EN中文

Journal of Arid Land >

2024 , Vol. 16 >Issue 6: 852 - 875

DOI: https://doi.org/10.1007/s40333-024-0101-4

  • CHEN Jiazhen 1, 2 ,
  • KASIMU Alimujiang , 1, 2, * ,
  • REHEMAN Rukeya 3 ,
  • WEI Bohao 1 ,
  • HAN Fuqiang 1 ,
  • ZHANG Yan 1
展开

收稿日期: 2024-02-26

  修回日期: 2022-04-25

  录用日期: 2024-05-01

  网络出版日期: 2025-08-13

收起

Temporal and spatial variation and prediction of water yield and water conservation in the Bosten Lake Basin based on the PLUS-InVEST model

  • CHEN Jiazhen 1, 2 ,
  • KASIMU Alimujiang , 1, 2, * ,
  • REHEMAN Rukeya 3 ,
  • WEI Bohao 1 ,
  • HAN Fuqiang 1 ,
  • ZHANG Yan 1
Expand
  • 1School of Geography and Tourism, Xinjiang Normal University, Urumqi 830054, China
  • 2Xinjiang Key Laboratory of Lake Environment and Resources in Arid Zone, Urumqi 830054, China
  • 3School of Geography and Tourism, Shaanxi Normal University, Xi'an 710062, China
*KASIMU Alimujiang (E-mail: )

Received date: 2024-02-26

  Revised date: 2022-04-25

  Accepted date: 2024-05-01

  Online published: 2025-08-13

Fold

本文引用格式

CHEN Jiazhen , KASIMU Alimujiang , REHEMAN Rukeya , WEI Bohao , HAN Fuqiang , ZHANG Yan . [J]. Journal of Arid Land, 2024 , 16(6) : 852 -875 . DOI: 10.1007/s40333-024-0101-4

Abstract

To comprehensively evaluate the alterations in water ecosystem service functions within arid watersheds, this study focused on the Bosten Lake Basin, which is situated in the arid region of Northwest China. The research was based on land use/land cover (LULC), natural, socioeconomic, and accessibility data, utilizing the Patch-level Land Use Simulation (PLUS) and Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) models to dynamically assess LULC change and associated variations in water yield and water conservation. The analyses included the evaluation of contribution indices of various land use types and the investigation of driving factors that influence water yield and water conservation. The results showed that the change of LULC in the Bosten Lake Basin from 2000 to 2020 showed a trend of increasing in cultivated land and construction land, and decreasing in grassland, forest, and unused land. The unused land of all the three predicted scenarios of 2030 (S1, a natural development scenario; S2, an ecological protection scenario; and S3, a cultivated land protection scenario) showed a decreasing trend. The scenarios S1 and S3 showed a trend of decreasing in grassland and increasing in cultivated land; while the scenario S2 showed a trend of decreasing in cultivated land and increasing in grassland. The water yield of the Bosten Lake Basin exhibited an initial decline followed by a slight increase from 2000 to 2020. The areas with higher water yield values were primarily located in the northern section of the basin, which is characterized by higher altitude. Water conservation demonstrated a pattern of initial decrease followed by stabilization, with the northeastern region demonstrating higher water conservation values. In the projected LULC scenarios of 2030, the estimated water yield under scenarios S1 and S3 was marginally greater than that under scenario S2; while the level of water conservation across all three scenarios remained rather consistent. The results showed that Hejing County is an important water conservation function zone, and the eastern part of the Xiaoyouledusi Basin is particularly important and should be protected. The findings of this study offer a scientific foundation for advancing sustainable development in arid watersheds and facilitating efficient water resource management.

Key words： PLUS model; InVEST model; Bosten Lake Basin; water yield; water conservation; land-use simulation; Geodetector

1 Introduction

Water resources are indispensable for maintaining ecosystem stability and are pivotal for human survival and development. The 6th meeting of the Intergovernmental Panel on Climate Change (IPCC) underscored the potential risks of global water supply and posed the serious threats of water resource scarcity to the sustainable advancement of humanity (Ming et al., 2021). In China, freshwater resources constitute merely 6.00% of the global total, with the arid zone in Northwest China grappling with even more acute scarcity. Consequently, the burgeoning disparity between water resources and societal needs in this region has emerged as a bottleneck hindering its developmental progress. Hence, the assessment of water ecosystem services is highly important for orchestrating the judicious allocation of regional water resources and safeguarding ecosystem equilibrium.
Water yield and water conservation are pivotal components of water ecosystem services, exerting direct influence on hydrological conditions, water cycle regulation, water purification, and flood prevention within a watershed while also regulating ecosystems. Water yield is typically considered as the total amount of water generated from a specific area, encompassing surface runoff, groundwater discharge, and direct precipitation contributions; and water conservation denotes the volume of stored water maintained at a given spatial and temporal scale (Hu et al., 2022). In recent years, scholars have conducted research on water yield at various scales, including the global, national, urban, and watershed levels (Xiao and Ouyang, 2019; Pokhrel et al., 2021; Duolaiti et al., 2023; Reheman et al., 2023). Water-scarce areas in the Asia-Pacific region have the potential to enhance water yield and mitigate drought risks through large-scale afforestation (Teo et al., 2022). Research conducted in the semiarid areas of China suggested that while the influence of climate change on runoff has intensified, human activities remain the principal factor in runoff reduction (Wu et al., 2020). Aghsaei et al. (2020) who conducted water yield research in the Anzali wetland watershed in Iran revealed that the expansion of agricultural land resulted in an increase in water yield. Significant disparities exist in the mechanisms through which the climatic environment, land-use patterns, and human activities influence water yield across diverse regions. Hence, a comprehensive understanding and assessment of regional water yield and water conservation, coupled with elucidating the mechanisms of their changes, are crucial for the efficient management of regional aquatic ecosystems.
In the contemporary evaluation of water ecosystem services, numerous studies have been conducted by employing diverse methodologies to assess water yield and water conservation. The Soil and Water Assessment Tool (SWAT) model has been extensively employed for basin-scale runoff simulation and hydrological cycle analysis owing to its high precision in simulation outcomes. However, the complexity and extensive input data requirements of SWAT model render it less suitable for application in regions with data scarcity (Janjić and Tadić, 2023). The ARtificial Intelligence for Ecosystem Services (ARIES) model can analyze and forecast the conditions, functions, and services of aquatic ecosystems (Vigerstol and Aukema, 2011). However, few hydrological studies have been conducted using the ARIES model in China, and its applicability requires validation. Compared with the aforementioned models, the water yield module within the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model requires less input data and demonstrates greater applicability. It has been extensively utilized in water production assessment worldwide (Li et al., 2021; Wang et al., 2021). Water yield is influenced by both climate and land use/land cover (LULC) changes. LULC alterations can directly impact the changes in watershed ecosystem services in short term, unlike climate change, which is characterized by its long-term nature (Guo et al., 2014). Within the framework of climate change, we can anticipate future LULC alterations and thereby formulate ecological conservation policies, such as optimizing land use structures. The Patch-level Land Use Simulation (PLUS) model has been extensively utilized for its capacity to predict the geographical distribution of future land use accurately and rapidly (Gao et al., 2022; Wei et al., 2023a, b). In view of this, the utilization of PLUS model to forecast the future land use of catchment areas serves as a cornerstone for subsequent planning and management initiatives within catchment.
This study was conducted in the Bosten Lake Basin in Northwest China, an area where the ecosystem is exceedingly fragile and sensitive to environmental changes (Yu et al., 2015). The Bosten Lake Basin, which is situated in the middle and upper reaches of the Tarim River Basin, plays a significant role in enhancing the ecological environment of the entire Tarim River Basin. Therefore, we integrated the PLUS model with the InVEST model to compute the spatial and temporal variations in water yield and water conservation in the Bosten Lake Basin from 2000 to 2020 and predict the distribution of water yield and water conservation in 2030. This study aimed to compensate for the absence of calculations regarding of water yield and water conservation in this region. Specifically, this study aimed to accomplish the following objectives: (1) analyzing the spatial-temporal dynamic pattern of LULC in the Bosten Lake Basin and conducting multi-scenario prediction simulation of LULC in 2030 to elucidate the trend of LULC change; (2) investigating the spatial-temporal pattern changes in water yield and water conservation to enhance people's understanding of the alterations in water ecosystem services within the basin; and (3) quantifying the contribution of various land use types by investigating the driving factors affecting water yield and water conservation and delineating water conservation function zones. This study will assist in formulating more effective conservation and management measures to ensure the sustainable development of the ecological environment in the Bosten Lake Basin.

2 Study area and data sources

2.1 Study area

The Bosten Lake Basin (41°00′-43°32′N, 82°28′-88°20′E) is located in the Bayingol Mongolian Autonomous Prefecture, Xinjiang Uygur Autonomous Region, China. It is in the northeastern part of the Tarim Basin, with a total area of 8.03×104 km2, and is administratively divided into Hesud County, Hejing County, Yanqi Hui Autonomous County, Bohu County, Korla City, Tiemenguan City, and part of Yuli County and Luntai County (Fig. 1). The Bosten Lake Basin is surrounded by the Tianshan Mountains on three sides. The overall topography is elevated to the north and the elevation ranges from 860 to 4739 m (Zhang et al., 2021). The Bosten Lake is located in the southeastern part of the basin and has had an average water level of 1047 m for many years (Hu et al., 2019). The Bosten Lake Basin is situated in the interior region of the Eurasian continent and is characterized by a continental arid climate. The average annual precipitation of the whole basin is 47.40-68.10 mm, and the average annual temperature is 8.20℃-11.50℃ (Kadeer et al., 2017).
Fig. 1 Location (a) and distribution of land use/land cover (LULC) (b) of the Bosten Lake Basin in Xinjiang Uygur Autonomous Region, China. Note that Figure 1a is based on the standard map (新S(2023)064) of the Map Service System (http://xinjiang.tianditu.gov.cn/bzdt_code/bzdt.html) marked by the Xinjiang Uygur Autonomous Region Platform for Common Geospatial Information Services, and the standard map has not been modified. DEM, digital elevation model.

2.2 Data sources

The data used in this study covered multiple dimensions, including land use type, natural, socioeconomic, and accessibility factors (Table 1). The land use type data were obtained from the Resource and Environmental Science and Data Center (https://www.resdc.cn). According to Liang et al. (2022), we identified seven land use types in the Bosten Lake Basin, including cultivated land, forest, grassland, water body, glacier, construction land, and unused land. Digital elevation model (DEM), slope, average annual precipitation, average annual temperature, average annual potential evapotranspiration (PET), normalized difference vegetation index (NDVI), and soil characteristics were selected as natural factors (Li et al., 2021; Reheman et al., 2023). The slope data were obtained via DEM processing. The socioeconomic factors included population density (POP), gross domestic product (GDP), and nighttime light (NTL) (Duolaiti et al., 2023). Accessibility factors included the distance from each pixel to the nearest roadways and transportation infrastructures such as railway, national highway, provincial highway, county highway, city-level road, administrative quarter, and railway station. The data of accessibility factors were derived from the Open Street Maps (https://openmaptiles.org) and the distance vectors were obtained by calculating Euclidean distance. We referred to the guidelines outlined in Food and Agriculture Organization (FAO) Irrigation and Drainage Paper No. 56 and other studies to obtain the data of root depth and crop evapotranspiration coefficient for the calculation of water yield and water conservation (Table 2) (Allen et al., 1998; Li et al., 2021; Yang et al., 2021).
Table 1 Data sources of factors involved in this study
Dimension Factor Data source Spatial
resolution		 Year
Land use type Land use/land cover (LULC) https://www.resdc.cn 90 m 2000, 2010, and 2020
Natural factor Average annual precipitation https://www.geodata.cn 1 km 2000-2020
Average annual temperature https://www.resdc.cn 1 km 2000-2020
Average annual potential evapotranspiration (PET) http://data.tpdc.ac.cn 1 km 2000-2020
Digital elevation model (DEM) https://www.gscloud.cn 90 m /
Slope https://www.gscloud.cn 90 m /
Normalized difference vegetation index (NDVI) http://www.nesdc.org.cn 1 km 2020
Soil texture https://www.fao.org/soils-portal/data-hub/soil-maps-and-databases/harmonized-world-soil-database-v12 1 km /
Soil type https://www.resdc.cn 1 km /
Soil organic matter https://data.tpdc.ac.cn 1 km /
Soil depth https://www.isric.org 1 km 2016
Socioeconomic factor Population density (POP) https://www.gscloud.cn 1 km 2020
Gross domestic product (GDP) https://www.resdc.cn 1 km 2020
Nighttime light (NTL) https://www.ngdc.noaa.gov 500 m 2020
Accessibility factor Euclidean distance from each pixel to the nearest railway https://openmaptiles.org 90 m /
Euclidean distance from each pixel to the nearest national highway https://openmaptiles.org 90 m 2020
Euclidean distance from each pixel to the nearest provincial highway https://openmaptiles.org 90 m 2020
Euclidean distance from each pixel to the nearest county highway https://openmaptiles.org 90 m 2020
Euclidean distance from each pixel to the nearest city-level road https://openmaptiles.org 90 m 2020
Euclidean distance from each pixel to the nearest administrative quarter https://openmaptiles.org 90 m 2020
Euclidean distance from each pixel to the nearest railway station https://openmaptiles.org 90 m 2020

Note: /, no data time.

Table 2 Value of parameters used in the InVEST model for the calculation of water yield and water conservation
Parameter Land use type
Cultivated land Forest Grassland Water body Glacier Construction land Unused land
Crop evapotranspiration coefficient 0.65 1.00 0.65 1.00 0.50 0.30 0.30
Root depth (mm) 300.00 2000.00 500.00 1000.00 10.00 10.00 10.00

3 Methods

Our study consisted of four steps: first, analyzing LULC alterations within the Bosten Lake Basin from 2000 to 2020, and developing a multi-scenario LULC forecast for 2030; second, analyzing the variations in water yield and water conservation; third, investigating the driving factors influencing water yield and water conservation, and analyzing the effects of different land use types on water yield and water conservation; and fourth, delineating water conservation function zone by using water conservation distribution data of 2020 (Fig. 2).
Fig. 2 Flow chart of this study. PLUS, Patch-level Land Use Simulation; InVEST, Integrated Valuation of Ecosystem Services and Tradeoffs; LULC, land use/land cover; LEAS, land expansion analysis strategy; CI, contribution index.

3.1 Land-use projection

3.1.1 PLUS model

This study modeled various scenarios of LULC in the Bosten Lake Basin in 2030. The use of PLUS model is widespread within the realm of LULC modeling and prediction (Liang et al., 2021). The fundamental principle underlying the PLUS model involves extracting the expansion for each specific land use type during transitions between distinct periods of land use change. Subsequently, the random forest algorithm is utilized to determine the extent of influence exerted by each factor on the expansion of various land use types, thereby yielding the probability of development for each land use type. The determination of the simulation benchmark for the spatiotemporal dynamics of land use changes was contingent upon the distribution of LULC in the beginning year. Moreover, the development probability serves as a constraint for simulation. The PLUS model integrates the rule-mining framework based on a land expansion analysis strategy (LEAS) module with the cellular automata based on multi-type random patch seeds (CARS) to improve the capacity to replicate realistic landscape patterns.

3.1.2 Influence factors and simulation process of LULC

First, the PLUS model employed the Markov chain to compute future land use demand. Following this, the LEAS module within the PLUS model was used to enhance the land-use variation forecasting approach by integrating variables. In light of data accessibility, a suite of 14 variables including DEM, slope, average annual precipitation, average annual temperature, NDVI, soil type, POP, GDP, and normalized distances from each pixel to various kinds of roads, administrative quarters, and railway stations were selected as influence factors for LULC variation prediction (Fig. 3). Finally, the CARS module within the PLUS model was employed to predict the LULC of the Bosten Lake Basin in 2030.
Fig. 3 Spatial distribution of the selected influence factors of LULC variation in the Bosten Lake Basin

3.1.3 Accuracy verification

Prior to the LULC forecast simulation of the Bosten Lake Basin in 2030, the PLUS model was utilized to forecast land utilization in 2020 by employing the LULC data of 2000 and 2010 as a basis. The validation phase of PLUS model involved the comparison of the simulated data with the real LULC data of 2020. The validation results included a kappa coefficient of 0.911 and an overall accuracy of 0.946, demonstrating that the precision of the simulation of PLUS model satisfied the necessary degree of precision.

3.1.4 Projection of LULC in 2030

Based on the growth trends and policies in the Bosten Lake Basin, this study established and simulated three development scenarios: a natural development scenario (S1), an ecological protection scenario (S2), and a cultivated land protection scenario (S3).
The estimated land demand in 2030, based on scenario S1, was determined by analyzing the historical pattern of changes in LULC. This trend was forecasted using the Markov chain module within the PLUS model, which analyzed the LULC change pattern between 2010 and 2020. The next step involved simulating these scenarios by adjusting the LULC transfer matrix within the CARS module.
The scenario S1 maintained default settings. To regulate the rapid expansion of construction land and cultivated land, as well as to increase the development potential of forest and grassland, strict restrictions were placed on the conversion of forest, grassland, and water body to cultivated land, construction land, and unused land in the scenario S2. The scenario S3 imposed significant restrictions on the transfer of cultivated land to other land use types.

3.2 Calculation of water yield

3.2.1 InVEST model

The assessment of water yield in the Bosten Lake Basin was conducted using the water yield module within the InVEST model. The InVEST model is an ecosystem service assessment and trade-off model jointly developed by the Stanford Woods Institute for the Environment and the World Wildlife Fund (WWF). The water yield module within the InVEST model assesses the water yield at each grid within one region by utilizing the Budyko curve and water balance equations. The fundamental formula employed for this purpose is as follows:
WY x = 1 – AET x P x × P x
where WYx is the water yield in grid cell x (mm); AETx is the actual evapotranspiration in grid cell x (mm); and Px is the average annual precipitation in gird cell x (mm).
The water balance equations account for the actual evapotranspiration of vegetation across different land use types. Actual evapotranspiration, which relates actual evapotranspiration to average annual precipitation and PET within a region, was estimated using the Budyko curve (Baw-puh, 1981; Zhang et al., 2004).
AET x P x = 1 + PET x P x − 1 + PET x P x ω x 1 / ω x  
where PETx is the potential evapotranspiration in grid cell x (mm); and ωx is the nonphysical parameter that characterizes natural climate-soil properties, expressed as (Yang et al., 2008; Donohue et al., 2012):
ω x = Z × AWC x P x + 1.25
where Z is an empirical constant, which captures the local precipitation pattern and additional hydrogeological characteristics, and its value ranges from 1 to 30; and AWCx is the effective water conservation capacity of plants in grid cell x (mm), which is calculated as:
AWC x = min r e s t l a y e r d e p t h , r o o t d e p t h × PAWC x
where restlayerdepth is the depth of root restricting layer where root penetration is strongly inhibited (mm); rootdepth is the depth at which 95% of the root biomass of a plant species is discovered (mm); and PAWCx is the available amount of water for plants in grid cell x (mm). The PAWCx can be estimated based on the physical and chemical properties of soil and its computation is as follows:
PAWC x = 54.509 − 0.132 × sand x − 0.003 × sand x 2 − 0.055 × silt x   − 0.006 × silt x 2   −                                                                             0.738 × clay x + 0.055 × silt x   − 0.006 × silt x 2 − 0.738 × clay x     + 0.007 × clay x 2 −                                                                             2.688 × OM x     + 0.501 × OM x 2
where sand, silt, and clay are the proportion of sand, silt, and clay in the soil in grid cell x, respectively (%); and OM is the proportion of organic matter of soil in grid cell x (%). All the values of the four soil variables in Equation 5 were normalized to fractions ranging from 0 to 1 in this study (An et al., 2022).

3.2.2 Validation methods

The process of simulating water yield is a spatialization of the sum of surface water and groundwater in one region, excluding inflow from external areas. Therefore, the total volume of simulated water yield from the InVEST model may be more closely related to the sum of basin surface runoff and groundwater minus inflow from external areas. Due to the lack of observed data in the Bosten Lake Basin, the relatively consistent and data-rich Kaidu River Basin was selected for validation (Table 3).
Table 3 Validation result of water yield in the Kaidu River Basin simulated by the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model
Result Area of
catchment
(km2)		 Average annual runoff
(×108 m3)		 Average annual groundwater
(×108 m3)		 Average annual snowmelt
(×108 m3)		 Total volume of water yield
(×108 m3)
The reference volume of water yield (Li et al., 2003; Chen et al., 2022; Chen et al., 2023) 18,541 36.59 13.00 9.96 39.63
The simulated volume of water yield by InVEST model 39.64
By referencing observed data in the Kaidu River Basin from previous studies, an approximation of actual volume of water yield was obtained. The reference volume of water yield was derived by summing surface runoff and groundwater and subtracting snowmelt (Li et al., 2003; Chen et al., 2022; Chen et al., 2023). We simulated the total volume of water yield of the Kaidu River for the years of 2000, 2010, and 2020 by InVEST model, and then averaged the results for comparison with the reference water yield. As an empirical constant input into the InVEST model, Z can fine-tune the simulated result to better match reality; in this study, when Z=10, the absolute error between the simulated result and the reference water yield reached approximately 0.25%, meeting the precision criteria.

3.3 Calculation of water conservation

Water conservation was determined by the InVEST model, with adjustments made for the saturated hydraulic conductivity of soil, topographic index, and flow coefficient (Hu et al., 2020), and the formula is:
WC x = min 1 , 249 υ × min 1 , 0.9 × TI x 3 × min 1 , K x 300 × WY x
where WCx is the water conservation in grid cell x (mm); υ is the flow coefficient; TIx is the topographic index in grid cell x; and Kx is the hydraulic conductivity of saturated soil in grid cell x (cm/d). The flow coefficient (υ) was calculated by multiplying the value from the National Engineering Handbook Chapter 9 by 1000 (Kent, 1972; Donohue et al., 2012).
TIx can be obtained through DEM calculation, and the formula is:
TI x = log n d × Q x
where n is the number of girds in the study area; d is the depth of soil layer (mm); and Qx is the slope percentage in grid cell x (%).
The hydraulic conductivity of saturated soil was determined by employing the formula as follows (Cosby et al., 1984):
K x = 60.96 × 10 – 0.6 + 0.0126 × sand x − 0.0064 × clay x

3.4 Calculation of contribution index

The spatial pattern of LULC exhibits substantial spatial heterogeneity; consequently, the mechanisms that influence water yield also demonstrate spatial variability (Wu et al., 2023). By employing a contribution index (CI), the extent to one land use type influencing water yield can be quantified:
CI i = WY i ¯ − WY ¯ × S i S
where CIi is the contribution index of ith land use type to water yield;
WY i ¯
 is the average water yield of the ith land use type (mm);
WY ¯
 is the average water yield of the whole study area (mm); Si is the area of the ith land use type (m2); and S is the area of the whole study area (m2). CIi≥0 indicates that the ith land use type contributes positively to the increase of water yield; otherwise, it contributes negatively to the increase of water yield. Similarly, CI also can be employed to determine the contribution degree of one land use type to water conservation.

3.5 Geodetector

The core feature of Geodetector is the detection of spatial anisotropy, which can be used to elucidate the factors that influence water yield (Wang and Xu, 2017). This study utilized a factor detector and interaction detector to determine the primary components and their interactions that contribute to the geographical variability of water yield and water conservation in the Bosten Lake Basin.
q = 1 − ∑ h = 1 L N h σ h 2 N σ 2
where q is the influence degree of one driving factor on the spatial heterogeneity of water yield; h is the stratum of water yield and driving factors, which is obtained through reclassification (h=1, 2, …, L); Nh is the number of units in stratum h, which corresponds to the number of grids in stratum h of the study area; N is the total number of units in the entire study area, which includes all grids within the study area;
σ h 2
 is the variance of water yield in stratum h; and
σ 2
is the variance of water yield across the entire study area.
The purpose of interaction detection is to ascertain the nature of the interaction between pairs of distinct driving factors and to determine whether the combined explanatory power of each pair of driving factors on dependent variables is independent, mutually reinforcing, or mutually weakening (Table 4). The degree of interaction between two driving factors is characterized by q-value ([q(X1∩X2)], where X1 and X2 represent the driving factors involved in the interaction detection).
Table 4 Criterion and type of interaction between pair of driving factors
Criterion Interaction type
q(X1∩X2)<Min[q(X1), q(X2)] Nonlinear weakening
Min[q(X1), q(X2)]<q(X1∩X2)<Max[q(X1), q(X2)] Single-factor nonlinear weakening
q(X1∩X2)>Max[q(X1), q(X2)] Dual-factor enhancement
q(X1∩X2)=q(X1)+q(X2) Independence
q(X1∩X2)>q(X1)+q(X2) Nonlinear enhancement

Note: q represents the degree of interaction between two factors; X1 and X2 represent the factors involved in interaction.

4 Results

4.1 Spatial and temporal changes in LULC

The predominant land use types in the Bosten Lake Basin were grassland and unused land during 2000-2020 (Figs. 4 and 5). Grassland covered approximately 50.12%, 49.70%, and 48.61% of the whole basin area in 2000, 2010, and 2020, respectively, closing to half of the basin's total area. Unused land accounted for 41.06%, 39.82%, and 39.74% of the whole basin area in 2000, 2010, and 2020, respectively. Cultivated land constituted the third largest portion of the basin, while water body, forest, glacier, and construction land comprised negligible portions. Throughout this span of two decades, the area of cultivated land and construction land showed noticeable increase. The area of cultivated land substantially increased by 77.17% over a span of twenty years, i.e., an increase of 2525.13 km2 (Table 5). Construction land experienced the greatest growth of 150.25% (329.56 km2); while unused land, forest, grassland, and glacier exhibited a declining trend on an annual basis. The predominant conversion of grassland to farmland occurred mostly in the southern part of the Bosten Lake Basin, as well as in the western and northern areas around the Bosten Lake.
Fig. 4 Spatial distribution of LULC in the Bosten Lake Basin in 2000 (a), 2010 (b), and 2020 (c)
Fig. 5 LULC transfer in the Bosten Lake Basin from 2000 to 2020. (a), 2000-2010; (b) 2010-2020; (c), 2000-2020.
Table 5 Area of each land use type in the Bosten Lake Basin in 2000, 2010, 2020, and the three predicted scenarios of 2030
Land use type Aera (km2)
2000 2010 2020 2030
S1 S2 S3
Cultivated land 3272.37 4968.81 5797.50 6589.69 5595.73 6589.69
Forest 1519.75 1134.75 1049.79 982.94 1089.44 982.89
Grassland 40,250.41 39,913.98 39,040.11 38,305.29 39,240.16 38,317.19
Water body 1243.58 1322.97 1363.04 1402.40 1374.12 1402.42
Glacier 829.35 599.90 597.82 600.74 598.86 601.37
Construction land 219.34 389.20 548.90 602.58 585.32 590.09
Unused land 32,977.25 31,982.45 31,914.88 31,828.43 31,828.43 31,828.42

Note: S1, natural development scenario; S2, ecological protection scenario; S3, cultivated land protection scenario.

The spatial land utilization in 2030 under the three scenarios was determined using the PLUS model. In accordance with the three designated scenario settings, the analysis revealed a discernible upward trajectory in the extent of water body and construction land by the year 2030 (Fig. 6). Conversely, there was a noticeable decline in the area of unused land, while the area occupied by glacier remained relatively stable. In comparison with 2020, the area of cultivated land, water body, and construction land increased by 792.19, 39.36, and 53.68 km2, respectively in scenario S1 (Table 5). In contrast, the area of forest, grassland, and unused land decreased by 66.85, 734.82, and 86.45 km2, respectively. Notably, the area of glacier remained largely unchanged (2.92 km2). Compared with 2020, the area of forest and grassland increased by 39.65 and 200.05 km2, respectively in scenario S2 of 2030. Conversely, a slight decrease was observed in construction land (36.42 km2). In accordance with scenario S3, there was a decrease area of 722.92 km2 in grassland but an increase of 41.19 km2 in construction land. Additionally, the remaining land use types experienced changes similar to those observed in scenario S1.
Fig. 6 Spatial distribution of LULC in the Bosten Lake Basin in the three predicted scenarios of 2030. (a), scenario S1 (natural development scenario); (b), scenario S2 (ecological protection scenario); (c), scenario S3 (cultivated land protection scenario).

4.2 Spatial and temporal variations in water yield and water conservation

4.2.1 Spatial and temporal variations in water yield

The water yield in the Bosten Lake Basin was calculated by incorporating the InVEST model and LULC data. From Figure 7, we can find that water yield in the Bosten Lake Basin exhibited interannual variability while maintaining consistent spatial distribution characteristics. Specifically, high water yield values were observed in Hejing County, the northern part of Luntai County, and the northern part of Hesud County. Moreover, the water yield in the Bosten Lake Basin displayed a north-south gradient, i.e., higher values were seen in the northern part of the basin, whereas lower values were observed in the southern part. Additionally, water yield was found to be greater in elevated regions and lower in regions with lower elevations. According to the results, the average water yield in the Bosten Lake Basin in 2000, 2010, and 2020 was 112.19, 110.43, and 110.61 mm, respectively, demonstrating a decreasing and then slightly increasing temporal trend. The average water yield in scenarios S1, S2, and S3 of 2030 was 110.64, 110.58, and 110.64 mm, respectively. The average water yield among the three scenarios did not significantly differ.
Fig. 7 Spatial distribution of water yield in the Bosten Lake Basin in 2000 (a), 2010 (b), 2020 (c), and the three predicted scenarios of 2030 (d, e, and f). The glacier and lake areas, which are shown as white areas, were excluded from the calculation of water yield (CAO et al., 2023).

4.2.2 Spatial and temporal variations in water conservation

The spatial distribution of water conservation in the Bosten Lake Basin exhibited a similar north-south gradient pattern to water yield; specifically, the water conservation values in the northern part of the Bosten Lake Basin (especially the northeastern part) were greater than that in the southern part (Fig. 8). The average water conservation of the Bosten Lake Basin in 2000, 2010, and 2020 was 4.92, 4.80, and 4.79 mm, respectively, exhibiting a decreasing and then stabilizing temporal trend. The average water conservation of scenarios S1, S2, and S3 of 2030 was 4.80, 4.79, and 4.79 mm, respectively.
Fig. 8 Spatial distribution of water conservation in the Bosten Lake Basin in 2000 (a), 2010 (b), 2020 (c), and the three predicted scenarios of 2030 (d, e, and f). The glacier and lake areas, which are shown as white areas, were excluded from the calculation of water conservation (CAO et al., 2023).

4.3 Contribution of each land use type to water yield and water conservation

4.3.1 Contribution of each land use type to water yield

The quantification of the contribution of each land use type to water yield in the Bosten Lake Basin from 2000 to 2020 was achieved by aggregating land use types into a statistical unit. As depicted in Figure 9, the CI for grassland, water body, and glacier exhibited values greater than zero across all three time periods, indicating that these land use types made a positive contribution to the increase of water yield in the Bosten Lake Basin. Grassland had the highest contribution to the increase of water yield in the Bosten Lake Basin. Cultivated land, forest, and construction land had CI values less than zero, indicating that these three land use types made a negative contribution to the increase of water yield in the Bosten Lake Basin. Crops in cultivated land and trees in forest need plenty of water to grow, thus affecting the water yield. The rapid expansion of cultivated land and construction land led to an increase in negative contributions. The shift from a positive contribution of unused land in 2000 to a negative contribution in 2010 and 2020 was primarily attributable to a reduction in the quantity of unused land in the northern part of the basin, which is situated in an area with high water yield. The CI of grassland increased from 3.22 in 2000 to 9.44 in 2010, which, along with the natural factors, may have also been due to the change in grassland cover, thus made a more positive contribution.
Fig. 9 Contribution index (CI) of each land use type to water yield in the Bosten Lake Basin in 2000, 2010, and 2020
The average annual water yield for each land use type in the Bosten Lake Basin from 2000 to 2030 was determined (Table 6). Glacier had the highest average annual water yield, while cultivated land and construction land had the lowest. Additionally, the average annual water yield of grassland and water body was estimated to be greater than that of forest and unused land.
Table 6 Average annual water yield of each land use type in the Bosten Lake Basin in 2000, 2010, 2020, and the three predicted scenarios of 2030
Land use type Average annual water yield (mm)
2000 2010 2020 2030
S1 S2 S3
Cultivated land 1.76 1.83 2.19 2.22 2.17 2.24
Forest 55.52 74.34 68.30 69.91 69.29 69.82
Grassland 118.41 128.78 128.58 130.99 127.84 130.94
Water body 131.21 205.75 198.26 200.96 196.90 200.69
Glacier 409.89 421.49 420.93 419.50 419.51 419.28
Construction land 5.81 0.55 0.34 1.41 0.66 1.27
Unused land 117.03 104.45 109.36 109.57 109.57 109.57

4.3.2 Contribution of each land use type to water conservation

Grassland, glacier, and water body made positive contributions to water conservation in the Bosten Lake Basin from 2000 to 2020 (Fig. 10). Negative contributions to the increase of water conservation were made by cultivated land, construction land, and unused land. Unused land made the most significant negative contribution. The CI value of forest was less than zero in 2000, and then greater than zero thereafter. The CI values of forest, water body, glacier, and construction land fell within a range of -0.05-0.02. However, these values were not considered statistically significant due to the relatively tiny fraction of land area they occupied.
Fig. 10 CI of each LULC land use type to water conservation in the Bosten Lake Basin in 2000, 2010, and 2020
According to the data presented in Table 7, the average annual water conservation in the Bosten Lake Basin from 2000 to 2030 varied across different land use types. The average annual water conservation of grassland, water body, and glacier were at a relatively high level, while the average annual water conservation of cultivated land and construction land were at a low level. The average annual water conservation of forest showed a certain upward trend.
Table 7 Average annual water conservation of each land use type in the Bosten Lake Basin in 2000, 2010, 2020, and the three predicted scenarios of 2030
Land use type Average annual water conservation (mm)
2000 2010 2020 2030
S1 S2 S3
Cultivated land 0.32 0.32 0.35 0.37 0.35 0.36
Forest 2.48 4.93 4.84 5.00 4.84 4.99
Grassland 6.05 6.62 6.67 6.79 6.64 6.78
Water body 5.37 7.13 6.83 7.56 6.78 7.09
Glacier 6.25 5.16 5.08 5.25 5.09 5.26
Construction land 0.96 0.32 0.30 0.49 0.30 0.50
Unused land 4.10 3.21 3.32 3.32 3.32 3.32

4.4 Analysis of the driving factors of water yield and water conservation variation

Based on the selection of driving factors in existing research, encompassing considerations of topography, climatic conditions, vegetation coverage, and socioeconomic factors, alongside an assessment of data availability and its relevance to the model, we chose DEM, slope, NDVI, temperature, precipitation, PET, POP, GDP, and NTL as driving factors of changes in water yield and water conservation (Reheman et al., 2023; Yang et al., 2023). As depicted in Figure 11, the q-values pertaining to the impacts of different driving factors on water yield within the Bosten Lake Basin decreased as the following order: PET (0.8979)> precipitation (0.8949)>DEM (0.8757)>temperature (0.8747)>slope (0.2976)>NDVI (0.1383)>GDP (0.0839)>NTL (0.0071)>POP (0.0010). Based on the findings of factor detector, the primary factor influencing the spatial variability of water yield in the Bosten Lake Basin was PET. Subsequently, precipitation, DEM, and temperature exhibited comparable level of influence. Conversely, the remaining factors demonstrated relatively minor contributions to the spatial variations in water yield. The results of the interaction detection analysis revealed that there were nonlinear enhancements of the interaction between NDVI∩POP and NDVI∩GDP, and other interaction detection results exhibited dual-factor enhancement. Among these factors, temperature, precipitation, DEM, and PET exhibited the most significant interactions with other variables.
Fig. 11 Interactive detection of driving factors of water yield variation in the Bosten Lake Basin. PET, potential evapotranspiration; NDVI, normalized difference vegetation index; NTL, nighttime light. ***, significant at P<0.001 level; **, significant at P<0.01 level.
As depicted in Figure 12, the q-values pertaining to the impacts of different driving factors on water conservation in the Bosten Lake Basin decreased as the following order: precipitation (0.5621)>temperature (0.5561)>PET (0.5525)>DEM (0.5367)>NDVI (0.1582)>slope (0.0873)> GDP (0.0634)>NTL (0.0050)>POP (0.0007). According to the results of factor detection analysis, precipitation emerged as the primary determinant influencing the spatial variations in water conservation. Equally significant contributions were made by temperature, DEM, and PET, while the remaining driving factors exhibited comparatively minor effects. The findings from interaction detector indicated that there were nonlinear enhancements in the relationships among NDVI∩POP, NDVI∩GDP, NDVI∩NTL, and NDVI∩slope. Furthermore, other interaction detection results exhibited dual-factor enhancement. Among these factors, temperature, precipitation, DEM, and PET exhibited the most significant interactions with the other variables.
Fig. 12 Interactive detection of driving factors of water conservation variation in the Bosten Lake Basin. ***, significant at P<0.001 level; **, significant at P<0.01 level.
Geodetector analysis revealed that precipitation, PET, temperature, and DEM were the primary driving factors influencing the changes in water yield and water conservation in the Bosten Lake Basin. The factor detector and the interaction detector had significantly greater q-values for these four driving factors than the other factors, suggesting that climatic conditions and elevation played important roles in determining water yield and water conservation in this basin.

4.5 Delineation of water conservation function zone

There was significant variation in the allocation of water conservation efforts across diverse regions in the Bosten Lake Basin. The significance of water conservation function in a given area is determined by the volume of water it can conserve. Consequently, it is feasible to categorize regions into different levels of water conservation function importance based on their varying water conservation value. We referred to the classification method outlined in the ''Technical Guidelines for the Delineation of Ecological Red Line'' and employed the natural break method to delineate the water conservation function zones in the Bosten Lake Basin (Ministry of Environmental Protection of the People's Republic of China and National Development and Reform Commission of the People's Republic of China, 2017). Water conservation function usually refers to the ability of an ecosystem to keep water in the system under certain time and space conditions. We classified five levels of importance of water conservation function (I: generally important level, Ⅱ: mildly important level, Ⅲ: moderately important level, Ⅳ: highly important level, and Ⅴ: extremely important level) for water conservation function zoning, based on the values of water conservation of the Bosten Lake Basin in 2020 (Table 8).
Table 8 Classification of water conservation function important level
The level of importance of water conservation function Level Water conservation (mm)
Generally important Ⅰ [0, 5.00)
Mildly important Ⅱ [5.00, 10.00)
Moderately important Ⅲ [10.00, 20.00)
Highly important Ⅳ [20.00, 35.00)
Extremely important Ⅴ [35.00, ∞)
The areas designated as generally important level of water conservation function were the largest, with a total area of 48,897.00 km2. Influenced by topography, PET, and precipitation, the areas designated as generally important level of water conservation function were predominantly located in the southern part of the basin (Fig. 13). Areas with water conservation higher than 35 mm are regarded as extremely important water conservation function areas and should be prioritized in policies and management decisions to ensure their protection. The majority of areas that designated as extremely important level of water conservation function were located in the northeastern part of Hejing County, with a distribution area of 509.00 km2. The areas classified as mildly important, moderately important, highly important, and extremely important levels of water conservation function are primarily concentrated in Hejing County; however, the areas also extended to Korla City, Luntai County, the northern portion of Hesud County, and the western portion of Yanqi Hui Autonomous County. The Bosten Lake Basin is critical water conservation function area due to relatively high average annual precipitation, low temperature, and low PET.
Fig. 13 Spatial distribution of the level of importance of water conservation function. The glacier and lake areas, which are shown as white areas, were excluded from the calculation of water conservation (CAO et al., 2023).

5 Discussion

5.1 Driving factors of water yield and water conservation

In this study, we estimated the water yield and water conservation and analyzed their spatial and temporal variations by the InVEST model from 2000 to 2020. This study showed that precipitation, temperature, DEM, and PET had significant effects on changes in water yield and water conservation. These findings provide new contributions to the understanding of water ecosystem services and valuable insights for sustainable water resource management for the Bosten Lake Basin. In the Bosten Lake Basin, grassland is a key contributor to water yield and water conservation. However, from 2000 to 2020, the area of grassland continued declining; meanwhile, the water yield and water conservation did not show a continuous downward trend. The main reason for this phenomenon is that a large amount of grassland has been converted to cultivated land in the Yanqi Basin and Korla area in the southern part of the Bosten Lake Basin (Wang et al., 2015). It is worth noting that the areas where grassland disappeared were not areas with high value of water yield. These locations are characterized by oasis environments and exhibit a significant level of human activity. Significant portions of grassland were allocated for economic development as a consequence of the implementation of the Great Western Development Strategy and the Belt and Road Initiative, resulting in a significant shrinkage of grassland (Reheman et al., 2023). The annual fluctuation of precipitation showed remarkable consistency with the trends in water yield and water conservation (Fig. 14). At the same time, changes in PET and temperature likewise play a key role in fluctuations of water yield and water conservation. Existing research results have clearly indicated that precipitation, temperature, and PET are the main factors affecting water yield (Lu et al., 2013; Sun et al., 2022). These findings were consistent with the results of our geodetector analyses, further reinforcing the centrality of these variables to the regional hydrological cycle. This study also considered the influence of topographic factors on water yield and water conservation, which have a significant impact on the climatic conditions, not only in its regulation of hydrothermal factors, but also in its regulation of vegetation types, the accumulation of understory vegetation, and the physical and chemical properties of soils (Wang and Shen, 2013; Maurya et al., 2016). Altitude influenced ecological and hydrological processes to a certain extent, and in the Bosten Lake Basin, the distribution pattern of elevation was similar to that of water yield.
Fig. 14 Average annual precipitation, temperature, and PET in the Bosten Lake Basin from 2000 to 2020

5.2 Water yield and water conservation under different scenarios of 2030

Multi-scenario-based projections of water yield and water conservation can provide a scientific basis for the development of regional water ecosystem protection policies. The projected water yield and water conservation of 2030 changed little comparing with 2020. Among the different prediction scenarios, the water yield under scenario S2 was surprisingly lower than that under the other two scenarios. The differences of water yield during 2000-2030 were mainly attributed to the change of LULC based on the simulated results of InVEST model. We compared the distribution of LULC under different scenarios and found that the change of LULC was concentrated in the southern part of the basin. The results of site expansion under scenario S1 were close to those under the combination of scenario S2 and scenario S3 (Fig. 15). The scenarios S1 and S3 have similar water yield and were higher than scenario S2, reflecting that even though grassland has expanded in the southern region under scenario S2, its contribution to water yield might not be significant. The grassland in the southern part of the basin was mainly distributed sparsely on the edges of oases, which are common landscape in Xinjiang Uygur Autonomous Region, China and Central Asia (Dixon et al., 2014). Changes in grassland in the northern part of the basin were not significant based on prediction of PLUS model. Therefore, the influence of grassland in the northern part of the basin on water yield and water conservation still needs in-depth research. The distribution pattern of predicted water yield under the three scenarios remains largely unchanged compared with that in 2020, demonstrating that the current management policy remains valid in the short term.
Fig. 15 LULC transfer in the Bosten Lake Basin between 2020 and the three predicted scenarios of 2030. (a), from 2020 to the scenario S1 of 2030; (b), from 2020 to the scenario S2 of 2030; (c), form 2020 to the scenario S3 of 2030.

5.3 Limitations and future research

Although a series of valuable research results were achieved by coupling the PLUS model with the InVEST model, there is still some uncertainty in the research of water yield and water conservation in the Bosten Lake Basin. Some of the general factors focus on the limitations of the InVEST model design, such as the failure to adequately account for the effects of complex subsurface factors and the potential introduction of errors in the use of uniform Z coefficient for study areas that span climatic zones (Vigerstol and Aukema, 2011; Redhead et al., 2016; Wei et al., 2022). In addition, the lack of precision in the resolution of the model input data may also lead to errors, although this may not have much impact on the macro interpretation of simulated results (Yang et al., 2021). The purpose of the coupled PLUS-InVEST model in this study is to better understand the changes in water yield and water conservation in the Bosten Lake Basin under multiple prediction scenarios. However, when using the InVEST model for water yield projections for future scenarios, we noted that the LULC in the model was obtained based on projections; while precipitation and PET were obtained using averages from 2000 to 2020, which does not consider the effects of global climate change. Coupled Model Intercomparison Project (CMIP) has now made multiple models publicly available for users to simulate climate change (O'Neill et al., 2016). Therefore, in the future research, the multi-scenario impacts of climate change should be considered and integrated with the latest climate models. This will help us to provide more accurate and informative projections of water yield and water conservation.

6 Conclusions

This study utilized coupled InVEST and PLUS models to dynamically evaluate the spatial and temporal fluctuations in water yield and water conservation in the Bosten Lake Basin, which is located in the arid area of Northwest China. A replicable paradigm is proposed for land-use decision-making in environmentally sensitive places, considering the viewpoint regarding alterations in ecosystem services. The findings showed that, there was a significant increase in cultivated land and construction land, and a decrease in grassland, forest, and unused land in the Bosten Lake Basin from 2000 to 2020. In the three projected scenarios of 2030, there was a discernible upward trajectory in the extent of water body and construction land and a noticeable decline in the area of unused land. Specifically, cultivated land expanded while grassland contracted under both scenario S1 and scenario S3; conversely, both forest and grassland experienced growth in scenario S2, whereas cultivated land showed a decline. The LULC changes mainly occurred in the oasis area in the southern part of the basin, but did not change the spatial patterns of water yield and water conservation significantly. Grassland contributed significantly to water yield and water conservation in the Bosten Lake Basin, but did not contribute much in the low-value areas, i.e., the southern part of the basin. The distributions of both water yield and water conservation in the Bosten Lake Basin were higher in the north and lower in the south, and the spatial distribution characteristics remained relatively stable over a multi-year time span. The water yield in the Bosten Lake Basin showed a declining trend from 112.19 mm in 2000 to 110.43 mm in 2010 and then a slightly increasing to 110.61 mm in 2020. Similarly, the water conservation in the Bosten Lake Basin exhibited a decreasing trend from 4.80 mm in 2000 to 4.79 mm in 2010, and then stabilizing at 4.79 mm in 2020. Majority of Hejing County was an important water conservation function area, with the Xiaoyouledusi Basin, which is located in the eastern part of the county, being particularly important and thus in need of adequate attention and protection. This study exists limitations, including the inability of the InVEST model to account for complex subsurface factors and limitations in data precision. In future research, incorporating future climate change factors into multi-scenario predictions will enable more accurate forecasts. In summary, the findings of this study provide a reference for the sustainable management and conservation of the watersheds in arid areas, offering insights and guidance for future research and practice.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This research was supported by the Special Project for the Construction of Innovation Environment in the Autonomous Region (2022D04007) and the National Natural Science Foundation of China (42361030).

Author contributions

Conceptualization: CHEN Jiazhen, KASIMU Alimujiang, REHEMAN Rukeya; Methodology: CHEN Jiazhen, KASIMU Alimujiang, REHEMAN Rukeya; Formal analysis: CHEN Jiazhen; Writing - original draft preparation: CHEN Jiazhen, KASIMU Alimujiang; Writing - review and editing: KASIMU Alimujiang; Funding acquisition: KASIMU Alimujiang; Visualization: CHEN Jiazhen; Investigation: WEI Bohao, HAN Fuqiang, ZHANG Yan; Supervision: KASIMU Alimujiang. All authors approved the manuscript.

参考文献

原文顺序 | 文献年度倒序 | 文中引用次数倒序
[1]
Aghsaei H, Mobarghaee Dinan N, Moridi A, et al. 2020. Effects of dynamic land use/land cover change on water resources and sediment yield in the Anzali wetland catchment, Gilan, Iran. Science of the Total Environment, 712: 136449, doi: 10.1016/j.scitotenv.2019.136449.

[2]
Allen R G, Pereira L S, Raes D, et al. 1998. FAO Irrigation and Drainage Paper No. 56—Crop Evapotranspiration (Guidelines for Computing Crop Water Requirements). Rome: FAO (Food and Agriculture Organization of the United Nations), 156.

[3]
An L, Zhong S, Shen L. 2022. Dynamic effects of climate and land use policies on water yield in drylands—A case study in the northwest of China. Water, 14: 3940, doi: 10.3390/w14233940.

[4]
Baw-puh F. 1981. On the calculation of the evaporation from land surface. Chinese Journal of Atmospheric Science, 5(1): 23-31. (in Chinese)

[5]
Cao Y J, Ma Y G, Bao A M, et al. 2023. Evaluation of the water conservation function in the Ili River Delta of Central Asia based on the InVEST model. Journal of Arid Land, 15(12): 1455-1473.

[6]
Chen H B, Wang T X, An J, et al. 2023. Changes of glacier area in Kaidu River Basin and its contribution to runoff during melting seasons. Journal of China Hydrology, doi: 10.19797/j.cnki.1000-0852.20230131. (in Chinese)

[7]
Chen S X, Rusuli Y, Zhang F, et al. 2022. Impacts of climate change and human activities on runoff in the upper reaches of Kaidu River based on SWAT model. Hubei Agricultural Sciences, 61(5): 171-176. (in Chinese)

[8]
Cosby B J, Hornberger G M, Clapp R B, et al. 1984. A statistical exploration of the relationships of soil moisture characteristics to the physical properties of soils. Water Resources Research, 20(6): 682-690.

[9]
Dixon A P, Faber-Langendoen D, Josse C, et al. 2014. Distribution mapping of world grassland types. Journal of Biogeography, 41(11): 2003-2019.

[10]
Donohue R J, Roderick M L, McVicar T R. 2012. Roots, storms and soil pores: Incorporating key ecohydrological processes into Budyko's hydrological model. Journal of Hydrology, 436-437: 35-50.

[11]
Duolaiti X, Kasimu A, Reheman R, et al. 2023. Assessment of water yield and water purification services in the arid zone of Northwest China: The case of the Ebinur Lake Basin. Land, 12(3): 533, doi: 10.3390/land12030533.

[12]
Gao L N, Tao F, Liu R R, et al. 2022. Multi-scenario simulation and ecological risk analysis of land use based on the PLUS model: A case study of Nanjing. Sustainable Cities and Society, 85: 104055, doi: 10.1016/j.scs.2022.104055.

[13]
Guo J T, Zhang Z Q, Wang S P, et al. 2014. Appling SWAT model to explore the impact of changes in land use and climate on the streamflow in a watershed of Northern China. Acta Ecologica Sinica, 34(6): 1559-1567. (in Chinese)

[14]
Hu C M, Narengerile, You L. 2019. Study of ecological water levels of Bosten Lake for water quality management. Acta Ecologica Sinica, 39(2): 748-755. (in Chinese)

[15]
Hu W M, Li G, Gao Z H, et al. 2020. Assessment of the impact of the Poplar Ecological Retreat Project on water conservation in the Dongting Lake wetland region using the InVEST model. Science of the Total Environment, 733: 139423, doi: 10.1016/j.scitotenv.2020.139423.

[16]
Hu Y X, Yu X X, Liao W, et al. 2022. Spatio-temporal patterns of water yield and its influencing factors in the Han River Basin. Resources and Environment in the Yangtze Basin, 31(1): 73-82. (in Chinese)

[17]
Janjić J, Tadić L. 2023. Fields of application of SWAT hydrological model—A review. Earth, 4(2): 331-344.

[18]
Kadeer R, Rusuli Y, Wufu A, et al. 2017. MODIS data-based study of the spatial distribution of land surface temperature in Bosten lake basin area. Acta Scientiarum Naturalium Universitatis Sunyatseni, 56(5): 127-138. (in Chinese)

[19]
Karnieli A, Ben-Asher J. 1993. A daily runoff simulation in semi-arid watersheds based on soil water deficit calculations. Journal of Hydrology, 149(1-4): 9-25.

[20]
Kent K M. 1972. Chapter 15:Travel time, time of concentration and lag. SCS National Engineering Handbook, Section 4, Hydrology. [2024-01-05]. https://babel.hathitrust.org/cgi/pt?id=mdp.39015012634476&seq=9.

[21]
Li M Y, Liang D, Xia J, et al. 2021. Evaluation of water conservation function of Danjiang River Basin in Qinling Mountains, China based on InVEST model. Journal of Environmental Management, 286: 112212, doi: 10.1016/j.jenvman.2021.112212.

[22]
Li Y A, Tan Y, Jiang F Q, et al. 2003. Study on hydrological features of the Kaidu River and the Bosten Lake in the second half of 20th century. Journal of Glaciology and Geocryology, 2: 215-218. (in Chinese)

[23]
Liang H W, Kasimu A, ZHAO H M, et al. 2022. Analysis of the spatial and temporal differences in surface temperature and the contribution of surface coverage in the urban agglomeration on the northern slope of the Tianshan Mountains. Arid Zone Research, 39(2): 388-399. (in Chinese)

[24]
Liang X, Guan Q F, Clarke K C, et al. 2021. Understanding the drivers of sustainable land expansion using a patch-generating land use simulation (PLUS) model: A case study in Wuhan, China. Computers, Environment and Urban Systems, 85: 101569, doi: 10.1016/j.compenvurbsys.2020.101569.

[25]
Lu N, Sun G, Feng X M, et al. 2013. Water yield responses to climate change and variability across the North-South Transect of Eastern China (NSTEC). Journal of Hydrology, 481: 96-105.

[26]
Maurya S, Srivastava P K, Gupta M, et al. 2016. Integrating soil hydraulic parameter and microwave precipitation with morphometric analysis for watershed prioritization. Water Resources Management, 30(14): 5385-5405.

[27]
Ming A, Rowell I, Lewin S, et al. 2021. Key Messages from the IPCC AR6 Climate Science Report. Cambridge Centre for Climate Science, University of Cambridge. Cambridge, UK.

[28]
Ministry of Environmental Protection of the People's Republic of China, National Development and Reform Commission of the People's Republic of China. 2017. Technical Guidelines for the Delineation of Ecological Red Line. [2024-01-12]. https://www.mee.gov.cn/gkml/hbb/bgt/201707/W020170728397753220005.pdf.

[29]
O'Neill B C, Tebaldi C, Vuuren D P, et al. 2016. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geoscientific Model Development, 9(9): 3461-3482.

[30]
Pokhrel Y, Felfelani F, Satoh Y, et al. 2021. Global terrestrial water storage and drought severity under climate change. Nature Climate Change, 11(3): 226-233.

[31]
Redhead J W, Stratford C, Sharps K, et al. 2016. Empirical validation of the InVEST water yield ecosystem service model at a national scale. Science of the Total Environment, 569-570: 1418-1426.

[32]
Reheman R, Kasimu A, Duolaiti X, et al. 2023. Research on the change in prediction of water production in urban agglomerations on the northern slopes of the Tianshan Mountains based on the InVEST-PLUS model. Water, 15(4): 776, doi: 10.3390/w15040776.

[33]
Sun S L, Liu Y B, Chen H S, et al. 2022. Causes for the increases in both evapotranspiration and water yield over vegetated mainland China during the last two decades. Agricultural and Forest Meteorology, 324: 109118, doi: 10.1016/j.agrformet.2022.109118.

[34]
Teo H C, Raghavan S V, He X G, et al. 2022. Large-scale reforestation can increase water yield and reduce drought risk for water-insecure regions in the Asia-Pacific. Global Change Biology, 28(21): 6385-6403.

[35]
Vigerstol K L, Aukema J E. 2011. A comparison of tools for modeling freshwater ecosystem services. Journal of Environmental Management, 92(10): 2403-2409.

[36]
Wang J, Xu C D. 2017. Geodetector: Principle and prospective. Acta Geographica Sinica, 72: 116-134. (in Chinese)

[37]
Wang X F, Chu B Y, Feng X M, et al. 2021. Spatiotemporal variation and driving factors of water yield services on the Qingzang Plateau. Geography and Sustainability, 2(1): 31-39.

[38]
Wang X X, Shen H T, Li X Y, et al. 2013. Concepts, processes and quantification methods of the forest water conservation at the multiple scales. Acta Ecologica Sinica, 33(4): 1019-1030. (in Chinese)

[39]
Wang Y, Chen Y N, Ding J L, et al. 2015. Land-use conversion and its attribution in the Kaidu-Kongqi River Basin, China. Quaternary International, 380: 216-223.

[40]
Wei B H, Kasimu A, Fang C L, et al. 2023a. Establishing and optimizing the ecological security pattern of the urban agglomeration in arid regions of China. Journal of Cleaner Production, 427: 139301, doi: 10.1016/j.jclepro.2023.139301.

[41]
Wei B H, Kasimu A, Reheman R, et al. 2023b. Spatiotemporal characteristics and prediction of carbon emissions/absorption from land use change in the urban agglomeration on the northern slope of the Tianshan Mountains. Ecological Indicators, 151: 110329, doi: 10.1016/j.ecolind.2023.110329.

[42]
Wei Q Q, Abudureheman M, Halike A, et al. 2022. Temporal and spatial variation analysis of habitat quality on the PLUS-InVEST model for Ebinur Lake Basin, China. Ecological Indicators, 145: 109632, doi: 10.1016/j.ecolind.2022.109632.

[43]
Wu K S, Wang D, Lu H Y, et al. 2023. Temporal and spatial heterogeneity of land use, urbanization, and ecosystem service value in China: A national-scale analysis. Journal of Cleaner Production, 418: 137911, doi: 10.1016/j.jclepro.2023.137911.

[44]
Wu L Y, Zhang X, Hao F H, et al. 2020. Evaluating the contributions of climate change and human activities to runoff in typical semi-arid area, China. Journal of Hydrology, 590: 125555, doi: 10.1016/j.jhydrol.2020.125555.

[45]
Xiao Y, Ouyang Z Y. 2019. Spatial-temporal patterns and driving forces of water retention service in China. Chinese Geographical Science, 20(1): 100-111.

[46]
Yang H B, Yang D W, Lei Z D, et al. 2008. New analytical derivation of the mean annual water-energy balance equation. Water Resources Research, 44(3), W03410, doi: 10.1029/2007WR006135.

[47]
Yang H F, Nie S N, Deng S Q, et al. 2023. Evaluation of water yield and its driving factors in the Yangtze River Basin, China. Environmental Earth Sciences, 82: 429, doi: 10.1007/s12665-023-11113-9.

[48]
Yang J, Xie B P, Zhang D G, et al. 2021. Climate and land use change impacts on water yield ecosystem service in the Yellow River Basin, China. Environmental Earth Sciences, 80: 72, doi: 10.1007/s12665-020-09277-9.

[49]
Yu Z T, Wang X J, Zhao C Y, et al. 2015. Carbon burial in Bosten Lake over the past century: Impacts of climate change and human activity. Chemical Geology, 419: 132-141.

[50]
Zhang F, Rusuli Y, Tuersun A. 2021. Spatio-temporal change of ecosystem service value in Bosten Lake Watershed based on land use. Acta Ecologica Sinica, 41(13): 5254-5265. (in Chinese)

[51]
Zhang L, Hickel K, Dawes W R, et al. 2004. A rational function approach for estimating mean annual evapotranspiration. Water Resources Research, 40(2): WR002710, doi: 10.1029/2003WR002710.

Options
文章导航

/

〈 〉