Download stxb201409051755 Spatio-temporal distribution pattern of

stxb201409051755
Spatio-temporal distribution pattern of vegetation coverage in Junggar Basin,
Xinjiang
Cheng Duan1, Ling Wu1*, Lingyun He2, Shaoming Wang1
(1.College of Life Science, Shihezi University, Shihezi 832003, Xinjiang, China;
2. College of Sciences, Shihezi University, Shihezi 832003, Xinjiang, China)
Abstract: Vegetation coverage is an indicator used for exploring the growth of vegetation, which has attracted
attention from ecologists owing to its significant role in ecological conservation and restoration. As an important
component of the terrestrial ecosystem, changes in vegetation coverage reflect changes in the environment,
especially with respect to arid areas. A change of vegetation coverage will trigger desertification, degradation of
the eco-environment, and regional climate change. Hence, we studied the spatio-temporal distribution pattern of
vegetation coverage and ephemerals in the Junggar basin during the last few decades based on Normalized
Difference Vegetation Index. The following conclusions were drawn. The vegetation coverage of the basin shows
an overall upward trend, mainly because of the expansion of farmland. In particular, the southern rim has already
formed an apparent continuous oasis belt, and the central part of the basin has low vegetation coverage due to a
rugged environment but is relatively stable, whereas the northern part has a relatively small oasis distribution
pattern. With respect to annual variation, the vegetation coverage of the basin reached the peak in July of every
year, whereas the peak in the desert was in May or June; ephemeral plants clearly flourished during this phase.
Ephemeral plants were distributed in the whole basin; the most prosperous region was the southern of the basin
edge, and the plants decreased gradually from the edge to the center of the basin. The desert area of the central
basin was the least prosperous area, but the southern and mid-eastern parts of the desert were prosperous. The
expansion of the oasis has mainly had an impact on the 50-km range of the oasis edge, and the greatest impact is at
the 5- to 20-km range of the oasis edge.
Key words: vegetation coverage; spatio-temporal distribution; cropland expansion; ephemerals; Junggar Basin
1. Introduction
With the intensifying trend of global change in recent decades, vegetation has attracted
attention from ecologists owing to its significant role in ecological conservation and restoration
and impact on climate change [1]. As an important component of the terrestrial ecosystem,
vegetation has played a crucial role in material cycle and energy flow between atmosphere and
terrestrial ecosystem via a series of complex biogeochemistry and geophysical processes [2,3].
As an index, vegetation coverage can reflect the status of vegetation growth, and its changes
also with a series of complex ecological processes. Especially in arid areas, where changes in
vegetation not only mirror evolvement of ecological environment, but also can trigger
desertification [4], regional climate change [5], and environmental degradation [6]. For the
investigation of vegetation coverage, traditional field study has been impossibly satisfied
quantitative study at a large scale. In recent decades, with the popularization of remote sensing, it
has been a convenient means using remote sensing data to monitor and analyze vegetation
coverage [7]. Currently, most of the researches using remote sensing data to monitor changes of
*Corresponding author. College of Life Science, Shihezi University, Shihezi 832003, Xinjiang, China
E-mail address：lingw@shzu.edu.cn(L.Wu)
vegetation coverage were based on normalized difference vegetation index (NDVI) to estimate
and analyze [8,9].
Junggar Basin is an economically important agricultural region in Xinjiang, so the use of
remote sensing to monitor spatio-temporal changes in vegetation coverage at a macro-scale has an
important meaning for the sustainable development of the economy and eco-environment. Hence,
in this paper, using NDVI from 2000 to 2013 to analyze spatio-temporal distribution pattern of
vegetation coverage and ephemerals in Junggar Basin of Xinjiang, our objectives were to provide
some scientific basis for the protection and restoration of regional eco-environment.
2. Study area and methods
2.1 Study area
Junggar Basin is located in the northern Xinjiang Uyghur Autonomous Region of China. It is
the second largest interior basin in China and is a representative area in the arid zone. It was found
between the Tianshan Mountains and Altai Mountains, forming an irregular triangle, and its terrain
tilts to the west. The southern edge of the basin is mainly oasis, and the vast Gurbantunggut Sandy
Desert is distributed in the central part of the basin. The area has a typical temperate arid climate,
with a low rate of annual rainfall and substantial evaporation, the winter and spring precipitation
accounted for 30%-40% of the total annual precipitation. The soil is mainly desert lime soil and
brown soil, with a low organic matter.
According to the 1:4,000,000 administrative map of China, topography map of Xinjiang, and
descriptions of structure of Junggar Basin [10], we ascertain the scope of study area is a irregular
triangle, with a 910 km east-west long and 426 km north-south wide.
2.2 Data sources and preprocessing
The data used in the study is MOD13Q1(https://lpdaac.usgs.gov/) data provided by the Land
Processes Distributed Active Archive Center(LP DAAC) of NASA, at a spatial resolution of 250m
with a 16-day time interval from 2000 to 2013. The dataset considered effects such as atmospheric
calibration and geometric corrections.
For data preprocessing, the MODIS Reprojection Tool (MRT) provided by NASA was used to
transform the format (HDF to TIFF) and projection (from Sinusoidal projection to Geographic
Lat/Lon, WGS1984). The original DN values were transformed to NDVI. We also manipulated
(prepossessed) the MOD13Q1 data subset via regions of interest (ROIs, vector border of study
area) based on ENVI4.8 version software. Finally, the NDVI data was analyzed to determine the
correlation between interannual trends and variances from 2000 to 2013. Furthermore, we used
Maximum value composite (MVC) model to obtain the highest value for each pixel of 23 images
during a year, and binary pixel model to estimate vegetation cover based on NDVI [11].
3. Results and analysis
3.1 Spatio-temporal distribution pattern of vegetation coverage in Junggar Basin
We preprocessed the MOD13Q1-NDVI data from the 2000 to 2013, and then estimated the
vegetation cover of images based on the binary pixel model and MVC. According to the
cumulative percentage of vegetation cover, we applied the decision tree method to classify the
study region into the five threshold value ranges (0-0.09, 0.09-0.13, 0.13-0.2, 0.2-0.56, >0.56),
which showed the distribution of vegetation cover for 2000 and 2013 in the Junggar Basin (Fig.
1).
90° E
84° E
86° E
88° E
90° E
84° E
86° E
88° E
90° E
84° E
86° E
88° E
90° E
(2000)
46° N
44° N
46° N
46° N
44° N
44° N
44° N
46° N
48° N
88° E
48° N
48° N
86° E
48° N
84° E
(2013)
Fig. 1 Vegetation coverage in Junggar Basin classification figure. Classification of vegetation cover: 0-0.09 represents fixed desert with
almost no vegetation; 0.09-0.13: fixed desert with little or sparse vegetation; 0.13-0.2: semi-fixed desert with low vegetation cover;
0.2-0.56: region with medium vegetation cover; >0.56: region with high vegetation cover, such as forests, grassland, and artificial oases.
It can be seen from Fig.1, the highest vegetation coverage area of the basin was in the south
and north, followed by the eastern region, with the midwest showing the lowest cover area. The
highest cover area in the basin was in the south, which was due to the artificial oases distributed in
the agricultural areas and the northern Tianshan Mountains from east to west, which are an
important agricultural zone in Xinjiang. The western artificial oases distributed along the Manas
River basin.
In addition, we found that artificial oases area obviously expanded to the central desert from
2000 to 2013. In 2000, the highest cover area mainly distributed along the national road and
Manas River basin, with a zonal distribution, while in 2013, artificial oases in Manas River basin
has been increased obviously into a patchy distribution, especially expansion of artificial oases in
Karamay. According to data statistics, artificial oases in Xinjiang increased from 13,000 km2 in
1953 to 61,900 in 2000(about 5 times increase). During the research period, red region in
vegetation coverage classification figure increased gradually, which may be related to the
reclamation of new cropland. Since 1998, with the popularization of drip irrigation, as well as the
increased price of cotton and policy-induced economy, farmers in Manas River basin have planted
large amounts of cash crops to increase economic returns [12]. Qin et al [13] studied
spatio-temporal changes in cropland in northern China, and showed that significant cropland
expansion occurred in Xinjiang from 2000 to 2010. Moreover, we found that the other classes
were not fixed, with no distribution pattern being apparent. In particular, the 0.2-0.56 and 0.13-0.2
classes in the central desert increased obviously from 2000 to 2013, it illustrated that natural
vegetation in the central desert increased during this period, which was likely due to the increase
of precipitation or ecological conservation, and it is hard to make this conclusion only according
to images.
3.2 The inter-annual variability of vegetation coverage
The highest cover area of the basin mainly distributed in artificial oases area, changes of
vegetation cover in artificial oases area were mainly affected by the human activities, while
changes of natural vegetation in the basin were mainly related to climate change. Hence,
according to observational data of rainfall provided by National Meteorological Information
Center of China, we counted the average annual precipitation of six meteorological stations
(including Urumchi, Hebukesaier, Karamay, Jinghe, Wulanwusu, Qitai) in study area (Fig.2b). For
the inter-annual variability of vegetation coverage, we take two cases into consideration, firstly,
the inter-annual variability of vegetation coverage in basin, secondly, the inter-annual variability
of natural vegetation in basin after excluding cropland area. The results are as shown in Fig.2a.
Fig. 2 (a) Change curve of vegetation in Junggar Basin from 2000 to 2013.(b) Precipitation in Junggar Basin from 2000 to 2013.
As shown in Fig.2(a), the change trend of two cases were same, the inter-annual variation
generally presented an upward trend, the increased trend were coincident with the results Li et al.
[14]studied in 2010. According to results we analyzed above, the reason why the inter-annual
variation presented an increased trend was that expansion of cropland in desert area. The increased
trend also can be seen from images, pixels above 0.56 made up 8.87% of the total area of the basin
in 2000, Which increased to about 12% by 2013(a 1.35 times increase). Moreover, natural
vegetation coverage also presented an upward trend, it indicated that natural vegetation coverage
had increased during research period, which may be related to increase of rainfall and ecological
conservation in recent decades. According to result in Fig.2(b), precipitation showed a slight
increased trend during study period, with study confirming that precipitation in northern Xinjiang
has increased in recent decades[15]. In arid area, change in precipitation has a great impact on
growth of natural vegetation, especially for vegetation in desert, which was highly sensitive to
rainfall, increase in precipitation was beneficial to increase of vegetation coverage.
3.3 The annual variance of vegetation coverage
For the annual variance of vegetation coverage, we analyzed average NDVI value of a 16-day
time interval in different years. According to results, we found that change trend were basically
same in different years, the change characteristics of NDVI accorded with climate change during
the four seasons in Xinjiang. Since the spring, the NDVI had a gradual increased trend and
increased to top point in summer, then had a gradual decreased trend in autumn, finally, decreased
to lowest point in winter. However, we wondered that whether the change characteristics of
different areas in basin were same or not? Therefore, we analyzed average NDVI value in the
central desert area, after eliminating the northern and southern artificial oasis regions and the
interference of low-value eastern and western regions in the central desert region, with a 200 km
east-west long and 150 km north-south wide rectangle in the center. Finally, we analyzed average
NDVI value in the basin and in the central desert area.
Fig. 3 NDVI average from April to July in Junggar Basin
According to results in Fig.3, we found that average NDVI value in the basin increased to top
point in July or August, and then had a gradual decreased trend. However, in the central desert
region, NDVI value presented a fast increased trend in March and April, even faster than in the
basin, while during this period, most vegetation (including cropland) in the basin did not start to
growth, NDVI value increased to top point at the end of May, decreased suddenly at followed 16
days, then began to increase to a stable level. The change characteristics of vegetation in desert
region, which may related to growth of ephemerals in desert area. In spring, ephemerals began to
germinate rapidly using snowmelt, with the vegetation index increased to top point at May, and
then with the death of ephemerals, the vegetation index reduced, in June, with growth of other
plants in the desert, the vegetation index increased to a stable level. In May, most of vegetation in
the desert has just started to grow, the max NDVI value was related to contribution of ephemerals.
In April and May, the basin has a strong windy and dusty period, while ephemerals appeared a
bloom phase during this period, which has a important meaning for wind prevention and sand
fixation [16].
3.4 Spatio-temporal distribution pattern of ephemerals in Junggar Basin
Ephemerals has a crucial meaning for eco-environment and sand fixation in Gurbantünggüt
Desert. According to results analyzed, at end of May or beginning of June, there was an image
showed high NDVI, and then decreased after 16 days, which may be related to short life cycle of
ephemerals. If we use high value image to subtract low value image, then area show high value in
the image can regard as area of high vegetation coverage of ephemerals.
Consequently, we used high value image at end of May to subtract low value image, after that,
we found that some high value areas were the same in different years, but some were not, it
indicated that spatio-temporal distribution pattern of ephemerals was not constant, which was
related to precipitation and temperature[17]. Of course, there may be noise and error in images
caused the change, excluding the noise and error, we added together images in different years and
88° E
90° E
84° E
86° E
88° E
90° E
48° N
86° E
44° N
44° N
46° N
84° E
46° N
48° N
divided by total years (14), then obtained average value of images, the distribution of high value
areas in the images were the same with many of years, as shown in Fig.4.
Fig. 4 The subtract and add diagram of high value and low value images from 2000 to 2013
As shown in Fig.4, the red areas in the image were negative value areas, which were cropland
and water body, distributed along northern of Tianshan Mountain, Ebinur Lake and Fuhai. The
yellow areas mainly distributed in the central desert, the other colors represented the positive
value areas, which NDVI value reflected the coverage of ephemerals. With respect to distribution
of ephemerals, ephemeral plants were distributed in the whole basin; the most prosperous region
was the southern of the basin edge, and the plants decreased gradually from the edge to the center
of the basin. The desert area of the central basin was the least prosperous area, but the southern
and mid-eastern parts of the desert were prosperous, which was consisted with field study of
previous research [18]. With the advent of mechanized agriculture, technological innovations, and
market-oriented economic modernization, expansion of the artificial oasis has caused
fragmentation of the vegetation, which made the distribution of ephemerals reduced.
3.5 Impact of artificial oases expansion on natural vegetation
In recent decades, because of influence of human activities, a large number of natural
vegetation landscapes around the desert had been destructed, which had caused a great impact on
distribution of natural vegetation and eco-environment. Hence, we used buffer analysis to analyze
the impact of artificial oasis expansion on natural vegetation. First, we extracted pixels of NDVI
greater than 0.56 in the image as cropland area, then to made 0-5km, 5-10km, 10-20km, 20-50km,
50-100km buffer zones, which including the natural vegetation in the buffer zone and excluding
cropland.
Fig. 5 The NDVI average data of variety regions of the artificial oasis edge
As shown in Fig. 5, the impact presented the same tendency in different years. The NDVI
values gradually declined until the 20 km buffer zone, then increased and stabilized at the 50 km
buffer zone, and finally converged at the 100 km buffer zone. Artificial oasis expansion mainly
impacted the 50 km buffer zone range, with the greatest influence in the 10-20 km range and the
least influence in the 0-5 km range.
The vegetation in the buffer zone 0-5 km from the artificial oases mainly included Haloxylon
ammodendron, H. persicum and shrubs, which grew well even when the groundwater level was
4-5 m under the land surface, so the influence of groundwater level on vegetation growth was
insignificant [19]. However, plants in 10-20 km buffer zone mainly included ephemeral and
ephemeroid plants with root systems only 0-30 cm under the land surface [20]. Due to the lack of
surface runoff and deep groundwater, the growth of plants in this region relied on atmospheric
precipitation, so the influence of groundwater level on vegetation growth was greatest.
4. Conclusion and discussion
In conclusion, the following conclusions were drawn:
(1). The vegetation coverage in the basin has a gradual increased trend from 2000 to 2013, in
particular, the artificial oasis has been increased obviously into a patchy distribution in
south of the basin, the central part of the basin has low vegetation coverage but is
relatively stable, whereas the northern part has a relatively small oasis distribution
pattern.
(2). The inter-annual variation presented an upward trend from 2000 to 2013, which was due
to expansion of artificial oasis in desert area. The natural vegetation also presents an
increased trend, because of the increase of rainfall. With respect to annual variation, the
vegetation coverage of the basin reached the peak in July of every year, whereas the peak
in the desert was in May or June; ephemeral plants clearly flourished during this phase.
(3). Ephemeral plants were distributed in the whole basin; the most prosperous region was the
southern of the basin edge, and the plants decreased gradually from the edge to the center
of the basin. The desert area of the central basin was the least prosperous area, but the
southern and mid-eastern parts of the desert were prosperous.
(4). The expansion of the oasis has mainly had an impact on the 50-km range of the oasis
edge, and the greatest impact is at the 5- to 20-km range of the oasis edge. However,
beyond the range of 50 km, the vegetation cover does not appear to be affected. This is
because the expansion of cropland has uplifted the level of underground water; therefore,
the vegetation of this region far away from the oasis is mainly affected by precipitation.
Discussion: conclusion in this paper only analyzed based on remote sensing data, which may
have noise and the trace of splicing in the images, and impact of NDVI itself (including soil
background effect). Therefore, these impact factors may cause some errors in the results.
Furthermore, the precipitation data used was sourced from only six meteorological stations, which
did not reflect the practical situation of the whole basin, notably climate change in the central
desert regions. More detailed and accurate data is required to analyze climatic impact. For
spatio-temporal distribution of ephemerals in this paper, there are some errors in the images and
we yet unable to estimate the impact of error. Hence, further effort is required to quantify the
different factors mentioned above and combine with field study at a long time-scale to reduce
uncertainties.
Acknowledgements
This work was funded by the National Youth Science Foundation of China (31300406) and
Social Development of Scientific and Technological Research of Xinjiang Production and
Construction Corps (2015AD023).We also thank the anonymous reviewers for their constructive
comments.
References
[1] Fang J Y, Liu G H, Xu S L. Biomass and net production of forest vegetation in China. Acta Ecologica Sinica,
1996, 16(5): 498-508.
[2] Zuo L J, Zhang Z X, Zhao X L, Wang X, Wu W B, Yi L, Liu F. Multitemporal analysis of cropland transition in
a climate-sensitive area: a case study of the arid and semiarid region of northwest China. Regional Environmental
Change, 2014, 14(1): 75-89.
[3] Yao J, He X Y, Li X Y, Chen W, Tao D L. Monitoring responses of forest to climate variations by MODIS
NDVI: a case study of Hun River upstream, northeastern China. European Journal of Forest Research, 2012,
131(3): 705-716.
[4] Lakes T, Muller D, Krüger C. Cropland change in southern Romania: a comparison of logistic regressions and
artificial neural networks. Landscape Ecology, 2009, 24(9): 1195-1206.
[5] Dubovyk O, Menz G, Conrad C, Kan E, Machwitz M, Khamzina A. Spatio-temporal analyses of cropland
degradation in the irrigated lowlands of Uzbekistan using remote-sensing and logistic regression modeling.
Environmental Monitoring Assessment, 2013, 185(6): 4775-4790.
[6] Baeza A, Bouma M J, Dobson A P, Dhiman R, Srivastava H C, Pascual M. Climate forcing and desert malaria:
the effect of irrigation. Malaria Journal, 2011, 10: 190-190.
[7] Yang Z P, Gao J X, Zhou C P, Shi P L, Zhao L, Shen W S, Ouyang H. Spatio-temporal changes of NDVI and
its relation with climatic variables in the source regions of the Yangtze and Yellow rivers. Journal of Geographical
Sciences, 2011, 21(6): 979-993.
[8] Bayramov E, Buchroithner M F, McGurty E. Determination of main climate and ground factors controlling
vegetation cover regrowth along oil and gas pipelines using multiple, spatial and geographically weighted
regression procedures. Environmental Earth Sciences, 2012, 66(7): 2047-2062.
[9] Thavorntam W, Tantemsapya N. Vegetation greenness modeling in response to climate change for Northeast
Thailand. Journal of Geographical Sciences, 2013, 23(6): 1052-1068.
[10] Qu G S, Ma Z J, Chen X F et al. Theory of structure and its evolution in the Junggar basin. Xinjiang
petroleum geology, 2009, 30(1): 1-5.
[11] Li M M, Wu B F, Yan C Z et al. Estimation of vegetation fraction in the upper basin of Miyun Reservoir by
remote sensing. Resource Science, 2004, 26(4): 154-158.
[12] Gu L F. The Xinjiang production and construction corps cotton drip irrigation under membrane formation and
development. Water saving irrigation, 2003, 1: 27-29.
[13] Qin Y W, Liu J Y, Shi W J, Tao F L, Yan H M. Spatial-temporal changes of cropland and climate potential
productivity in northern China during 1990-2010. Food Security, 2013, 5(4): 499-512.
[14] Li Y, Liu Y, Zhang P. Desert vegetation coverage changes and their impacts on sand-dust weather in
Gurbantunggut desert. Journal of Arid land Resource and Environment, 2010, 24(3): 141-146.
[15] Xu G Q, Wei W T. Climate Change of Xinjiang and its Impact on Eco-Environment. Arid land geography,
2004, 27(1): 14-18.
[16] Wang X Q, Jiang J, Lei J Q et al. The Distribution of Ephemeral Vegetation on the Longitudinal Dune Surface
and Its Stabilization Significance in the Gurbantunggut Desert. Acta Geograpgica Sinica, 2003, 58(4): 598-605.
[17] Yuan S F, Tang H P. Patterns of ephemeral plant communities and their adaptations to temperature and
precipitation regimes in Dzungaria Desert, Xinjiang. Biodiversity Science, 2010, 18(4): 346-354.
[18] Qian Y B, Wu Z N, Zhang L Y et al. Spatial distribution characteristics of the short-lived Gurbantunggut
desert plants. Chinese science bulletin, 2007, 52(19): 2299-2306.
[19] Qian Y B, Wu Z N et al. 2010. Environments of Gurbantunggut Desert. Beijing: Science press. 101, 103.
[20] Li X Y. Preliminary studying the characteristic of roots and relations between roots and environment of
ephemerals in Xinjiang. Research in arid areas, 2000, 17(3): 28-34.