Spatial-Temporal Dynamics of Water Conservation in Gannan in the Upper Yellow River Basin of China

Abstract

Water resources are the material basis for human social development. They play an essential role in human production and life. They are vital for the healthy growth of ecosystem services. The quantitative assessment of the water conservation (WC) function is of great significance to the management of regional ecosystems and the construction of ecological civilization. Gannan is an influential water supply area in the upper Yellow River Basin. Understanding the current status and future changes of its WC function can provide scientific support for the stable use and sustainable development of water resources in the upper Yellow River Basin; its ecological status is highly important. The present study quantitatively assesses the WC function of Gannan based on the InVEST model. It aims to explore the regional ecological compensation mechanism, enhance regional ecological regulation capacity and increase water supply by studying the trend in changes in the WC function and the influence mechanism. The results show that: (1) The overall trend in WC capacity in Gannan increased during the period 2000–2020. (2) The spatial distribution of WC in Gannan varies from year to year, but there is some similarity, showing a high degree in the southwest and a low degree in the northeast. (3) Climatic factors are the dominant factors underlying the spatial and temporal distribution of WC in Gannan. (4) The important areas of WC are located in the southwest and southeast of Gannan.

1. Introduction

Water resources are indispensable for human daily life as well as the social and economic development of society. At present, the consumption of water resources is increasing year by year, the demand is increasing and the shortage of water resources in some parts of the world is very serious, so the rational use and allocation of water resources has become a common global concern [1,2]. The water cycle is an important part of ecosystem functioning. Precipitation is intercepted, stored and made available for use or function in the environmental system for a certain period of time known as water conservation (WC). As an important water ecosystem service, the demand for WC is increasing against the background of high social and economic demand, rapid population growth and urban progress and global climate change. Regional WC relies on ecosystem types, soil types etc., within the ecosystem to interact with atmospheric precipitation and evapotranspiration, thus maintaining the internal hydrological balance. Spatially, WC is the regulation of the water cycle in terrestrial ecosystems at the vertical level, i.e., the interaction of precipitation and evapotranspiration with the vegetation canopy, the understory litter layer and the soil layer, which ultimately recharges the regional water supply in the form of surface runoff, subsurface runoff and loamy midstream flow or is stored within the system. In time, this is mainly in the form of storage of precipitation and flood mitigation, and in the form of groundwater recharge to rivers during low precipitation seasons, which can effectively protect water sources, reduce flooding and maintain healthy regional ecosystems [3].

Adequate understanding of the regional water resource conservation situation is an important prerequisite for WC and ecological security. In the current management of regional WC functions, in order to balance socio-economic needs and environmental protection requirements as far as possible, policy-makers, experts and scholars tend to divide important ecological barriers and WC areas into large regions, and carry out WC function zoning for the corresponding areas, using these sub-level areas chosen to undertake major ecological functions as the main focus of conservation, and focus on treatment and development. The main means of improving WC capacity is the search for vegetation types and their allocation patterns that are suitable for the area and increase effective water yield (WY) [4]. In addition, to carry out simulations of regional ecological WC functions in different contexts, researchers can study the relationship between key factors such as tree species composition, hierarchical structure, configuration, reasonable density and the water transfer function of WC forests under different site conditions that can help predict the future trends of regional ecological functions, which are a strong guarantee for the sustainable development of the region.

With recent advances in theoretical methods and hydrological models, mathematical methods and hydrological models have been introduced for quantitative assessment. Currently, commonly used models include the Algorithms for Recovery and Isolation Exploiting Semantics (ARIES) [5], the Soil and Water Assessment Tool (SWAT) [6,7] and the Integrated Valuation of Ecosystem Services and Trade-off (InVEST) [8,9], among others. Compared to other models, the InVEST model has better data sources and accuracy of results [10,11,12]. The InVEST model, the Integrated Valuation of Ecosystem Services Functions and Trade-offs model, is a comprehensive valuation of ecosystem services. The model was jointly developed by Stanford University, The Nature Conservancy (TNC) and the World Wide Fund for Nature (WWF). The InVEST model has been successfully applied in many countries and regions in Africa, Latin America, the USA, China and Indonesia. For example, the InVEST model was used to measure WY in the Wildcat Creek watershed in Indiana and the Upatoi Creek watershed in Georgia, and was compared with the SWAT model to analyze hydrological accuracy [13]. The applicability of the model to the North Carolina sub-basin in the USA was explored and the factors influencing it were analyzed [14]. The InVEST model was specifically applied to study water retention, carbon storage, habitat quality and biodiversity in a UK sub-catchment [15]. Applying the WaterYield module to the WY of Korean forests, another study showed that the water content of Korean forests is much lower than that of Korea, and analyzed the intra-annual and spatial variability of water content [16]. The InVEST model was also applied to an assessment of the spatial and temporal variability of the Yellow River basin WY as well as its reaction to precipitation and land use change, and it was found that precipitation has a strong influence on WY, while land use/cover change has a smaller influence on WY [17]. This not only facilitates scholarly research, but also provides a scientific basis for policy guidance for decision-makers.

Gannan is located in the upper Yellow River Basin and is blessed with rich natural resources with diverse ecosystems. Accurate evaluation of the region’s WC function is crucial to the rational use of the region’s water resources. It also provides a reference for regional land planning, determining zoning and protecting and restoring the water environment [18]. Over the long term, the special environmental and ecological problems caused by the particular geographical conditions and traditional production patterns in southern Gannan have led to a series of environmental and ecological problems, such as soil erosion, grassland degradation and geological disasters. These have seriously affected and restricted the ecological environment and sustainable socio-economic development of Gannan, and ecological restoration and regional economic development face challenges [19,20,21]. Recently, there have a been number studies on ecosystem services in the Yellow River Basin of China. However, not many studies have been conducted on the WC function of the upper Yellow River Basin. They are mainly single studies focusing on function and value assessment. Comprehensive studies that combine function assessment and importance zoning, etc., are lacking. In addition, the existing studies on WC functions have a relatively short time scale. A longer time series of analyses is needed to identify the main influences on regional WC functions and to study their response to environmental change.

Gannan is located in the upper Yellow River Basin of China, which has large areas of grasslands, forests and wetlands. It recharges 11.4% of the total runoff of the Yellow River. It has become an important water recharge area for the Yellow River and is an important part of the “Chinese Water Tower” [20,22]. By quantifying its WC function, this study provides a scientific basis for regional environmental and ecological planning and establishment. The specific targets for this study are as follows: (1) To analyze the spatial and temporal patterns of change in WC function in Gannan from 2000 to 2020. (2) To explore the differences in the degree of contribution of different influencing factors to WC function. (3) To carry out a classification of WC importance and spatial delineation, and to identify priority WC and key protection areas.

2. Materials and Methods

2.1. Study Area

Gannan is located on the northeastern edge of the Qinghai-Tibet Plateau and in the upper Yellow River Basin (100°46′–104°44′ E, 33°06′–36°10′ N) (Figure 1). The total area of the region is 4.5 × 10⁴ km². The topography is higher in the northwest and lower in the southeast, sloping from northwest to southeast. The terrain is intricate and complex with mountains and valleys. The geological structure system mainly belongs to the eastern and western tectonics of the Qinling Mountains. Gannan is characterized by a continental seasonal climate. There is more precipitation, and the geographical distribution significantly varies. The average annual precipitation ranges from 400–700 mm. The average annual temperature is between 1–13 °C. The territory has abundant water resources. The main rivers are the Yellow River, the Tao River, the Daxia River and the Bailong River (collectively known as the Three Rivers and One River). The soil types are mainly brown, calcium, meadow and swampy soils. The vegetation types are vertically distributed, mainly alpine scrub, alpine meadows, subalpine coniferous forests and temperate coniferous broad-leaved forests [19].

2.2. Data Sources and Processing Methods

The main data required for this study were average annual rainfall, average annual potential evapotranspiration, soil depth, plant available water content, land use type, watershed boundary, the Zhang coefficient, topographic index, soil saturation hydraulic conductivity and flow coefficient [23,24]. The relevant underlying database sources are presented in

Table 1. Acquisition and preprocessing of the parameters.

Data	Data Sources and Processing Methods
Precipitation	Data obtained from the National Earth System Science Data Center (http://www.geodata.cn, accessed on 11 January 2023).
Potential evapotranspiration	Calculated from the modified Hargreaves equation: $E{T}_0=0.0013\times 0.408\times RA\times \left(Tav+17\right)\times {\left(TD-0.0123P\right)}^{0.76}$, where Tav is the average of the month-by-month mean daily maximum and month-by-month mean minimum temperatures, TD is the difference between the month-by-month mean daily maximum and month-by-month mean minimum temperatures, RA is the extraterrestrial radiation and P is the month-by-month precipitation [25,27].
Land use/cover	Data obtained from the Resource and Environmental Science Data Center (RESDC), Chinese Academy of Sciences (http://www.resdc.cn, accessed on 11 January 2023).
Soil data	Includes soil type, soil texture (% clay, % sand, % chalk, % organic carbon) and soil depth from the World Soil Database (https://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonized-world-soil-database-v12/en/, accessed on 11 January 2023).
Plant available water content (PAWC)	Calculation based on soil texture: $PAWC=54.509-0.132sand\%-0.003{\left(sand\%\right)}^2-0.055silt\%-0.006{\left(clay\%\right)}^2-0.738clay\%+0.007{\left(clay\%\right)}^2-2.6888OM\%+0.501{\left(OM\%\right)}^2$, where OM% indicates the soil organic matter content [28].
Digital elevation model	Geospatial Data Cloud Platform of Chinese Academy of Sciences (http://www.gscloud.cn, accessed on 11 January 2023).
Topographic index	Calculations were based on the number of watersheds rasters, soil depths and percentage slopes of the study area, combined with the use of ArcGIS spatial analysis tools.
Velocity coefficient	According to relevant studies and the InVEST model handbook [25,29].
Soil saturated hydraulic conductivity	Calculations were performed by SPAW software based on soil texture data.
Percentage slope	Based on the DEM, the calculation was performed using the slope tool in ArcGIS.
Watershed division map	Based on DEM, processed by ArcGIS hydrological analysis tool.
The parameter Z	Obtained by repeated verification based on relevant information from Gannan. Z = 4.1.

. All data were resampled to a 100 × 100 m resolution and were projected with the WGS_1984_UTM_Zone_48N coordinate system. Biophysical parameters for individual land use types used for this study are presented in

Table 2. Table of biophysical parameters.

Lucode	LULC_Desc	LULC_Veg	Kc	Root_Depth (mm)
1	Farmland	1	0.65	2000
2	Woodland	1	1.00	5200
3	Grassland	1	0.65	2300
4	Water	0	1.00	100
5	Residential area	0	0.30	100
6	Unused land	0	1.00	300

Note: Lucode is a unique integer for each LULC class. LULC_desc is the descriptive name of the land use/land cover class. LULC_veg is a value that must be 1 for vegetated land use except for wetlands and 0 for all other land uses. Kc is the plant evapotranspiration coefficient for each LULC class. It is calculated according to vegetation type by means of the leaf area index [30], referring to the InVEST manual and the results of existing studies [31]. Root_depth is the maximum root depth for the vegetated land use class. Taken from the China soil datasets checklist and related studies [31] and the InVEST model manual.

, which provides data on the name of each land use type, the presence or absence of vegetation, the plant evapotranspiration coefficient (Kc) and the root depth for each land use type [25,26].

2.3. Methodology for Evaluating the Water Conservation Function

2.3.1. InVEST Model—Water Conservation Calculation

The WY module of the InVEST model is an estimation method based on water balance. The amount of WY is obtained by subtracting the actual evapotranspiration from the amount of rainfall in the raster cell. The WY is corrected for topography, soil saturation hydraulic conductivity and other factors to obtain the WY of the area [12]. The InVEST model combines topographic factors and differences in soil permeability to assess the WC capacity, making the results more scientific. The model is flexible in terms of parameter adjustment and spatial expression of the assessment results [32]. The specific formula is as follows.

$$Y\left(x\right)=\left(1-\frac{AET\left(x\right)}{P\left(x\right)}\right)\times P\left(x\right)$$

(1)

where Y(x) is the WY of grid x in mm. AET(x) is the actual evapotranspiration of grid x in mm. P(x) is the precipitation of grid x in mm.

$$\frac{AET\left(x\right)}{P\left(x\right)}=\frac{1+\omega \left(x\right)R\left(x\right)}{1+\omega \left(x\right)R\left(x\right)+1/R\left(x\right)}$$

(2)

where ω(x) is a non-physical parameter that characterizes the watershed. R(x) is the ratio of potential evapotranspiration to rainfall, also known as the drought factor. The equations for each parameter in the equation are as follows.

$$R\left(x\right)=\frac{k\times E{T}_0}{P\left(x\right)}$$

(3)

$$\omega \left(x\right)=Z\frac{AWC\left(x\right)}{P\left(x\right)}$$

(4)

$$AWC\left(x\right)=Min\left(Rest.layer.depth, root.depth\right)\times PAWC$$

(5)

where ET₀ is the potential evapotranspiration, or reference evapotranspiration (mm). AWC(x) is the effective water content of vegetation (mm), $Rest.layer.depth$ is the maximum soil depth (mm), $root.depth$ is the maximum root restriction depth (mm) and PAWC is the plant available water content (mm). Once the WY has been obtained, the WC capacity can be obtained. The model for calculating WC capacity is as follows.

$$Retention=min\left(1,\frac{249}{Velocity}\right)\times min\left(1,\frac{0.9\times TI}{3}\right)\times min\left(1,\frac{Ksat}{300}\right)\times Yield$$

(6)

$$TI=lg\left(\frac{Drainage\_area}{Soil\_depth\times Percent\_slope}\right)$$

(7)

where Retention is the water content (mm). Velocity is the flow rate coefficient. TI is the topographic index. Ksat is the soil saturation hydraulic conductivity (mm/d). Drainage_area is the number of grids in the catchment. Soil_depth is the soil depth (mm). Percent_slope is the percentage slope.

The WC factor is the ratio of WC to precipitation. It indicates the effect of precipitation, vegetation type and different surface features on water availability. The higher the coefficient, the better the area’s ability to support water [33]. It is calculated as follows.

$$WCC=\frac{Retention}{P}$$

(8)

where WCC is the WC coefficient. Retention is the WC capacity (mm). P is the precipitation (mm).

2.3.2. Change Trend Analysis

The one-dimensional linear regression method was used to assess WC trends in Gannan from 2000 to 2020. For the raster data, each image element has a corresponding 20-year time series value, and the trend of individual image element reflects the spatial evolution characteristics [34]. It is calculated as follows.

$$Slope=\frac{n{\sum }_{i=1}^{n}i\times Wi-{\sum }_{i=1}^{n}i{\sum }_{i=1}^{n}Wi}{n{\sum }_{i=1}^n{i}^2-{\left({\sum }_{i=1}^{n}i\right)}^2}$$

(9)

where Slope is the linear trend rate of change. Values greater than 0 indicate an improvement of the WC function over the time series under study. Values less than 0 indicate a decline. n is the number of numerical years. i is the year series.

2.3.3. Spatial Delineation of Water Conservation Importance

The classification method of the importance for WC in Gannan was based on “Technical Guidelines for the Delineation of Ecological Protection Red Line” [35,36,37]. In addition, the classification of the importance of WC in Gannan was carried out by combining the quantile classification method in ArcGIS software [28,38]. Five importance grades were defined, such as areas of general importance (Grade I; <151 mm), areas of mild importance (Grade II; 151–232 mm), areas of moderate importance (Grade III; 232–293 mm), areas of high importance (Grade IV; 293–368 mm) and areas of extreme importance (Grade V; >368 mm).

3. Results

3.1. Inter-Annual Trends in Water Conservation

The precipitation, potential evapotranspiration, actual evapotranspiration, WY and WC of Gannan from 2000 to 2020 are shown in Figure 2. The mean annual precipitation in the study area during the study period was 791.44 mm, with an overall interannual trend of a slight decrease in precipitation at a rate of 1.79 mm/a. The mean potential evapotranspiration was 774.78 mm, with an increasing trend over the last 20 years at a rate of 3.29 mm/a. The lowest value of potential evapotranspiration occurred in 2003 (673.20 mm) and the highest value occurred in 2018 (925.86 mm). The mean annual actual evapotranspiration in the basin was 555.22 mm, with relatively stable inter-annual variability and an increasing trend of 0.83 mm/a. Thus, more than 70.15% of the annual precipitation in the basin was returned to the atmosphere through actual evapotranspiration.

From 2000 to 2020, the interannual variation in the average annual WC in Gannan and the regions is shown in Figure 3. During this period, the annual average WC in Gannan was about 90.04 × 10⁸ m³, with an overall increasing trend (13.70 × 10⁸ m³/10a). The maximum value was 110.77 × 10⁸ m³ in 2020 and the minimum value was 73.96 × 10⁸ m³ in 2015. The annual mean WC considerably varied between regions, with the highest in the southwest (23.37 × 10⁸ m³) and the lowest in the northeast (2.75 × 10⁸ m³). From 2000 to 2020, there was an overall upward trend in WC in all regions of Gannan. Southwest, northwest and east had more obvious upward trends (3.065 × 10⁸ m³/10a, 2.325 × 10⁸ m³/10a and 2.219 × 10⁸ m³/10a, respectively).

The changes in the average annual WC coefficient for Gannan and the regions from 2000 to 2020 are illustrated in Figure 4. During this period, the average annual WC coefficient for Gannan was 0.462, with an interannual trend similar to the amount of WC, with an overall upward trend (1.48 × 10⁻²/10a). The maximum value was 0.494 in 2020 and the minimum value was 0.425 in 2010. The average annual WC coefficient considerably varied from region to region. South (0.518) > southeast (0.490) > east (0.480) > northwest (0.473) > west (0.454) > north (0.427) > southwest (0.424) > northeast (0.411). In terms of interannual trends, seven regions showed an increasing trend and one region showed a decreasing trend (Figure 4). Among them, the WC coefficients of the southwest and west showed a more obvious upward trend (0.83 × 10⁻²/10a and 1.66 × 10⁻²/10a, respectively), indicating that the WC function in these two regions gradually increased. In contrast, the WC coefficient for the southeast showed a decreasing trend (−6.45 × 10⁻²/10a), indicating a gradual weakening of the WC function in this region.

3.2. Spatial Distribution of Water Yield and Water Conservation

3.2.1. Spatial Distribution Characteristics of Water Yield

A spatial distribution of WY in Gannan from 2000 to 2020 is shown in Figure 5. During this period, the spatial distribution of WY did not change much, showing a trend of high in the southwest and low in the northeast. However, the average annual WY considerably varied from region to region. The area with the highest average annual WY was the southwest (498.12 mm) and the area with the lowest average annual WY was the southeast (352.48 mm). Areas of higher WY were mainly in the southwest (498.12 mm) and west (438.25 mm). Areas of medium WY were generally in the northwest (399.13 mm), north (388.01 mm), south (387.13 mm) and east (385.25 mm). Areas of low WY were located in the northeast (354.60 mm) and southeast (352.48 mm).

3.2.2. Spatial Distribution Pattern of Water Conservation

An indication of the spatial distribution of the average annual WC in Gannan from 2000 to 2020 is shown in Figure 6. During this period, the spatial distribution did not change much, similar to that of WY. There was a downward trend from southwest to northeast. The high-value areas of WC were mainly in the south (287.14 mm) and southeast (271.94 mm). The middle-value areas were mainly in the southwest (255.75 mm), west (252.94 mm) and east (252.43 mm). Low-value areas were mainly in the northwest (243.33 mm), north (216.64 mm) and northeast (204.85 mm).

3.3. Responses of Water Conservation to Key Factors

3.3.1. Responses of Water Conservation to Climatic Factors

The influence of climatic elements on the regional WC is mainly reflected in precipitation and evapotranspiration. The highest levels of precipitation (631.51 mm), WY (512.44 mm) and WC (312.27 mm) were recorded in 2020, as shown in Figure 6. The year with the lowest precipitation was 2015 (472.26 mm), which was also the year with the lowest WY (346.92 mm) and WC (208.54 mm), indicating that precipitation is an important factor affecting WC. This is consistent with the correlation analysis results (Figure 7). The correlation coefficients between the WC and each influence factor are P (0.77) > Ksat (0.14) > V (0.12) > ASL (1.4 × 10⁻¹⁶) > TI (1.3 × 10⁻¹⁶) > PAWC (−0.02) > SD (−0.2) > PET (−0.62) > AET (−0.63). The spatial distribution of mean annual precipitation in Gannan is in accordance with that of the WC. It shows a tendency to be high in the southwest and low in the northeast. It also shows a decreasing trend from the upper to the lower reaches.

The actual evapotranspiration in Gannan remained relatively stable during this period, as shown in Figure 2. Spatially, actual evapotranspiration was higher in the southeast and south; i.e., sub-basins 1 and 7 (198.95 mm and 164.33 mm, respectively). The west and north, i.e., sub-basins 2, 4, 5 and 6, were relatively low, with the southwest, i.e., sub-basin 5, in particular having the lowest (103.61 mm) (Figure 8). Actual evapotranspiration showed an increasing trend from upstream to downstream, with a negative correlation with WC.

3.3.2. Responses of Water Conservation to Land Use Patterns

From 2000 to 2020, the land use structure of Gannan did not change much, as shown in Figure 9, with grassland and forest land dominating. Farmland, woodland, grassland, water, residential area and unused land account for 4.49%, 30.79%, 56.74%, 0.49%, 0.33% and 7.16% of Gannan, respectively.

In this study, the average WC and total WC for various land use types in Gannan were calculated for the period 2000–2020. The average annual WC of the six land use types was woodland > grassland > unused land > farmland > residential area > water (

Table 3. Total amount of WC of different land use types in Gannan from 2000 to 2020.

Year	WC Of Each Land Use Type						Total Amount of WC (108 m3)
	Farmland (108 m3)	Woodland (108 m3)	Grassland (108 m3)	Water (106 m3)	Residential Area (106 m3)	Unused Land (108 m3)
2000	1.51 (1.81%)	35.70 (42.82%)	40.95 (49.12%)	0.07 (0.08%)	0.04 (0.05%)	5.11 (6.13%)	83.37
2005	1.77 (1.79%)	42.75 (43.13%)	48.59 (49.02%)	0.08 (0.08%)	0.06 (0.06%)	5.87 (5.92%)	99.12
2010	1.48 (1.78%)	34.39 (41.44%)	42.12 (50.76%)	0.07 (0.08%)	0.05 (0.06%)	4.87 (5.87%)	82.98
2015	1.28 (1.73%)	30.85 (41.71%)	37.25 (50.37%)	0.06 (0.08%)	0.05 (0.07%)	4.47 (6.04%)	73.96
2020	1.91 (1.72%)	48.85 (44.10%)	53.40 (48.21%)	0.11 (0.10%)	0.10 (0.09%)	6.38 (5.76%)	110.76

). The highest WC of the six land use types was found in 2020, with woodland (443.04 mm) > grassland (266.34 mm) > unused land (258.55 mm) > farmland (123.10 mm) > residential area (64.66 mm) > water (53.69 mm). The lowest WC was found in 2015 for woodland (282.95 mm) > grassland (184.84 mm) > unused land (175.32 mm) > farmland (81.63 mm) > residential area (45.58 mm) > water (35.07 mm). The total and average WC results for different land use types were similar, grassland > woodland > unused land > farmland > water > residential area (Table 3). Of these, grassland and woodland accounted for over 91% of the total WC in Gannan.

3.4. Spatial Classification of the Importance of Water Conservation

The spatial distribution of the importance of WC in Gannan is shown in Figure 10, with the western, northwestern, northern and northeastern parts being mainly Grade I and Grade II. The southwest is mainly Grade III and Grade IV. The eastern, southern and southeastern parts are mainly Grade IV and Grade V. Of these, Grade I is of general importance, Grade II is of relative importance, Grade III is of moderate importance, Grade IV is of high importance and Grade V is of extreme importance. Grade I, II and III areas are of relatively low importance for WC. These areas are mainly in the western and northern parts of Gannan. Areas with a WC capacity greater than 368 mm are considered to be of higher WC importance and require priority protection. These areas are mainly in the southeastern part of Gannan. Meanwhile, southwestern Gannan lies at the headwaters of the upper Yellow River Basin. It is the main flow area of the Yellow River. It covers a large area. The importance of WC is mainly tertiary and quaternary. It should also be protected as a priority (Figure 10).

4. Discussion

4.1. Spatial and Temporal Patterns of Water Yield and Water Conservation in Gannan

The average annual WY of Gannan gradually decreased from southwest to northeast from 2000 to 2020 (Figure 4). The area with higher WY is located in the southwest. This region is located at the source of the Yellow River, with abundant precipitation, relatively large grassland and wetland areas, and surface runoff easily formed after rain, leading to higher WY. Medium WY areas are located in the west, northwest, north, east and northeast. In these areas, forest vegetation cover is high. The forest canopy, dead branches and soil layers effectively intercept precipitation, thus reducing water production. The low WY area is located in the southeast. Low precipitation and frequent drought disasters in this region contribute to low WY [39]. In this study, it was found that as the vegetation cover in Gannan gradually increased from west to east, the interception of surface runoff also increased, which led to a gradual decrease in WY.

The spatial distribution of WC in Gannan is basically consistent with water production (Figure 5). Most of the areas with high WC are in the south, where the terrain is relatively flat. Since these areas are gently sloping, the water flow velocity decreases when it rains, which is conducive to soil loosening and water infiltration, and therefore, the WC is higher. The areas with low WC values are mainly located in the north and east. Due to the high mountains and steep slopes in these areas, the soil cover is low. In addition, the local farmland is large and the surface is exposed for a long period of time throughout the year, thus it is not conducive to WC. With the change in topography, surface runoff increases from north to south, soil absorption time is prolonged and its WC gradually increases [40,41].

4.2. Driving Factors Affecting the Regional Water Conservation Function

Regional WC function can be influenced by multiple factors [11,42]. Previous studies have shown that WC is a complex ecological process driven by the coordination of multiple factors. This is consistent with the results of the present study [43,44,45]. Considering the indicators and processes used in this study to calculate WC, precipitation, evapotranspiration, land use type and soil properties were considered as important factors influencing WC. This study further analyzed the factors affecting the spatial and temporal distribution patterns of WC in Gannan. According to the change characteristics of precipitation and WC in Gannan from 2000 to 2020 (Figure 2), the response of WC to precipitation in sub-basins of the study area (Figure 8), as well as the relationship between WC and other influencing factors (Figure 7), this study proves that precipitation is an important driving force of WC. These results are consistent with many other studies. For example, a global assessment of the sensitivity of runoff to changes in precipitation, potential evapotranspiration and other factors is based on the Budyko method. The study found that surface water resources are most sensitive to changes in precipitation on a global scale [43]. As WFcrop’s WFblue was higher than the world average, it was reduced by changing crop patterns or synchronizing crop planting and harvesting dates to periods favorable to precipitation [46]. A bottom-up approach was used to analyze the blue and green water footprints of the Upper Tigris River Basin (UTRB, Turkey) from 2010 to 2018. The results of the study found that reducing evaporation rates, capturing more rainwater and providing more advanced irrigation methods are the main preventive measures to reduce the water footprint of any basin [47]. Based on a comparative study of different regions, climatic factors, i.e., precipitation and evaporation, were found to be the most critical drivers of change in WC [32]. Precipitation in Gannan decreases from southwest to northeast, with the average annual precipitation in most areas being around 549 mm [48]. The overall precipitation in the Yellow River Basin has increased since the 21st century. The tendency to reduce runoff from the area has slowed. Evapotranspiration has decreased in 56.5% of the area [49,50,51]. With the significant increase in precipitation and the significant decrease in evapotranspiration, the WC is bound to increase. The correlation coefficient between WC and meteorological elements in the Gannan region is very significant, as shown in Figure 7. On the other hand, the correlation coefficients with other elements are not as strong or weak. This indicates that the change in WC is a complex combination of various elements. Of these, precipitation has a positive effect on WC, presenting a positive correlation to WC (0.77). Actual and potential evapotranspiration had a negative effect on WC, showing a negative correlation with WC (−0.63 and −0.62, respectively).

According to the water balance principle, WC can be seen to be the discrepancy between precipitation and evaporation. The close relationship between potential and actual evapotranspiration affects terrestrial water ecosystems [52,53,54,55]. The present study obtained a negative relationship between the two and WC. The results of this study found that different land use types tend to have different soil texture characteristics, which leads to differences in WC [56,57]. Water is transported through connected pores in the soil and the hydraulic properties of the soil are essential for the transport of water and solutes. The saturated hydraulic conductivity of soils increases with improved vegetation and community succession, and improved forest structure further enhances WC capacity [58,59]. The driving effect of vegetation on WC is complex and comprehensive, and the underlying reason is that the effect of vegetation on WC is a combination of soil water depletion and precipitation retention. The depletion of soil water by vegetation is mainly through the uptake of soil water by the root system and the transfer of water to the atmosphere through transpiration. Therefore, for land use types such as woodlands, the maximum root restriction depth and transpiration coefficient are usually high, thus weakening the WC function. The retention of precipitation by vegetation is mainly reflected in the retention of precipitation by the vegetation canopy and the retention of water by the apoplastic and soil layers, which is reflected in the flow rate coefficients of the WC calculation model. Apart from making use of these major factors, topographic factors were also considered in this study. Topography influences climate by impacting regional hydrothermal conditions. In turn, hydrothermal conditions affect the composition of the forest vegetation, the aggregation of bottomland plankton and the physical and chemical characteristics of the soil. Ultimately, this affects the WC capacity of the region [60,61].

4.3. Ecological Priority and Regional Management

The ecological environment of the upper Yellow River Basin is at a stage of partial degradation and overall improvement. This study assessed the WC of Gannan in the upper Yellow River basin from 2000–2020, with important implications for sustainable ecological conservation and water resource management in the region. The WC function was evaluated according to land type, ecological function and environmental value, and key protection areas in Gannan were identified. The potential areas could be highlighted in planning decisions and provide a reference for subsequent priority protection. The southwestern part is rich in rainfall, with little impact from human activities and with a large area covered by grasslands and wetlands. By optimizing the vegetation type and structure, the WC capacity of the area can be improved. The northeastern part is mostly arable land and construction land with flat terrain, which is more influenced by human activities. In view of the natural and socio-economic factors in Gannan, afforestation and optimization of vegetation in the area can improve the WC capacity of the area. As a central water source in the upper Yellow River Basin, Gannan has a very significant role to play in mitigating the increasingly severe water resource problems in the upper Yellow River Basin. Its WC function is of strategic importance to the strategic development and water resources management schemes of the region [62]. In this study, it was found that Gannan’s WC has generally improved in the past 20 years, indicating that Gannan’s WC function is being optimized.

By analyzing the spatial and temporal characteristics of WC in Gannan and its influencing factors, this study proposes the following countermeasures for promoting natural environmental management and economic and social development in the upper Yellow River Basin. (1) Build grass squares, create shrub belts, fence and seal, replant grass species and zone rotational grazing to improve grassland ecological benefits. (2) Strengthen the restoration and management projects of forest ecosystems, consolidate the achievements of returning farmland to forests and increase the area of vegetation cover. (3) Protect and restore wetland ecosystems, stabilize water recharge in the Yellow River and prevent downstream disruptions and flood storage. (4) Adhere to the ecological red line, implement key natural forest management and conservation projects in the ecological red line area and carry out the pilot construction of a comprehensive biodiversity observatory and grassland greening observation area.

5. Conclusions

Gannan is one of the important areas in the upper Yellow River Basin and is of strategic importance in terms of ecological environmental protection and water resources management in China. In this study, we quantitatively evaluated the WC in Gannan from 2000 to 2020, analyzed the spatial and temporal variation of WC, discussed its response to various factors and spatially divided the important areas. The results show that:

Changes in WC function are the result of a combination of climate, soil and land use. The correlation coefficients between the WC and each influence factor are P (0.77) > Ksat (0.14) > V (0.12) > ASL (1.4 × 10⁻¹⁶) > TI (1.3 × 10⁻¹⁶) > PAWC (−0.02) > SD (−0.2) > PET (−0.62) > AET (−0.63). The southwest and southeast are extremely important WC areas in Gannan (mainly Grade IV and V). We need to give adequate attention and priority to conservation. In the last 20 years, the amount of WC in Gannan has generally shown an increasing trend, and effective measures should continue to be taken to improve the quality of the ecological environment and enhance water resources protection. Therefore, this study recommends the implementation of an artificial climate project based on ecological restoration in the Gannan region of the upper Yellow River Basin to effectively monitor climate change dynamics. At the same time, a national nature reserve should be established in the southwestern part of Gannan to expand the ecological protection area centered in the southwestern part.