Ecological–Economic Assessment and Managerial Signiﬁcance of Water Conservation in the Headwaters of the Yellow River

: The water conservation function plays a vital role in the land–water cycle. As the “Chinese water tower”, the headwaters of the Yellow River are of great signiﬁcance to the safety of the Yellow River basin and even the global ecosystem. Taking the grassland ecosystem in the Yellow River source area as the research object, the InVEST water yield model with modiﬁed parameters and the ecological value evaluation of the modiﬁed equivalent factor method were used to explore the simulated spatio-temporal changes and the value of grassland water conservation from 2001 to 2020. The results show that: (1) the average total amount of water conservation in the source area is 549 × 10 8 m 3 , which is 16% of the runoff in the Yellow River basin, with a growth rate of 7.5 mm/year 1 and a contribution rate of 30%; (2) the total ecological value of grassland water conservation in 2020 is USD 340.03 × 10 8 . The proportion of improved grassland in ecological restoration and management is only 0.51%, while the proportion of original alpine meadow reaches 67% and its ecological function and value are irreplaceable; (3) based on the comprehensive indicators of water conservation capacity, value and importance, Qumalai, Chengduo and Maduo counties are ranked as priority areas for the ecological protection of water resources.


Introduction
Over the past decades, a substantial contradiction concerning the development of human society and the protection of the ecological environment has become increasingly exacerbated and 60% of global ecosystems have been degraded to varying degrees [1]. The headwater of the Yellow River (HYR) is a major ecological and functional area for water conservation in China. The safety of the ecosystem is related to the ecological, environmental and economic development of the Yellow River basin and even East Asia, focusing the interest of scientists from all walks of life [2]. However, the vegetation type of the ecosystem in the headwater is mainly grassland, which is complex, diverse and highly fragile. Due to the stress of global climate change and the influence of anthropological factors, the ecological environment has been damaged and the grasslands' degradation is serious; this leads to a reduction in the water conservation capacity, deterioration of the soil and production loss, all of which impact the lifestyle of local herdsmen in the region [3][4][5] For the regional rotational grazing and grass stand management, a comprehensive grazing ban was implemented on moderately degraded grassland in the HYR, under the Qinghai Sanjiangyuan Nature Reserve Construction Phase I and Phase II. The project achieved remarkable results and the ecological situation has improved significantly [6][7][8].
Using the alpine meadow in the HYR, three key problems will be solved using the modified parameter investment water yield model and the modified equivalent factor method: (1) the spatio-temporal pattern of water yield distribution; (2) the evaluation of the ecological function value of water conservation; (3) the comprehensive evaluation of the water conservation function based on grassland type, ecological function and ecological value, and the determination of an ecological priority conservation area to theoretically support the ecological conservation and management of the headwater area.

Study Area
The HYR (32 • 30 N-35 • 0 N-95 • 50 E-103 • 30 E) is the largest runoff area and accounts for 16.20% of the total Yellow River basin, with a drainage area of 12.37 × 10 4 km 2 [46]. The study area is between 2680 and 6248 m above sea level (a.s.l.) and increases gradually with an average elevation of 4123 m (a.s.l.) from east to west. The area under study characterizes a typical continental highland climate and the annual average temperature ranges from −12.7 • C to 5.6 • C. Under the influence of the southwest monsoon, the annual average precipitation ranges from 281.8 mm to 1058.6 mm. However, the summer (June to September) generally accounts for 90% of the total annual precipitation. The average annual evaporation ranges from 730 mm to 1700 mm. In the HYR, there are various types of grassland, of which the alpine meadow accounts for 51.55% of the total area of the source area, while improved grassland accounts for only 0.01%. The data of vegetation cover types involved in this study generated the spatial distribution map of vegetation types in the study area based on the IGBP and the Chinese grassland type survey map (CERES land classification from the International Geosphere Biosphere Program IGBP) ( Figure 1). ever, for the regional averaging of the national equivalence factor in the HYR, it is necessary to take into account the vegetation cover in the study area, to improve the equivalence factor value based on the NPP data and to develop the reliability of the equivalence factor method at the regional scale [45].
Using the alpine meadow in the HYR, three key problems will be solved using the modified parameter investment water yield model and the modified equivalent factor method: (1) the spatio-temporal pattern of water yield distribution; (2) the evaluation of the ecological function value of water conservation; (3) the comprehensive evaluation of the water conservation function based on grassland type, ecological function and ecological value, and the determination of an ecological priority conservation area to theoretically support the ecological conservation and management of the headwater area.

Study Area
The HYR (32°30′ N-35°0′ N-95°50′ E-103°30′ E) is the largest runoff area and accounts for 16.20% of the total Yellow River basin, with a drainage area of 12.37 × 10 4 km 2 [46]. The study area is between 2680 and 6248 m above sea level (a.s.l.) and increases gradually with an average elevation of 4123 m (a.s.l.) from east to west. The area under study characterizes a typical continental highland climate and the annual average temperature ranges from −12.7 °C to 5.6 °C. Under the influence of the southwest monsoon, the annual average precipitation ranges from 281.8 mm to 1058.6 mm. However, the summer (June to September) generally accounts for 90% of the total annual precipitation. The average annual evaporation ranges from 730 mm to 1700 mm. In the HYR, there are various types of grassland, of which the alpine meadow accounts for 51.55% of the total area of the source area, while improved grassland accounts for only 0.01%. The data of vegetation cover types involved in this study generated the spatial distribution map of vegetation types in the study area based on the IGBP and the Chinese grassland type survey map (CERES land classification from the International Geosphere Biosphere Program IGBP) ( Figure 1).

Data Source and Preprocessing
Environmental variables and socioeconomic data are needed to evaluate water yield and ecological value for the time period 2001-2020 (Table 1). The environmental variables' data are divided into four categories: soil physical and chemical factors (S, 9), terrain factors (T, 5), meteorological factors (A, 4) and vegetation factors (B, 2). The raster images of all Using the water yield module of the InVEST model, the total annual water yield (mm) was calculated by formula (Equation (1)), where AET (x) is the annual actual evapotranspiration of grid unit X, and P (x) is the annual precipitation of grid unit X. The formula is as follows. AET (x)/P (x) is the approximate Budyko curve (Equation (2)) in terms of annual precipitation and actual evapotranspiration [47]. ω (x) is a nonphysical empirical adjustment parameter that describes the natural climate and soil properties of the watershed (Equation (3)) [48]. The Zhang coefficient is a climatic seasonal factor that represents local hydrogeological properties and is an empirical constant. The standard value ranges from 1 to 30 [18,19]. PAWC effective water retention capacity (Equation (4)), where sand, silt, clay and soil organic matter (om) are described in percent (%).
AET(x) P(X) Calculation formula for water retention (mm) (Equation (5)), topographic index (TI), soil saturated hydraulic conductivity (Ksat) and velocity coefficient of total annual water storage [49]. The calculation formula is as follows: Water Retention Spatio-Temporal Dynamics and Trend Analysis The InVEST model is based on ArcGIS and MATLAB 2017 software for analyzing the annual water retention and the spatio-temporal dynamic pattern of the HYR from 2001 to 2020 and for conducting geographic mapping. The trend analysis uses Theil Sen (SEN) median trend analysis and nonparametric (Mann-Kendall, MK) to test each pixel to determine water retention trend and mutation point detection. When the Sen slope > 0, it shows an upward trend, whereas in the opposite situation it shows a downward trend (SEN slope < 0). With a confidence level α = test the significance of the change trend of water retention at 0.05 and confidence level α = 0.05 to test the significance of the change trend of the water yield. If z > 1.96, the results increase significantly; if z < −1.96, the results decrease significantly; if −1.96 ≤ z ≤ 1.96, the results do not change significantly ( Table 2).
The Mann-Kendall method (MK) estimates the point of annual change in the simulation value of the annual water production from 2000 to 2020. UF k is a standard normal distribution (Equation (6)), calculating the statistical sequence by x 1 , x 2 , . . . , x n ; UB k = −UF k generates the inverse order x n , x n−1 , . . . , x 1 (Equation (7)).
Confidence level α = 0.05 to check the change trend of time series of water retention. When UF or UB > 0, it indicates that the series shows an upward trend, and when UF or UB < 0, it shows a downward trend. The time point corresponding to the intersection of UF and UB curves is the time point at which mutation starts.

Ecological Value Evaluation of Regional Water Yield
Based on the Chinese terrestrial ecological service value equivalent table [30], the ecological service value equivalent factor in the HYR is calculated using the national average grain and unit price data from 2001 to 2020, according to the principle that the economic value of a standard ecosystem service equivalence factor is 1/7 of the economic value of food production per unit area of farmland [35,42,50]. Based on the NPP data of grassland in the source area of the Yellow River, an ESV equivalent factor of USD 14,461.68/(hm 2 ·year −1 ) was calculated for the study area and the total value of aquatic ecological services of grassland in the HYR was calculated in combination with the distribution area of grassland species in the headwaters.

Division of Ecological Priority Protection Areas
For the delineation method of the water conservation area, we followed the guidelines for the delineation of the ecological red line of protection [51][52][53][54]. Since the delineation of ecological red line certainty, only the water protection amount and importance level in 2020 were extracted and superimposed to delineate the ecological red line and the protection area.
According to the evaluation and classification steps of the guidelines, the grid values corresponding to 30%, 50%, 80% and 90% of the total water conservation capacity are classified as evaluation and classification nodes of water conservation capacity. The analytical hierarchy process (AHP) is used to build the hierarchical structure of importance rating. At the same time, the significance of the data is standardized using the correlation test between the variable environmental factors and the water production. The AHP is used to build the evaluation matrix. After the consistency test, the weighting of the influencing factors is determined. Combined with the comprehensive index algorithm, the water retention importance index (WRI) is obtained. The value range of the WRI ranges from 0-1 and the closer it is to 1, the higher the importance. Finally, the overlay analysis in ArcGIS is used and the natural segment points method is employed for classification.

Spatio-Temporal Variation Characteristics
The annual average water retention from 2001 to 2020 in the study area was 443.86 mm, and the overall interannual variation trend of water conservation showed a slight increase, with a rate of 7.5 mm/year −1 . The total annual average water retention was 5.49 × 10 10 m 3 .
The estimated water retention was mainly between 300 and 700 mm, which accounted for 93.65% of HYR, and the areas with less than 300 mm and more than 700 mm accounted for 3.21% and 3.15%, respectively. The high value areas were located in Maduo, Toronto and Qumai counties, with heavy rainfall and narrow alpine grassland areas. After rainfall, it is easy to form surface runoff and the regional water retention capacity is high. The median water conservation was observed in Ruoergai, Aba and Hongyuan counties, with water retention of more than 600 mm/year −1 , whereas Maqu and Jiuzhi counties also showed high water retention, at 540 mm and 587.79 mm/year −1 , respectively. The median area has a high forest cover and the forest canopy, litter and surface soil effectively intercept the rainfall, thus reducing the water conservation capacity. The least water retention of Qumalai county was 229.69 mm/year -1 , with less rainfall, flat terrain and dense forest vegetation coverage intercepting rainfall and surface runoff, resulting in less water retention ( Figure 2). increase, with a rate of 7.5 mm/year −1 . The total annual average water retention was 5.49 × 10 10 m 3 . The estimated water retention was mainly between 300 and 700 mm, which accounted for 93.65% of HYR, and the areas with less than 300 mm and more than 700 mm accounted for 3.21% and 3.15%, respectively. The high value areas were located in Maduo, Toronto and Qumai counties, with heavy rainfall and narrow alpine grassland areas. After rainfall, it is easy to form surface runoff and the regional water retention capacity is high. The median water conservation was observed in Ruoergai, Aba and Hongyuan counties, with water retention of more than 600 mm/year −1 , whereas Maqu and Jiuzhi counties also showed high water retention, at 540 mm and 587.79 mm/year −1 , respectively. The median area has a high forest cover and the forest canopy, litter and surface soil effectively intercept the rainfall, thus reducing the water conservation capacity. The least water retention of Qumalai county was 229.69 mm/year -1 , with less rainfall, flat terrain and dense forest vegetation coverage intercepting rainfall and surface runoff, resulting in less water retention ( Figure 2).

Trend Change Characteristics of Water Yield
The water conservation capacity of most regions in the HYR showed an upward trend from 2001 to 2020 (Figure 3). The area with significant growth accounted for 9.44%, mainly in the southeast of Zeku and Henan counties, the middle of Maqin and the north of Qumalai, with 90.56% of regional water conservation, which showed a slight upward trend.

Trend Change Characteristics of Water Yield
The water conservation capacity of most regions in the HYR showed an upward trend from 2001 to 2020 ( Figure 3). The area with significant growth accounted for 9.44%, mainly in the southeast of Zeku and Henan counties, the middle of Maqin and the north of Qumalai, with 90.56% of regional water conservation, which showed a slight upward trend.  The water retention in the source area in the last 20 years is statistically described in four time periods (Table 3) and the water production in each period shows an increasing trend. Among them, the annual water production from 2015 to 2020 is the largest. It should be noted that the water production of Chengduo, Gande, Maqu, Dari, Ruoergai, Jiuzhi, Aba and Hongyuan counties increased consistently from 2001 to 2020, while that of the other counties increased from 2001 to 2010, decreased from 2011 to 2015 and increased from 2016 to 2020. Analysis of the water retention data by the MK test shows that 2005 is a mutation year. Therefore, we take 2005 as the cut-off point to analyze the change in water retention ( Figure 3). The annual water retention increased from 347.14 mm in 2001 to 510.17 mm in The water retention in the source area in the last 20 years is statistically described in four time periods (Table 3) and the water production in each period shows an increasing trend. Among them, the annual water production from 2015 to 2020 is the largest. It should be noted that the water production of Chengduo, Gande, Maqu, Dari, Ruoergai, Jiuzhi, Aba and Hongyuan counties increased consistently from 2001 to 2020, while that of the other counties increased from 2001 to 2010, decreased from 2011 to 2015 and increased from 2016 to 2020. Analysis of the water retention data by the MK test shows that 2005 is a mutation year. Therefore, we take 2005 as the cut-off point to analyze the change in water retention ( Figure 3). The annual water retention increased from 347.14 mm in 2001 to 510.17 mm in 2005 and the lowest annual water retention was recorded in 2002 with 306.06 mm. From 2005 to 2015, the annual water production was 435.77 mm, with a growth rate of 2.71 mm/year −1 . Another abrupt change occurred in 2015, which shows the water production increased significantly from 435.77 mm to 518.75 mm. In 2018, the water yield in the HYR was the highest since 2001. Overall, the simulation results show that the water yield in the HYR continues to increase (Figure 4).  (Figure 4). In general, the spatial and temporal changes of water storage in grassland ecosystems are related to rainfall, vegetation cover, surface runoff and local ecological governance policies, and have seen high values in 2005 and 2015 [55][56][57][58]. The ecological protection project of the HYR is an important part of the Sanjiangyuan ecological protection program. The "Sanjiangyuan ecological protection project of China" is divided into two phases, every eight years; 2005 and 2014 are the two phases of the project, respectively. Affected by the policy, the ecological environment in the source area has been significantly improved, the vegetation coverage has increased, the overall production and water storage capacity has shown a steady growth trend, the ecological service function of water conservation in the source area has been improved and the ecological management has achieved remarkable results.

Ecological Service Value of Water Conservation in Grassland Ecosystem
The service value of grassland water conservation in the HYR is divided into five The ecological value of water conservation apparently differs between different grassland types. The low-value area improves grassland, but the value of ecological service also shows a weak growth trend from USD 3.3 × 10 8 in 2001 to USD 5.74 × 10 8 in 2020, with an annual contribution rate of 5%, distributed in the north of Zeku County. Alpine grassland in the high-value area was from USD 146.15 × 10 8 in 2001 to USD 227.59 × 10 8 in 2020, with an annual contribution rate of 80%. From high to low, the average ecological values and contribution rates over the years are: alpine meadow (66.93%) > temperate grassland (11.30%) > mountain meadow (7.85%) > lowland meadow (7.19%) > temperate desert (5.17%) > alpine grassland (1.03%) > improved grassland (0.51%) ( Figure 5). In general, the spatial and temporal changes of water storage in grassland ecosystems are related to rainfall, vegetation cover, surface runoff and local ecological governance policies, and have seen high values in 2005 and 2015 [55][56][57][58]. The ecological protection project of the HYR is an important part of the Sanjiangyuan ecological protection program. The "Sanjiangyuan ecological protection project of China" is divided into two phases, every eight years; 2005 and 2014 are the two phases of the project, respectively. Affected by the policy, the ecological environment in the source area has been significantly improved, the vegetation coverage has increased, the overall production and water storage capacity has shown a steady growth trend, the ecological service function of water conservation in the source area has been improved and the ecological management has achieved remarkable results.

Ecological Service Value of Water Conservation in Grassland Ecosystem
The service value of grassland water conservation in the HYR is divided into five intervals: 2001, 2005, 2010, 2015 and 2020. The total ecological value was USD 173.06 × 10 8 , USD 247.20 × 10 8 , USD 296.17 × 10 8 , USD 296.26 × 10 8 and USD 340.03 × 10 8 during 2001, 2005, 2010, 2015 and 2020, respectively, with the value of water conservation services increasing steadily over the past 20 years. The ecological value of water conservation apparently differs between different grassland types. The low-value area improves grassland, but the value of ecological service also shows a weak growth trend from USD 3.3 × 10 8 in 2001 to USD 5.74 × 10 8 in 2020, with an annual contribution rate of 5%, distributed in the north of Zeku County. Alpine grassland in the high-value area was from USD 146.15 × 10 8 in 2001 to USD 227.59 × 10 8 in 2020, with an annual contribution rate of 80%. From high to low, the average ecological values and contribution rates over the years are: alpine meadow (66.93%) > temperate grassland (11.30%) > mountain meadow (7.85%) > lowland meadow (7.19%) > temperate desert (5.17%) > alpine grassland (1.03%) > improved grassland (0.51%) ( Figure 5).
The value of water conservation in grassland ecosystems is increasing year after year, which further illustrates the importance of ecological protection. Monetary value cannot only make people intuitively feel the ecological value and improve the awareness of ecological protection but can also provide a reference for the formulation of local ecological protection and management policies and planning and design according to ecological values and land uses. The first phase of the ecological protection project invested RMB 7.5 billion, reaching one third of the assessed value in 2005; the second phase investment reached RMB 16 billion, reaching one half of the assessed value in 2015. Ecological restoration planning was classified and managed according to different ecosystems and land use types with remarkable achievements.
The value of water conservation in grassland ecosystems is increasing year after year, which further illustrates the importance of ecological protection. Monetary value cannot only make people intuitively feel the ecological value and improve the awareness of ecological protection but can also provide a reference for the formulation of local ecological protection and management policies and planning and design according to ecological values and land uses. The first phase of the ecological protection project invested RMB 7.5 billion, reaching one third of the assessed value in 2005; the second phase investment reached RMB 16 billion, reaching one half of the assessed value in 2015. Ecological restoration planning was classified and managed according to different ecosystems and land use types with remarkable achievements.

Spatial Division of Water Conservation Capacity and its Importance
Five levels of water conservation capacity were identified. The lowest value area (grade I) is concentrated in the northwest of Qumalai, Maduo and Chengduo counties; the low-value area (grade II) is in the southeast of Maduo and Chengduo counties and the northwest of Xinghai and Maqu counties. The medium-grade area (grade III) is in the northeast of Gande, Maqin, Dari and Xinghai counties; the high-grade area (grade IV) is in the northwest of Zeku, Henan, Maqu and Jiuzhi counties. The highest areas (grade V) are in Ruoergai, Aba and Hongyuan counties and account for 14.30%, 13.40%, 21.40%, 34.20% and 16.70% of the HYR, respectively. In general, the spatial distribution of water yield in the source area of the Yellow River is imbalanced and decreases from the southeast to the northwest, showing a step-like distribution ( Figure 6).

Spatial Division of Water Conservation Capacity and Its Importance
Five levels of water conservation capacity were identified. The lowest value area (grade I) is concentrated in the northwest of Qumalai, Maduo and Chengduo counties; the low-value area (grade II) is in the southeast of Maduo and Chengduo counties and the northwest of Xinghai and Maqu counties. The medium-grade area (grade III) is in the northeast of Gande, Maqin, Dari and Xinghai counties; the high-grade area (grade IV) is in the northwest of Zeku, Henan, Maqu and Jiuzhi counties. The highest areas (grade V) are in Ruoergai, Aba and Hongyuan counties and account for 14.30%, 13.40%, 21.40%, 34.20% and 16.70% of the HYR, respectively. In general, the spatial distribution of water yield in the source area of the Yellow River is imbalanced and decreases from the southeast to the northwest, showing a step-like distribution ( Figure 6).
The patterns of the water conservation importance are: the northwest of Qumalai, Maduo and Chengduo counties (mainly grade I); the northwest of Qumalai, Maduo and Chengduo districts (mainly grade II); the southeast of Maduo and Chengduo counties and the northwest of Xinghai and Maqu counties (mainly grade III); the higher-grade area (grade IV) is located in the northwest of Zeku, Henan, Maqu and Jiuzhi counties. The highest areas (grade V) are concentrated in Ruoergai, Aba and Hongyuan counties (mainly grade V) and account for 14.30%, 13.40%, 21.40%, 34.20% and 16.70% of the HYR, respectively. In general, the spatial division of water conservation importance in the HYR is uneven and decreases from the southeast to the northwest, showing a step-like distribution (Figure 7). The patterns of the water conservation importance are: the northwest of Qumalai, Maduo and Chengduo counties (mainly grade I); the northwest of Qumalai, Maduo and Chengduo districts (mainly grade II); the southeast of Maduo and Chengduo counties and the northwest of Xinghai and Maqu counties (mainly grade III); the higher-grade area (grade IV) is located in the northwest of Zeku, Henan, Maqu and Jiuzhi counties. The highest areas (grade V) are concentrated in Ruoergai, Aba and Hongyuan counties (mainly grade V) and account for 14.30%, 13.40%, 21.40%, 34.20% and 16.70% of the HYR, respectively. In general, the spatial division of water conservation importance in the HYR is uneven and decreases from the southeast to the northwest, showing a step-like distribution ( Figure 7).
Overall, the spatial division of water conservation capacity and its importance is an indicator of the ecosystem assessment of water conservation capacity. In this study, by dividing the importance level of the water conservation capacity, the spatial mapping of the water conservation function facilitates the division and management of the regional ecological reserves.

Regionalization of Z-Value Parameters in the Model
At present, only a few researchers have used the water production module of the InVEST model to analyze the ecological service function of water conservation in the HYR. Yang et al., used the InVEST model to obtain the total water production and to trend change dynamics of the grassland in the Yellow River basin from 1990 to 2018. Our results demonstrated that, when Z is taken as 3.6, the average annual water production is simulated as (534 × 10 8 m 3 ) and the actual average water production in 1995, 2005 and 2018 was (596 × 10 8 m 3 ), which is relatively close to their study results [59]. A recent study conducted Overall, the spatial division of water conservation capacity and its importance is an indicator of the ecosystem assessment of water conservation capacity. In this study, by dividing the importance level of the water conservation capacity, the spatial mapping of the water conservation function facilitates the division and management of the regional ecological reserves.

Regionalization of Z-Value Parameters in the Model
At present, only a few researchers have used the water production module of the InVEST model to analyze the ecological service function of water conservation in the HYR. Yang et al., used the InVEST model to obtain the total water production and to trend change dynamics of the grassland in the Yellow River basin from 1990 to 2018. Our results demonstrated that, when Z is taken as 3.6, the average annual water production is simulated as (534 × 10 8 m 3 ) and the actual average water production in 1995, 2005 and 2018 was (596 × 10 8 m 3 ), which is relatively close to their study results [59]. A recent study conducted by Zhao et al., calculated the water volume of grassland in the source area of the Yellow River, and the value of Z is 3.5. It is obtained that the average water production for many years from 2001 to 2020 is 509 × 10 8 m 3 , which is similar to the results of the present study (549 × 10 8 m 3 ) [60]. Kang et al., analyzed the annual water production and its change trend in the source area of the Yellow River from 1990 to 2010. They calculated the value of Z as 5, and the simulated total water production in 2010 was 0.27 × 10 8 m 3 , lower than the actual total water production in 2010 (497.27 × 10 8 m 3 ). The reason for this unreasonable value is that the simulated and measured values are not verified [61]; therefore, Z = 5 is not applicable to the source area of the Yellow River. In this study, based on the annual average water resources of the Tangnaihai hydrological station in the Yellow River Water Resources Bulletin [62], the water yield results of the InVEST model were repeatedly verified ( Figure 8). The InVEST model is based on ArcGIS software for data processing and mapping of the environmental variable grid. We take Z = 3.5 as the starting value [60], design the difference (every 0.5 is an interval and gradually increase the difference to run the model. When Z = 5, the average annual water production in 2001, 2005, 2010 and 2020 is close to the value of the actual water production in the bulletin (Figure 8).

Improvement Measures of Water Conservation Function
The contribution rate of improved grassland to the value of water conservation function in the source area is only 0.51%, which plays a greater role in maintaining biodiversity

Improvement Measures of Water Conservation Function
The contribution rate of improved grassland to the value of water conservation function in the source area is only 0.51%, which plays a greater role in maintaining biodiversity and productivity, while the restoration ability of water conservation and ecological function is lower [63]. On the other hand, the original alpine grassland plays a significant role, ensuring 60% of the water conservation and ecological value of the source area. Ecological protection projects and policies play a positive role in the restoration of regional ecological functions [59]. The previous "one size fits all" management model is no longer applicable. Due to the change of management ideas, the goal of restoring grassland productivity is changed to the goal of restoring ecological functions. According to the classification standard of ecological service value, ecological management is being carried out on the grassland in the study area and the high-value area is fenced. Generally, the vegetation coverage has increased by 10% and the biomass has increased by 50g/m 2 in the past three years, which can effectively enhance the water conservation function [64]. The median area is divided into rotational grazing and grassland livestock balance and the appropriate livestock carrying capacity of herdsman. The grasslands in the warm season (livestock carrying capacity = their own grassland area/10.62) and the cold season (livestock carrying capacity = their own grassland area/25.21) are checked and verified, respectively, in order to curb the phenomenon of overgrazing and to reduce the ecological pressure of the grasslands [65,66]. In low-value areas, grassland fertilization can increase the content of soil organic carbon and phosphorus, change the structure of soil particle size, increase soil water content, reduce soil erosion and improve the water conservation function of grassland [66].

Ecological Priority and Regional Management
From 2005 to 2010, the implementation of China's "phase I project of ecological protection and construction in the HYR" effectively curbed the trend of ecological degradation in the HYR and the ecological situation in the source area showed a stable improvement. In 2018, the water production of the source area was 498.79 mm, reaching the highest value in the recent 20 years. In January 2018, Sanjiangyuan National Park, the source of the Yellow River Park, carried out key ecological restoration projects. According to the remote sensing monitoring results of 2020, the vegetation coverage of sandy land in the source area increased from 33.36% to nearly 40%. The wetland vegetation coverage increased by more than 4%, the grassland vegetation coverage reached 74%, the grass yield reached 3082 kg/hm 2 , the water yield reached 587.5 mm and the water conservation function was significantly improved [67]. However, the ecological environment in the source area of the Yellow River has a stage of overall improvement and local degradation. Based on the grassland type, ecological function and ecological value, evaluating the water conservation function determines the priority protection areas and determines that Qumalai, Chengduo and Maduo counties are in extremely sensitive areas. The National Nature Reserve of the Yellow River source park is established in Maqu County to expand the ecological reserve to Qumalai and Chengduo counties, with Maqu County as the center. These three counties and districts are set as the priority protection areas for the ecological functions of water conservation.

Conclusions
The stability of the water conservation function and the evaluation of the ecological value in the HYR are related to the protection and management of water resources in the Yellow River basin. The quantitative analysis of water retention and the management of ecological spatial planning are the focus of the research. Based on the calculation of water retention using improved parameters of the InVEST model, this study analyzes the spatio-temporal dynamic characteristics of water conservation in the headwaters of the Yellow River from 2001 to 2020; the modified NPP equivalent factor method is used to evaluate the ecological value of the water production services and to classify the ecological conservation areas. Finally, the following three conclusions are drawn: From 2001 to 2020, the average annual water yield of the study area was 443.86 mm and the average annual total water yield was 4.29 × 10 10 m 3 , with an increase in water retention in most areas. The area with significant growth accounted for 9.44% and is mainly located in the southeast of Zeku County, Henan, the middle of Maqin and the north of Qumalai; 90.56% of the regional water retention showed a slight upward trend, with an average annual growth rate of 7.5 mm/year −1 . The water production capacity of the study area is divided into five levels. Low-value areas are in the northwest of Qumalai County, Maduo County and Chengduo County, while high value areas are in Ruoergai, Aba and Hongyuan counties.
The total value of water conservation ecological services in the grassland ecosystem increased from USD 146.15×10 8 in 2001 to USD 227.59 × 10 8 in 2020, showing a steady growth trend. The areas with low value are in the improved grasslands, while the areas with high value are distributed in the alpine grassland. The ecological value of alpine grassland increased from USD 146.15 × 10 8 in 2001 to USD 227.59 × 10 8 in 2020, with an average annual growth rate of 80%.
The Chinese government has carried out ecological protection projects in the HYR. The implementation of the projects has achieved remarkable results, the grassland degradation has been alleviated and the water conservation function has been significantly improved. Qumalai, Chengduo and Maduo counties are in extremely important areas. It is proposed to classify them into priority areas for the ecological protection of water resources. In areas with high ecological values such as alpine meadows, temperate grasslands and mountain meadows in the headwaters, regional rotational grazing and grass balance should be introduced, and a comprehensive grazing ban should be implemented for lowland meadows, temperate deserts and alpine grasslands. Our suggestions are based on the policies and measures implemented in previous ecological projects, which are reasonable for the restoration and management of the water conservation function in the HYR.
This study evaluates the ecological function and value of water conservation in the HYR, provides reference data and policy suggestions for the formulation of ecological conservation measures and technical project investments and is of great significance for the sustainable development of water resources in the headwaters and watershed of the Yellow River.