- Email: [email protected]
Quantifying isolation effect of urban growth on key ecological areas
Ecological Engineering 69 (2014) 46–54
Contents lists available at ScienceDirect
Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng
Quantifying isolation effect of urban growth on key ecological areas Bin Xun a,b,c , Deyong Yu a,c,∗ , Yupeng Liu a,c , Ruifang Hao a,c , Yun Sun a,c a
State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China Department of Environmental Science, Northwest University, Xi’an 710069, China c Human-Environment System Sustainability (CHESS), Beijing Normal University, Beijing 100875, China b
a r t i c l e
i n f o
Article history: Received 22 April 2013 Received in revised form 17 January 2014 Accepted 29 March 2014 Keywords: Urbanization Ecological restoration Fragmentation Urban isolation index Ecological processes Landscape ecology
a b s t r a c t Urbanization may threaten multiple ecosystem processes as well as biodiversity in ecological reserves due to the loss, fragmentation and degradation of habitats in unprotected areas, making it necessary to measure the interference effect of urban growth on the protected areas. This study develops an urban isolation index (UII) to quantify the isolation effects of land-use conversion on key ecological areas (KEAs) and accounts for urban patch size, morphology, location, and quality loss. Based on the five remote sensing Landsat images taken between 1980 and 2005, UII is applied to analyze the spatio-temporal dynamics of the isolation effects caused by rapid urbanization in Shenzhen, China. The results showed that the habitat isolation constantly intensified during the urbanization process. We identified spatially explicit information of critical urban domains that exerted an interference effect on KEAs. There are considerable differences of isolation effects resulting from different urban growth types. Urban edge expansion caused the greatest habitat isolation in Shenzhen. The outlying and infilling urban patches contributed minor isolation effects to the KEAs. This paper furthers the understanding of the ecological effects of urbanization processes and provides a spatially explicit identification for ecological conservation and restoration in rapidly urbanized regions. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The urban population of the world grew almost exponentially during the late 18th century. Today, half of the world’s population resides in urban areas. It is predicted that the global urban population will increase to 69% by 2050 (UNPFA, 2008). Urbanization, a complex process that transforms landscapes formed by rural life styles into urban landscapes (Antrop, 2000), is becoming increasingly universal and irresistible throughout the world. Urbanization dramatically shapes new landscape patterns and significantly influences local and global ecosystem functions and dynamics. In most urban areas, urban spatial diffusion can fragment, isolate, and degrade natural habitats (Alberti, 2005). These negative effects on
Abbreviations: UII, urban isolation index; CUII, cumulative urban isolation index; NPP, net primary productivity; PSI, patch shape index; KEA, key ecological areas; SEZ, special economic zone; ERA, ecosystem restoration assessment. ∗ Corresponding author at: State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China. Tel.: +86 13681176750. E-mail address: [email protected] (D. Yu). http://dx.doi.org/10.1016/j.ecoleng.2014.03.041 0925-8574/© 2014 Elsevier B.V. All rights reserved.
habitats not only involve the direct extinction and reduction of native wildlife but also alter multiple ecological processes, such as animal movement, seed dispersal, genetic flow and nutrient cycling (Harrison and Bruna, 1999; Alberti, 2005; Crooks and Sanjayan, 2006; Grimm et al., 2008), and, ultimately, influence ecosystem services and human well-being (Alberti, 2005). Furthermore, urban development is facing the dilemma of urban land intransigence or natural habitat degradation. The only way to face this challenge is to develop sustainable practices, i.e., economically, socially and ecologically acceptable solutions, to mitigate the conflicts between urban development and ecological conservation in urbanized areas (Barot et al., 2012). Ecological engineering is defined as the design of sustainable ecosystems that integrate human society with the natural environment for the benefit of both (Mitsch, 2012). Its goals involve the creation or restoration of ecosystems that can provide sustainable services for humans and other forms of life (Mitsch, 2012). Since the 20th century, establishing protected areas has become the fundamental strategy of regional ecological conservation and restoration (Howard et al., 2000). Ecological reserves may be the most direct and effective measure to maintain species persistence and biodiversity, especially in urban areas. During
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
urbanization processes, however, ecological reserves are likely to cause ecological isolation among protected areas if the surrounding unprotected vegetation habitats undergo anthropogenic disturbances, and can even lead to a conversion to human-dominated landscapes (Grumbine, 1994). In other words, from the viewpoint of ecosystem integrity, whether reserves can provide sustained and effective protection for organisms depends not only on the protected areas themselves but also on the natural context outside of the protected boundaries (Hansen and DeFries, 2007). This is largely because some of the key areas excluded from the protected domains are important to protect wildlife from exposure to a hostile landscape matrix and to maintain nutrient flows, organism movements and population processes (Grumbine, 1990). It is therefore essential and implicative for regional-scale ecological conservation and restoration, especially in a large-range urbanization context, to evaluate the isolation effect of urban growth on the ecological reserves in a quantitative and comprehensive manner. The quantification method of landscape fragmentation and connectivity has always been a hot topic for ecologists (Lindenmayer et al., 2008). Early proposed landscape metrics, such as the patch density, mean patch size, perimeter-area ratio and fractal index, can only capture the separate aspects of fragmentation (Davidson, 1998). Recently, some new metrics, such as the landscape division, splitting index and effective mesh size, have been developed to characterize landscape fragmentation from a geometric perspective (Jaeger, 2000; Moser et al., 2007). Landscape connectivity is defined as the degree to which the landscape facilitates or impedes the dispersal movement of species among habitat patches that exist in the landscape (Taylor et al., 1993). Graph theory is thought to be a promising methodology in the assessment of landscape connectivity (Urban et al., 2009). Some scholars have proposed a series of graph-based metrics to quantify the network connectivity related to species dispersal processes (e.g., Urban and Keitt, 2001; PascualHortal and Saura, 2006; Saura and Pascual-Hortal, 2007; Yu et al., 2013). In general, the metrics used to assess landscape fragmentation and connectivity are well developed; some of them have even been successfully applied to biological conservation and ecological restoration (e.g., Cabeza, 2003; Zetterberg et al., 2010; Yu et al., 2012). However, most of the metrics concentrate on the fragmentation features of a single landscape type and thus cannot reflect the interaction among landscape types during land-use conversion processes. Habitat fragmentation may cause functional isolation due to the weak exchange of genes for both individuals and populations, which explains why fragmented habitats often contain fewer species than contiguous habitats (Forman et al., 1976). Habitat isolation related to interference effects is a more complex process than fragmentation in the context of urbanization. Recently, two metrics, the insulation degree (ID) (Su et al., 2010) and urbanization isolation effect (UIE) (Ng et al., 2011), were proposed to examine the isolation effect caused by urban areas. These metrics were described by the distance of urban patches to the natural patch and the area of urban patches within a specified radius. In the current study, we developed a new urban isolation index (UII) by comprehensively considering the interaction effects of urban size, morphology and quality losses on key ecological areas (KEAs) during land-use conversion processes. We evaluated the spatiotemporal dynamics of the isolation effect of urban growth on KEAs in Shenzhen, an area of China that has been rapidly urbanized. Our objective is to quantitatively answer two questions: (1) What degree of isolation do KEAs undergo during the urbanization process? (2) Which type of urban growth causes the most serious habitat isolation?
47
2. Material and methods 2.1. Study area Shenzhen is a coastal city in southern China, neighboring Dongguan and Huizhou in the north and Hong Kong in the south, and is flanked by the Daya Gulf in the east and the Pearl River Estuary in the west. The city consists of a special economic zone (SEZ), Bao’an District and Longgang District (Fig. 1). It lies between 22◦ 26 N and 22◦ 51 N latitude and between 113◦ 45 E and 114◦ 37 E longitude, and its total area is approximately 2020 km2 . With a subtropical marine climate, the annual average temperature, precipitation and sunshine hours are 22.4 ◦ C, 1933.3 mm and 2020 h, respectively. The rainy season is mainly concentrated from April to September. The dominant vegetation is an evergreen broad-leaved mixed forest. There are 1889 types of wild vascular plants, among which 22 belong to the nationally rare and endangered plants list (Xing and Yu, 2000). There are five types of nationally protected Class I wild animals and 43 nationally protected Class II wild animals (SWCA, 2009). Shenzhen was municipalized in 1979 and established as the first SEZ in China in 1980. Over the past 30 years, it has evolved from a border fishing village into a metropolitan region with a population of 10.35 million and a GDP of 951.10 billion Yuan in 2010 (SSB, 2011). Meanwhile, Shenzhen has experienced significant land-use changes. In 1980, urban lands and forests occupied 0.63% and 38.67% of the whole city area, respectively. By 2005, urban land had increased to 33.52% and forests had decreased to 29.81% of the total area (see the remote sensing Landsat data in Section 2.2). Because of the habitat fragmentation caused by this urban growth, biodiversity has only remained in isolated mountain areas. In 2005, the Shenzhen government promulgated a legally binding regulation stating that KEAs need to provide a survival environment for wild species, maintain ecosystem processes and regional biodiversity, and promote environmental quality for human well-being. KEAs contain (Fig. 1): 1) water source protection areas, scenic spots, nature reserves, and primary farmland protection areas, forests and country parks; 2) mountains and woodlands with a slope over 25◦ and uplands with an elevation over 50 m; 3) main trunk rivers, reservoirs and wetlands; 4) green space for the maintenance of ecosystem integrity; and 5) peninsulas and other ecologically important coastal areas. However, the impacts of urban growth on these ecological areas have not undergone scientific assessment until now.
2.2. Data processing We used five cloud-free Landsat images (including a multispectral scanner (MSS) satellite image with a resolution of 79 m × 79 m from 1980; three thematic mapper (TM) satellite images with a resolution of 30 m × 30 m from 1988, 1994 and 2005; and an enhanced thematic mapper plus (ETM+) satellite image with a resolution of 30 m × 30 m from 2000) to produce the land-use thematic maps. Other auxiliary data included topographic maps at a 1:50,000 scale from the 1960s and 1970s, an administrative boundary map at a 1:100,000 scale from 2005, and aerial photographs with high resolution and field survey data to carry out the geometric correction and land-use classification. We obtained 389 evenly distributed ground control points (GCPs) with exact locations and land cover information to perform the geometrical correction and classification. Different land-use types were categorized using both unsupervised classification and supervised classification algorithms. Based on the normalized
48
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Fig. 1. Location of the study area and spatial distribution of the KEAs (1: Luotian KEA, 2: Qiankeng KEA, 3: Maluan Mountain KEA, 4: Xichong-Qiniang Mountain KEA, 5: Baguang KEA, 6: Meilin-Tanglang Mountain KEA, 7: Tiegang-Yangtai Mountain KEA, 8: Guangming KEA, 9: Huangzhukeng KEA, 10: Fenghuang Mountain KEA, 11: Wutong Mountain KEA, 12: Silver Lake KEA).
difference vegetation index (NDVI) calculated from Landsat Bands 3 and 4, ground objects were categorized into three groups by the unsupervised classification method: non-vegetation, vegetation and water. This process maximally reduced artificial errors and performed the appropriate classification for further interpretation. Each group was further classified by using a maximum likelihood supervised classification to obtain the land-use thematic map. Six land-use types were classified as forest, orchard, urban land, farmland, water and vacant land (Fig. 2). The classification results were considered reasonable, with an overall accuracy of 83% in 1980, 84% in 1988, 86% in 1994, 87% in 2000 and 89% in 2005. 2.3. Urban growth types Generally, there are three urban growth types — edge expansion, infilling and leapfrog. Edge expansion refers to an urban sprawl in a contiguous manner that proceeds outward from the existing urban area. Infilling refers to urban expansion in the gap-infilling type among historical urban centers. Leapfrog means that new urban patches are formed and have no common boundary with the existing urban area. This study identified the three urban growth types through a quantitative method (Xu et al., 2007) to detect their individual isolation effect on ecological reserves. The equation is given by: S=
Lc P
(1)
where Lc is the length of the common boundary between the newly grown and pre-growth urban patch and P is the perimeter of this newly grown patch. Urban growth type is identified as infilling when 0.5 < S ≤ 1, edge expansion when 0 < S ≤ 0.5, and leapfrog when S = 0, which indicates no common boundary. 2.4. Landscape isolation index The new metric of UII was developed to evaluate the isolation effect of the newly grown urban patches on KEAs. Here, UII is calculated by the parameters of the quality loss, shape, growth types and location of new urban patches within a specified radius of KEAs to quantify the comprehensive effect of land-use conversion on KEAs. UII is defined as follows: UIIi =
Qi L L − Lci × i × i × Qt Lt 2 Ai
d −1 i
d
(2)
where Qi is the quality loss caused by a newly transformed urban patch i. We use net primary productivity (NPP) loss as a proxy of the quality reduction, which equals the product of the area of a newly transformed urban patch and the NPP of previous land-use type per unit. Qt is the total NPP loss resulting from urban growth. Primary production can provide a reliable basis for stratifying surveys of biodiversity and environmental quality by highlighting the areas of potentially high biodiversity across large areas (Bailey et al., 2004). In this study, NPP is calculated by our previously improved
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
49
Fig. 2. Land-use maps in Shenzhen City from 1980 to 2005.
CASA-NPP model (Yu et al., 2009, 2011). di is the distance from the urban patches i to the nearest KEA j . d is the mean nearest distance of all urban patches to all habitats in the study area. The distance is measured by the edge-to-edge distance. Ai is the area of the urban patch i. Li is the perimeter of the urban patch i. Lci is the length of common boundary between the newly grown and the previous urban patches. Lt is the total perimeter of the newly grown urban patches. Li /2 Ai is the patch shape index (PSI) (McGarigal et al., 2002) and represents the shape of the urban patch i in this study. When the natural landscape is transformed into urban land with significant NPP loss, the suitable habitats or available resources for species will be reduced, and the isolation degree among habitats will be strengthened. The interference or isolation effect of newly grown urban patches is also related to the patch shape and edge adjacent to the non-urban matrix. The more complex and longer the common boundaries between the urban and non-urban areas, the stronger the isolation effect of urban expansion becomes. When the urban growth belongs to the leapfrog type (S = 0 in Eq. (1)), meaning that the patch edge is completely adjacent to other landuse types, UII is relatively higher. New urban patches in full infilling growth (S = 1 in Eq. (1)) have the smallest UII (UII = 0). The distance between the urban patch and the KEA is negatively related to the isolation effect, namely, the closer the distance, the higher the UII value.
The cumulative urban isolation index (CUII) represents the total isolation effect of ecological reserves undergoing urban growth, which is defined as: CUIIj =
n
UIIi
(3)
i=1
where CUIIj is the sum of the isolation effect of newly grown urban patches on KEAj within a certain radius. UII and CUII with no dimension facilitate the comparison of the results in different periods and regions. In this study, we defined the mean nearest distance of all urban patches to all habitats (d) as the buffer radius. 3. Results 3.1. Spatiotemporal dynamics of the urban isolation effect The radius d sharply decreased from 1988 to 2000, indicating that urban development constantly extended toward KEAs (Table 1). The urban area increased by 251.1%, from 68.18 km2 in 1988 to 239.37 km2 in 2005 within the radius d around KEAs. There was a constant loss in the NPP resulting from the newly converted urban land. The NPP was 0.63 × 108 gC a−1 in 1988 and reached 2.52 × 108 gC a−1 in 2005. The rising trend in PSI indicated that the
50
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Table 1 Main features of new growth urban patches within a radius of d¯ and the cumulative impacts of UII and CUII on the baseline of land use in 1980. Years
d¯ (m)
Urban growth area (km2 )
NPP loss (108 gC a−1 )
PSI
UII Max
UII Mean
UII Sum
CUII Max
CUII Mean
1988 1994 2000 2005
2386 2182 2042 2041
68.18 149.35 213.66 239.37
0.63 1.54 2.27 2.52
1.46 1.55 1.58 1.67
10.66 85.82 118.84 368.27
0.02 0.31 0.85 3.23
23.56 505.51 1326.45 4534.72
11.99 85.98 119.66 368.66
1.58 17.52 41.80 127.31
Note: The earliest available time of the MODIS product during one year was in 2000, thus we used the NPP in this year to replace the NPPs from 1980, 1988 and 1994 in places where the land-use types did not change. When the land covers were altered, the NPPs from 1980, 1988 and 1994 were obtained by calculating the average NPP of each land-use type in 2000.
shape of the new urban patches became increasingly irregular and complex. The cumulative isolation impacts represented by UII and CUII over different times are listed in Table 1. The maximum, mean and sum of UII in 2005 are 33, 161 and 197 times larger than those in 1988, respectively. The greatest increments for the mean and sum UII occurred between 1988 and 1994, with a value of 1450% and 2046%, respectively. Similarly, the mean and sum of CUII increased by 30 and 80 times from 1988 to 2005, respectively. The largest increment of the mean CUII was 1009%, which occurred from 1988 to 1994. The spatiotemporal dynamics of the isolation effect are shown in Fig. 3. In 1988, the urban patches with a stronger isolation effect were mainly concentrated in the middle part of the SEZ. Silver Lake Forest Park (KEA12) was the most seriously affected KEA, with the CUII of 11.99%. In 1994, the areas with the higher isolation effect presented a remarkable spatial expansion from the SEZ toward other areas. Meilin-Tanglang Mountain Forest Park (KEA 6) underwent the strongest amount of isolation, and Silver Lake Forest Park ranked second in this metric. The interference caused by the urban sprawl enlarged further throughout the entire city after 2000. In contrast to the Longgang District, the SEZ and Bao’an had a higher UII and CUII. In 2005, except for KEA 4 and 5, the other KEAs were strongly isolated. 3.2. Isolation effect caused by the three urban growth types The three urban growth types were identified and compared for their isolation effect within the radius d around KEAs during the four periods from 1980–1988, 1988–1994, 1994–2000 and 2000–2005 (Tables 2 and 3). Urban growth was dominated by edge
expansion, occupying 60–70% of the total new urban areas in the four periods from 1980 and 2005. NPP loss was mainly caused by edge expansion, amounting to 64–70% of the total loss in the four periods from 1980 and 2005. Compared with the other three periods, leapfrog growth resulted in a relatively higher NPP loss of 34% from 1980 and 1988. Urban edge expansion ranked the highest in shape complexity during the four periods, followed by infilling and then leapfrog growth. Edge expansion had the greatest isolation influence on KEAs, accounting for over 90% of the total UII from 1980 and 2005 (Table 3 and Fig. 4). According to the maximum, mean and sum of UII for the newly developed urban patches, the stronger isolation effect caused by urban edge expansion mainly occurred in the SEZ during the periods from 1980–1988 and 1988–1994. The isolation effect caused by the infilling and leapfrog growth was less. Compared with the other three periods, the isolation caused by leapfrog urban growth was relatively larger from 1980 to 1988, which mainly occurred in the middle and west of Shenzhen. The contribution ratios of the three urban growth types to habitat isolation are illustrated in Fig. 5. Edge expansion was the dominant type that caused the isolation effect on KEAs, with an average contribution of over 85% in the periods from 1988–1994, 1994–2000 and 2000–2005. The habitat isolation caused by leapfrog growth was higher in the first period from 1980 and1988, with an average contribution of approximately 35.3%. Spatially, KEAs in the Bao’an and Longgang Districts were more seriously affected by leapfrog growth from 1980 and1988. This was particularly evident in Luotian and Qiankeng Forest Parks, where leapfrog growth accounted for up to 94% of the total isolation effect. The infilling growth contributed less to the isolation effect on KEAs
Fig. 3. Spatiotemporal variations of UII and CUII.
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
51
Table 2 ¯ Area and number of newly grown urban patches for the three urban growth types within a radius of d. Periods
Leapfrog
Edge-expansion 2
1980–1988 1988–1994 1994–2000 2000–2005
6
−1
Area (km )
NPP loss (10 gC a
22.38 17.54 9.95 6.85
21.95 20.39 10.19 7.91
)
2
Infilling growth 6
−1
PSI
Area (km )
NPP loss (10 gC a
1.42 1.42 1.41 1.43
45.79 67.55 59.51 43.42
40.94 61.33 44.69 36.04
)
PSI
Area (km2 )
NPP loss (106 gC a−1 )
PSI
2.00 2.28 2.12 1.93
0.22 10.29 21.54 17.96
0.08 6.95 14.42 10.06
1.28 1.80 1.81 1.69
Table 3 UII for the three urban growth types. Periods
UII for leapfrog growth
1980–1988 1988–1994 1994–2000 2000–2005
UII for edge-expansion
UII for infilling growth
Max
Mean (10−4 )
Sum
Max
Mean
Sum
Max
Mean (10−4 )
Sum
1.2093 0.0081 0.0083 0.0078
18.54 0.75 1.12 0.70
1.5908 0.0902 0.0871 0.0425
10.6572 7.5126 2.8810 0.1347
0.1434 0.0318 0.0115 0.0011
21.8016 21.6761 12.3907 1.4507
0.0001 0.0206 0.0956 0.0096
0.21 3.59 7.79 0.73
0.0002 0.0927 0.5675 0.0700
Fig. 4. Spatiotemporal variations of UII for the three urban growth types.
Fig. 5. Contributions of the three urban growth types to the isolated effect around the 12 KEAs. The twelve bars in each period represent KEA 1 to 12 in Fig. 1 in turn and the abbreviation of each KEA is shown at the bottom of the bar.
52
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
from 1980 to 1994 and subsequently contributed more from 1994 to 2005. The infilling growth had a considerable isolated influence on the Silver Lake Forest Park from 1994 and 2000, with a contribution rate of 48% of the total isolation effect.
4. Discussion Urbanization has been deemed as one of the most critical driving forces of land-use conversion (Long et al., 2007). A large number of papers have shown the changes of composition and configuration of natural habitats as well as their ecological consequences during urbanization processes, such as vegetation habitat fragmentation (Li et al., 2010), vegetation productivity reduction (Yu et al., 2009) and biodiversity loss (McKinney, 2002). However, habitat isolation, a more complex process associated with the interference effects of urban land use on natural habitats, has been less investigated in the literature. With the requirements of habitat conservation and restoration in cities, the interpretation of the ecological mechanisms of isolation effects linking protected areas to surrounding lands is a fundamental step. As portions of the whole ecosystem, the protected and unprotected habitats are indispensable for maintaining organism movements, nutrient flows and population processes within them (Fig. 6a) (Hansen and DeFries, 2007). Due to urban development enforcements, unprotected areas around reserves are prone to conversion into urban-dominant lands. This may directly result in the loss of ecological benefits for rare habitats and species or undermine the connectivity among protected areas via ‘stepping stones’ for species movement or ecological flows (Fig. 6b). Urban development may indirectly affect multiple ecological processes because newly transformed urban patches are likely to disturb the predictable pathways of species dispersal, nutrient cycling and energy flow (Fig. 6c). It is also probable that human activities affect neighboring environments beyond urban boundaries and exert interference on ecological flows among reserves (Fig. 6d). Generally, land-use changes in unprotected habitats may cause the functional isolation of protected areas via multiple ecological mechanisms, which are relevant to the quantity, composition and configuration of land-use conversion. Therefore, there is a need to quantify habitat isolation in a comprehensive way by which the insulation or interference factors of urban land use are integrated. In this paper, we proposed that the urban growth area, morphology and types associated with habitat quality losses due to urban land-use conversion are important factors for assessing the role of urban development. Ecological restoration, as “ecological engineering of the best kind” (Bradshaw, 1997), is the process of assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed (Lewis, 2005). An ecosystem restoration assessment (ERA) should include restoration objectives and changes in the landscape context of land adjacent to protected areas. The measurement of the isolation effect, focusing on the impact of surrounding land-use conversion on protected areas, can provide an alternative for ERA and offer practical references regarding ecosystem restoration. In this study, we attempted to use the UII to illustrate the significant impact of urban sprawl on habitat isolation. As illustrated in the case study of Shenzhen, we found that the isolated degree of KEAs exhibited a notable increase with the expansion of the urban regions and the increase of complexity of urban growth patterns. To be specific, the UII has the ability to characterize the spatial-temporal dynamics of habitat isolation effects. The results revealed that the degree of habitat isolation constantly increased during Shenzhen’s rapid urbanization process. KEAs in the southwest part of Shenzhen were seriously affected by urban land-use changes. These findings provide a spatially explicit identification
about where the habitat isolation occurred or were increasing and which elements or patches created this disturbance. Such detailed information is useful for ecologists and urban planners to adjust and restore the ecosystem structure and function. For example, they can establish buffer zones around the KEAs where the serious isolation effect occurred to protect the KEAs from anthropogenic disturbance. The construction of corridors or vegetation patches acting as connecting elements among the seriously isolated KEAs may be an effective solution to enhance the habitat connectivity. The vegetation conservation, restoration and rehabilitation in KEAs can also contribute to the improvement of habitat quality. In addition, we benefit from a temporal analysis that predicts the potential variation of isolation effects and then scientifically delimits the range of KEAs. If necessary, we can simulate the potential risks of habitat isolation by creating scenarios of future land-use changes. Furthermore, our major improvement of the proposed index is its ability to detect the spatiotemporal dynamics of the isolation effect related to different urban growth types and to measure the contribution of each type to habitat isolation. As illustrated in this study, there are tremendous differences in the isolation effect caused by the three urban growth types. Urban edge expansion results in the most serious amount of habitat isolation, especially in the initial stage of urbanization. In contrast, leapfrog and infilling urban patches make minor contributions to the isolation effect on KEAs. These findings provide details to investigate the ecological effect of urbanization processes. Different urban growth types may have various influences on ecological processes and, in turn, cause functional isolation among KEAs in different ways. In landscape ecology, one of the key research topics is to understand the relationship between spatial patterns and ecological processes (Wu and Hobbs, 2002). Currently, there are two types of metrics related to isolation implications in the literature. One is a set of distance-based isolation metrics that measure the shortest or average distance between habitat patches (Tischendorf et al., 2003). This type of metric, which only has one parameter, can convey the limited ecological connotation to characterize habitat isolation. Another type is represented by area-based isolation metrics, such as the proximity index (Gustafson and Parker, 1994) and similarity index (McGarigal et al., 2002), which only involve isolated habitats themselves in a certain spatial radius without considering the impact of interference patches on natural ones. Indeed, habitat isolation is relevant to the interactions between both urban and natural landscapes. Urban expansion may lead to direct habitat losses and poor matrix permeability around protected areas, which are embodied by the NPP losses in this study. NPP, as the basic provisioning service in the framework of ecosystem services (MEA, 2005), has the ability to represent the survival quality or available resources for heterotrophic organisms. Thus, NPP loss has an important role in the evaluation of habitat isolation effects. The interference effect (or edge effect) caused by human activities may go beyond urban edges and exacerbate the isolated degree around the protected areas. The urban patch shape and edge length adjacent to the non-urban matrix are closely relevant to this edge effect and serve as the other two key factors in quantifying the isolation effect. Our work is the first step to explore the relationship between urban growth patterns and the isolation effect associated with ecological processes. Quantitative methods that integrate patterns with processes are needed in future research. The index and its quantitative results also provide practical guidelines for urban planning and design. A comprehensive optimization can be performed through adjusting the size, shape and position of urban patches with different growth types to mitigate the negative impacts of urban growth and other human activities on KEAs. Integrating three key variables, the proposed index of UII provides
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
53
Fig. 6. Illustration of isolation effects of urban growth on the KEAs (modified from Hansen and DeFries (2007). (a) Materials, energy, or organisms flowing between ecological reserves and the surrounding environment free from the interference of urban areas. (b) Urban growth invades species habitats and disrupts the source-sink dispersal. (c) Urban growth obstructs or impairs ecological flows. (d) Edge effects from urban growth negatively influence the material, energy flow or species exchange between reserves and the surroundings.
a new concept for linking the isolation effect on natural landscapes to the urban growth patterns in an urbanization context. In this study, we evaluated the habitat isolation effect from the point of land-use composition and configuration as well as the context. In fact, functional isolation is species-specific. For example, organisms with stronger dispersal abilities are able to cross more hostile landscapes as the functional isolation is relatively weak. Likewise, edge species are suitable for surviving in landscape transitional zones and are less isolated than interior species. When more information on species behavior becomes available, our results will be more effectively interpreted for ecological conservation and restoration. We defined the mean nearest distance of all urban patches to all habitats (d) as the calculation scope of the isolation effect without considering the extent beyond this radius. Ideally, if the buffer radius is determined by the dispersal distance related to the movement process of target species, then the results would be better able to reveal the isolated effects of urbanization on KEAs and make it possible for designers and planners to integrate UII and CUII with the empirical data of specific species to establish an ecological infrastructure in urbanization areas. 5. Conclusions Urbanization is one of the major threats to biodiversity and ecological processes because of the loss, fragmentation and degradation of natural habitats. Unprotected habitats compared with protected areas are easily converted into human-dominant landscapes by urban developed enforcement, which may disturb or cut off species dispersal and ecological flows among reserves. All of these factors ultimately threaten biodiversity and degrade ecosystem services for human well-being. Therefore, the interference or isolation effects of land-use conversion on protected habitats should receive a great deal of attention, and effective measures should be taken for ecological conservation and restoration. Because of the practical requirements, ecological designers are in critical need of a quantitative tool to detect isolation effects. In this study, we account for the interactions between urban and natural landscapes and develop the UII through integrating urban scale, morphology, location, and quality loss factors in the land-use
conversion process. Our results can identify the discrepancy of habitat isolation caused by three urban growth types and quantify the isolated effect of each type on a KEA in Shenzhen. We also identify the most seriously isolated KEAs in the urbanization process. The index proved effective to capture the spatiotemporal characteristics of habitat isolation related to different urban growth patterns. This assessment and spatially explicit identification can help to optimize urban landscape patterns and determine conservation priorities for isolated habitats. We believe that habitat conservation and restoration based on an isolation effect assessment and spatial diagnosis can shape sustainable ecosystems in urbanized regions. Acknowledgments This research was funded by the Program of National Natural Science Foundation of China (41171404), the Program of Introducing Talents of Discipline to Universities (Grant no. B08008) and the Fundamental Research Fund for the Central Universities. Special thanks are given to the referees and editors for their instructive comments, suggestions and editing for the manuscript. References Alberti, M., 2005. The effects of urban patterns on ecosystem function. Int. Reg. Sci. Rev. 28, 168–192. Antrop, M., 2000. Changing patterns in the urbanized countryside of Western Europe. Landsc. Ecol. 15, 257–270. Bailey, S.A., Horner-Devine, M.C., Luck, G., Moore, L.A., Carney, K.M., Anderson, S., Betrus, C., Fleishman, E., 2004. Primary productivity and species richness: relationships among functional guilds, residency groups and vagility classes at multiple spatial scales. Ecography 27, 207–217. Barot, S., Lata, J.C., Lacroix, G., 2012. Meeting the relational challenge of ecological engineering within ecological sciences. Ecol. Eng. 45, 13–23. Bradshaw, A., 1997. Restoration of mined lands using natural processes. Ecol. Eng. 8, 255–269. Cabeza, M., 2003. Habitat loss and connectivity of reserve networks in probability approaches to reserve design. Ecol. Lett. 6, 665–672. . Crooks, K.R., Sanjayan, M. (Eds.), 2006. Connectivity Conservation, Cambridge University Press, New York. Davidson, C., 1998. Issues in measuring landscape fragmentation. Wildl. Soc. Bull. 26, 32–37. Forman, R.T.T., Galli, A.E., Leck, C.F., 1976. Forest size and avian diversity in NewJersey woodlots with some land-use implications. Oecologia 26, 1–8.
54
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, J.M., 2008. Global change and the ecology of cities. Science 319, 756–760. Grumbine, R.E., 1990. Protecting biological diversity through the greater ecosystem concept. Nat. Areas J. 10, 114–120. Grumbine, R.E., 1994. What is ecosystem management? Conserv. Biol. 8, 27–38. Gustafson, E.J., Parker, G.R., 1994. Using an index of habitat patch proximity for landscape design. Landsc. Urban Plan. 29, 117–130. Hansen, A.J., DeFries, R., 2007. Ecological mechanisms linking protected areas to surrounding lands. Ecol. Appl. 17, 974–988. Harrison, S., Bruna, E., 1999. Habitat fragmentation and large-scale conservation: what do we know for sure? Ecography 22, 225–232. Howard, P.C., Davenport, T., Kigenyi, F.W., Viskanic, P., Baltzer, M.C., Dickinson, C.J., Lwanga, J., Matthews, R.A., Mupada, E., 2000. Protected area planning in the tropics: Uganda’s national system of forest nature reserves. Conserv. Biol. 14, 858–875. Jaeger, J.A.G., 2000. Landscape division, splitting index, and effective mesh size: new measures of landscape fragmentation. Landsc. Ecol. 15, 115–130. Lewis, R.R., 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecol. Eng. 24, 403–418. Li, T., Shilling, F., Thorne, J., Li, F., Schott, H., Boynton, R., Berry, A.M., 2010. Fragmentation of China’s landscape by roads and urban areas. Landsc. Ecol. 25, 839–853. Lindenmayer, D., Hobbs, R.J., Montague-Drake, R., et al., 2008. A checklist for ecological management of landscapes for conservation. Ecol. Lett. 11, 78–91. Long, H.L., Tang, G.P., Li, X.B., Heilig, G.K., 2007. Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China. J. Environ. Manage. 83, 351–364. McGarigal, K., Cushman, S.A., Neel, M.C., Ene, E., 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. University of Massachusetts, Amherst, MA, http://www.umass.edu/landeco/research/fragstats/fragstats.html (Available on: 28.09.12). McKinney, M.L., 2002. Urbanization, biodiversity, and conservation. BioScience 52, 883–890. MEA, 2005. Ecosystems and Human Well-being: Synthesis. World Resources Institute, 488 Washington, DC. Mitsch, W.J., 2012. What is ecological engineering. Ecol. Eng. 45, 5–12. Moser, B., Jaeger, J.A.G., Tappeiner, U., Tasser, E., Eiselt, B., 2007. Modification of the effective mesh size for measuring landscape fragmentation to solve the boundary problem. Landsc. Ecol. 22, 447–459. Ng, C.N., Xie, Y.J., Yu, X.J., 2011. Measuring the spatio-temporal variation of habitat isolation due to rapid urbanization: a case study of the Shenzhen River crossboundary catchment, China. Landsc. Urban Plan. 103, 44–54. Pascual-Hortal, L., Saura, S., 2006. Comparison and development of new graph-based landscape connectivity indices: towards the priorization of habitat patches and corridors for conservation. Landsc. Ecol. 21, 959–967.
Saura, S., Pascual-Hortal, L.A., 2007. A new habitat availability index to integrate connectivity in landscape conservation planning: Comparison with existing indices and application to a case study. Landsc. Urban Plan. 83, 91–103. SSB (Shenzhen statistics Bureau), 2011. Shenzhen Statistical Yearbook 2010. China Statistics Press, Beijing, China (in Chinese). Su, W., Gu, C., Yang, G., Chen, S., Zhen, F., 2010. Measuring the impact of urban sprawl on natural landscape pattern of the Western Taihu Lake watershed, China. Landsc. Urban Plan. 95, 61–67. SWCA (Shenzhen Wildlife Conservation Administration), 2009. Situation of wildlife conservation in Shenzhen, http://www.szgreen.gov.cn/news. aspx?newsid=5041.html (Available on: 20.08.12). Taylor, P.D., Fahrig, L., Henein, K., Merriam, G., 1993. Connectivity is a vital element of landscape structure. Oikos 68, 571–573. Tischendorf, L., Bender, D.J., Fahrig, L., 2003. Evaluation of patch isolation metrics in mosaic landscapes for specialist vs. generalist dispersers. Landsc. Ecol. 18, 41–50. UNPFA, 2008. State of world population 2007: unleashing the potential of urban growth, United Nation, http://www.unfpa.org/public/home/ publications/pid/408.html (Available on: 28.09.12). Urban, D., Keitt, T., 2001. Landscape connectivity: a graph-theoretic perspective. Ecology 82, 1205–1218. Urban, D.L., Minor, E.S., Treml, E.A., Schick, R.S., 2009. Graph models of habitat mosaics. Ecol. Lett. 12, 260–273. Wu, J.G., Hobbs, R., 2002. Key issues and research priorities in landscape ecology: an idiosyncratic synthesis. Landsc. Ecol. 17, 355–365. Xing, F.W., Yu, M.E., 2000. Wild Plants of Shenzhen China. China Forestry Publishing House, Beijing (in Chinese). Xu, C., Liu, M., Zhang, C., An, S., Yu, W., Chen, J., 2007. The spatiotemporal dynamics of rapid urban growth in the Nanjing metropolitan region of China. Landsc. Ecol. 22, 925–937. Yu, D.Y., Shao, H.B., Shi, P.J., Zhu, W.Q., Pan, Y.Z., 2009. How does the conversion of land cover to urban use affect net primary productivity? A case study in Shenzhen city, China. Agric. Forest Meteorol. 149, 2054–2060. Yu, D.Y., Shi, P.J., Han, G.Y., Zhu, W.Q., Du, S.Q., Xun, B., 2011. Forest ecosystem restoration due to a national conservation plan in China. Ecol. Eng. 37, 1387–1397. Yu, D.Y., Xun, B., Shi, P.J., Shao, H.B., Liu, Y.P., 2012. Ecological restoration planning based on connectivity in an urban area. Ecol. Eng. 46, 24–33. Yu, D.Y., Xun, B., Liu, Y.P., 2013. Measuring landscape connectivity in an urban area for biological conservation. CLEAN-Soil Air Water, http://dx.doi.org/10.1002/clen.201200448 (in press). Zetterberg, A., Mörtberg, U.M., Balfors, B., 2010. Making graph theory operational for landscape ecological assessments, planning, and design. Landsc. Urban Plan. 95, 181–191.
Contents lists available at ScienceDirect
Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng
Quantifying isolation effect of urban growth on key ecological areas Bin Xun a,b,c , Deyong Yu a,c,∗ , Yupeng Liu a,c , Ruifang Hao a,c , Yun Sun a,c a
State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China Department of Environmental Science, Northwest University, Xi’an 710069, China c Human-Environment System Sustainability (CHESS), Beijing Normal University, Beijing 100875, China b
a r t i c l e
i n f o
Article history: Received 22 April 2013 Received in revised form 17 January 2014 Accepted 29 March 2014 Keywords: Urbanization Ecological restoration Fragmentation Urban isolation index Ecological processes Landscape ecology
a b s t r a c t Urbanization may threaten multiple ecosystem processes as well as biodiversity in ecological reserves due to the loss, fragmentation and degradation of habitats in unprotected areas, making it necessary to measure the interference effect of urban growth on the protected areas. This study develops an urban isolation index (UII) to quantify the isolation effects of land-use conversion on key ecological areas (KEAs) and accounts for urban patch size, morphology, location, and quality loss. Based on the five remote sensing Landsat images taken between 1980 and 2005, UII is applied to analyze the spatio-temporal dynamics of the isolation effects caused by rapid urbanization in Shenzhen, China. The results showed that the habitat isolation constantly intensified during the urbanization process. We identified spatially explicit information of critical urban domains that exerted an interference effect on KEAs. There are considerable differences of isolation effects resulting from different urban growth types. Urban edge expansion caused the greatest habitat isolation in Shenzhen. The outlying and infilling urban patches contributed minor isolation effects to the KEAs. This paper furthers the understanding of the ecological effects of urbanization processes and provides a spatially explicit identification for ecological conservation and restoration in rapidly urbanized regions. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The urban population of the world grew almost exponentially during the late 18th century. Today, half of the world’s population resides in urban areas. It is predicted that the global urban population will increase to 69% by 2050 (UNPFA, 2008). Urbanization, a complex process that transforms landscapes formed by rural life styles into urban landscapes (Antrop, 2000), is becoming increasingly universal and irresistible throughout the world. Urbanization dramatically shapes new landscape patterns and significantly influences local and global ecosystem functions and dynamics. In most urban areas, urban spatial diffusion can fragment, isolate, and degrade natural habitats (Alberti, 2005). These negative effects on
Abbreviations: UII, urban isolation index; CUII, cumulative urban isolation index; NPP, net primary productivity; PSI, patch shape index; KEA, key ecological areas; SEZ, special economic zone; ERA, ecosystem restoration assessment. ∗ Corresponding author at: State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China. Tel.: +86 13681176750. E-mail address: [email protected] (D. Yu). http://dx.doi.org/10.1016/j.ecoleng.2014.03.041 0925-8574/© 2014 Elsevier B.V. All rights reserved.
habitats not only involve the direct extinction and reduction of native wildlife but also alter multiple ecological processes, such as animal movement, seed dispersal, genetic flow and nutrient cycling (Harrison and Bruna, 1999; Alberti, 2005; Crooks and Sanjayan, 2006; Grimm et al., 2008), and, ultimately, influence ecosystem services and human well-being (Alberti, 2005). Furthermore, urban development is facing the dilemma of urban land intransigence or natural habitat degradation. The only way to face this challenge is to develop sustainable practices, i.e., economically, socially and ecologically acceptable solutions, to mitigate the conflicts between urban development and ecological conservation in urbanized areas (Barot et al., 2012). Ecological engineering is defined as the design of sustainable ecosystems that integrate human society with the natural environment for the benefit of both (Mitsch, 2012). Its goals involve the creation or restoration of ecosystems that can provide sustainable services for humans and other forms of life (Mitsch, 2012). Since the 20th century, establishing protected areas has become the fundamental strategy of regional ecological conservation and restoration (Howard et al., 2000). Ecological reserves may be the most direct and effective measure to maintain species persistence and biodiversity, especially in urban areas. During
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
urbanization processes, however, ecological reserves are likely to cause ecological isolation among protected areas if the surrounding unprotected vegetation habitats undergo anthropogenic disturbances, and can even lead to a conversion to human-dominated landscapes (Grumbine, 1994). In other words, from the viewpoint of ecosystem integrity, whether reserves can provide sustained and effective protection for organisms depends not only on the protected areas themselves but also on the natural context outside of the protected boundaries (Hansen and DeFries, 2007). This is largely because some of the key areas excluded from the protected domains are important to protect wildlife from exposure to a hostile landscape matrix and to maintain nutrient flows, organism movements and population processes (Grumbine, 1990). It is therefore essential and implicative for regional-scale ecological conservation and restoration, especially in a large-range urbanization context, to evaluate the isolation effect of urban growth on the ecological reserves in a quantitative and comprehensive manner. The quantification method of landscape fragmentation and connectivity has always been a hot topic for ecologists (Lindenmayer et al., 2008). Early proposed landscape metrics, such as the patch density, mean patch size, perimeter-area ratio and fractal index, can only capture the separate aspects of fragmentation (Davidson, 1998). Recently, some new metrics, such as the landscape division, splitting index and effective mesh size, have been developed to characterize landscape fragmentation from a geometric perspective (Jaeger, 2000; Moser et al., 2007). Landscape connectivity is defined as the degree to which the landscape facilitates or impedes the dispersal movement of species among habitat patches that exist in the landscape (Taylor et al., 1993). Graph theory is thought to be a promising methodology in the assessment of landscape connectivity (Urban et al., 2009). Some scholars have proposed a series of graph-based metrics to quantify the network connectivity related to species dispersal processes (e.g., Urban and Keitt, 2001; PascualHortal and Saura, 2006; Saura and Pascual-Hortal, 2007; Yu et al., 2013). In general, the metrics used to assess landscape fragmentation and connectivity are well developed; some of them have even been successfully applied to biological conservation and ecological restoration (e.g., Cabeza, 2003; Zetterberg et al., 2010; Yu et al., 2012). However, most of the metrics concentrate on the fragmentation features of a single landscape type and thus cannot reflect the interaction among landscape types during land-use conversion processes. Habitat fragmentation may cause functional isolation due to the weak exchange of genes for both individuals and populations, which explains why fragmented habitats often contain fewer species than contiguous habitats (Forman et al., 1976). Habitat isolation related to interference effects is a more complex process than fragmentation in the context of urbanization. Recently, two metrics, the insulation degree (ID) (Su et al., 2010) and urbanization isolation effect (UIE) (Ng et al., 2011), were proposed to examine the isolation effect caused by urban areas. These metrics were described by the distance of urban patches to the natural patch and the area of urban patches within a specified radius. In the current study, we developed a new urban isolation index (UII) by comprehensively considering the interaction effects of urban size, morphology and quality losses on key ecological areas (KEAs) during land-use conversion processes. We evaluated the spatiotemporal dynamics of the isolation effect of urban growth on KEAs in Shenzhen, an area of China that has been rapidly urbanized. Our objective is to quantitatively answer two questions: (1) What degree of isolation do KEAs undergo during the urbanization process? (2) Which type of urban growth causes the most serious habitat isolation?
47
2. Material and methods 2.1. Study area Shenzhen is a coastal city in southern China, neighboring Dongguan and Huizhou in the north and Hong Kong in the south, and is flanked by the Daya Gulf in the east and the Pearl River Estuary in the west. The city consists of a special economic zone (SEZ), Bao’an District and Longgang District (Fig. 1). It lies between 22◦ 26 N and 22◦ 51 N latitude and between 113◦ 45 E and 114◦ 37 E longitude, and its total area is approximately 2020 km2 . With a subtropical marine climate, the annual average temperature, precipitation and sunshine hours are 22.4 ◦ C, 1933.3 mm and 2020 h, respectively. The rainy season is mainly concentrated from April to September. The dominant vegetation is an evergreen broad-leaved mixed forest. There are 1889 types of wild vascular plants, among which 22 belong to the nationally rare and endangered plants list (Xing and Yu, 2000). There are five types of nationally protected Class I wild animals and 43 nationally protected Class II wild animals (SWCA, 2009). Shenzhen was municipalized in 1979 and established as the first SEZ in China in 1980. Over the past 30 years, it has evolved from a border fishing village into a metropolitan region with a population of 10.35 million and a GDP of 951.10 billion Yuan in 2010 (SSB, 2011). Meanwhile, Shenzhen has experienced significant land-use changes. In 1980, urban lands and forests occupied 0.63% and 38.67% of the whole city area, respectively. By 2005, urban land had increased to 33.52% and forests had decreased to 29.81% of the total area (see the remote sensing Landsat data in Section 2.2). Because of the habitat fragmentation caused by this urban growth, biodiversity has only remained in isolated mountain areas. In 2005, the Shenzhen government promulgated a legally binding regulation stating that KEAs need to provide a survival environment for wild species, maintain ecosystem processes and regional biodiversity, and promote environmental quality for human well-being. KEAs contain (Fig. 1): 1) water source protection areas, scenic spots, nature reserves, and primary farmland protection areas, forests and country parks; 2) mountains and woodlands with a slope over 25◦ and uplands with an elevation over 50 m; 3) main trunk rivers, reservoirs and wetlands; 4) green space for the maintenance of ecosystem integrity; and 5) peninsulas and other ecologically important coastal areas. However, the impacts of urban growth on these ecological areas have not undergone scientific assessment until now.
2.2. Data processing We used five cloud-free Landsat images (including a multispectral scanner (MSS) satellite image with a resolution of 79 m × 79 m from 1980; three thematic mapper (TM) satellite images with a resolution of 30 m × 30 m from 1988, 1994 and 2005; and an enhanced thematic mapper plus (ETM+) satellite image with a resolution of 30 m × 30 m from 2000) to produce the land-use thematic maps. Other auxiliary data included topographic maps at a 1:50,000 scale from the 1960s and 1970s, an administrative boundary map at a 1:100,000 scale from 2005, and aerial photographs with high resolution and field survey data to carry out the geometric correction and land-use classification. We obtained 389 evenly distributed ground control points (GCPs) with exact locations and land cover information to perform the geometrical correction and classification. Different land-use types were categorized using both unsupervised classification and supervised classification algorithms. Based on the normalized
48
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Fig. 1. Location of the study area and spatial distribution of the KEAs (1: Luotian KEA, 2: Qiankeng KEA, 3: Maluan Mountain KEA, 4: Xichong-Qiniang Mountain KEA, 5: Baguang KEA, 6: Meilin-Tanglang Mountain KEA, 7: Tiegang-Yangtai Mountain KEA, 8: Guangming KEA, 9: Huangzhukeng KEA, 10: Fenghuang Mountain KEA, 11: Wutong Mountain KEA, 12: Silver Lake KEA).
difference vegetation index (NDVI) calculated from Landsat Bands 3 and 4, ground objects were categorized into three groups by the unsupervised classification method: non-vegetation, vegetation and water. This process maximally reduced artificial errors and performed the appropriate classification for further interpretation. Each group was further classified by using a maximum likelihood supervised classification to obtain the land-use thematic map. Six land-use types were classified as forest, orchard, urban land, farmland, water and vacant land (Fig. 2). The classification results were considered reasonable, with an overall accuracy of 83% in 1980, 84% in 1988, 86% in 1994, 87% in 2000 and 89% in 2005. 2.3. Urban growth types Generally, there are three urban growth types — edge expansion, infilling and leapfrog. Edge expansion refers to an urban sprawl in a contiguous manner that proceeds outward from the existing urban area. Infilling refers to urban expansion in the gap-infilling type among historical urban centers. Leapfrog means that new urban patches are formed and have no common boundary with the existing urban area. This study identified the three urban growth types through a quantitative method (Xu et al., 2007) to detect their individual isolation effect on ecological reserves. The equation is given by: S=
Lc P
(1)
where Lc is the length of the common boundary between the newly grown and pre-growth urban patch and P is the perimeter of this newly grown patch. Urban growth type is identified as infilling when 0.5 < S ≤ 1, edge expansion when 0 < S ≤ 0.5, and leapfrog when S = 0, which indicates no common boundary. 2.4. Landscape isolation index The new metric of UII was developed to evaluate the isolation effect of the newly grown urban patches on KEAs. Here, UII is calculated by the parameters of the quality loss, shape, growth types and location of new urban patches within a specified radius of KEAs to quantify the comprehensive effect of land-use conversion on KEAs. UII is defined as follows: UIIi =
Qi L L − Lci × i × i × Qt Lt 2 Ai
d −1 i
d
(2)
where Qi is the quality loss caused by a newly transformed urban patch i. We use net primary productivity (NPP) loss as a proxy of the quality reduction, which equals the product of the area of a newly transformed urban patch and the NPP of previous land-use type per unit. Qt is the total NPP loss resulting from urban growth. Primary production can provide a reliable basis for stratifying surveys of biodiversity and environmental quality by highlighting the areas of potentially high biodiversity across large areas (Bailey et al., 2004). In this study, NPP is calculated by our previously improved
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
49
Fig. 2. Land-use maps in Shenzhen City from 1980 to 2005.
CASA-NPP model (Yu et al., 2009, 2011). di is the distance from the urban patches i to the nearest KEA j . d is the mean nearest distance of all urban patches to all habitats in the study area. The distance is measured by the edge-to-edge distance. Ai is the area of the urban patch i. Li is the perimeter of the urban patch i. Lci is the length of common boundary between the newly grown and the previous urban patches. Lt is the total perimeter of the newly grown urban patches. Li /2 Ai is the patch shape index (PSI) (McGarigal et al., 2002) and represents the shape of the urban patch i in this study. When the natural landscape is transformed into urban land with significant NPP loss, the suitable habitats or available resources for species will be reduced, and the isolation degree among habitats will be strengthened. The interference or isolation effect of newly grown urban patches is also related to the patch shape and edge adjacent to the non-urban matrix. The more complex and longer the common boundaries between the urban and non-urban areas, the stronger the isolation effect of urban expansion becomes. When the urban growth belongs to the leapfrog type (S = 0 in Eq. (1)), meaning that the patch edge is completely adjacent to other landuse types, UII is relatively higher. New urban patches in full infilling growth (S = 1 in Eq. (1)) have the smallest UII (UII = 0). The distance between the urban patch and the KEA is negatively related to the isolation effect, namely, the closer the distance, the higher the UII value.
The cumulative urban isolation index (CUII) represents the total isolation effect of ecological reserves undergoing urban growth, which is defined as: CUIIj =
n
UIIi
(3)
i=1
where CUIIj is the sum of the isolation effect of newly grown urban patches on KEAj within a certain radius. UII and CUII with no dimension facilitate the comparison of the results in different periods and regions. In this study, we defined the mean nearest distance of all urban patches to all habitats (d) as the buffer radius. 3. Results 3.1. Spatiotemporal dynamics of the urban isolation effect The radius d sharply decreased from 1988 to 2000, indicating that urban development constantly extended toward KEAs (Table 1). The urban area increased by 251.1%, from 68.18 km2 in 1988 to 239.37 km2 in 2005 within the radius d around KEAs. There was a constant loss in the NPP resulting from the newly converted urban land. The NPP was 0.63 × 108 gC a−1 in 1988 and reached 2.52 × 108 gC a−1 in 2005. The rising trend in PSI indicated that the
50
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Table 1 Main features of new growth urban patches within a radius of d¯ and the cumulative impacts of UII and CUII on the baseline of land use in 1980. Years
d¯ (m)
Urban growth area (km2 )
NPP loss (108 gC a−1 )
PSI
UII Max
UII Mean
UII Sum
CUII Max
CUII Mean
1988 1994 2000 2005
2386 2182 2042 2041
68.18 149.35 213.66 239.37
0.63 1.54 2.27 2.52
1.46 1.55 1.58 1.67
10.66 85.82 118.84 368.27
0.02 0.31 0.85 3.23
23.56 505.51 1326.45 4534.72
11.99 85.98 119.66 368.66
1.58 17.52 41.80 127.31
Note: The earliest available time of the MODIS product during one year was in 2000, thus we used the NPP in this year to replace the NPPs from 1980, 1988 and 1994 in places where the land-use types did not change. When the land covers were altered, the NPPs from 1980, 1988 and 1994 were obtained by calculating the average NPP of each land-use type in 2000.
shape of the new urban patches became increasingly irregular and complex. The cumulative isolation impacts represented by UII and CUII over different times are listed in Table 1. The maximum, mean and sum of UII in 2005 are 33, 161 and 197 times larger than those in 1988, respectively. The greatest increments for the mean and sum UII occurred between 1988 and 1994, with a value of 1450% and 2046%, respectively. Similarly, the mean and sum of CUII increased by 30 and 80 times from 1988 to 2005, respectively. The largest increment of the mean CUII was 1009%, which occurred from 1988 to 1994. The spatiotemporal dynamics of the isolation effect are shown in Fig. 3. In 1988, the urban patches with a stronger isolation effect were mainly concentrated in the middle part of the SEZ. Silver Lake Forest Park (KEA12) was the most seriously affected KEA, with the CUII of 11.99%. In 1994, the areas with the higher isolation effect presented a remarkable spatial expansion from the SEZ toward other areas. Meilin-Tanglang Mountain Forest Park (KEA 6) underwent the strongest amount of isolation, and Silver Lake Forest Park ranked second in this metric. The interference caused by the urban sprawl enlarged further throughout the entire city after 2000. In contrast to the Longgang District, the SEZ and Bao’an had a higher UII and CUII. In 2005, except for KEA 4 and 5, the other KEAs were strongly isolated. 3.2. Isolation effect caused by the three urban growth types The three urban growth types were identified and compared for their isolation effect within the radius d around KEAs during the four periods from 1980–1988, 1988–1994, 1994–2000 and 2000–2005 (Tables 2 and 3). Urban growth was dominated by edge
expansion, occupying 60–70% of the total new urban areas in the four periods from 1980 and 2005. NPP loss was mainly caused by edge expansion, amounting to 64–70% of the total loss in the four periods from 1980 and 2005. Compared with the other three periods, leapfrog growth resulted in a relatively higher NPP loss of 34% from 1980 and 1988. Urban edge expansion ranked the highest in shape complexity during the four periods, followed by infilling and then leapfrog growth. Edge expansion had the greatest isolation influence on KEAs, accounting for over 90% of the total UII from 1980 and 2005 (Table 3 and Fig. 4). According to the maximum, mean and sum of UII for the newly developed urban patches, the stronger isolation effect caused by urban edge expansion mainly occurred in the SEZ during the periods from 1980–1988 and 1988–1994. The isolation effect caused by the infilling and leapfrog growth was less. Compared with the other three periods, the isolation caused by leapfrog urban growth was relatively larger from 1980 to 1988, which mainly occurred in the middle and west of Shenzhen. The contribution ratios of the three urban growth types to habitat isolation are illustrated in Fig. 5. Edge expansion was the dominant type that caused the isolation effect on KEAs, with an average contribution of over 85% in the periods from 1988–1994, 1994–2000 and 2000–2005. The habitat isolation caused by leapfrog growth was higher in the first period from 1980 and1988, with an average contribution of approximately 35.3%. Spatially, KEAs in the Bao’an and Longgang Districts were more seriously affected by leapfrog growth from 1980 and1988. This was particularly evident in Luotian and Qiankeng Forest Parks, where leapfrog growth accounted for up to 94% of the total isolation effect. The infilling growth contributed less to the isolation effect on KEAs
Fig. 3. Spatiotemporal variations of UII and CUII.
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
51
Table 2 ¯ Area and number of newly grown urban patches for the three urban growth types within a radius of d. Periods
Leapfrog
Edge-expansion 2
1980–1988 1988–1994 1994–2000 2000–2005
6
−1
Area (km )
NPP loss (10 gC a
22.38 17.54 9.95 6.85
21.95 20.39 10.19 7.91
)
2
Infilling growth 6
−1
PSI
Area (km )
NPP loss (10 gC a
1.42 1.42 1.41 1.43
45.79 67.55 59.51 43.42
40.94 61.33 44.69 36.04
)
PSI
Area (km2 )
NPP loss (106 gC a−1 )
PSI
2.00 2.28 2.12 1.93
0.22 10.29 21.54 17.96
0.08 6.95 14.42 10.06
1.28 1.80 1.81 1.69
Table 3 UII for the three urban growth types. Periods
UII for leapfrog growth
1980–1988 1988–1994 1994–2000 2000–2005
UII for edge-expansion
UII for infilling growth
Max
Mean (10−4 )
Sum
Max
Mean
Sum
Max
Mean (10−4 )
Sum
1.2093 0.0081 0.0083 0.0078
18.54 0.75 1.12 0.70
1.5908 0.0902 0.0871 0.0425
10.6572 7.5126 2.8810 0.1347
0.1434 0.0318 0.0115 0.0011
21.8016 21.6761 12.3907 1.4507
0.0001 0.0206 0.0956 0.0096
0.21 3.59 7.79 0.73
0.0002 0.0927 0.5675 0.0700
Fig. 4. Spatiotemporal variations of UII for the three urban growth types.
Fig. 5. Contributions of the three urban growth types to the isolated effect around the 12 KEAs. The twelve bars in each period represent KEA 1 to 12 in Fig. 1 in turn and the abbreviation of each KEA is shown at the bottom of the bar.
52
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
from 1980 to 1994 and subsequently contributed more from 1994 to 2005. The infilling growth had a considerable isolated influence on the Silver Lake Forest Park from 1994 and 2000, with a contribution rate of 48% of the total isolation effect.
4. Discussion Urbanization has been deemed as one of the most critical driving forces of land-use conversion (Long et al., 2007). A large number of papers have shown the changes of composition and configuration of natural habitats as well as their ecological consequences during urbanization processes, such as vegetation habitat fragmentation (Li et al., 2010), vegetation productivity reduction (Yu et al., 2009) and biodiversity loss (McKinney, 2002). However, habitat isolation, a more complex process associated with the interference effects of urban land use on natural habitats, has been less investigated in the literature. With the requirements of habitat conservation and restoration in cities, the interpretation of the ecological mechanisms of isolation effects linking protected areas to surrounding lands is a fundamental step. As portions of the whole ecosystem, the protected and unprotected habitats are indispensable for maintaining organism movements, nutrient flows and population processes within them (Fig. 6a) (Hansen and DeFries, 2007). Due to urban development enforcements, unprotected areas around reserves are prone to conversion into urban-dominant lands. This may directly result in the loss of ecological benefits for rare habitats and species or undermine the connectivity among protected areas via ‘stepping stones’ for species movement or ecological flows (Fig. 6b). Urban development may indirectly affect multiple ecological processes because newly transformed urban patches are likely to disturb the predictable pathways of species dispersal, nutrient cycling and energy flow (Fig. 6c). It is also probable that human activities affect neighboring environments beyond urban boundaries and exert interference on ecological flows among reserves (Fig. 6d). Generally, land-use changes in unprotected habitats may cause the functional isolation of protected areas via multiple ecological mechanisms, which are relevant to the quantity, composition and configuration of land-use conversion. Therefore, there is a need to quantify habitat isolation in a comprehensive way by which the insulation or interference factors of urban land use are integrated. In this paper, we proposed that the urban growth area, morphology and types associated with habitat quality losses due to urban land-use conversion are important factors for assessing the role of urban development. Ecological restoration, as “ecological engineering of the best kind” (Bradshaw, 1997), is the process of assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed (Lewis, 2005). An ecosystem restoration assessment (ERA) should include restoration objectives and changes in the landscape context of land adjacent to protected areas. The measurement of the isolation effect, focusing on the impact of surrounding land-use conversion on protected areas, can provide an alternative for ERA and offer practical references regarding ecosystem restoration. In this study, we attempted to use the UII to illustrate the significant impact of urban sprawl on habitat isolation. As illustrated in the case study of Shenzhen, we found that the isolated degree of KEAs exhibited a notable increase with the expansion of the urban regions and the increase of complexity of urban growth patterns. To be specific, the UII has the ability to characterize the spatial-temporal dynamics of habitat isolation effects. The results revealed that the degree of habitat isolation constantly increased during Shenzhen’s rapid urbanization process. KEAs in the southwest part of Shenzhen were seriously affected by urban land-use changes. These findings provide a spatially explicit identification
about where the habitat isolation occurred or were increasing and which elements or patches created this disturbance. Such detailed information is useful for ecologists and urban planners to adjust and restore the ecosystem structure and function. For example, they can establish buffer zones around the KEAs where the serious isolation effect occurred to protect the KEAs from anthropogenic disturbance. The construction of corridors or vegetation patches acting as connecting elements among the seriously isolated KEAs may be an effective solution to enhance the habitat connectivity. The vegetation conservation, restoration and rehabilitation in KEAs can also contribute to the improvement of habitat quality. In addition, we benefit from a temporal analysis that predicts the potential variation of isolation effects and then scientifically delimits the range of KEAs. If necessary, we can simulate the potential risks of habitat isolation by creating scenarios of future land-use changes. Furthermore, our major improvement of the proposed index is its ability to detect the spatiotemporal dynamics of the isolation effect related to different urban growth types and to measure the contribution of each type to habitat isolation. As illustrated in this study, there are tremendous differences in the isolation effect caused by the three urban growth types. Urban edge expansion results in the most serious amount of habitat isolation, especially in the initial stage of urbanization. In contrast, leapfrog and infilling urban patches make minor contributions to the isolation effect on KEAs. These findings provide details to investigate the ecological effect of urbanization processes. Different urban growth types may have various influences on ecological processes and, in turn, cause functional isolation among KEAs in different ways. In landscape ecology, one of the key research topics is to understand the relationship between spatial patterns and ecological processes (Wu and Hobbs, 2002). Currently, there are two types of metrics related to isolation implications in the literature. One is a set of distance-based isolation metrics that measure the shortest or average distance between habitat patches (Tischendorf et al., 2003). This type of metric, which only has one parameter, can convey the limited ecological connotation to characterize habitat isolation. Another type is represented by area-based isolation metrics, such as the proximity index (Gustafson and Parker, 1994) and similarity index (McGarigal et al., 2002), which only involve isolated habitats themselves in a certain spatial radius without considering the impact of interference patches on natural ones. Indeed, habitat isolation is relevant to the interactions between both urban and natural landscapes. Urban expansion may lead to direct habitat losses and poor matrix permeability around protected areas, which are embodied by the NPP losses in this study. NPP, as the basic provisioning service in the framework of ecosystem services (MEA, 2005), has the ability to represent the survival quality or available resources for heterotrophic organisms. Thus, NPP loss has an important role in the evaluation of habitat isolation effects. The interference effect (or edge effect) caused by human activities may go beyond urban edges and exacerbate the isolated degree around the protected areas. The urban patch shape and edge length adjacent to the non-urban matrix are closely relevant to this edge effect and serve as the other two key factors in quantifying the isolation effect. Our work is the first step to explore the relationship between urban growth patterns and the isolation effect associated with ecological processes. Quantitative methods that integrate patterns with processes are needed in future research. The index and its quantitative results also provide practical guidelines for urban planning and design. A comprehensive optimization can be performed through adjusting the size, shape and position of urban patches with different growth types to mitigate the negative impacts of urban growth and other human activities on KEAs. Integrating three key variables, the proposed index of UII provides
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
53
Fig. 6. Illustration of isolation effects of urban growth on the KEAs (modified from Hansen and DeFries (2007). (a) Materials, energy, or organisms flowing between ecological reserves and the surrounding environment free from the interference of urban areas. (b) Urban growth invades species habitats and disrupts the source-sink dispersal. (c) Urban growth obstructs or impairs ecological flows. (d) Edge effects from urban growth negatively influence the material, energy flow or species exchange between reserves and the surroundings.
a new concept for linking the isolation effect on natural landscapes to the urban growth patterns in an urbanization context. In this study, we evaluated the habitat isolation effect from the point of land-use composition and configuration as well as the context. In fact, functional isolation is species-specific. For example, organisms with stronger dispersal abilities are able to cross more hostile landscapes as the functional isolation is relatively weak. Likewise, edge species are suitable for surviving in landscape transitional zones and are less isolated than interior species. When more information on species behavior becomes available, our results will be more effectively interpreted for ecological conservation and restoration. We defined the mean nearest distance of all urban patches to all habitats (d) as the calculation scope of the isolation effect without considering the extent beyond this radius. Ideally, if the buffer radius is determined by the dispersal distance related to the movement process of target species, then the results would be better able to reveal the isolated effects of urbanization on KEAs and make it possible for designers and planners to integrate UII and CUII with the empirical data of specific species to establish an ecological infrastructure in urbanization areas. 5. Conclusions Urbanization is one of the major threats to biodiversity and ecological processes because of the loss, fragmentation and degradation of natural habitats. Unprotected habitats compared with protected areas are easily converted into human-dominant landscapes by urban developed enforcement, which may disturb or cut off species dispersal and ecological flows among reserves. All of these factors ultimately threaten biodiversity and degrade ecosystem services for human well-being. Therefore, the interference or isolation effects of land-use conversion on protected habitats should receive a great deal of attention, and effective measures should be taken for ecological conservation and restoration. Because of the practical requirements, ecological designers are in critical need of a quantitative tool to detect isolation effects. In this study, we account for the interactions between urban and natural landscapes and develop the UII through integrating urban scale, morphology, location, and quality loss factors in the land-use
conversion process. Our results can identify the discrepancy of habitat isolation caused by three urban growth types and quantify the isolated effect of each type on a KEA in Shenzhen. We also identify the most seriously isolated KEAs in the urbanization process. The index proved effective to capture the spatiotemporal characteristics of habitat isolation related to different urban growth patterns. This assessment and spatially explicit identification can help to optimize urban landscape patterns and determine conservation priorities for isolated habitats. We believe that habitat conservation and restoration based on an isolation effect assessment and spatial diagnosis can shape sustainable ecosystems in urbanized regions. Acknowledgments This research was funded by the Program of National Natural Science Foundation of China (41171404), the Program of Introducing Talents of Discipline to Universities (Grant no. B08008) and the Fundamental Research Fund for the Central Universities. Special thanks are given to the referees and editors for their instructive comments, suggestions and editing for the manuscript. References Alberti, M., 2005. The effects of urban patterns on ecosystem function. Int. Reg. Sci. Rev. 28, 168–192. Antrop, M., 2000. Changing patterns in the urbanized countryside of Western Europe. Landsc. Ecol. 15, 257–270. Bailey, S.A., Horner-Devine, M.C., Luck, G., Moore, L.A., Carney, K.M., Anderson, S., Betrus, C., Fleishman, E., 2004. Primary productivity and species richness: relationships among functional guilds, residency groups and vagility classes at multiple spatial scales. Ecography 27, 207–217. Barot, S., Lata, J.C., Lacroix, G., 2012. Meeting the relational challenge of ecological engineering within ecological sciences. Ecol. Eng. 45, 13–23. Bradshaw, A., 1997. Restoration of mined lands using natural processes. Ecol. Eng. 8, 255–269. Cabeza, M., 2003. Habitat loss and connectivity of reserve networks in probability approaches to reserve design. Ecol. Lett. 6, 665–672. . Crooks, K.R., Sanjayan, M. (Eds.), 2006. Connectivity Conservation, Cambridge University Press, New York. Davidson, C., 1998. Issues in measuring landscape fragmentation. Wildl. Soc. Bull. 26, 32–37. Forman, R.T.T., Galli, A.E., Leck, C.F., 1976. Forest size and avian diversity in NewJersey woodlots with some land-use implications. Oecologia 26, 1–8.
54
B. Xun et al. / Ecological Engineering 69 (2014) 46–54
Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, J.M., 2008. Global change and the ecology of cities. Science 319, 756–760. Grumbine, R.E., 1990. Protecting biological diversity through the greater ecosystem concept. Nat. Areas J. 10, 114–120. Grumbine, R.E., 1994. What is ecosystem management? Conserv. Biol. 8, 27–38. Gustafson, E.J., Parker, G.R., 1994. Using an index of habitat patch proximity for landscape design. Landsc. Urban Plan. 29, 117–130. Hansen, A.J., DeFries, R., 2007. Ecological mechanisms linking protected areas to surrounding lands. Ecol. Appl. 17, 974–988. Harrison, S., Bruna, E., 1999. Habitat fragmentation and large-scale conservation: what do we know for sure? Ecography 22, 225–232. Howard, P.C., Davenport, T., Kigenyi, F.W., Viskanic, P., Baltzer, M.C., Dickinson, C.J., Lwanga, J., Matthews, R.A., Mupada, E., 2000. Protected area planning in the tropics: Uganda’s national system of forest nature reserves. Conserv. Biol. 14, 858–875. Jaeger, J.A.G., 2000. Landscape division, splitting index, and effective mesh size: new measures of landscape fragmentation. Landsc. Ecol. 15, 115–130. Lewis, R.R., 2005. Ecological engineering for successful management and restoration of mangrove forests. Ecol. Eng. 24, 403–418. Li, T., Shilling, F., Thorne, J., Li, F., Schott, H., Boynton, R., Berry, A.M., 2010. Fragmentation of China’s landscape by roads and urban areas. Landsc. Ecol. 25, 839–853. Lindenmayer, D., Hobbs, R.J., Montague-Drake, R., et al., 2008. A checklist for ecological management of landscapes for conservation. Ecol. Lett. 11, 78–91. Long, H.L., Tang, G.P., Li, X.B., Heilig, G.K., 2007. Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China. J. Environ. Manage. 83, 351–364. McGarigal, K., Cushman, S.A., Neel, M.C., Ene, E., 2002. FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. University of Massachusetts, Amherst, MA, http://www.umass.edu/landeco/research/fragstats/fragstats.html (Available on: 28.09.12). McKinney, M.L., 2002. Urbanization, biodiversity, and conservation. BioScience 52, 883–890. MEA, 2005. Ecosystems and Human Well-being: Synthesis. World Resources Institute, 488 Washington, DC. Mitsch, W.J., 2012. What is ecological engineering. Ecol. Eng. 45, 5–12. Moser, B., Jaeger, J.A.G., Tappeiner, U., Tasser, E., Eiselt, B., 2007. Modification of the effective mesh size for measuring landscape fragmentation to solve the boundary problem. Landsc. Ecol. 22, 447–459. Ng, C.N., Xie, Y.J., Yu, X.J., 2011. Measuring the spatio-temporal variation of habitat isolation due to rapid urbanization: a case study of the Shenzhen River crossboundary catchment, China. Landsc. Urban Plan. 103, 44–54. Pascual-Hortal, L., Saura, S., 2006. Comparison and development of new graph-based landscape connectivity indices: towards the priorization of habitat patches and corridors for conservation. Landsc. Ecol. 21, 959–967.
Saura, S., Pascual-Hortal, L.A., 2007. A new habitat availability index to integrate connectivity in landscape conservation planning: Comparison with existing indices and application to a case study. Landsc. Urban Plan. 83, 91–103. SSB (Shenzhen statistics Bureau), 2011. Shenzhen Statistical Yearbook 2010. China Statistics Press, Beijing, China (in Chinese). Su, W., Gu, C., Yang, G., Chen, S., Zhen, F., 2010. Measuring the impact of urban sprawl on natural landscape pattern of the Western Taihu Lake watershed, China. Landsc. Urban Plan. 95, 61–67. SWCA (Shenzhen Wildlife Conservation Administration), 2009. Situation of wildlife conservation in Shenzhen, http://www.szgreen.gov.cn/news. aspx?newsid=5041.html (Available on: 20.08.12). Taylor, P.D., Fahrig, L., Henein, K., Merriam, G., 1993. Connectivity is a vital element of landscape structure. Oikos 68, 571–573. Tischendorf, L., Bender, D.J., Fahrig, L., 2003. Evaluation of patch isolation metrics in mosaic landscapes for specialist vs. generalist dispersers. Landsc. Ecol. 18, 41–50. UNPFA, 2008. State of world population 2007: unleashing the potential of urban growth, United Nation, http://www.unfpa.org/public/home/ publications/pid/408.html (Available on: 28.09.12). Urban, D., Keitt, T., 2001. Landscape connectivity: a graph-theoretic perspective. Ecology 82, 1205–1218. Urban, D.L., Minor, E.S., Treml, E.A., Schick, R.S., 2009. Graph models of habitat mosaics. Ecol. Lett. 12, 260–273. Wu, J.G., Hobbs, R., 2002. Key issues and research priorities in landscape ecology: an idiosyncratic synthesis. Landsc. Ecol. 17, 355–365. Xing, F.W., Yu, M.E., 2000. Wild Plants of Shenzhen China. China Forestry Publishing House, Beijing (in Chinese). Xu, C., Liu, M., Zhang, C., An, S., Yu, W., Chen, J., 2007. The spatiotemporal dynamics of rapid urban growth in the Nanjing metropolitan region of China. Landsc. Ecol. 22, 925–937. Yu, D.Y., Shao, H.B., Shi, P.J., Zhu, W.Q., Pan, Y.Z., 2009. How does the conversion of land cover to urban use affect net primary productivity? A case study in Shenzhen city, China. Agric. Forest Meteorol. 149, 2054–2060. Yu, D.Y., Shi, P.J., Han, G.Y., Zhu, W.Q., Du, S.Q., Xun, B., 2011. Forest ecosystem restoration due to a national conservation plan in China. Ecol. Eng. 37, 1387–1397. Yu, D.Y., Xun, B., Shi, P.J., Shao, H.B., Liu, Y.P., 2012. Ecological restoration planning based on connectivity in an urban area. Ecol. Eng. 46, 24–33. Yu, D.Y., Xun, B., Liu, Y.P., 2013. Measuring landscape connectivity in an urban area for biological conservation. CLEAN-Soil Air Water, http://dx.doi.org/10.1002/clen.201200448 (in press). Zetterberg, A., Mörtberg, U.M., Balfors, B., 2010. Making graph theory operational for landscape ecological assessments, planning, and design. Landsc. Urban Plan. 95, 181–191.