Spatial Pattern Analysis for Monitoring of National Parks: Sorkhe Hesar National Park, Iran
Afshin Alizadeh Shabani,
In the developing countries, habitat loss and fragmentation severely affect
protected area, due to the rapid expansion of anthropogenic activities that
has occurred in the last decades. Landscape metrics are commonly used to define
landscape patterns and to evaluate fragmentation processes. In this study, remote
sensing data and landscape metrics are used to investigate the changes in land
use and landscape pattern in Sorkhe Hesar national park, Iran over a study period
from 1986-2010. Landsat Thematic Mapper (TM) images are processed using ENVI
and landscape metrics are calculated at landscape and class levels using FRAGSTATS.
Landscape metrics analysis shows that land use patterns in Sorkhe Hesar have
been undergoing rapid changes over the past 25 years. Changes were mostly due
to the conversion of poor range and orchard to urban built up area which in
2010 had increased by more than 18.9 times in 1986. An evaluation of landscape
pattern based on landscape metrics analysis indicates rapid urbanization from
1986-2010. Due to rapid economic development and urbanization, the landscape
became more fragmented from 1986-2010. The study demonstrates that landscape
metrics can be a useful indicator in land use change analysis and are vital
for integrated landscape evaluation.
Received: September 12, 2013;
Accepted: October 29, 2013;
Published: March 08, 2014
Land cover change is one of the most important aspects of environmental change and represents the largest threat to ecological systems over the past several decades (Sala et al., 2000; Hansen et al., 2004; Foody, 2003). Habitat loss and fragmentation have been directly linked to biodiversity loss and species extinction (Laurance et al., 1997; Pimm, 1998; Brook et al., 2003; Hanski et al., 2007) and can severely alter the physical or biotic conditions of habitat, directly influence species distribution patterns and even induce species loss (Laurance et al., 1998; Williams-Linera et al., 1998; Sala et al., 2000; Hanski et al., 2007). Over the recent decade, protected areas have expanded globally in order to conserve biological and cultural resources (Rodriguez and Young, 2000; Chape et al., 2003; Zimmerer et al., 2004). It is recognized that protected areas are effective in protecting some kinds of biodiversity (Bruner et al., 2011) although extinction of native species still occurs inside reserves (Woodroffe and Ginsberg, 1998). National parks land its cover has been rapidly deteriorating during the past several decades, with considerable impact on biodiversity and ecosystem services (Xu et al., 2006). Causes of landscape change of protected area and conservation strategies to protect biodiversity will vary, but it is apparent that economic, political and conservation forces at local, national and global scales are drivers of land cover change (Homewood et al., 2001; Jokisch and Lair, 2003) and influence the designation and management of protected areas (Zimmerer et al., 2004; Bonta, 2005).
Developing concepts and tools to describe and quantify land cover or landscape patterning is essential to the study of change in land cover patterns (Forman, 1995; Turner et al., 2001; Ivits et al., 2005). Due to the difficulty of accessing many protected area, remote sensing has been employed as an important tool to carry out ecological studies and for mapping, monitoring and modeling protected area (Aplin, 2005; Joshi et al., 2001; Dorner et al., 2002; Munsi et al., 2010). Remote sensing is a proven technology that is effective for mapping and characterizing natural resources (Jensen, 1996; Campbell, 1997; Welch et al., 2002). Remote sensing allows observation and measurement of biophysical characteristics of the landscape and tracking of changes in landscapes over time (Parmenter et al., 2003; Wang and Moskovits, 2001). Remote sensing change detection can be used to discern and simulate areas that have been altered by natural or anthropogenic processes (Jantz et al., 2003; Hansen et al., 2002). Understanding the magnitude and pattern of land-cover change helps establish a landscape context for the national parks and protected areas and offers resource managers a better understanding of how park ecosystems fit into the broader landscape.
The integration of landscape ecology and remote sensing through using landscape metrics may provide the key to unraveling landscape and land cover pattern dynamics from temporally discrete data (Crews-Meyer, 2002). Landscape metrics have been increasingly employed to address whether and how land cover and landscape patterns have changed through time (Turner et al., 2001; Batistella et al., 2003) and to link these changes to patterned ecological and environmental processes such as disturbance, fragmentation and deforestation (Griffiths and Lee, 2000; Turner et al., 2001; Batistella et al., 2003; Narumalani et al., 2004; Frohn and Hao, 2006). Currently, landscape metrics are widely used and have been incorporated into widely used landscape analysis software packages such as FRAGSTATS (McGarigal et al., 2002). The objective of the present work was to quantify and characterize land cover Change and monitoring of Sorkhe Hesar national park using remote sensing and landscape metrics from 1986-2010.
MATERIALS AND METHODS
Sorkheh Hesar National Park (Fig. 1) lies with an area of 9,380 ha (23,200 acres) and stands at an altitude of 1,547 m (5,075 ft) above sea-level, nearby Ray-Tehran 20th District, Tehran, Iran. Whole this park, except the northeastern part has been protected by Iran Environmental Protection Organization since 1980. The major portion of this territory is a base for immigrant birds during winter. The area has semi-arid climate and the day-night temperature difference is very high. The maximum temperature annual average in the area is 18.9°C (66.0°F) (Bumabad, 2001).
For this study, we acquire nine predominantly cloud-free Landsat TM images of the Sorkhe Hesar in Iran (WRS2 path 165, row 35). Cloud-free images are obtained from June to July.
Three images are georeferenced to the master image with a topographic map provided by the Iranian National Cartographic Center at 1:25000 scale.
|Fig. 1:|| Study area
The images are resampled to 30 by 30 m pixels using a nearest neighbor resampling algorithm with a first-order binomial. The number of Ground Control Points (GCPs) used for the registration varies by image and in all but one case, the Root Mean Square Error (RMSE) of the registration process is less than a pixel. The RMSE of the 1984, 2000 and 2010 images were 0.45, 0.48, 0.42 pixel respectively. To correct for changes in atmospheric conditions, illumination angles and seasonal variation across the images, a relative radiometric normalization technique is used (Song et al., 2001). Other techniques are available (Schott et al., 1988; Hall et al., 1991) but they are not suitable for images of the Sorkheh Hesar because there is a dearth of well defined spectrally stable dark and bright ground features. Different land use types are then categorized by using both an unsupervised classification and a supervised classification algorithm. The land use classification system of land use survey is chosen and used to form the classification system for this study. The classes specified in this study consist of good range, poor range, orchard and urban built up area. The ISODATA and K-MEANS algorithms are firstly used to perform unsupervised classification to obtain different land use clusters using different colour composites. This preliminary interpretation can maximally reduce the artificial errors and select the most appropriate clusters for further processing. According to the land use classification system specified above, the Maximum Likelihood algorithm is then used to improve the accuracy of the land use classification for the images on all three dates. At least 40 Areas of Interest for each class are created for editing signatures and their error matrixes are compared. The signature with more than 90% accuracy for each class will be saved for processing.
|Table 1:|| Result of accuracy assessment
The spatial resolution of Landsat TM data is 30 m, in which urban built up area can be clearly visualized but cannot be well classified using supervised classification. Thus, urban built up area is not categorized in the classification system. However, landscape metrics such as MESH are very sensitive to human made class (Herzog and Lausch, 2001; Lausch and Herzog, 2002). Thus, urban built up area should not be omitted when calculating landscape metrics. In order to resolve this problem, a new vector layer is built by digitizing the road element using a manual method (Contrast Stretch). It is then recoded as a new class and merged with the classified images. The new image obtained is then comprised of four land use classes: Good range, poor range, orchard and urban built up area (Fig. 2).
Accuracy assessment is critical for a map generated from any remote sensing data (Congalton and Green, 1999; Lunetta and Lyon, 2004). In order to ensure that the dates of images and referenced data are obtained from the corresponding periods, aerial photography and ground truth method using GPS are employed to conduct accuracy assessment. A confusion matrix is used to evaluate the classification accuracy of each image. For the 1984 and 2000 images, classification accuracies are evaluated using aerial photography taken in the corresponding years obtained from Iranian National Cartographic Center. For the 2010 image classification accuracies are evaluated using GPS control point that obtained from ground survey (including about 2% area). The overall accuracy is 82.60% for the 1984 image and 87.29% for the 2000 image and 96.73% for the 2010 image (Table 1).
Landscape metrics can be categorized into four groups, namely patch area metrics, edge and shape metrics, diversity metrics and configuration metrics (Herzog and Lausch, 2001; Herzog et al., 2001). These metrics can reflect the composition and configuration of landscape pattern. Metrics at the class level are helpful for the understanding of landscape change. The criteria used to select landscape metrics are to fully reflect their conceptual basis and to well describe landscape changes in the study area. Metrics selected in this study are: Class Area (CA), Number of Patches (NP), Mean Patch Size (MPS), Largest Patch Index (LPI), Landscape Shape Index (LSI), Mean nearest neighbor distance (ENN). The short descriptions and reasons selected in this study are listed in Table 2. All these metrics are calculated using the FRAGSTATS software.
RESULTS AND DISCUSSION
Class area (CA): Class Area (CA) is the sum of the area of all patches of the corresponding class (habitat type) and its unit is hectare. CA is useful when comparing different study areas with the same extent. In 1986, Sorkhe Hesar national park was dominated by poor range and good range which together account for 97.9% of the total area (Table 3). In contrast, built up area covered only 90.36 ha, a mere 0.99% of the total area.
||Sorkheh Hesar Natinal Park in (a) 1984, (b) 2000 and (c) 2010
|Table 2:|| Calculation of landscape ecology metrics
|Table 3:|| Landscape metrics result
This indicated that Sorkhe Hesar was still an range dominated area and urban development had just begun at the time. Good range increased slightly since 1986-2000 and then increased severe in 2010. Poor range decreased from 1986-2000 and then decreased severe until 2010. Orchard increased slightly since 1986-2000 and then from 1986-2000, orchard reduction started at a rapid speed and a total of 73.89 ha was transferred. Many orchard areas were converted into built up area. Built up area increased severe since 1986-2000 (614.97 ha) and from 2000-2010, urban expansion continued at a rapid speed and a total of 1013.31 was added. Built up area had a great development trend number (18.96 times). From 1986-2010. Built up area increasing have a direct relation with orchard decreasing, because the majority of built up patch begin from orchard area.
NP: Number of Patches (NP) has limited interpretive value by itself; however, it can provide useful information when compared with other metrics such as Patch Density (PD) or Mean Patch Size (MPS). PD has the same basic utility as NP except that it facilitates comparisons among landscapes of varying sizes. Based on the principles of landscape ecology, the increasing number of patches is an index of ecosystems degradation (Gergel and Turner, 2002). From 1986-2010, the value trend of the NP metric decreased constantly in good range. This meant fragmentation reduction in good range in the study period. A similar changing trend was found between poor range and orchard. The NP of both of them increased from 1986-2000 and decreased from 2000-2010 (Table 3). The NP of built up area increased from 1986-2000 and decreased from 2000-2010. This meant built up patches have a rapid expansion and mixed together in the study period.
MPS: MPS metric determines the average size of the patches. Changes in the metric are proportional to the area and number of patches in the class or landscape (McGarigal and Marks, 1995). The most increase in MPS metric is for built up area in the study period. This meant growth of built up patches from 1986-2010. However, MPS of built up area increased constantly which reflected the process of aggregation during urban development. This is the result of rapid economic development leading to an increasing demand for urban development and housing. Urban sprawl and many newly built areas were established and then extended to the former poor range and orchard that resulted in the increase of MPS. A similar increasing trend was found in good range that regarding to MPS metric, small patches of good range represent a profile that there is a trend to form larger patches of good range through integrating the smaller once. The witnessed trend can be evaluated as a suitable trend mainly due to conservation value of larger patches. There is no significant change in Orchard since 1986-2010. The value of the metric in the poor range has a decreasing trend from 1986-2010. This meant fragmentation in the poor range.
LPI: LPI metric decreasing meant larger patches fragmentated to small patches and patchiness in landscape. Patchiness in integrated landscape has a intense effect on ecological process (Veldcamp and Lambin, 2001). LPI metric value increase in built up and good range over study time. Increasing in LPI metric in built up meant developemrnt in built up patches and this is not a favorite trend in national park. LPI metric have a decreasing trend for poor range and orchard over study time. However, the poor ranges LPI consistently declined during the whole study period which was the result of gradual encroachment of poor range by urbanization.
LSI: In the landscape shape index, the sum of all patch perimeters is divided by a figure equivalent to the perimeter of a circle with the same area as the landscape area. LSI increases with increasing overall complexity of patch shapes. Increase of the metric meant irregularity in patches form. The metric values have a decrease trend in all of classes. This trend meant irregularity in area.
ENN: Measures edge-edge distance between a patch and the nearest neighboring patch of the same class (m). The metric values have decreasing trend in good range and built up area and also increasing trend in poor range and orchard from 1986-2010.
This study demonstrates the applicability of remote sensing and GIS analysis in evaluating landscape change in Sorkhe Hesar, a rapidly developing national park in Iran. Result of this study shows, the increase in built up development and increasing land use changes in the study area during the last 25 years has been changed the landscape of the area. Due to urbanization and industrialization, urban and built up areas have significantly increased at the expense of poor range and orchard. In 1986, built up was only 0.99% of total area. Urban built up area had increased by more than 18.9 times during the study period from 1986-2010. On the other hand, poor range and orchard had reduced by more than 2 times during this period.
There has been a rapid change of landscape pattern at Sorkhe Hesar too. The landscape matrix was dominated by poor range landscape before 1986, then gradually urban built up area became more prominent and resulted in urban built up area. Meanwhile, the landscape has been undergoing homogenization along with urbanization. However, whether this trend would continue through time will require further investigation based on the latest remote sensing images. If this trend continues, the lack of proper planning and management of land use can cause irreversible damage to the ecosystem structure. The result shows that controlling the residential class in the area is necessary for conservational goals and decreases the natural resources degradation. Results also demonstrate that landscape metrics can be used to determine the landscape matrix where land use statistics cannot identify it well. In conclusion, this study demonstrates that landscape metrics analysis can be a useful tool for land use monitoring and assessment. It can provide essential information on the temporal and spatial changes of land use and can be used to examine the landscape matrix which is based on land use statistics only. Furthermore, the functional change of landscape, especially in a human dominated and rapidly developing area can also be synchronously revealed via analyzing landscape pattern changes.
Aplin, P., 2005.
Remote sensing: Ecology. Prog. Phys. Geog., 29: 104-113.
Batistella, M., S. Robeson and E.F. Moran, 2003.
Settlement design, forest fragmentation and landscape change in Rondonia, Amazonia. Photogramm. Eng. Rem. S., 69: 805-812.Direct Link |
Bonta, M., 2005.
Becoming-forest, becoming-local: Transformations of a protected area in Hondouras. Geoforum, 36: 95-112.CrossRef | Direct Link |
Leitao, B.A. and J. Ahern, 2002.
Applying landscape ecological concepts and metrics in sustainable landscape planning. Lands. Urban Plann., 59: 65-93.CrossRef |
Brook, B.W., N.S. Sodhi and P.K.L. Ng, 2003.
Catastrophic extinctions follow deforestation in Singapore. Nature, 424: 420-426.Direct Link |
Bruner, A.G., R.E. Gullison, R.E. Rice and G.A.B. de Fonseca, 2011.
Effectiveness of parks in protecting tropical biodiversity. Science, 291: 125-128.CrossRef |
Study of the environmental management plan for Jajrud 2001. Office of Habitat and Areas, Environmental Consulting Engineers, Environmental Protection Organization of Iran, Tehran, Iran.
Campbell, J.B., 1997.
Introduction to Remote Sensing. Guilford Press, New York
Chape, S., S. Blyth, P. Fox and M. Spalding, 2003.
United Nations list of protected areas. IUCN, Gland, Switzerland and Cambridge, UK and UNEP-WCMC, Cambridge, UK.
Congalton, R. and K. Green, 1999.
Assesssing the Accuracy of Remotely Sensed Data: Principles and Practices. Lewis Publications, Boca Raton, FL
Crews-Meyer, K.A., 2002.
Characterizing landscape dynamism using paneled-pattern metrics. Photogramm. Eng. Rem. S., 68: 1031-1040.Direct Link |
Cushman, S.A., K. McGarigal and M.C. Neel, 2008.
Parsimony in landscape metrics: Strength, universality and consistency. Ecol. Indic., 8: 691-703.CrossRef |
Dorner, B., K. Lertzman and J. Fall, 2002.
Landscape pattern in topographically complex landscapes: Issues and techniques for analysis. Landscape Ecol., 17: 729-743.CrossRef | Direct Link |
Foody, G.M., 2003.
Remote sensing of tropical forest environments: towards the monitoring of environmental resources for sustainable development. Int. J. Remote Sensing, 24: 4035-4046.CrossRef |
Forman, R.T.T., 1995.
Land Mosaics: The Ecology of Landscapes and Regions. Cambridges University Press, Cambridge, England, ISBN: 9780521479806, Pages: 632
Frohn, R.C. and Y. Hao, 2006.
Landscape metric performance in analyzing two decades of deforestation in the Amazon Basin of Rondonia, Brazil. Remote Sens. Environ., 100: 237-251.CrossRef |
Gergel, S.E. and M.G. Turner, 2002.
Learning Landscape Ecology: A Practical Guide to Concepts and Techniques. Spriger Publisher, New York, USA Pages: 316
Griffiths, G.H. and J. Lee, 2000.
Landscape pattern and species richness; regional scale analysis from remote sensing. Int. J. Remote Sens., 21: 2685-2704.CrossRef | Direct Link |
Hall, F.G., D.E. Strebel, J.E. Nickeson and S.J. Goetz, 1991.
Radiometric rectification toward a common radiometric response among multidate, multisensor images. Remote Sens. Environ., 35: 11-27.CrossRef | Direct Link |
Hansen, A.J., R. Rasker, B. Maxwell, J.J. Rotella, A. Wright and U. Langner, 2002.
Ecology and socioeconomics in the New West: A case study from Greater Yellowstone. Bioscience, 52: 151-168.
Hansen, A.J., R. De Fries and W. Turner, 2004.
Land use Change and Biodiversity: A Synthesis of Rates and Consequences during the Period of Satellite Imagery. In: Land Change Science: Observing, Monitoring and Understanding Trajectories of Change on the Earth`s Surface, Gutman, G. and C. Justice (Eds.). Springer Verlag, New York, ISBN: 10: 1402025610, pp: 277-299
Hanski, I., H. Koivulehto, A. Cameron and P. Rahagalala, 2007.
Deforestation and apparent extinctions of endemic forest beetles in Madagascar. Biol. Lett., 3: 344-347.Direct Link |
Hargis, C.D., J.A. Bissonette and J.L. David, 1998.
The behavior of landscape metrics commonly used in the study of habitat fragmentation. Landscape Ecol., 13: 167-186.CrossRef | Direct Link |
Herzog, F. and A. Lausch, 2001.
Supplementing land-use statisticswith landscape metrics: Some methodological considerations. Environ. Monit. Assess., 72: 37-50.PubMed |
Herzog, F., A. Lausch, E. Muller, H.H. Thulke, U. Steinhardt and S. Lehmann, 2001.
Landscape metrics for assessment of landscape destruction and rehabilitation. Environ. Manage., 27: 91-107.PubMed |
Homewood, K., E.F. Lambin, E. Coast, A. Kariuki and I. Kikula et al
Long-term changes in Serengeti-Mara wildebeest and land cover: Pastoralism, population, or policies? Proc. Natl. Acad. Sci. USA., 98: 12544-12549.Direct Link |
Ivits, E., B. Koch, T. Blaschke, M. Jochum and P. Adler, 2005.
Landscape structure assessment with image grey-values and object-based classification at three spatial resolutions. Int. J. Remote Sensing, 26: 2975-2993.Direct Link |
Jantz, C.A., S.J. Goetz and M.A. Shelley, 2003.
Using the SLEUTH urban growth model to simulate the land use impacts of policy scenarios in the Baltimore-Washington metropolitan area. Environ. Plann. B., Plan. Design, 31: 251-271.Direct Link |
Jensen, J.R., 1996.
Introductory Digital Image Processing: A Remote Sensing Perspective. 2nd Edn., Prentice-Hall, Upper Saddle River, New Jersey, Pages: 316
Jokisch, B.D. and B.M. Lair, 2003.
One last stand? Forests and change on Ecuador's Eastern Cordillera. Geogr. Rev., 92: 235-256.Direct Link |
Joshi, P.K., S. Singh, S. Agarwal and P.S. Roy, 2001.
Forest cover assessment in Western Himalayas Himachal Pradesh using IRS 1C/1D WiFS data. Curr. Sci. India, 80: 941-947.Direct Link |
Laurance, W.F., L.V. Ferreira, M.J. Rankin-De-Merona and S.G. Laurance, 1998.
Rain forest fragmentation and the dynamics of Amazonian tree communities. Ecology, 79: 2032-2040.
Laurance, W.F., S.G. Laurance, L.V. Ferreira, M.J. Rankin-De-Merona, C. Gascon and T.E. Lovejoy, 1997.
Biomass collapse in Amazonian forest fragments. Science, 278: 1117-1118.
Lausch, A. and F. Herzog, 2002.
Applicability of landscape metrics for the monitoring of landscape change: Issues of scale, resolution and interpretability. Ecol. Indic., 2: 3-15.CrossRef |
Williams-Linera, G., V.D. Gastelu and M.E.G. Zurita, 1998.
Microenvironment and floristics of different edges in fragmented tropical rainforest. Conserv. Biol., 12: 1091-1102.Direct Link |
Lunetta, R.S. and L.G. Lyon, 2004.
Remote Sensing and GIS Accuracy Assessment. CRC Press, Boca Raton, FL., USA
McGarigal, K., S.A. Cushman, M.C. Neel and E. Ene, 2002.
FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Version 3.0, University of Massachusetts, Amherst, MA
Munroe, D.K., H. Nagendra and J. Southworth, 2007.
Monitoring landscape fragmentation in an inaccessible mountain area: Celaque National Park, Western Honduras. Landscape Urban Plann., 83: 154-167.Direct Link |
Munsi, M., S. Malaviya, G. Oinam and P.K. Joshi, 2010.
A landscape approach for quantifying land-use and land-cover change (1976-2006) in middle Himalaya. Reg. Environ. Change, 10: 145-155.CrossRef | Direct Link |
Narumalani, S., D.R. Mishra and R.G. Rothwell, 2004.
Change detection and landscape metrics for inferring anthropogenic process in the greater EFMO area. Remote Sens. Environ., 91: 478-489.Direct Link |
Parmenter, A.P., A. Hansen, R. Kennedy, W. Cohen and U. Langner et al
Land use and land cover change in the Greater Yellowstone Ecosystem: 1975-95. Ecol. Appl., 13: 687-703.
Pimm, S.L., 1998.
Ecology: The forest fragment classic. Nature, 393: 23-24.Direct Link |
Rodriguez, L.O. and K.R. Young, 2000.
Biological diversity of Peru: Determining priority areas for conservation. Ambio, 29: 329-337.Direct Link |
Sala, O.E., F.S. Chapin III, J.J. Armesto, E. Berlow and J. Bloomfield et al
Global biodiversity scenarios for the year 2100. Science, 287: 1770-1774.CrossRef | Direct Link |
Schott, J.R., C. Salvaggio and W.J. Volchok, 1988.
Radiometric scene normalization using pseudoinvariant features. Remote Sens. Environ., 26: 1-16.CrossRef | Direct Link |
Song, C., C.E. Woodcock, K.C. Seto, M.P. Lenney and S.A. Macomber, 2001.
Classification and change detection using Landsat TM data when and how to correct atmospheric effects. Photogrammetric Eng. Remote Sens., 75: 230-244.Direct Link |
Soverel, N.O., N.C. Coops, J.C. White and M.A. Wulder, 2010.
Characterizing the forest fragmentation of Canada's national parks. Environ. Monit. Assess., 164: 481-499.CrossRef | Direct Link |
Turner, M.G., R.H. Gardner and R.V. O'Neill, 2001.
Landscape Ecology in Theory and Practice: Pattern and Process. Springer-Verlag, New York, USA., ISBN: 9780387951232, Pages: 401
Veldcamp, A. and E.F. Lambin, 2001.
Predicting land use change. Agric. Ecosyst. Environ., 85: 1-6.CrossRef | Direct Link |
Wang, Y. and D.K. Moskovits, 2001.
Tracking fragmentation of natural communities and changes in land cover: Applications of landsat data for conservation in an urban landscape (Chicago Wilderness). Conserv. Biol., 15: 835-843.CrossRef | Direct Link |
Welch, R., M. Madden and T. Jordan, 2002.
Photogrammetric and GIS techniques for the development of vegetation database of mountainous areas: Great Smoky Mountains National Park. ISPRS. J. Photogramm., 57: 53-68.CrossRef | Direct Link |
Woodroffe, R. and J.R. Ginsberg, 1998.
Edge effects and the extinction of populations inside protected areas. Science, 280: 2126-2128.CrossRef | Direct Link |
Xu, J., J. Fox, D. Melick, Y. Fujita and A. Jintrawet et al
Land use transition, livelihood and environmental services in mountane mainland southeast Asia. Mt. Res. Dev., 26: 278-284.Direct Link |
Yu, X. and C. Ng, 2006.
An integrated evaluation of landscape change using remote sensing and landscape metrics: A case study of Panyu, Guangzhou. Int. J. Remote. Sens., 27: 1075-1092.Direct Link |
Zimmerer, K.S., R.E. Galt and M.V. Buck, 2004.
Globalization and multi-spatial trends in the coverage of protected- area conservation (1980-2000). Ambio, 33: 520-529.PubMed |
McGarigal, K. and B.J. Marks, 1995.
FRAGSTATS: Spatial pattern analysis program for quantifying landscape structure. Gen. Tech. Rep. PNW-GTR-351, United States Department of Agriculture, Pacific Northwest Research Station, Portland, OR., USA., pp: 62.