Vegetation cover dynamics in arid lands of western China from 1982 to 2000 Текст научной статьи по специальности «Науки о Земле и смежные экологические науки»

АРИДНЫЕ ЭКОСИСТЕМЫ, 2005, том 11, №26-27

================================ ДОКЛАДЫ ==================================

ДИНАМИКА РАСТИТЕЛЬНОГО ПОКРОВА В АРИДНЫХ ЗЕМЛЯХ ЗАПАДНОГО КИТАЯ С 1982 ПО 2000

© 2005 г. Куингдонг Ши1, Джиагво Ки12, Ксиаолинг Пэн1, Гуангуи Льви1, Кингсан Ши1, Хейян Лу1

1 Лаборатория экологии Оазиса Ксинжианг Ки, Международный центр Устойчивого развития в аридных и семиаридных землях,

Урумки, Ксинжианг, Китай 2Центр исследований глобальных изменений климата и земли, Географический отдел, Мичиганский

Государственный Университет, Ист Лензинг, Мичиган, США

Доминирующее влияние на состояние и динамику растительного покрова в рассматриваемых регионах оказывают антропогенные факторы. Изменение растительного покрова в аридных регионах западного Китая было проанализировано в период от 1982 до 2000 гг. Данные о фракционном составе растительного покрова (ФРП) были получены NOAA/AVHRR в разное время, методом воздушного деления, (Shi, 2003), который был использован для анализа изменений в фракциональном покрове в данный период. Были установлены движущие факторы изменения растительности в аридных территориях Западного Китая, путем анализа температурных изменений для каждого типа растительных сообществ в различных регионах. Таким путем были установлены характерные различия растительности в Аридных территориях Западного Китая. Было выявлено, что между различными регионами в пределах Западного Китая существуют характерные различия. В Северном Ксинжианге общая территория опустынивания уменьшилась, в то время, как площадь лесов увеличилась за последние два десятилетия, результатом чего стало улучшение растительного покрова прилегающих территорий. Однако на высоких равнинах Квайдам, общая территория пустыни увеличилась на границе исчезающих лугов, что в результате привело к уменьшению продуктивности всего растительного покрова.

На равнине, общая территория пустыни не изменилась, в то время как оазис продолжает расширяться через Ксинжианг, в результате чего сокращаются экотоны перехода к средней и высокой продуктивности травяного покрова. Выяснилось, что человеческая деятельность является одним из важнейших факторов, влияющих на изменение типа растительности в аридных территориях Ксинжианга, хотя ученые говорят об изменении регионального климата в этой части Китая.

VEGETATION COVER DYNAMICS IN THE ARID LANDS OF WESTERN CHINA FROM

1982 TO 2000

© 2005. Qingdong Shi1, Jiaguo Qi1,2, Xiaoling Pan1, Guanghui Lv1, Qingsan Shi1, Haiyan Lu1

1Xinjiang Key Laboratory of Oasis Ecology; International Center for Sustainable Development in Arid

and Semi-arid Land, Urumqi, Xinjiang, China 2Center for Global Change and Earth Observations; Dept. of Geography, Michigan State University,

East Lansing, Michigan, USA

Introduction

The amount of data and remote sensing technologies provide an unmatched superiority to study the vegetation cover change over large or middle scale areas (Zhang, 2001). A few large area studies of vegetation cover dynamics utilizing remote sensing data over a long time period have been made in the arid western China. However, most recent studies in this region focused primarily at local scales (Wang et al., 1998, Liu et al., 2001, Meng et al., 2003, Zhao et al., 2004). Moreover, these studies (Ma et al., 2003, Yan et al., 2003, and Piao et al., 2001) on vegetation change in the arid western China were less

quantitative as these studies focused primarily on the spectral properties of this region. There were limited number of quantitative studies but their study areas were primarily in desert oasis ecotone (middle and high grassland cover). The oasis ecotone is the core area of human productivity and activity in the arid lands.

Studies on vegetation changes over time focus on the spatial dynamics as well as on the total areal changes of different vegetation types. Changes in areas over space and time directly are a collective result of many causes, ranging from natural climate extreme events to human activities. At local scale, the spatial pattern is often not enough to represent changes at large scales. Only when we look at changes at a regional scale can we understand vegetation changes and their tendencies of an entire ecosystem. Continual expansion of the desert areas indicate that the ecosystem's environment has been degraded, while shrinking of desert would suggest an improvement of the ecosystem environment. Therefore in western China, it is necessary to quantitatively analyze the spatial and temporal patterns of different vegetation types over a period of time. In this study, we first derived fractional vegetation cover (FVC) images from NOAA/AVHRR time series data from 1982-2000 and used the areal division method (Shi, 2003) to study the spatial and temporal patterns of land use and land cover change. The temporal changes in fractional cover for different regions within the western China were analyzed to arrive at the patterns and drivers of these changes.

Study Area The northern boundary of the study area follows the Altai watershed (Appendix 1, Fig.1). The Yin Mountains along the border between China and Mongolia bounds the study area on the east. The southern boundary ranges from south of the Tianshan Mountains to the Helan Mountains in the Kunlun Mountains. The western part of China is characterized by mountains (Tianshan, Kunlun, and Altai) and arid basins. The precipitation ranges from 20mm in the arid basins to 250mm in the mountain regions, while the annual temperature ranges from 0 to 50oC. There is a general gradient in precipitation along the longitude ranging from 200-250mm in the eastern part of the study area to 20-50mm in the western part. There is also a latitudinal gradient at the elevation of 300m-800m that ranges from 300500mm in Jungger Basin in the north to 100mm-1500m in Taripan Basin in the south. Most of rainfall events occur in the summer time and are so spatially heterogeneous that some areas receive no rains in years.

Vegetation distribution is primarily a function of elevation. On the Altay Mountains (north of Part I in Appendix 1, Fig. 1), the desert areas are distributed below 800m in elevation. On the Tianshan Mountains there is a markedly difference in vegetation types between north and south slopes. The north slopes (shaded slope) are primarily needleleaf forests while on the south slopes are steppes. The deserts are located below an altitude of 1700m on the north slopes and above 2400m on the south slope. On the south Xinjiang (Part III) and the north slopes of Kunlun Mountains and Alijin, the deserts are located primarily below an elevation of 3200m (Academic Research Xinjiang Term of Chinese Academy of Science & Institute of Botany of Chinese Academy of Sciences, 1978).

The major human land uses include rangelands, deserts, forests, oasis, and urban. There is a narrow band of oases around the mountains between forest/snowcaps and deserts, fed primarily by the seasonal snow melting from mountains. The major agricultural activities include grazing (70%) and farming (30%). This region is rich in petroleum from the basins and rich in minerals from the mountains. The major economy is petroleum production and livestock. This region is also a major tourist attraction as it has preserved its natural beauty due to sparse population and limited economic developments in China. With the implementation of the Grand West Development Strategy by the Chinese central government since the later 1990s, this region has been identified as one of the most desirable areas for near future development. In recent years, the area has seen increased population and urbanization as investors come in for petroleum and livestock/dairy productions from the coastal region.

Data Processing and Classification 3 In order to analyze the changes in vegetation dynamics in the past 20 years, a time 4 series of fractional vegetation cover (FVC) for the period of 1982-2000 was constructed 5 using the AVHRR data (8km spatial resolution) from the University of Arizona (Zeng et al, 6 2000). Ancillary data used in the analysis included a DEM with a 1 km spatial resolution, a 7

КУИНГДОНГ ШИ, ДЖИАГВО КИ, КСИАОЛИНГ ПЭН, ГУАНГУИ ЛЬВИ,

КИНГСАН ШИ, ХЕЙЯН ЛУ

Chinese environment resource dataset with a scale of 1:4000000, and a world water 8 borderline provided in the ENVI software. These images were geometrically registered 9 using more than 60 ground control points (GCPs) that were previously collected. The images were then transferred to the ARC/INFO grid format for subsequent processing using ARC/GIS 8.3 software.

Table 1. Classification Rule in the Three Zones of the Study Region

Zones (Figure 1) Codes Range of FVC Value Vegetation types Vegetation cover

Plain Low altitude plain area 1. Plain of North Xinjiang DEM<=1300m 2.Plain of East South Xinjiang and Hexi Corridor—Alxi Plain, DEM<=2000m T1 0-0.045 Desert, Gobi without vegetation <10%

T2 0.046-0.095 Plain low grassland 10%-30%

T3 0.096-0.23 Plain middle grassland cover 30%-50%

T4 0.231-0.445 Plain high grassland 50%--70%

T5 0.456-1 Oasis, pasture, meadow >70%

Al tiplano 3.Qaidamu High Basin 3300m>=DEM>2 000m T11 0.456-1 Meadow, High Basin Oasis >70%

Mountain Mountainous l.Mt. of North Xinjiang DEM>1300m 2.Mt. of East South Xinjiang and Hexi Corridor—Alxi Plain, DEM>2000m 3. Qaidamu High Basin DEM>3300m T6 0-0.045 Mountain without vegetation <10%

T7 0.046-0.095 Mountain low grassland cover 10%-30%

T8 0.096-0.23 Mountain middle grassland cover 30%-50%

T9 0.231-0.445 Mountain high grassland cover 50%--70%

T10 0.456-1 Mountainous steppe, forest >70%

Water area T0 1.11 Water area and glaciers 0%

Based on previous studies and reports (Academic Research Xinjiang Term of Chinese Academy of Science & Institute of Botany of Chinese Academy of Sciences, 1978; Editorial Board of Vegetation Map of China, China Academy of Sciences, 2001; Pan et al., 2001, and Shi, 2003a, 2003b), the study area was further divided into three zones: Altai-West Tianshan zone (I), East Tianshan-Tarim Basin zone (II) and Qaidamu high basin-Alxa desert zone (III) as shown in Figure 1. These sub-regions were

divided based primarily on the characteristics of the vegetation type and geographic characteristics (altitude, longitude and latitude).

Different classification methods (supervised maximum likelihood and unsupervised methods) were used for different sub-regions/zones according to their ecological and biophysical properties as shown in Table 1. Since there are large differences in total areas between different vegetation types (for example, the desert occupies about 70% of the total study area while oasis makes up approximately 10%), we used the phenological information to show the vegetation change and to smooth the data in the subsequent change analysis. The vegetation classification maps produced for the research area are shown in Figures 2, 3, 4, and 5 (Appendix 2, Fig. 2,3,4,5). These classification maps were used in change detection analysis by examining their total areas from 1982 - 2000 for each vegetation cover

type.

4. Change Analysis and Results 5 In the Hexi Corridor-Alxi Plain the total oasis and grassland areas went through little 6 changes during the 18 years from 1982-2000, as only 70 pixels changed (Appendix 3, Fig. 6), which is 7 equivalent to a total area of 4200km2 (8km x 8km x 70). The middle grassland cover increased by around 300 pixels (~20,000km2), while the low grassland cover and the desert decreased by about 1200 pixels (~80,000km2). For the mountain vegetation type change in the Qaidamu High Basin, all of the change was much less than in the plain vegetation types (Appendix 3, Fig. 7). Interestingly, the amount of desert in this area is in an upward trend while the meadow areas had a downtrend from 1982-2000. Additionally, the extent of low and middle cover grassland areas decreased whereas the high cover grassland areas increased by approximately 50 pixels (~3,000km2).

In the north Xinjiang there was a decrease of 223 pixels (15,000km2) in the desert area and of 177 pixels (12,000km2) in the high cover grasslands (Appendix 3, Fig. 8). Moreover, the low and middle grassland cover almost disappeared. In contrast, the areas of oasis showed a steady growth throughout 1982-2000 with 414 pixels (29,000km2).

The desert, low, and middle grassland covers exhibited a stability with little changes (less than 50-pixel range) for the mountain vegetation type in the north Xinjiang (Appendix 3, Fig. 9). In contrast, the high grassland cover and mountain forest changed equally in total areas (forest areas increased while the high grassland cover decreased). The desert and low grassland covers of the plain vegetation type in south Xinjiang did not have a significant change over the time period of this study (Appendix 3, Fig. 10). While both the middle and high grassland covers decreased, the oasis increased from 1982 to 2000.

Significant changes in vegetation types were found to be in the mountain vegetation types in the south Xinjiang (Appendix 3, Fig. 11). The desert and the low grassland covers had the largest undulating change while the middle grassland change was in the opposite direction. High grassland and forest covers had a tendency of increases from 1982-2000.

Discussion In the plain areas, there was an overall slight decrease in deserts from 1982 to 2000, which is more obvious in the north Xinjiang. Deserts and low grasslands in the south Xinjiang and Qaidamu High Basin showed an increasing trend, resulting a balance in the total area changes. One reason for the decrease in desert areas in north Xinjiang is likely 6 related to human activities. From 1955-1970, residents in Xinjiang developed new farms on the natural oases, and then abandoned them due to either water shortage or salinity problems. Regrowth of the natural vegetation occurred after these lands were abandoned. These observations agree with well with our field surveys in this study. The fact that the climate in north Xinjiang is propitious to vegetation growth may be another reason for the decrease in the total desert areas (Hu et al, 2001). An additional characteristic in the plain is the continued increase in the total man-made oases, especially in Xinjiang. There were no significant increases in the Hexi Corridor-Alxi Plain. This contrast between the two sub-regions is most likely related to the quality of the farmlands. Xinjiang has the one of the highest quality areas for the production of cotton and the quality of the cotton in this sub-region is much higher than those in other sub-regions. Consequently the cotton farmland expanded quickly near the existing oases to boost the economic profits. In the Hexi Corridor-Alxi plain, the major crop in these old

КУИНГДОНГ ШИ, ДЖИАГВО КИ, КСИАОЛИНГ ПЭН, ГУАНГУИ ЛЬВИ,

КИНГСАН ШИ, ХЕЙЯН ЛУ

oases is wheat. Since the price of wheat has not changed and its value is not as competitive as cotton, there was not motivation to expand wheat production, resulting in little changes in wheat dominated oasis areas.

The vegetation on the mountains was less influenced by human activities and, therefore, the change in vegetation covers in this region most likely reflects the natural variations such as climate change over the study period. Owing to the similarities in the desert, low and middle grassland covers between the mountain region in south Xinjiang and Qaidamu High Basin, it is possible that the same factors caused the changes in the Qaidamu High Basin and the mountains in south Xinjiang. The forest areas increased in Xinjiang due to both human activities and natural climate changes. The logging activities are now regulated by the governments when new policy was in the later 1980s. This resulted in a reduced logging and deforestation. In addition, the regional climate changed over the region and the precipitation increased by approximately 10% over the study period, especially in the northern Xinjiang (The National People's Congress of the People's Republic of China, 1979; Hu Ruji et al, 2001). The coupled consequence is an increase in forested areas in the north Xinjiang. However, in Qaidamu High Basin, the total meadow areas decreased and the amount of desert increased, indicating that the vegetation has degraded in those regions, as there was little increases in precipitation and logging regulations were not enforced.

There was a very strong inter-transitional relationship among the oasis, the high and the middle grassland in the plain areas. In the mountain areas, the same relationship was found to occur between the desert, the low and middle grassland. Moreover, the relationship was found to be the strongest between the high grasslands and forest meadows. Analysis of the changes in vegetation suggested an obvious increasing trend in oases in the plain areas. Particularly from 1980 to 1990, during the period of "rural land reform" olicy in place, oasis areas expanded rapidly. The farmland increased as the cost of the decreases in the high and middle grasslands. In arid and semi-arid regions, ecological functionality of the natural oasis systems is superior to the man-made oases. However, little attention has been paid to the differences between natural and man-made oases. Clearly, the results from this study suggest that the human activity was one of the most important factors resulting in changes in vegetation types in the arid lands of Xinjiang. In the plain areas, man-made oasis is relatively stable while the middle and high grasslands continually decreased. The man-made oasis systems are increasingly subjected to desertification due to reductions in regional precipitation and in ground water supplies. It is important to ensure that the oasis is stable as there is a need to maintain a balance between desert, ecotone and oasis (Pan, 2001). At the same time, it is also important to protect the middle and high grasslands. By protecting the ecotone the oasis can be protected and this may be the best way to sustain the arid and semi-arid lands while maintaining economic development.

Acknowledgment

This research was partially supported by NASA's LCLUC project at Michigan State University (Contract #) and the key project of fundamental research "The ecological environment evolution and control in the arid western China" (G1999043503), "Chun Hui Plan" of Ministry of Education P. R. C (Z2004-1-65004).

REFERENCES

1. Academic Research Xinjiang Term of Chinese Academy of Science & Institute of Botany of Chinese Academy of Sciences, Vegetation and Their Utilized in Xinjiang region. Beijing: Science Press, 1978(in Chinese). Editorial Board of Vegetation Map of China, China Academy of Sciences, 1:1000000 Vegetation Atlas of China, Science Press, Beijing, 2001. (In Chinese) Hu Ru-ji, Fan Zi-li, Wang Ya-jun. Assessment about the Impact of Climate Change on Environment in Xinjiang Since Recent 50 Years. 2001. Р. 97-103.

2. Liu Chuan-sheng, Tashpolat.Tiyip, Ding Jian-li.Landcover Change of Yutian Oasis Based on Remote Sensing and GIS. 2001. Volume 23. P. 59-62.

3. Ma Ming-gou, Dong Li-xin, Wang Xue-mei.Study on the Dynamically Monitoring and Simulating the Vegetation Cover in Northwest China in the Past Years. Journal of Glaciology and Geocryology. 2003. Volume 25. P. 232-236.

4. Meng Ji-jun, Li Zheng-gou, Wu Xiu-qin. Land use changes of Hexi Corridor between 1995 and 2000. 2003. Volume 18. P. 645-651.

5. Pan Xiao-ling, Dang Rong-li. Geographical Distribution and Resource Utilized of 24 Vegetation in West-North China Arid-desert Area[M], Science Press, Beijing, 2001. 25 Pan Xiaoling. A Preliminary Study on the Stability of Oasis Ecosystem in Arid Area, 26 Quaternary Sciences. 2001. Volume 21. P. 345-351.

6. Piao Shilong, Fang Jingyun. Dymamic Vegetation Cover Change over the Last 18 Years in 28 China. Quaternary Sciences. 2001. Volume 21. P. 294-302.

7. Qi J, R.C. Marsett, M.S. Moran et al. Spatial and temporal dynamics of vegetation in the San Pedro River Basin area, Agricultural and Forest Meteorology. 2000. Volume 105. P. 55-68.

8. Shi Qingdong, Lv Guanghui, Pan Xiaoling, et al. Remote sensing classification method for divided areas and applied to Xinjiang vegetation classification. Proceedings of 6 SPIE. Volume 4890, published by SPIE, Bellingham, WA. USA. 2003. Volume 4890, 7. P. 254-261.

9. Shi Qingdong, Lv Guanghui, Wei Ruyi et al. The Vegetation Classification Using FVC and DEM in South Xinjiang of China. Journal of Xinjiang University (Natural Science 10 Edition). 2003. Volume 20(3). P. 280-284.

10. The National People's Congress of the People's Republic of China, Forest Law of P.R.C, 1979 Wang Rang-hui, Fan Zi-li. Study on Land Desertification with RS and GIS Techniques in Alagan, the Lower Reaches of Tarim River[J]. JOURNAL OF REMOTE SENSING. 15. 1998. Volume 2. P. 137-142.

11. Yan Fu-li, Li Zhen, Shao Yun,et al. Detection Methods and Land Cover Change Analysis Using NOAA/AVHRR NDVI Datasets, Remote Sensing Information. 2003. 3. P. 15-18.

12. ZengXubin, Robert E. Dickinson Alison Walker et al. Derivation and Evaluation of Global 19 1-km Fractional Vegetation Cover Data for Land Modeling, Journal of Applied 20 Meteorology. 2000. Volume 39. P. 826-839.

13. Zhang Dian-min, Liu Xiao-yun, Liu Su. New Partitioned Scheme of Xinjiang Vegetation, Arid Zone Research. 1990, 7(1): 1-10. 23 Zhang Wenbo, Fu Suhua, Liu Baoyuan. Error Assessment of Visual Estimation Plant 24 Coverage, Journal of Beijing Normal University. Natural Science. 2001. 25 37(3). P. 402-408.

14. Zhao Chen-yi, Song Yu-dong, Wang Yu-chao, et al., Analysis on the Dynamics of desert-oasis vegetation in the Sangong River Basin. Chinese Journal of Applied Ecology.2004. Volume 15(2). P. 249-254.