Ecosystem Assessment in Altay Region

Spatiotemporal Pattern and Driving Force Analysis of Vegetation Variation in Altay Prefecture based on Google Earth Engine

  • HE Yuchuan , 1 ,
  • XIONG Junnan 1 ,
  • CHENG Weiming 4, 5, 6, 7 ,
  • YE Chongchong 1 ,
  • HE Wen 1 ,
  • YONG Zhiwei 1 ,
  • TIAN Jie 1
  • 1. School of Civil Engineering and Geomatics, Southwest Petroleum University, Chengdu 610500, China
  • 2. School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190, China
  • 3. Altay Regional Committee of the Communist Youth League, Altay, Xinjiang 836500, China
  • 4. State Key Laboratory of Resources and Environmental Information System, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
  • 5. University of Chinese Academy of Sciences, Beijing 100049, China
  • 6. Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, China
  • 7. Collaborative Innovation Center of South China Sea Studies, Nanjing 210093, China
*ABUDUMANAN·Ahemaitihali, E-mail:

HE Yuchuan, E-mail:

Received date: 2021-04-01

  Accepted date: 2021-05-30

  Online published: 2021-11-26

Supported by

The Science and Technology Project of Xizang Autonomous Region(XZ201901-GA-07)

The Key Research and Development Project of Sichuan Science and Technology Department(2021YFQ0042)

The Science and Technology Bureau of Altay Region in Yili Kazak Autonomous Prefecture(Y99M4600AL)


Quantitative evaluation and driving mechanism analysis of vegetation dynamics are essential for promoting regional sustainable development. In the past 20 years, the ecological environment in Altay Prefecture has changed significantly due to global warming. Meanwhile, with increasing human activities, the spatiotemporal pattern and driving forces of vegetation variation in the area are uncertain and difficult to accurately assess. Hence, we quantified the vegetation growth by using the Normalized Difference Vegetation Index (NDVI) on the Google Earth Engine (GEE). Then, the spatiotemporal patterns of vegetation from 2000 to 2019 were analyzed at the pixel scale. Finally, significance threshold segmentation was performed using meteorological data based on the correlation analysis results, and the contributions of climate change and human activities to vegetation variation were quantified. The results demonstrated that the vegetation coverage in Altay Prefecture is mainly concentrated in the north. The vegetation areas representing significant restoration and degradation from 2000 to 2019 accounted for 24.08% and 1.24% of Altay Prefecture, respectively. Moreover, spatial correlation analysis showed that the areas with significant correlations between NDVI and temperature, precipitation and sunlight hours accounted for 3.3%, 6.9% and 20.3% of Altay Prefecture, respectively. In the significant restoration area, 18.94% was dominated by multiple factors, while 3.4% was dominated by human activities, and 1.74% was dominated by climate change. Within the significant degradation area, abnormal degradation and climate change controlled 1.07% and 0.17%, respectively. This study revealed the dynamic changes of vegetation and their driving mechanisms in Altay Prefecture, and can provide scientific support for further research on life community mechanism theory and key remediation technology of mountain-water-forest-farmland-lake-grass in Altay Prefecture.

Cite this article

HE Yuchuan , XIONG Junnan , CHENG Weiming , YE Chongchong , HE Wen , YONG Zhiwei , TIAN Jie . Spatiotemporal Pattern and Driving Force Analysis of Vegetation Variation in Altay Prefecture based on Google Earth Engine[J]. Journal of Resources and Ecology, 2021 , 12(6) : 729 -742 . DOI: 10.5814/j.issn.1674-764x.2021.06.002

1 Introduction

Vegetation dynamic change is a hot topic in current global climate change research, which mainly includes ecological processes, socio-economic processes, and others, and it is easily disturbed by human activities (Jorda-Capdevila et al., 2018; Philippe et al., 2018). Identifying and separating the driving forces of climate change and human activities on vegetation are crucial for understanding the landscape pattern (Ye et al., 2013). The Intergovernmental Panel on Climate Change (IPCC) released the fifth comprehensive report on climate change in 2014, which pointed out that the global temperature has increased by 0.85 ℃ in the past 130 years and the intensity of precipitation has also increased, and that report predicted that the trend of climate warming would continue (IPCC, 2014a; IPCC, 2014b). Additionally, since the 1950s, economic construction activities, the population index and human activities have been increasingly intensified. In arid and semi-arid areas, such as the Altay Prefecture, the vegetation coverage is relatively sparse, the soil is poor, and the ecosystem is extremely fragile, which make such areas very sensitive to the changes of climatic factors such as precipitation, temperature, sunshine duration and so on (Chen et al., 2019). Therefore, vegetation dynamic analysis in Altay Prefecture is vital for maintaining regional ecological balance.
Many scholars have studied the feedback responses of vegetation to climate change, and found that vegetation activity increased in China. This is mainly due to the increases in temperature and precipitation, which provide good hydrothermal conditions and promote the growth of vegetation (Yao et al., 2018). Moreover, the vegetation change in response to temperature is higher than the response to precipitation, and the response of vegetation to hydrothermal changes has a certain lag (Sun et al., 2021). However, in the arid and semi-arid regions such as Xinjiang, precipitation is one of the main meteorological factors driving vegetation change (Du et al., 2015). Although numerous studies have clarified the impact of climate change and human activities on vegetation change, few studies have quantified the rates at which these two factors have contributed to vegetation change.
For separating the driving forces of vegetation change, some studies have tried to conduct quantitative evaluation of climate change and human factors in a specific region (Mahmoud et al., 2018; Wang et al., 2019), but they use different indicators, and there is still a lack of complete separation in the process and evaluation criteria. At present, there are three methods for separating the driving forces of vegetation change: 1) Based on the regression model, the regression relationships between vegetation change, climate change and human activities are established to realize the separation of the driving forces (Mueller et al., 2014). This method is relatively simple to operate and obtaining the needed data is easy, but it is difficult to describe the nonlinear relationships between vegetation change and driving factors (Turner and Carpenter, 2017). 2) Based on residual model analysis, the impact of human activities is estimated indirectly by simulating the difference between vegetation change without human disturbance and actual change with human disturbance (Jiang et al., 2017). The drawback of this method is that model calibration in the year assuming no human disturbances would introduce errors into the model itself. 3) Based on the modelling method of biophysical indicators, the various driving forces are separated through the study of the mechanisms of vegetation change. For example,
the terrestrial ecosystem model (Melillo et al., 1993), the global vegetation model (Foley et al., 1996) and the Carnegie Ames Stanford model (Potter, 1996) (CASA model) were used to simulate the potential Net Primary Productivity (NPP) and actual NPP, and then the relative contribution rate of each driving force index was determined. Although this process-based model can accurately reflect the process of vegetation change, it requires a large number of measured physiological indicators, many of which are difficult to obtain and so the process has great uncertainty.
As an alternative, we used the method of significance threshold segmentation to separate the contributions of climate change and human activities to vegetation change (Jiang et al., 2019). This method does not require a large amount of physical, chemical or human activity data, but it can indirectly quantitatively assess the impact of human activities, the process is relatively complete, and it is a relatively new research method based on correlation statistical models. Multiple climatic factors serve as the assessment index, and a simpler significant correlation serves as the criterion to separate the coupled effects, in order to detect vegetation activities in arid areas (Tian et al., 2015).
In this study, we used trend analysis, correlation analysis and threshold segmentation methods to quantitatively analyze the spatial-temporal patterns and driving forces of vegetation changes in Altay Prefecture from 2000 to 2019 on a 250 m grid scale based on the Google Earth Engine (GEE). This paper mainly answers the questions of whether vegetation has been restored or degraded, and whether climate change or human activities are the main driving factors of regional vegetation changes in Altay Prefecture over the past 20 years.

2 Materials and methods

2.1 Study area

Altay Prefecture (85°31′36″- 91°04′23″E, 45°00′00″- 49°10′45″N) is located in northern Xinjiang, China, and includes the eight counties and cities of Altay, Jimunai, Fuyun and others, with a total area of 1.18×105 km2 (Fu et al., 2017) (Fig. 1). It extends from the northern part of the Gurbantunggut Desert in the south to the southern portion of the Altay Mountains. The center of Altay Prefecture is the Irtysh River Valley, and the terrain is higher in the northern than the southern altitudes. This area has a temperate continental monsoon climate, and is a typical arid and semiarid region with limited precipitation and cold weather. Specifically, the annual average temperature and sunlight hours in the north of Altay Prefecture are lower than those in the south, while on the contrary, the annual precipitation in the north is higher than in the south (Zhang et al., 2019). Affected by hydrothermal conditions, climatic change and human activities, the land cover in 2020 is mainly cropland (1.91%), forest (4.92%), grassland (11.39%), meadow (10.94%) and desert (61.84%).
Fig. 1 Overview of the study area

2.2 Data

2.2.1 NDVI data

Vegetation dynamics is an important ecological process, and the Normalized Difference Vegetation Index (NDVI) is supposed to change in concert with it. Therefore, as a fundamental indicator of the vegetation canopy, NDVI can reflect vegetation dynamics (Lou et al., 2021). The NDVI datasets (MOD13Q1) used in this study were from NASA LP DAAC at the USGS EROS Center (, with spatial resolution of 250 m and temporal resolution of 16 days. These data have been widely used to explore the vegetation variation, based on the Google Earth Engine (GEE). For this study, we extracted the time-series NDVI data for Altay Prefecture from 2000 to 2019, using the maximum value composite technique (MVC) (Holben and Brent, 1986), and then synthesized the annual scale data, which can reflect the best vegetation growth of the whole year and avoid the interferences caused by clouds, aerosols, and solar altitude angles (Li et al., 2020a).

2.2.2 Meteorological data

The meteorological data were provided by the Chinese Meteorological Administration ( We selected monthly meteorological data (temperature, precipitation (Islam et al., 2021), sunlight hours (Ozkaynak et al., 2013) compiled from 34 weather stations of Xinjiang covering from 2000 to 2019. Based on these data, we used the professional interpolation package ANUSPLIN (Stefanie et al., 2017) to interpolate the annual meteorological data for the entire region to surfaces with the same spatial resolution as the NDVI data.

2.2.3 Other data

Land use, land cover data and digital elevation model (DEM) data were collected from GlobeLand30 ( and NASA (, respectively, each with a spatial resolution of 30 m. The above-mentioned data products were resampled (via the bilinear interpolation method) at 250 m to match the NDVI data.

2.3 Methods

2.3.1 Vegetation dynamics assessment

We used the linear regression method to analyze the time trend of NDVI. The slope was estimated by least squares fitting, which could more reasonably and precisely reflect the variation trend of NDVI (Ma and Frank, 2006). The formula is as follows:
$Slope=\frac{n\times \sum\limits_{i=1}^{n}{\left( i\times NDV{{I}_{i}} \right)-\left( \sum\limits_{i=1}^{n}{i} \right)\times \left( \sum\limits_{i=1}^{n}{NDV{{I}_{i}}} \right)}}{n\times \sum\limits_{i=1}^{n}{{{i}^{2}}-{{\left( \sum\limits_{i=1}^{n}{i} \right)}^{2}}}}$
where i is the time series, from 1 to n, n is 20 in this paper; NDVIi is the NDVI value in year i; and Slope is the time trend of NDVI. Slope>0 represents an increasing trend, or vegetation restoration; while Slope<0 represents a decreasing trend, or vegetation degradation. In this study, we used the F-test method (P<0.05) to identify areas of significant vegetation variation. The F-test is given as:
$F=U\times \frac{n-2}{Q}$
whereis the explained sum of squares, is the sum of the squared errors, yi is the NDVI value in year i, $\hat{y}_{i}$ is the return value of NDVI, and $\bar{y}$is the mean value of NDVI.

2.3.2 Correlation analysis

A geographic system is a complex multi-element system, so the fluctuation of one element will inevitably affect the changes of the other elements. The variation of NDVI is affected by various factors, and air temperature, precipitation and sunlight hours are three important abiotic factors for vegetation growth. We used the Pearson correlation coefficient to measure the linear relationships between NDVI and the three meteorological factors on both annual and pixel scales. The formula to calculate the correlation coefficient is shown in equation (3):
$r=\frac{\sum\limits_{i=1}^{n}{\left( {{x}_{i}}-\bar{x} \right)\left( {{y}_{i}}-\bar{y} \right)}}{\sqrt{\sum\limits_{i=1}^{n}{{{\left( {{x}_{i}}-\bar{x} \right)}^{2}}}\sum\limits_{i=1}^{n}{{{\left( {{y}_{i}}-\bar{y} \right)}^{2}}}}}$
where r is the correlation coefficient; $\bar{x}$, $\bar{y}$ are the mean values of NDVI and meteorological factors; xi is the NDVI value in year i, yi is the value of meteorological factors in year i; i is the time series, and n is 20. The t-test was used to carry out the significance test, with P < 0.05 being regarded as significant. The t-test is given as:
$t=r\times \sqrt{\frac{n-2}{1-{{r}^{2}}}}$

2.3.3 Identification of driving forces

In this study, we used the threshold segmentation method to effectively distinguish the effects of climate change and human activities on regional NDVI where a significant variation was noted. This method assumes that short term climate changes and disturbances of natural factors usually do not have significant effects on vegetation variation, but human activities and sudden natural disasters will significantly change vegetation coverage (Li et al., 2012). At the same time, the study area is located in arid and semi-arid alpine regions, and the vegetation is obviously limited by hydrothermal conditions (Ye et al., 2020). Therefore, continuous improvement in hydrothermal conditions will significantly improve the vegetation (i.e., precipitation increase, temperature increase, and sunlight hours increase), while the sustained deterioration of the climate (precipitation decrease, temperature decrease, sunlight hours decrease) can lead to significant degradation of the vegetation. For areas that are not significantly positively correlated with climate change, the significant improvement or significant degradation of vegetation is more likely to be caused by human activities. Therefore, we only consider the positive correlation as the condition to identify the cases where climate significantly affects the vegetation. Furthermore, human restoration activities such as returning farmland to forest and reclaiming oasis in desert, which can transfer the land cover form sparse vegetation land or even bare land to denser vegetation, mean the NDVI value should be significantly increased and the slope of NDVI should be higher than the mean value (threshold) of entire region (Tian et al., 2015).
Figure 2 shows the principles and processes for distinguishing the driving forces. Firstly, we identified areas of significant restoration and significant degradation in NDVI, then we analyzed the correlations between NDVI and the three meteorological factors (temperature, precipitation, sunlight hours). 1) For areas of significant degradation, if NDVI is significantly positively correlated with at least one of the three meteorological factors, then we define this situation as “significant degradation triggered by climate change”, otherwise, we define it as “abnormal degradation”, which may be caused by human activities or sudden natural disasters (fire, mudslide). 2) For areas of significant restoration, if NDVI is significantly positively correlated with at least one of the three meteorological factors, then we define this situation as “significant restoration triggered by climate change”; but if NDVI is not significantly positively correlated with any of the three meteorological factors, and the NDVI slope is higher than the average slope of the entire region, then we define it as “significant restoration triggered by human activities”, otherwise, we define it as “significant restoration triggered by multiple factors (the collective effects of climate change and human activity)”.
Fig. 2 Flowchart of significance threshold segmentation

3 Results

3.1 Spatiotemporal pattern of annual average NDVI

3.1.1 Spatial heterogeneity and temporal dynamics analysis

As shown in Fig. 3a and Table 1, the spatial pattern of NDVI from 2000 to 2019 represents a significant spatial heterogeneity in Altay Prefecture, depicting a trend that decreased from north to south, and the annual average of NDVI across Altay Prefecture was 0.3115. The high-value areas of NDVI were mainly distributed in Altai Mountains, Sawuer Mountains and Irtysh River Valley. The low-value areas of NDVI were mainly located in the Gurbantunggut Desert. According to the trend analysis results (Fig. 3b, Table 2), the NDVI in Altay Prefecture showed a restoration trend with an average increase of 0.0019 yr‒1. The area of vegetation restoration accounted for 50.4% of the total area, and 24.08% of the area was significantly restored (Slope > 0, P<0.05). The area of vegetation showing significant restoration was mainly distributed in Irtysh River Valley, where the land cover was mainly cropland and grassland. By comparison, the areas of vegetation degradation were rare and scattered, accounting for 4.9% of the total area, and the significant degradation area accounted for only 1.24% (Slope < 0, P<0.05). Vegetation showing significant degradation was mainly distributed in the Gurbantunggut Desert and sporadic areas of the Altai Mountains. Meanwhile, the regions where the NDVI slope was higher than the average slope of the entire region were mainly concentrated in the cultivated land in the Irtysh River Valley, and the regions where NDVI slope was lower than the average slope of the entire region were mainly concentrated in the Gurbantunggut Desert and the woodlands in the north.
Fig. 3 Spatial distribution of NDVI in Altay Prefecture

Note:The bottom left in Fig. 3b is the F test result (P<0.05); The statistical graphs in the upper right corner of Fig. 3 are the percentage of each data level in the legend on the right in the entire area.

Table 1 Vegetation cover classification of Altay Prefecture
NDVI Vegetation cover classification Area (km2) Proportion (%)
NDVI<0.2 Low coverage 64716 56.1
0.2≤NDVI<0.4 Medium and low coverage 14253 12.4
0.4≤NDVI<0.6 Medium coverage 8363 7.2
0.6≤NDVI<0.8 Medium and high coverage 7833 6.8
0.8≤NDVI High coverage 20146 17.5

3.1.2 NDVI trend analysis on the county scale

Figure 4 shows the multi-year average NDVI and vegetation variation for the eight counties and cities. In the north, the average NDVI of Altay, Beitun, Habahe and Burqin were higher than the mean NDVI of the entire region. However, in the south, the average NDVI of Jimunai, Fuhai, Fuyun and Qinghe were lower than the mean NDVI of the entire region. The annual average NDVI trend increased in all regions and showed the lowest values around 2008, while the maximum slope of NDVI was in Beitun (0.005 yr‒1), and the minimum slopes of NDVI were in Habahe and Fuyun.
Table 2 NDVI trend grading table
Slope NDVI trend grading Area (km2) Proportion (%)
< -0.0100 Serious degradation 183 0.2
-0.0100 - -0.0050 Moderate degradation 451 0.4
-0.0050 - -0.0010 Slight degradation 5084 4.3
-0.0010 - 0.0010 Essentially unchanged 52531 44.8
0.0010 - 0.0050 Slight restoration 51090 43.6
0.0050 - 0.0100 Moderate restoration 4279 3.6
> 0.0100 Obvious restoration 3586 3.1
Fig. 4 The NDVI trends in different counties of Altay Prefecture from 2000 to 2019

3.2 Spatio-temporal patterns of climatic factors

In this study, we adopted the temperature, precipitation and sunlight hours to reflect the influence of meteorological factors on vegetation variation. As displayed in Fig. 5, these three meteorological factors in Altay Prefecture demonstrated heterogeneous geographical distributions, and their spatial distributions are extremely uneven. The annual average temperature and annual sunlight hours are lower in the mountains than in the river basin (Fig. 5 A1 and B1). However, the distribution of annual precipitation is inverted (Fig. 5 C1). Over the past 20 years, the temperature in most of area presented an increasing trend with fluctuations (91.7%, 0.023 ℃ yr‒1, Fig. 5 A3 and A4), and the high slopes are mainly in the southeast and northwest (Fig. 5 A2). In addition, the precipitation of most regions (61.2%) experienced a downward trend, while 38.8% of the study regions displayed increasing trends (Fig. 5 B4). On the whole, the precipitation showed a slight increase (0.0984 mm yr‒1, Fig. 5 B3), and the high positive trend was mainly concentrated in the northwest, while the low negative trend was mainly concentrated in the southeast. Furthermore, compared with the temperature and precipitation, the sunlight hours in all regions showed a decreasing trend (15.489 h yr‒1, Fig. 5 C3 and C4), and the sharpest declining trends were primarily concentrated in the north (Fig. 5 C2).
Fig. 5 The variation trends of meteorological factors. A, B and C respectively represent air temperature, precipitation and sunshine hours; 1, 2, 3 and 4 respectively represent the multi-year mean, temporal trend, broken line chart of average trend and histogram at the pixel scale of the change rates.

3.3 Relationships between climatic factors and NDVI

Figure 6 shows spatial correlations between NDVI and the meteorological factors of temperature, precipitation and sunlight hours. In order to provide a more visual view of the effects of climatic factors on NDVI in different regions, we used the Maxwell color triangle (RGB) to compound the significant spatial correlations (P < 0.05) for the effects of the three meteorological factors on NDVI (Fig. 7). We found the significant correlations between NDVI and meteorological factors were mainly distributed in the north, however, except for air temperature, the significant correlation almost did not occur at all in the south. The relationships between NDVI and air temperature were mainly positive correlations (accounting for more than 50%), but the area significantly correlated with air temperature was only 3.3%, and among them, significant positive correlations were mainly concentrated in the southern region, and significant negative correlations are mainly concentrated in the northern region (P<0.05, Fig. 6a). Similarly, the relationships between NDVI and precipitation were mainly positive correlations (accounting for more than 60%), with only 6.9% of the regions showing significant correlations, and among them, 6.6% showed significant positive correlations which were chiefly concentrated in the northwest (r > 0, P < 0.05, Fig. 6b). In contrast, negative correlations were the main relationship between NDVI and sunlight hours, which accounted for about 75%, and the area significantly affected by sunlight hours accounted for 20.3%, with 19.97% showing a significant negative correlation and mainly located in the northeast and northwest (r < 0, P < 0.05, Fig. 6c).
Fig. 6 Spatial distribution diagrams of correlation coefficients between NDVI and meteorological factors of temperature (a), precipitation (b), and sunlight hours (c).

Note: The inset maps in the bottom left corner of each show the t test results for each factor (P<0.05) and the statistical graphs in the upper right corner of Fig. 4 are the percentage of each data level in the legend on the right in the entire area.

Fig. 7 Spatial distribution diagrams of significantly correlated r values between NDVI and temperature (T, red), precipitation (P, green) and sunlight hours (S, blue).

3.4 Identification of driving forces

Climate change and diverse human activities could be responsible for the vegetation variations. In this study, we used the threshold segmentation method to determine the contributions of climate change and human activities to the observed vegetation variations. As shown in Fig. 8, in Altay Prefecture during the time-frame of 2000-2019, zones of vegetation with significant restoration accounted for 24.08%, whereas 1.24% of the regions showed significant degradation. The vegetation with significant restoration caused by multiple factors accounted for 18.94% of the entire region, and this phenomenon was widely distributed in Altay, especially in the south of Fuhai County. Zones of vegetation with significant restoration triggered by human activity accounted for 3.4% of the entire region, and were predominately distributed in the Irtysh River Valley. The effects of climate change resulted in the vegetation with significant restoration in a relatively small area (1.74%), and mainly occurred in the northwest of Altay. Simultaneously, zones of vegetation with degradation caused by climate change accounted for 1.07%, and were unevenly distributed in sporadic areas of Altay. However, there were very few areas of anomalous degradation (0.17%), which can be explained by human activities.
Fig. 8 Spatial distribution and proportions of the driving forces

4 Discussion

4.1 Spatial heterogeneity of vegetation variation

Overall, the multi-year (2000-2019) average NDVI in Altay Prefecture presented a significant spatial heterogeneity (Fig. 3a), with a gradient of variation which decreased from north to south that was consistent with the regional elevation changes. Meanwhile, the NDVI increased in most regions of the Altay Prefecture, suggesting that vegetation variation in Altay demonstrated an overall increasing tendency from 2000 to 2019 (Slope > 0, Fig. 3). This result is similar with the observations of vegetation turning green from ongoing studies in Eurasia, arid and semiarid regions and other areas (Zhao et al., 2011). Vegetation variation is inseparable from the effects of climatic changes, anthropogenic activities and geography. In recent years, a great deal of surveillance and research have verified that northwestern China is becoming warmer and wetter (Peng and Zhou 2017; Yang et al., 2017), and this result was demonstrated by the trend analysis of climatic factors in this study. The beneficial hydrothermal conditions provided an advantageous foundation for vegetation growth (Ye et al., 2020). Some artificially implemented positive policies in Altay Prefecture could increase the regional NDVI, such as returning the grain plots to forestry (Li et al., 2020b), reclaiming the oasis in a desert (Su et al., 2007) and the “Three North” Shelterbelt Project (Deng et al., 2019). Concurrently, the grazing and construction of water conservancy facilities also profoundly impact the vegetation of the Altay Prefecture. Previous studies suggested that artificial oases created by agricultural activities contributed to the improvement of vegetation cover (Piao et al., 2003), and the obvious vegetation improvement areas in Altay were mainly concentrated in the oases of the central river basin, although they also have sporadic distribution in the mountainous areas. Meanwhile, with the passage of time, an increase in NDVI also occurred in the areas surrounding the oasis. Such vegetation dynamic variations have also appeared in the relevant studies of Xinjiang (Du et al., 2015).

4.2 The response of the NDVI to climatic factors and human activity

Climate change is one of the important factors affecting vegetation dynamic variations. The hydrothermal condition is the primary limiting factor for vegetation growth (Ye et al., 2020), which is also affected by radiation (Yao et al., 2018). Vegetation variation (Fig. 3a) and its response to climate change (Fig. 7) in Altay Prefecture can be explained from the following aspects. An appropriate temperature increase can promote the enzymatic activities related to photosynthesis (Davidson and Janssens, 2006; Sun et al., 2020), accelerate the decomposition of organic matter (Kim et al., 2012) and extend the growth period of vegetation (Piao et al., 2007), with an evident promotion of photosynthesis. But higher temperatures will significantly affect the enzymatic activities of photosynthesis (Bao et al., 2015) as well as the transpiration of vegetation and the transport of nutrients (Bao et al., 2014), thus inhibiting the growth of vegetation. The increase of precipitation can increase soil moisture content, and soil water is a necessary medium for guaranteeing soil nutrient transportation (Sun et al., 2020), so when light and CO2 concentrations are sufficient, the soil water content determines the rate of photosynthesis. Therefore, the increase of precipitation played a crucial role in the restoration of vegetation. In addition, the sunlight hours can affect the amount of solar radiation the vegetation receives, and the radiation input would change the anaerobic soil conditions caused by high soil moisture content, thus indirectly affecting the photosynthesis of vegetation. In arid and semi-arid regions, precipitation is the main factor affecting vegetation growth, while temperature has little influence, which can explain why the area of NDVI is more significantly affected by precipitation than temperature in Altay (Fig. 6). The continuous increase in temperature may promote the circulation of the atmosphere, leading to an increase in precipitation (Qin et al., 2005), which leads to the hydrothermal synchronization in the Altay Prefecture (Fig. 5). Taking Jimunai County as an example, due to beneficial hydrothermal conditions, the NDVI of 58.68% of the areas improved, and significant improvement accounts for 19.95%. However, 5.39% of degradation occurred in this region, which may be caused by the fact that although the increase of precipitation improved the drought stress of the vegetation (Stefanie et al., 2017), the continuous increase of temperature also led to more intense evaporation in the arid region, which reduced the soil moisture content and inhibited the growth of vegetation (Shi et al., 2014). For northern Altay in the mountains, the regional NDVI showed overall improvement which is mainly controlled by multiple factors, but since human activity is relatively difficult in the mountains, the NDVI dynamics is dominated by climate change. This may be due to rising temperatures in the region, leading to the permafrost melting, the precipitation increased and sunshine time decreased, the soil moisture was supplemented, the germination period of the vegetation was advanced, and the deciduous period was pushed back, resulting in an increase in the vegetation growth period (Oliva et al., 2018), which was conducive to the vegetation prosperity. In addition, the vegetation type in this region is primarily coniferous forest, which absorbs water mainly through shallow root systems and does not depend on deep groundwater, so the increase of soil water is more conducive to its growth (Jiang et al., 2017). The southern region of Altay is mainly desert with poor vegetation coverage, so in the past 20 years, as the temperature rose and the precipitation decreased, it was difficult to improve the vegetation situation, but it was also difficult for it to undergo extreme degradation relative to its initially poor state.
Although NDVI is closely related to climatic factors, the influence of human activities cannot be ignored. Human activities can change the vegetation cover in a short period of time, and their impact is more significant. The impacts of human beings on vegetation mainly include two types: one is improvement including afforestation, returning the grain plots to forestry and reclaiming oasis in desert, etc.; and the other is destruction including overgrazing, construction of reservoirs and deforestation. Studies have shown that the afforestation projects widely implemented around the world have played a positive role in vegetation restoration, improvement of ecosystem services (Gao and Bian, 2019) and soil erosion control (Wang et al., 2014). In addition, grazing is one of the main ways of grassland utilization, and livestock activities can change the biomass, soil structure and nutrients of the grassland. Therefore, long-term overgrazing will also reduce the resilience of grassland vegetation and increase the risk of land degradation (Hao et al., 2018). Combined with the analysis of the results obtained here, we found that the reclamation of the Irtysh River Valley had a remarkable effect: The large area of cultivated land has been increased and the original desert area has been turned into oasis, so the NDVI is worth improving. According to the land use data, the cultivated land in Altay Prefecture is mainly distributed in this region. In 2000, the total area of arable land in Altay Prefecture was 4864.35 km2, which increased to 6201.58 km2 in 2020, for an increase of 27.5%. From the perspective of stocking rate, the overall livestock amount in Altay Prefecture showed a decreasing trend from 2000 to 2019 (Fig. 9), while the actual NDVI during this period showed an increasing trend. Therefore, the decline in stocking rate and the improvement of land use practices have exerted profound effects on the vegetation variation.
Fig. 9 Temporal trend of livestock production

4.3 Evaluation of the significance threshold segmentation method

In order to verify the reliability of driving force separation by the significance threshold segmentation method, especially separating the influences of climate factors and human factors, we compared the NDVI values of the human factors leading to
significant changes in area (Fig. 10). Based on the previous assumption, the disturbance of climate change has no significant effect on vegetation change in a short time, but human activity or sudden disasters will lead to significant changes in the vegetation coverage condition. If the vegetation is significantly improved or degraded by human factors, then the NDVI value of that area will also have a significant change. In Fig. 10, three verification zones were selected to verify the significant changes led by human factors. Regions A and C are the areas with increased cultivated land near the river, and the result of driving force separation is the significant restoration triggered by human activities. In contrast, region B is the man-made reservoir, and the result of driving force separation is abnormal degradation.
Fig. 10 The evaluation chart of significant changes led by human factors. A and C indicate areas where cultivated land has increased, B indicates artificial reservoir areas, 1, 2 and 3 respectively represent NDVI distribution in 2000, NDVI distribution in 2019, and the numbers of pixels with different NDVI values.
As can be seen from Fig. 10, the NDVI values of validation regions A and C have significantly increased in the past 20 years, while those of validation area B have significantly decreased, indicating that the vegetation coverage of validation areas A and C show significant restoration, while the vegetation coverage of validation area B shows significant degradation. Therefore, the method of significance threshold segmentation is reliable for separating the driving forces.

4.4 Uncertainties

In this study, we analyzed the spatiotemporal distribution characteristics of vegetation in the Altay Prefecture and its responses to climate change and human activities, including quantitatively separating the contributions of climate change and human activities to the observed vegetation changes. Nevertheless, there are some deficiencies in this study. First of all, in addition to the three meteorological factors selected, other factors, such as soil nutrients, soil microorganisms and carbon dioxide concentration (Zhang et al., 2021), will also have a decisive impact on vegetation changes. Therefore, in future studies, more attention should be paid to the influences of soil properties and other factors on vegetation change. Of course, the influences of evaporation (Hoffman et al., 2011), acquisition date, sensors, etc., on the inter-annual NDVI should also be considered in future research. Also, although the NDVI index can reflect the vegetation productivity to a certain extent (Piao et al., 2005), the net primary production (NPP) can better reflect the situation of vegetation photosynthesis and ecosystem vitality. As a result, NPP is more sensitive to climate change and human activities and can more accurately reflect the vegetation change response mechanism (Liu et al., 2019), thus NPP can be used to reflect ecosystem productivity in future studies. In addition, this study only conducted a simple quantitative separation and analysis of the driving factors, and subsequent in-depth research on the influencing mechanisms of each driving factor can be carried out, to further clarify the size and change of their contribution rates.

5 Conclusions

In the present study, based on MODIS13Q1 NDVI and meteorological data, and using the maximum value composite, tendency estimation of Slope, Pearson correlation coefficient, significance test, and threshold segmentation methods, we analyzed the spatial and temporal heterogeneity of NDVI in Altay Prefecture from 2000 to 2019, and quantitatively separated the contributions of climate change and human activities to the observed NDVI changes. The spatial and temporal distribution characteristics of NDVI showed that the vegetation coverage in Altay Prefecture was low and uneven, showing a pattern of “north > south, mountain > basin > desert”. Over the past 20 years, the climate change evident in the Altay Prefecture showed weak rises in temperatures and precipitation, and sunlight hours were reduced. Combined with spatiotemporal characteristics of NDVI, and correlation analysis results between NDVI and meteorological factors, we found a significant vegetation restoration area with the proportion of 48.77%, in which multiple factors dominated 18.94%, human factors dominated 3.40%, and climate change dominated 1.74%. For 1.24% of the areas with significant degradation of the vegetation, the abnormal degradation accounted for 1.07%, and the climate change dominated degradation accounted for 0.17%. Based on the unique geographical structure of the Altay area, human activities are more difficult in most of the northern and southern areas, so the multiple factors were dominated by climate change. The significant restoration caused by human factors can be explained by ecological restoration projects and land reclamation, while the abnormal degradation can be explained by population growth, urbanization expansion and abandoned land. Meanwhile, the decrease of livestock carrying capacity also promotes the restoration of vegetation.
The authors express their gratitude to the editors and reviewers for their efforts, and thank the National Geomatics Center of China and Chinese Meteorological Administration for data sets.

The authors express their gratitude to the editors and reviewers for their efforts, and thank the National Geomatics Center of China and Chinese Meteorological Administration for data sets.

Bao G, Bao Y H, Sanjjava A, et al. 2015. NDVI-indicated long-term vegetation dynamics in Mongolia and their response to climate change at biome scale. International Journal of Climatology, 35(14): 4293-4306.


Bao G, Qin Z H, Bao Y H, et al. 2014. NDVI-based long-term vegetation dynamics and its response to climatic change in the Mongolian Plateau. Remote Sensing, 6(9): 8337-8358.


Chen T, Bao A M, Jiapaer G, et al. 2019. Disentangling the relative impacts of climate change and human activities on arid and semiarid grasslands in Central Asia during 1982-2015. Science of the Total Environment, 653: 1311-1325.


Davidson E A, Janssens I A. 2006. Temperature sensitivity of soil carbon decomposition and feedbacks to climate change. Nature, 440(7081): 165-173.


Deng C L, Zhang B Q, Cheng L Y, et al. 2019. Vegetation dynamics and their effects on surface water-energy balance over the Three-North Region of China. Agricultural and Forest Meteorology, 275: 79-90.


Du J Q, Shu J M, Yin J Q, et al. 2015. Analysis on spatio-temporal trends and drivers in vegetation growth during recent decades in Xinjiang, China. International Journal of Applied Earth Observation and Geoinformation, 38: 216-228.


Foley J A, Prentice I C, Ramankutty N, et al. 1996. An integrated biosphere model of land surface processes, terrestrial carbon balance, and vegetation dynamics. Global Biogeochemical Cycles, 10(4): 603-628.


Fu Q, Li B, Hou Y, et al. 2017. Effects of land use and climate change on ecosystem services in Central Asia’s arid regions: A case study in Altay Prefecture, China. Science of the Total Environment, 607-608: 633-646.

Gao J, Bian H Y. 2019. The impact of the Plains afforestation program and alternative land use scenarios on ecosystem services in an urbanizing watershed. Urban Forestry & Urban Greening, 43: 126373. DOI: 10.1016/j.ufug.2019.126373.


Hao L, Pan C, Fang D, et al. 2018. Quantifying the effects of overgrazing on mountainous watershed vegetation dynamics under a changing climate. Science of the Total Environment, 639: 1408-1420.


Hoffman M T, Cramer M D, Gillson L, et al. 2011. Pan evaporation and wind run decline in the Cape Floristic Region of South Africa (1974-2005): Implications for vegetation responses to climate change. Climatic Change, 109(3-4): 437-452.


Holben B N. 1986. Characteristics of maximum-value composite images from temporal AVHRR data. International Journal of Remote Sensing, 7(11): 1417-1434.


IPCC. 2014a. Climate Change 2013: The physical science basis: Summary for policymakers. Geneva, Switzerland, IPCC.

IPCC. 2014b. Climate change 2014:Synthesis report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Geneva, Switzerland: IPCC.

Islam A R M T, Islam H T, Shahid S, et al. Spatiotemporal nexus between vegetation change and extreme climatic indices and their possible causes of change. Journal of Environmental Management, 289: 112505. DOI: 10.1016/j.jenvman.2021.112505.


Jiang H L, Xu X, Guan M X, et al. 2020. Determining the contributions of climate change and human activities to vegetation dynamics in ago-pastural transitional zone of northern China from 2000 to 2015. Science of the Total Environment, 718: 134871. DOI: 10.1016/j.scitote nv.2019.134871.


Jiang L L, Guli J, Bao A M, et al. 2017. Vegetation dynamics and responses to climate change and human activities in Central Asia. Science of the Total Environment, 599-600: 967-980.

Jorda-Capdevila D, Gampe D, Huber García V, et al. 2019. Impact and mitigation of global change on freshwater-related ecosystem services in Southern Europe. Science of the Total Environment, 651: 895-908.


Kim D G, Vargas R, Bond-Lamberty B, et al. 2012. Effects of soil rewetting and thawing on soil gas fluxes: A review of current literature and suggestions for future research. Biogeosciences, 9(7): 2459-2483.


Li A, Wu J G, Huang J H. 2012. Distinguishing between human-induced and climate-driven vegetation changes: A critical application of RESTREND in Inner Mongolia. Landscape Ecology, 27(7): 969-982.


Li K W, Tong Z J, Liu X P, et al. 2020a. Quantitative assessment and driving force analysis of vegetation drought risk to climate change: Methodology and application in Northeast China. Agricultural and Forest Meteorology, 282-283: 107865. DOI: 10.1016/j.agrformet.2019.107865.


Li W Q, Chen J H, Zhang Z M. 2020b. Forest quality-based assessment of the Returning Farmland to Forest Program at the community level in SW China. Forest Ecology and Management, 461: 117938. DOI: 10.1016/j.foreco.2020.117938.


Liu Y Y, Wang Q, Zhang Z Y, et al. 2019. Grassland dynamics in responses to climate variation and human activities in China from 2000 to 2013. Science of the Total Environment, 690: 27-39.


Lou J P, Xu G Y, Wang Z J, et al. 2021. Multi-year NDVI values as indicator of the relationship between spatiotemporal vegetation dynamics and environmental factors in the Qaidam basin, China. Remote Sensing, 13(7): 1240. DOI: 10.3390/rs13071240.


Ma M G, Veroustraete F. 2006. Interannual variability of vegetation cover in the Chinese Heihe River Basin and its relation to meterological parameters. International Journal of Remote Sensing, 27(22): 5127. DOI: 10.1080/01431160601103475.


Mahmoud S H, Gan T Y. 2018. Impact of anthropogenic climate change and human activities on environment and ecosystem services in arid regions. Science of the Total Environment, 633: 1329-1344.


Melillo J M, McGuire A D, Kicklighter D W, et al. 1993. Global climate change and terrestrial net primary production. Nature, 363(6426): 234-240.


Mueller T, Tucker C J, Dressler G, et al. 2014. Human land-use practices lead to global long-term increases in photosynthetic capacity. Remote Sensing, 6(6): 5717-5731.


Oliva M, Pereira P, Antoniades D. 2018. The environmental consequences of permafrost degradation in a changing climate. Science of the Total Environment, 616: 435-437.

Ozkaynak E. 2013. Effects of air temperature and hours of sunlight on the length of the vegetation period and the yield of some field crops. Ekoloji, 22(87): 58-63.

Peng D D, Zhou T J. 2017. Why was the arid and semiarid northwest China getting wetter in the recent decades? Journal of Geophysical Research: Atmospheres, 122(17): 9060-9075.

Philippe A K R, Phoebe K, Ebun A, et al. 2019. Down scaling of climate change scenarii to river basin level: A transdisciplinary methodology applied to Evrotas River Basin, Greece. Science of the Total Environment, 660: 1623-1632.


Piao S L, Fang J Y, Liu H Y, et al. 2005. NDVI—Indicated decline in desertification in China in the past two decades. Geophysical Research Letters, 32(6): 347-354.

Piao S L, Fang J Y, Zhou L M, et al. 2003. Interannual variations of monthly and seasonal normalized difference vegetation index (NDVI) in China from 1982 to 1999. Journal of Geophysical Research: Atmospheres, 108(D14): 4401. DOI: 10.1029/2002JD002848.


Piao S L, Pierre F, Ciais P, et al. 2007. Growing season extension and its impact on terrestrial carbon cycle in the Northern Hemisphere over the past 2 decades. Global Biogeochemical Cycles, 21(3): 3018. DOI: 10.1029/2006GB002888.


Potter C S, Davidson E A, Verchot L V. 1996. Estimation of global biogeochemical controls and seasonality in soil methane consumption. Chemosphere, 32(11): 2219-2246.


Qin D H, Dong G G, Lu Q, et al. 2005. Assessment of climate and environment changes in China (I): Climate and environment changes in China and their projection. Climate Change Research, 1(2): 4-9. (in Chinese)

Shi P J, Sun S, Wang M, et al. 2014. Climate change regionalization in China(1961-2010). Science China Earth Sciences, 57(11): 2676-2689.


Herrmann S M, Anyamba A, Tucker C J. 2005. Recent trends in vegetation dynamics in the African Sahel and their relationship to climate. Global Environmental Change, 15(4): 394-404.


Su Y Z, Zhao W Z, Su P X, et al. 2007. Ecological effects of desertification control and desertified land reclamation in an oasis-desert ecotone in an arid region: A case study in Hexi Corridor, Northwest China. Ecological Engineering, 29(2): 117-124.


Sun H Z, Wang J Y, Xiong J N, et al. 2021. Vegetation change and its response to climate change in Yunnan Province, China. Advances in Meteorology, 2021(12): 1-20.

Sun J, Zhou T C, Liu M, et al. 2020. Water and heat availability are drivers of the aboveground plant carbon accumulation rate in alpine grasslands on the Tibetan Plateau. Global Ecology and Biogeography, 29(1): 50-64.


Tian H J, Cao C X, Chen W, et al. 2015. Response of vegetation activity dynamic to climatic change and ecological restoration programs in Inner Mongolia from 2000 to 2012. Ecological Engineering, 82(4): 276-289.


Turner M G, Carpenter S R. 2017. Ecosystem modeling for the 21st century. Ecosystems, 20(2): 211-214.


Wang F Y, Duan K Q, Fu S Y, et al. 2019. Partitioning climate and human contributions to changes in mean annual streamflow based on the Budyko complementary relationship in the Loess Plateau, China. Science of the Total Environment, 665: 579-590.


Wang Z J, Jiao J Y, Su Y, et al. 2014. The efficiency of large-scale afforestation with fish-scale pits for revegetation and soil erosion control in the steppe zone on the hill-gully Loess Plateau. Catena, 115: 159-167.


Yang P, Xia J, Zhang Y Y, et al. 2017. Temporal and spatial variations of precipitation in Northwest China during 1960-2013. Atmospheric Research, 183: 283-295.


Yao Y T, Wang X H, Li Y, et al. 2018. Spatiotemporal pattern of gross primary productivity and its covariation with climate in China over the last thirty years. Global Change Biology, 24(1): 184-196.


Ye C C, Sun J, Liu M, et al. 2020. Concurrent and lagged effects of extreme drought induce net reduction in vegetation carbon uptake on Tibetan Plateau. Remote Sensing, 12(15): 2347. DOI: 10.3390/rs12152347.


Ye X C, Zhang Q, Liu J, et al. 2013. Distinguishing the relative impacts of climate change and human activities on variation of streamflow in the Poyang Lake Catchment, China. Journal of Hydrology, 494: 83-95.


Zhang G P, Yan J J, Zhu X T, et al. 2019. Spatio-temporal variation in grassland degradation and its main drivers, based on biomass: Case study in the Altay Prefecture, China. Global Ecology and Conservation, 20: e00723. DOI: 10.1016/j.gecco.2019.e00723.


Zhang M Y, O’Connor P J, Zhang J Y, et al. 2021. Linking soil nutrient cycling and microbial community with vegetation cover in riparian zone. Geoderma, 384: 114801. DOI: 10.1016/j.geoderma.2020.114801.


Zhao X, Tan K, Zhao S, et al. 2011. Changing climate affects vegetation growth in the arid region of the northwestern China. Journal of Arid Environments, 75(10): 946-952.