Spatiotemporal Changes in Vegetation Cover and Soil Moisture in the Lower Reaches of the Heihe River Under Climate Change

Abstract

As global climate change intensifies, arid land ecosystems face increasing challenges. Vegetation, a key indicator of climate variation, is highly responsive to meteorological factors such as temperature (Tem), precipitation (Pre), and soil moisture (SM). Understanding how fractional vegetation cover (FVC) responds to climate change in arid regions is critical for mitigating its impacts. This study utilizes MOD13Q1-NDVI data from 2000 to 2022, alongside corresponding Tem, Pre, and SM data, to explore the dynamics and underlying mechanisms of SM and FVC in the context of climate change. The results reveal that both climate change and human activities exacerbate vegetation degradation, underscoring its vulnerability. A strong correlation between FVC and both Tem and Pre suggests that these factors significantly influence FVC variability. In conclusion, FVC in the lower reaches of the Heihe River is shaped by a complex interplay of Tem, Pre, SM, and human activities. The findings provide a scientific basis and decision-making support for ecological conservation and water resource management in the lower reaches of the Heihe River, aiding in the development of more effective strategies to address future climate challenges.

1. Introduction

Global climate change is one of the most significant challenges facing the world today, with its widespread impact on Earth’s ecosystems drawing considerable attention from the international community [1,2]. In arid regions, the effects of climate change are especially pronounced, as these vulnerable ecosystems are highly sensitive to climate fluctuations. Vegetation, a critical component of terrestrial ecosystems, is highly susceptible to variations in climate conditions [3,4]. The concepts of “Carbon Farming” and “Soil Carbon Sequestration” involve improving farming practices to increase the organic carbon content in the soil, which helps mitigate climate change. [5]. Furthermore, forests in different regions play a significant role in absorbing atmospheric carbon dioxide. [6]. By analysing the effects of forest growth, forest management, and natural disturbances (such as wildfires and insect infestations) on the carbon sequestration capacity of forests, this study provides valuable insights into forest carbon dynamics. Therefore, studying vegetation dynamics in arid regions against the backdrop of climate change holds significant ecological and practical importance [7].

Vegetation in these regions is heavily dependent on water availability, with soil moisture (SM) being a key factor influencing vegetation growth [8,9]. SM serves as the link between the atmosphere, vegetation, soil, and groundwater, which reflects the Earth’s water utilisation. SM significantly impacts vegetation growth and fractional vegetation cover (FVC) [10]. Under global climate change, variations in SM and vegetation growth are expected to significantly impact FVC. FVC is a key quantitative indicator for studying vegetation cover dynamics and effectively reflects vegetation trends and ecosystem health [11,12]. Amilcare Porporato and his team studied how vegetation cover in arid regions responds to changes in soil moisture, as well as the adaptation mechanisms of different vegetation types under drought stress [13]. Sonia Seneviratne’s team investigated how global warming affects soil moisture and vegetation cover in arid regions [14]. By analysing climate models and observational data, they examined the impact of extreme climate events (such as heatwaves and prolonged droughts) on soil moisture and vegetation distribution in arid regions, particularly the constraints that water limitations impose on vegetation growth. Therefore, researching the spatiotemporal characteristics of SM and FVC in arid regions is essential for addressing global climate change, enhancing regional ecological protection, and promoting the sustainable development of natural resources.

In recent years, significant progress has been made in research on vegetation cover changes in arid regions. Particularly in Northwest China, many scholars have analysed the spatiotemporal changes in vegetation cover and its relationship with climate factors using remote sensing data and climate models [15]. These studies have shown that vegetation cover in China is highly sensitive to climate change, with variations in temperature (Tem) and precipitation (Pre) leading to significant changes in SM and FVC. However, existing studies are limited in terms of geographic coverage and data time scales. Longer time series of remote sensing data and higher spatial resolution are needed to better understand how different vegetation types respond to climate change across regions [16,17].

The methods used in this study are widely applied in vegetation change studies. The methods help to reveal long-term trends in vegetation cover in arid regions and their underlying climate drivers. In addition to the direct relationship between SM and FVC, global climate change also indirectly affects vegetation by altering regional hydrological cycles [18,19]. For example, changes in precipitation patterns and temperature owing to climate change alter evapotranspiration and water vapor exchange, which in turn affect SM and its ability to support vegetation growth. Studies have shown that, under the influence of climate change, the arid regions of Northwest China have experienced both warming and increased humidity. Water evaporation caused by rising temperatures and greater precipitation can promote vegetation growth, which enhances vegetation cover [20,21]. However, this process exhibits non-linear patterns, with warming and humidification demonstrating complex spatiotemporal variability.

The Heihe River basin serves as a prime example. As the second-largest inland river in Northwest China, the downstream areas of the Heihe River are characterised by a typical arid climate, with a highly fragile ecosystem [22]. The low vegetation cover of the region is particularly vulnerable to the impacts of climate change [23,24]. Therefore, studying the spatiotemporal characteristics of SM and FVC in the Heihe River basin and surrounding areas not only enhances our understanding of how climate change affects arid ecosystems but also provides scientific support for regional ecological protection and resource management [25,26]. In conclusion, this study aims to uncover the driving mechanisms behind vegetation cover changes in the context of climate change by analysing the spatiotemporal characteristics of SM and FVC in the arid regions of Northwest China [27,28]. This research explores how different vegetation types respond to climate change by integrating remote sensing data with various analytical methods. The research provides scientific evidence to support ecosystem management and policy-making in the face of climate change challenges [29].

Water is a key factor influencing vegetation growth, community characteristics, and FVC in the lower reaches of the Heihe River, with climate change directly impacting its eco-hydrological processes [30,31]. Many researchers have studied groundwater, SM, salinity, and ecological water transmission to examine their effects on vegetation growth in this region, which has led to a range of valuable findings [32,33]. However, there is still a lack of sufficient research on climate change as a critical factor affecting the vegetation–soil–hydrological processes in the lower reaches of the Heihe River. In particular, the characteristics of FVC changes and their relationship to climate change remain unclear [34,35]. Furthermore, the natural vegetation in the lower reaches of the Heihe River serves as a primary barrier against wind and sand encroachment, which safeguards the ecosystem, and acts as a crucial ecological buffer zone that influences the ecological security of Northwest and North China [36,37,38]. This study investigates the spatiotemporal variations and responses of SM and vegetation cover in the arid zone of Northwest China in the context of climate change.

This study provides a detailed analysis of the relationship between vegetation cover changes in arid regions and global climate change. By elucidating how climate factors and soil moisture influence vegetation dynamics, it offers valuable insights into the complex interactions between climate and ecosystems. This paper specifically presents predictive insights into how climate change may impact ecosystems in arid regions, particularly in terms of changes in vegetation cover. These predictions are crucial for understanding potential future scenarios under various climate conditions. Based on the research findings, adaptive strategies for ecosystem management in response to climate change are proposed. These strategies, grounded in scientific evidence, aim to help policymakers, land managers, and local communities better adapt to the challenges posed by changing environmental conditions.

2. Materials and Methods

2.1. Study Area

The lower reaches of the Heihe River are located in Ejina Banner (39°52′–42°47′ N, 97°10′–103°7′ E) at the westernmost point of the Inner Mongolia Autonomous Region. Ejina Banner is bordered by Alxa Right Banner to the southeast, Gansu Province to the west and southwest, and Mongolia to the north. The elevation in this area ranges from 898 m to 1598 m and encompasses approximately 114,600 km² (Figure 1). Ejina Banner experiences a typical continental arid climate, characterised by dryness, low precipitation, high evaporation, abundant sunshine, and frequent sandstorms. The average annual temperature is 8.3 °C, while the average precipitation is 37.0 mm, and evaporation reaches 3841.5 mm. The region experiences frequent winds, often accompanied by sandstorms. The topography is predominantly characterised by Gobi, low mountains, deserts, rivers, lakes, and oases, with most of the area comprising Gobi deserts, which creates a transitional zone between oases and deserts. Vegetation is relatively sparse and primarily consists of desert riparian forests, desert shrublands, and desert meadows [39,40].

2.2. Data Sources and Processing

The NDVI data used in this study were obtained from NASA (https://ladsweb.modaps.eosdis.nasa.gov/) (accessed on 19 September 2024). The data type is MOD13Q1 and covers the period from 2000 to 2022, with a spatial resolution of 250 m and a temporal resolution of 16 days. The maximum value composite method was applied to process the data, resulting in a 23-year interannual sequence dataset (

Table 1. Data sources and formats used in this study.

Data Type	Source	Calculation	Time Series	Resolution	Abbreviation
NDVI	https://ladsweb.modaps.eosdis.nasa.gov/ (accessed on 19 September 2024)	Maximum value composite	2000–2022	250 m	NDVI
The mean annual temperature	Peng, S. (2020). 1 km monthly maximum temperature dataset for China (1901–2023). National Tibetan Plateau/Third Pole Environment Data Center. https://doi.org/10.5281/zenodo.3185722 (accessed on 19 September 2024)	Monthly data average	1901–2022	1 km	Tem
The mean annual precipitation	Peng, S. (2020). 1 km monthly precipitation dataset for China (1901–2023). National Tibetan Plateau/Third Pole Environment Data Center. https://doi.org/10.5281/zenodo.3114194 (accessed on 19 September 2024)	Monthly data summing	1901–2022	1 km	Pre
Soil moisture	https://data.tpdc.ac.cn (accessed on 19 September 2024)	Monthly data average	2000–2020	1 km	SM

).

Meteorological data, including Tem and Pre, were obtained from the China National Tibetan Plateau Data Center (http://data.tpdc.ac.cn) (accessed on 19 September 2024) covering the period from 1901 to 2022. The Pre data features a monthly dataset with a resolution of 1 km for China, while the Tem data consists of a monthly maximum temperature dataset with the same resolution. The meteorological data were converted to Tif format, with monthly Pre data summed to derive annual precipitation and monthly Tem data averaged to calculate annual average temperature. Projection, resampling, and clipping operations were performed in ArcGIS to align the annual average temperature and annual average precipitation with the NDVI dataset in terms of pixel size and projection (Table 1).

The SM data used in this study were derived from satellite inversion of the global daily surface SM dataset, which has a resolution of 1 km and was released by the National Tibetan Plateau Science Data Centre (https://data.tpdc.ac.cn/) (accessed on 19 September 2024). This dataset provides monthly data spanning from 1 February 2000 to 31 December 2020.

Using the lower reaches of the Heihe River as a case study, the research utilises MOD13Q1-NDVI (Normalised Difference Vegetation Index) data from 2000 to 2022, along with Tem, Pre, and SM data from the same period. These data are employed to map the spatiotemporal dynamics of vegetation cover, assess its evolutionary trends, and explore the relationship between vegetation growth and Tem, Pre, and SM.

2.3. Research Methods

2.3.1. Calculation of Vegetation Cover

FVC exhibits a strong linear correlation with NDVI [41]. The integration of vegetation indices with the pixel dichotomy model effectively mitigates the impact of atmospheric and soil background factors and is widely used for estimating vegetation cover at both regional and global scales (Figure 2). Therefore, this study employed the pixel dichotomy model based on NDVI to estimate vegetation cover. The calculation of FVC is provided in Equation (1) [42,43].

$$FVC={\displaystyle \frac{NDVI-{NDVI}_{soil}}{{NDVI}_{veg}-{NDVI}_{soil}}}$$

(1)

FVC represents the fractional vegetation cover, while NDVI refers to the normalised difference vegetation index for a unit grid. One of the variables denotes the NDVI value for bare soil or non-vegetated areas, which is theoretically close to 0, and indicates the NDVI value for fully vegetated areas, which is theoretically close to 1. Owing to the comprehensive influence of lighting conditions, image quality, vegetation types, and other factors, the actual values may vary. Values within the confidence interval of [5%, 95%] were selected as the values for the year. This study references the relevant literature on vegetation cover classification in arid regions [44,45]. According to field survey data, vegetation cover is classified into four levels: 0 ≤ FVC < 0.05 for no vegetation cover; 0.05 ≤ FVC < 0.25 for low vegetation cover; 0.25 ≤ FVC < 0.50 for medium vegetation cover; 0.50 ≤ FVC < 1.00 for high vegetation cover.

2.3.2. Stability Analysis

The coefficient of variation ($C_v$) quantifies the degree of data dispersion and is commonly used to represent the variability of a data series over a given study period. We can reveal the stability distribution characteristics of vegetation growth in the lower reaches of the Heihe River by calculating the pixel-by-pixel coefficient variation of vegetation cover data during the study period. The calculation of is provided in Equation (2) [46].

$$C_v={\displaystyle \frac{\sqrt{\frac{1}{n}{\sum _{i=1}^n\left({FVC}_i-\overline{FVC}\right)}^2}}{\overline{FVC}}}$$

(2)

where one of the variables is the vegetation cover value in the i-th year, another variable is the annual average vegetation cover during the study period, and n is the number of monitoring years.

2.3.3. Trend Analysis

Methodologically, the Mann–Kendall trend analysis, Theil–Sen slope estimator, and Hurst exponent are classical tools for studying the spatiotemporal trends of climate change and vegetation dynamics. The Mann–Kendall trend analysis is a non-parametric test commonly used to detect trends in time series data, while the Theil–Sen slope estimator assesses trend intensity by calculating the median slope between two points in a time series. The Hurst exponent is used to quantify long-term memory and autocorrelation in time series.

This study utilised the Theil–Sen trend analysis, the Mann–Kendall test, and the Hurst index to investigate the spatiotemporal changes, evolutionary trends, and sustainability characteristics of vegetation cover.

(1)

Theil–Sen Trend Analysis and Mann–Kendall Test

Theil–Sen trend analysis and the Mann–Kendall test are non-parametric statistical methods that effectively mitigate issues related to data distribution and missing values. These methods are also resistant to the influence of outliers, leading to more robust and reliable results. The trend of vegetation cover change is defined by Equation (3) [47].

$$S_{FVC}=Median({\displaystyle \frac{{FVC}_j-{FVC}_i}{j-i}}),2000\le i<j\le 2022$$

(3)

When >0, it indicates an increasing trend in FVC, while <0 indicates a decreasing trend.

The Mann–Kendall test is used to assess the significance of trends [48]. The calculation formula is as follows:

For a set {} [36,49], I = 2000, 2001, …, 2022, the Z statistic is defined as follows:

$$Z=\left\{\begin{array}{c}{\displaystyle \frac{S-1}{\sqrt{Var\left(S\right)}}} (S>0)\\ 0 (S=0) \\ {\displaystyle \frac{S+1}{\sqrt{Var\left(S\right)}}} (S<0)\end{array}\right. \mathrm{Where} S={\sum }_{i=1}^{n-1}{\sum }_{j=i+1}^{n}sign(x_j-x_i)$$

(4)

$$Sign(x_j-x_i)=\left\{\begin{array}{c}1 x_j-x_i\\ 0 x_j-x_i\\ -1 x_j-x_i\end{array}\right.$$

(5)

$$Var={\displaystyle \frac{n(n-1)(2n+5)}{18}}$$

(6)

where two of the variables represent the FVC values in years j and i, respectively, and n is the length of the time series. The value range of Z is (−∞, +∞). At a given significance level α, when $\left|\mathrm{Z}\right|>{\mathrm{u}}_{1-\propto /2}$, it indicates a significant change in the study series at the α level [50,51].

(2)

Hurst Index

The Hurst index is a useful method for quantitatively describing the dependence in a time series and is widely applied in studying vegetation cover changes to predict the developmental trends of FVC. The calculation steps are as follows [52]:

$$\overline{{\xi }_{\tau }}={\displaystyle \frac{1}{\tau }}\sum _{\tau =1}^{\tau }\xi \left(t\right) 1\le t\le \tau =\mathrm{1,2},3\dots ,n$$

(7)

$$\mathrm{X}\left(t,\tau \right)=\sum _{n=1}^t(\xi \left(n\right)-\overline{{\xi }_{\tau }})$$

(8)

$$\mathrm{R}\left(\tau \right)=\underset{1\le t\le \tau }{\max }\mathrm{X}\left(t,\tau \right)-\underset{1\le t\le \tau }{\min }\mathrm{X}\left(t,\tau \right)$$

(9)

$$\mathrm{S}\left(\tau \right)={\left[{\displaystyle \frac{1}{\tau }}\sum _{t=1}^{\tau }{(\xi \left(t\right)-\overline{{\xi }_{\tau }})}^2\right]}^{{\displaystyle \frac{1}{2}}}$$

(10)

$${\displaystyle \frac{\mathrm{R}\left(\tau \right)}{\mathrm{S}\left(\tau \right)}}={\left(c\tau \right)}^H$$

(11)

If the ratio follows the relation ${\mathrm{R}}_{\left(\mathrm{\tau }\right)}/{\mathrm{S}}_{\left(\mathrm{\tau }\right)}\triangleq \mathrm{R}/\mathrm{S}\propto {\mathrm{\tau }}^{\mathrm{H}}$, the time series exhibits the Hurst phenomenon, with H being the Hurst index. The value of H can take three forms: If 0 < H < 0.5, it indicates anti-persistence in the FVC. If H = 0.5, it suggests that the FVC follows a random sequence. If 0.5 < H < 1, it indicates persistence in the FVC.

3. Results

3.1. Spatiotemporal Distribution Characteristics

The spatial distribution of vegetation cover in the lower reaches of the Heihe River (Figure 3) shows that overall vegetation cover is relatively low, with areas of low vegetation cover being predominant. These low vegetation cover zones make up approximately 93.37% of the study area and are mainly concentrated in the Gobi desert region. Medium and high vegetation cover areas were smaller and covered 4.18% and 1.71% of the study area, respectively, and were mainly concentrated around the Ejina Oasis and Gurinai Lake. Figure 3 shows vegetation cover in the upper and lower sections of the East and West Rivers and the East and West Juyan Lakes regions. In contrast, the middle section of the East River and other Gobi regions display poorer vegetation cover. Thus, Ejina Banner, located in the westernmost part of Inner Mongolia, is predominantly desert and semi-desert. Vegetation growth largely depends on water replenishment from the middle and upper reaches of the Heihe River, which makes the ecological environment extremely fragile and highly vulnerable to climate change.

Between 2000 and 2022, the vegetation coverage in the lower reaches of the Heihe River showed a slow upward trend with fluctuations (Figure 3). The annual mean FVC value fluctuated within the range of 0.095 to 0.122, indicating relatively small fluctuations. However, the interannual variation trends in the areas of different vegetation coverage levels varied. The areas of medium and high vegetation coverage zones both showed an increasing trend, with increase rates of 37.14 km²/a and 24.19 km²/a, respectively (Figure 3e). In contrast, the areas of no coverage and low vegetation coverage zones showed a decreasing trend, with reduction rates of 54.87 km²/a and 6.45 km²/a, respectively. This indicates that areas of no vegetation and low vegetation coverage are transitioning towards zones with medium and high vegetation coverage.

Furthermore, through transition matrix analysis of FVC from 2020 to 2022, it was found that the areas of medium and high coverage increased the most, with increases of 854.19 km² and 556.37 km², respectively. The increase in medium coverage was mainly contributed to by the transition from low coverage, while the increase in high coverage was primarily due to the transition from medium coverage (Figure 3d). Overall, since the implementation of ecological water delivery in the Heihe River in 2000, the vegetation in Ejina has shown significant improvement and recovery, which is consistent with related studies [40].

3.2. Stability Evaluation of Vegetation Cover Changes

The coefficient of variation (CV) was used as an indicator to evaluate vegetation stability, and the variation index was divided into five levels using the quantile method: high (0 < CV ≤ 0.11), relatively high (0.11 < CV ≤ 0.16), medium (0.16 < CV ≤ 0.21), relatively low (0.21 < CV ≤ 0.31), and low (0.31 < CV ≤ 2.83). ArcGIS was used to calculate the area of each stability level. The majority of the study area’s vegetation coverage fell within the relatively high or medium stability categories, accounting for 30.09% and 43.17% of the region, respectively. These areas were mostly located in desert regions with low vegetation coverage, consistent with the trend of minimal changes in the area of low vegetation coverage. Regions with relatively low and low stability accounted for 15.00% and 8.71% of the area, respectively, mainly distributed in the northern part of the lower reaches of the Heihe River, the lower reaches of the Donghe River, and the areas around East and West Juyan Lake (Figure 4a). This shows that the vegetation’s ecological stability in the core oasis area of the lower reaches of the Heihe River is weak and highly susceptible to climate change and human activities, indicating high ecological vulnerability. Overall, the lower reaches of the Heihe River basin contain a large number of areas with low vegetation stability, particularly in the light green and dark green regions, indicating high ecological vulnerability in these areas. They are more susceptible to the impacts of climate change and human activities.

3.3. Dynamic Evolution Trends of Vegetation Cover

The Theil–Sen median trend analysis and Mann–Kendall test were employed to create a spatial distribution map of vegetation cover change trends in the lower reaches of the Heihe River (Figure 4b). The analysis revealed that vegetation improvement occurred over an area of 31,865.81 km² and accounted for 53.34% of the region, with significantly improved areas covering 19,112.38 km² (31.99%), mainly concentrated around the Ejina Oasis and Gurinai Lake. In contrast, the area of vegetation degradation was much smaller. It covered only 4180.50 km² in area, or 7% of the region. Therefore, over the past 23 years, the overall trend of vegetation cover in the lower reaches of the Heihe River has been one of improvement.

Overall, most areas of the Heihe River Basin exhibit low to medium Hurst index values (orange, yellow, and light green), indicating poor vegetation cover stability in these regions, especially in certain parts of the lower reaches of the Heihe River where the long-term trends of vegetation change are relatively uncertain. In contrast, the relatively high Hurst index in green areas reflects the relative stability of vegetation in a small portion of the Heihe River Basin. (Figure 5a). This suggests that the region overall displays weak anti-persistence, which indicates that future changes may reverse past trends. Coupling the linear fit slope with the Hurst index (Figure 5b) reveals that 4.81% of the area experienced continuous degradation, while 55.65% showed non-persistent improvement. Therefore, this study predicts that, without effective ecological protection measures, more than half of the vegetation in the lower reaches of the Heihe River is at risk of continuous degradation in the future.

3.4. Response of Vegetation Cover to Climate Change

Existing studies have shown that climate change significantly affects vegetation cover in arid regions, with Tem and Pre being the most direct and sensitive climatic factors [44,53]. To examine how vegetation cover in the lower reaches of the Heihe River responds to climate change, this study analyses the correlation and impact trends using monthly Tem and Pre data from 2000 to 2022. As illustrated in Figure 6a, Tem exhibits a notable upward trend over the past 23 years, at a rate of 0.418 °C per decade, which is significant at the 0.05 level. This indicates a noticeable warming trend in the Ejina region, consistent with the warming patterns observed in Northwestern China. The annual Pre trend rate is −1.195 mm per decade, which signifies a slight but decreasing trend, accompanied by significant fluctuations in the curve. This suggests that precipitation in Ejina exhibits unstable interannual variability. Over the past 23 years, the Ejina region has experienced a gradual ‘warming and drying’ trend [54].

Pearson correlation analysis was applied to calculate the relationships among FVC, Tem, and Pre. The correlation coefficients between FVC and Tem in the study area range from 0.84 to −0.79 (Figure 6b). Regions with a significant positive correlation account for 9.98% of the area, mainly located in the southeastern part of the study area and along the Ejina River, where warming promotes vegetation growth. Conversely, regions with a significant negative correlation make up 1.06% of the area, scattered and discontinuously distributed, which indicates a threshold effect of Tem on vegetation growth, where further increases in Tem exacerbate drought and inhibit vegetation growth. The largest proportion of regions, approximately 88.96%, exhibit a non-significant correlation.

Compared with Tem, FVC exhibits a weaker correlation with Pre, with correlation coefficients ranging from 0.85 to −0.78 (Figure 6c). Regions with significant positive (p < 0.01), significant negative (p < 0.05), and non-significant correlations (p > 0.05) account for 6.94%, 0.98%, and 92.08% of the area, respectively. The areas with a positive correlation are mainly located in the southern part of the region, where the dry climate and herbaceous plant types make vegetation more sensitive to changes in precipitation, which enhances growth with increased Pre. Significant negative correlation areas are sporadically distributed in the central part of the study area and tend to decrease in size.

The lower reaches of the Heihe River are situated in an arid to extremely arid climate region, where the average annual Pre is only 48.03 mm, and evaporation far exceeds Pre. Consequently, vegetation growth in these extremely water-scarce areas is highly dependent on the limited Pre, leading to a positive correlation between FVC and Pre. However, this correlation can be easily disrupted by natural and human factors, such as intensive agricultural irrigation near the Ejina River, which may weaken the relationship between climatic factors and FVC. In addition to Tem and Pre, vegetation cover in the lower reaches of the Heihe River is influenced by factors such as surface water, groundwater, and soil water–salt content [36,39]. This indicates that the vegetation ecosystem in this region remains fragile and is highly susceptible to climate change and human activities.

3.5. Relationship Between SM and NDVI

To analyse the impact of Tem and Pre on the seasonal variation of SM, a seasonal spatial distribution map of SM was created (Figure 7). The results indicate that the seasonal dynamics of SM in the lower reaches of the Heihe River are pronounced, with the highest SM observed in autumn at 10%, and the lowest levels occurring in summer and winter at 9.71%. The seasonal dynamics of SM in this region are mainly influenced by factors such as Tem, rainfall, and vegetation. The relationship between climatic factors and changes in SM is illustrated in Figure 8b,c. Spring (March to May) is identified as a critical recovery period for SM in the lower reaches of the Heihe River. As Tem gradually rises, winter snow and ice begin to melt, leading to increased rainfall and a replenishment of SM. As the demand for water by vegetation increases, SM is rapidly consumed. In spring, SM fluctuates owing to the combined influences of Tem, rainfall, and vegetation. Summer (June to August) serves as the primary replenishment period for SM. Although high temperatures accelerate SM evaporation and intensify plant transpiration, increased rainfall steadily replenishes SM, resulting in a gradual increase. In autumn (September to November), as Tem and rainfall decrease, SM enters a period of stable consumption and exhibits a slow declining trend. In winter (December to February of the following year), as Tem and Pre decrease, snow is preserved owing to the low temperatures, keeping SM relatively stable. Although changes in SM during winter are minimal, they play an important role in replenishing SM in the following spring. Snowmelt is a crucial source of SM during this season.

A correlation analysis was conducted between the NDVI during the growing season (May to September) and the corresponding SM. The correlation coefficients were classified into six levels: strong positive (0.8 < R ≤ 1), moderate positive (0.3 < R ≤ 0.8), weak positive (0 < R ≤ 0.3), weak negative (−0.3 ≤ R < 0), moderate negative (−0.8 ≤ R < −0.3), and strong negative (−1 ≤ R < −0.8). This analysis resulted in a distribution map illustrating the correlation between NDVI and SM in the lower reaches of the Heihe River (Figure 8a). The findings revealed that NDVI is weakly positively correlated with SM in most areas, particularly in regions with low vegetation cover, primarily in desert and Gobi areas, which account for approximately 66.1% of the study region. This suggests that in these remote desert areas, vegetation is significantly constrained by hydrological conditions, such as surface water and groundwater, which makes it more susceptible to variations in SM and results in a positive correlation [55,56]. However, the correlation is relatively weak. Although an increase in SM can promote vegetation growth to some extent, it is also influenced by climatic conditions, soil texture, and soil salinity.

4. Discussion

4.1. Climate Change in the Lower Reaches of the Heihe River Under Global Climate Change

Global climate change is one of the most pressing challenges of our time, which profoundly affects climate systems worldwide. Arid regions, characterised by unique climatic conditions and environmental sensitivity, respond markedly to global climate change [57]. Global warming has generally resulted in rising temperatures in these areas, accelerated moisture evaporation, and exacerbated drought conditions. In recent decades, temperatures in arid regions have increased significantly at a rate of 0.032 °C per year, which reflects an accelerating warming and drying trend. The impact of global climate change on precipitation in arid regions is more complex, with varying effects across different areas [54]. Some areas may experience reduced precipitation, which exacerbates drought conditions, while others may see increased precipitation, which indicates a ‘wetting’ trend. Additionally, climate change may result in more pronounced seasonal and spatial variations in precipitation, particularly in terms of timing and intensity. Some arid regions might experience more concentrated precipitation and more frequent extreme rainfall events, leading to flooding, even as the overall climate becomes drier.

The lower reaches of the Heihe River, located in the high-latitude arid region of Northwest China, exhibit strong continental climate characteristics and are significantly affected by global climate change [58]. Since the 1980s, particularly after 1986, temperatures in this region have shown a marked warming trend, although the rate of increase varies by season. Winter temperatures have increased the most, while spring temperatures have increased the least, which is somewhat inconsistent with the broader climatic changes observed in Northwest China and Inner Mongolia [59]. Precipitation in the lower reaches of the Heihe River has shown significant decadal variability. The 1960s and 1990s recorded higher precipitation than the overall average, while other decades exhibited a decreasing trend. Since 2000, annual precipitation has alternated between positive and negative anomalies, with an overall declining trend, particularly after 2018, when precipitation consistently decreased. This indicates that, while global climate change impacts the lower reaches of the Heihe River, regional climate changes also exhibit local peculiarities. The significant temperature increase, high evaporation rates, and reduced precipitation of the region have led to a warmer, drier climate, which poses serious threats to vegetation growth.

4.2. Relationship Between Vegetation Dynamics and Climate Change in the Lower Reaches of the Heihe River

Arid regions, among the most vulnerable ecosystems on Earth, display complex interactions between vegetation dynamics and climate change. Rising temperatures can accelerate plant transpiration and increase water stress, while reduced precipitation limits SM replenishment. These changes may lead to reduced vegetation cover, ecosystem degradation, and an increased risk of desertification. The study results show a generally weak correlation between vegetation cover changes and Tem and Pre in the lower reaches of the Heihe River. However, a unique feedback mechanism has developed between regional vegetation and climate. In this extremely arid continental climate, the vegetation—comprising species such as Populus euphratica, Haloxylon ammodendron, Elaeagnus angustifolia, and desert shrubs such as Reaumuria soongorica and Calligonum mongolicum—has adapted to drought conditions [39]. With ongoing warming and intensified evaporation, SM dissipation accelerates, which increases the risk of surface vegetation survival. Some ecosystems in the region have shifted from functional to structural decline, as riparian forests gradually transform into shrub and semi-shrub desert landscapes, and hydrophytic herbaceous plants are replaced by xerophytic species. This underscores the significant impact of climate change on the vegetation ecosystems in the lower reaches of the Heihe River.

As a crucial component of the ecosystem, vegetation changes have feedback effects on regional climate. Research shows that vegetation cover changes can influence climate by regulating the surface energy balance, which alters surface albedo, and affects evapotranspiration [60]. When vegetation cover declines, surface albedo increases, which reduces the absorption of shortwave radiation and creates a cooling effect. However, vegetation degradation also exposes more surface area, intensifies evaporation, and exacerbates regional aridity. Vegetation changes further impact local and regional water cycles through transpiration, with reduced cover potentially lowering transpiration rates and altering precipitation patterns. Therefore, the feedback effect of vegetation cover changes on temperature is a complex process with disputed mechanisms [61]. Nonetheless, it is widely agreed that changes in vegetation cover and precipitation have a mutually reinforcing feedback effect.

5. Conclusions

This study examines the spatiotemporal characteristics of vegetation cover and its response to climate change in the lower reaches of the Heihe River, an arid region in Northwest China. Using MOD13Q1-NDVI data from 2000 to 2022, along with Tem and Pre data for the same period, the main conclusions are as follows:

(1)

The lower reaches of the Heihe River are predominantly characterised by low vegetation cover or areas lacking vegetation. Medium to high vegetation cover is mainly concentrated along riverbanks and in the Juyan Lake region and forms the Ejina River Oasis landscape. From 2000 to 2022, overall vegetation cover in the region exhibits a slow but fluctuating upward trend, with signs of degradation stabilising and showing improvement;

(2)

The ecological stability of vegetation in the core oasis area of the lower reaches of the Heihe River is weak and highly susceptible to climate change and human activities. The vegetation ecosystem exhibits a high degree of vulnerability. In terms of evolutionary trends, vegetation cover has generally improved over the past 23 years. However, future changes are expected to show weak non-persistence, meaning that the region’s vegetation still faces a significant risk of degradation;

(3)

Climate change in the lower reaches of the Heihe River is influenced by global patterns but also exhibits local climatic peculiarities. The region has experienced a notable warming trend, consistent with global and Northwestern China trends, though the rate of warming varies seasonally. Precipitation has shown a slight decreasing trend with highly unstable decadal variations. Overall, the regional climate demonstrates a ‘warming and drying’ trend, which poses severe threats and challenges to vegetation growth;

(4)

The changes in vegetation cover are generally weakly correlated with temperature and precipitation, although temperature has a slightly greater impact on vegetation cover changes. An increase in temperature to some extent promotes the growth of plants along the river, while it inhibits the growth of drought-tolerant desert plants. An increase in precipitation can also promote plant growth to a certain degree. However, the evaporation rate in the study area far exceeds the amount of precipitation, and factors such as surface water, groundwater, and soil salinity are more sensitive to plant growth. Therefore, the contribution of the area’s minimal rainfall to plant growth is relatively weak.