CC BY
26
АРИДНЫЕ ЭКОСИСТЕМЫ, 2005, том 11, №26-27
============================= ДОКЛАДЫ ====================================
ОТНОСИТЕЛЬНОЕ ИЗМЕНЕНИЕ ОКРУЖАЮЩЕЙ СРЕДЫ В АРИДНЫХ ЗОНАХ ЦЕНТРАЛЬНОЙ АЗИИ ПО ДАННЫМ НАБЛЮДЕНИЙ ЗА СОСТОЯНИЕМ ЗЕМЛИ И НАБЛЮДЕНИЮ ПОЧВ
©2005 г. Е. Любимцева
Отдел географии и планирования, Государственный Университет Грэнд-Вэлли 1150 ОуСейбл Холл, 1 Кампус драйк Оллендэил, MI49401, США
Климатические изменения свидетельствуют о том, что окружающая среда аридных зон может подвергнуться радикальным изменениям в результате глобального потепления и растущей концентрации углекислоты в атмосфере.
Анализы временных и пространственных колебаний N0"^, полученных NOAA метеорологическими спутниками с 1981 до 2001 гг. вместе с температурой и данными выпадания осадков метеорологических станций показали огромные изменения в ландшафтном покрове Центральной Азии в течение последних двух десятилетий. N0"! является необходимым индикатором для определения растительного покрова, производства зеленой биомассы и поэтому, тесно связана с климатическими факторами. Изменчивость этого показателя взаимосвязана с изменением осадков и температуры в этом средне-опустыненном регионе. Согласно нашим предварительным данным, в центре аридных земель наблюдалась тенденция расширения растительного покрова в период 1982-1996 гг., за исключением территории окрестностей Аральского моря, сменившаяся обратной тенденцией в период 1996-2001 гг. С начала 1980-х годов более чем 2/з территории аридных зон в Центральной Азии, подверглись расширению зеленого покрова приблизительно на 10%. Эти изменения, происходящие в растительном покрове связаны с изменением количества осадков и, вероятно, климатическими различиями, связанными с увеличением растительности и биогенетических слоев вследствии накопления в атмосфере СО2. Большая неопределенность существует в изменчивости наблюдаемой N0""! Указанный факт является ответной реакцией флоры и фауны на глобальные изменения климата, интенсивного сельского хозяйства с применением орошения, оптимизацией выпаса скота.
ENVIRONMENTAL CHANGES IN ARID CENTRAL ASIA INFERRED FROM REMOTE SENSING DATA AND GROUND OBSERVATIONS.
© 2005. E. Lioubimtseva
Geography and Planning Department, Grand Valley State University 1150 Аи Sable Hall, 1 Campus Drike Allendale, MI 49401, USA Lioubime@gvsu.edu
Introduction
Meteorological records indicate that arid and semi-arid zones of Central have experienced a significant warming signal since the last decades (Neronov, 1997; Chub, 2000; Lioubimtseva, 2004). Regional responses to the global warming trend include melting of the high-mountain glaciers in the Tianshan and Pamir, dramatic lake-level fluctuations, changes in the river' discharge, species migration, and numerous changes in ecosystem dynamics and land-use. The arid Central Asia region is very vulnerable to human disturbances and regional climate changes, because ecosystems in the region may be the first to reach tipping points under current human disturbances and climate change (Lioubimtseva and Adams, 2004).
To fully understand the impact of human activities, it is also necessary to consider the extent to which anthropogenic effects have modified the background level of carbon storage, and whether change in the intensity of either process has any evident potential to take up or release carbon from the desert-zone carbon reservoir. There is significant uncertainty regarding the possible impacts of global climate change on the sequestration of carbon in the vegetation and soils of the arid zones in general, and in those of Central Asia in particular. It is possible that global climate change could result in significant changes in carbon reservoirs in these areas. Estimates of the carbon pools in the desert soils are still very uncertain (Lioubimtseva and Adams, 2002; Lioubimtseva et al, 2005).
Current and predicted climate trends
The Central Asian arid region comprises the Turan Lowland and the southern margin of the Kazakh Hills (Figure 1) and is bounded by the Middle Asian mountains on its southern and southeastern edges.
Fig. 1. The study region and its climate
In the southwest the somewhat lower mountains of the Kopet Dagh allow monsoon precipitation to reach the western slopes of the Tian Shan and Pamir-Alai ranges. To the north the Turanian plain descends progressively northward and westward and opens out towards the Caspian lowland. The northern boundary of this vast arid zone is rather poorly defined but it lies at approximately 48 DN. Continental climate of this region implies that summers are hot, cloudless and dry, and winters are moist and relatively warm in the south and cold with severe frosts in the north. Precipitation has a distinctive
spring maximum, which is associated with the northward migration of the Iranian branch of the Polar front. Most frequently rain is brought by the depressions which develop over the Eastern Mediterranean, migrate north-eastwards, and regenerate over the Caspian Sea.
The IPCC report Regional Impacts of Climate Change, 2001 addresses Central Asian republics in Chapter 7 "Middle East and Arid Asia" but provides very limited information about climate change in arid Central Asia (IPCC 2001). "There were no discernible trends in annual precipitation during 190095 for the region as a whole, nor in most parts of this region" (IPCC 2001, Chapter 7). The aridity index shows no consistent trends for Central Asia as a whole (IPCC, 2001). The report does point out a likely 1-2 degree C/century temperature increase for Central Asia.
Meteorological data series show a steady increase of annual and winter temperatures in this region since the beginning of the past century (Neronov, 1997, Chub, 2000; Lioubimtseva et al., 2005). Unfortunately only a few stations in central Asia have a period of observations spanning more than a century. Most stations have records for a relatively short time, roughly 50-60 years. In addition, many meteorological stations in the region practically stopped functioning after the collapse of the USSR as a result of severe funding cuts (Chub, 2000).
Fig. 2. Temperature increase from 1900 to 2000
a. Toshkent station (adapted from Chub, 2000)
b. temperature anomaly for 1880 to 2000 for arid zones of Central Asia (based on GHCN dataset, Peterson & Vose, 1997).
Climate models predict that the temperature in arid Central Asia will increase by 1-2° C by 20302050, with the greatest increase in winter. Precipitation projections vary from one model to another and projected changes in the aridity index for different model runs show no consistent trend for this region (Figure 3). Some models project greater aridity in the future and some predict less; it is becoming increasingly apparent that climate change modelling in arid zones is extremely uncertain, partly because of the extreme natural variability (both temporal and spatial) of the desert climate and partly because of inherent uncertainties in global and regional climate modelling (Lioubimtseva and Adams, 2004). Atmospheric dynamics are known to be very sensitive to natural climate variability at relatively short time-scales and the effect of short-time variability on longer (decadal-to-millennial) time-scales are not fully understood.
Both theoretical considerations and numerical models have shown a significant sensitivity of the climate of arid regions to vegetation distribution. Ground-cover parameters can significantly alter the modelled climate (Zolotokrylin, 2002; Wang and Eltahir, 2000).
LG
Ъ. 1
и
е
В
ШИП ХШ МИ Ж ЙК> 2№t SJ» W№ 30J0
Ym
HftDCM2 esfnewts | Г GCHMM "1 1 CCGM1
им 2CSO зоаа X H j к Г:'. 53 202D 2050 голо 332D 2« iC
1
Fig. 3. GCM scenarios for arid zones of Central Asia, 2020 to 2080 (from Lioubimtseva et al., 2005).
a. Temperature range (degrees C).
b. Precipitation range (mm per day).
Climate change scenarios, unfortunately, do not incorporate regional controls on climate. The impacts of the extensive redirection of montane and lacustrine water resources to irrigated agriculture in Central Asia and the degradation of the Aral Sea remain unmeasured in current climate models yet may be of significant importance in regional climate change. Although it is clear that both observed and predicted climate changes might be partly caused by global climate change and partly by local anthropogenic processes it is extremely difficult, if not impossible to delineate the boundary between these two factors.
Remote sensing approach
In the regions with limited amount of ground observations long temporal series of remote sensing data offer a very useful and often the only possible approach to climate change. Such monitoring techniques are based on the estimation of the statistical relationship between climate aridity and vegetation phytomass estimations from satellite imagery (Kogan, 1995; Lambin, 1997; Nicholson et al., 1998). Most of remote sensing research on arid lands has been focused on climate change in the Sudan-Sahelian zone of Africa and the Western USA. Zolotokrylin (2002) developed an new empirical aridity index for Central Asian deserts and semideserts, which can be defined as a duration of the period with a normalised difference vegetation index (NDVI) less than 0.07. This indicator reflects the zonality of heat exchange between arid land and atmosphere as the relation of radiation and evapotranspiration mechanisms in the regulation of thermal conditions of soil surface and the lower layer of atmosphere (Zolotokrylin, 2002) .
Methodology
This study is based on the analysis of satellite imagery from the Pathfinder Advanced Very High Resolution Radiometer Land dataset. Parameters produced as a part of this dataset include reflectances and brightness temperatures derived from the five-channel cross-track scanning AVHRR aboard the NOAA Polar Orbiter 'afternoon' satellites (NOAA-07: Jul 81 to Jan 85, NOAA-09: Feb 85 to Oct 88, NOAA-11: Nov 88 to Sep 94, and NOAA-14: Jan 85 to Oct 01), along with a derived Normalized Difference Vegetation Index (NDVI), cloud and quality control indicators, and ancillary data. These data are derived from the NOAA Global Area Coverage (GAC) Level 1B data spanning a period of more than 20-years (1981-2001).
Pigments in green leaves (notably chlorophyll) absorb strongly at red and blue wavelengths. Lack of such absorption at near-infrared wavelengths results in strong scatter from leaves. The contrast between red and near-infrared reflectance of vegetation is captured by NDVI, a commonly used greenness index (near infrared-RED)/(near infrared + RED), which is often used as a proxy for biomass, net primary productivity, and leaf area index (Tucker, 1979; Kogan, 1995; Qi et al., 2000). Although NDVI poses some serious problems and more advanced indices and NDVI modifications have been developed during the past decade, NDVI still represent a very useful tool for landcover monitoring since
it is robust, easily available from the AVHRR and other sensors, can be easily ground-truthed and demonstrates good correlation with the amount of photosynthesizing vegetation. Advantages and disadvantages of NDVI have been thoroughly discussed in the international remote sensing literature during the past decade (Qi et al., 2000; De Beurs and Henebry, 2004).
Six hundred ninety eight (698) 10 -day maximum value composites available since July 10, 1981 to September 30 2001, gridded at a resolution 8 km by 8 km were used to establish the NDVI temporal trends in arid central Asia. There is a permanent data gap in 1994 from the middle of September to the end of December due to satellite failure. September 30, 2001 is currently the last day of data available.
Simple statistical analyses of these data series involved 1) review of temporal variations revealed in the 10-day and monthly NDVI composites with particular focus on annual trends in spring NDVI (associated with precipitation peaks), 2) analyses of statistical relationships between NDVI and precipitation for four key area (Eastern Kara Kum, Plateau Usturt; Central Kyzyl Kum, and an area around and eastward from the Aral Sea), and 3)computation and analyses of the spatial and temporal patterns of an empirical aridity index, derived from the NDVI.
Results and discussion
Smoothed NDVI series revealed a substantial increase in NDVI between 1986 and 1994 with a very prominent peak in 1993-1994, followed by a slight decline. Interestingly, this trend is well observable in all parts of Central Asia, despite the great variability the local landscape and meteorological conditions (Figure 4). The 1993-1994 greenness peak has been reported also in other vegetation indices- based studies in the adjacent arid and semi-arid regions of Eurasia (DeBeurs and Henebry, 2004; DeBeurs et al., this issue; Bajargardal and Karnieli, this issue). The reason of this peak, however, is still unclear. While many studies attribute this increase in NDVI and related indices to higher vegetation biomass due to higher levels of precipitation in early 1990s (Bayarjargal and Karnieli, 2004; Zolotokrylin, 2002; Kharin et al., 1998), others suggest that the most plausible explanation of higher NDVI might be an artifact caused by miscalibration of the NOAA-11 imagery (DeBeurs and Henebry, 2004).
Our study revealed fairly strong correlation between NDVI and precipitation for most of the region in the 1980s and 1990s that is in a good agreement with the earlier NOAA-based studies in Uzbekistan and Kazakhstan (Kharin et al., 1998; Zolotokrylin, 2002).
Based on the assumption of the steady relationship between NDVI and precipitation a simple empirical index of aridity derived from NDVI was used in order to address the spatial and temporal climate variability:
A= NNDVI<a /T, where N is number of 10-day intervals with NDVI less than a threshold value a (climate-related variable that was set to 0.07 for arid Central Asia) and T is a number of 10-day intervals between April 1 and October 31 during the period of observation. This index of aridity, computed both for the 1980s and 1990s reveals a decrease of aridity in most of Central Asian region except for the Aral Sea area (table 1). The increase of the number of 10-day intervals with arid NDVI in a vicinity of the Aral Sea is apparently caused by the severe human-induced desertification of this area and is well confirmed by the ground observations.
Table 1 Changes in the NDVI-based aridity index
Years/location East Karakum Usturt KyzylKum East Aral
1981-1990 0.033854 0.035024 0.0294б3 0.023627
1991-2000 0.028701 0.025141 0.038037 0.044112
One must keep in mind, however, that compare to other biomes, desert and semi-desert landscapes are featured by a very sparse vegetation cover and that more than 65% of the surface reflectance signal in this very coarse-resolution pixel data is coming not from vegetation but from the soil. If one assumes that the barren soil signal is fairly constant and the data misregistration from different satellites is probably not significant to cause this trend, we should assume that the revealed NDVI difference should be attributed to vegetation changes. On the other hand, one should also ask a question, how barren is barren soil in the desert, or in other words, what is the proportion of the observed NDVI signal coming
from the microphytic communities (mosses, fungi, algae and cyanobacteria) on the soil surface compare to the signal of vegetation?
Microphytic communities form biogenic crusts on the soil surface varying from a few millimetres to several centimetres in thickness and play a significant role in the desert ecosystems controlling such processes as water retention and carbon and nitrogen fixation in soils. In the Kara Kum desert of Turkmenistan the accelerated growth of such biogenic crusts has been observed during the past 40-50 years and usually has been attributed to the undergrazing caused by the decrease of the wild fauna and insufficient pressure on the desert rangelands. It is possible, however, that the main reason for the accelerated growth of "black mosses" in the Kara Kum is a response of microphytes to increasing concentrations of CO2 in the atmosphere (Lioubimtseva et al., 2005). Remote sensing data series including the blue channel of electromagnetic spectrum (such as Landsat TM and VEGEATION-SPOT), which is absent in AVHRR NOAA, might provide more insights on delineation of the signal produced by the biogenic crusts from the one of the higher vegetation (Karnieli et al., 1999). Unfortunately VEGETATION imagery of the spatial resoution comparable to that of AVHRR is available only since 1998.
The CO2 fertilisation effects not only microphytic communities but also higher vegetation. An increased atmospheric CO2 concentration has direct and relatively immediate effects on two important physiological processes in plants: it increases the photosynthetic rate, but decreases stomatal opening and therefore the rate at which plants lose water. The combination of these two factors, increased photosynthesis and decreased water loss, implies a significant increase of water efficiency (the ratio of carbon gain per unit water loss) and productivity and a reduction in the sensitivity to drought stress in desert vegetation as a result of elevated atmospheric CO2 (Smith et al., 2000).
No regional modelling studies have been conducted in the Central Asian region but global biogeography models (Melillo et al., 1993; Woodward et al., 1998) predict relatively strong responses of arid ecosystems to global climatic change.
The Kara Kum and Kyzyl Kum deserts of Central Asia are dominated by vegetation with the C3 photosynthetic pathway with only few C4 and CAM species. It is often expected that plants using the C3 photosynthetic pathway will respond more strongly to raised CO2 than species with the more water-efficient and CO2-efficient C4 photosynthetic system. The significance of different photosynthetic pathways in the adaptation of perennial plants to life in extreme desert environments is still hotly debated (Graybill and Idso, 1993; Whitford, 2002; Grunzweig, and Korner, 2000). Most publications on this subject are based on chamber experiments and the recent Free-Air CO2 Enrichment experiments studying responses of desert vegetation to increased CO2 levels conducted in the south-western USA where desert vegetation cover is dominated by C4 species (Smith, 2000; Huxman et al., 2000).
Conclusions
The preliminary results of this study suggest that there have been a greening trend in arid central between 1982 and 1995 with a prominent peak in 1994, followed by the opposite trend between 1996 and 2001. Since the early 1980s more than 70% of arid zones of Central Asia have become greener by about 10 %. These vegetation changes are due to precipitation changes and probably, to easing of climatic constraints to growth of plants and biogenic crusts caused by CO2 increase in the atmosphere.
The complexities of precipitation changes, vegetation-climate feedbacks, and direct physiological effects of CO2 on vegetation present particular challenges for understanding and modelling climate change in temperate arid regions. Great uncertainties exist in the prediction of responses of arid landscapes of Central Asia to elevated CO2, as well as to global and regional, natural and human-induced climate change.
There has been a general warming trend in Central Asian republics on the order of 1-2 degrees C since the beginning of the 20th century that might have a strong potential impact on the regional temperature and precipitation regimes and also on natural ecosystems, agricultural crops and human health.
Climate change projections in this region vary from one global model to another. Despite the great progress in global climate modelling, the GCMs give very variable results, with large spatial differences
in the areas forecast to receive higher or lower precipitation. The lack of integration of such factors as dust aerosols, biophysical and biochemical feedbacks caused by land-cover changes, as well as the regional factors of human-induced climate change, such as irrigation, are the major sources of uncertainties. For example, the dust aerosols from the drying Aral Sea bottom might have a very significant impact on regional climate but they are not taken into account by the models.
Projections based on biogeographic models suggest considerable changes in desert and semi-desert vegetation due to a combination of greenhouse-related climate change and direct physiological CO2 effects on vegetation, such as changes in photosynthesis and water-use-efficiency over the coming century. This could likewise have implications for crop growth in desert-marginal areas, favouring greater productivity, and perhaps increase productivity and biomass of natural desert vegetation and soil organic matter. However, the very limited number of CO2 enrichment experiments in the Kara Kum and Kyzyl Kum do not always confirm this thesis. Moreover, the results of CO2 fertilisation experiments in other arid regions of the world, such as the Negev desert in Israel (Grünzweig and Körner, 2000) and Mojave desert in the US (Smith et al., 2000) are rather mixed - which only contributes to the uncertainty about the implications of the doubled CO2 concentrations for desert ecosystems. The accelerated growth of biogenic crusts (observed during the past 40-50 years) in this arid region might be a response of microphytes to increasing concentrations of CO2 in the atmosphere. However, because such responses occur only on the undergrazed desert rangelands it is still unclear if the CO2 increase really is the major cause or such growth, rather than land-use change.
One of the major sources of uncertainty about vulnerability and impacts of climate and land-cover changes in arid lands of Central Asia is the lack of reliable and accurate data on climate and ecosystems necessary for regional climatic and biogeographic modelling. While the local responses to global climate change have been a source of major uncertainties, it is clear that there have been very intensive human-induced regional climate and environmental changes in Central Asia, primarily associated with massive irrigation schemes, the desertification crisis in the Aral Sea area, and changes in the grazing pressure on desert rangelands. Local and regional human impacts in arid zones can significantly modify surface albedo, as well as water exchange and nutrient cycles that could potentially have impacts on the climatic system both at the regional and global scales. On the other hand, improved management techniques can increase the carbon sequestration capacity of semi-desert rangelands and arable lands. Acknowledgements:
The GCM scenarios were obtained from the IPCC Data Distribution Centre [http://ipcc-ddc.cru.uea.ac.uk, last time accessed September 2004]. The author would like to thank Philip Micklin and Roy Cole for their comments on the earlier versions of this paper, as well as Kin Ma for the editorial help.
References
1. Bayarjargal Y.and A. Karnieli, 2004. Remote sensing change detection of land degradation/rehabilitation in Mongolia's desert-steppe. The first international workshop on Human Dimensions of Climate and Environmental Changes in Central Asia, Grand Valley State University, Grand Rapids, MI, USA
2. Chub, V.E., 2000. Climate change and its impact on the natural resources potential of the Republic of Uzbekistan. (Izmenenija klimata i ego vlijanije na prirodno-resursnij potensial respubliki Uzbekistan), Gimet, Tashkent, 253 p. (in Russian).
3. De Beurs, K.M., and G.M. Henebry. 2004. Land surface phenology, climatic variation, and institutional change: Analyzing agricultural land cover change in Kazakhstan. Remote Sensing of Environment, 89:497-509.
4. Grünzweig, J. and Ch. Körner, 2000. Growth and reproductive responses to elevated CO2 in wild cereals of the northern Negev of Israel. Global Change Biology 6:631-638.
5. Graybill, D.A. and Idso, S.B. 1993. Detecting the aerial fertilization effect of atmospheric CO2 enrichment in tree-ring chronologies. Global Biogeochemical Cycles 7: 81-95.
6. Huxman, T.E., Nowak, R.S., Redar, S., Loik, M.E., Jordan, D.N., Zitzer, S.F., Coleman, J.S., Seeman, J.R. and Smith, S.D. 2000: Photosynthetic responses of Larrea tridentata to a step-increase
in atmospheric CO2 at the Nevada Desert FACE Facility. Journal of Arid Environments, 44: 425436.
7. IPCC, 2001. Climate Change: The Scientific Basis Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) J. T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P. J. van der Linden and D. Xiaosu (Eds.) Cambridge University Press, UK. 944 pp.
8. IPCC Data Distribution Centre 2004. http://ipcc-ddc.cru.uea.ac.uk, last time accessed September 2004.
9. Karnieli, A., Kidron, G.J., Glaesser, C. and Ben-Dor, E. 1999. Spectral characteristics of cyanobacteria soil crust in semiarid environment. Remote Sensing of Environment, 69, pp 67-75.
10. Kharin N.G., R. Tateishi, I.G. Gringof, (1998). Use of NOAA AVHRR data for assessment of precipitation and land degradation in Central Asia. Arid Ecosystems, 4 (8): 25-34.
11.Kogan, F. N., 1995. Application of vegetation index and brightness temperature for drought detection. Advances in Space Research 15: 91-100.
12. Lambin, E., 1997. Modelling and monitoring land-cover change processes in tropical regions, Progress in Physical Geography, 21 (3): 375-393.
13. Lioubimtseva, E. 2004. Climate Change in Arid Environments: revisiting the past to understand the future. Progress in Physical Geography, 28 (4): 1-25
14. Lioubimtseva, E. and J.M. Adams, 2002. Carbon Content in Desert and Semidesert Soils in Central Asia. In: J.M. Kimble, R. Lal, and R.F. Follett (eds.) pp.409-256, Agricultural Practices and Policies for Carbon Sequestration in Soil. CRC Press Lewis Publishers, Boca Raton, London, New York, Washington DC, 512 pp.
15. Lioubimtseva, E. and J.M. Adams, 2004. Possible implications of increased carbon dioxide levels and climate change for desert ecosystems. Environmental Management, 33 (S1): S388-S404
16. Lioubimtseva, E., R. Cole, J. M. Adams and G. Kapustin, 2005. Impacts of Climate and Land-Cover Changes in Arid Lands of Central Asia. Journal of Arid Environments, in press.
17. Melillo, J.M., A.D. McGuire, D.W. Kicklighter, B. Moore III, C.J. Vorosmarty, and A.L. Schloss, 1993. Global climatic change and terrestrial net primary production. Nature, 363: 234-240.
18. Neronov V.V., 1997. Landshaftnije osobennosti vekovih izmenenij uvlaznennosti v yugo-vostochnom Turkmenistane (Landscape features of century changes in wetting of south-eastern Turkmenistan), Arid Ecosystems, 3(6-7): 141-150 (in Russian with summary in English).
19. Nicholson S.E., M.B. Ba, and Tucker C.J., 1998. Desertification, Drought and Surface Vegetation: An example from the West African Sahel. Bulliten of the American Meteorolgical Society, 79, 816829.
20. Peterson, T.C. and Vose, R.S.,1997: An overview of the Global Historical Climatology Network temperature data base. Bulletin of the American Meteorological Society, 78: 2837-2849.
21. Qi, J., Kerr, Y., Moran, M., Weltz, M., Huete, A., Sorooshian, S., and Bryant, R., 2000. Leaf area index estimates using remotely sensed data and BRDF models in a semi-arid region, Remote Sens. Environ. 73: 18-30.
22.
23. Smith, S.D., T.E. Huxman, S.F. Zitzer, T.N. Charlet, D C. Housman, J.S. Coleman, L.K. Fenstermaker, J.R. Seemann, and R.S. Nowak, 2000. Elevated CO2 increases productivity and invasive species success in an arid ecosystem. Nature 408: 79-82.
24. Tucker, C. J.,1979 Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environments 8:127-150.
25. Wang G. and E.A.B. Eltahir, 2000. Biosphere-atmosphere interactions over West Africa. II: Multiple climate equilibria. Q.J.R. Meteorol.Soc., 126:1261-1280.
26. Whitford, W., 2002. Ecology of Desert Systems. Academic Press, London, 343 pp.
27. Woodward, F.I., M.R. Lomas and R.A. Betts, 1998. Vegetation-climate feedbacks in a greenhouse world. Philosophical Transactions of the Royal Society of London, Series B, 353, 29-38.
28. Zolotokrylin, A.N., 2002. The indicator of climate aridity (Indicator klimnaticheskoj aridnnosti). Arid Ecosystems 8 (16): 49-57 (in Russian with summary in English).
