Mongolian Journal of Biological Sciences 2006 Vol. 4(2): 33-40

Habitat Classifi cation Using Landsat 7ETM+ Imagery of the Ikh Nart Nature Reserve and Surrounding Areas in Dornogobi and Dundgobi Aimags, Mongolia

Daniel S. Jackson1, James D. Murdoch2 and Bayart Mandakh3

1Department of Biology, Oberlin College, Oberlin, Ohio 44074, USA* 2Wildlife Conservation Research Unit, University of Oxford, Tubney House, Abingdon Road, Tubney, Abingdon OX13 5QL, UK . 3Mongolian Academy of Sciences, Institute of Botany, Ulaanbaatar, Mongolia

Abstract

Wildlife studies increasingly employ thematic information, such as habitat classes, extracted from remotely sensed image data. Thematic information allows researchers to characterize biological processes of interest at various spatial scales. In and around Ikh Nart Nature Reserve of Mongolia, few large-scale habitat and landscape maps exist. Such maps are needed to support several long-term biological studies. Here we report the results of a maximum likelihood supervised classifi cation of a fi ve-band multispectral composite Landsat 7ETM+ image of the Dalanjargalan Soum of the reserve and surrounding areas in Dornogobi and Dundgobi Aimags. We classifi ed the image into seven habitat classes: dense rock, low-density shrub, high-density shrub-rock mix, semi-shrub steppe, forb- dominated short grass steppe, tall vegetation, and ephemeral standing water bodies. Overall classifi cation accuracy was 90.5% with a Khat statistic of 88.8%, and a user’s accuracy of >85% per class, suggesting strong agreement between map and ground reference information. As such, the map presents a detailed distribution of habitats that may be suitable for a variety of research applications including the analysis of wildlife ranging behavior and the identifi cation of priority areas for conservation.

Key words: Ikh Nart Nature Reserve, Landsat, map, Mongolia, supervised classifi cation, habitat, wildlife

Introduction plan protected areas (Smith et al., 1997). As such, habitat classifi cation represents a valuable tool for Studies that describe biological processes at landscape conservation and research (Johnson et the local, regional, and global scale increasingly al., 2004). employ thematic information extracted from In Mongolia, satellite imagery has been used to remotely sensed satellite image data of the earth survey the country’s ecological landscape in several (Sampson & Delgiudice, 2006). Information studies. It has aided in linking seasonal patterns in from satellite imagery has been applied to diverse primary productivity with the migratory patterns contexts, from the quantifi cation of landscape of Mongolian gazelles (Procapra gutturosa, change (Lunetta et al., 2006) to the assessment of Leimgruber et al. 2001), and also been used to alternative energy viability (Dudhani et al., 2006). assess Mongolia’s vegetation cover (Naidansuren, In particular, habitat information from satellite 2003), and mitigate the growing impact of imagery has been applied to a variety of wildlife livestock (Rasmussen et al. 1999). Studies have studies (Hansen et al., 2001; Braun, 2005; Newton yet to employ high resolution imagery to explore Cross, 2007; Stevens et al., 2007). Johnson et al. the biogeographical characteristics of the Gobi- (2004), for example, used multispectral Landsat steppe ecosystem. images to map (and predict) suitable habitat for Ikh Nartiin Chuluu Nature Reserve in endangered mountain caribou (Rangifer tarandus Dornogobi Aimag (province) sits at the unique caribou) at multiple scales in British Columbia, confl uence of grassland and semi-desert steppe Canada. Other studies have used habitat zones of the Gobi-steppe ecosystem (Reading et information from satellite imagery to examine al., 2006) and is the site of several wildlife studies patterns of habitat use (Jepsen et al., 2002), monitor focusing on ungulates, carnivores, insectivores, variations in richness (Gould, 2000), and and raptors (Reading et al., in press). The reserve

33 http://dx.doi.org/10.22353/mjbs.2006.04.13 34 Jackson et al. Ikh Nart Habitat Classifi cation is also currently in the initial stages of developing reserve along with a few freshwater springs, short a comprehensive, long-term, management plan (i.e., <50 m) streams, oases, alkaline pools, and (Reading et al., in press). To our knowledge, ephemeral ponds. The climate is arid and highly despite growing interest in the region’s fl ora and variable, with temperatures ranging from -40 °C fauna, no large-scale habitat or landscape feature in winter to +43 °C in summer. Precipitation is maps of the reserve exist (with the exception of low (<100 mm/year), falling mostly as rain from low resolution topographic maps). Accurate June through August. habitat maps would provide researchers with a standardized reference for studying wildlife Habitat classifi cation ecology and reserve managers with a tool to more Our classifi cation strategy involved 1) effectively plan conservation measures. identifying major, predominant habitat classes in We identifi ed major habitat classes in the the study area, 2) using known, ground-truthed northern section of Ikh Nart and mapped them sites to train a software classifi er on representative over a broad area by conducting a supervised pixels of a composite image (i.e., multiple overlaid classifi cation of a multispectral Landsat 7ETM+ bands), 3) classifying each pixel of the image into satellite image. Here we report the results of one of the habitat classes, and 4) evaluating the the classifi cation and describe the major habitat accuracy of the classifi cation by examining random classes. We aimed to generate an accurate and test sites on the ground. For the classifi cation, detailed habitat map suitable for use in current we used a Landsat 7ETM+ image (ID: 010-406: wildlife and conservation studies. WRS-2, Path 130, Row 028: eight spectral image bands; pixel resolution = 30 m x 30 m for bands Materials and Methods 1 to 5 and band 7) taken on July 26, 2002 from the Global Land Cover Facility (College Park, Study Area Maryland, USA). The Ikh Nart Nature Reserve (Ikh Nart) is a We identifi ed major habitat classes based on small protected area located in the northwest exploratory unsupervised classifi cations of the region of Dornogobi Aimag (N45.723° E108.645°; Figure 1). Established in 1996, the reserve protects ~66,000 ha of rocky outcrops and steppe habitats that harbor threatened argali sheep (Ovis ammon) and a myriad of other steppe fl ora and fauna (Myagmarsuren, 2000; Reading et al., 2006; Murdoch et al., 2006). Two soums (counties) jointly manage the reserve including Dalanjargalan and Airag. Our study area included the Dalanjargalan Soum section of the reserve (37,919 ha representing the northern 57%) and surrounding regions in Dornogobi and Dundgobi Aimags (total study area = 72,937 ha enclosed by N45.838943°- N45.545245°; E108.489732°- E108.731806°). The study area encompassed the reserve’s ‘core’ zone (7,120 ha; established to protect the argali sheep population) where several wildlife research projects are based. Ikh Nart typifi es the Gobi-steppe ecosystem, characterized by a mosaic of desert, semi-desert, and grassland habitats (Reading et al., 2006). Semi- arid vegetation dominates the reserve, including short grasses, shrubs, semi-shrubs, and forbs clustered among dense igneous and metamorphic Figure 1. Locaton of the Ikh Nart Nature Reserve, rock outcroppings (Reading et al., 2006). Several Mongolia relative to country (inset), aimag (black ephemeral river valleys and creek beds occur in the line), and soum (dashed line) borders. Mongolian Journal of Biological Sciences 2006 Vol. 4(2) 35

Landsat image, previous botanical surveys of We assessed the accuracy of the classifi cation Ikh Nart, and local knowledge of the area. We with fi eld test sites by generating a stratifi ed fi rst conducted unsupervised classifi cations with random sample of 60 test sites (pixels) per class MultiSpec software (version 3.0.2, Laboratory for (Fitzpatrick-Lins, 1981; Congalton & Green, Applications of Remote Sensing, Indiana, USA) 1999) using a Geographic Information System to classify various combinations of composite (ArcGIS v. 9.0, ESRI, Redlands, California, USA). images (three to fi ve bands) into fi ve to ten spectral Visiting each test site, we recorded the ‘actual’ classes (Jensen, 2005). We then referenced the habitat class of the site and the eight surrounding results with on-the-ground knowledge of the pixels (Jensen, 2005). major vegetation and landscape features of the We based classifi cation accuracy on overall, area and previous botanical surveys. We used user’s, and producer’s accuracy measures, and a this initial exploratory information to estimate target Coeffi cient of Agreement, or Khat statistic the number of distinct habitat types that would be (Jensen, 2005). Producer’s accuracy describes potentially resolved by the images in a supervised the probability of a correctly mapped reference classifi cation, but not to classify the fi nal habitat sample. User’s accuracy describes the probability image itself. that a sample from a classifi cation map actually

For the habitat classifi cation, we performed a refl ects what is on the ground. The Khat statistic supervised classifi cation with MultiSpec software considers all class accuracies, providing a more using maximum likelihood methods and the comprehensive measure of map accuracy (see Decision Boundary Feature Extraction (DBFE) Jensen, 2005). We considered scores of at least 85% algorithm (Lee & Landgrebe, 1993; Jensen, for user’s and producer’s accuracies acceptable,

2005). The DBFE algorithm is a commonly used and a Khat score of at least 80% acceptable (Jensen, algorithm in classifi cation that extracts features 2005). (image bands) from the available image bands in order to maximize the spectral differences Class description between habitat classes. This preliminary step To more thoroughly describe the habitat in feature selection considers the possibility that classes, we surveyed the vegetation community the optimal set of features may, in fact, be some at 25 randomly selected training sites per class. linear combination of bands rather than a subset We identifi ed the species present in a 1 m2 plot of those available. Five linearly transformed, at each chosen site. We estimated percent cover normally distributed features were used for the of vegetation and recorded the dominant fi nal classifi cation. species in terms of biomass within a 20 m radius To train the software, we selected 60 training of each plot. We calculated dry biomass (g/m2) of sites (pixels) in each habitat class separated by a new growth for all , and species diversity >10 pixel distance from one another, ensuring a using Simpson’s Diversity Index (Krebs, 1999) selection of at least 25% of training pixels in each for each plot. class at least 12 km outside the central 50% of the study area—a strategy suggested by Jensen (2005) Results to reduce spatial autocorrelation. We stratifi ed training pixels geographically, segregating areas Habitat classifi cation with different spectral characteristics within each Based on exploratory unsupervised habitat class. We selected image bands for the classifi cations, previous botanical surveys of Ikh fi nal composite image used in the classifi cation Nart, and local knowledge of the area, we identifi ed based on an analysis of the distribution of each seven habitat classes including: dense rock (DR), band’s spectral frequencies. Spectral frequencies low-density shrub rock mix (LDS), high-density for all bands except 5 were normally distributed. shrub (HDS), forb-dominated short grass steppe Because maximum likelihood methods assume (SGS), semi-shrub steppe (SSS), tall vegetation normally distributed spectral data (Jensen, 2005), (TV), and ephemeral standing water bodies (WA). we omitted band 5 from the classifi cation. We also Outcrops of granitic and metamorphic rock with omitted bands 6 and 8 because they had different sparse vegetation characterized areas of the dense resolutions than the other bands. We used a DBFE rock class. The low-density shrub rock mix class transformed composite of bands 1, 2, 3, 4, and 7. described areas dominated by woody shrubs at 36 Jackson et al. Ikh Nart Habitat Classifi cation densities of ≤100/ha interspersed with patchy rock outcrops and talus. High-density shrub habitat, by contrast, included open areas with shrubs at densities of >100/ha. Semi-shrub steppe habitat included areas dominated by turfy semi- shrubs and the forb-dominated short grass steppe described areas containing mostly perennial forbs and grasses. Tall grasses and trees (>1 m in mean height in late summer/autumn) characterized the tall vegetation class. The ephemeral standing water body class described ponds, pools, and springs with seasonally variable standing water. The fi nal supervised classifi cation image covered a total area of 794,995 pixels (Figure 2). Low-density shrub rock mix, forb-dominated short grass steppe, and high-density shrub dominated habitats comprised most of the study area, accounting for 27.1%, 24.1%, and 21.2%, respectively (Figure 2). Semi-shrub steppe, dense rock, tall vegetation, and ephemeral water bodies accounted for 9.5%, 8.9%, 6.6%, and 2.6%, respectively (Figure 2). The classifi cation was 90.5% accurate overall, Figure 2. Map of seven habitat classes generated from a maximum likelihood supervised classifi cation of a with a Khat of 88.8% (Table 1). User’s accuracy exceeded 85% in all classes (Table 1). Producer’s Landsat 7ETM+ satellite image of the northern Ikh accuracy varied from 80.0% to 93.3% (Table 1). Nart Nature Reserve and surrounding areas in Dor- The semi-shrub steppe class exhibited the lowest nogobi and Dundgobi Aimags (total area shown= 72,937.3 hectares). The seven habitat classes include: degree of accuracy (Table 1). We observed a dense rock (DR), low-density shrub (LDS), high-den- relatively high degree of commission error (i.e., sity shrub (HDS), forb-dominated short grass steppe when a pixel is included in an incorrect category) (SGS), semi-shrub steppe (SSS), tall vegetation (TV), and omission error (i.e., when a pixel is excluded and ephemeral water bodies (WA). The black line rep- from the category in which it belongs) between resents the reserve boundary. ESRI GIS Shapefi les of the semi-shrub steppe and high-density shrub each habitat are available for free download at http:// classes. www.ikhnart.com.

Class description Atraphaxis frutescens (shrub; Polygonaceae) for Among 150 plots surveyed (WA class excluded), high-density shrub; Reaumuria soongorica (semi- we identifi ed 48 plant species representing 21 shrub; Tamaricaceae) and Salsola passerina (semi- families (Table 2). The most commonly occurring shrub; Chenopodiaceae) for semi-shrub steppe; species across all plots included Stipa gobica polyrrhizum (perennial forb; ) and (perennial grass; Poaceae; 49% of plots), Carex Stipa gobica (perennial grass; Poaceae) for forb- duriuscula (perennial sedge; Cyperaceae; 27% of dominated short grass; Achnatherum splendens plots), and Allium polyrrhizum and A. mongolicum (perennial grass; Poaceae), Ulmus pumila (tree; (perennial forbs; Liliaceae; 44% of plots) (Table Ulmaceae), and Salix ledeubouriana (shrub; 2). The most represented families included Salicaceae) for tall vegetation; and Spiraea Asteraceae (N = 8 spp.), Chenopodiaceae (N = 5 aquilegifolia (shrub; Rosaceae) and for dense spp.), Liliaceae (N = 5 spp.), Poaceae (N = 9 spp.) rock. (Table 2). Percent vegetation cover varied from 0 to Dominant species identifi ed around plots 100% among plots. Mean ± SE % cover for included: Amygdalus pedunculata (shrub; habitat classes ranged from 7.6 ± 2.7 % (dense Rosaceae) for low-density shrub rock mix; rock) to 46.0 ± 8.1 % (tall vegetation) (Table 2). Caragana pygmaea (shrub; Leguminosae) and Mean ± SE dry biomass was highest for the tall Mongolian Journal of Biological Sciences 2006 Vol. 4(2) 37

Table 1. Accuracy assessment of a supervised classifi cation of a Landsat 7ETM+ composite image (bands 1, 2, 3, 4, and 7; DBFE transformed) of Ikh Nart Nature Reserve, Mongolia into seven habitat classes: dense rock (DR), low-density shrub rock mix (LDS), high-density shrub (HDS), forb-dominated short-grass steppe (SGS), semi- shrub steppe (SSS), tall vegetation (TV), and ephemeral water bodies (WA). The assessment compared the actual habitat of random test sites (60/class) to the habitat identifi ed by the classifi cation. In the matrix, values represent the number of correctly classifi ed sites (where two habitats meet) and incorrectly classifi ed sites. For example, 56 actual DR sites were correctly classifi ed as DR and four were incorrectly classifi ed (three as LDS and one as WA). Producer’s refers to the probability that the researcher correctly mapped a reference sample while user’s refers to the probability that a sample from a particular class refl ects what is on the ground. Overall refers to the overall accuracy of the classifi cation (number test sites correctly classifi ed/total number of test sites).

Classifi cation Habitat Actual % DR LDS HDS SSS SGS TV WA Total % User’s Habitat Overall

DR 56 3 0 0 0 0 3 62 90.3 LDS 3 56 0 0 4 0 0 63 88.9 HDS 0 0 53 6 0 2 1 62 85.5 SSS 0 0 5 48 0 3 0 56 85.7 SGS 0 1 2 3 56 0 0 62 90.3 TV 0 0 0 3 0 55 0 58 94.8 WA 1 0 0 0 0 0 56 57 98.2 Total 60606060606060420 % Producer’s 93.3 93.3 88.3 80.0 93.3 91.7 93.3 90.5

vegetation class (8.1 ± 3.7 g/m2) and lowest for the proportion of spectral data available for capture forb-dominated short grass class (1.2 ± 0.5 g/m2) by the satellite and use by the software classifi er. (Table 2). Semi-shrub class represented the most Given that the classes were often adjacent, similar species diverse habitat (Simpson’s index = 0.90), soil characteristics likely contributed equally to whereas the tall vegetation class represented the the spectral character of each class, resulting in least diverse habitat (Simpson’s index = 0.45). similar overall spectral qualities. In our study, the Landsat imagery resolved most of the boundaries Discussion between the SSS and HDS classes, but appeared to have diffi culty when the size of an uninterrupted The level of classifi cation accuracy suggests habitat patch dropped to some threshold below the that there was negligible overlap in the spectral 30 m pixel scale. characteristics of our classes, and strong agreement In most cases, errors between these two classes between the classifi cation and actual ground occurred at the often fuzzy areas of confl uence habitat information. Although overall accuracy between the two habitat classes in the northwestern of the classifi cation was high, we observed a region of the study area. Hansen et al. (2001) relatively high degree of producer’s error for the reported similar patterns of error, and Knick & semi-shrub steppe class. The lower accuracy, Rotenberry (1998) described similar diffi culty which at 80% we believe is still reasonable, mapping arid regions. Though we achieved higher may be due to the spectral similarity and spatial accuracy overall than comparable studies using relationship between semi-shrub steppe and high- Landsat satellite imagery (e.g., Knick et al. 1997), density shrub. Although both classes contained our study area was relatively smaller, allowing us distinct plant communities with unique spectral to conduct supervised training on the ground over characteristics, they exhibited similar vegetation a majority of the area. densities and soil types. For both classes, The high user’s accuracy of all habitat classes then, in a given pixel, soil composed a similar indicates that the classifi cation map is suitable for 38 Jackson et al. Ikh Nart Habitat Classifi cation

Table 2. Plant species identifi ed in 1 m2 plots in six habitat classes (N = 25 plots/class) along with mean (± SE) percent vegetation cover and dry biomass (g/m2), and species diversity (Simpson’s index) across all plots in Ikh Nart Nature Reserve, Mongolia. Habitat classes include: low-density shrub rock mix (LDS), high-density shrub (HDS), semi-shrub steppe (SSS), forb-dominated short grass steppe (SGS), tall vegetation (TV), and dense rock (DR) – ephemeral water body habitat class excluded. Values represent the number of plots in which we found a particular plant species.

Number of plots Family Species LDS HDS SS FSG TV DR

Asclepiadaceae Vincetoxicum sibiricum 000001 Asteraceae Ajania fruticulosa 404101 Artemisia dracunculus 000011 Artemisia frigida 322110 Artemisia pectinata 100100 Artemisia ruthifolia 000013 Artemisia sp. 101010 Asterothamnus central-asiaticus 003600 Scorzonera divaricata 100001 Caryophyllaceae Arenaria capillaris 000003 Stelleria dichotoma 000002 Gypsophila desertorum 200000 Chenopodiaceae Bassia dasyphylla 031000 Chenopodium album 001001 Kochia prostrata 400211 Salsola passerina 0117200 Salsola pestifera 010010 Convolvulaceae Convolvulus ammanni 0014220 Crassulaceae Orostachys fi mbriata 000100 Cyperaceae Carex duriuscula 8591101 Ephedraceae Ephedra equisetina 000001 Ephedra sinica 300100 Iridaceae Iris potaninia 000001 Iris tenuifolia 192100 Labiatae Thymus gobicus 000001 Leguminosae Caragana pygmaea 71001 1 5 Liliaceae Allium anisopodium 001001 Allium eduardii 000009 01850 2 0 Allium polyrrhizum 9111702 Asparagus gobicus 024110 Plumbaginaceae Limonium aureum 001000 Poaceae Achnatherum splendens 0111120 Agropyron cristatum 100001 Cleistogenes soongorica 086000 Cleistogenes squarrosa 250011 Eragrostis minor 000001 Stipa brevifolia 000013 Stipa gobica 16 11 4 11 1 18 Stipa krylovii 500101 Tripogon chinensis 000009 Polygonaceae Atraphaxis frutescens 030000 Rosaceae Amygdalus pedunculata 000001 Rutaceae Haplophyllum dahuricum 110010 Scrophulariaceae Cymbaria dahurica 010100 Tamaricaceae Reaumuria soongorica 0014100 Verbenaceae Caryoptris mongolica 200002 Zygophyllaceae Peganum nigellastrum 010110 Mean ± SE % cover 10.0 20.9 16.0 13.1 46.0 7.6 ±2.7 ± 3.6 ± 2.6 ± 2.9 ± 8.1 ± 2.7 Mean ± SE dry biomass 1.4 2.8 5.7 1.2 8.1 4.4 ± 0.3 ± 0.8 ± 1.1 ± 0.5 ± 3.7 ± 1.4 Simpson’s Diversity Index 0.889 0.885 0.901 0.875 0.453 0.888 Mongolian Journal of Biological Sciences 2006 Vol. 4(2) 39 most current research and conservation studies Hansen, M.J., Franklin, S.E., Woudsma, C.G. & in the Ikh Nart reserve. Potential applications Peterson, M. 2001. Caribou habitat mapping of the map include: analysis and modeling and fragmentation analysis using Landsat of wildlife movement patterns and resource MSS, TM, and GIS data in North Columbia selection functions, quantifi cation of landscape Mountains, British Columbia, Canada. Remote change, and identifi cation of conservation priority Sensing of Environment 77: 50-65. areas. The spectral character of our classes may Jensen, J.R. 2005. Introductory Digital Image also be representative of the reserve as a whole. Processing: a Remote Sensing Perspective. The classifi cation scheme adopted here may be Prentice Hall, New Jersey. robust enough, with some minor changes, for use Jepsen, J.U., Eide, N.E., Prestrud, P. & Jacobsen, in mapping habitats of the southern Airag Soum L.B. 2002. The importance of prey distribution section of the reserve. in habitat use by arctic foxes (Alopex lagopus). Canadian Journal of Zoology 80: 418-429. Acknowledgments Johnson, C.J., Seip, D.R. & Boyce, M.S. 2004. A quantitative approach to conservation We thank the Fulbright Commission, Rufford planning: using resource selection functions Foundation, and Denver Zoological Foundation to map the distribution of mountain caribou for supporting the project. We also thank M. at multiple spatial scales. Journal of Applied Ondra and Earthwatch volunteers for help with Ecology 41: 238-251. plant surveys, as well as the assistance of S. Knick, S.T. & Rotenberry, J.T. 1998. Limitations Amgalanbaatar, T. Munkhzul, G. Otgonbayar, N. to mapping habitat use areas in changing Enkhtuvshin and N. Narantuya. Richard Reading landscapes using the Mahalanobis distance provided valuable comments on early versions statistic. Journal of Agricultural, Biological, of the manuscript. We conducted the project in and Environmental Statistics 3: 311-322. collaboration with the Mongolian Academy of Krebs, C. 1999. Ecological Methodology. Second Sciences under permit #3/451 and are grateful for edition. Addison-Welsey, Longman, New the support of Janchiv and B. Lhagvasuren of the York. Institute of Biology. ESRI GIS shapefi les of each Lee, C. & Landgrebe, D.A. 1993. Feature habitat class are available for free download at extraction based on decision boundaries. IEEE www.ikhnart.com. Transactions on Pattern Analysis and Machine Intelligence 15: 388-400. References Leimgruber, P., McShea, W.J., Brookes, C.J., Bolor-Erdene, L., Wemmer, C. & Larson, Braun, C. (ed.) 2005. Techniques for Wildlife C. 2001. Spatial patterns in relative primary Investigation and Management. Sixth edition. productivity and gazelle migration in the The Wildlife Society, Bethesda, Maryland. eastern steppes of Mongolia. Biological Congalton, R.G. & Green, K. 1999. Assessing the Conservation 102: 205-212. Accuracy of Remotely Sensed Data: Principles Lunetta, R.S., Knight, J.F., Ediriwickrema, and Practices. Lewis Publishers, New York. J., Lyon, J.G. & Worthy, L.D. 2006. Land Dudhani, S., Sinha, A.K. & Inamdar, S.S. 2006. cover change detection using multi-temporal Assessment of small hydropower potential MODIS NDVI data. Remote Sensing of the using remote sensing data for sustainable Environment 105: 142-154. development in India. Energy Policy 34: 3195- Myagmarsuren, D. 2000. Special protected 3205. areas of Mongolia. Mongolian Environmental Fitzpatrick-Lins, K. 1981. Comparison of sampling Protection Agency and the German Technical procedures and data analysis for a land-use and Advisory Cooperation (GTZ), Ulaanbaatar, land-cover map. Photogrammetric Engineering Mongolia. and Remote Sensing 47: 343-351. Murdoch, J., Munkhzul, T., Amgalanbaatar, S. & Gould, W. 2000. Remote sensing of vegetation, Reading, R. In Press. Checklist of mammals in plant species richness, and regional biodiversity Ikh Nart Nature Reserve. Mongolian Journal hotspots. Ecological Applications 10: 1861- of Biological Sciences.4: 69-74. 1870. Naidansuren, E. 2003. Time-series satellite data 40 Jackson et al. Ikh Nart Habitat Classifi cation

analysis for assessment of vegetation cover in Õóðààíãóé Mongolia. Mongolian Journal of Biological Sciences. 1: 25-32. Зайнаас тандах судалгааны өгөгдлөөс Newton Cross, G., White, P.C.L. & Harris S. 2007. гарган авсан амьдрах орчны ангилал зэрэг Modeling the distribution of badgers Meles сэдэвчилсэн мэдээллийг зэрлэг амьтдын meles: comparing predictions from fi eld-based судалгаанд өргөн ашиглах боллоо. Сэдэвчилсэн and remotely derived habitat data. Mammal мэдээлэл нь биологийн процессыг орон Review 37: 54-70. зайн янз бүрийн масштабад тодорхойлох Rasmussen, M.S., James, R., Adiyasuren, T., боломжийг судлаачдад олгоно. Монгол Улсын Khishigsuren, P., Naranchimeg, B., Gankhuyag, тусгай хамгаалалттай газар болох Их нартын R. & Baasanjargal, B. 1999. Supporting байгалийн нөөц газрын болон түүний эргэн Mongolian pastoralists by using GIS to тойрны амьдрах орчны ба ландшафтын том identify grazing limitations and opportunities масштабын цөөн зураг хийгдсэн байдаг. Эдгээр from livestock census and remote sensing data. зураг биологийн удаан хугацааны хэд хэдэн GeoJournal. 47: 563-571. судалгааны үндэс болдог. Энд бид Их нартын Reading, R., Kenny, D., Wingard, G., Mandakh, байгалийн нөөц газрын Даланжаргалан сумын B. & Steinhauer-Burkart, B. 2006. Ikh хэсэг болон түүний эргэн тойрны Дорноговь, Nart Nature Reserve. ECO Nature Edition, Дундговь аймгуудын зарим нутгийн Landsat Ulaanbaatar, Mongolia. 7ETM+-ийн таван сувгийн нийлмэл зургаас Reading R., Amgalanbaatar, S., Murdoch, J., хамгийн их магадлалын хяналттай ангиллын Wingard, G. & Kenny, D. In Press. Working аргыг ашиглан амьдрах орчныг ангилсан үр to conserve Mongolia’s natural heritage. In: дүнг толилуулж байна. Бид зургаас хад чулуут, Proceedings of the Association of Zoos and бага нягтшилтай сөөгт, өндөр нягтшилтай Aquariums Annual Conference. (Ed. by T. сөөг-хад чулуу хосолсон, заримдаг сөөгт хээр, Lewthwaite). AZA, Bethesda, Maryland. алаг өвс давамгайлсан богино үетэнт хээр, Sampson, B.A. & Delgiudice, G.D. 2006. Tracking өндөр ургамал, түр зуурын тогтмол усан сан the rapid pace of GIS-related capabilities and гэсэн долоон амьдрах орчны ангийг ялгалаа. their accessibility. Wildlife Society Bulletin 34: Ангиллын нийт нарийвчлал 90.5% (анги тус 1446-1454. бүрийн Khat статистик 88.8% ба хэрэглэгчийн Smith, A.P., Horning N. & Moore, D. 1997. нарийвчлал >85%) байсан нь зургаас хийсэн Regional biodiversity planning and lemur ангилал ба газарт хийсэн бодитой ажиглалтын conservation with GIS in western Madagascar. хоорондын тохирол өндөр байгааг илтгэж Conservation Biology 11: 498-512. байна. Иймд амьдрах орчны тархалтыг Stevens, C.E., Paszkowski, C.A. & Foote, A.L. нарийвчлан харуулсан энэхүү зураг нь 2007. Beaver (Castor canadensis) as a surrogate зэрлэг амьтны орон зайн зан төрхийг судлах, species for conserving anuran amphibians on хамгаалалтын тэргүүлэх ач холбогдол бүхий boreal streams in Alberta, Canada. Biological газар нутгийг тодорхойлох зэрэг судалгаанд Conservation 134: 1-13. хэрэглэгдэх боломжтой гэж дүгнэв.

Received: 20 January 2008 Accepted: 20 February 2008