CC BY
43
АРИДНЫЕ ЭКОСИСТЕМЫ, 2007, том 13, № 33-34
ОТРАСЛЕВЫЕ ПРОБЛЕМЫ ОСВОЕНИЯ ЗАСУШЛИВЫХ ЗЕМЕЛЬ =
УДК 581.5
MAPPING THE VEGETATION OF SOUTHERN MONGOLIAN PROTECTED AREAS: APPLICATION OF GIS AND REMOTE SENSING TECHNIQUES
H. von Wehrden* **, K. Wesche*
* University of Halle/Wittenberg, Germany, E-Mail: henrikvonwehrdenaweh. de ** Research Institute of Wildlife Ecology, Savoy en Strasse 1, Vienna, 1160 Austria
Abstract. We present initial results of a vegetation survey of the Great Gobi A Strictly Protected Area. The plant biodiversity within this area is closely related to the average precipitation pattern. Thirteen plant communities were derived, and we chose plots for the nine zonal communities as ground truth data for a supervised classification of Landsat data. Accuracy of the final habitat map was > 90 %. We additionally give an overview of all remotely sensed data which we compiled to create a complex GIS data base. The GIS can serve as a tool for the protection and conservation of Przewalski Horses and Khulans.
Key words. Mongolia, Khulan, Takhi, Landsat, rangelands, biodiversity, DCA
Introduction
The protected areas of the southern Mongolian Gobi host many endangered animals and plants. At a total size of more than 90 000 square kilometres these reserves cover vast regions, which are appropriate because of the large home range of several focal species, mainly mammals. The range of habitats extends from high mountain peaks downwards into depressions, but current land use and climate change are thought to threaten some of these ecosystems (Gunin et al., 1999; Batkhishig, Lehmkuhl, 2003; Christensen et al., 2004).
Montane and extrazonal habitats show patchy mosaics of different plant communities, whereas the semi-desert and desert vegetation is rather homogeneous. Several nationwide overviews of the Mongolian vegetation have already been compiled (Anonymous, 1990; Gunin, Vostokova, 1995; Hilbig, 2000), but higher resolution data are needed to understand the habitat requirements of the wildlife in order to improve protection schemes. Mapping based on remote sensing and GIS techniques offers an alternative to classic survey methods, which may be too time-consuming with respect to the size of regions to be covered (Gunin et al., 1999). Landsat is a standard platform for vegetation mapping, offering a suitable spatial scale for ecological applications (Cohen, Goward, 2004) even in dry and species poor environments (Nagendra, 2001).
Interannual changes are pronounced in these non-equilibrium ecosystems (Fernandez-Gimenez & Allen-Diaz, 1999). These can be quantified by 4-dimensional analysis based on timelines of MODIS or NOAA data, as has already been done for central Asia (Lee et al., 2002; Kogan et al., 2004). Thus, a comprehensive assessment of the vegetation of the southern Mongolian Gobi depends on a combination of several data sources. Our project aims at eventually compiling a complex GIS database for all southern Mongolian nature reserves, which should be readily applicable to both nature conservation and wildlife management. Here we present initial results from the driest part of our working area, the Great Gobi A Strictly Protected Area (map 1), and additionally present an overview of all data which we will finally compile into our GIS. In terms of nature conservation, our primary focus is currently on protection of the Khulan (Equus hemionus hemionus); however, other applications are equally possible.
Methods
Vegetation sampling followed a modified Braun-Blanquet approach, using 281 plots as ground truth data (see Wesche et al., 2005 for details). All plots were 10 x 10 metres in size. Topsoil samples were taken from each plot and later analyzed for pH and conductivity. Basic environmental data collected in the field were supplemented by a public domain climate model (Hijmans et al., 2005), and by standard parameters derived from SRTM datasets (Rabus et al., 2003), e.g. altitude, slope and aspect. The vegetation data was digitalized using TABWIN. CA and DCA ordinations were used to evaluate relationships between vegetation data and environmental background information (Jongman et al., 1987)
Fig. 1. Overview of the protected areas mapped within the framework of our project. Рис. 1. расположение охраняемых заповедных территорий, исследуемых в рамках проекта.
Manual classification was aided by employing statistical fidelity measures (COCKTAIL, Bruelheide, 2000; implemented in the JUICE software package, Tichy, 2002). The large data sets were initially split into major groups using a cluster analysis (UPGMA based on Sorensen similarity). Plant communities were finally classified in accordance with those provided by W. Hilbig (1995; 2000), although some modifications were necessary (Wesche et al., 2005; von Wehrden et al., in prep; von Wehrden et al., 2006a). All 281 plots were assigned to a phytosociological group and then served as ground truth data for a supervised classification of the Landsat ETM+ raster datasets (von Wehrden et al., 2006b). Classifications were made using a Maximum Likelihood classifier, with a mean 7x7 pixel, nearest neighbour filter (Campbell, 1996). The results were evaluated using an independent dataset with -300 plots assessing covariance matrices, kappa statistics (Foody, 2002) and the overlap between neighbouring scenes. All GIS analyses were made using Arc Map 9.0, BLACKART, 3DEM and Diva-GIS.
Description of the mapped units. Descriptions of the vegetation of the Transaltay region including АРИДНЫЕ ЭКОСИСТЕМЫ, 2007, том 13, № 33-34
0 125 250
500
115'
Results and Discussion
the Great Gobi A Strictly Protected Area were provided by E.I. Rachkovskaya and E.A. Volkova (1977). Based on our own data and classifications, we compiled a description of plant communities and habitats, details of which are available in H. von Wehrden and K.Wesche (2006), and details on the plant community composition are given by H. von Wehrden et al. (2006a). We therefore present only a brief summary here.
As is typical for arid environments, plant species diversity is mainly driven by available rainfall (fig- 1).
precipitation in rrrrVa
Fig. 1. Regression model plotting the number of species per 100 m2 plot against the estimated mean annual precipitation in mm/a, based on 229 vegetation samples from the zonal vegetation. (r=0.4214: p<0.00001). Рис. 1. Регрессионная модель, связывающая осадки, мм/а и число видов на 100 м2, базирующаяся на 229 ключевых участках с зональной растительностью (г2=0.4214; р<0.00001).
The higher mountain sites and hills are covered by species-poor regional variants of communities belonging to the Caraganion and the Anabasietum sensu Hilbig (2000). Compared to other regions of the southern Mongolian Gobi (Wesche et al., 2005; von Wehrden et al., 2007), both vegetation cover and number of species are lower in the Transaltay Gobi. A few stands lacked both Anabasis brevifolia and Caragana leucophloea and were preliminarily mapped as a distinct unit, which is restricted to the higher forelands of the Atas Bogs and Tsagaan Bogd. It is clear that all of these three habitats contain a comparably high number of species per plot, and most stands are characterized by a stony or even rocky habitat, which is typical for the higher hills and mountain sites where they are found.
The pediments of these mountain ranges are often populated by an Ephedra przewalskii-Zygophyllum xanthoxylon community that mediates towards the lower semi-desert habitats, and which is clearly indicated by the central positions of the respective samples in the graph (fig. 2). Stands grow mainly along dry ravines, and the shrub-layer grows higher than in the previous communities. The presence of Reaumuria songarica hints at a high salt content in the soils.
Haloxylon ammodendron semi-deserts are common in the Transaltay region, and are mainly
found at the lower pediments. Although Saxaul can grow as a tree, we encountered mainly shrubs with average heights of around one metre. The relatively low habit may be due to grazing (Helmecke, Schamsran, 1979); however the almost complete lack of livestock in the study region implies that the availability of groundwater is a more important determinant. As in the previous unit, the shrubby Reaumuria songarica indicates again a higher salt content in the soils. At saline depressions mainly covered by a clayey soil matrix, mono-specific stands of this Tamaricaceae were sampled as well, but were mapped as a separate plant community. The shrub layer can be rather dense, but the amount of bare soil increases along the salinity gradient. The first axis indicates that the Haloxylon stands without Reaumuria songarica reach down to comparatively lower elevations than the Saxaul scrub with this Tamaricaceae.
The driest stands at the lower pediments and depressions are characterized by the drought-tolerant Chenopodiaceae Iljinia regelii, which may form dominant stands. The unique composition of these samples is evidenced by their position within the ordination (fig. 2). This group is rather heterogeneous, as shown by the wide scatter in the ordination, with some stands linking to the Anabasis brevifolia stands while others indicate affinities to the Haloxylon stands. The lowermost depressions of the Transaltay receive an average precipitation of below 40 mm/a, and rainfall may even only reach these areas in extraordinarily wet years. Not surprisingly, these sites are devoid of perennial plants. Table 1 gives an overview of the zonal plant communities.
Table 1. Overview of the zonal vegetation units and their running number, which are likewise used in figure 2 and map 2. Таблица 1. Зональные растительные сообщества и их порядковый номер, который используется на рисунке 3 и 4.
Community running number
Caragana 1
Stipa-Allium 2
Stipa-Anabasis 3
Ephedra-Zygophyllum 4
Haloxylon 5
Haloxylon-Reaumuria 6
Reaumuria 7
Iljinia 8
vegetationless 9
The heterogeneous extrazonal vegetation had to be excluded from the DCA (fig. 2), in order to derive a clearer pattern for the zonal vegetation. The oases of the Transaltay Gobi provide a range of different habitats characterized by rather luxurious growth, which is in stark contrast to the surrounding semi-deserts and deserts. Several stands of poplar forest were sampled, which were often mixed or bordered by Tamarix stands. At the innermost parts of some oases, extensive reed beds were found, and few salt meadows of very limited spatial extent were also sampled.
Supervised classification and distribution of plant communities. The Tasselled Cap transformation proved a most valuable tool for post-sampling enlargement of training areas, which was necessary to guarantee a sufficient classification basis. The Tasselled Cap compensates for the soil signal, which is inevitably mixed with the vegetation's reflectance. NDVI transformations were instead only helpful in the dry steppe communities (Schmidt, Karnieli, 2001). The final classification showed an overall accuracy of well above 90%, which is comparable with results from the Gobi Gurvan Saykhan National Park (von Wehrden et al., 2006b) and the Great Gobi B Strictly Protected
Area (von Wehrden, 2005).The accuracy was higher in the relatively homogenous semi-deserts (-94%), which have a lower vegetation cover. The Stipa glareosa stands found at higher elevations have a somewhat more small-scale patchy distribution, which may be the reason for the lower mapping accuracy for these habitats (-84%). The extrazonal salt-adapted vegetation showed the highest vegetation signal, and the highest accuracy related to the characteristic and readily distinguishable spectral signatures of these stands (-97%). The overall accuracy of the vegetation map is comparable to maps already compiled for the Great Gobi B Strictly Protected Area (=GGB SPA) and the Gobi Gurvan Saykhan National Park (=GGS NP). In the latter area, spectral differences within the high mountain ranges were clearer due to an overall higher number of relevés and the denser vegetation cover found there. Therefore 17 communities were mapped in the GGS SPA (Wesche el a/., 2005; von Wehrden et al., 2006b), and 12 in the GGB SPA (von Wehrden, 2005). With respect to the much drier conditions in the Great Gobi SPA (table 2), a number of 13 mapping units is sufficient for that region (von Wehrden, Wesche, 2007). Map 2 gives an example of a part of the final vegetation map.
SAMPLES
• Reaumuria(7)
О Halo-Reau (6)
о Halo(5)
□ Anabasis(3)
□ Stipa-Al!ium(2)
X Сагадапа(1 )
♦ Ephe-Zygo(4)
о lljinia(S)
о veg-less(9)
ENV. VARIABLES
-
Fig. 2. Detrended Correspondence Analysis of the zonal vegetation (cover values log-transformed, down-weighting of rare species; length of gradient: 1st axis 6.103, 2nd axis 4.905; Eigenvalues: 1st axis 0.809, 2n axis 0.480). The lowest circle indicates the units 5-7; the right circle includes units 1-3; the upper circle indicates units 8 and 9. Рис.2. Анализ соответствия с удаленным трендом зональной растительности (растительный покров log-трансформированный, оцененный через веса редких видов; линии градиентов: 1ые оси 6.103, 2ь1е оси 4.905); собственные значения: 1ая ось 0.809, 2ая ось 0.480). Самый нижний круг показывает единицы 5-7; правый круг - 1-3; верхний - 8 и 9.
Rainfall largely originates from two directions. Western disturbances cross the Turanic highlands and dominate the climate of the Dzungarian Gobi, whereas the Gobi-Altay is mainly influenced by the eastern monsoon. The Transaltay region, situated in between both circulation systems, receives the lowermost precipitation of all protected areas in the southern Mongolian Gobi.
Hence, dry desert steppes, which dominate in the Zuun Saykhan in the eastern Gobi Altay, become rarer towards the west and are completely missing in the Transaltay ranges. Nevertheless, the southern ranges within the Great Gobi В strictly protected area are covered by plant communities that are comparable to the drier mountains of the Gobi-Altay (von Wehrden, 2005; von Wehrden et al., 2007).
Table 2: Estimates of mean annual precipitation rates in mm/a (Hijmans et al., 2005) as derived for the five protected areas in southern Mongolia (""min" and "max"' refer to the driest and moistest sites in the given region). Таблица 2. Оценка значений годовых осадков в мм/а (Hijmans et al., 2005) как показателя различий пяти участков в Южной Монголии ("min'' and "max"' относятся к самому сухому и самому увлажненному участкам в районе исследований).
Area min max range mean std
Great Gobi "B" 69 177 108 96 17
Great Gobi "A" 33 124 91 54 11
Gobi Gurvan Saykhan 39 222 183 103 35
Small Gobi "A" 63 115 52 84 8
Small Gobi "B" 113 173 60 139 14
Fig. 4. Example of the vegetation map, showing the region west of the Tsagaan Bogd. Рис. 4. Пример карты растительности, показывающий район к западу от Тасаган Богдо.
The climatic differences regarding the origin of the rainfall are important for the distribution of the semi-desert and true desert communities as well. In the Transaltay, the lowest depressions are almost devoid of plants, save a few mostly annual Chenopodiaceae (von Wehrden et al., 2006a). In the other protected areas within southern Mongolia, similar depressions are covered by shrubby
semi-deserts or extrazonal salt-tolerant vegetation (Wesche et al., 2005; von Wehrden et al., in prep; Hilbig, Tungalag, 2006).
Outlook and implications for nature conservation. Table 3 summarizes all data sources currently used within our project's framework. In total, 1518 vegetation checks were collected covering all reserves in the Mongolian Gobi. A revised complete plant community classification system is currently compiled for the entire Mongolian Gobi (von Wehrden et al., in prep). Samples will serve as a ground reference basis for the classification processes. More than 2000 additional locations were sampled, where we recorded only coordinates and the important diagnostic species; these data will be used for independent accuracy assessments. Our vegetation survey aims at a scale of approximately 1:100 000, based on the resolution of Landsat TM & Landsat ETM.
Initial classifications are currently improved using a stratification based on SRTM-datasets derived for the whole of central Asia. Additional information, such as locations of wells, springs, streams, roads, settlements etc. are compiled into the GIS. In order to understand the climatic determinants of the mapped units, climate data modelled by Hijmans et al. (2005) is used. Multi-temporal analyses will be attempted using all Landsat generations available, while gradual NDVI timelines are derived from MODIS and NOAA data. Seasonality will be assessed using rainfall models on a coarse spatial resolution (~8 km) based on a two week interval.
Table 3. List of data sources used within our study. Таблица 3. Список источников данных, использованных в данной работе.
data source resolution application
Vegetation checks own data 10x10 metres ground truth data
Landsat MSS, TM & GLCF, USGS, DLR 79x79 m, baseline data for vegetation
ETM 30x30 m mapping
MODIS, NOAA GLCF 250 m-8 km NDVI-timelines
Meteosat GLCF 8 km rainfall estimates
Gtopo 30 GLCF ~1 km digital elevation model
SRTM GLCF -90 m digital elevation model
Climate data Berkeley, California ~1 km averaged climate data model
Animal positions Surveys and ARCOR Point-information habitat modelling
All these datasets are being used to understand the habitat use of the Mongolian Khulan. Migrations of these species are currently monitored using satellite-collars, which yields thousands of spatially explicit position records. The available data have already been used for a habitat correlation analysis for Przewalskii Horses and Khulans (Kaczensky et al., submitted). Several wild horses were already set free at a location in the western Great Gobi B Strictly Protected Area. Our vegetation maps indicated a suitable habitat comparable to the home ranges of already established groups, and this region was finally chosen for the reintroduction (see www.takhi.org). Other endangered species could also benefit from the improved knowledge on their habitats, including gazella (Milner-Gulland, Lhagvasuren, 1998), wild camel (Mix et al., 2002), argali (Reading et al., 1999; Schaller, 2000) and saiga (Milner-Gulland et al., 2001).
Acknowledgements
The beginnings of this project were made by G. & S. Miehe. Field work was aided by several people, namely K. Appel, M. Beckmann, A. Hilbig, E.J. Jäger, F. Riithrich, A. Tsolmon and
D. Walter. W. Hilbig also helped with the data analysis. The UNDP project "Conservation of the Great Gobi and its umbrella species" and the Protected Area bureau in Ulaanbataar provided logistical support, and our partners at the National University of Ulaanbataar assisted in organizing the fieldwork. The data source of the Global land cover facility is crucial for our work. H. Zimmermann and D. McCluskey proofread earlier versions of this manuscript and polished our English. Financial support was granted by the gtz, represented by S. Schmidt. The Takhi-project (funded by the FWF) supported work in the Great Gobi B strictly protected area. H. v. Wehrden was supported by the German Academic Exchange Service. Ongoing studies are financed by the FWF (project PI8624), in close cooperation with P. Kaczensky and C. Walzer.
We are grateful to members of the Russian-Mongolian complex biological expedition of the Russian Academy of Science and the Mongolian Academy of Science for laying the foundation for all plant ecological work in Mongolia.
LITERATURE CITED
1. Atlas of Mongolia. Ulaan Baatar. 1990. Pp. 140.
2. Batkhishig ()., Lehmkuhl F. Degradation und Desertifikation in der Mongolei // Petermanns Geographische Mitteilungen. 2003. No. 147. Pp. 48-49.
3. Bruelheide H A new measure of fidelity and its application to defining species groups // Journal of Vegetation Science. 2000. No. 11. Pp. 167-178.
4. Campbell J. B. Introduction to Remote Sensing. New York, 1996. Pp. 622.
5. Christensen L., Coughenour M.B., Ellis J.E., Chen Z.Z. Vulnerability of the Asian typical steppe to grazing and climate change // Climatic Change. 2004. No. 63. Pp. 351-368.
6. Cohen W.B., Goward S.N. Landsat's role in ecological applications of remote sensing // Bioscience. 2004. No. 54. Pp. 535-545.
7. Fernandez-Gimenez M.E., Allen-Diaz B. Testing a non-equilibrium model of rangeland vegetation dynamics in Mongolia // Journal of Applied Ecology. 1999. No. 36. Pp. 871-885.
8. Foody GM. Status of land cover classification accuracy assessment // Remote Sensing of Environment. 2002. No. 80. Pp. 185-201.
9. Gunin P.D., Vostokova E.A. Ecosystems of Mongolia. Moscow, 1995. Pp. 223.
10. Gunin P.D., Vostokova E.A., Dorofeyuk N.I. Vegetation dynamics of Mongolia. Boston: Dordrecht, 1999. Pp. 238.
11 .Helmecke K., Schamsran Z. Ergebnisse ökologischer Untersuchungen in der Gobi der Mongolischen Volksrepublik. 1. Untersuchungsgebiet, Vegetationseinheiten und Ergebnisse der mikroklimatischen Untersuchungen // Archiv fur Naturschutz und Landschaftsforschung. 1979. No. 19. Pp. 1-22.
12 Hijmans R.J., Cameron S.E., Parra J.L., Jones P.G., Jarvis A. Very high resolution interpolated climate surfaces for global land areas // International Journal of Climatology. 2005. No. 25. Pp. 1965-1978.
13. Hilbig W. The vegetation of Mongolia. Amsterdam, 1995. Pp. 258.
14. Hilbig W. Kommentierte Ubersicht über die Pflanzengesellschaften und ihre höheren Syntaxa in der Mongolei // Feddes Repert. 2000. No. 111. Pp. 75-120.
15.Hilbig W., TungalagR. Vegetationskundliche Untersuchungen in der Bordzongijn- und Gallbyn-Gobi (Omnogov Aimak, Mongolei) // Feddes Repertorium. 2006. No. 117. Pp. 399-429.
16 Jongman R.H.G., ter Braak C.J.G., van Tongeren O.F.R. Data analysis in community and
landscape ecology. Wageningen, 1987. Pp. 299. 17. KoganF., Stark R, GitelsonA., Jargalsaikhan L., Dugrajav C, Tsooj S. Derivation of pasture biomass in Mongolia from AVHRR-based vegetation health indices // International Journal
Remote Sensing. 2004. No. 25. Pp. 2889-2896.
18. Lee R, Yu F., Price K., Ellis ./., Shi P. Evaluating vegetation phenological patterns in Inner Mongolia using NDVI time-series analysis // International Journal of Remote Sensing. 2002. No. 23. Pp. 2505-2512.
19. Milner-Gulland E.J., Lhagvasuren B. Population dynamics of the Mongolian gazelle Procapra gutturosa: an historical analysis // Journal of Applied Ecology. 1998. No. 35. Pp. 240-251.
20. Milner-Gulland E.J., Kholodova M.V., Bekenov A., Bukreeva O.M., Grachev A., Amalan L., Lushchekina A. A. Dramatic declines in Saiga antelope populations // Oryx. 2001. 35 p.
21. Mix HM., Reading R.P., Blumer E.S., Badamjaviin L. Status and distribution of Wild Bactrian camels in Mongolia // Ecology and conservation of Wild Bactrian Camels. Eds.: Reading R., Dulamtserengiin E., Tuvdendorjiin G. Ulaanbaatar, 2002. Pp. 39-48.
22. Nagendra H. Using remote sensing to assess biodiversity // International Journal of Remote Sensing. 2001. No. 22. Pp. 1-12.
23. Rabus B., Eineder M., Roth A., Bamler R The shuttle radar topography mission - a new class of digital elevation models acquired by spaceborne radar. ISPRS II Journal of Photogrammetry and Remote Sensing. 2003. No. 57. Pp. 241-262.
24. Rachkovskaya E.I., Volkova E.A. Rastitelnost Zaaltayskoy Gobi // Biol. Res. prir. Uslov MNR 1977. Vol. 7. Pp. 46-74.
25. Reading RP., Amgalanbaatar S., Lhagvasuren L. Biological assessment of Three Beauties of the Gobi National Conservation Park, Mongolia // Biodiversity and Conservation. 1999. No. 8. Pp. 1115-1137.
26. Schaller G.B. Wildlife of the Tibetan Steppe. Chicago, 2000. 373 p.
27. Schmidt K, Karnieli A. Sensitivity of vegetation indices to substrate brightness in hyper-arid environment: the Makhtesh Ramon Crater (Israel) case study // International Journal of Remote Sensing. 2001. No. 22. Pp. 3503-3520.
28. Tichy L. JUICE, software for vegetation classification // Journal of Vegetation Science. 2002. No. 13. Pp. 451-453.
29. Von Wehrden H. Vegetation mapping in the Gobi Gurvan Saikhan national park and the Great Gobi B - a comparison of first results // Erforschung der biologischen Ressourcen der Mongolei. Halle: Martin-Luther-University, 2005. Vol. 9. Pp. 225-236.
30. Von Wehrden K, Wesche K. Mapping Khulan habitats - a GIS-based approach // Erforschung der bioogischen Ressourcen der Mongolei. Halle: Martin-Luther-University, 2007. Vol. 10. Pp. 31-44.
31. Von Wehrden K, Hilbig W., Wesche K. Plant communities of the Great Gobi A Special Protected Area in the Mongolian Transaltay Gobi // Feddes Repertorium. 2006 a. No. 117. Pp. 526-570.
32. Von Wehrden K, Wesche K., Reudenbach C., Miehe G. Mapping of large-scale vegetation pattern in southern Mongolian semi-deserts - an application of LANDS AT 7 data // Erdkunde. 2006 b. No. 60. Pp. 261-272.
33. Von Wehrden K, Wesche K., Tungalag R. Plant communities of the Great Gobi B // Mongolian Journal of Biological Sciences. 2007. No. 4. Pp. 3-18.
34. Wesche K., Miehe S., Miehe G. Plant communities of the Gobi Gurvan Sayhan National Park (South Gobi Aymak, Mongolia) // Candollea. 2005. No. 60. Pp. 149-205.
APHflHBIE 3K0CHCTEMBI, 2007, tom 13, № 33-34
КАРТОГРАФИРОВАНИЕ РАСТИТЕЛЬНОСТИ ОХРАНЯЕМЫХ ТЕРРИТОРИЙ ЮЖНОЙ МОНГОЛИИ: ОПЫТ ПРИМЕНЕНИЯ ДИСТАНЦИОННЫХ МЕТОДОВ И ГИС ТЕХНОЛОГИЙ
©2007 г. Г. фон Верден*'**, К. Веше**
* Университет Галле/Виттенберг, Германия, E-mail: henrikvonwehrden(a),web.de **Исследовательский Институт Естественной Экологии, Австрия, 1160 Вена, ул. Савоен, 1.
Охраняемые территории на юге Монгольской пустыни Гоби дают приют многим исчезающим животным и растениям. Эта территория представляет широкое разнообразие местообитаний от высокогорных пиков до низменностей, однако, современное использование и изменения климата ставит под угрозу экосистемы этого района.
Горные и экстразональные местообитания представляют собой пеструю мозаику сообществ, в то время как полупустынные и пустынные районы более однородны. Картографирование, основанное на дистанционных методах и ГИС технологиях - это альтернатива классическим методам исследований, которые могут быть слишком долговременными и трудоемкими, в особенности на обширных территориях. Ландсат дает хорошую основу для картографирования растительности, предлагая удобную пространственную шкалу для экологических интерпретаций, даже в засушливых местообитаниях и маловидовых сообществах.
Наш проект предлагает комплексную базу данных ГИС для всех природных резерватов южной Монголии, которую возможно будет применить как для природоохранных мероприятий, так и для управления естественными экосистемами. В настоящей работе мы даем результаты для наиболее засушливой части, строго охраняемой территории Большой Гоби, а также дополнительно представляем обзор всех данных, которые в конечном итоге войдут в нашу базу данных ГИС. Что касается охраны видов, мы сфокусировали наше внимание на кулане (Equus hemionus hemionus).
В общей сложности было обработано 1518 описаний растительности из всех охраняемых территорий Монгольской пустыни Гоби. Переработанная и дополненная классификация растительности представлена нами для всей территории Монгольской пустыни Гоби. Более 2000 дополнительных местообитаний было закартировано, в которых мы отмечали только координаты и важные диагностические виды, эта информация послужит для независимых оценок точности. Наши исследования растительности производились примерно в масштабе 1:100 000, основанном на разрешении Ландсат ТМ и Ландсат ЕТМ.
Все данные из базы использовались нами для понимания особенностей местообитания монгольского кулана. Миграции этого вида отслеживаются в настоящее время благодаря использованию спутниковых маяков-ошейников, которые дают сотни данных о пространственном распространении вида. Полученные данные уже были использованы для корреляционного анализа местообитаний лошади Пржевальского и куланов. Несколько лошадей уже были выпущены на волю на территории Б строго охраняемой зоны Большой Гоби. Наша карта растительности показала, что именно здесь существуют подходящие условия для реинтродукции этого вида. Для других исчезающих видов также будут полезны наши данные о местообитаниях, для таких как газели, дикий верблюд, аргали и сайгак.
