Quantitative assessment of water-related ecological security patterns based on ecosystem service flows: A case study of Jilin Province, China

Summary

The water-related ecological security pattern directly impacts the allocation and utilization efficiency of regional water resources, and its quantitative assessment is a core requirement for water resource management and planning. However, current assessments of water-related ecological security patterns suffer from insufficient dynamic quantitative modeling, oversimplified river network structures, and neglect of variations in cropland types. Therefore, this study employs a comprehensive approach integrating PLUS multi-scenario land-use simulation, InVEST ecosystem service flow assessment, and the MCR model to quantitatively analyze the supply-demand patterns of water ecosystem services in Jilin Province from 2000 to 2032. The results showed: 1. Persistent service deficits exist in the irrigated croplands of Baicheng and Songyuan in the northwest, while the mountainous areas of Tonghua, Baishan, and Liaoyuan in the southeast serve as primary supply units. This study identified 71 ecological corridors and 58 vulnerable nodes. These include 18 important corridors and 9 early warning nodes. 2. All three scenarios maintain the “southeast surplus-northwest deficit” pattern by 2032. Under the natural development and ecological conservation scenarios, early warning nodes clustered toward the northwest and southeast, respectively, primarily distributed along the Songhua River, the second Songhua River, and the surrounding areas of adjacent cities. Under the farmland conservation scenario, early warning nodes exhibited a strip-like distribution along the depth, with overall aquatic ecosystem health declining. 3. Model validation indicates PLUS simulation accuracy reaches 0.92 (Kappa = 0.87). The annual average aquatic production calculated by the InVEST model shows an error of approximately 2.35% compared to measured data, demonstrating reliable results. These findings contribute to optimizing the spatial configuration of aquatic ecosystems and hold significant implications for multi-scale water security assurance and sustainable water resource management.

Graphical abstract

Highlights

• •Established a quantitative analysis for the 23 years
• •Conducting ecological security pattern predictions under various future scenarios
• •Refined river classification to improve corridor and ecological node identification accuracy
• •Refined cropland classification to capture water-use differences among cropland types

Environmental science; Environmental management; Environmental monitoring; Ecology; Environmental resource; Hydrology

Introduction

Water security is the foundation for sustainable development and social stability. The core of water security lies in guaranteeing an adequate, standardized, and stable water supply. It must meet basic human needs, support the development process, and protect ecosystems, while at the same time responding effectively to extreme events such as droughts and floods.^1,2,3,4 The state of water security is affected by multiple factors, including the role of human activities and the constraints of natural factors, such as the specific utilization of land resources, the actual process of regional economic development, as well as the formulation and implementation of various types of management policies, which will have a practical impact on water security in different dimensions, and jointly influence the overall level of water security.^2,5 However, rapid urbanization, industrialization, and population growth, combined with climate change, have led to increasingly serious water-related ecological problems such as droughts, floods, and pollution in many parts of the world. The stable supply of water resources has decreased while the demand gap has widened, posing a serious challenge to water-related ecological security.^6,7 As an important tool for regional ecological protection, water-related ecological security patterns form a spatial network that maintains ecological processes and service functions by identifying and constructing elements such as ecological sources and ecological corridors,^8,9,10 providing a comprehensive perspective that can protect and enhance the structure, functions, and services of regional ecosystems.^11,12 Therefore, the delineation of the water-related ecological security pattern is of great significance for the synergistic development between ecological management and economic construction, and the realization of the sustainable use of water resources.

Early studies on water-related ecological security patterns used quantitative assessment methods such as water balance analysis to conduct macro-level evaluations of regional water supply and demand conditions.¹³ With further research, methods for delineating water-related ecological patterns based on supply and demand relationships have emerged. These methods identify key water sources and water demand areas and analyze their spatial matching to construct water security patterns to ease the conflict between supply and demand.¹⁴ For example, Xiang et al. evaluated the supply-demand relationship of ecosystem services in the northeast region and found that the overall supply and demand in the region were basically balanced, but there were local mismatches.¹⁴ In recent years, the theory of ecosystem service flows (ESFs) has been increasingly integrated into studies on the construction of water-related ecological security patterns to better harmonize ecological processes with human needs^{15,16,17,18,19} and connect supply and demand by quantifying the spatial transmission of services.^20,21 This progress has enabled researchers to identify the “source” and “sink” areas of ecological services and the flow paths between them, revealing the spatial patterns of key ecological processes. The field of effective coordination between water supply service flows and water-related ecological security patterns has also gradually attracted the attention of scholars.^{22,23,24,25,26,27} Chen et al. proposed a methodological framework for incorporating ESF into water-related ecological security patterns, identifying water supply areas, demand areas, and the key corridors and vulnerable nodes connecting the two.¹⁸ Similarly, Guan et al. constructed a “supply-flow-demand” model in a typical watershed to simulate the spatial flow paths and values of water supply services from supply areas to demand areas, and divided the watershed units into types such as catchment areas, transit areas, and outflow areas.¹⁶ Sun and An (2024) constructed a water security pattern for the northeast region based on the “supply-flow-demand” process, achieving an organic connection between water ecological service supply areas, flow processes, and demand areas.²⁴ Overall, research on water-related ecological security patterns is undergoing an evolution from static water balance assessment to dynamic ecological service flow patterns. This methodological shift provides more comprehensive spatial decision-making support for regional water security management.

Although research on water-related ecological security patterns has been developing, there are still some shortcomings in current research. First, existing research is mainly based on static identification and lacks dynamic simulation and quantitative analysis of the evolution process of ecological service patterns, making it difficult to truly reflect the coupling mechanism of water resource systems in the spatial and temporal dimensions. Second, during corridor identification, river network structures are often simplified, failing to adequately consider the hierarchical connectivity and hydrological importance of rivers, thereby affecting the spatial identification accuracy of ecological service flow paths and key nodes. Additionally, in studies on water supply services, most research ignores the core differences in water intensity and ecological services between different types of agricultural land, leading to discrepancies in ecological service estimation and resource regulation. Therefore, there is an urgent need to introduce methods such as process modeling, river network hierarchy classification, land heterogeneity identification, and multi-scenario simulation into the research paradigm to enhance the scientific and practical nature of water-related ecological security pattern research and improve the theoretical basis for the management of regional ecological security and the efficient allocation of water resources.

Jilin Province is located in the heart of Northeast China, featuring diverse topography and terrain, with an overall landscape characterized by “mountains in the east, plains in the central region, and grasslands and wetlands in the west.” It integrates multiple ecosystems such as forests, wetlands, farmlands, and grasslands, exhibiting distinct geographical unit transitional characteristics and ecosystem complexity.^{28,29,30,31,32} The province spans multiple river basins, with complex hydrological processes and significant spatial variation in water resources. The western region has long suffered from drought and water shortages, with restricted agricultural water use, while the eastern region has underutilized water resources, leading to prominent local water ecological conflicts.^{28,33,34,35,36} In recent years, ecological degradation in Jilin Province has become increasingly evident, with wetland areas continuing to shrink and ecosystem service functions declining, primarily driven by human development and climate change.³⁷ At the same time, the eastern Changbai Mountains are facing forest degradation, the western grasslands are shrinking rapidly, and grassland quality is declining, posing challenges to the overall stability of the ecosystem.³⁸ In summary, Jilin Province has typical water resource supply and demand contradictions, significant climate change responses, and human activity impacts. It is representative and exemplary for the study of the evolution of water-related ecological security patterns in the northern middle and high latitudes. It has an important reference value for revealing the dynamics of regional water security pattern under the background of climate change and land use change, and optimizing the spatial management strategy of water resources.

In view of this, this study focused on delineating the water-related ecological security pattern of Jilin Province. Based on the ecosystem service flow theory, the PLUS-InVEST-MCR model coupling method was used to study the impact of climate and land use type differences on the water resource supply and demand pattern, as well as the roles and changes of different cropland types in water resource distribution, ecological service provision, and water-related ecological security patterns. The study also revealed changes in water resource pressure and ecological corridors and vulnerable nodes under different development scenarios. The research findings are of significant importance for optimizing regional water resource management, enhancing water resource utilization efficiency, and formulating sustainable water resource policies. Overall, our objectives are: (1) to conduct a quantitative analysis of the evolution of the water-related ecological security pattern in Jilin Province from 2000 to 2022 and under three different development scenarios by 2032--a total of 26 years. This will be achieved by incorporating river classification regulations and refining the differentiation of cropland types to enhance the accuracy of corridor and ecological node identification; (2) to propose optimization schemes and implementation pathways for systematic water resource management based on the identified issues.

Study area

Jilin Province is located in the central part of Northeast China, bordering Heilongjiang Province to the north, Liaoning Province to the south, Inner Mongolia Autonomous Region to the west, and Russia to the east.³⁹ To the southeast, it is separated from North Korea by a river. It is the geographical center of Northeast Asia, spanning 121°38′to 131°19′E and 40°50′ to 46°19′N, with a total area of approximately 187,400 km².^28,40 The central and western parts of the province consist of low-lying terrain such as the Songnen Plain and the Liaohe Plain. Approximately 36% of the province is mountainous, 30% is plains, and the rest is terraces and hills.^40,41 Jilin Province has a temperate continental monsoon climate. The four seasons are distinct, with rainfall and heat occurring in the same season. The average annual temperature is approximately 2°C–6°C, with lower temperatures in mountainous areas and higher temperatures in plains. The frost-free period in the province generally lasts between 100 and 160 days.^28,40,42 The annual average precipitation is approximately 400–600 mm, with 80% concentrated during the summer months of June to August. The eastern mountainous regions receive the highest precipitation. The province is crisscrossed by rivers, with the five major river systems (Songhua River, Yalu River (Hunjiang section), Dongliao River, Xiliao River, and Tumen River) intersecting within its borders. The annual total water resources amount to approximately 37.09 billion cubic meters. For a regional overview, please refer to Figure 1.

Figure 1.

Overview of the study area (The map depicts China using the officially approved map review number GS (2024)0650 issued by the Ministry of Natural Resources, with no modifications made to the base map)

Figure 2 presents the technical workflow diagram for this study.

Figure 2.

Technical roadmap

Results

Land use simulation and prediction

Land use changes in Jilin Province from 2000 to 2022

This study is based on land use data obtained through remote sensing monitoring, with the original data comprising 35 land categories. To highlight the focus of the analysis, refine the differences between rainfed croplands and irrigated croplands, and simplify the overall classification system, the data were reclassified into nine land categories: rainfed croplands, irrigated croplands, forest land, shrubland, grassland, wetland, construction land, bare land, and water bodies. The figure shows that Jilin Province is mainly composed of three types of land: rainfed croplands, irrigated croplands, and forest land, which account for more than 90% of the total area. In addition, the land use change string diagram and the rise and fall line chart show that: ① From 2000 to 2008, land use in Jilin Province was generally stable, but some land types showed a clear trend of transformation. Among them, rainfed croplands and shrubs showed the strongest inertia, with year-on-year decreases/increases of less than 1%. The area of grassland increased significantly, mainly due to the conversion of rainfed croplands, irrigated croplands, and bare land, with a net increase of 1,994.01 km². The expansion of grassland during this period was closely related to the ecological restoration projects, such as “returning cropland to forests (grass)” implemented during the same period, reflecting the positive conversion of cropland land to ecological land driven by policy. ② From 2008 to 2016, land use types changed steadily, with bare land remaining the most active land conversion type and showing a large-scale reduction trend. However, the volume of farmland converted to other land types remained large. Among them, the area of rainfed croplands converted to construction land reached 1,606.95 km². This reflects the occupation of cropland resources by new urbanization construction. Irrigated croplands were extensively converted into grassland, with a conversion area of approximately 2,479.64 km², possibly related to measures such as wetland protection, returning cropland to wetlands, and agricultural structural adjustments. Forest land remains the largest land category in the province, with almost no outward conversion; water areas continue to expand, indicating the strong stability of the ecosystem. ③ From 2016 to 2022, land use changes across Jilin Province showed an increasing trend. Bare land and wetlands exhibit high mobility, with wetland areas converting to water bodies reaching up to 1,567.27 km², serving as the primary source of water body expansion. Meanwhile, forest land stability has improved, with a self-sustaining area of 153,850.59 km². Water bodies have seen a significant net increase, reaching 4,550.60 km² by 2022, thanks to the comprehensive implementation of water management policies such as the “River and Lake Chief System” and strict ecological protection red lines.

Multi-scenario land use simulation

This article simulates the land use pattern of Jilin Province in 2032 under three different scenarios (natural development, cropland protection, and ecological protection) using the PLUS model by adjusting the transition probability matrix, neighborhood weight parameters, and land demand forecasts. The simulation results were then compared and analyzed with the actual land use data from 2022, with the relevant results shown in Figure 3C. Overall, different development scenarios had a significant impact on the land use pattern, showing a trend of diversified spatial patterns and land use structures. Under natural development scenarios, by 2032, the forest area in Jilin Province will decrease slightly, while the area of rainfed croplands and water bodies will show an upward trend. Among them, although the area of rainfed croplands will increase by 2,213.07 km², it will only account for 0.02% of the original area, indicating that its expansion will be relatively slow. The area of water bodies will increase significantly by about 14%, showing the spatial succession characteristics of “compensating for development with rainfed croplands and expanding ecology with water bodies” driven by urban expansion. Under the cropland protection scenario, the land use transfer string diagram shows that the area of rainfed croplands and irrigated croplands in Jilin Province has increased significantly, with new cropland mainly converted from forest land, grassland, wetlands, and some construction land. Among them, rainfed croplands increased by 3,163 km², the largest increase among all land types, showing a trend of continuous expansion of agricultural space under the idea of “determining land use based on cultivation.” Under the ecological protection scenario, the area of wetlands and grasslands has expanded significantly, with a net increase of 1,587 km² in grassland area, making it the land type with the largest increase. At the same time, the area of rainfed croplands and irrigated croplands has decreased by a total of 1,187 km², showing a clear trend of cropland returning to natural ecological land use and demonstrating a clear ecological priority development orientation.

Figure 3.

Group of images shows land use transfer and change trends

(A) land use maps for the years 2000, 2008, 2016, and 2022; (B) land use projections for 2032 and transition patterns (2022–2032) under three scenarios; and (C) line chart of land use changes during 2000–2008, 2008–2016, and 2016–2022.

Quantification of water ecosystem service supply and demand

Quantitative assessment of water supply and demand in Jilin Province

Based on a spatial analysis of average water supply and demand in Jilin Province from 2000 to 2022, as shown in Figure 4, water supply is relatively high in the central and southern regions of Jilin Province, particularly in the Changbai Mountain area, which boasts favorable natural conditions, abundant precipitation, and rich water resources. In contrast, the western region of Jilin Province, characterized by a semi-arid climate, has relatively lower water supply. In terms of water demand, areas with high water demand are mainly concentrated in the central urban agglomeration and the irrigated cropland area in the northwest. Rapid urbanization, especially in Changchun, has led to a sharp increase in urban water consumption, while irrigation demand in irrigated cropland areas is also high. The spatial distribution of these high level of water demand areas highly overlaps with population density and economic development levels, reflecting the rigid constraints of human activities on water demand.

Figure 4.

Spatial patterns of average water production and water demand during 2000–2022

As shown in Figure 5, the annual average water production in Jilin Province exhibits a change curve characterized by “low-level fluctuations—mid-term peak—temporary decline—and subsequent recovery,” reflecting the high sensitivity of regional water resource supply to climate variability, extreme weather events, and changes in the spatiotemporal patterns of precipitation. In contrast, water demand changes exhibit more pronounced phased characteristics: driven by accelerated urbanization, increased industrial water demand, and intensified agricultural irrigation, water demand has steadily grown, reflecting the increasing reliance of socio-economic development on water resources. However, since 2014, water demand has shown a continuous decline trend, which to some extent indicates that Jilin Province has begun to see initial results from implementing water-saving society construction, water pricing mechanism reforms, and the promotion of agricultural water-saving technologies.

Figure 5.

Temporal variation in water supply and water demand during 2000–2022

The simulation results of the three development scenarios for 2032 are shown in Figure 6. The results indicate that, despite differing policy orientations, the spatial pattern of water resource production in Jilin Province remains stable, maintaining the ecological-geographical pattern formed over the past two decades. Particularly in the eastern mountainous regions and the densely river-networked southern areas, due to high vegetation coverage and favorable hydrological conditions, water production capacity remains prominent, with peak water production volumes maintained at approximately 221 × 10⁸ m³/yr (year), showing a slight increase, reflecting the sustainability and resilience of the ecosystem’s water regulation and storage functions. In contrast, changes in water demand patterns are more sensitive. The peak has decreased to 280 × 10⁸ m³/yr. This change is not only constrained by adjustments to urban water use intensity but is also closely related to improvements in irrigation efficiency in western farmlands. From a spatial perspective, urban areas remain the core of water demand intensity, while agricultural areas form a band of secondary high values. Mountainous forest areas, due to their lower development levels and better water conservation capabilities, exhibit generally low water demand characteristics. Overall, the future water resource supply and demand pattern in Jilin Province will be characterized by “stable supply and varying demand.” Spatially, this will manifest as the coexistence of high-yield resource areas and high-demand areas, further highlighting the importance of scientifically allocating water resources and optimizing the ecological spatial pattern.

Figure 6.

Water production and demand distribution in Jilin Province under three different scenarios

Quantitative assessment of the water supply-demand ratio in Jilin Province

The spatial distribution pattern of water resource supply and demand essentially reflects the coupling relationship between natural geographical conditions, regional hydrological processes, and human land use behavior. As shown in the figure, water resource surplus areas are mainly concentrated in mountainous forest areas with significant terrain variations, high vegetation coverage, and abundant precipitation, which possess stable water production capacity and good water conservation functions. Water-surplus areas are primarily distributed along major rivers such as the Second Songhua River, Nenjiang River, and Toudaojiang River, benefiting from dense river networks, stable flow rates, and high natural water supply intensity. In contrast, water-deficient areas (ESDR <0) are mostly found in the arid agricultural areas of northwestern Jilin Province, concentrated in Baicheng, Songyuan, Siping, and the plains of central Changchun. These areas have high land development intensity and high irrigation demand, coupled with relatively insufficient annual precipitation, leading to a prominent supply-demand imbalance and a long-term water deficit.

From a time-series perspective, the relationship between water supply and demand in Jilin Province from 2000 to 2022 exhibits distinct phased characteristics. Prior to 2015, the regional water shortage fluctuated at high levels due to the interplay between expanding agricultural and industrial water demands and the effectiveness of water conservation measures. The year 2015 marked a turning point. With the deepening of the national ecological civilization strategy and the implementation of the strictest water resource management system, significant progress was made in optimizing industrial structures and enhancing ecological protection. Consequently, the area affected by water shortages entered a sustained decline, and the supply-demand imbalance began to ease. Meanwhile, according to the results of multiple scenario simulations, the spatial pattern of water resource supply and demand in Jilin Province in 2032 will generally continue the current characteristics, basically maintaining the pattern of “surplus in the southeast and deficit in the northwest.” However, there are obvious differences in the evolution trends of deficit areas under different scenarios: under natural development scenarios, continuing the current management policy will help to continuously alleviate water resource pressure, and the deficit area will further shrink, showing that the current system has a certain regulatory effect; under cropland protection scenarios, the expansion of agricultural land and the increase in irrigation intensity will significantly increase water demand, leading to a significant expansion of the deficit area compared with 2022 and a significant increase in regional water security risks; In contrast, the ecological protection scenario, by strengthening the protection of ecological land such as forests and wetlands and limiting human activity interference, effectively reduces water demand intensity, leading to a further reduction in deficit areas and a moderate expansion of surplus areas, with the overall water resource supply-demand structure evolving toward a more stable and sustainable direction.The distribution of the ecological supply–demand ratio is shown in Figure 7.

Figure 7.

Distribution of the ecological supply and demand ratio

Quantification of water-related ecological security patterns

To identify key ecological zones within the water resource supply-demand structure of Jilin Province, this study employs regional ESDR to identify water ecological “source-sink” zones and incorporates land use structure analysis to further elucidate their ecological foundations and functional differences. ESDR is a critical indicator for assessing the balance between water resource supply and demand at the regional level, where ESDR >0 indicates a supply surplus, ESDR <0 signifies demand exceeding supply, and ESDR = 0 denotes supply-demand equilibrium. To exclude the interference of fragmentation, a patch size threshold of 100 km² was set. Through the cluster analysis of ESDR data from 2000 to 2022, areas with significant water resource surplus service capacity were identified as water ecological “source” zones, while areas with ESDR <0 were defined as “sink” zones.

In terms of spatial distribution, the “source” areas are mainly concentrated in the southern mountainous regions and densely river networked areas of Jilin Province, which are predominantly forested. Their excellent ecological conservation and hydrological regulation functions enable them to provide a stable water supply. Time-series analysis shows that since 2003, the area of forested land in the source areas has been steadily decreasing, reflecting the significant impact of human development and expansion on the ecosystem. However, with the promotion of ecological protection policies in recent years, the forest area has shown a rebound trend in 2022. Three simulated scenarios further validate the impact of different development pathways on the water supply function of forested areas: Under the natural development scenario for 2032, the forest land in the source region will recover to the 2006 level. Under the ecological conservation scenario, the forested land will increase further, reaching the highest level recorded over the past 23 years, highlighting the positive role of ecological governance in enhancing water supply security. Forest lands and rainfed croplands jointly constitute the two primary water source areas within the study region, contributing equally to the water supply. Rainfed croplands are predominantly distributed in the Second Songhua River Basin in central Jilin Province. This area has developed a local water supply surplus due to its favorable irrigation and drainage conditions and relatively abundant water resources. Under the cropland protection scenario, the area of rainfed croplands in the water source region expands significantly, further encroaching upon forest lands. This reflects the potential threat posed by cropland expansion to aquatic ecosystems. In contrast, the “sink” area is mainly concentrated in the highly irrigated agricultural areas in western Jilin Province and the central urban belt, where irrigated cropland accounts for 60%–70% of the total land area, rainfed cropland accounts for about 20%, and the rest is grassland in the west. Grasslands are mostly distributed in semi-arid areas with low precipitation, sparse vegetation, and weak natural water conservation functions, making it difficult to provide an effective water supply. Irrigated cropland, due to its frequent irrigation and high evapotranspiration characteristics, is a typical type of land use with high water consumption, further exacerbating the contradiction between water supply and demand and the risk of water shortages in the region.

The weights of resistance factors were calculated using principal component analysis. The results are shown in Table 1.

Table 1.

Weight calculation results table

	e	VE	CVE	SC	w
water demand	1.45	60.56	60.56	0.03	0.03
drainage	0.48	19.92	80.49	0.48	0.42
DEM	0.35	14.48	94.96	0.26	0.22
AWC	0.12	5.04	100.00	0.38	0.33

Resistance surfaces were comprehensively calculated based on weighted resistance factors, and corridors were identified using these resistance surfaces. The corridor identification results revealed the critical role of Jilin Province’s major river systems in ecological water conveyance. A total of 71 corridors were identified, with Figure 8D displaying the frequency histogram for corridors assessed ≥10 times. Over the 23-year analysis period from 2000 to 2022, the Second Songhua River, Dongliao River, and Nenjiang River were identified as corridors most frequently, recognized as major ecological corridors for 23 consecutive years. This is primarily due to their extensive basins, stable flow rates, and transcontinental flow from northeast to southwest Jilin Province, forming core transport channels connecting “source areas to convergence zones.” Next were the Huldah River, Liao River, and Erdao Baihe River, which primarily lie at the boundary between source and sink areas, achieving identification rates exceeding 90% throughout the 23-year period. In contrast, rivers such as the Dasha River, Huifa River, Santong River, and Hani River exhibited lower identification frequencies, mainly located within the core “source area.” Despite abundant water resources, these rivers lack the demand for transporting water to the “sink area,” thus lacking significant ecological connectivity functions.

Figure 8.

A series of images related to corridor identification and land use in source and sink areas in Jilin Province

(A) water-related ecological security patterns of Jilin Province, (B) water security-sensitive resistance surface in Jilin Province; and (C) bar charts show the proportional land use types in source and sink areas.

Identification of key corridors and vulnerable nodes

To reveal the dynamic changes in water resource supply and demand trends across different regions of Jilin Province over the long term and to characterize the spatial gradient of water security, this article used the natural breakpoint method to divide the regional supply-demand ratio into five security levels, including highly secure, relatively secure, generally secure, relatively insecure, and insecure areas, in order to identify key protection areas and potential risk areas. The spatial distribution results indicate that the western and northeastern regions of Jilin Province, particularly Baicheng City, Songyuan City, and Siping City, are primarily located in low-safety zones due to long-term water resource demand pressures and weak ecological supply capacity, forming a regional water ecological vulnerability belt. In contrast, the southern region, with its good forest coverage and water conservation capacity, exhibits a high-safety zone overall, possessing stable ecological support functions, primarily located within Tonghua City. The central region predominantly exhibits general safety levels, simultaneously fulfilling agricultural irrigation and urban water supply functions while maintaining some ecological buffer capacity. This area represents a potential frontier for the evolution of water resource supply-demand contradictions.

This article combines the MCR model to identify regional water ecological corridors and delineate scientific “source-sink” corridor levels, providing targeted references for regional river management. To scientifically classify ecological corridor levels, this study employs the Cumulative Contribution Rate Method based on the cumulative frequency of rivers identified as corridors over a 23-year period. This method emphasizes the dominant role of “key few” rivers within the system, making it suitable for identifying rivers that play a core role in ecological network connectivity. Rivers are categorized into three levels based on the distribution characteristics of their contribution rates: Rivers with a cumulative contribution rate of no more than 50% are classified as important corridors, representing the primary water systems that consistently fulfill migration and connectivity functions in most years; Rivers with a cumulative contribution rate between 50% and 85% are defined as key corridors, playing a secondary supporting role in the regional ecological network; Rivers with a cumulative contribution rate exceeding 85% are classified as secondary corridors, whose ecological connectivity functions are relatively weak and subject to annual fluctuations or regional characteristics. Among these, 18 rivers were designated as important water ecological corridors: Second Songhua River, Nen River, Dongliao River, Huerda River, Liao River, Erdaobai River, Erdao River, Yitong River, Fuer River, Songhua River, and so forth.25 rivers were designated as key corridors: Gaya River, Hailan River, Suifen River, Toudao River, Jinsha River, Huangni River, Wuli River, Piao River, Dawangqing River, Mudan River, Seconddao River, Aolong River, Lalin River, and so forth.

Meanwhile, a total of 58 vulnerable nodes were identified between 2000 and 2022, 9 of which were early warning nodes. To validate the vulnerable nodes identified in the model, this study utilized flood event occurrence data from 2000 to 2022. By statistically analyzing the timing and locations of flood events, the annual and seasonal flood occurrence frequencies were calculated. Annual flood frequency distribution maps were then generated to analyze the spatial distribution characteristics of floods. Subsequently, high-frequency flood zones were spatially compared with vulnerable nodes identified via the minimum cumulative resistance (MCR) model to verify whether these nodes were located in areas prone to frequent flooding. This validation process employed spatial analysis tools in ArcGIS, utilizing spatial overlay analysis to assess the overlap between high-frequency flood zones and vulnerable nodes. The analysis results are shown in the Figure 9, revealing a significant spatial overlap between the vulnerable nodes identified by the model and high-frequency flood zones. Specifically, early warning nodes are predominantly concentrated around these high-risk areas and exhibit consistent spatial distribution patterns with flood-prone regions. This indicates that the model effectively captures key flood-prone areas when identifying vulnerable nodes, providing robust support for water resource management and disaster prevention and mitigation.

Figure 9.

Overlapping distribution of early warning nodes, vulnerable nodes, and flood-prone areas

Although the spatial distribution of the nodes is relatively dispersed, there is a clear trend toward concentration, with most located in the areas around Baicheng, Changchun, and Jilin City. The spatial pattern exhibits a “vertically dense, horizontally dispersed” characteristic. Early warning nodes are primarily located in areas with lower-tier river networks and obstructed water supply channels, often situated at the “source-to-sink” interface zones, urban expansion peripheries, or near general safety zones. These nodes are primarily concentrated in Changchun City, Jilin City, and their surrounding areas. This region exhibits weak ecological connectivity and a continuously rising risk resistance value, making it a potential risk hotspot within the water ecological safety network. Based on this, this article identifies vulnerable nodes and early warning nodes for the year 2032 under three scenarios. The findings reveal that early warning nodes exhibit a spatial trend of outward diffusion from central urban areas, predominantly distributed in regions such as Baicheng, Songyuan, and Baishan, and concentrated at the direct junctions of “source-sink” zones. This reflects these areas' critical role in bridging ecological functions between source regions and high-water-consumption zones, making them susceptible to dual impacts from supply-demand imbalances and heightened ecological resistance. The disappearance of early warning nodes in the central region, on the other hand, suggests that ecological connectivity in this area has improved, with supply-demand relationships becoming more balanced. This is attributed to optimized land use structures, leading to a significant reduction in the vulnerability of the regional water ecosystem. Under natural development and ecological protection scenarios, early warning nodes tend to concentrate in the northwest and southeast directions. The spatial distribution in the northwest direction roughly overlaps with the boundary between the source and sink areas, mainly distributed in the upper reaches of the Songhua River and the Second Songhua River and their surrounding cities. At the same time, the disappearance of early warning nodes in the general safety zone indicates that the “source area” is continuously shrinking, and the water self-sufficient area at the junction of the source and sink is gradually expanding. Compared with the other two scenarios, under the farmland protection scenario, early warning nodes are predominantly concentrated along the main streams of the Songhua River, the Second Song River, the Toudao River, and the Erdao River, forming a belt-like distribution near the general safety zone. This phenomenon demonstrates that with the vigorous development of agriculture and farming, the overall water ecological health in the middle reaches of the study area—which previously maintained a relatively high level of water security—has declined as shown in Figure 10.

Figure 10.

Distribution of vulnerable nodes, early warning nodes, and corridors

Discussion

Driving mechanisms of climate and land use change on the evolution of water security patterns

Climate factors play an important role in the evolution of water-related ecological security patterns by regulating the temporal and spatial distribution of water resources and influencing all aspects of the water cycle. Climate elements such as precipitation and evaporation directly determine the supply of water resources and the effective availability of water sources.⁴³ In particular, climate change has a more prominent impact on water-related ecological security patterns in regions with water scarcity or significant climate change. This article selects precipitation (precip) and potential evapotranspiration (PET) as two key climate factors as research variables to explore their role in the evolution of water security patterns. By analyzing the correlation between these two climate factors and water security indicators such as total water production and source area, we reveal the impact mechanism of climate factor changes on water security patterns.

Using Pearson’s correlation analysis, the following results were obtained: As shown in Figure 11, there is a significant positive correlation between precipitation and total water yield, as well as between precipitation and source area (R = 0.964, p < 0.001; R = 0.742, p < 0.001). This indicates that an increase in precipitation has a significant promotional effect on regional water yield and source area expansion. Additionally, there is a significant correlation between water production and park area (R = 0.672, p < 0.001). Specifically, increased precipitation directly leads to greater water accumulation and inflow, thereby increasing the total water production in the region, promoting the expansion of the source area, enhancing the region’s effective water supply capacity, and facilitating the distribution of water resources across a broader area. This mechanism is particularly important in the evolution of water security patterns, especially under conditions of adequate precipitation, where water resource accessibility is significantly enhanced. However, the correlation between evaporation and water production as well as source area is weak (R = −0.423, p = 0.405), indicating that evaporation has a relatively minor direct impact on these variables. While increased evaporation may affect water resource availability through moisture loss, its influence on regional water production and source area is indirect and insufficient to rival the impact of precipitation.

Figure 11.

Pearson correlation analysis plot

Land use often redefines the relationship between water supply and demand by altering surface characteristics and water use patterns.^44,45 Che et al. noted that land use changes can alter watershed water cycle processes, such as affecting evapotranspiration and infiltration, thereby influencing water yield.⁴⁵ For example, large-scale deforestation can increase baseflow rapidly due to increased surface runoff, with its impact on water yield being approximately 2.7 times that of concurrent climate change. Land use change directly affects water demand and the balance between supply and demand.^46,47 This study set up three different development scenarios by restricting the conversion between land types: natural development scenario, cropland protection scenario, and ecological protection scenario. The results showed that in the cropland protection scenario, when large areas of forest land and wetlands are converted to cropland, the regional water production process and spacetime pattern will change: the function of water sources will weaken, the pressure on the catchment area will intensify, and the smoothness of the corridor for the transport of water resources and ecological services from the “source” area to the “catchment” area will decrease, reducing transmission efficiency. In the ecological protection scenario, large areas of wetlands, grasslands, and forest land increased, strengthening water conservation, expanding the area of water sources, reducing water supply pressure, improving water supply security, and significantly improving the connectivity and overall function of water ecosystems.

The driving mechanism of cropland types on the water supply and demand pattern

As the world’s largest developing country, agriculture serves as the cornerstone of China’s economic development, with the cultivation of food and cash crops playing a pivotal role in the national economy. Agricultural water uses account for 60%–70% of China’s total water consumption, with irrigation water demand being particularly prominent. However, there are significant differences in water demand among different types of crops. In order to more accurately assess water demand and distribution, this study subdivides cropland into irrigated cropland and rainfed fields to analyze their different water demands and their impact on water security patterns. Through a comparative analysis of the catchment area of nine types of land use (distinguishing between rainfed cropland and irrigated cropland) and eight types of land use from 2000 to 2022, we found that after refining the types of croplands, the catchment area of the study area increased significantly, reflecting the higher demand for water resources by irrigated cropland. As shown in Figure 12, the increase in the sink area was particularly significant in 2005 and 2020. In 2020, the increase in the sink area reached 39% of the original total sink area, further indicating that the refinement of irrigated cropland further expanded the sink area, resulting in a water supply deficit and promoting an increase in water demand. Especially in areas with relatively scarce water resources, the high demand for irrigated cropland further exacerbated the pressure on water supply, which in turn affected the identification and construction of ecological corridors.

Figure 12.

Bar chart of sink area changes with differentiated cropland types

To further explore the impact of cropland type refinement on regional water ecological supply and demand patterns, this article conducted geographical center of gravity migration and standard deviation ellipse analysis on the distribution of source areas based on 8-type and 9-type land use structures at five-year intervals, as shown in Figure 13. The results showed that the spatial morphology of source areas under the 9-type structure after cropland refinement exhibited significant directional characteristics. The standard deviation ellipse analysis results showed that although the long axis of the source area distribution in both structures was northwest-southeast, the ellipse contour in the 9-type structure was further stretched toward the southeast, indicating that after the refinement of cropland types, the spatial pattern of the source area showed a stronger southeast concentration, and the geographical center of gravity had shifted from Baicheng City to Songyuan City. This stretching not only reveals the increased spatial concentration of the source area distribution but also indirectly reflects the expansion effect of water-intensive irrigated cropland in the regional distribution, thereby affecting the geographical pattern of water supply capacity. From 2000 to 2020, the maximum horizontal and vertical displacements of the geographic origin centers for both the 8-category and 9-category land use structures were approximately 11 km, with the longest migration distance reaching about 20.3 km. Although the overall displacement amplitude is limited, existing research findings have shown that the migration of geographic centers often reflects significant spatial shifts in land use structures, resource allocation, or ecological service provision.⁴⁸ Combining the above analysis, the shift in the center of gravity of the source area in this study is highly consistent and consistent with the high level of water consumption characteristics of irrigated cropland. Therefore, it can be reasonably judged that this is a signal of the restructuring of the spatial distribution pattern of water supply capacity driven by changes in cropland types. In particular, considering the significant increase in the sink area and the spatial deviation between the demand area and the supply area, the refinement of irrigated cropland has not only increased regional water demand but also spatially reshaped the supply and demand pattern of water ecosystem services.

Figure 13.

Centroid migration and standard deviation ellipse maps of sink areas with differentiated cropland types

Suggestion

The article identifies three types of water security patterns. Based on the above results (3.3, 3.4) and the figure, it can be analyzed that among the 10 important water-related ecological corridors constructed in Jilin Province over the past 23 years, the Gudong River has been identified as an ecological corridor 21 times. Although the Gudong River is a fourth-level river with a relatively low grade, it has been repeatedly identified as a key ecological corridor, reflecting its stable and irreplaceable ecological connectivity. This phenomenon of “low grade but high importance” indicates that although the Gudong River has a limited scale, its ecological functions and connectivity value significantly exceed those of other rivers of the same grade. Therefore, it should be given special attention and priority protection in regional ecological conservation strategies. Specific recommendations include: Strictly implementing the “Implementation Plan for the “Return Rivers to Rivers” Initiative in Jilin Province (2023–2025)” issued by the Jilin Provincial People’s Government, precisely demarcating river boundaries, establishing a routine mechanism for clearing illegal structures, ensuring ecological flow and flood safety, strengthening coordinated protection of groundwater and surface water in the Gudong River basin, advancing ecological bank protection construction, restoring riverbank vegetation, and enhancing water self-purification capacity (Jilin Provincial People’s Government, 2020). Additionally, cross-regional collaborative governance should be strengthened, and cooperation with Liaoning Province should be established to implement the “Gudong River Basin Comprehensive Plan (2024–2035),” establish a joint governance platform for the basin, and improve the overall governance efficiency of the basin.

Early warning nodes are primarily distributed in areas such as Baicheng City, Songyuan City, and Changchun City. In the results of water ecological corridor identification over the past 23 years and across three future scenarios spanning 26 years, the early warning node located at the confluence of the Tao’er River and Nenjiang River in Baicheng City has consistently appeared, indicating that the comprehensive ecological resistance in this region has continued to increase and that it is situated in a water supply deficit zone with significant ecological vulnerability. The river sections upstream and downstream of this early warning node should be included in the priority protection zone. It is recommended to accelerate the construction of ecological buffer zones upstream and downstream, enhance river ecological functions through wetland restoration and shoreline ecological reconstruction, strictly control land development intensity along both sides of the river, establish water-retaining forests to improve local ecological hydrological conditions, and implement a long-term ecological safety dynamic monitoring mechanism.

Limitations of the study

This study employed the InVEST model’s annual average runoff module for regional runoff simulation. The simulation was based on an annual average water balance method, which did not explicitly account for surface-groundwater interactions or temporal variations in hydrological processes. This method assumes that precipitation and evapotranspiration are evenly distributed across the year, resulting in the model’s inability to reflect seasonal dynamics or hydrological changes under extreme climate events. Especially under extreme conditions such as prolonged droughts or sudden floods, the annual average method often fails to capture the reduction in runoff during dry periods or the surge in runoff during flood periods, potentially underestimating the risk of water scarcity during droughts or ignoring peak flood flows, thereby introducing biases in the assessment of regional total water yield.⁴⁹ Since the model does not distinguish between surface runoff and groundwater processes, the hydrological cycle is overly simplified, and the contribution of potential groundwater recharge to flow is ignored. This simplification of hydrological process details is believed to introduce significant uncertainty, weakening the model’s accuracy in depicting actual water production. In addition, the Z factor (i.e., Zhang parameter) in the InVEST water supply module is an empirical constant introduced based on the Budyko hypothesis to characterize the seasonal distribution of local precipitation and hydrogeological characteristics. It is not a directly measurable parameter and typically requires empirical or calibration-based settings. This empirical calibration process is highly subjective: different basins often require different values to match measured runoff or water resources, and the Z values selected by different researchers may vary significantly. For example, a study in the South China monsoon region calibrated the Z value to approximately 1.55 to align model water production with observed total water resources, but optimized Z values in adjacent regions reached as high as 2.32 and 7.0,⁵⁰ indicating that even under similar climatic conditions, suitable values can vary by several orders of magnitude across different regions. The uncertainty of climate inputs such as precipitation, combined with errors in the selection of Z parameter values, constitutes the primary source of uncertainty in the results of the InVEST water production model. If the Z factor is improperly selected, the simulated water production results may exhibit significant deviations.

Innovation and prospects

This study conducted a quantitative assessment of the water-related ecological security pattern in Jilin Province from 2000 to 2022, and classified regional rivers to more accurately identify the characteristics of the water-related ecological pattern and optimize ecological corridors. It confirmed that the eastern Changbai Mountain area has dense forests and wetlands and abundant water resources, while the western agricultural irrigation plains have a persistent water-related ecological service deficit due to excessive water use. Considering that it is crucial to comprehensively consider the hierarchical structure of river networks in ecological network planning, Large rivers often serve as the backbone of regional ecological corridors and should be protected and utilized as natural corridors to enhance regional ecological connectivity. However, lower-level tributaries also make important contributions to the ecosystem of the river basin. Therefore, drawing on relevant research methods, incorporating the river network into the construction of a water-related ecological security pattern according to its level and distinguishing between corridor levels will help to accurately identify corridors that need to be protected and comprehensively enhance the function and connectivity of water-related ecological corridors. It also emphasized the importance of the differences in water use intensity and ecological services between irrigated cropland and rainfed cropland for regional water security assessments. Many previous studies tended to treat cropland as a single land type, ignoring the high water consumption characteristics and ecological functions of irrigated cropland as artificial wetlands.^51,52To address this shortcoming, we distinguished between irrigated cropland and rainfed cropland, thereby improving the accuracy of the assessment of water ecosystem service supply and demand and meeting the requirements of detailed management in water-intensive agricultural areas.⁵³ The spatial differentiation characteristics of the supply and demand patterns of water ecological services in Jilin Province were revealed: the northwestern agricultural irrigation plains have formed a persistent water ecological service deficit zone, while the eastern Changbai Mountains, due to their dense forests and wetlands, have become the primary water supply zone.⁵⁴ This confirms the universal law of water supply and demand imbalance and is consistent with the current situation of water resource scarcity in China’s northeastern grain-producing regions caused by excessive groundwater irrigation. The analysis of changes in water ecosystem services under different development scenarios provides a scientific basis for policy-making. This study incorporated future land use changes and dynamic changes in water ecosystem services into the assessment, whereas traditional studies have rarely considered the impact of multiple land use scenarios on the spatial distribution of water ecosystem services.⁵⁵ Under the cropland protection scenario, further expansion of cropland will significantly exacerbate regional water resource pressure, consistent with the conclusions of other studies that cropland expansion will lead to a decline in water output and exacerbate water shortages.⁵⁵ Conversely, under the ecological protection scenario, a large-scale increase in forest and wetland areas can enhance the regional water production service function. The simulation shows that the water supply deficit area is reduced by about 22.3% compared with the natural development scenario. The results showed that in areas with high water consumption in agriculture, ecological protection measures can still increase the net water supply by reducing irrigation water withdrawal and other means.

In summary, this study comprehensively used PLUS land use scenario simulation, InVEST hydrological service assessment, and MCR minimum cumulative resistance model to construct a “source-sink” ecological network method. Based on ecosystem service flow theory, it improved the accuracy and scenario adaptability of water security pattern assessment, provided a targeted and more dynamically adaptable scientific basis, and provided more scientific tool support for water resource management. The research findings hold significant practical value for regional water resource optimization and water security management.

Future research could further incorporate groundwater processes to address the shortcomings in current water production service simulations. If a monthly scale analysis is adopted, downscaling methods could be used to couple the InVEST model with the MODFLOW model to address the issue that InVEST does not consider the overlap between surface water and groundwater in water production service simulations, thereby enhancing the completeness of hydrological processes. Alternatively, the groundwater retention module in the InVEST model could be introduced to partially reflect groundwater processes without the need for temporal downscaling. However, this approach struggles to capture the dynamic interaction between rivers and groundwater. Therefore, future research should select appropriate supplementary mechanisms based on scale requirements to improve the model’s adaptability and simulation accuracy under the coupling of multi-source hydrological elements.

Conclusions

This study used Jilin Province as a typical research area and comprehensively applied PLUS land use multi-scenario simulation, InVEST ecosystem service flow assessment, and the MCR model to construct a technical framework for identifying water security patterns under the “supply-flow-demand” process of water ecosystem services in future scenarios. Overall, the research methodology demonstrates good applicability and innovation, providing scientific support for studies on the evolution of water resource spatial patterns under complex conditions. However, it also has certain limitations, and its reliability and applicability should be comprehensively analyzed in conjunction with actual application requirements.

• 1.Analysis based on the PLUS model indicates that Jilin Province’s land use pattern remained generally stable from 2000 to 2022, though localized changes were active. Grasslands significantly increased due to ecological projects, water bodies continued to expand, and frequent conversions occurred between farmland and construction land. Multi-scenario projections show that by 2032: - Water bodies will increase markedly under the natural scenario; - Rain-fed farmland will see the largest increase under the farmland protection scenario, while the ecological conservation scenario promotes grassland and wetland growth with corresponding reductions in farmland.
• 2.Using the InVEST model to assess water resource services reveals simulation errors below 2.35%, meeting regional analysis requirements. Water demand correlates closely with population density and farmland area, exhibiting uneven spatial distribution: mountainous areas show “high supply and low demand” while plains exhibit “low supply and high demand.” The ecological conservation scenario significantly reduces the water supply deficit area by 22.3% by 2032, delivering the most effective water security improvement.
• 3.The MCR model identifies 18 important ecological corridors and 25 key corridors. Additionally, 58 vulnerable nodes and 9 early warning nodes are determined. The spatial distribution of early warning nodes varied across scenarios. Under the farmland protection scenario, nodes shifted southward, leading to a decline in overall water ecological health.
• 4.Model validation showed that the PLUS simulation achieved an overall accuracy of 0.92 with a Kappa coefficient of 0.87. The InVEST simulation yielded a relative error of approximately 2.35% in runoff data. Both results met research requirements, confirming the models' applicability and accuracy in reflecting the interaction between land use and water resources.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Changlei Dai (hd_daichanglei@126.com).

Materials availability

This study did not generate new unique materials.

Data and code availability

• •Data: This article analyzes existing, publicly available data, including hydrometeorological information, surface information, water resources data, and socioeconomic data. Acquisition methods are listed in the key resources table.
• •Code: This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

This study was supported by funding from the project “Study on the Impact of Climate Change on the Hydrological Regime of the Heilongjiang (Amur River) Basin” (Project Number: 2022KF03), funded by the Institute of international rivers and eco-security,Yunnan University.

Author contributions

Conceptualization, Q.R., C.D., and X.Y.; methodology, Q.R., X.Y., Y.L., X.W., C.D., and Y.S.; investigation, Q.R., X.Y., Y.L., and X.W.; writing – original draft, Q.R.; writing – review and editing, Q.R., Y.X., C.D., and Y.S.; funding acquisition, C.D.; resources, C.D. and Y.S.; supervision, Y.X., CD, and YS.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Deposited data
Land use data	the Institute of Air and Space Information Innovation, Chinese Academy of Sciences56	https://data.casearth.cn/thematic/glc_fcs30/
DEM data	geospatial data cloud website	https://www.gscloud.cn/
GDP data	a published article by Prof. Nai-Chuo Zhao et al.57	https://github.com/thestarlab/ChinaGDP
Population density data	by Oak Ridge National Laboratory (ORNL) of the U.S. Department of Energy and provided by East View Cartographic	https://landscan.ornl.gov
Nighttime lighting data	a published article by Prof. Yizhen Wu et al.58	https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/GIYGJU
Precipitation;Potential evapotranspiration;Edrock depth;Plant Available Water Content	the Resource and Environment Data Center of the Chinese Academy of Sciences	https://www.resdc.cn/
Urban flood data	a published article by Shengnan Fu et al.9	https://github.com/shengnan0218/China-urban-flood-dataset
Total Water Resources data	Jilin Province Water Resources Bulletin	https://slt.jl.gov.cn/yw/szyglybh/szygb/
Software and algorithms
ArcGIS	ESRI	https://www.arcgis.com/index.html
PLUS	High-Performance Spatial Computational Intelligence Lab	https://www.urbancomp.net/archives/plus
InVEST	Stanford University	https://naturalcapitalproject.stanford.edu/software/invest
Origin 2021	OriginLab	https://www.originlab.com/OriginProLearning.aspx
SPSS	IBM SPSS	https://www.ibm.com/cn-zh/spss

Method details

Data

A detailed description of the data materials and their sources, including geographical (meteorological and topographical) and socioeconomic (population and gross domestic product) data, is provided in data sources table.

Table 2.

Data sources

Data types	Resolution	Source
Land use data	30 m	The first set of global 30-meter resolution land cover data from 1985 to 2022 released by Liu Liangyun’s team, a researcher at the Institute of Air and Space Information Innovation, Chinese Academy of Sciences (CAS)56 (https://data.casearth.cn/thematic/glc_fcs30/)
DEM data	30 m	geospatial data cloud website (https://www.gscloud.cn/)
GDP data	1 km	a published article by Prof. Nai-Chuo Zhao et al.57 in 2016 (https://github.com/thestarlab/ChinaGDP)
Population density data	1 km	The LandScan population dataset (https://landscan.ornl.gov) developed by Oak Ridge National Laboratory (ORNL) of the U.S. Department of Energy and provided by East View Cartographic
Nighttime Lighting Data	1 km	The data originates from a article titled “Developing improved time-series DMSP-OLS-like data (1992–2019) in China by integrating DMSP-OLS and SNPP-VIIRS” published by Wu Yizhen (2021),58 and other scholars in the journal IEEE Transactions on Geoscience and Remote Sensing.
Precipitation	1 km	The Resource and Environment Data Center of the Chinese Academy of Sciences (https://www.resdc.cn/)
Potential evapotranspiration	30 m
Edrock depth	30 m
Available Water Capacity	1 km
Plant Available Water Content	1 km

Land-use simulation under multiple scenarios

The land use changes in the study area were simulated and predicted using PLUS model (Patch-generating Land Use Simulation Model). The PLUS model is a metacellular automata (CA) model based on raster data, which is mainly used for patch-scale land use/land cover change simulation. The model is able to effectively simulate the land use evolution process under the combined effect of human activities and natural factors through two core modules: the Land Expansion Analysis Strategy (LEAS) and the CA model based on multi-class stochastic patch seeds (CARS), combined with machine learning algorithms.⁵⁹ The model sets up land use simulations under different scenarios by adjusting the transfer probability matrix, neighborhood weights, land demand forecasts, and multi-objective optimization (MOP) methods to balance expansion capacity, land demand, and environmental protection.

This study simulated the land use distribution pattern of Jilin Province in 2022 and used the Kappa coefficient to verify the accuracy of the simulation results. The results showed that the Kappa coefficient reached 0.85, with an overall accuracy of 0.88. This result indicates that the model has high reliability and meets basic scientific research requirements. Then, using land use data from 2002, 2012, and 2022 as baseline land use data, the study simulated the demand for various land uses in 2032. By combining prior experience to determine the conversion of different land types under various development scenarios, the study obtained the land use situation in 2032 under each scenario.

Land-use conversion methodology

The transition matrix comprehensively and specifically describes the structural changes and transformation directions between different land use types within a region. This method is based on state transitions and quantitative descriptions of system states in systems analysis. It can accurately calculate the inflows and outflows of various land use types, thereby revealing the evolution process of land use structure between the beginning and end of the study period and the dynamic changes in land use type transitions during different time periods.⁶⁰ Its mathematical expression is as follows:

Let there be n types of land use. The land use transition matrix can be represented as an n×n square matrix.

$$T_{ij}=\left[\begin{array}{ccc}{T}_{11}& \cdots & T_{1n}\\ ⋮& \ddots & ⋮\\ T_{n1}& \cdots & T_{nn}\end{array}\right]$$ (Equation 1)

In the formula, the symbol T represents area, n represents the total number of land use types, and i and j represent the land use types at the beginning and end of the study period, respectively. In practical applications, to facilitate visualization and interpretation of results, the land use transition matrix is typically presented in tabular form.

The land use change analysis in this study is divided into two time periods. The first period (2000-2022) is further subdivided into three stress periods: 2000-2008, 2008-2016, and 2016-2022. Corresponding land use transition matrices are constructed for each period to quantitatively identify the conversion relationships and structural evolution characteristics between land use types within each stage. The second period is the future projection period (2022-2032). Based on simulation results, transition matrices are constructed for 2032 under three scenarios (natural development, ecological protection, and cropland protection) and compared with the 2022 land use patterns. This analysis examines the transformation of land use types under different development pathways, thereby revealing the potential trends in land system changes driven by these scenarios.

Quantitative assessment method for water supply

The InVEST model was developed by the Natural Capital Project at Stanford University in the United States to assess and quantitatively simulate how natural capital supports human well-being. This study employs the Water Yield Module within the InVEST model to quantify the water supply capacity within a given region. This module is based on the Budyko hydrological hypothesis,⁶¹ which assumes that there is a stable empirical relationship between annual precipitation and actual evapotranspiration consumption in a watershed on an annual time scale. The formula is as follows

$$\begin{array}{c}\frac{{AET}_{\left(x\right)}}{P_{\left(x\right)}}=1+\frac{{PET}_{\left(x\right)}}{P_{\left(x\right)}}-{\left[1+{\left(\frac{{AET}_{\left(x\right)}}{P_{\left(x\right)}}\right)}^{⍵}\right]}^{\frac{1}{⍵}}\end{array}$$ (Equation 2)

In the formula, AET_(x)represents actual evapotranspiration (mm), PET_(x)represents potential evapotranspiration (mm), P_(x)represents annual precipitation (mm), and $⍵$ is a non-physical empirical parameter representing natural climate and soil properties. This is used to estimate annual water production:

$$\begin{array}{c}{Y}_{\left(x\right)}=\left[1-\left(\frac{{AET}_{\left(x\right)}}{P_{\left(x\right)}}\right)\right]\times P_{\left(x\right)}\end{array}$$ (Equation 3)

In the formula, Y_(x)represents annual water production (mm).

This study simulated water production in Jilin Province from 2000 to 2022 and calibrated the Z factor in the model using actual annual water resource data published in the regional water resource bulletin to ensure the accuracy of the simulation results. By comparing the measured and simulated results, it was determined that a Z value of 3.2 yields the best fit. Based on this, the calibrated parameters were applied to the land use simulation results for three different scenarios in 2032, thereby estimating the spatial distribution and supply capacity of water production services in the region under future scenarios.

Quantitative assessment method for water demand

The demand for water supply ecosystem services is equivalent to the total amount of water resources actually consumed by humans,⁶² including water used for agriculture, industry, and domestic purposes. Ecological water use is not counted toward water supply service demand since it is not directly consumed by humans.

This study utilizes data on population density, industrial GDP, and nighttime light values, combined with land use reclassification methods, to assign residential, agricultural irrigation, and industrial water consumption values to each pixel grid.^63,64 This process ultimately constructs a comprehensive water demand grid dataset for Jilin Province at a 1 km × 1 km resolution, averaging multiple years. This dataset will be used for subsequent analysis and assessment of water resource surplus/deficit patterns. The specific calculation method is as follows:

$$\begin{array}{c}WD=Ard+Id+Dd\end{array}$$ (Equation 4)

$$\begin{array}{c}Rcd\left(x\right)=\frac{Rca\left(x\right)}{{Rca}_{sum}}\times {Rcd}_{sum}\end{array}$$ (Equation 5)

$$\begin{array}{c}Icd\left(x\right)=k\times Rcd\left(x\right)\end{array}$$ (Equation 6)

$$\begin{array}{c}Ard={Rcd}_{sum}+{Icd}_{sum}\end{array}$$ (Equation 7)

$$\begin{array}{c}{GDP}_i\left(x\right)=\frac{DN\left(x\right)}{{DN}_{sum}}\times {GDP}_i\end{array}$$ (Equation 8)

$$\begin{array}{c}Id\left(x\right)=\frac{{GDP}_i\left(x\right)}{{GDP}_i}\times {Id}_{sum}\end{array}$$ (Equation 9)

$$\begin{array}{c}Dd\left(x\right)=\frac{POP\left(x\right)}{{POP}_{sum}}\times {Dd}_{sum}\end{array}$$ (Equation 10)

Among these, WD denotes total water demand, Ard represents agricultural water demand, Id signifies industrial water demand, and Dd indicates domestic water demand; Rcd(x) and Icd(x)denote grid-based water demand for rainfed cropland and irrigated cropland respectively, Rca(x) represents the area of rainfed cropland grids, Rcd_sum denotes the total area of rainfed cropland, while Rcd_sum and Icd_sum signify the total water consumption for rainfed and irrigated cropland respectively; GDP_i(x) denotes grid-level industrial GDP, GDP_i denotes total industrial GDP, DN(x) and DN_sum denote grid-level nighttime light values and total nighttime light values, Id(x) denotes grid-level industrial water consumption, and Idsum denotes total industrial water consumption; Dd(x) and Dd_sum denote grid-level domestic water demand and total domestic water demand, respectively, while POP(x) and POP_sum denote grid-level population density values and total population density, respectively.

Supply-demand ratio method

This study uses ESDR (ecological supply and demand ratio) to assess the spatial supply-demand relationship of water supply ecosystem services. The specific formula is as follows:

$$\begin{array}{c}ESDR=\frac{S-D}{\left(S_{\mathit{max}}+D_{\mathit{max}}\right)/2}\end{array}$$ (Equation 11)

In the formula: S、D represent regional water resource supply and demand, respectively;S_max、D_maxrepresent maximum regional water resource supply and demand, respectively. ESDR> 0、ESDR< 0 indicate surplus and deficit states of water ecosystem supply and demand services, respectively; ESDR=0 indicates that supply and demand are nearly balanced.

Identifying “source” and “sink” areas method

In this study, the “source area” of the water-related ecological security pattern is defined as a spatial unit with strong water supply service capacity and service surplus that can provide replenishment to other areas on the basis of meeting local water demand.^65,66 After screening regions with ESDR > 0, this study conducted cluster analysis to identify areas with high surplus water supply services as water ecological “source” regions, which are water-surplus areas. Conversely, ESDR < 0 indicates a water supply deficit, meaning the region requires water supplementation and is thus a “sink” region. Considering the number and scale of source regions, the minimum area threshold for source and sink regions was set at 100 km².

Resistance surface establishment method

During the transmission of water supply services from the “source area” to the “sink area,” the flow path is subject to comprehensive influences from factors such as terrain, river flow, and water resource consumption, resulting in a certain degree of obstruction. However, ecosystem service flows tend to diffuse along paths of least resistance, thereby forming ecological corridors.²⁰ Therefore, this study utilizes the D8 flow direction algorithm combined with flow distribution and river grade information to determine the dominant flow direction of water supply services.²⁴ A resistance surface was constructed using key factors including drainage network density, terrain elevation, available water content (AWC), and water demand intensity. Given the poor connectivity and susceptibility to service interruptions in lower-order river segments, higher resistance values were assigned to these areas¹⁸; non-channel areas were assigned a value of 10 to ensure water supply services flow along the river channels. After standardizing each resistance factor using ArcGIS standardization tools, principal component analysis (PCA) was applied to calculate weights.⁶⁷ Annual composite water ecological resistance surfaces were then generated by weight-weighted overlaying these factors, providing support for subsequent ecological corridor identification and service flow simulation. The specific calculation process is as follows:

$$\begin{array}{c}{PC}_{ij}=\frac{{WL}_{ij}}{\sqrt{e_j}}\end{array}$$ (Equation 12)

$$\begin{array}{c}{SC}_i=\frac{{{\sum }_{j=1}^k\left({PC}_{ij}\times {VE}_j\right)WL}_{ij}}{{CVE}_k}\end{array}$$ (Equation 13)

$$\begin{array}{c}{w}_{ij}=\frac{\left|{SC}_i\right|}{{\sum }_{k=1}^n\left|{SC}_k\right|}\end{array}$$ (Equation 14)

Among these, PC denotes principal component score coefficient, VE represents variance explained, CVE signifies cumulative variance explained, SC indicates score coefficient,e represents eigenvalue, WL stands for loading coefficient, and w denotes weight.

Source-to-sink corridor construction method

The water-related ecological security pattern constructed in this paper comprehensively considers the physical flow characteristics of water bodies and the spatial distribution patterns of water ecosystem services. The former mainly refers to the natural flow process of water bodies along river networks from high to low, from upstream to downstream, through farmland and urban areas under the combined effects of terrain, climate, and human activities. The latter refers to the service flow path of water resources being transferred spatially from areas with surplus supply to areas with scarce resources to meet actual water demand. Based on water resource supply and demand analysis, this paper first identifies the “source areas” and “sink areas” within the study region. Subsequently, relying on the Minimum Cumulative Resistance (MCR) model, it simulates the optimal ecological service transmission paths during the diffusion process from source to sink. Since the construction of the resistance surface fully integrates topographical and hydrological characteristics, the generated ecological corridors exhibit high spatial consistency with actual river paths, effectively reflecting the flow channels of water ecosystem services from supply areas to demand areas.

Key corridor and vulnerable node extraction method

Based on the MCR model, this paper identifies the “source-sink” ecological corridor paths in Jilin Province from 2000 to 2022, overlays the simulation results for each year, and combines them with the actual river network distribution to delineate ecosystem service transmission paths that closely match the direction of the rivers. To further select key corridors, this study mapped each corridor to real rivers, counted the frequency with which river segments were identified as ecological corridors in 23 simulations, and conducted a comparative analysis based on river grades. Ultimately, river segments with high frequencies and lower grades were defined as rivers requiring priority protection, while segments with higher grades and frequent service flows were categorized as important ecological corridors to support regional water resource coordination and allocation.

To identify potential risk points in the service flow pathways, this study uses the ArcGIS Network Analyst tool to extract ecological corridor intersections as vulnerable nodes, as these areas are located in zones of concentrated material and energy transmission and are susceptible to disturbance and structurally sensitive.⁶⁸ Using the “Extract Values to Points” tool to track changes in resistance surfaces across different periods, nodes with consistently increasing resistance values are identified as water ecological safety warning nodes, with recommendations for prioritizing protection and management efforts.

Quantification and statistical analysis

This study quantifies the impact of precipitation (precip) and potential evapotranspiration (PET) as key climatic factors on water security patterns. Pearson correlation analysis revealed strong positive correlations between precipitation and both total water yield (R = 0.964, p < 0.001) and catchment area (R = 0.742, p < 0.001). This indicates that increased precipitation significantly enhances regional water resource availability and its spatial distribution. In contrast, the correlation between potential evapotranspiration and the aforementioned water security indicators was weaker and non-significant (R = -0.423, p = 0.405). All statistical results, including detailed correlation coefficients and significance levels, are presented in the discussion section and corresponding figure captions (Figure 11).

Additional resources

No additional external resources, datasets, or code repositories were used or developed for this study.

Published: January 29, 2026