2010, decreasing by 11.09%, 2.82%, 18.20% and 31.27%, respectively. By the end of 2010, the area of garden land, forestland and construction land all increased, expanding by 5.71%, 20.91% and 38.54%, respectively, relative to 1989. Over the two decades, farmland, forestland, construction land and unused land underwent marked changes. Specifically, farmland and unused land lost 1208.56 km2 and 679.17 km2, respectively, but forestland and construction land increased significantly by 1913.40 km2 and 385.49 km2, respectively.
Table 2 Land use changes in Zhangjiakou city throughout 1989-2010
Land use types 1989 2010 1989-2010
Area
(km2)		 Proportion (%) Area
(km2)		 Proportion (%) Change area
(km2)		 Change rate
(%)
Farmland 10900.30 29.61 9691.74 26.33 -1208.56 -11.09
Garden land 1412.81 3.84 1493.49 4.06 80.68 5.71
Forest land 9149.40 24.86 11062.80 30.06 1913.4 20.91
Grassland 11207.94 30.45 10891.42 29.59 -316.52 -2.82
Construction land 1000.24 2.72 1385.73 3.77 385.49 38.54
Water body 963.16 2.62 787.84 2.14 -175.32 -18.20
Unused land 2171.62 5.90 1492.45 4.05 -679.17 -31.27
The rate of land use change in different periods is illustrated by the land use dynamic degree (Table 3). From 1989 to 2000, the value of the land use integrated dynamic degree was 2.21%, which indicated that the rate of land use change was relatively slow. Meanwhile, from 2000 to 2010, the value of the land use integrated dynamic degree was 3.96%, which demonstrated that the rate of land use change had accelerated compared with that of the previous period and that the effects of socio-economic activities on the land use pattern had intensified. At the same time, the area of farmland, grassland, water body and unused land decreased gradually: the rate changed from 0.27%, 0.15%, 0.88% and 0.73% to 0.83%, 0.11%, 0.94% and 2.53%, respectively, while the area of garden land, forestland and construction land increased continuously. Furthermore, the rate of garden land decreased from 0.51% during the former period to 0.01% during the latter period, and that of forestland and construction land increased from 0.56% and 1.32% to 1.39% and 2.10%, respectively.
Table 3 Rate of land use change during different periods in Zhangjiakou city
Periods Land use dynamic degree (%) Integrated
dynamic
degree (%)
Farmland Garden land Forest land Grassland Construction land Water body Unused land
1989-2000 -0.27 0.51 0.56 -0.15 1.32 -0.88 -0.73 2.21
2000-2010 -0.83 0.01 1.39 -0.11 2.10 -0.94 -2.53 3.96
On the spatial scale, the expansion of construction land was strikingly clear from the urban center to the outskirts (Figure 2). The direction of construction land expansion was primarily along the highways and national roads, from administrative districts (i.e, Qiaoxi, Qiaodong, Xiahuayuan and Xuanhua districts) to Zhangbei, Yangyuan and Yuxian counties in the suburbs. Meanwhile, the area with growth in forestland mainly occurred in Bashang Plateau and its neighboring mountainous areas. In addition, the reduction in farmland was mainly located in administrative districts and their surrounding regions in Zhangjiakou city. This indicated that the urban-rural fringe was the area most sensitive to urbanization. With the socio-economic development, the conflict between urban development and farmland preservation was found to be gradually increasing. The administrative districts and their surrounding regions were the hot spots witnessing much faster land use changes.
4.1.2 Change in land use types
By overlaying the 1989 and 2010 land use maps, the area changes in land use types in the study area over the two decades were further calculated (Table 4). Investigating from the conversion direction of land use types revealed that there was an obvious interaction among farmland, forestland, grassland, construction land and unused land. Changes mainly occurred in Bashang Plateau, river valley and mountainous areas. The total changed area throughout 1989-2010 was 4759.14 km2, accounting for 12.93% of the study area. As shown in Table 4, land use transformation types were characterized by grassland to forestland, and by farmland to forestland and grassland. Among the transformation types of land use, the area of grassland to forestland, farmland to forestland, grassland to farmland, unused land to forestland, unused land to grassland, farmland to construction land, and farmland to garden land represented more than 80% of the area undergoing land use change. It was obvious that land use changes in 1989-2010 were dramatic and dominated by these seven transformation types. During 1989-2000, the conversions of grassland to forestland, and farmland to grassland, comprised the principal transformations in land use. Construction land expansion was mainly derived from farmland, and unused land was changed into forestland and grassland. However, throughout 2000-2010, the conversions of farmland and grassland to forestland accounted for 45.52% of the total area of land use changes, which was mainly affected by the Grain for Green Project. The conversion of farmland to forestland mainly occurred in Bashang Plateau and its neighboring mountainous areas around the junction between Chongli county and Chicheng county. The conversion of grassland to forestland mainly occurred in mountainous areas involving Chongli, Chicheng and Zhuolu counties. The conversion of farmland to construction land was distributed in administrative districts and their surrounding regions, and was caused by the urbanization process and the development of infrastructure.
Table 4 Main types of land use change in Zhangjiakou city throughout 1989-2010
1989-2000 2000-2010 1989-2010
Transformation type Ci (%) Transformation type Ci (%) Transformation type Ci (%)
Grassland to Forestland 29.61 Grassland to Forestland 27.69 Grassland to Forestland 29.15
Farmland to Grassland 18.67 Farmland to Forestland 17.83 Farmland to Forestland 16.26
Farmland to Forestland 8.05 Grassland to Farmland 10.96 Grassland to Farmland 11.30
Unused land to Forestland 6.65 Unused land to Grassland 10.09 Unused land to Forestland 8.44
Farmland to Construction land 6.21 Unused land to Forestland 8.55 Unused land to Grassland 8.21
Grassland to Farmland 5.17 Farmland to Construction land 5.67 Farmland to Construction land 6.04
Unused land to Grassland 4.16 Farmland to Garden land 3.65 Farmland to Garden land 4.21
Other 31 types 21.48 Other 43 types 15.56 Other 36 types 16.39
Total 100.00 Total 100.00 Total 100.00

4.2 Regression analysis of land use change

The CLUE-S model requires setting the relationship between location factors and different land use (Zheng et al., 2015). In this study, a logistic regression model was used to explore the relationship between land use change and the related driving factors (Braimoh and Onishi, 2007). Certain driving factors - including landform (X1), soil organic matter (X2), elevation (X3), slope (X4), aspect (X5), distance to the nearest river (X6), distance to the nearest town center (X7), distance to the nearest county center (X8), distance to the nearest main road (X9), distance to the nearest railway (X10), population density (X11), urbanization rate (X12), economic density (X13), fixed assets investment (X14) and per-capita net income of farmers (X15) - were selected for the purpose of analyzing the location suitability of a certain grid to be devoted to a land use type.
As shown in Table 5, the ROC test statistics for various land use types were all above 0.70. This revealed that the factors selected performed well in terms of explaining the land use pattern of the study area, but different driving factors resulted in some differences in various land use types. Slope was a negative explanatory variable for farmland, suggesting that steep areas with higher elevation and slope would experience a smaller probability of farmland. The probability distribution of farmland decreased 0.2729 times when slope increased by one unit. Distance to the nearest river and distance to the nearest town center were positive explanatory variables of garden land, which was located in river valley, as well as the area around Guanting Reservoir. Slope became an important factor restricting the development of garden land in Zhangjiakou city. Soil was also an important factor shaping land use patterns. However, soil organic matter and slope played an important role in the distribution of forestland and grassland. Forestland was mainly distributed in areas with high elevation and slope as well as in certain remote regions, while grassland was distributed in Bashang Plateau, i.e., Kangbao, Shangyi, Zhangbei and Guyuan counties. Thus, many factors affected the distribution of construction land. Moreover, elevation, slope, distance to the nearest river, distance to the nearest town center, and distance to the nearest county center were negative explanatory variables of construction land. However, economic density and fixed assets investment were positive determinants of construction land. In addition, physical variables had different effects on the distribution of water body.
Table 5 Logistic regression results of the spatial distribution of land use types in Zhangjiakou city throughout 2000-2010
Drivers Farmland Garden land Forestland Grassland Construction land Water body Unused land
β EXP(β) β EXP(β) β EXP(β) β EXP(β) β EXP(β) β EXP(β) β EXP(β)
X1 -0.0228 0.9775 0.0117 1.0118 0.0250 1.0253 0.0117 1.0118 -0.0453 0.9557 -0.0730 0.9296 — —
X2 — — — — -0.0340 0.9666 0.0441 1.0451 — — — — — —
X3 -0.0009 0.9991 -0.0093 0.9908 0.0008 0.9992 0.0227 1.0229 -0.0026 0.9974 -0.0020 0.9980 0.0185 1.0187
X4 -0.2729 0.7612 -0.4639 0.6288 0.0345 1.0351 -0.1669 0.8463 -0.197 0.8212 0.0276 1.0280 0.0415 1.0424
X5 0.0006 1.0006 — — — — -0.0005 0.9995 — — — — — —
X6 -0.0065 0.9935 0.0568 1.0585 -0.0120 0.9881 0.0154 1.0155 -0.0163 0.9839 -0.0545 0.9470 — —
X7 -0.0083 0.9917 0.0435 1.0445 0.0100 1.0100 -0.0066 0.9935 -0.0189 0.9813 — — -0.0024 0.9976
X8 -0.0285 0.9719 — — 0.0269 1.0272 — — -0.1232 0.8841 0.0999 1.1051 -0.1026 0.9025
X9 — — -0.0472 0.9539 0.0055 1.0055 0.0019 1.0019 — — 0.0106 1.0106 -0.0159 0.9843
X10 — — — — — — 0.009 1.0091 — 0.9998 — — -0.0852 0.9183
X11 -0.0009 0.9991 0.0015 1.0015 -0.0008 0.9992 -0.0007 0.9993 -0.0002 — — — — —
X12 -0.0282 0.9722 0.0354 1.0360 -0.0297 0.9707 -0.0046 0.9954 — — — — -0.1562 0.8554
X13 0.0007 1.0007 -0.0018 0.9982 -0.0002 0.9998 -0.0004 0.9996 0.0004 1.0004 — — 0.0014 1.0014
X14 0.0044 1.0045 0.0074 1.0074 -0.0034 0.9966 -0.0038 0.9962 0.0014 1.0014 — — 0.0243 1.0246
X15 -0.0001 0.9999 0.0004 1.0004 — — — — — — 0.0002 1.0002 -0.0002 0.9998
Constant -0.1458 0.8643 -12.7820 0.0001 -8.7846 0.0002 -9.8221 0.0001 -4.4890 0.0112 -2.0367 0.1305 -9.9562 0.0014
ROC value 0.896 0.878 0.870 0.862 0.915 0.835 0.798

Note: The mark “—” denotes the factors removed. X1-landform, X2-soil organic matter, X3-elevation, X4-slope, X5-aspect, X6-distance to the nearest river, X7-distance to the nearest town center, X8-distance to the nearest county center, X9-distance to the nearest main road, X10-distance to the nearest railway, X11-population density, X12-urbanization rate, X13-economic density, X14-fixed assets investment, X15-per-capita net income of farmers.

4.3 Simulation of land use demand

Based on the land use change from 1989 to 2010, the Markov model, GM (1, 1) model, Grey prediction model and multi-objective programming model were used to predict the demand for different land use types in 2020 according to the four future scenarios defined above. The prediction results were presented in Table 6. As shown in the table, there was significant difference in the demand for various land use types under different scenarios.
Table 6 The prediction results of land use under different scenarios (km2)
Year Scenario design Farmland Garden
land		 Forestland Grassland Construction land Water
body		 Unused land
2010 Actual land use 9691.74 1493.49 11062.80 10891.42 1385.73 787.84 1492.45
2020 Land use planning scenario 8721.00 1500.49 12932.90 10634.90 1535.00 725.48 755.70
Natural development
scenario		 8288.79 1521.89 13061.68 10819.07 1744.20 722.96 646.88
Ecological-oriented
scenario		 8844.12 1495.98 12478.14 10756.97 1521.87 728.66 979.73
Farmland protection
scenario		 9429.17 1562.82 11543.47 10712.17 1516.48 712.91 1328.45

4.4 Simulation of land use scenarios

Based on the results of the logistic regression model and the transition matrix, the CLUE-S model was used to simulate the distribution of land use pattern in 2010. The accuracy of the temporal predictions of the CLUE-S model was analyzed based on the Kappa index by comparing the actual land use map in 2010 with the simulated map of land use pattern for Zhangjiakou city in 2010 (Figure 3). The Kappa statistic was employed to evaluate the accuracy of the model (Pontius, 2000). The value of the Kappa statistic was 0.902 throughout the temporal scale of 10 years during 2000-2010, which indicated that the consistency was good between the simulated map and the actual land use map. Moreover, this showed that the model was reliable for Zhangjiakou city and could be used to predict future land use patterns under different scenarios (Pan et al., 2011).
Figure 3 Land use map (a) and simulation map (b) of Zhangjiakou city in 2010
Figure 4 Land use simulation maps of Zhangjiakou city in 2020
(a) Land use planning scenario; (b) Natural development scenario; (c) Ecological-oriented scenario; (d) Farmland protection scenario
Figure 4, farmland, forestland and grassland were the main land use types. Farmland was Combined with the land use demand under different scenarios, land transfer rules, related driving factors and the calibrated model settings, the CLUE-S model was used to simulate the distribution of land use patterns of Zhangjiakou city in 2020 (Figure 4). As shown in primarily concentrated in mid river valley. Forestland and grassland were mainly distributed in Bashang Plateau and its neighboring mountainous areas, as well as southern mountainous areas. In addition, land use types underwent different dynamics under the four scenarios, as follows:
(1) Land use planning scenario (reference scenario, Figure 4a). Apparently, farmland, grassland, water body and unused land shrank significantly in 2020 compared with 2010, while garden land, forestland and construction land increased. During 2010-2020, forestland increased by 1900.61 km2. The increase in forestland was mainly derived from farmland and unused land. Thus, the conversion of farmland to forestland was expected to be distributed in the Bashang Plateau and the southern mountainous areas, accounting for 49.63% of the total area in new forestland. In addition, the conversion area of unused land to forestland was 664.20 km2, and was mainly located in the mountainous areas of Wanquan and Yuxian counties. Moreover, the expansion of construction land was concentrated in administrative districts and their surrounding regions.
(2) Natural development scenario (Figure 4b). There was an obvious expansion of construction land with a concentrated distribution in 2020. Compared with the land use pattern in 2010, construction land increased by 358.47 km2, which had expanded by occupying farmland and unused land. The conversion of farmland to construction land was 195.48 km2, which mainly occurred in districts and their surrounding regions. The conversion of unused land to construction land was 128.79 km2, and had been concentrated in the regions along roads and river valley of Yangyuan county. Forestland increased by 1998.88 km2. The area of the new forestland from farmland was 1142.62 km2, which mostly occurred in the Bashang Plateau and mountainous areas around the junction among Zhangbei, Guyuan, Chongli and Chicheng counties as well as the junction between Zhuolu and Yuxian counties with high elevation. The area of new forestland from unused land in the mountainous areas of Huai’an and Yuxian counties, occurring mostly from 2010 to 2020, was 710.19 km2.
The results of the simulation in the natural development scenario were similar to those in the land use planning scenario, but the changes in land use types were more fundamental. There were significant differences in most land use types, including farmland, forestland, grassland, construction land and unused land. In the natural development scenario, the area of farmland and unused land was 432.21 km2 and 108.82 km2 less, respectively, than that in the land use planning scenario, while the area of forestland, grassland and construction land was 128.78 km2, 184.17 km2 and 209.20 km2, respectively, more than that in the reference scenario. On the spatial scale, some differences were found to be distributed in the Bashang Plateau and mountainous areas between the two scenarios. More farmland converted to forestland than the reference scenario in the Bashang Plateau and its neighboring mountainous areas, and much more farmland was used for construction in Qiaodong district. In addition, more conversion of unused land to construction land primarily occurred in Zhangbei, Yangyuan and Yuxian counties under the natural development scenario.
(3) Ecological-oriented scenario (Figure 4c). In this scenario, farmland and unused land shrank by 847.62 km2 and 512.72 km2, respectively, while construction land experienced slow expansion and forestland underwent an obvious expansion. During 2010-2020, forestland increased by 1415.34 km2, which was mainly derived from farmland, grassland and unused land, with 709.20 km2, 120.51 km2, 455.22 km2, respectively. The conversion of farmland to forestland was mostly located in the hilly areas, river valleys and mountainous areas. The predicted grassland/unused land-forestland conversion areas mainly occurred in the following regions: the hilly area in Bashang Plateau, the northeast mountainous areas of Yanshan Mountain, and the southern mountainous area of Xiaowutai.
Less change in land use occurred under the ecological-oriented scenario from 2010 to 2020. This showed that there was less demand for garden land, forestland and construction land in this scenario. The area of forestland was 454.76 km2 less than that in the land use planning scenario, which was mainly located in Shangyi, Zhangbei, Huai’an and Yuxian counties. Meanwhile, more demand for farmland, grassland and unused land was required in the ecological-oriented scenario. The area of farmland, grassland and unused land was 123.12 km2, 122.07 km2 and 224.03 km2, respectively, more than that in the reference scenario. There was more grassland, which was mainly distributed in mountainous areas within Chicheng, Yuxian and Zhuolu counties. Moreover, more unused land was scattered in high-altitude areas in Shangyi, Zhangbei, Huai’an and Yuxian counties.
(4) Farmland protection scenario (Figure 4d). This study assumed that the land use was under a strict farmland protection policy from 2010 to 2020. Comparing the simulated results in 2020 with the actual map, the changes in farmland and forestland were relatively obvious. In particular, farmland was primarily converted to forestland, with an area of 225.63 km2. Moreover, grassland/unused land-farmland conversion mostly occurred in regions of the surrounding districts (i.e., Huai’an county, Xuanhua county) and the mountainous area within Yuxian county. Similar to the above scenarios, farmland was concentrated in the flat areas. Construction land was scattered in Zhangjiakou city, which was mainly surrounded by farmland.
Compared with the land use planning scenario in 2020, this showed that there was less demand for forestland and more demand for farmland and unused land under the farmland protection scenario. Much more unused land existed in the mountainous areas of Huai’an and Yuxian counties. However, more farmland was scattered in the region around the junction among Guyuan, Zhangbei, Chongli and Chicheng counties. However, the conversion of farmland to construction land and forestland still existed in the mountainous areas of Huai’an, Chicheng, Chongli and Yuxian counties, which indicated that Zhangjiakou would continue to face pressure in farmland protection.

4.5 Landscape metrics analysis

The heterogeneity of landscape can be used to characterize the stability and safety of an ecological system (Feng et al., 2010; Jones et al., 2013; Wang et al., 2014), which can be used to reflect the simulation results of land use patterns under different scenarios. Compared with the landscape pattern of land use in 2010, the study area will exhibit more homogeneity and its landscape pattern will become less marked in 2020, evidenced by the reduction in the fragmentation index (F) and landscape shape index (LSI), and the increase in the contagion (CONTAG) and aggregation index (AI) (Table 7). Grassland and unused land will become more fragmented, while other land use types will become less fragmented in 2020. A reduction in PAFRAC for farmland and water body indicates that patch shapes will tend to be simpler. However, an increase in PAFRAC and LSI of construction land will occur from 2010 to 2020 (Table 8).
Table 7 Landscape metrics of Zhangjiakou city under different scenarios
Land use scenario F LSI PAFRAC SHDI IJI CONTAG AI
2010 Actual land use 0.0097 131.4746 1.5738 1.5691 73.8408 29.6276 59.4424
Land use planning scenario
(Reference scenario)		 0.0090 126.4539 1.5730 1.5875 73.9858 32.2885 61.0127
Natural development scenario 0.0089 126.6271 1.5768 1.5519 73.2345 32.5534 60.9583
Ecological-oriented scenario 0.0093 128.0227 1.5744 1.5706 72.5546 31.1907 60.5223
Farmland protection scenario 0.0095 129.8600 1.5750 1.5733 70.5517 30.0865 59.9485
Table 8 Class metrics of Zhangjiakou city under different scenarios
Farmland Garden land Forestland Grassland Construction land Water body Unused land
2010 Actual land use F 0.0055 0.0177 0.0068 0.0076 0.0447 0.0377 0.0181
PAFRAC 1.6025 1.5709 1.5594 1.6022 1.4809 1.5045 1.4974
LSI 127.6094 67.2558 126.8260 141.3391 82.5360 51.7500 57.0814
Land use planning
Scenario
(Reference scenario)		 F 0.0048 0.0176 0.0057 0.0079 0.0427 0.0345 0.0204
PAFRAC 1.5989 1.5712 1.5613 1.6038 1.4983 1.5014 1.5112
LSI 118.6083 66.3977 123.0988 141.0610 85.2759 47.5944 44.5054
Natural development
scenario		 F 0.0048 0.0170 0.0054 0.0076 0.0417 0.0340 0.0182
PAFRAC 1.6003 1.5693 1.5667 1.6028 1.5134 1.4991 1.5189
LSI 116.2158 63.9960 125.3124 141.2334 90.3011 46.9441 40.1000
Ecological-oriented
scenario		 F 0.0049 0.0177 0.0059 0.0077 0.0434 0.0339 0.0214
PAFRAC 1.5992 1.5727 1.5622 1.6033 1.5106 1.4971 1.5078
LSI 119.4530 66.4690 123.6671 141.1055 89.5018 46.3911 50.6555
Farmland protection
scenario		 F 0.0052 0.0170 0.0066 0.0077 0.0435 0.0342 0.0195
PAFRAC 1.6008 1.5731 1.5595 1.6031 1.4988 1.4941 1.5091
LSI 124.5201 67.0455 124.7448 140.7728 85.5211 46.1638 56.3827
4.5.1 Landscape-level metric analysis
The results of landscape-level metrics analysis are shown in Table 7. There are many superior indices in land use planning scenarios, including the landscape shape index (LSI), fractal dimension index (PATRAC), Shannon’s diversity index (SHDI), interspersion and juxtaposition index (IJI) and aggregation index (AI). There are significant characteristics of landscape pattern at the landscape level in this scenario. LSI and PATRAC are at their lowest value in this scenario, indicating that the landscape is more regular. The natural development scenario reflects the relevant difference in the landscape pattern of land use. The fragmentation index (F) and contagion index (CONTAG) are the superior indices in this scenario, whose landscape is less fragmented with better spatial connectivity. It is indicated that appropriate human interventions can regulate the spatial form of land use effectively to avoid excessive dispersion of the landscape. The increases in SHDI, IJI and AI suggest that landscape diversity becomes more significant and the distribution of landscape patches tends to be accumulative. However, there is no superior index under an ecological-oriented scenario or farmland protection scenario, in which the landscape pattern exhibits more fragmented and homogeneous. Given all that, the combination of land use types in land use planning scenarios is less marked than other scenarios. In this section, reductions in F and LSI and increases in CONTAG and AI indicate that landscapes become less fragmented under different scenarios. At the same time, the spatial structures of landscapes tend to be simpler, while shapes of landscape will become more regular in 2020 than in 2010. Moreover, the aggregation and connectivity of land use patterns will increase gradually. The complex changes in landscape patterns are examined with three metrics (i.e., PAFRAC, SHDI and IJI). In addition to SHDI, its increase indicates that the heterogeneity and diversity of landscapes increases under different scenarios, except with the natural development scenario.
4.5.2 Class-level metric analysis
Table 8 shows the results from the landscape pattern changes at class level under different scenarios. The land-use classes incorporated into the class-level metric analysis were farmland, garden land, forestland, construction land, water body and unused land. Fragmentation index (F), fractal dimension index (PATRAC) and landscape shape index (LSI) reveal important information about the landscape characteristics of each land use type under different scenarios. In the land use planning scenario, F and PAFRAC are at the highest values for landscapes of grassland and water body, indicating that their landscape patches become fragmented and more complex during 2010-2020. Furthermore, landscapes of forestland and construction land are characterized by the lowest values of LSI. This finding shows that the effective protection of landscapes of grassland and water body cannot be achieved by the total amount of control in the transition zone between the cropping area and the nomadic area. Considering the quantity demand of farmland and construction land, some guidance on the spatial changes in the major land use types by human factors must be enhanced to realize reasonable allocation of land resources and promote the coordinated development of socio-ecological systems (S-ESs). Under the natural development scenario, the landscape is characterized by low fragmentation of land use types and more complex shapes of construction land and unused land with high PAFRAC value. In addition to forestland and grassland, LSI is at its highest value, indicating the concentrated condition. Owing to the economic development and Grain for Green Project in the study area, farmland is occupied by construction land and ecological land. The changes in land use types dominated by human activities have an important influence on the landscape pattern of land use in Zhangjiakou city. Under the ecological-oriented scenario, there is low fragmentation of landscapes of garden land and unused land. The landscape of ecological land gets effective protection, which contributes to the policy of the Grain for Green Project as well as the Sandstorm-Control Program. Under the farmland protection scenario, the landscapes of farmland, garden land, forestland and construction land exhibit more fragmented and heterogeneous. Moreover, patch shapes of farmland and garden land will become more complex and irregular than other scenarios in 2020. However, a reduction in the fractal dimension index (PATRAC) of grassland and water body suggests that their structures present simpler and more regular. According to the above analysis, the strict measures of farmland protection cannot effectively relieve the potential impact of the landscape pattern, which was caused by human activities in the environmental fragile zone.

5 Discussion

Land use changes are caused by both natural factors and human activities (Vitousek et al., 1997; Liao et al., 2011). Since 1989, human activities, such as population growth, rapid economic development, increased urbanization and land use policy, have had a major impact on land use changes in Zhangjiakou city. In addition, administrative policies and natural factors have played important roles in land use changes. From 1989 to 2010, the land use structure of Zhangjiakou city changed dramatically, as shown in Table 2. Land use changes were characterized by the decline in farmland, grassland, water body and unused land, and the increase in garden land, forestland and construction land. Throughout 1989 to 2000, ecological deterioration and farmers’ poverty were the two critical problems of Zhangjiakou city. Owing to low productivity, population pressure and unreasonable human activities, the irrational land use mode of over-reclamation, overgrazing and deforestation prevailed in this area. The conversions of grassland to forestland and farmland to grassland constituted the main direction of land use changes. The economic development was at a lower level owing to unreasonable industrial structure and serious land desertification during the same period. As a result, the land use intensity and rate of land use change were much lower in Zhangjiakou city. All these indicated the characteristics of land use changes in farming-pastoral zones. At the same time, throughout 2000-2010, Zhangjiakou city experienced rapid socio-economic development and urbanization. The gross domestic product (GDP) increased by 296.06%, and the urbanization rate increased from 24.00% to 32.65%, during the same period. The poor transport infrastructures were further improved from 2000 to 2010, and the total roadways mileage increased from 5368 km to 19,234 km, with an increase of 13,866 km. Construction land (including residential land and transportation land) increased by 240.30 km2 over the same period. The implementations of reform and opening-up policy in 1995 and the Grain for Green Project in 2000 had profound impacts on land use changes in Zhangjiakou city. The changes in land use intensified. Forestland and construction land increased dramatically, at a rate of 1.32% and 2.10%, respectively. As a result, grassland and farmland were more likely to convert to forestland in Bashang Plateau and mountainous areas. Construction land expanded by occupying farmland and grassland, and mostly occurred in administrative districts and their surrounding regions. Overall, the land use integrated dynamic degree increased from 2.21% during 1989-2000 to 3.96% during 2000-2010. Considering the socio-economic development and eco-environmental protection, as well as local government policies, four scenarios (land use planning scenario, natural development scenario, ecological-oriented scenario, farmland protection scenario) were designed to simulate land use changes in the study area. With the rapid growth in population and urbanization, and further implementation of the Grain for Green Project, forestland and construction land increased significantly, while farmland and unused land decreased dramatically under the natural development scenario. With the land use planning scenario, land use demands were related to the Land Use Master Plan of Zhangjiakou city (2006-2020), which highlighted farmland protection by strictly controlling the expansion of construction land. Forestland and construction land increased steadily, while the areas of farmland, grassland, water body and unused land were on the decline under this scenario. With the ecological-oriented scenario, there were wide changes in land use in Zhangjiakou city, which were mainly reflected in farmland, forestland, grassland and unused land. The conversions of farmland, grassland and unused land to forestland were the principal transformation of land use under this scenario. In addition, policies aimed at the protection of farmland suppressed the increase in land use intensity and the expansion of construction land under the farmland protection scenario.
Landscape metrics can quantify landscape patterns and facilitate the comparison between different land use scenarios. Moreover, metrics can reveal important information about landscape patterns. As a result, landscape metrics have been widely used to analyze landscape patterns of land use changes as well as urban expansion, i.e., metropolitan areas (Jenerette and Wu, 2001; Luck and Wu, 2002; Weng, 2007; He et al., 2011; Wu et al., 2015), desert-oasis areas (Meng et al., 2005; Li et al., 2013; Li et al., 2014), coastal areas (Wu et al., 2013; Fan et al., 2013; Hu and Dong, 2013) and arid inland river basins (Tian et al., 2014; Zhang et al., 2014c). In this study, landscape metrics such as F, LSI, PAFRAC, SHDI, IJI, CONTAG and AI were used to analyze landscape patterns under different scenarios. This revealed that the changes in landscape patterns under different scenarios in 2020 will tend to be less marked as human activities intensify. Fragmentation and irregularity of landscapes decreased from 2010 to 2020, while concentration and connectivity of landscapes were further improved. Owing to the rapid economic development, and further implementation of the Grain for Green Project, the landscape was characterized by homogeneity and connectivity under the natural development scenario. Land use pattern was more closely related to land use master plan and policies under the land use planning scenario. With increasing urbanization, the landscape became more regular in shape, and landscape diversity increased as well as aggregation under this scenario during 2010-2020. However, in the ecological-oriented scenario and farmland protection scenario, consideration of the responses to human activities, ecological land and farmland protections were the major goals. Fragmentation, heterogeneity and complexity were the most obvious characteristics of landscape changes. Overall, there were also spatial differences in landscape patterns in Zhangjiakou city, which were consistent with the basic pattern of socio-economic conditions and the trend toward changes in future land use pattern.
Owing to the complexity of the land use system, it is necessary to consider and combine different land use processes within a single modeling framework to model land use dynamics (Luo et al., 2010). The CLUE-S model has been recognized as an excellent tool for simulating land use changes (Overmars and Verburg, 2007; Sohl and Claggett, 2007; Pan et al., 2011; Hu et al., 2013; Wang et al., 2014; Zeng et al., 2014). This model has been widely applied to the dynamic studies of land use changes in smaller-scale regions, especially the hot spots of LUCC research in China, such as sandy areas (Zhang et al., 2003), tourism areas (Wang et al., 2014), urban areas (Sheng et al., 2008; Wu et al., 2011; Zhang et al., 2013b; Zheng et al., 2015) and ecologically fragile areas (Li et al., 2011b; Huang et al., 2012; Xu et al., 2013a). In this study, the Markov-CLUE-S model was used to analyze land use changes from both spatial and temporal dimensions. This model obtained the spatial characteristic information of land use changes by analysis tools using GIS. The model parameters were based on the historical land use data from 2000 to 2010 in Zhangjiakou city and were tested by comparing the simulation scenarios of land use with the actual land use in 2010 of the study area. The Kappa value was 0.902, indicating the reliability of the model simulation. The results of land use simulation in 2020 showed the land use alternatives under four different scenarios in the future. The simulation results indicated that land use change was influenced by physical, locational, political and socio-economic factors. Moreover, to a large extent the selected factors can be used to explain land use allocation.
By using multi-temporal remote sensing data and statistical analysis, the CLUE-S model and landscape metrics analysis are combined to simulate the changes in landscape patterns under different scenarios in 2020, which is an efficient method, as proved by certain case studies (Wang et al., 2014; Zhang et al., 2014b; Zeng et al., 2014). This research is a pilot study for applying a simulation model to explain land use dynamics in an urban fringe. The results of combining a Markov model, CLUE-S model and landscape pattern indices (LPIs) suggest that this integrated model has the capacity to reflect the complex changes in land use in Zhangjiakou city, which indicate that the approach can serve as a useful tool for analyzing related driving factors and estimating related effects of the land use dynamic. The integrated model developed in this research considers land use allocation factors including physical, locational and socio-economic attributes, which are comparatively comprehensive. However, the effects of various policy factors and unexpectedness of natural disasters are not considered in the simulation process. Consequently, improving the accuracy of socio-economic factors and reducing the influences caused by the uncertainty of natural disasters and regional policies are likely to become increasingly important in terms of simulating land use changes. Furthermore, it is well known that the land use change process is dynamic and results from the interaction between natural and socio-economic elements at different scales. The simulation of land use change and landscape metrics analysis rely heavily on scale, and results will vary with particle sizes (Røpke, 2005; Hu and Dong, 2013; Wu et al., 2015). In this study, good results were achieved at a scale of 300 m×300 m in Zhangjiakou city. A further study could employ regional spatial factors and policy-related factors together in the simulation of further land use change in Zhangjiakou city, which could guide more informed decision making. In addition, multi-scale analysis of land use simulation and landscape pattern changes should be carried out to enhance the reliability of simulation in future research.

6 Conclusions

Based on the remote sensing data of Landsat TM images in 1989, 2000 and 2010, using a Logistic regression model, CLUE-S model and landscape metrics, this paper selected Zhangjiakou city - a representative area of the Poverty Belt around Beijing and Tianjin - as an example to simulate land use scenarios driven by different socio-economic and ecological policies and analyzed the landscape pattern dynamic change, aiming to provide useful information for decision-makers and planners to take appropriate land management measures in the study area. The results can be shown as follows:
(1) The land use integrated dynamic degree increased from 2.21% during 1989-2000 to 3.96% during 2000-2010. Comparing the period 1989-2000 with 2000-2010, the rate of farmland, grassland, water body and unused land changed from 0.27%, 0.15%, 0.88% and 0.73% to 0.83%, 0.11%, 0.94% and 2.53%, respectively. At the same time, the rate of garden land, forestland, and construction land changed from 0.51%, 0.56% and 1.39% to 0.01%, 1.32% and 2.10%, respectively. The administrative districts and their surrounding regions were the hot spots in land use changes, where land use changed much faster.
(2) Land use changed significantly throughout 1989-2010. The total area experiencing land use change was 4759.14 km2, accounting for 12.53% of the study area. Land use transformation was characterized by grassland to forestland, and by farmland to forestland and grassland. During 1989-2000, the conversions of grassland to forestland, and farmland to grassland, constituted the principal transformation of land use. However, throughout 2000-2010 the conversions of farmland and grassland to forestland constituted the principal transformations of land use. The land use changes mostly occurred in Bashang Plateau and river valleys from 1989 to 2010.
(3) The spatial distribution of land use types could be explained, to a large extent, by the driving factors. The results of land use simulation showed that the CLUE-S model conforms to reality with an accuracy of 90.2%, implying that this model is appropriate for simulating the temporal and spatial variations in the land use of Zhangjiakou city. Under the natural development scenario, the simulation of land use in 2020 indicated significantly rapid reductions in farmland and unused land due to human activities. Additionally, increases in forestland and construction land resulted from the Grain for Green Project and rapid socio-economic development. Under the land use planning scenario, farmland, grassland, water body and unused land shrank significantly, while garden land, forestland and construction land increased. The expansion of construction land had been concentrated in administrative districts and their surrounding regions. Under the ecological-oriented scenario, forestland increased by occupying other landscapes, such as farmland, grassland and unused land. The ecological land is mostly distributed in the river valleys and the mountainous areas. Under the farmland protection scenario, farmland was well protected and stable, and was concentrated in the flat areas. The expansion of construction land was restricted to some extent by farmland in this scenario.
(4) Compared with the actual landscape in 2010, the landscape patterns under the four scenarios in 2020 were found to be less marked. The fragmentation of land use landscapes decreased and landscape shapes became more regular, while the aggregation of landscape patches and connectivity of landscapes increased. There were significant differences among land use patterns under different scenarios. In the land use planning scenario, the landscape pattern tended to be more optimized. Shapes of landscape became more regular and the degree of landscape diversity and aggregation increased. The landscape became less fragmented and heterogeneous under the natural development scenario, as evidenced by the reductions in F and SHDI and the increase in CONTAG. At the same time, under the ecological-oriented scenario and farmland protection scenario, landscapes were characterized by fragmentation and spatial heterogeneity. Spatial differences in landscape patterns also existed in Zhangjiakou city in 2020.

The authors have declared that no competing interests exist.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Batisani N, Yarnal B, 2010. Rainfall variability and trends in semi-arid Botswana: Implications for climate change adaptation policy. Applied Geography, 30(4): 483-489.Rainfall variability is an important feature of semi-arid climates, and climate change is likely to increase that variability in many of these regions. An understanding of rainfall variability and trends in that variability is needed to help vulnerable dryland agriculturalists and policymakers address current climate variation and future climate change. The goals of this paper are to examine this climatic phenomenon in semi-arid Botswana, to investigate these results for signs of climate change, and to explore the policy implications for climate adaptation. To reach these goals, the paper determines rainfall variability and monthly and annual trends in that variability. The results agree with earlier work showing gradients in rainfall and rainfall variability across Botswana. The results also identify a trend towards decreased rainfall throughout the nation, which is associated with decreases in the number of rainy days. Both the drying trend and decrease in rainy days agree with climate change projections for southern Africa. The paper discusses policies that the government could adopt to help its farmers adapt to climate change.

DOI

[2]
Braimoh A K, Onishi T, 2007. Geostatistical techniques for incorporating spatial correlation into land use change models.International Journal of Applied Earth Observation and Geoinformation, 9(4): 438-446.Land use modeling requires large amounts of data that are typically spatially correlated. This study applies two geostatistical techniques to account for spatial correlation in residential land use change modeling. In the first approach, we combined generalized linear model (GLM) with indicator kriging to estimate the posterior probability of residential development. In the second approach, generalized linear mixed model (GLMM) was used to simultaneously model spatial correlation and regression fixed effects. Spatial agreement between actual and modeled land use change was higher for the GLM incorporating indicator kriging. The GLMM produced more reliable estimates and could be more useful in analyzing the effects of driving factors of land use change for land use planning.

DOI

[3]
Britz W, Verburg P H, Leip A, 2011. Modeling of land cover and agricultural change in Europe: Combining the CLUE and CAPRI-Spat approaches.Agriculture Ecosystems & Environment, 142(1): 40-50.

[4]
Chen H, Wang T, Liang Xet al., 2009. Simulation and application of household’s LUCC based on a multi-agent system: A case study for Mengcha village of Mizhi county of Shaanxi province.Acta Geographica Sinica, 64(12): 1448-1456. (in Chinese)<p>Multi-Agent Systems (MAS) offer a conceptual approach to include multi-actor decision making into models of land use change. Through the simulation based on the MAS, this paper tries to show the application of MAS in the micro scale LUCC. This paper starts with a description of the context of MAS research. Then, based on the BDI (belief, desire, intention) decision-making architecture, it provides a model for the household decision- making. Based upon this framework, Section 4 reports a case study for Mengcha Village of Mizhi County of Shaanxi Province. Section 5 gives the simulated LUCC of 2008, and verifies the result. Section 6 discusses the potentials and drawbacks of the following approach. From our design and implementation of the MAS in a micro scale model, a number of observations and conclusions can be drawn on the implementation and future research directions. (1) The use of BDI decision-making to represent individual households provides a more realistic modeling of the making-decision process. (2) The use of continuous function, not discrete function, which constructs the interaction among households, is more realistic to reflect the effect. (3) This paper attempts to give a quantitative method to analyze the household interaction. And it provides the premise and foundation for researching the communication and learning. (4) The BDI decision-making architecture constructed in this paper helps to accumulate theoretical and practical experience for the interaction research between the micro land use decision-making and the macro land use landscape. Our future research work will focus on the following aspects. (1) The scale issues should be paid attention to the transformation of the household land use decision-making to the collective decision-making. (2) The methods should be explored on the researches into the household decision-making in a longer period of time, so as to build the bridge between long-term LUCC data and short-term household making-decision. (3) The quantitative method and molder, especially the scenario analysis molder, should be researched, which may reflect the interaction among the different household types.</p>

[5]
Couclelis H, 2005. Where has the future gone? Rethinking the role of integrated land-use models in spatial planning. Environment and Planning A:Environment and Planning, 37(1): 1353-1371.The focal concern of this article is the investigation of the transfer and sustainability of the reflective process into the work environment. Specifically, the identification of the variables which support or challenge practitioners to continue the ongoing process of reflection in practice contexts is addressed. The article describes a study carried out over a seven-year period with students/graduates from a master's in social work (MSW) professional qualification programme in Ireland. The research involved gathering data on participants' experiences of reflective teaching and learning while on the course and in the initial years of their work as practitioners. In the early phases of the data collection, participants referred to developing epistemological awareness through the reflective process while on the course. This was in the context of a scaffold for reflection through journal writing and mentored portfolio inquiry. The outcomes of the study offer considerable insight into

DOI

[6]
Dai S, Zhang B, 2013. Land use change scenarios simulation in the middle reaches of the Heihe river basin based on CLUE-S model: A case of Ganzhou District of Zhangye City.Journal of Natural Resources, 28(2): 336-348. (in Chinese)<p>Spatial data, like land-use data, have a tendency to be dependent (spatial autocorrelation), which means that when using spatial models, a part of the variance may be explained by neighboring values. Through incorporating components describing the spatial autocorrelation into a classic logistic model, this study sets up a regression model (Autologistic model), and uses the model to simulate and analyze the spatial land use patterns in Ganzhou District of Zhangye City. Then the scenario simulation of the land use/cover change in 2020 in the study area was modeled by CLUE-S (The Conversion of Land Use and its Effects at Small Region Extent) model. The results show that: (1) Through comparison with the classic logistic model without considering the spatial autocorrelation, the Autologistic model showed better goodness of fitting and higher accuracy of fitting. The distribution of land use types of cultivated land, forestland, water area and unused land yielded areas under the ROC curves (AUC) were improved to 0.924, 0.892, 0.766, 0.716 and 0.835 respectively when using autologistic model. (2) The land use type spatial pattern of 2009 was simulated based on the LUCC data of 2005. The Kappa index based on Autologistic model and logistic model is 0.9354 and 0.8880, respectively,which means the accuracy of CLUE-S model was increased by Autologistic model. (3) This paper simulates the land use pattern of 2020 based on the five scenarios indicated that there existed an obvious spatial difference during different scenarios of LUCC model. 1) Under natural growth scenario, the increase of arable land is important for food security, and the expansion of construction land area will promote the economic development, but the woodland and grassland conversion to unused land will exacerbate land degradation and ecological environmental degradation. 2) Under three water resource restriction scenarios, the water resources serve as very important restriction factors for land use/cover change, and improvement of the utilization of water resources is an important measure to improve the ecological environment in arid areas. 3) Under the land use structure optimization scenario, the land use was more rational because of taking into account the needs of food security, ecological protection and economic development in the study area. 4) Under the scenario of economic development, the construction land was rapidly expanded, and food security would be threatened due to occupation of lots of high quality cultivated land and grassland. 5)Under the ecological protection scenario, the regional eco-environmental quality could be improved to a large extent in woodland, grassland and water area.</p>

DOI

[7]
Deng H, He Z, Chen Yet al., 2013. The land use spatio-temporal pattern simulation in metropolitan fringe of the Sichuan Basin: A case study in Yanjian District, Ziyang City, Sichuan.Scientia Geographica Sinica, 33(12): 1524-1530. (in Chinese)<p>Nowadays with the rapid development of China's economy, the urbanization process has been accelerating significantly. Due to the obvious advantages of geographical location and led by large cities, the urbanization process of many satellite cities located in urban fringes of large cities have been speeding up. How to achieve the sustainable utilization of land resources has become the key issue that must be addressed to keep the urban development. Locating in the middle of the Sichuan basin, Yanjiang District, Ziyang is in Chengdu one-hour economical circle. Taking Yanjiang District as the study area and using the CLUE-S model (The conversion of Land Use and its Effects at Small Region Extent), this article made a dynamic simulation of spatial-temporal pattern of land use from 2005 to 2020 in study area and analyzed the characteristics of its land use and cover change. In the simulation research, by using ETM image and ALOS image as data source, land use data of 2005 and 2009 was obtained through decision tree classification. Thereinto, the data of 2005 was land use data of simulation initial year, and data of 2009 was verification data of simulation result. Gray model is applied to predict the demand for land use types of future years based on the data of land use in previous years.On this basis, combined with the general plan of regional land use, the paper selected socio-economic drive factors and physical geography drive factors which are relating to land use and land cove change, and simulated the spatial-temporal pattern of study area land use from 2005 to 2020 by using CLUE-S model.Logistic model is used to calculate the correlation between types of land use of study area and socio-economic, physical geography drive factors.Then the validity of the prediction results in the study area is gotten through the ROC test.Afterwards based on the land use data of 2005, by inputting correct parameters into the CLUE-S model, The map of land use spatial pattern distribution of Jianyang County in 2009 and 2020 is simulated.Then the paper verified the accuracy of 2009 land use simulation result, of which the Kappa coefficient is 0.887. That showed the simulation results are of high precision. By using GIS（Geographic Information System）, the analysis of simulation results of 16 years(2005-2020) showed that in the study area,cultivated land area reduced significantly, garden plot area decreased slightly, and construction land and woodland area increased. Due to the strict water protection, the water area changed little. The simulation results and the land layout of general land use planning coincided highly with each other. That meant the CLUE-S model could simulate the land use and cover change of Sichuan basin hilly area more accurately, and could be used for reference in research of land use change of other cities in this area.</p>

[8]
Ellis E A, Baerenklau K A, Marcos-Martínez Ret al., 2010. Land use/land cover change dynamics and drivers in a low-grade marginal coffee growing region of Veracruz, Mexico.Agroforestry Systems, 80(1): 61-84.In the state of Veracruz, Mexico, lowland and marginal coffee growing regions have been particularly vulnerable since the 1989 coffee crisis. Government programs have promoted production diversification as a strategy to improve local incomes and conserve environmentally beneficial shade-tree coffee agroforests. We present results on land use/land cover dynamics in the municipality of Zozocolco de Hidalgo from 1973 to 2006. The municipality is recognized for its indigenous population and poverty, and currently, diversification efforts are being implemented. Our study combines remote sensing and GIS analyses, binary logistic regression and econometric modeling, as well as socioeconomic surveys to evaluate land use/land cover change (LULCC) dynamics and explore potential environmental and socioeconomic drivers. Results show that tree cover and coffee agroforests had largely been conserved during the first decade after the coffee crisis. But, recent trends indicate loss of tree cover in coffee agroforests and their conversion mostly to pasture. Land use/land cover drivers are largely explained by spatially explicit environmental variables such as slope and elevation. Relevant socioeconomic variables such as distance to markets and land use profitability were not significantly related to land use changes in Zozocolco. Surveys revealed that many households had converted coffee agroforests to pasture or agriculture in the past decade and others intended on renting or selling their agroforest plots, mostly for conversion to pasture. Diversification programs may not be sufficient to stem deforestation in lowland and marginal coffee growing regions. Moreover, information about locally varying socioeconomic and cultural contexts needs to be strongly considered in order to formulate effective strategies.

DOI

[9]
Evans T P, Manire A, de Castro Fet al., 2001. A dynamic model of household decision-making and parcel level land cover change in the eastern Amazon.Ecological Modelling, 143(1): 95-113.The region around Altamira, Brazil, located in the Eastern Amazon, has experienced rapid landcover change since the initiation of government sponsored colonization projects associated with the construction of the Trans-Amazon Highway. The 30 years since colonization (1971) have been marked by a net loss of forest cover and an increase in the amount of cultivated/productive land, particularly for pasture and annual/perennial crop production. This research presents a parcel-level model of landcover change for smallholders in the Altamira study area. The utility of specific land-use activities is calculated to identify those land-uses that are most optimal at each time point, and labor is allocated to these activities based on the availability of household and wage labor. The model reports the proportion of the parcel in the following landcover classes at each time point using a 1-year interval: mature forest, secondary successional forest, perennial crops, annual crops and pasture. A graphical user interface is used for scenario testing, such as the impact of high/low (population) fertility, the increase of out-migration to urban areas, or changes in cattle and crop prices. The model shows a rapid reduction in the amount of mature forest in the 30 years following initial settlement, after which the parcel is composed of a mosaic of secondary succession, pasture and crops. The nature and rapidity of this landcover change is the function of a variety of household and external variables incorporated in the model. In particular, the model produces different landcover compositions as a function of demographic rates (fertility, mortality) and agricultural prices.

DOI

[10]
Fan C, Myint S, 2014. A comparison of spatial autocorrelation indices and landscape metrics in measuring urban landscape fragmentation.Landscape and Urban Planning, 121(1): 117-128.The combined use of remote sensing based land cover classification and landscape metrics has provided a positive step toward gaining a comprehensive understanding of the features of landscape structure. However, numerous limitations of land cover classification indicate that the utilization of classified thematic maps in landscape pattern analysis is questionable and may even lead to large errors in subsequent analyses. Instead of generating and employing detailed land cover classification maps, the utility of local spatial autocorrelation indices derived directly from satellite imagery to measure landscape fragmentation was examined. Since local spatial autocorrelation can capture spatial pattern at a local scale, it can be expected to detail the spatial heterogeneity for various parts of a landscape instead of providing a single value as in the case with the global measures. This study compares the traditional landscape metrics to the use of satellite imagery based local spatial autocorrelation measures in quantifying landscape structure over Phoenix urban area. Two local spatial autocorrelation indices, the Getis statistic and the local Moran's I were employed in evaluating landscape pattern, using normalized indices as the inputs. Results show that there is a clear relationship between local spatial autocorrelation indices and FRAGSTATS metrics. Both the Getis statistic and the local Moran's I can serve as useful indicators of landscape heterogeneity, for the entire landscape, and for different land use and land cover types. The paper provides a feasible methodology for urban planners and resource managers for effectively characterizing landscape fragmentation using continuous dataset.

DOI

[11]
Fan Q, Yang J, Wu Net al., 2013. Landscape patterns changes and dynamic simulation of coastal tourism town: A case study of Dalian Jinshitan national tourist holiday resort.Scientia Geographica Sinica, 33(12): 1467-1475. (in Chinese)lt;p>Based on multi-temporal land use data and remote sensing data, and also using Dalian Jinshitan National Tourist Holiday Resort as an example, through quantitative analysis of landscape ecology and the simulative method of CA-Markov model, this research systematically analyzes evolution characteristics of the landscape pattern in study area from 1998 to 2009, and simulates and predicts the landscape pattern in 2020. The results show that: 1) From 1998 to 2009, the main performance of landscape pattern changes is that the area of the tourism landscape increased 2.30 km<sup>2</sup> auxiliary tourism landscape area increased 2.22 km<sup>2</sup> the tourism landscape area reduced 5.27 km<sup>2</sup>; 2) The process of landscape pattern change shows that a gradual change from a single type of villagers living to various types which can satisfy the need of tourism&rsquo;s entertainment, sightseeing, business and other complex landscape. And change areas are mainly distributed in Longshan village, Manjiatan village, Chenjia village and Miaoshang village in southeast, from west to east in the study area. In addition, the priority change type is the &quot;traditional agriculture&rarr;artificial entertainment recreation&quot; model. 3) After analyzing and testing the availability of CA-Markov model, we get the landscape pattern simulation result of Jinshitan in 2020 that the area increased in artificial entertainment leisure landscape, public infrastructure landscape, traditional industry and resident landscape. By contrast the area of natural biological landscape, traditional agricultural landscape and other landscape decreased. In addition, the area reminded stable of the water landscape and transportation land use landscape. Furthermore distinct areas are mainly concentrated in Longshan village, Manjiatan village, Chenjia village and Miaoshang village in southeast from west to east in the study area.</p>

[12]
Feng S, Gao X, Gu Jet al., 2013. Land use spatial distribution modeling based on CLUE-S model in the Huangshui River Basin.Acta Ecologica Sinica, 33(3): 985-997. (in Chinese)In this paper,taking the Huangshui river basin of Qinghai province located in Qinghai-Tibetan Plateau and Loess Plateau as the study area,based on CLUE-S model,using land use data in 1987,1996 and 2007,land use spatial distribution pattern and future scenarios was simulated.Firstly,18 key driving factors including elevation,slope of between 0°—5°,5°—15°,15°—25° and 25°,aspect(flat aspect,eastern aspect,southern aspect,western aspect and northern aspect),distance to rivers,distance to roads,distance to urban and rural residential areas,population,local financial revenue and so on,were selected by using logistic step-regression method;Secondly,land use data in 1987 and 1996 were used respectively to simulate the spatial pattern of land use in 2007(the temporal scale was 20 years and 11 years respectively) for the Huangshui river basin.In the end,in order to validate simulation results of two temporal scale,simulated land use map in 2007 was respectively compared with actual land use map in 2007.The results showed that using CLUE-S model,at 250m spatial resolution level,the simulation accuracy reached respectively 88.31% and 89.96% at two temporal scale,and Kquality indices attained all 0.999,Klocation indices were 0.826 and 0.851 respectively,Kstandard indices were 0.826 and 0.851 respectively,Knoindices were 0.826 and 0.851 respectively,all Kappa indices were more than 0.82,which suggesting that CLUE-S model has a good applicability in simulating land use change in the Huangshui river basin and can be used to predict future land use change. Focusing mainly on slope cultivated land change of the study area,four kinds different scenarios of land use change for future 20 years(from 2008 to 2027) were constructed,and land use spatial pattern in 2027 in the study area was stimulated under above different scenarios by using the CLUE-S model.The simulation results indicated that in 2027 urban and built-up land was significantly increased under all four kinds different scenarios,mainly distributing in valley regions of the study area,whereas there was obvious spatial difference in the slope cultivated land change under scenario Ⅱ,Ⅲ and Ⅳ,that is scenario Ⅱ scenario Ⅲ scenario Ⅳ.The returning slope cultivated land will distribute in Ledu county,Minhe county,Datong county and Huangyuan county.The increased area of forest land will mainly distribute in south and north mountains of Xining city.The study conclusions will provide data reference and basic information of decision support for watershed future land use planning,management and policy-making.

DOI

[13]
Feng Y, Luo G, Zhou Det al., 2010. Effects of land use change on landscape pattern of a typical arid watershed in the recent 50 years: A case study on Manas River Watershed in Xinjiang.Acta Ecologica Sinica, 30(16): 4295-4305. (in Chinese)

[14]
Frazier A E, 2014. A new data aggregation technique to improve landscape metric downscaling.Landscape Ecology, 29(7): 1261-1276.Scale is a fundamental concept in landscape ecology and considerable attention has been given to the scale-dependent relationships of landscape metrics. Many metrics have been found to exhibit very consistent scaling relationships as map resolution (i.e., pixel or grain size) is increased. However, these scaling relationships tend to break down when attempting to ‘downscale’ them, and the scaling function is often unable to accurately predict metric values for finer resolutions than the original data. The reasons for this breakdown are not well understood. This research examines the downscaling behavior of metrics using various data aggregation techniques in an attempt to better understand the characteristics of metric scaling behavior. First, downscaling performance is examined using the traditional method of aggregation known as ‘majority rules’. Second, a new data aggregation technique is introduced that utilizes fractional land cover abundances obtained from sub-pixel remote sensing classifications in order to capture a greater amount of the spatial heterogeneity present in the landscape. The goal of this new aggregation technique is to produce a more accurate scaling relationship that can be downscaled to predict metric values at fine resolutions. Results indicate that sub-pixel classifications have the potential to transform data aggregation to allow more accurate downscaling for certain landscapes, but accuracy is linked to the spatial heterogeneity of the landscape.

DOI

[15]
Gong J, Liu Y, Xia B, 2009. Spatial heterogeneity of urban land-cover landscape in Guangzhou from 1990 to 2005.Journal of Geographical Sciences, 19(2): 213-224.

DOI

[16]
Han H, Yang C, Song J, 2015. Scenario simulation and the prediction of land use and land cover change in Beijing, China.Sustainability, 7(4): 4260-4279.Land use and land cover (LULC) models are essential for analyzing LULC change and predicting land use requirements and are valuable for guiding reasonable land use planning and management. However, each LULC model has its own advantages and constraints. In this paper, we explore the characteristics of LULC change and simulate future land use demand by combining a CLUE-S model with a Markov model to deal with some shortcomings of existing LULC models. Using Beijing as a case study, we describe the related driving factors from land-adaptive variables, regional spatial variables and socio-economic variables and then simulate future land use scenarios from 2010 to 2020, which include a development scenario (natural development and rapid development) and protection scenarios (ecological and cultivated land protection). The results indicate good consistency between predicted results and actual land use situations according to a Kappa statistic. The conversion of cultivated land to urban built-up land will form the primary features of LULC change in the future. The prediction for land use demand shows the differences under different scenarios. At higher elevations, the geographical environment limits the expansion of urban built-up land, but the conversion of cultivated land to built-up land in mountainous areas will be more prevalent by 2020; Beijing, however, still faces the most pressure in terms of ecological and cultivated land protection.

DOI

[17]
He D, Jin F, Zhou J, 2011. The changes of land use and landscape pattern based on Logistic-CA-Markov model: A case study of Beijing-Tianjin-Hebei metropolitan region.Scientia Geographica Sinica, 31(8): 903-910. (in Chinese)

[18]
Hu R, Dong S, 2013. Land use dynamics and landscape patterns in Shanghai, Jiangsu and Zhejiang.Journal of Resources and Ecology, 4(2): 141-148. (in Chinese)Land use change and landscape patterns have a large effect on land productivity and ecosystem biodiversity. Based on geographical information system technology and remote sensing data related to land use and land cover of Jiangsu and Zhejiang provinces and Shanghai (Jiang-Zhe-Hu area), we analyzed patterns of landscape change and predicted land use dynamics using the CA-MARKOV model. We also analyzed the conversion rate and area among landscape classes using the CA-Markov model. We found that from 1980 to 2005, there was a significant decrease in the area of farmland, and much of this landscape was transformed into settlements. Most of the landscape classes have become fragmented and isolated. The areas of farmland, settlement land and water tend to be complex in their shape and spatial clustering. The shapes of other land class patches have become simpler, and overall landscape fragmentation has increased. Landscape diversity and heterogeneity have increased. The CAMARKOV model predicted that settlement land will continue to grow from 2005 to 2015, but the speed of conversion will be reduced. The speed of the reduction in farmland and forest has increased, and increased settlement areas are clustered along the Yangtze River. Land use dynamics and change in the landscape pattern have affected land productivity and made the ecosystem more sensitive and fragile in this study region.

DOI

[19]
Hu Y, Zhang Y, Zheng X, 2013. Simulation of land-use scenarios for Beijing using CLUE-S and Markov composite models.Chinese Geographical Science, 23(1): 92-100.This study investigated and simulated land use patterns in Beijing for the year 2000 and the year 2005 from the actual land use data for the year 1995 and the year 2000,respectively,by combining spatial land allocation simulation using the CLUE-S model,and numerical land demand prediction using the Markov model.The simulations for 2000 and 2005 were confirmed to be generally accurate using Kappa indices.Then the land-use scenarios for Beijing in 2015 were simulated assuming two modes of development:1) urban development following existing trends;and 2) under a strict farmland control.The simulations suggested that under either mode,urbanized areas would expand at the expense of land for other uses.This expansion was predicted to dominate the land-use conversions between 2005 and 2015,and was expected to be accompanied by an extensive loss of farmland.The key susceptible to land-use changes were found to be located at the central urban Beijing and the surrounding regions including Yanqing County,Changping District and Fangshan District.Also,the simulations predicted a considerable expansion of urban/suburban areas in the mountainous regions of Beijing,suggesting a need for priority monitoring and protection.

DOI

[20]
Huang M, Zhang X, Zhang Jet al., 2012. A multi-scale simulation of land use change in Luoyugou Watershed based on CLUE-S model.Resources Science, 34(4): 769-776. (in Chinese)The CLUE-S model is a special model, which has already been widely applied into the dynamic studies about the spatial pattern of smaller-scale regional land use change. To investigate the applicability and the best applicable scale of the CLUE-S model in small watershed, this paper has simulated the spatial pattern of land use change in 2008 in Luoyugou watershed of TianShui City, Gansu Province with this model at five different scales (25m×25m, 50m×50m, 75m×75m, 100m×100m,and 125m×125m) based on the land use data of 2001(as the base year) and 2008 and seven driving factors for land use change (digital elevation model, slope, slope direction, curvature, the distance to road, the distance to town center and the distance to river). Besides, the simulated maps were verified by the present land use map of 2008. The results show that the application of the CLUE-S model has got quite favorable results in the study area, and the best simulation scale is 50m. Under this scale, the Kappa indices of cultivated land, orchard, forest land, meadow, residential area, water area and unused land are 95.71%, 88.97%, 90.68%, 74.66%, 74.53%, 96.89%, 81.94% respectively and the aggregated Kappa index value of all study areas is 92.34%. Therefore, it has been proved that the CLUE-S model can accurately simulate the land use change in small watershed. The spatial patterns of land use under three different scenarios: the natural increase scenario, the economic growth scenario and the ecological conservation scenario in 2020 were also simulated by this model. At last, by comparing the simulation maps under three different scenarios, we get a result that the land use changes mainly happen on the hillside land with a slope larger than 15 degree. Therefore, during the process of development and utilization of Luoyugou watershed, a series of appropriate measures about the ecological protection and the prevention of water and soil erosion should be taken at the same time to prevent the ecological environment of the watersheds from serious destruction.

[21]
Jenerette G D, Wu J, 2001. Analysis and simulation of land-use change in the central Arizona Phoenix region, USA.Landscape Ecology, 16(7): 611-626.lt;a name="Abs1"></a>To understand how urbanization has transformed the desert landscape in the central Arizona &#x2013; Phoenix region of the United States, we conducted a series of spatial analyses of the land-use pattern from 1912&#x2013;1995. The results of the spatial analysis show that the extent of urban area has increased exponentially for the past 83 years, and this urban expansion is correlated with the increase in population size for the same period of time. The accelerating urbanization process has increased the degree of fragmentation and structural complexity of the desert landscape. To simulate land-use change we developed a Markov-cellular automata model. Model parameters and neighborhood rules were obtained both empirically and with a modified genetic algorithm. Land-use maps for 1975 and 1995 were used to implement the model at two distinct spatial scales with a time step of one year. Model performance was evaluated using Monte-Carlo confidence interval estimation for selected landscape pattern indices. The coarse-scale model simulated the statistical patterns of the landscape at a higher accuracy than the fine-scale model. The empirically derived parameter set poorly simulated land-use change as compared to the optimized parameter set. In summary, our results showed that landscape pattern metrics (patch density, edge density, fractal dimension, contagion) together were able to effectively capture the trend in land-use associated with urbanization for this region. The Markov-cellular automata parameterized by a modified genetic algorithm reasonably replicated the change in land-use pattern.

[22]
Jones K B, Zurlini G, Kienast Fet al., 2013. Informing landscape planning and design for sustaining ecosystem services from existing spatial patterns and knowledge.Landscape Ecology, 28(6): 1175-1192.ABSTRACT

DOI

[23]
Li A, Wang A, Liang Set al., 2006. Eco-environmental vulnerability evaluation in mountainous region using remote sensing and GIS: A case study in the upper reaches of Minjiang River, China.Ecological Modelling, 192(1): 175-187.The upper reaches of Minjiang River-valley, located on the eastern edge of Qinghai鈥揟ibet Plain, is characterized by the complex distribution of hills and valleys. It is a typical and key mountainous region with apparent upland ecosystem vulnerability and sensitivity according to National Eco-environmental Renovating Scheme of china. In order to analyze eco-environmental vulnerability, remote sensing (RS) and geographical information system (GIS) technologies are adopted, and an environmental numerical model is developed using spatial principle component analysis (SPCA) method. The model contains nine factors including elevation, slope, accumulated temperature, drought index, land use, vegetation, soil, water-soil erosion, and population density. Using the model, the integrated eco-environmental vulnerability index (EVI) of study area in 1972, 1986, and 2000 are computed. According to the numerical results, the vulnerability is classified into five levels: potential, slight, light, medial, and heavy level by means of the cluster principle. The eco-environmental vulnerability distribution and its dynamic change in the last 30 years from 1972 to 2000 are analyzed and discussed. The results show that the eco-environmental vulnerability in study area is at medial level, and presents apparent vertical-belt distribution, and that driving forcings of dynamic change are mainly caused by human social economic activities and the contribution of late national eco-environmental protection policies, such as Natural Forest Protection and Grain for Green. According to these results, the study area is regionalized into three sub-regions, which may serve as a base for decision-making for eco-environmental recovering and rebuilding. The results of this study indicate that the method that integrates RS, GIS, and the SPCA to evaluate eco-environment vulnerability in mountainous region, cannot only distinctly represent the input subject spatial distribution of mountain vertical-belt feature, but also respect the river-valley as a whole system.

DOI

[24]
Li C, Yu F, Liu Jet al., 2011a. Research on land use/cover change and its driving force in midstream of the Heihe mainstream basin during the past 20 years.Journal of Natural Resources, 26(3): 353-362. (in Chinese)Since water allocation scheme in the Heihe River Basin (HRB) between Gansu Province and Inner Mongolia was executed after 2000, water use in the midstream in the HRB was limited by space-time restriction. Therefore, it is important to analyze the land use/cover change and driving force, and the problem of eco-environment. The results were as follows: 1) The land use structure changed greatly in the study area. From 1985 to 2000, farmland, woodland, residential and industrial land increased while grassland, water area and unused land decreased; from 2000 to 2005, farmland, residential and industrial land continued to increase, while the remaining land use types decreased. 2) Since the execution of the water allocation scheme in the HRB, the ecological environment deteriorated more quickly, the main reasons of which may be that the available water decreased and the farmland area increased quickly in such a short period. Grassland, woodland and water area decreased obviously, and the change ratios of grassland, woodland and water area were 2.14%, 7.36% and 3.69%, respectively. 3) From 1985 to 2000, the major patterns of land use change were mutual conversion of farmland and residential and industrial land, mutual conversion of farmland and unused land, the conversions of grassland to farmland; from 2000 to 2005, the major patterns of land use change were the conversions of unused land to farmland, mutual conversion of grassland and farmland. 4) There were obvious spatial differences in land use degree changes, and the greatest change took place in Ganzhou County. 5) The main forces of land use changes in the study area are population growth and economic development. However, the amount of the available water resources of the HRB influenced the land use changes more significantly after 2000.

DOI

[25]
Li P, Li X, Liu X, 2001. Macro-analysis on the driving forces of the land-use change in China.Geographical Research, 20(2): 129-138. (in Chinese)Land use is one complex system, which is affected by many factors, including both socio economic elements and natural resources and environment This paper firstly established a framework for macro analysis on the driving forces of land use change in China, in which economic welfare, environmental welfare, the need for food security and the advancement of science and technology are the main forces contributing to the land use changes Then the paper analyzed the basic contradictions in the land use change in recent years on the basis of current situation of land use and economic development in China In the third section, according to the analytical framework established firstly and the basic contradictions in the land use change, and on the basis of the land use data and other related socio economic documents in recent years, this paper analyzed and computed the indexes of the driving forces of the land use change and represented their spatial distribution in China By the distribution of these driving forces, the whole China can be seen as three distinct sub areas: Tibet southwest sub area, central northwest sub area and east sub area, each of which takes on different features in respect of the driving forces of land use changes As indicated in this paper, the economic welfare is the fundamental impetus to the land use changes, and the environmental welfare and the need for food security are also important factors in China The future land use change is ultimately decided by the performance of these factors and their interactions

DOI

[26]
Li S, Zhang X, Li Set al., 2013. Spatial and temporal evolution and mechanism of landscape pattern of oasis of urban in Xinjiang.Economic Geography, 33(12): 161-168. (in Chinese)The oasis of urban land use landscape is the main part of the city system,Xinjiang city is a typical representative of the oasis city,land use landscape evolution has its unique regularity,the sequence and the object of study is 18 oasis cities in 1995-2010 in Xinjiang as the study of,constructing the urban land use landscape diversity index,evenness index,dominance index,distribution reflect more urban land use multiple,complexity,stability,the degree of monopoly,And from 21 influence on the land use evolution of the landscape factors,through the statistical correlation screen out four dominant factors.The results of the study show that:①15 a Xinjiang oasis between urban land use landscape diversity showed a rising trend,the land use structure tend to be complex and changeable situation;Land use landscape uniformity evolution process presents spiral,differences between the cities;Urban land landscape monopolistic still exists,but tend to go down.②Oasis city land utilization landscape evolution process complex factors,multiple,and at the same time,cities because of the natural environment,social economic environment,policy environment,difference is bigger in output efficiency of urban land,Urban space expansion and ecological degradation of landscape of land use change present bidirectional recursive model,which is effect between the urban space external extension and internal erosion extrusion of ecological degradation.③Land economy into effective and urban expansion into impact oasis of urban land use landscape evolution of the main factors shows that human factors of urban land use become oasis landscape diversity,evenness and dominance in the evolution of the leading power.

[27]
Li X, Ding J, Wang Get al., 2014. Change of LUCC and characteristics of landscape pattern in a typical oasis in Turkmenistan.Journal of Desert Research, 34(1): 260-267. (in Chinese)Using four sets of remote sensing images in 1976, 1988, 2001 and 2011, and with ecological quantity analytical method, we analyzed the change of land use patterns and landscape pattern in about 40 years in the Murghab-Tejen Oasis in Turkmenistan. The results were as follows: (1) There happened a significant land use/cover change in the studied area. The areas of woodland, arable land, desert and saline soil showed increased, while that of grassland, water region and other area decreased. The fastest change rates of land use occurred in grassland (in 1976-1988), water region (in 1988-2001), grassland (in 2001-2011), and arable land (in 1976-2011), respectively. (2) The Total Area(CA), Percentage of Landscape(PLAND), Number of Patch(NP), Edge Density(ED), Landscape Division Index (LPI), Interspersion and Juxtaposition Index(IJI), Splitting Index(SPLIT) and Patch Cohesion Index (COHESION) at class level, and the Shape Area-Weighted Mean(SHAPE_AM), Patch Density(PD), Landscape Division Index (LPI), (ENN_MN), Contagion Index (CONTAG), Shannon's Diversity Index (SHDI), Simpson's Diversity Index (SIDI), Interspersion and Juxtaposition Index(IJI), Splitting Index(SPLIT) and Patch Cohesion Index (COHESION) at landscape level all showed that the number of the patches increased and landscape fragmentation presented an increasing degree. The connectivity of patches reduced. It would be essential to optimize the land-use pattern and to maintain the continuity of landscape ecological process and pattern in the Murghab-Tejen Oasis if sustainable development of land resources and ecological environment protection were expected.

DOI

[28]
Li Y, Deng Q, Zhang Det al., 2011b. Land use and ecosystem service value scenarios simulation in Danjiangkou reservoir area.Transactions of the Chinese Society of Agricultural Engineering, 27(5): 329-335. (in Chinese)Danjiangkou Reservoir is the water soure of the middle route of south-to-north water diversion project.Not only water flooded area will increase after the normal storage level in 2015,but also land-use will change.Based on the analysis of gray forecast,3 target situation including ecological security,economic development and integrated development were designed.Meanwhile,temporal and spatial simulations were conducted on the land use and its value of ecosystem services in the reservoir area with CLUE-S model.The results indicate that:the change of each ecosystem type,size and spatial distribution pattern which caused by land use change before and after raising the water level can directly affect the type and intensity of services provided by ecosystem.The maximum total value is the objectives scenario of ecological security,which is 16.19 billion Yuan and 16.72 billion Yuan respectively before and after raising the water level.The minimum total value is the economic development target scenarios,which is 15.903 billion Yuan to 15.373 billion Yuan corresponding.The total values of ecosystem services are improved in all scenarios.The value raises 525 million Yuan,530 million Yuan,530 million Yuan and 542 million Yuan respectively in gray prediction,ecological security and economic development target scenarios and integrated development goals scenarios.

DOI

[29]
Liang Y, Xu Z, Zhong F, 2011. Land use scenario analyses based on system dynamic model and CLUE-S model at regional scale: A case study of Ganzhou District of Zhangye City.Geographical Research, 30(3): 564-576. (in Chinese)Recently,scientists have developed different models of land use/cover change(LUCC) depending on their objectives and background.However,no single model is able to capture all of key processes essential to explore land use change at different scales and make a full assessment of driving factors and impacts.In this paper,we would like to make our efforts to develop an approach in combination of SD model and CLUE-S model to deal with some shortcomings of the existing LUCC models and to properly address the processes at different scales that give rise to the land use dynamics.The approach presented in this study will be helpful to understand the complexity of land use change and provide scientific support for land use planning and managements,and also can be used as data source in scenario analysis of different hydrological processes based on different underlying surfaces of LUCC.The objectives of the study are:(1) to develop an SD model to calculate and predict demands for different land use types at the macro-scale as a whole during the period 2000~2035,(2) to improve the characterization and presentation of the land use change processes by developing a CLUE-S model that will transfer and allocate land demands from SD model to spatially explicit land use patterns at a finer spatial scale(at 500 m resolution in our study),and the Kappa value of the land use map simulation in 2000 is 0.86 and the Kappa value is 0.81 in 2005,and(3) to discuss the advantages and disadvantages of combining and integrating the current land use change models.The further objective of this study is to find the key driving factors of LUCC(e.g.,human factors,including social capital,different cultural types and so on),and these factors should be represented as different spatial maps and integrated into the model analysis to improve land use change modeling and projection.

DOI

[30]
Liao W, Li L, Wu Yet al., 2011. Land use change and eco-environmental vulnerability evaluation in the Danjiangkou Reservoir Area.Journal of Natural Resources, 26(11): 1879-1889. (in Chinese)The Danjiangkou Reservoir is the source of water for the Middle Route Project under the South-to-North Water Transfer Scheme(SNWT) in China.The eco-environment of Danjiangkou Reservoir Area(DRA) plays an important role in water conservation and purification and would have significant implications for the economic prosperity in Hanjiang Basin as well as for the SNWT.In order to analyze the relationships between different land use patterns and eco-environment in the DRA,RS and GIS technologies were adopted,and an environmental numerical model was developed using spatial principal component analysis(SPCA) method.The land use and eco-environmental vulnerability dynamic change in the last 18 years were analyzed and discussed.The general tendency of land use change in the study area was that the proportion of forest,shrub,grassland,urban and rural land increased constantly from 48.76%,10.13%,5.32%,0.35%,0.47% to 50.41%,12.43%,6.91%,0.8%,1.49% between 1990 and 2007,and that of cropland and paddy field decreased gradually from 28.26%,0.64% to 20.90%,0.55% between 1990 and 2007.During this period,the land use pattern in the DRA was under the tremendous pressure from the conflict between the rapid urbanization,economic development and the eco-environmental protection and recovering.From 1990 to 2007,the average eco-environmental vulnerability synthetic index(EVSI) in the study area decreased from 5.96 to 5.56,which showed that the eco-environment of the DRA had been improved.However,the eco-environment in some areas even went worse despite of large-scale eco-environmental protection and recovering.In the DRA,most of forest,shrub and grassland were at potential or slight vulnerable levels,and the rest of the land use types basically at slight,moderate or heavy vulnerable levels.Land use types were closely correlated with eco-environment.The order of EVSI was: urban landrural residential landwaste landcroplandorchardgrasslandpaddy fieldshrub and forest land,which indicated that measures taken based on forest and shrub are the priority and important means for ecological restoration.Furthermore,the urban and rural residential land and cropland were the essential problems of the eco-environmental protection and recovering.

DOI

[31]
Liu J, Liu M, Zhuang Det al., 2002. Analysis of China’s recent change of land use spatial pattern. Science in China:Earth Sciences, 32(12): 1031-1040. (in Chinese)

[32]
Locantore N W, Tran L T, O’Neill R Vet al., 2004. An overview of data integration methods for regional assessment.Environmental Monitoring and Assessment, 94(1-3): 249-261.The U.S. Environmental Protections Agency's (U.S. EPA) Regional Vulnerability Assessment (ReVA) program has focused much of its research over the last five years on developing and evaluating integration methods for spatial data. An initial strategic priority was to use existing data from monitoring programs, model results, and other spatial data. Because most of these data were not collected with an intention of integrating into a regional assessment of conditions and vulnerabilities, issues exist that may preclude the use of some methods or require some sort of data preparation. Additionally, to support multi-criteria decision-making, methods need to be able to address a series of assessment questions that provide insights into where environmental risks are a priority. This paper provides an overview of twelve spatial integration methods that can be applied towards regional assessment, along with preliminary results as to how sensitive each method is to data issues that will likely be encountered with the use of existing data.

DOI PMID

[33]
Luck M, Wu J, 2002. A gradient analysis of urban landscape pattern: A case study from the Phoenix metropolitan region, Arizona, USA.Landscape Ecology, 17(4): 327-339.lt;a name="Abs1"></a>Urbanization is arguably the most dramatic form of land transformation that profoundly influences biological diversity and human life. Quantifying landscape pattern and its change is essential for the monitoring and assessment of ecological consequences of urbanization. Combining gradient analysis with landscape metrics, we attempted to quantify the spatial pattern of urbanization in the Phoenix metropolitan area, Arizona, USA. Several landscape metrics were computed along a 165 km long and 15 km wide transect with a moving window. The research was designed to address four research questions: How do different land use types change with distance away from the urban center? Do different land use types have their own unique spatial signatures? Can urbanization gradients be detected using landscape pattern analysis? How do the urban gradients differ among landscape metrics? The answers to these questions were generally affirmative and informative. The results showed that the spatial pattern of urbanization could be reliably quantified using landscape metrics with a gradient analysis approach, and the location of the urbanization center could be identified precisely and consistently with multiple indices. Different land use types exhibited distinctive, but not necessarily unique, spatial signatures that were dependent on specific landscape metrics. The changes in landscape pattern along the transect have important ecological implications, and quantifying the urbanization gradient, as illustrated in this paper, is an important first step to linking pattern with processes in urban ecological studies.

DOI

[34]
Luo G, Yin C, Chen Xet al., 2010. Combining system dynamic model and CLUE-S model to improve land use scenario analyses at regional scale: A case study of Sangong watershed in Xinjiang, China.Ecological Complexity, 7(2): 198-207.<h2 class="secHeading" id="section_abstract">Abstract</h2><p id="">Uses of models of land use change are primary tools for analyzing the causes and consequences of land use changes, assessing the impacts of land use change on ecosystems and supporting land use planning and policy. However, no single model is able to capture all of key processes essential to explore land use change at different scales and make a full assessment of driving factors and impacts. Based on the multi-scale characteristics of land use change, combination and integration of currently existed models of land use change could be a feasible solution. Taken Sangong watershed as a case study, this paper describes an integrated methodology in which the conversion of land use and its effect model (CLUE), a spatially explicit land use change model, has been combined with a system dynamic model (SD) to analyze land use dynamics at different scales. A SD model is used to calculate area changes in demand for land types as a whole while a CLUE model is used to transfer these demands to land use patterns. Without the spatial consideration, the SD model ensures an appropriate treatment of macro-economic, demographic and technology developments, and changes in economic policies influencing the demand and supply for land use in a specific region. With CLUE model the land use change has been simulated at a high spatial resolution with the spatial consideration of land use suitability, spatial policies and restrictions to satisfy the balance between land use demand and supply. The application of the combination of SD and CLUE model in Sangong watershed suggests that this methodology have the ability to reflect the complex behaviors of land use system at different scales to some extent and be a useful tool for analysis of complex land use driving factors such as land use policies and assessment of its impacts on land use change. The established SD model was fitted or calibrated with the 1987&ndash;1998 data and validated with the 1998&ndash;2004 data; combining SD model with CLUE-S model, future land use scenarios were analyzed during 2004&ndash;2030. This work could be used for better understanding of the possible impacts of land use change on terrestrial ecosystem and provide scientific support for land use planning and managements of the watershed.</p>

DOI

[35]
Luo Y, Yang S, Liu Xet al., 2014. Land use change in the reach from Hekouzhen to Tongguan of the Yellow River during 1998-2010.Acta Geographica Sinica, 69(1): 42-53. (in Chinese)In order to evaluate the effect of Grain for Green Policy in the reach from Hekouzhen to Tongguan of the Yellow River, based on dryness/wetness, geomorphic type, slope gradient and aspect, from macroscopic to microcosmic, three indicators, including (a) a land use change significance index, (b) a land use change proportion (c) a vegetation change index, are used to characterize the primary transformation types, the intensity of land use changes, and the degree of vegetation restoration in the period 1998-2010. The results show that: (1) The area of land use change accounted for 19.19% of the study area. High coverage grassland, forest, and other woodland increased significantly, while low coverage grassland, farmland decreased markedly. (2) Spatially, land use change was found primarily west of the Yellow River, between 35 and 38 degrees north latitude, including Malian River basin, Beiluo River basin, Yanhe River basin. (3) The transformation types, including low coverage grassland to moderate coverage grassland, moderate coverage grassland to high coverage grassland, farmland to other woodland, shrub to forest were the primary types resulting from land use change. (4) The effect of dryness/wetness, geomorphic type and slope gradient on land use change was significant, but that of aspect on land use change was not so clear.

DOI

[36]
Meng J, Wu X, Li Z, 2005. Land use/cover changes and its landscape ecological effects in the middle-western Hexi Corridor: A case study of Suzhou oases.Acta Ecologica Sinica, 24(11): 2535-2541. (in Chinese)In recent years, landscape pattern is often combined with land use/land cover changes to study the tendency of structure changes of land use/land cover and their ecological effects. Therefore, it can provide foundation for the planning and designing of sustainable land use. Landscape pattern is the result of different disturbing factors, which in turn, affects the ecological processes of regions. Suzhou was selected as the study area. It is located in the middle western Hexi Corridor. Studies on land use/land cover changes and their landscape ecological effects were carried out by GIS and Fragstats 3.3 according to landscape ecology principles on the basis of two composite Landsat 5 TM (Thematic Mapper)images of band 4,3,2 taken in 1988 and 2000. The results show the following occurred in the past 12 years: Firstly, landscape types are spatially different because of various different natural geographical conditions. For example, Cropland is mainly situated on both sides of the river and on the flood plain, while grassland is mainly located in the Qilian mountains and on the edge of the oasis. In addition, forestland is mainly distributed on both sides of the rivers and on the north slope belt of Qilian mountain. The Gobi and desert are broadly situated among oases, landscape is particularly characteristic patterns of Gobi. The oasis and desert are connected to each other tightly on the background of Gobi matrix. Secondly, great changes have taken place both in land use types and landscape pattern indices. Cropland and water area have increased evidently, and the same as urban land and cultural residences. While grassland, woodland and unused land had decreased, the high-covered grassland had disappeared completely. Thirdly, though the patch fragmentation was decreasing, and so the heterogeneity, and the ecological environment was degrading. For example, most grassland and woodland had been cultivated as cropland, which weaken the capability of maintaining ecological balance of oases. Also, it made the over-loaded oases more urgent in water. Fourthly, the patch dominance index decreased. Different patch types tended to distribute evenly. The decline of heterogeneity and the tendency of evenness certainly will lead to the decline of landscape stability. Fifthly, the edge effect of woodland and grassland decreased. As a result, the function of those kinds of patches in landscape declined. On the other hand, the edge effect of water, saline alkali and barren land increased, which showed the further degradation of landscape.

[37]
Ou W, Xiao J, Li W, 2014. Spatial pattern optimization simulation of coastal wetland use based on BP neural network and cellular automata: A case of Dafeng Coastal Wetland.Journal of Natural Resources, 29(5): 744-756. (in Chinese)In this paper, using BP neural network model and cellular automata(CA) model,23 spatial variables including natural, economic and social aspects which drive the beach reclamation of coastal vegetation are selected and the spatial pattern simulation model of wetland is established by using MATLAB R2010b software. With remote sensing image of Landsat TM/ ETM + in Dafeng, Yancheng coastal wetland in three periods of 1988, 2002and 2009, and the population and economic data of the respective year and based on the simulation thought of optimizing pattern, firstly, we use the wetland function niche model to diagnose historical excessively exploiting areas. Then we train the conversion rules of CA model belonging to the rational development rule. Finally, we simulate the space optimization pattern of Dafeng coastal wetland in 2016. To improve operational efficiency and accuracy of the model, we use two new methods in this study. One method is based on the cover quantity structure in the target year, which limits the proportion of training samples randomly, so as to improve 10000 samples of the effective information amount. The second is through comparing with simulating accuracy under different hiding layer nodes of BP neural network,"23-13-8"and"23-17-8"are determined as the optimal BP structures in two experiments. These two improvements enhance the efficiency of simulation, and the overall accuracy is higher than 93.6%. This research shows that, during these years, the succession among the cover types of Dafeng coastal wetland tends to be frequent, and the speed of reclamational year of natural wetland is obviously faster. At the same time, the development focus is constantly moving eastward, which shows the characteristics of"rolling development" to the sea. Finally, it is recommended to slacken the disorderly development in the surrounding core areas of Jiangsu Yancheng National Nature Reserve and Dafeng Milu National Nature Reserve, and to promote the orderly reclamation of tidal flats near Dafeng Port.

DOI

[38]
Overmars K P, Verburg P H, 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.In this paper, two research approaches to specify the relation between land use types and their explanatory factors are applied to the same modelling framework. The two approaches are used to construct land suitability maps, which are used as inputs in two model applications. The first is an inductive approach that uses regression analysis. The second applies a theoretical, actor decision framework to derive relations deductively using detailed field data. Broadly speaking, this classification coincides with the distinction between empirical and theoretical models and the distinction between deriving process from pattern and pattern from process. The two modelling approaches are illustrated by a scenario analysis for a case study in a municipality in the Philippines. Goodness-of-fit of the deductive approach in predicting current land use is slightly lower compared to the inductive approach. Resulting land use projections from the modelling exercise for the two approaches differ in 15 percent of the cells, which is caused by differences in the specification of the suitability maps. The paper discusses the assumptions underlying the two approaches as well as the implications for the applicability of the models in policy-oriented research. The deductive approach describes processes explicitly and can therefore better handle discontinuities in land use processes. This approach allows the user to evaluate a wide range of scenarios, which can also include new land use types. The inductive approach is easily reproducible by others but cannot guarantee causality. Therefore, the inductive approach is less suitable to handle discontinuities or additional land use types, but is well able to rapidly identify hotspots of land use change. It is concluded that both approaches have their advantages and drawbacks for different purposes. Generally speaking, the inductive approach is applicable in situations with relatively small land use changes, without introduction of new land use types, whereas the deductive approach is more flexible. The choice of modelling approach should therefore be based on the research and policy questions for which it is used.

DOI

[39]
Pan Y, Liu Y, Wang Jet al., 2011. Non-point pollution control for landscape conservation analysis based on CLUE-S simulations in Miyun County.Acta Ecologica Sinica, 31(2): 529-537. (in Chinese)Miyun County located in the North-East of Beijing City.Miyun reservoir in the Miyun County is the major drinking water source of the Beijing residents.Non-point pollution control of Miyun reservoir is the primary task in the Miyun County,in the mean while,agricultural development is also important for the income increase of the local people.This study was seeking to optimize the land-use spatial distribution which is one of the important factors influencing the watershed hydrology.Land-use change scenarios for non-point pollution control adapted to Miyun County were generated.Integrating landscape security pattern analysis with a land-use explicit model based on the 'Source-sink' theory was chosen as one method.Landscape security pattern is one of the potential spatial patterns in this region.This potential pattern which composed of strategic portions and distributions of the landscape has crucial importance in controlling certain ecological processes.In our case study,landscape security pattern in the Miyun area was focused and analyzed for non-point pollution control of the Miyun reservoir."Source-sink" theory assumed the land-use can be classified into two types: 'source' and 'sink' land types,based on their functions in pollutant transport and detention.The land-use change processes which will potentially contribute to the non-point pollution were identified in the low security area.Next two scenarios were simulated using the CLUE-S model.Scenario one investigates the land-use change based on historical land-use dynamics.Scenario two examines the development obeyed to a negative planning approach,which focus on the conservation of low security area.In the model operation,applied land-use dynamics are identical for both scenarios except for a land-use change restriction within the low security pattern area which will probably increase the non-point pollution.Results for scenario one depict two processes,one is the unused land reclamation by arable land and second a transformation of arable land to orchard.These two processes might have a great potential to contribute to non-point pollution.In scenario two,these two processes were restricted to the low security pattern area.The increased arable land and orchard patterns were transferred to a moderate security pattern area by the model in the simulation process.Probably vegetation recovery within the low security pattern area may enhance as major process the interception of the pollutant of this region.Although the total area of "risk source land" keep unchanged in the study region,the optimization of land-use patterns reduces non-point pollution risks.

DOI

[40]
Pontius R G, Schneider L C, 2001. Land-cover change model validation by an ROC method for the Ipswich watershed, Massachusetts, USA. Agriculture,Ecosystems & Environment, 85(1): 239-248.Scientists need a better and larger set of tools to validate land-use change models, because it is essential to know a model&rsquo;s prediction accuracy. This paper describes how to use the relative operating characteristic (ROC) as a quantitative measurement to validate a land-cover change model. Typically, a crucial component of a spatially explicit simulation model of land-cover change is a map of suitability for land-cover change, for example a map of probability of deforestation. The model usually selects locations for new land-cover change at locations that have relatively high suitability. The ROC can compare a map of actual change to maps of modeled suitability for land-cover change. ROC is a summary statistic derived from several two-by-two contingency tables, where each contingency table corresponds to a different simulated scenario of future land-cover change. The categories in each contingency table are actual change and actual non-change versus simulated change and simulated non-change. This paper applies the theoretical concepts to a model of deforestation in the Ipswich watershed, USA.

DOI

[41]
Røpke I, 2005. Trends in the development of ecological economics from the late 1980s to the early 2000s.Ecological Economics, 55(2): 262-290.By Inge Ropke; Trends in the development of ecological economics from the late 1980s to the early 2000s

DOI

[42]
Sheng S, Liu M, Xu Cet al., 2008. Application of CLUE-S model in simulating land use changes in Nanjing metropolitan region.Chinese Journal of Ecology, 27(2): 235-239. (in Chinese)lt;FONT face=Verdana>Land use and cover change (LUCC) models are the important tools in researching regional landscape dynamics and its driving mechanisms. With the application of CLUE-S (Conversion of Land Use and Its Effects at Small Regional Extent) model and under the support of Landsat remotely sensed data, this paper simulated the land use changes in Nanjing metropolitan region from 1998 to 2006. The results indicated that the land use changes were mostly affected by topography, and the distribution of urban land and agricultural land were significantly related with GDP per capita. Moreover, the urbanrural trunk roads made a much greater contribution to land use changes than provincial roads. Generally, highaltitude region tended to benefit the odds-ratio of woodland, while flat and low terrain benefited all the odds ratios of farmland and settlement. The accuracy of the simulation approached to 85.7% in 2003 and 84.1% in 2006 at 300 m spatial resolution, while as the parameters were recalculated according to the partial conditions and given divisionally, the accuracy of the model improved remarkably to 89.7% in 2003 and 88.3% in 2006. The results suggested that CLUE-S had a strong capability of predicting the changes of land use types, and even, the spatial structure of landscape, being available in urban planning and in researching the driving mechanisms of land use change.<BR></FONT>

[43]
Sohl T L, Claggett P R, 2013. Clarity versus complexity: Land-use modeling as a practical tool for decision-makers.Journal of Environmental Management, 129(16): 235-243.The last decade has seen a remarkable increase in the number of modeling tools available to examine future land-use and land-cover (LULC) change. Integrated modeling frameworks, agent-based models, cellular automata approaches, and other modeling techniques have substantially improved the representation of complex LULC systems, with each method using a different strategy to address complexity. However, despite the of new and better modeling tools, the use of these tools is limited for actual planning, decision-making, or policy-making purposes. LULC modelers have become very adept at creating tools for modeling LULC change, but complicated models and lack of transparency limit their utility for decision-makers. The complicated nature of many LULC models also makes it impractical or even impossible to perform a rigorous analysis of modeling uncertainty. This paper provides a review of land-cover modeling approaches and the issues causes by the complicated nature of models, and provides suggestions to facilitate the increased use of LULC models by decision-makers and other stakeholders. The utility of LULC models themselves can be improved by 1) providing model code and documentation, 2) through the use of scenario frameworks to frame overall uncertainties, 3) improving methods for generalizing key LULC processes most important to stakeholders, and 4) adopting more rigorous standards for validating models and quantifying uncertainty. Communication with decision-makers and other stakeholders can be improved by increasing stakeholder participation in all stages of the modeling process, increasing the transparency of model structure and uncertainties, and developing user-friendly decision-support systems to bridge the link between LULC science and policy. By considering these options, LULC science will be better positioned to support decision-makers and increase real-world application of LULC modeling results.

DOI PMID

[44]
Su S, Xiao R, Jiang Zet al., 2012. Characterizing landscape pattern and ecosystem service value changes for urbanization impacts at an eco-regional scale.Applied Geography, 34(4): 295-305.This paper qualitatively examined urbanization impacts at an eco-regional scale by analyzing landscape pattern and ecosystem service value changes in four eco-regions in the Hang-Jia-Hu region (China): the Hang-Jia-Hu Plains agricultural eco-region (region 1), the Tianmu Mountain forest eco-region (region 2), the Hangzhou urban eco-region (region 3), and the Qiandao Lake watershed forests-wetlands eco-region (region 4). Our results showed that the four eco-regions exhibited a similar urbanization process of rapid population growth, economic development and urban expansion. The considerable urban expansion led to a loss of 8.5 billion RMB yuan ecosystem service values per year on average between 1994 and 2003. The problems associated with urbanization were namely that the level of the landscapes within regions 1, 2, and 3 became increasingly more diverse, irregular, fragmented and isolated. Region 4 presented the opposite trend. Multivariate regression further qualitatively explored the dynamics of landscape changes in response to urbanization as well as the interactions between landscape pattern and ecosystem service values. It was found that landscape fragmentation, configuration and diversity, which were induced by urbanization, could significantly impair the provisions of ecosystem services. By discussing the inside meaning of the obtained regression models, we also discussed some implications for landscape planning. Our results highlighted the significance of joint application of landscape metric analysis and ecosystem service values assessment in landscape planning.

DOI

[45]
Sun P, Xu Y, Wang S, 2014. Terrain gradient effect analysis of land use change in poverty area around Beijing and Tianjin.Transactions of the Chinese Society of Agricultural Engineering, 30(14): 277-288. (in Chinese)Abstract: Being the core of the study on global environment change and sustainable development, land use/cover change (LUCC) gains more and more attention. The pattern of land use and its spatial-temporal change plays an important role in the study on land use/cover change. As global environment deteriorates, the study on land use change and its driving forces becomes an important research subject in the field of land science. The relationship between terrain variables and land use has become an important part of land use/cover change. Based on the remote sensing data of TM images in 1985, 1995, 2000, and 2010, we took the poverty area around Beijing and Tianjin as an example to analyze the terrain gradient characteristics of land use change by geo-informatics map analysis. In order to provide some recommendations on land use planning and land use management, we also explored the impact of topography on the selection of land use pattern and the causes of the terrain gradient effect of land use change. The results showed that 1) From 1985 to 2010 the land use types mainly distributed on the area of terrain relief under 200 m as well as slope of slope (SOS)) less than 15°. During the past 25 years, arable land, forestland and grassland were the dominant land use types. The areas of arable land, grassland and unused land decreased constantly, while the areas of forestland and construction land increased gradually. 2) The hierarchy was shown obviously in the distribution of land use types along the terrain gradient in poverty area around Beijing and Tianjin. Arable land, water area, construction land mainly were distributed on the area of low terrain gradient. The grassland mainly was distributed on the area of medium and high terrain gradient, while forestland and unused land mainly covered the high terrain gradient. 3) The construction land was obviously confined by the terrain niche, which was mainly located in the low terrain gradient area. While arable land acted as the main resources for transforming to construction land in the same area, it expanded to higher terrain gradient area in order to make up the area occupied by the construction of infrastructure facilities. The forestland was well restored in the area of higher terrain gradient, which expanded to mid-high terrain gradient area with the project of returning arable land to forestland. In addition, the predominant distribution area of unused land and grassland became gradually diminished. 4) The terrain gradient effect of land use change was the result of comprehensive effect of natural, socio-economic and policy factors. The natural factors played a key role in the formation of the terrain gradient effect of land use change. Socio-economic factors and policy factors were important push-drivers, such as human migration, construction of infrastructure facilities, and the project of returning arable land to forestland. This research was important in easing the conflict between human and the surrounding environment, and it was meaningful to the sustainable development of land resources in order to keep the ecological balance of the studied region. The study provided some alternatives for the dynamic optimal allocation of land use and the construction of eco-environment of the poverty area around Beijing and Tianjin.

DOI

[46]
Tang H, Wu W, Yang Pet al., 2009. Recent progresses of land use and land cover change (LUCC) models.Acta Geographica Sinica, 64(4): 457-468. (in Chinese)Land use and land cover change(LUCC) is a major part and also a main cause of global environmental changes,and it has emerged recently as an important focus for land change studies.Based on the systematic summary of the progress of studies in LUCC in the latest decade,including its theories,methods and applications,a series of problems that should be urgently resolved in the study are put forward,and some important study directions and priorities for future are reviewed.Results show that LUCC model plays an important role and is an efficient tool to support the analysis of the causes,processes and consequences of land use systems and to support land use planning and policy.Second,spatio-temporal patterns of LUCC are the research core of LUCC models.The development of models has experienced an evolvement from single non-spatial to the combination of non-spatial and spatial models,however,at present most models are static models and ignore the temporal dimension of land-use change.Third,feedback is one of the important characteristics of LUCC;however,the majority of the existing LUCC models are very weak in analyzing and presenting the feedbacks of LUCC.In this regard,how to get a better understanding of the feedbacks at different time and space scales will be one of new tasks in LUCC models.Fourth,the objective of LUCC models is to study the dynamic relations of a coupled human-environment.Currently,most LUCC models are partial-equilibrium ones.Future LUCC models will focus on studies on the human-environment system from a systematic and holistic point of view.Fifth,multi-scale analysis in LUCC models is needed for a better understanding of land use change.Early LUCC models used to take a single scale or level of analysis into account.Recently,a number of LUCC models which implement multiple scales can be distinguished.The scaling will be a key issue in future LUCC models.Finally,although many methods of model validation are available,there is not a uniform standard and criterion of model validation.The weakness in reference data also limits the performance of model validation.All these will challenge the development of future LUCC models.

DOI

[47]
Tian X, Wang X, Kadipov K Get al., 2014. Land use/cover dynamic change and landscape pattern analysis in Kayrakkum reservoir area during past 40 years.Transactions of the Chinese Society of Agricultural Engineering, 30(6): 232-241. (in Chinese)For discussing the change of land use/cover and landscape patterns in Kayrakkum Reservoir during the past 40 years, remote sensing images, i.e., MSS in 1975, TM in 1990, ETM+ in 2000, and TM in 2011, were used as the data source. The transfer matrix of land use/cover type map superposition was obtained using ENVI software, to study the land use area change, the single type of land use dynamic degree, and the indices of land use dynamic degree model of regional land use. Furthermore, the spatial-temporal characteristics of land use and landscape pattern changes in the Kayrakkum Reservoir area were systematically analyzed by using the landscape pattern quantity analysis method and by selecting different indices in type and landscape levels. Results showed that tremendous changes had taken place in the past 40 years for the land use/cover in Kayrakkum Reservoir area. Construction land area showed a sustainable increase; while low coverage grassland and unutilized land area tended a 鈥榁' shape over time. However, water, arable land and middle coverage grassland areas were all with inverted 鈥榁' shapes. Overall, the construction land and arable land expanded largely, while the middle coverage grassland, low coverage grassland and unutilized land decreased. Also, a large number of middle coverage grassland and unused land were changed into cultivated land. At the same time, part of the cultivated land was changed into construction land, which resulted in a significant increase in construction land and farmland. The study area changed with a 鈥榁' shape, with a gradually accelerating trend, The active land use/cover type changed from water at early and middle time to low coverage grassland recently, whereas the stable type of land use/cover changed from middle coverage grassland and arable land at an early time, to unutilized land at the middle time, and then recently to arable land. Landscape pattern analysis indicated that middle coverage grassland and arable land hold dominant positions in the whole landscape. More landscape elements were detected adjacent to aquatic land. Construction land showed more patch numbers and was relatively scattered in the trend of concentration distribution. The patch size tended to be uniform, landscape fragmentation and diversity of landscape types tended to increase, and the landscape aggregation degree tended to decrease under the influence of human activity. Moreover, low coverage grassland and arable land played a controlling role of the whole landscape. Because water is the source of an oasis' s life, Kayrakkum reservoir has occupied an important position in the region's ecological system. With the changes in the economic system and the impact of human activities in nearly 40 years, the area of the reservoir and the water level had experienced tremendous changes, which had an important impact on the dynamic change of land cover and landscape pattern. The result could provide a scientific basis for ecological environment protection, as well as sustainable use of water and land resources.

DOI

[48]
Tian Y, Jim C Y, Wang H, 2014. Assessing the landscape and ecological quality of urban green spaces in a compact city.Landscape and Urban Planning, 121(1): 97-108.The interrelated nature of landscape metrics calls for their joint application in analyzing complicated landscape patterns and associated ecological processes. Using geographic information system, remote-sensing and factor-analysis techniques, the landscape patterns of urban green spaces (UGS) in the compact city of Hong Kong were analyzed for their landscape-ecological quality in different land uses and districts. Using the Fragstat software, some key indices were selected to characterize the landscape mosaic with reference to patch size, patch shape, proximity relationship and edge configuration. Some old districts have smaller and more heterogeneous UGS than newer ones due to relatively low-quality landscape attributes. The landscape patterns of UGS have not improved with old-town renewal and new-town developments. In land uses with less human disturbance, UGS tend to be larger, and more homogeneous to enhance ecosystem services, and are closer to each other with more green cover to enhance connectivity and facilitate movements of organisms and people between proximal patches. Furthermore, vegetation-dominated land uses often have more complicated and hence longer UGS edges than other land uses to augment interfacial benefits. Of the 11 land uses, Government, institution and community and open space have more complicated UGS edges than more human-dominated types. The findings could inform the landscape-pattern design of UGS in compact cities to optimize their ecological qualities and benefits to both nature and residents, and to reinforce urban nature conservation.

DOI

[49]
Tian Y, Ren Z, 2012. Land use change simulations in loess hilly areas based on CLUE-S model: A case study in Xianyang loess tableland areas of Shaanxi Province.Progress in Geography, 31(9): 1224-1234. (in Chinese)Based on the remote sensing image interpretation in 2000, and combined with hydrology and water resources data, this paper evaluates optimal allocation of the soil and water resources in quantity and spatial configuration. On this basis, it simulates the land use change scenarios in Xianyang loess hilly areas through the application of binary logistic regression, CLUE-S model and SPSS 19.0 statistical analysis software. The results are shown as follows. (1) From the quantity results of optimal allocation of land and water, it is indicated that arable land, grassland, water area and unused land are decreasing, while the woodland and construction land areas are showing an increasing trend. (2) From the point of view of ecology, the results of optimal allocation of the soil and water resources in the loess hilly areas can reflect ecological value and economic benefits of land ecosystem. It is estimated that the ecological benefits would increase from 153.13 million yuan in 2000 to 154.45 million in 2020, and the land economic benefits would rise from 6.84901 billion to 7.19021 billion yuan in the 20 years. (3) From the results of spatial optimal allocation of water and land resources in the past 10 years and the next 10 years, we can come to a conclusion that the construction land has increased dramatically, and the increased area is mainly concentrated in the surrounding areas of original construction land. The areas with most significant construction land changes are mainly concentrated in the Qindu District of Xianyang city and Sanyuan county.

DOI

[50]
Veldkamp A, Lambin E F, 2001. Predicting land-use change.Agriculture, Ecosystems & Environment, 85(1): 1-6.Land use change modelling, especially if done in a spatially-explicit, integrated and multi-scale manner, is an important technique for the projection of alternative pathways into the future, for conducting experiments that test our understanding of key processes in land use changes. Land-use change models should represent part of the complexity of land use systems. They offer the possibility to test the sensitivity of land use patterns to changes in selected variables. They also allow testing of the stability of linked social and ecological systems, through scenario building. To assess current progress in this field, a workshop on spatially explicit land-use/land-cover models was organised within the scope of the Land-Use and Land Cover Change project (LUCC). The main developments presented in this special issue concern progress in: 1) Modelling of drivers of land-use change; 2) modelling of scale dependency of drivers of land use change; 3) modelling progress in predicting location versus quantity of land-use change; 4) the incorporation of biophysical feedbacks in land-use change models.

DOI

[51]
Verburg P H, Overmars K P, 2009. Combining top-down and bottom-up dynamics in land use modeling: Exploring the future of abandoned farmlands in Europe with the Dyna-CLUE model.Landscape Ecology, 24(9): 1167-1181.lt;a name="Abs1"></a>Land use change is the result of interactions between processes operating at different scales. Simulation models at regional to global scales are often incapable of including locally determined processes of land use change. This paper introduces a modeling approach that integrates demand-driven changes in land area with locally determined conversion processes. The model is illustrated with an application for European land use. Interactions between changing demands for agricultural land and vegetation processes leading to the re-growth of (semi-) natural vegetation on abandoned farmland are explicitly addressed. Succession of natural vegetation is simulated based on the spatial variation in biophysical and management related conditions, while the dynamics of the agricultural area are determined by a global multi-sector model. The results allow an exploration of the future dynamics of European land use and landscapes. The model approach is similarly suitable for other regions and processes where large scale processes interact with local dynamics.

DOI

[52]
Verburg P H, Soepboer W, Veldkamp Aet al., 2002. Modeling the spatial dynamics of regional land use: The CLUE-S Model.Environmental Management, 30(3): 391-405.lt;a name="Abs1"></a><div class="AbstractPara"> <div class="">Land-use change models are important tools for integrated environmental management. Through scenario analysis they can help to identify near-future critical locations in the face of environmental change. A dynamic, spatially explicit, land-use change model is presented for the regional scale: CLUE-S. The model is specifically developed for the analysis of land use in small regions (e.g., a watershed or province) at a fine spatial resolution. The model structure is based on systems theory to allow the integrated analysis of land-use change in relation to socio-economic and biophysical driving factors. The model explicitly addresses the hierarchical organization of land use systems, spatial connectivity between locations and stability. Stability is incorporated by a set of variables that define the relative elasticity of the actual land-use type to conversion. The user can specify these settings based on expert knowledge or survey data. Two applications of the model in the Philippines and Malaysia are used to illustrate the functioning of the model and its validation.

DOI PMID

[53]
Verburg P H, Veldkamp A, 2004. Projecting land use transitions at forest fringes in the Philippines at two spatial scales.Landscape Ecology, 19(1): 77-98.This paper presents two applications of a spatially explicit model of land use change at two spatial scales: a nation-wide application for the Philippines at relatively coarse resolution and an application with high spatial resolution for one island of the Philippines: Sibuyan island, Romblon province. The model is based on integrated analysis of socio-economic and biophysical factors that determine the allocation of land use change in combination with the simulation of the temporal dynamics (path-dependence and reversibility of changes), spatial policies and land requirements. Different scenarios of near-future developments in land use pattern are simulated illustrating the effects of implementing spatial policies. Results from the coarse scale model with national extent mainly serve to identify the overall pattern of land use change and 鈥榟ot zones鈥 of deforestation. The detailed application provides more insight in the pattern of land use change and its consequences for ecological processes. The use of the results for environmental assessments is illustrated by calculating spatial indices to assess the impact of land use change on forest fragmentation. It is concluded that spatially explicit modeling of land use change yields important information for environmental management and land use planning. The applications illustrate that the scale of analysis is an important determinant of the model configuration, the interpretation of the results and the potential use by stakeholders. There is no single, optimal, scale for land use change assessments. Each scale enables different types of analysis and assessment: applications at multiple scales therefore give complementary information needed for environmental management.

DOI

[54]
Vitousek P M, Mooney H A, Lubchenco Jet al., 1997. Human domination of earth’s ecosystems.Science, 277(5325): 494-499.Human alteration of Earth is substantial and growing. Between one-third and one-half of the land surface has been transformed by human action; the carbon dioxide concentration in the atmosphere has in

DOI

[55]
Wang Q, Meng J, Mao X, 2014. Scenario simulation and landscape pattern assessment of land use change based on neighborhood analysis and auto-logistic model: A case study of Lijiang River Basin.Geographical Research, 33(6): 1073-1084. (in Chinese)Tourism is identified to be capable of stimulating economic development in certain regions. However, tourist regions are experiencing a series of adverse effects with regard to local ecosystem associated with boosts in tourism. That is, the development of tourist region is able to manage a series of trade-offs between societal, economic and environmental goals. Given its integrated and dynamic land use form, tourist regions offer great samples to link physical and human systems and understand the anthropogenic effects and implications of land changes. Building on this premise, this paper seeks to model the physical-and human-induced landscape change by using Lijiang River Basin as a case. We employed a hybrid model of neighborhood analysis and Auto-logistic regression to project the likelihood map of land use distribution. Then, CLUE-S model is used to simulate future land use pattern under four policy-based scenarios: natural growth scenario, land planning scenario, resource conservation scenario and tourism development scenario. Several landscape indices are introduced to reveal the features of each pattern and to compare advantages of each scenario, which can provide scientific basis for future policy-making. Key results emerged: (1) Hybrid model of neighborhood analysis and auto-logistic regression is a more active and effective way, than the traditional logistic regression model is, to project the likelihood map of land use distribution. It helps to deal with the weakness of CLUE-s model on representing self-organizing character of land use change; (2) The Lijiang River Basin is a landscape dominated by the woodland and cultivated land. Vulnerability of landscape in the basin is primarily ascribable to the fragmentation of grassland and construction land and the fluctuated amount of water. To achieve an efficient land use in tourist region, a reasonable spatial regulation is more important than quantitative limits; (3) Tourism development requires a stable and diverse landscape. It indicates a necessity of controlling the interference of anthropogenic driven land use change on landscapes and managing the trade-off between socio-economic and eco-environmental land demand simultaneously. Result of this study may inform regional ecosystem management of ways to adapt sustainably to future change.

[56]
Wang S Y, Liu J S, Ma T B, 2010. Dynamics and changes in spatial patterns of land use in Yellow River Basin, China.Land Use Policy, 27(2): 313-323.Rapid land use change has taken place in many arid and semi-arid regions of China over the past decade, such as in Yellow River Basin. In this paper, changes in the land use pattern of the Yellow River Basin were analyzed using Landsat TM data in 1990, 1995 and 2000. The aim was to improve the understanding of changes in land use with a view to identifying potentially more sustainable systems o...

DOI

[57]
Wang X L, Bao Y H, 1999. Study on the methods of land use dynamic change research.Progress in Geography, 18(1): 81-87. (in Chinese)From the respect of the meaning and the researching contents of land use and cover change, the methods for land use change were analyzed. Further, the methods of establishing land use dynamic change model were mainly introduced. These models include land resources quantity change model, land resources environmental quality change model, land use degree change model, land use change regional diversity model, land use spatial change model and land requirement forecasting model.

DOI

[58]
Wassenaar T, Gerber P, Verburg P Het al., 2007. Projecting land use changes in the Neotropics: The geography of pasture expansion into forest.Global Environmental Change, 17(1): 86-104.

[59]
Weng Y C, 2007. Spatiotemporal changes of landscape pattern in response to urbanization.Landscape and Urban Planning, 81(4): 341-353.Specifically, at the class-level, residential land-use type shows the strongest positive relationship to the degree of urbanization in all of the class-level metrics adopted. Changes in residential land-use pattern were further analyzed with the number of housing units. The analyses revealed that there are different patterns of residential development along the transect in the study area鈥攚ith the core urban area expanding outward in a contiguous manner while the rural areas have scattered development. This study demonstrated the additional insights into landscape change by integrating the spatial and the temporal perspectives and by targeting the forms of residential developments.

DOI

[60]
Wu C F, Lin Y P, Chiang L Cet al., 2014. Assessing highway’s impacts on landscape patterns and ecosystem services: A case study in Puli Township, Taiwan.Landscape and Urban Planning, 128(8): 60-71.Highway construction facilitates urban growth in Taiwan. However, the long-term effects of transportation infrastructure are not well understood; these include land-use changes, changes in landscape patterns, and the alteration of ecosystem services. To assess the effects of different land-use scenarios under various agricultural and environmental conservation policy regimes, this study applies an integrated approach to analyze the effects of Highway 6 construction on Puli Township. Interviews with neighborhood leaders of Puli Township, along with remote sensing analysis, reveal that both biophysical and socioeconomic factors are the major forces driving land-use change. The effects of these land-use changes are varied. An example is the road-effect zone, which for Puli Township extends 400 m perpendicular to the length of the highway; however, due to differing spatial patterns it is highly asymmetric; indirect effects include the spatial restructuring of certain landscapes, which can drastically influence habitat dynamics. Land-use simulation results indicate that agricultural and environmental conservation policies have significant effects on projected land-use patterns in the southern part of Puli's downtown area and in areas along major roads. Specifically, highway construction and subsequent urbanization under various land-use policies result in varying degrees of isolation and fragmentation in the overall landscape pattern. A habitat quality assessment using the InVEST model indicates that the conservation of agricultural and forested lands improves habitat quality and preserves rare habitats. In summary, appropriate environmental policies will mediate both the direct and indirect impacts of Highway 6 on landscape patterns and ecosystem services in Puli Township.

DOI

[61]
Wu J, Feng Z, Huang Let al., 2011. CLUE-S based scenario prediction on sustainable land use: A case study of suburban district, Yangquan City.Resources Science, 33(9): 1699-1707.Determination of the most ecologically sound configuration for towns and cities inhabiting millions of people is of importance in China.Improving land use efficiency and optimizing the allocation of land resources are the key measures to promote regional economic sustainable development.This paper presents a land use planning framework for land supply and demand in suburban districts of Yangquan City,Shanxi Province.The situation of land use change was analyzed based on land use and socio-economic data.By applying the Analytic Hierarchy Process(AHP) method,the weights of driving factors of land-use change were calculated.By setting different development goals,different development paths of land use change on a 15-year timescale were analyzed.A possible way in land use planning field under the framework of CLUE-S model was proposed.Results show that the land priority principle,ecological protection principle,and efficiency principle could in turn be considered in land use planning.The ecological protection path,speeding plunder path,and balanced development path were set as three prediction scenarios.Compared to the ecological protection path and speeding plunder path,the balanced development path requires a farm land area of 18 131.57 hm2 and a construction land area of 14 069.40 hm2 to fulfill the goals that the self-sufficiency rate of food increase at a rate of 27% and the GDP grows at a rate of 13% per year.The balanced development path was the most appropriate development path of the land use planning scenario among the three scenarios formulated.Under the framework of the CLUE-S model,driving factors,conversion factors,and land demand jointly determined changes in land use,providing reference for land use planning.The three characteristics are able to describe the natural environment,social impact,and use cost,respectively.Therefore,urban land use planning should be made from the viewpoint of integrated study so as to ensure cultivated land quantity and quality,regional food security,and sustainable development among population,resources,and environment.In summary,the framework of CLUE-S was successfully applied to land use planning for the study area.A balanced development strategy for towns,based on land space analysis and pattern optimization,can provide an effective approach for space development and management in small-scale regions.This framework can also guide regions with similar spatial scales,improving the effectiveness and operability of land use planning.

[62]
Wu L, Hou X, Xu Xet al., 2013. Land use and landscape pattern changes in coastal areas of Shandong province, China.Transactions of the Chinese Society of Agricultural Engineering, 29(5): 207-216. (in Chinese)In this paper, the Spatial-Markov model, which was based on the theory of Markov process and spatial analysis techniques, was proposed to simulate land use change and landscape dynamics. By the Spatial-Markov model, the study area could be divided into numerous lattices and land use change in each lattices was simulated separately by the Markov process model. The outputs of the model include a set of ratio scale images and a nominal scale image. The whole process of the model was fulfilled by compiling programs with AML in ArcGIS 9.3. The coastal area of Shandong province was selected as the case study area. Land use maps were extracted based on Landsat TM/ETM+ images captured in 2000, 2005, and 2010 respectively. Firstly, characteristics of land use change and landscape dynamics were analyzed. It showed that, from 2000 to 2010, urban area and rural settlement expanded dramatically by massively occupying farmland, which, in turn, drove grassland reclaimed to farmland. At the landscape level, the landscape fragmentation increased, and both the diversity and evenness of the landscape increased. Secondly, using land use maps in 2000 and 2005, the Spatial-Markov model was developed to simulate the land use map in 2010 at a spatial scale of 500m. At the same time, the CA-Markov model was selected for model comparison, in specific, eleven driving factors were selected and the Logistic regression method was used to create the transitional maps for CA. Both Kappa coefficient and landscape indices were introduced to evaluate and compare the two models. It showed that the Spatial-Markov model not only achieved much higher Kappa coefficient, but also much better landscape indices than the CA-Markov model. Therefore, the Spatial-Markov model was applied to predict land use change and landscape dynamics in the next decade. Moreover, the prediction result shows that, from 2010 to 2020, areas of urban area and rural settlement will go on increasing, while areas of farmland will continue to decline. At the landscape level, all the landscape indices will follow their historical trend except for fractal dimension. As to the landscape indices at the class level, all landscape types will follow the same trend as before except for water and unused land.

DOI

[63]
Wu W, Zhao S, Zhu Cet al., 2015. A comparative study of urban expansion in Beijing, Tianjin and Shijiazhuang over the past three decades.Landscape and Urban Planning, 134(2): 93-106.Detailed comparative studies on spatiotemporal patterns of both urbanized area and urban expansion over a relatively long timeframe are rare. Here, we compared spatiotemporal patterns of urbanization in three major cities (i.e., Beijing, Tianjin and Shijiazhuang) in the Jing-Jin-Ji Urban Agglomeration using multi-temporal Landsat MSS, TM, and ETM+ images data of circa 1980, 1990, 1995, 2000, 2005 and 2010 integrated with Geographic Information System (GIS) techniques and landscape analysis approaches. A multi-scale analysis on the landscape responses to urban expansion from regional landscape to city and within city levels was performed. Results showed that urban area in Beijing, Tianjin and Shijiazhuang has expanded from 80102km 2 , 79502km 2 and 68202km 2 to 245202km 2 , 334302km 2 and 169902km 2 , increasing annually at 3.7%, 4.7% and 3.2%, respectively. Spatially, Beijing, Tianjin and Shijiazhuang have presented a mononuclear concentric polygon pattern, a double-nucleated polygon-line pattern, a sectorial point pattern, respectively, resulting primarily from their respective topographic constraints as well as urban planning and policy. Landscape responses to urban expansion varied with time and scale investigated, suggesting a general understanding on landscape metrics at regional or city level may fail to reveal detailed within city landscape dynamics under the impacts of urban expansion. The Jing-Jin-Ji Urban Agglomeration faces a great challenge to manage trades-offs between narrowing down intra-regional disparity and maintaining regional economic and ecological benefits.

DOI

[64]
Xu Y, Luo D, Guo Het al., 2013a. Multi-simulation of spatial distribution of land use based on CLUE-S model: A case study of Yuzhong County, Gansu Province.Acta Scientiarum Naturalium Universitatis Pekinensis, 49(3): 523-529. (in Chinese)Taking Yuzhong County of Gansu Province as a case study, based on land-use data in 1996, the authors adopted CLUE-S model to simulate land-use distribution in 2008 at the basic grid level (200 m×200 m), with the Kappa index 0.82. Then three scenarios of land-use spatial allocation in Yuzhong County in 2020, namely, farmland protection scenario, economic-oriented scenario and ecologically-oriented scenario, were established through designing different restrictions on land-use transition when running CLUE-S model in GIS environment. The results show that the scenario schemes have strongly influence on land-use distribution. Under farmland protection scenario, the farmland keep higher stability and tend to concentrated distribution characteristics. Arable land transition mainly occur in the north of the county and transit into grassland and forestland. While under economic-oriented scenario, construction land mainly distribute in Yuzhong Basin, where has better geological conditions. Ecological land such as farmland, grassland and forestland, take on characteristics under ecolo- gically-oriented scenario. Therefore, CLUE-S model is a powerful tool to simulate land-use spatial distributionof Yuzhong County, which provides a scientific basis for land-use planning and urban planning in the future.

[65]
Xu Y, Luo D, Peng J, 2011. Land use change and soil erosion in the Maotiao River watershed of Guizhou Province.Journal of Geographical Sciences, 21(6): 1138-1152.Due to the extremely poor soil cover, a low soil-forming rate, and inappropriate intensive land use, soil erosion is a serious problem in Guizhou Province, which is located in the centre of the karst areas of Southwest China. In order to bring soil erosion under control and restore environment, the Chinese Government has initiated a serious of ecological rehabilitation projects such as the Grain-for-Green Programme and Natural Forest Protection Program and brought about tremendous influences on land-use change and soil erosion in Guizhou Province. This paper explored the relationship between land use and soil erosion in the Maotiao River watershed, a typical agricultural area with severe soil erosion in central Guizhou Province. In this study, we analyzed the spatio-temporal dynamic change of land-use type in Maotiao River watershed from 1973 to 2007 using Landsat MSS image in 1973, Landsat TM data in 1990 and 2007. Soil erosion change characteristics from 1973 to 2007, and soil loss among different land-use types were examined by integrating the Revised Universal Soil Loss Equation (RUSLE) with a GIS environment. The results indicate that changes in land use within the watershed have significantly affected soil erosion. From 1973 to 1990, dry farmland and rocky desertified land significantly increased. In contrast, shrubby land, other forestland and grassland significantly decreased, which caused accelerated soil erosion in the study area. This trend was reversed from 1990 to 2007 with an increased area of land-use types for ecological use owing to the implementation of environmental protection programs. Soil erosion also significantly varied among land-use types. Erosion was most serious in dry farmland and the lightest in paddy field. Dry farmland with a gradient of 6掳鈥25掳 was the major contributor to soil erosion, and conservation practices should be taken in these areas. The results of this study provide useful information for decision makers and planners to take sustainable land use management and soil conservation measures in the area.

DOI

[66]
Xu Y, McNamara P, Wu Yet al., 2013b. An econometric analysis of changes in arable land utilization using multinomial logit model in Pinggu district, Beijing, China.Journal of Environmental Management, 128(15): 324-334.Arable land in China has been decreasing as a result of rapid population growth and economic development as well as urban expansion, especially in developed regions around cities where quality farmland quickly disappears. This paper analyzed changes in arable land utilization during 1993鈥2008 in the Pinggu district, Beijing, China, developed a multinomial logit (MNL) model to determine spatial driving factors influencing arable land-use change, and simulated arable land transition probabilities. Land-use maps, as well as social-economic and geographical data were used in the study. The results indicated that arable land decreased significantly between 1993 and 2008. Lost arable land shifted into orchard, forestland, settlement, and transportation land. Significant differences existed for arable land transitions among different landform areas. Slope, elevation, population density, urbanization rate, distance to settlements, and distance to roadways were strong drivers influencing arable land transition to other uses. The MNL model was proved effective for predicting transition probabilities in land use from arable land to other land-use types, thus can be used for scenario analysis to develop land-use policies and land-management measures in this metropolitan area.

DOI PMID

[67]
Yang X, Zheng X Q, Chen R, 2014. A land use change model: Integrating landscape pattern indexes and Markov-CA.Ecological Modelling, 283(13): 1-7.Landscape pattern indexes are quantitative descriptions of the spatial composition and configuration of land use, which can influence a variety of ecological phenomena. In this paper, we propose a land use change simulation model based on landscape pattern indexes, Markov chain and cellular automata. In the model, Markov Chain is applied to predict the amount of land use change; transition potential maps generated from natural and socioeconomic indexes are used to control the spatial distribution of land use; landscape pattern indexes in the start year are used to differentiate the transition probabilities of land use classes within different sub-regions of the study area. First, the principles and implementation of the model were described. Then the model was successfully applied to the simulation of land use change in Changping, a district of Beijing. Based on land use maps in years 1988 and 1998, the land use map in year 2008 was simulated. By analyzing the simulation result, the effectiveness of the model for land use change simulation was demonstrated. By comparing results simulated by this model and the results simulated by Markov-CA model with the actual land use map, the advantage of this model in spatial accuracy was shown.

DOI

[68]
Zeng Y, Jin W, Wang Het al., 2014. Simulation of land-use changes and landscape ecological assessment in eastern part of Qinghai Plateau.Transactions of the Chinese Society of Agricultural Engineering, 30(4): 185-194. (in Chinese)West Development and a series of ecological environmental regulation projects have been carried out in West China since 1999. In order to study the effects of West Development and ecological projects on future land use change, this paper combines Gray-Markov model and CLUE-S model to simulate land use/cover change in 2020 in the Eastern part of Qinghai Plateau. The validation of the CLUE-S model is verified by comparing the predictive model to a null model, and the former is higher in agreement due to the quantity and location of the latter, while they have the same agreement due to chance. According to actual conditions in the Eastern part of Qinghai Plateau three scenarios(natural scenario, farmland protection scenario and planning scenario) are designed, then it analyzes the changes of land use and land cover in these scenarios and assesses them by landscape pattern index and landscape ecological risk index. The forecast of land structure based on Gray-Markov model show that construction land and forest increase greatly due to West Development and a series of ecological environmental regulation projects, especially in the planning scenario, while irrigation farmland and dry farmland descend in the study period from 2009 to 2020. The development mode of the natural scenario is the same as that from 1999 to 2009, and there will be 633.98km2 farmland returning to forest and grassland and 117.66km2 unused land turning into forest and grassland by 2020; the farmland protection scenario strictly protects arable land, with142.00km2 farmland returning to forest and grassland and 130.71km2 unused land turning into forest and grassland; the planning scenario which integrates development of economic and environmental protection, will have 444.18km2 farmland returning to forest and grassland while 333.75km2 unused land for afforestation. The predictive results are assessed by four class-level indexes including number of patches(NP), percent of landscape(PLAND), largest patch index(LPI), landscape shape index(LSI), and eight landscape-level indexes comprised by NP, patch density(PD), edge density(ED), LSI, contagion(CONTAG), Shannon's diversity index(SHDI),Shannon's evenness index(SHEI), and aggregation index(AI). Class-level indexes indicate that woodland has the least landscape fragmentation in the planning scenario. Five of eight landscape-level indexes, containing ED, LSI,SHDI, SHEI and AI, indicate that the planning scenario is the best scenario of the three in intensive use of land. In conclusion, landscape evaluation and landscape ecological risk index show that the simulation results of the land planning scenario is reasonable. The planning scenario comprehensively coordinates the superior index, the goal of regional economic development and ecological protection, paying attention to the development of economic construction as well as the ecological environment construction, and improving or restoring the damaged ecosystem function to improve the overall productivity and stability of the landscape ecological system. Although new construction land expands strongly, strengthening ecological protection can still guarantee the regional landscape ecological security. Therefore, to 2020 the Eastern part of Qinghai plateau should continue to carry out the conversion of farmland to forest, but slightly lower than the strength during the period of 1999 to 2009.This study can provide evidence for the planning and land use policy formulation in the Western part of China.

[69]
Zhang D, Fu M, Tao Jet al., 2013a. Scenario simulation of land use change in mining city based on CLUE-S model.Transactions of the Chinese Society of Agricultural Engineering, 29(12): 246-256. (in Chinese)Abstract: The LUCC model is an important way to understand the process of land-use change, driving mechanisms, dynamic changes, ecological effects, and environmental impact assessment. Studies of land use models on land use prediction in mining cities are relatively rare. The CLUE-S model, as the representative of experience-based statistical models, has high simulation accuracy and application value in land use change prediction from spatial and temporal aspects. In this paper, the CLUE-S model was applied to land use change simulation in Wu'an, a typical mining city relying on GIS technology. First, selecting the correct driving factors is necessary to the accuracy of prediction map. 15 driving factors of land use were selected from 28 driving factors according to RDA(redundancy analysis) and factor analysis. Using 15 driving factors not only reduces the complexity of the problem, but also preserves simulation accuracy. In consideration of sustainable development, the free market scenario is more suitable compared with the alternatives. The expansion trend of the free-market mode is towards east and southeast, which conforms to the development planning of Wu'an city. Second, we set the corresponding land use quantity change in 2020 under three development modes by a Markov model and GM (1, 1) grey model, and then predicted the land use distribution map under a free market scenario, a cultivated land protection scenario, and an ecological protection scenario. The result showed that forest land and construction land increased under a free market scenario, reflecting the effects of economic development and environmental protection.. In the cultivated land protection scenario, cultivated land distribution restricted the free expansion of construction land. In the ecological security scenario, forestland grows obviously, and mining land reduces sharply influenced by environmental policy. Comprehensive consideration of sustainable development in Wu′an from the perspective of social, economic, ecological development and cultivated land protection, the free market scenario is relatively reasonable, and the other two scenarios can be a supplement for regional land optimal allocation. The results further verify that the CLUE-S model can simulate the future land use change of mining cities under different scenarios. Meanwhile, a new method to predict future land use under different scenario using the CLUE-S model can be applied in the implementation and management of land use planning, which can guide the land use change in the implementation process with reference to different simulation results by adjusting the land use objective and ultimately achieve the purpose of land use optimization.

DOI

[70]
Zhang H, Qi Z, Ye Xet al., 2013b. Analysis of land use/land cover change, population shift, and their effects on spatiotemporal patterns of urban heat islands in metropolitan Shanghai, China.Applied Geography, 44(5): 121-133.Using time series Landsat TM/ETM+ imagery and demographic data of Shanghai for 1997 and 2008, the relationship between land use/land cover (LULC) change and population shift and their effects on the spatiotemporal patterns of urban heat islands (UHIs) were quantitatively examined using an integrated approach of remote sensing, geographical information systems (GIS), and statistical analysis. The results showed that this city has experienced unprecedented urban growth and sprawl during the study period. The developed land increased by 219.50%, approximately 72.52% of which was converted from former cropland (24.79%), fallow land (21.21%), forest and shrub (18.97%), bare land (6.62%), and water (0.93%). Furthermore, in combination with the detection of LULC change, an analysis of the spatially differential growth rates for developed land area and population size revealed an urban鈥搒uburban鈥揺xurban gradient pattern of population shifting, as evidenced by a sharp increase in developed land area within the middle sub-zones at the urban fringe and the exurban sub-zones beyond the outer traffic ring. Consequently, changes in LULC and population shifts resulted in significant variation in the spatiotemporal patterns of the UHIs due to the loss of water bodies and vegetated surfaces. In the foreseeable future, substantial population growth and urban expansion will continue, especially in the rapidly urbanizing suburban and exurban areas, and thus, the extent and magnitude of UHI effects will continue expanding as well. The relationships between land use, the UHI effect, and regional climate change require that the underlying mechanisms, patterns, and processes of land conversion as well as the response of urban climate should be addressed throughout official decision-making processes. Thus, the planners and decision-makers could fully evaluate the environmental consequences of different land development scenarios and therefore improve the scientific basis of future planning and regulations.

DOI

[71]
Zhang L, Yang G, Liu J, 2014a. The dynamic changes and hot spots of land use in Fushun City from1986 to 2012.Scientia Geographica Sinica, 34(2): 185-191. (in Chinese)lt;p>This article uses GIS and RS technologies to study the dynamic changes and driving forces in land use of Fushun City in the two periods of 1986-2000 and 2000-2012. Its main findings are as follows: from 1986 to 2012 the main types of land use in Fushun City are woodland, farmland, water area and residential land, among which woodland gradually reduced, farmland and residential land gradually increased, and water area reduced first and then increased.The dynamic degree of the single land use in residential land is the highest, followed by water area and grassland, the last is woodland. It also finds out that the dynamic degree of the integrated land use from 1986 to 2000 in Fushun City is higher and spatial difference is little; while the dynamic degree of the integrated land use from 2000 to 2012 in Fushun City is lesser and spatial difference is obvious. This indicates that from 1986 to 2000 the whole city was being developed and its development degree was very high. But from 2000 to 2012 development activities of large-scale was only carried out in some local areas. There are three hot spots of dynamic changes in land use from 1986 to 2000 and they distribute in a larger range. While there is only one hot spot of dynamic changes in land use from 2000 to 2012 and it distributes in a smaller range. All of these four hot spots are located in the concentration area of Fushun residents. Natural factors are the essential conditions in change of land use. Contrary to natural factors, human activities have the decisive impact on the space and time change of land use. Human activities are also the main reasons for the rapid concentration change of land use in Fushun City. This article aims to put forward the best model for humans to use the land resources reasonably and provide a scientific basis for regional sustainable development.</p>

[72]
Zhang W, Liu M, Qi Y, 2014b. Land-use scenarios simulation based on the CLUE-S model in Kunming.Chinese Journal of Ecology, 33(6): 1655-1662. (in Chinese)lt;div style="line-height: 150%">Land-use model plays an important role in land-use change analysis, simulation and prediction. Land-use maps in 1986, 1996 and 2006 in Kunming were interpreted based on remote sensing images. The suitability of the CLUES model was estimated in Kunming with complex topography. Three scenarios were designed considering different polices and development trends for land use from 2007 to 2020. In the &ldquo;historic development trend scenario&rdquo; forestland area (matrix of the landscape) would decrease constantly, while build-up land and grassland area would increase. The landscape pattern would be more fragmental. In the &ldquo;planning scenario&rdquo; buildup land would increase rapidly, while farmland and forest land would decrease. The trend of landscape fragment would be less severe than that in the &ldquo;historical development trend scenario&rdquo;. In the &ldquo;ecologypriority scenario&rdquo; forestland would increase, while farmland and grassland would decrease, and the trajectory of landscape fragment and the landscape pattern would be more optimized. The simulated results of CLUE-S model provide a scientific support for land-use planning and policymaking in Kunming.</div><div style="line-height: 150%">&nbsp;</div>

[73]
Zhang X, Shi P, Luo Jet al., 2014c. The ecological risk assessment of arid inland river basin at the landscape scale: A case study on Shiyang River Basin.Journal of Natural Resources, 29(3): 410-419. (in Chinese)It is significant to make a reasonable assessment of ecological risk to optimize the landscape pattern, establish the ecological risk alarm mechanisms, minimize the risk of ecological environment and maintain the ecological function in river basin. The study, based on the remote sensing data of 1987, 2000 and 2010, chooses the typical arid inland river basin as the subject, divides the study area into 20 km× 20 km risk area and analyzes the temporal-spatial distribution pattern of ecological risk in Shiyang River Basin. Proceeding from the structure of landscape ecological system, the landscape disturbance degree index, the fragile index and the loss degree index are used to build the integrated ecological risk index(ERI) in the study with the help of spatial analysis of GIS. The results show that: 1) Great changes of landscape have taken place in the study area during 1987-2010, arable, woodland and grassland area have respectively reduced by 2.46×104hm2, 4.77×104hm2and 12.73 × 104hm2, while the residential land, water and unused land area have increased by 1.79×104hm2, 2.58 ×104hm2and 15.59 ×104hm2, which mean that the predominant of arable land, woodland and grassland decrease gradually, while the predominant of water, unused land and residential land increase gradually. 2) According to scope of ERI, 5 ecological risk grades are separated by ‘natural breaks'. If 0.12 ≤ERI ≤0.17, then the ecological risk grade is extremely low; if 0.17 ERI ≤0.22, the ecological risk grade is low; if 0.22 ERI ≤ 0.27, the ecological risk grade is medium; and if 0.27 ERI ≤0.32, then the ecological risk grade is high, or the ecological risk grade is extremely high. 3) At the early stage of the study period, the main ecological risk grades are extremely low and high, with the elapse of time, three ecological risk models coexist in the study area, i.e., extremely low, low, and extremely high, which means that the threaten of the ecosystem is increasing. It embodies in the space that the extremely low ecological risk areas shrink to upstream, and the area reduces 31.89×104hm2; the low ecological risk areas spread to the upper and middle stream, and the area increases 29.30 × 104hm2; the extremely high ecological risk areas expand to the downstream, and the area increases 58.69 × 104hm2. 4) There are seven conversion modes of ecological risk, among which the overall performance is from low-grade ecological risk to high-grade ecological risk. The conversion area from low-grade to high-grade is 122.56 × 104hm2, while conversion area from high-grade to low-grade is 6.12×104hm2.

DOI

[74]
Zhang Y, Zhao S, Zhang K, 2003. Simulation of changes in spatial pattern of land use in Horqin Desert and its outer area.Journal of Beijing Forestry University, 25(3): 68-73. (in Chinese)Due to over cultivation, over grazing, and other unreasonable land use methods, Horqin desert and its outer area in northern China have experienced significant land use changes, among which the decrease of grassland and agriculture expansion were the most important, which are contributing to severe environmental problems such as extension of desertification. The objective of this paper is to simulate changes in the land use pattern of Horqin desert and its outer area between 1985 and 2000. So, a land use change model, CLUE S, is presented for this simulation. It is based upon an empirical analysis of the spatial distribution of land use types in the study area which takes into account socioeconomic as well as geographical variables. The empirical analysis indicates that a reasonably complete description of land use distribution can be made by including slope, soil types, the distribution of settlement, and proximity to roads as variables. The quantity of land use types from 1985 to 2000 is obtained by linear interpolation of the real area in 1985 and 2000. The result shows that CLUE S is capable of classifying correctly 91 5% of the grid cells, for four categories: cultivated land and settlement (including villages and towns), forest, grassland, unused land and water. Kappa equals 0 90 in the simulation at the basic grid level (1 500 m脳1 500 m). It can be concluded that the model, CLUE S, constructed in this study is an important tool to explore near future land use change and is valuable for environmental management in Horqin desert and its outer area.

DOI

[75]
Zheng H W, Shen G Q, Wang Het al., 2015. Simulating land use change in urban renewal areas: A case study in Hong Kong.Habitat International, 46(2): 23-34.A considerable amount of research has been conducted on land use change, as it is extremely helpful when it comes to decision-making and policy formulation. Although land use change in urban renewal areas differs from that in new towns, very little research has focused on urban renewal and even less at the local or district level where most decisions need to be made. This study therefore developed a model for simulating land use change in urban renewal districts by combining the conversion of land use and its effects at small regional extent (CLUE-S) model and the Markov chain prediction model. The Yau Tsim Mong district of Kowloon in Hong Kong was the study area for the simulation, and historical land utilization data from 2000 to 2009 was used to validate the proposed model. By applying the validated model, four future land use scenarios were simulated for 2018 (the baseline scenario, the open space scenario, the residential scenario and the balanced scenario). The results not only indicate the effectiveness of the proposed model but also provide alternatives for future urban renewal based on different policy directions taken in the land use planning process.

DOI

Options
Outlines

/

〈 〉