1. Introduction
Since the beginning of the 21st century, the combination of human activities and natural processes has caused significant changes in land use and global ecosystems. This has brought attention to the ecological risks associated with changes in land use [
1,
2], which are a major concern for both developed and developing nations [
3]. Human alterations to land use patterns in recent decades have significantly disrupted ecosystems, resulting in ecological functional degradation, soil erosion, land desertification, environmental pollution, and diminishing biodiversity [
4,
5]. These alterations have markedly increased ecosystem risks [
6] and pose severe threats to human well-being [
7,
8,
9,
10]. Assessing ecological risks associated with changes in land use and identifying their causes is essential for establishing an ecological risk warning system. This system can help accurately and effectively control ecological risks, guide human behavior, and provide a scientific basis for ecological construction.
The concept of ecological risk assessment originated in the United States Environmental Protection Agency (EPA), which defines ecological risk as the probability of adverse ecological impacts resulting from exposure to one or more stressors [
10]. Since the 1990s, ecological issues have become increasingly prominent, leading to a shift in the focus of risk assessment from human health to ecological risk assessment [
11]. This shift has extended to populations, communities, and entire ecosystems. Ecological risk is typically defined as the probability and magnitude of adverse effects on ecosystem structure, function, stability, and sustainability caused by external stressors [
12]. As research scales expand from local to regional, a significant branch of ecological risk research has emerged, known as regional ecological risk assessment. This primarily evaluates the probability and extent of adverse effects of environmental pollution, human activities, or natural disasters on multiple risk receptors at the regional scale [
13]. Currently, ecological risk sources include natural variability and human activities. Within the domain of natural disasters, a theoretical and methodological framework for ecological risk assessment has been established [
14,
15]. Changes in land use can reflect the impact of human activities on natural ecosystems [
16,
17]. Such changes can have significant ecological impacts on atmospheric, soil, aquatic, and biological systems [
18]. These impacts are cumulative and can have regional effects.
The assessment of ecological risks associated with changes in land use primarily focuses on urban areas, watersheds, coastal regions, administrative districts, and nature reserves. In urban areas, frequent conversions among different land use types due to excessive population density and rapid spatial expansion have resulted in increasingly prominent ecological issues [
19,
20]. Watershed areas are ecologically sensitive due to their poor natural endowments, particularly water resources. The landscape in these areas experiences both ecological improvement and deterioration. However, the deterioration trend outweighs the improvement trend, leading to an intensification of landscape ecological deterioration [
21,
22]. During the development of coastal areas, uncontrolled urban expansion in the early stages of economic development can lead to landscape fragmentation, which increases overall ecological risk and tends to concentrate spatially. However, as expansion is restrained, overall ecological risk decreases [
23,
24]. Establishing conceptual models for ecological risk assessment and determining quantitative assessments of ecological risk present significant challenges due to the complexity of ecosystems and the uncertainty of risk occurrence.
Two mature theories and methods currently exist in the field of ecological risk assessment. The first is the traditional assessment approach based on the source–pathway–receptor theory, known as “source analysis–receptor assessment–exposure and hazard assessment–risk characterization’ [
25]. The Relative Risk Model (RRM) is widely employed for conceptual model construction within this framework. For example, Muditha K. and Heenkenda used this assessment system to rank and classify pressure sources and habitats in a specific area. They modeled the interactions between them using exposure and effect filters, revealing the spatio-temporal distribution of ecological risks in a port setting [
26,
27]. Yu et al. [
28] constructed a “source–pressure source–ecosystem–ecological adverse endpoint” assessment framework based on the principles of RRM to predict and rank the potential ecological risks of various sub-regions in Xiamen Bay. In comparison to RRM, the Landscape Pattern Risk Assessment and Evaluation (LPRAE) method is more commonly used to assess single risk sources. LPRAE quantitatively evaluates ecological risk by assessing the likelihood of ecological risk occurrence within a region and the losses caused by risk sources to risk receptors [
29]. Another method involves ecological risk assessment based on landscape ecology theory. This method emphasizes the impact of landscape patterns on specific ecological functions or processes and focuses on the overall loss of landscape in providing ecosystem services and ecological functions [
30,
31]. Evaluation indicators include landscape fragility, resilience, and stability. Research outcomes emphasize the comprehensive characterization and spatial visualization of multiple risk sources, supporting sustainable landscape planning, design, and ecological management. Landscape ecological risk assessments often overlook the structure and functionality of the ecosystem, focusing solely on the landscape perspective. Despite the fact that the landscape represents only a small portion of the entire ecosystem, assessments are typically static, failing to account for the dynamic nature of land use changes. As a result, landscape ecological risk assessments do not adequately capture the fluctuations in ecological risks associated with land use changes. Current risk assessment methodologies only consider the impact of baseline land use changes, without exploring the potential transitions between different land use types and ignoring future risks. Therefore, a new approach is needed to evaluate the ecological risks associated with land use changes, capable of addressing the dynamic alterations in land use and the complex nature of ecosystems.
The Tarim River Basin (TRB) is located in an extremely arid region of China and is heavily influenced by dry climates and intense human activities. As a result, significant alterations in ecological processes have occurred, leading to a heightened ecological vulnerability (EV). Furthermore, rapid urbanization and extensive cropland development in the region have resulted in dramatic changes to the ecosystem. The study’s logical framework focuses on analyzing and simulating land use changes, assessing ecosystem service functions, and spatially identifying and evaluating ecological risks. To achieve this, we coupled the PLUS and Invest models [
32] to simulate potential future scenarios of land use changes in the TRB under three development scenarios for the year 2035: baseline development, economic development, and ecological conservation. The economic Sharpe ratio [
33] is subsequently introduced to integrate land use simulation results and ecosystem services into ecological risk assessment. This study provides empirical evidence for ecological risk assessment in the TRB by analyzing the spatiotemporal differentiation of ecological risks associated with land use changes under the three development scenarios for 2035 and exploring their attribution. It also offers technical support and a multi-level research approach for similar study areas.
4. Results
4.1. Analysis of the Evolution of Spatial and Temporal Patterns of Land Use
From 2000 to 2020, the land use types in the study area were predominantly unused land, followed by grassland and cropland, with comparatively fewer areas covered by grassland, water, and construction land (refer to
Figure 3). Unused land is concentrated mainly in the central part of the basin, belonging to the Taklimakan Desert region. Cropland is primarily distributed in the central-western and southwestern central zones, while forestland is concentrated in the Awat area in a strip-like pattern. Grassland predominantly occupies the northern and southern edge areas, with substantial area coverage, while water and construction land are relatively dispersed.
Table 4 shows that from 2000 to 2020, cropland, construction land, and unused land increased, while forestland, grassland, and water decreased. The proportion of unused land increased significantly by 1.72%, and cropland increased by 1.54%. In contrast, the grassland area experienced the most significant decrease, with its proportion declining by 1.87%. This was followed by a decrease in water proportion by 1.45% and forestland area proportion by 0.10%. However, the area of construction land only saw a marginal increase of 0.17%. Therefore, from 2000 to 2020, there were minimal changes in the area of construction land and forestland.
4.2. Spatial Increase or Decrease in Different Land Use Types
The study analyzes the spatial dynamics of different land use types during historical periods, showing the areas of increase and decrease for each category from 2000 to 2020.
Figure 4 illustrates that cropland has increased mainly in the northwest of Kashgar and the central region of Awat, with a relatively concentrated distribution. Conversely, the areas experiencing loss of cropland are sparse and scattered. Regions with increased forestland exhibit a sporadic distribution patterns, whereas areas with decreased forestland are primarily concentrated in the central region of Awat. Increased grassland areas are sporadically distributed along the southern fringe, whereas areas of decrease are mainly concentrated in the northwest of Kashgar, the central region of Awat, and the northern part of Manas, displaying a relatively concentrated distribution. There are no substantial changes observed in water. Areas of increased construction land are predominantly distributed in the eastern part of Korla, while areas of decrease are primarily concentrated in the northern region of Korla, exhibiting a relatively concentrated distribution. Regions of increased unused land are mainly situated in the northern part of Korla, with a relatively concentrated distribution, whereas areas of decrease in unused land are predominantly characterized by sparse and scattered distribution patterns.
4.3. Multi-Scenario Model of Land-Use Change
Using the 2020 baseline image, and based on the expansion changes in land use between 2010 and 2020 in the study area, three scenarios of land use in the TRB for 2035 were simulated (refer to
Figure 5).
Compared to 2020, according to the statistical data under the baseline development scenario, the area of cropland increased by 15,976.63 km2, with a growth rate of 2.46%; the area of water increased by 945.19 km2, with a growth rate of 0.13%. Meanwhile, the areas of forestland, grassland, and unused land decreased by 480.31 km2, 13,352 km2, and 2461.81 km2, respectively, with reduction rates of 0.12%, 2.64%, and 1.40%. Under the economic development scenario, the areas of cropland, water, and construction land increased by 24,545.79 km2, 996.13 km2, and 702.31 km2, respectively, with growth rates of 2.95%, 1.40%, and 0.21%; whereas the areas of forestland, grassland, and unused land decreased by 499.90 km2, 19,622.50 km2, and 6121.84 km2, respectively, with reduction rates of 0.13%, 3.00%, and 1.37%.
Figure 5 depicts the expansion of cropland towards the northwest in the Kashgar area under the baseline development scenario (
Figure 5b). The encroachment of cropland and unused land on grasslands is significant, particularly in the northwest region. Unused land primarily expands towards the marginal zones, while changes in construction land, forestland, and water are less pronounced. Under the economic development scenario (
Figure 5d), cropland expands more actively in the northwest of Kashgar, the central region of Awat, and the northern region of Korla. In particular, there is extensive encroachment of grassland by cropland in the northwest of Kashgar and the central region of Awat. Unused land mainly encroaches upon grassland at the margins and continues to expand towards the periphery, while construction land predominantly exhibits scattered patchy distributions in the northern region. Under the economic development scenario (see
Figure 5c), the expansion of cropland is notably active in the northwest of Kashgar, the central region of Awat, and the northern region of Korla. Extensive encroachment of cropland into grassland is observed in the northwest of Kashgar and the central region of Awat. Forestland expansion is concentrated in the eastern region of Awat and the western region of Korla, displaying patchy distributions. Water exhibits more active expansion in the southwest of Hotan and the northern region of Yining.
4.4. Characterizations of Ecosystem Services under Different Model Scenarios
The total ecosystem service quantity in the study area exhibits significant spatial heterogeneity, showing a general increasing trend in the northern and southern marginal zones. Specifically, under the baseline development scenario, the total ecosystem service quantity decreased by 15.247% compared to 2020, with a decrease of 3.818% in water yield, 0.02% in habitat quality, and 0.061% in carbon storage, and an increase of 0.036% in soil retention. Under the economic development scenario, the total ecosystem service quantity decreased by 13.358% compared to 2020, with a decrease of 4.81% in water yield, 0.253% in habitat quality, and 0.005% in carbon storage, and an increase of 0.451% in soil retention. Under the urban expansion scenario, the total ecosystem service quantity decreased by 19.852% compared to 2020, with a decrease of 3.628% in water yield and 0.516% in habitat quality, an increase of 0.011% in carbon storage, and a decrease of 7.113% in soil retention.
There is significant regional variation in ecosystem services in the study area, with ecosystem service conditions in the marginal areas significantly better than those in the plains (see
Figure 6). Specifically, significant changes are observed in the northwestern, southern, and western parts of the study area, while minimal changes are noted in the central and eastern regions, with the lowest values of ecosystem service functions. Particularly, the spatial distribution of soil retention function predominantly exhibits low values, with relatively higher values observed in the northern and western marginal areas. Water yield appears to be clustered in large areas of high-value regions in the northern area, while the spatial distribution characteristics of habitat quality and carbon storage remain consistent with the overall ecosystem services, showing a pattern of high values in the periphery and low values in the central region.
4.5. Characteristics of Changes in the Distribution of Ecological Risks under Different Scenarios
Based on the statistics of changes in risk zones at various levels, it is clear that in 2035, the ecological risk level structure in the study area will be dominated by medium risk, with other risk levels accounting for a small proportion of the total area. It is important to note that this analysis is objective and does not include any subjective evaluations. The medium-risk zones cover the largest and most widespread area, accounting for over 97% of the total ecological risk level area. In contrast, the low-risk zones are the smallest, each accounting for less than 0.2%. Under the baseline development scenario, the area distribution of high-risk levels is the widest, accounting for 1.14%. Meanwhile, the area of medium-risk levels reaches its maximum under the economic development scenario, accounting for 98.17%, followed by the baseline development scenario, accounting for 98.03%. In the medium-high risk levels, the economic development scenario increases by 0.36% compared to the ecological protection scenario. Meanwhile, the ecological protection scenario has the largest proportions of low-risk and medium–low-risk levels, accounting for 0.15% and 1.32%, respectively, with increases of 0.12% and 1.05%.
The spatial distribution of risk zones remains consistent across the three scenarios, mainly dominated by medium risk (see
Figure 7). However, there are significant changes in the medium–low-risk and high-risk levels. Under the baseline development scenario (see
Figure 7a), high-risk areas are primarily located in the northwest of Kashgar, the central part of Awate, the central part of Hotan, and the northwest direction of Korla. The widest distribution of high-risk areas is in the northwest of Kashgar, while other regions exhibit scattered and patchy distributions. In the economic development scenario (see
Figure 7b), the main changes occur in the medium–high-risk and medium–low-risk aspects. Medium–low-risk areas are concentrated mainly in the northwest of Kashgar and the central part of Awate. Medium–high-risk areas are mainly distributed in the central part of Kashgar, the north-central part of Awate, the northwest of Korla, and the central part of Hotan. Other areas exhibit scattered patchy distributions. Regarding ecological protection (see
Figure 7c), significant changes occur mainly in the medium–low-risk level. These changes are mainly distributed in Atushi, the central and northern parts of Kashgar, the northern part of Awate, the central and southern parts of Hotan, and the central part of Korla. The distribution in Kashgar is the most concentrated, while the high-risk level is mainly distributed in scattered patches, mainly concentrated in the Kashgar and Awate areas.
5. Discussion
5.1. Spatial Heterogeneity of Ecological Risk Indices and Their Formation Mechanisms
Land is a crucial element for socio-economic activities, and serves as a tangible representation of human development and utilization of the natural environment. Changes in land structure and patterns are closely related to the spatiotemporal distribution of ecological risk. By assessing the spatial patterns of ecological risk, we can reveal the impacts of land use changes on the structure and function of ecosystems. In the TRB report of 2035, the spatial distribution of ecological risk remained largely unchanged across all three scenarios. The risk tends to increase from the central part of the basin towards the periphery, with medium risk being the most common. The study area is characterized by peripheral regions with high- and medium-high risk areas, while low- and relatively low-risk areas are mainly distributed around the basin. The most extensive distribution of areas with medium risk is observed. Unused land is the predominant land use type in the study area, especially in the basin’s central region, which includes the Taklamakan Desert, China’s largest and the world’s tenth-largest desert, as well as the second-largest mobile desert globally. As a result of the land use types, the overall ecosystem services in the basin’s central region are relatively low [
41]. The transitional zones adjacent to the desert have extensive grasslands and forests. However, these areas are experiencing a decrease in forest, grassland, and water due to the encroachment and expansion of unused land [
42]. As a result, ecosystem services are declining. In this environmentally harsh region, the types of land available for human development and utilization are relatively limited. Therefore, the expansion of construction land will not be significant over the next 20 years, although it will be relatively concentrated in distribution.
Under the baseline development scenario, land use changes mainly follow historical developmental trends. The expansion of cropland and unused land predominates, while there is a trend of decreasing ecological land areas such as forests, grasslands, and water. Construction land experiences concentrated expansion, with relatively insignificant land encroachment. However, the proportion of cropland and unused land encroaching on other ecological areas increases, leading to significantly reduced vegetation cover, sparse vegetation, severe soil erosion, and intense desertification. Consequently, ecosystem service indices, such as water yield, habitat quality, and carbon storage, are relatively low, while soil conservation service indices increase. Under the baseline development scenario, the transitional zones at the periphery of the basin present higher ecological risks. Therefore, it is necessary to control agricultural and ecological spaces. Protection measures should be improved in the agroforestry transition zones, as well as in the grassland and unused land transition zones, to enhance forest and grass cover and strengthen ecosystem stability. In the scenario of economic development, rapid urbanization and agricultural expansion worsen human activities, which significantly contribute to the decline in regional ecological environment quality [
43]. The expansion of construction and cropland increases the size of areas classified as medium- to high-risk. Urbanization accelerates soil erosion and water loss, while cropland expansion encroaches upon other ecological land, resulting in decreased grassland and forest areas. This threatens ecological balance [
44]. Therefore, ecological risks in transitional areas between cropland and grassland or forest are worsened. In the ecological protection scenario, the proportion of cropland, forest land, and water increases, resulting in significant changes in areas classified as low-ecological-risk and medium–low-ecological-risk. These changes are mainly concentrated in the transitional zones between grassland and cropland, which are characterized by high vegetation cover, high organic matter content in soil, and fertile soil, thus exhibiting strong ecosystem service functions. However, the study area’s unique location and fragile ecological environment exacerbate the pressures on cropland, forest land, and unused land due to intensified human disturbances. Therefore, it is crucial to nurture mountain vegetation and water resources in this region for sustainable development.
Simulations of multiple future scenarios indicate an urgent need to optimize land use structures to provide decision support for sustainable development and high-quality ecological environments. Considering the extensive desertification in the hinterland of this basin, which may exacerbate desertification in peripheral areas, it is imperative to rationally plan urban development boundaries, limit uncontrolled expansion, and improve land use efficiency at the same time. Urban development boundary planning is significant in mitigating the exacerbation of desertification and improving land use efficiency [
45]. To enhance the ecological benefits of forest and grassland areas with low to medium–low ecological risks, it is essential to promote the positive feedback evolution of forest–grassland ecosystems and urgently strengthen land remediation efforts. The goal is to intensify cropland and grassland use, limit the disorderly expansion of unused and construction land, and harmonize relationships among various ecological landscapes. This will reduce ecological risk levels and enhance the stability of the land ecosystem. Additionally, measures can be taken to increase soil carbon sequestration and fertility potential. This can be achieved by actively implementing ecological protection projects [
46], optimizing species richness, enhancing ecosystem stability and resilience, and promoting synergies among ecosystem services. These actions can enhance ecosystem productivity and sustainability [
47,
48].
The TRB is a crucial ecological barrier in western China. The degradation of its ecological environment poses significant challenges to the high-quality development of ecological civilization. The rapid development and transformation of urban construction have inevitably altered land use patterns and functions, leading to ecological risks. In the core area of the study region, the activation speed of marginal sand dunes has accelerated, posing a severe threat to cropland. This is particularly concerning as artificial oases continue to expand, replacing natural ones. As a result, the buffer zone between oases and deserts is continuously shrinking, which has a negative impact on desert-edge vegetation. Therefore, to implement the concept of guided restoration, it is essential to create artificial ecosystems in the Taklamakan Desert by constructing artificial oases. Additionally, adjusting the industrial structure and optimizing the allocation of agriculture, forestry, animal husbandry, and subsidiary industries can help mitigate the risks of vegetation destruction caused by human activities. Under the conditions of economic and technological priority, broad-scale afforestation is an effective method for protecting vegetation on the edges of deserts. Therefore, it is necessary to strictly adhere to policies and regulations to protect the ecological environment, mitigate the negative impacts of human activities on the environment, and promptly initiate ecological restoration efforts.
5.2. Comparison with Previous Research
Currently, the ecological risks of land use changes in the TRB have not been assessed. This study integrates simulated land use results and ecosystem services into ecological risk assessment by introducing the economic Sharpe ratio [
33].
Extensive research has been conducted on the ecosystem services of the TRB. The ecological security assessment, based on ecological footprint, indicates that water scarcity is a significant constraint on the socio-economic development of the TRB. This is exacerbated by climate change and rapid cropland expansion. Furthermore, the significant growth of artificial ecosystems, referred to as artificial oases, resulting from the transformation of natural oases or deserts, worsens landscape fragmentation and the ongoing degradation of ecological security levels in the basin [
49]. The increase in ecological risks due to water scarcity is consistent with the findings of this study, primarily resulting in a gradual rise in ecological risks in the transitional zones between cropland and forests. The TRB indicates an overall increase in ecosystem vulnerability levels, with some areas experiencing extremely severe vulnerability [
50]. This finding is consistent with the gradual rise in ecological risk levels discussed in this study. The evaluation of ecological risks resulting from ecosystem degradation shows a significant correlation between these risks and the swift urbanization and expansion of cropland, which are the primary drivers of ecological risks in the TRB [
13].
The study shows that the spatial distribution of ecological risk resulting from land use change in the TRB is closely aligned with the spatial distribution pattern of ecosystem service functions in the TRB after the implementation of ecological restoration projects, as described in other studies [
51]. This suggests that human activities significantly influence land use change-induced ecological risks.
5.3. Shortcomings and Prospects
The PLUS model’s Markov module relies exclusively on past land use changes for quantitative prediction. However, accurately predicting the transfer of land use types in the future is challenging due to the influence of multiple factors, limiting the model’s accuracy. To enhance accuracy, it is advisable to comprehensively consider policy and natural economic factors for precise quantitative analysis. Furthermore, the spatiotemporal evolution of ecological risk is influenced by multiple factors, which requires additional research. To clarify the impact of land use change on ecological risk, longer time series studies are necessary, along with a thorough investigation of the relationship between ecosystem services and ecological risk. In summary, future research on the ecological risk of land use change should consider multiple variables, including social, economic, and environmental factors, and conduct more empirical studies. Furthermore, ecological protection should be a dynamic adaptive response, focusing on the trends of land use change in high-risk areas, and intensifying land restoration efforts to enhance resilience against ecological risks.
6. Conclusions
(1) Between 2000 and 2020, the primary land use types in the study area were unused land, followed by grassland and arable land. Grassland, water, and construction land had relatively less coverage. There was a significant increase in unused and arable land, while grassland, forest land, and water exhibited a declining trend. From 2020 to 2035, the main trend in the three development scenarios was the expansion of arable and unused land, while the grassland area decreased significantly. There were minimal changes in other land types. Unused land was mainly clustered in the central part of the basin, while arable land was concentrated in the central areas of Kashgar and Awat. The Awat region had a strip-like pattern of forest land, while the northern and southern marginal areas had mainly grassland. Water and construction land areas were relatively dispersed in comparison.
(2) Between 2020 and 2035, the total quantity of ecosystem services in the study area showed significant spatial differentiation under various scenarios. There was an overall increasing trend in the total quantity of ecosystem services in the northern and southern marginal areas. Under the baseline development scenario, the total quantity of ecosystem services decreased by 15.247% compared to 2020. Under the economic development scenario, it decreased by 13.358% compared to 2020. Under the ecological protection scenario, it decreased by 19.852% compared to 2020.
(3) Between 2020 and 2035, various scenarios showed that ecological risk levels had similar characteristics. The majority of the areas fell under the moderate-risk category, with other risk levels accounting for a smaller proportion of the total area. Moderate-risk areas had the largest and most widespread distribution, covering over 97% of the total ecological risk level area. Conversely, low-risk areas had the smallest area, with proportions all below 0.2%. In various scenarios, the distribution of high-risk areas varied. In the baseline development scenario, high-risk areas were mainly located in the northwest of Kashgar, the central part of Awate, the central part of Hotan, and the northwest part of Korla. Under the economic development scenario, there were significant changes in the distribution of moderate–high-risk and moderate–low-risk areas. Under the ecological protection scenario, there were noticeable changes in the distribution of moderate- to low-risk areas. These changes were mainly concentrated in specific regions, with Kashgar having the most concentrated distribution.