Remote-Sensing-Based Assessment of the Ecological Restoration Degree and Restoration Potential of Ecosystems in the Upper Yellow River over the Past 20 Years

Abstract

The Upper Yellow River is the most important area for water retention and flow production in the Yellow River basin, and the statuses of the ecosystems in this region are related to the ecological stability of the whole Yellow River basin. In this paper, the fractional vegetation cover (FVC), net primary productivity (NPP) of vegetation and water retention, soil retention, and windbreak and sand fixation services of the Upper Yellow River ecosystems were analysed from 2000 to 2019 with the trend analysis method. Ecological restoration degree evaluation indices were constructed to comprehensively assess the ecological restoration situation and restoration potential of the ecosystems in the Upper Yellow River region over the past 20 years and to quantitatively determine the contribution rates of climate factors and human activities to these ecosystem changes. The results showed that the settlement ecosystem area exhibited the greatest increase, while the grassland ecosystem area decreased significantly over the study period. In the Upper Yellow River region, the ecosystem quality and ecosystem services generally remained stable or improved. Areas with moderately, strongly and extremely improved ecological restoration degrees accounted for 32.9%, 21.0% and 2.8% of the entire Upper Yellow River region, respectively. Areas with strongly improved and extremely improved ecological restoration degrees were mainly distributed in the Loess Plateau gully areas and on the eastern Hetao Plain. The contribution rates of climatic factors and human activities to the NPP changes measured in the Upper Yellow River were 81.6% and 18.4%, respectively, while the contribution rates of these processes to soil erosion modulus changes were 77.6% and 22.4%, respectively. The restoration potential index of the FVC in the Upper Yellow River was 22.7%; that of the forest vegetation coverage was 14.4%; and that of the grassland vegetation coverage was 23.0%. Over the past 20 years, the ecosystems in the Upper Yellow River region have improved and recovered significantly. This study can provide scientific support for the next stage of ecological projects in the Upper Yellow River region.

1. Introduction

Ecological systems and the service functions they provide form the foundation of human survival and sustainable social development. As the global population increases and human development causes further destruction to the environment, ecological systems are disturbed, and eventually degraded, issues that have become major problems faced when considering the global environment. Ecological restoration refers to a process by which damaged or degraded ecosystems are restored to a certain or improved structure or functional level through restoration, improvement or renewal processes performed with certain biotechnologies or engineering technologies at the ecosystem level [1]. The restoration potential was developed to measure the conceptual distance between a habitat and a corresponding reference target [2]. Thus, comprehensively maintaining existing healthy ecosystems and restoring degraded ecosystems have become important research topics.

In 2003, the United Nations launched the Millennium Ecosystem Assessment (MA) [3], in which a framework and index systems describing ecosystem assessments were proposed for the first time alongside strategies for curbing ecosystem degradation. The United States [4], the United Kingdom [5] and other countries have successively conducted national ecosystem assessments; China was one of the first countries to carry out ecosystem assessment and restoration research [6]. Remote sensing methods have been widely applied in ecological status change research performed at the global or regional scale due to advantages associated with the macro-scale sampling abilities, rapid sampling abilities and economic characteristics of these techniques. Scholars from various countries have utilized remote sensing indices mainly to evaluate ecosystem status changes, including by applying individual indices and multiple indices. Some past evaluations have been based on the use of single indices, such as the normalized difference vegetation index (NDVI) [7] or the net primary productivity (NPP) [8]. However, the abilities of these individual indices to reflect the ecological status of an entire region are insufficient, and entire ecosystems thus cannot be comprehensively evaluated using single evaluators. To combat this, some evaluations have been performed based on multiple indices, such as the global disturbance index (MGDI) derived from Moderate-resolution Imaging Spectroradiometer (MODIS) sensor measurements [9] and the remote-sensing-based ecological index (RSEI) [10]. Although these remotely sensed multi-indices are constructed by adopting multiple indicators, they still face a shortcoming in that they cannot comprehensively reflect the information of regionally integrated ecosystems. Thus, monitoring ecological status changes based on these comprehensive remote sensing indices is still challenging, and a complete remote sensing index evaluation system must be established to facilitate comprehensive evaluations of regional ecosystem status changes.

The Yellow River basin is an ecological corridor that connects the Qinghai–Tibet Plateau, Loess Plateau and North China Plain. This basin serves as an important ecological security barrier and is a critical area for human activities and economic development in China. The headwaters of the Yellow River, the Gannan region and the southern Qilian Mountains are important areas for water retention in the Yellow River basin. The Toudaoguai hydrological station records the annual runoff of the Upper Yellow River, accounting for more than 60% of the water in the Yellow River basin [11]; this region is the most important runoff-producing area within the Yellow River basin. The Loess Plateau region in the Upper Yellow River watershed accounts for 53.3% of the areal extent of the region, and the Hetao Plain is one of the main agricultural production areas in the Yellow River basin. In recent decades, under the influence of climate change and human activities, ecological problems have been frequently observed in the Upper Yellow River region; for example, at the turn of the century, land fragmentation and overgrazing caused grassland degradation in the Yellow River source area and the Qilian Mountain region; indiscriminate mining has left the Qilian Mountains riddled with holes; soil erosion has intensified in the gully areas of the Loess Plateau; water consumption for agricultural irrigation use on the Hetao Plain has increased; the runoff of the Upper Yellow River has presented a decreasing trend [12,13,14]; and a series of problems have seriously threatened the ecological security and sustainable development of the social economy in the Upper Yellow River region. To this end, China has carried out many ecological projects in the Upper Yellow River region, including ecological conservation and restoration projects in the Three-River Sources Region [15], ecological protection and construction projects aimed at protecting the Yellow River region in the Gannan Tibetan Autonomous Prefecture (a region that provides important ecological functions in the form of water supply), comprehensive ecological protection and construction projects developed in the Qilian Mountains and soil and water retention projects sited on the Loess Plateau. At the same time, the Upper Yellow River region has been involved in key national ecological projects, such as the Natural Forest Protection Project [16], Grain for Green Project [17] and Grazing Withdrawal Project [18]. Many scholars have performed extensive research on the ecological environment of the Yellow River basin. Some scholars have used remotely sensed time-series data, mainly in combination with trend analyses or partial correlation analyses, to unilaterally study vegetation status changes [19], climate change [20] and land use changes [21] in the Yellow River basin. Some scholars have performed comprehensive assessments of the ecosystem types and vegetation [22,23] present in the Yellow River basin. These studies have revealed the changes that have occurred in the ecological status of the Yellow River basin, but no relevant research on ecological restoration or the restoration potential has been carried out in this region. Other scholars have focused on the ecological protection and restoration of important ecological function areas in the Yellow River basin, such as the Three-River Region [15], the Qilian Mountains [24], the Ruoergai grasslands and wetlands [25] and the Loess Plateau [26]. However, to date, few comprehensive studies have explored the overall ecological environment in the Upper Yellow River region. Many scholars have attempted to study the ecosystem restoration potential in this region using quantitative evaluation methods, but at present, most of these published studies involve qualitative evaluations of empirical indicators, while relatively few quantitative measurements have been performed.

In this paper, we selected three categories, the ecosystem macrostructure, the ecosystem quality and ecosystem services, along with six first-level indicators and nine s-level indicators. Using multisource remote sensing data and ground-measured data in combination with the model simulation method, we generated a parameter dataset containing ecosystem status assessment indicators characterizing the Upper Yellow River region from 2000 to 2019 and quantitatively evaluated the ecological restoration degree and potential in this region. Through this work, we aim to provide supporting data and a decision-making reference for further ecological protection, high-quality development and horizontal ecological compensation measures in the Yellow River basin.

2. Materials and Methods

2.1. Study Area

The Upper Yellow River is located in the alternating zone of the first and second terrain steps in China; the length of the river is 3472 km, and the basin area is 429,000 km² (Figure 1). The climate types in this region are complex and diverse, spanning the transition from alpine and humid areas to arid desert regions. The evaporation of rainfall is unevenly distributed throughout the year. Approximately 75% of rainfall is concentrated from June to September, while evaporation is mainly concentrated from April to August [27]. The Upper Yellow River can be divided into three zones according to the data published in the China Hydrological Yearbook. Zone I is upper section of the Upper Yellow River, includes the source region of the Yellow River on the eastern Qinghai–Tibet Plateau and part of the Loess Plateau; this region is the main water-producing area, containing rich hydropower resources spread over its area of 250,000 km². The average annual precipitation in this zone is 408 mm, the average annual temperature is −1.5 °C and the actual annual evapotranspiration is 414 mm. Zone II is lower section of the Upper Yellow River, includes the Kubuqi Desert and Hetao Plain on the western Loess Plateau and Inner Mongolia Plateau. This zone is the main irrigated agricultural region in the Upper Yellow River, covering an area of 134,000 km² with an annual average precipitation total of nearly 300 mm, an annual average temperature of 1.5 °C and an actual annual evapotranspiration total of 317 mm. Zone III is the inner flow region, including the Maowusu Sandy Land and the northern Loess Plateau, spanning an area of 45,000 km². The annual precipitation total in this region is only 173 mm, the annual mean temperature is 7.6 °C and the actual annual evapotranspiration total is 127 mm.

2.2. Acquisition and Validation of Remote Sensing Index Parameter Data

We selected three categories for our analysis, the ecosystem macrostructure, the ecosystem quality and ecosystem services, alongside six first-level indicators and nine s-level indicators [28] to evaluate the ecosystem status changes that have occurred in the Upper Yellow River region over the past 20 years by applying remote sensing techniques (

Table 1. Indices used to assess the ecological restoration degree.

Category	Index
	First-Level	Second-Level
Ecosystem macrostructure	Area of ecosystem	Classified area of ecosystem
		Rate of change
Ecosystem quality	Net primary productivity (NPP)	NPP
	Fractional vegetation cover (FVC)	FVC
Ecosystem services	Water retention	Water retention
	Soil retention	Soil erosion modulus
		soil retention
	Windbreak and sand fixation	Wind erosion modulus
		Windbreak and sand fixation

).

2.2.1. Acquisition and Verification of the Ecosystem Macrostructure Data

In this study, we chose to use land use data from the China multiperiod land use and land cover remote sensing monitoring dataset (CNLUCC) [29,30]. These data were obtained by the Landsat 8 satellite and verified using GaoFen 2 (GF-2) satellite and unmanned aerial vehicle (UAV) data. The classification accuracy and total accuracy of these data were evaluated using a confusion matrix, and the comprehensive evaluation accuracy reached 93%. We selected the datasets comprising land use type information in the Upper Yellow River in 2000, 2010 and 2018. According to the terrestrial ecosystem macrostructure classification system and the corresponding conversion relationships with the land use/land cover classification system [31], the land use data characterizing 2000, 2010 and 2018 were used to generate a 1 km gridded dataset of ecosystem type percentages.

2.2.2. Acquisition and Validation of the Ecosystem Quality Data

In this study, we calculated the fractional vegetation cover (FVC) in the Upper Yellow River using MOD13Q1 products. These products have a temporal resolution of 16 days and a spatial resolution of 250 m. We carried out Savitzky–Golay filtering on the MOD13Q1 products, calculated the FVCs characterizing 16 days through a pixel dichotomy method and generated the annual FVCs with the maximum synthesis method. The accuracy of the annual FVC data was verified by calculating the FVCs using UAV data derived from field surveys conducted in the corresponding years (R² = 0.52). Then, we resampled the annual FVC data spanning the 2000 to 2019 period to a spatial resolution of 1 km.

The net primary productivity (NPP) was measured using the MOD17A3HGF product derived from MODIS. This product has a one-year temporal resolution and a spatial resolution of 500 m. We verified the accuracy of the MODIS NPP data by using biomass data collected from 168 grassland quadrangles [32] and productivity data obtained from typical forest ecosystems in China (R² = 0.75) [33]. Finally, we resampled the 2000 to 2019 NPP dataset to a spatial resolution of 1 km.

2.2.3. Estimation and Validation of the Ecosystem Service Data

(1)

Water Retention

In this study, we used the water balance method to estimate the magnitude of the water retention ecosystem service in the Upper Yellow River. To this end, the integrated valuation of ecosystem services and trade-offs (InVEST) model and the methods described in the “Guide to Delimiting Ecological Red Lines” were comprehensively applied. The following equation was used to assess the water retention service:

$$Q_{\mathrm{wr}}=P-ET-R=(1-\frac{ET}{P})\times P-u\times P$$

(1)

where $Q_{\mathrm{wr}}$ is the amount of water conserved (mm); P is precipitation (mm); R is surface runoff (mm), obtained by multiplying runoff coefficient by precipitation [34]; ET is the actual evapotranspiration (mm) [35,36,37]; and u is the runoff coefficient [34].

We verified the water retention simulation results using annual runoff data monitored at the Jimai and Tangnaihai hydrological stations in the source region of the Yellow River from 1990 to 2018, and the R² values of the datasets constructed for these two hydrological stations were 0.61 and 0.44, respectively (p < 0.01). Finally, we generated a water retention service dataset spanning the period from 2000 to 2019 with a spatial resolution of 1 km.

(2)

Soil Retention

The ecosystem soil retention service represents the difference between the potential soil erosion under bare-soil conditions and the soil erosion that occurs under the true vegetation conditions. In this study, we used the revised universal soil loss equation (RUSLE) [38] to estimate the soil erosion modulus in the Upper Yellow River as follows:

$$SC=R\times K\times LS\times (1-C)\times P$$

(2)

where R is the rainfall erodibility factor (MJ·mm·ha⁻¹·h⁻¹·yr⁻¹) [34]; K is the soil erodibility factor (t·h·MJ⁻¹·mm⁻¹) [39,40]; LS is the slope length and slope factor [41,42]; C is the vegetation coverage factor; P is the soil and water retention measure factor; and LS, C and P are dimensionless parameters [34].

Through comparisons with the sediment concentration monitoring data recorded at the Jimai and Tangnaihai hydrological stations in the Yellow River watershed, we verified the simulation results derived using the soil erosion modulus and obtained R² values of 0.79 and 0.57 (p < 0.01), respectively. Finally, we generated a soil retention service dataset spanning the 2000–2019 period at a spatial resolution of 1 km.

(3)

Windbreak and Sand Fixation

The windbreak and sand fixation magnitudes in an ecosystem represent the difference between the potential wind erosion that would occur under bare-soil conditions and the actual wind erosion that occurs under the real vegetation conditions. We used the modified revised wind erosion equation (RWEQ) [43] to estimate the wind erosion modulus in the Upper Yellow River as follows:

$$SL=\frac{Q_{\max }\left[1-e^{{\left(\frac{X}{S}\right)}^2}\right]}{X}$$

(3)

$$Q_{max}=109.8\times (WF\times EF\times SCF\times K^{\prime }\times COG)$$

(4)

$$S=150.71\times {(WF\times EF\times SCF\times K^{\prime }\times COG)}^{-0.3711}$$

(5)

where SL is the soil wind erosion modulus (kg·m⁻²); X is the block length (m); $Q_{\max }$ is the maximum sand transport capacity of wind (kg·m⁻¹); S is the length of the key block (m); WF is the meteorological factor (kg·m⁻¹); EF is the soil erodibility factor; SCF is the soil crust factor; K^′ is the soil roughness factor; COG is a comprehensive vegetation factor; and EF, SCF, K^′ and COG are dimensionless parameters [34].

In this paper, the wind erosion modulus values estimated using the RWEQ model were compared with the values derived from an empirical wind erosion prediction model based on a wind tunnel test [44], and the resulting R² value was 0.40. Finally, we generated a serviceable dataset containing windbreak and sand fixation information at a spatial resolution of 1 km covering the period from 2000 to 2019.

2.3. Assessment Methods of the Ecological Restoration Degree and Restoration Potential

2.3.1. Assessment Method of the Restoration Tendency and Ecological Restoration Degree

In this work, the restoration tendency refers to the continuous change tendencies of a single assessment index from 2000 to 2010 and from 2010 to 2019. We calculated the slope (P) of the FVC, NPP and water retention, soil retention, windbreak and sand fixation ecosystem services in the Upper Yellow River region by using the least-squares method from 2000 to 2010 and from 2010 to 2019. When p > 0.05, the restoration tendency was considered to have improved; when −0.05 ≤ p ≤ 0.05, the restoration tendency was considered to be basically stable; and when p < −0.05, the restoration tendency was considered to have worsened. For the soil erosion modulus and wind erosion modulus, the opposite was true [28]. The ecosystem quality and ecosystem service trends were thus judged in the Upper Yellow River region from 2000 to 2019 based on the categories listed in

Table 2. Basis for judging the restoration tendencies of the ecosystem quality and ecosystem service parameters.

Judgement Basis		Judgement Result
2000–2010	2010–2019	Overall Restoration Tendency from 2000 to 2019
Improved	Improved	Continuously improved
Improved	Worsened	Continuously worsened
Remained stable	Remained stable	Continuously remained stable
Improved	Worsened	First improved and then worsened
Improved	Remained stable	First improved and then remained stable
Worsened	Improved	First worsened and then improved
Worsened	Remained stable	First worsened and then remained stable
Remained stable	Improved	First remained stable and then improved
Remained stable	Worsened	First remained stable and then worsened

.

The ecological restoration degree is the degree of regional restoration generated by the superposition of different states of the multiple considered assessment indicators (this degree can either improve, remain stable or worsen). We selected the NPP, FVC, water retention service, soil erosion modulus and wind erosion modulus values derived from 2000 to 2019 and applied the Sen method to calculate the slope (P); then, we conducted a significance test by using the Mann–Kendall method. To obtain the spatial distribution data of the three kinds of change trends that each index could characterize (the improving, remaining stable and or worsening trends), we conducted a superposition analysis on the spatial data characterizing the slope changes of the five indicators. The more indicators exhibited improvements, the higher the ecological restoration degree was; thus, the overall ecological restoration degree of the Upper Yellow River could be obtained (

Table 3. Basis for judging the ecological restoration degree.

Ecological Restoration Degree	Judgement Condition 1
Basically remained stable	Si ≥ 3
Slightly worsened	Si < 3 and Wi = 2
Moderately worsened	Si < 3 and Wi = 3
Significantly worsened	Si < 3 and Wi ≥ 4
Extremely improved ecological restoration degree	Bi ≥ 4
Strongly improved ecological restoration degree	Bi = 3
Moderately improved ecological restoration degree	Si < 3 and Wi < 2 and Bi = 2
Some elements improved while some elements worsened	Si < 3 and Wi < 2 and Bi = 1

¹ In the table, W_i represents the number of indicators that exhibited deterioration, B_i represents the number of indicators that exhibited improvement, S_i represents the number of indicators that basically remained stable, and i ≤ 5.

).

2.3.2. Assessment Method of the Restoration Potential

On the basis of Shao’s improvement to the second-level national ecological-geographical division scheme [28,45], we regarded the ecosystems in national and provincial nature reserves as the climax ecological background corresponding to the same ecological geographical area and the same ecosystem type.

In this paper, the annual maximum FVC of the Upper Yellow River region from 2000 to 2019 was spatially superimposed on the regional ecological geographical zones and nature reserves. We extracted the FVCs of forest, grassland and desert ecosystems in nature reserves in each ecological-geographical zone and established linear regression equations with air temperature and precipitation:

$$EPRI=(\frac{E{R}_t-E{R}_c}{E{R}_t})\times 100\%$$

(6)

where ERPI is the ecological restoration potential index of the target area, ER_t is the upper ecological background value of the corresponding ecological-geographical zone and ER_c is the ecological restoration status value. The greater the ecological restoration potential index is, the greater the ecological restoration potential is.

2.4. Methods for Determining the Contributions of Climate Factors and Human Activities to Ecosystem Changes

We applied a residual analysis [46] to determine the contributions of human activities and climate factors to NPP changes from 2000 to 2019. Based on temperature, precipitation and NPP data, we constructed a binary linear regression model, calculated the deviation between the NPP data and regression-estimated values and quantitatively reflected the impacts of ecological engineering on NPP.

We used the model variable control method [34,47] to determine the contributions of human activities and climate factors to changes in soil retention services from 2000 to 2019. We controlled the climate parameters (temperature and precipitation) in the RUSLE equation by establishing constant values and thus estimated the soil erosion modulus under fixed climate conditions.

3. Results and Analysis

3.1. Temporal and Spatial Ecosystem Changes

3.1.1. Analysis of Ecosystem Macrostructure Changes

In 2018, the area of grassland ecosystems in the Upper Yellow River was the largest among all ecosystem types, accounting for 61.4% of the land area. The area of settlement ecosystems was the smallest, accounting for only 1.1%. The areas of farmland, forest, water and wetland and desert ecosystems accounted for 14.6%, 7.9%, 3.2% and 7.6% of the overall area, respectively, while other ecosystems accounted for 4.2% (Figure 2).

Over two time periods (2000–2010 and 2010–2018), the macrostructures of the terrestrial ecosystems in the Upper Yellow River changed significantly. The areas of forest, water and wetland and settlement ecosystems continuously increased, while the areas of grassland, farmland, desert and other ecosystems continuously decreased. From 2000 to 2018, the areas of settlement, forest, water and wetland ecosystems increased by 2661.6 km², 1246.1 km² and 1136.2 km², respectively, while the areas of grassland, farmland, desert and other ecosystems decreased by 2568.4 km², 1254.5 km², 885.8 km² and 335.2 km², respectively.

3.1.2. Temporal and Spatial Changes in Ecosystem Quality

In 2019, the average FVC in the Upper Yellow River region was 47.79%, while the average FVC over the 20 previous years was 44.22%. From 2000 to 2019, the FVC in the Upper Yellow River showed an overall increasing trend. In Zone I, the FVC increased the most in the Loess Plateau area. The FVCs decreased in some desert areas on the Mongolian Plateau in Zone II (Figure 3a).

In 2019, the average NPP of vegetation in the Upper Yellow River region was 251.17 gC·m⁻¹·a⁻¹, with an average value of 212.39 gC·m⁻¹·a⁻¹ calculated for the period comprising the previous 20 years. From 2000 to 2019, the annual NPP change rate in the Upper Yellow River was 3.22 gC·m⁻¹·a⁻¹, and the NPP generally increased steadily. In Zone I, the annual NPP increase rate on the Loess Plateau in central Gansu Province was the highest, reaching 17.81 gC·m⁻¹·a⁻¹(Figure 3b).

3.1.3. Temporal and Spatial Changes in Ecosystem Services

In 2019, the total amount of water conserved in the Upper Yellow River ecosystems was 461.50 × 108 m³·a⁻¹, higher than the average found for the past 20 years (

Table 4. Annual mean values of ecosystem services in the Upper Yellow River region.

Ecosystem Service		2000–2010	2010–2019	2000–2019
Water retention	Water retention per unit area (m3·hm−2·a−1)	786.82	869.47	825.76
	Total amount of water conserved (×108 m3·a−1)	336.79	372.16	353.45
Soil retention	Soil erosion modulus per unit area (t·hm−2·a−1)	9.69	9.93	9.85
	Total soil erosion modulus amount (×108 t·a−1)	3.65	3.75	3.71
	Soil retention per unit area (t·hm−2·a−1)	19.18	27.13	22.81
	Total amount of soil conserved (×108 t·a−1)	7.23	10.23	8.6
Windbreak and sand fixation	Wind erosion modulus per unit area (t·hm−2·a−1)	9.28	4.89	6.99
	Total wind erosion modulus amount (×108 t·a−1)	3.97	2.09	2.99
	Windbreak and sand fixation per unit area (·hm−2·a−1)	21.6	17.05	19.04
	Total windbreak and sand fixation amount (×108 t·a−1)	9.25	7.3	8.15

). From 2000 to 2019, the water retention change rate was 105.46 m³·hm⁻²·a⁻¹. The water retention in the source region of the Yellow River in Zone I increased rapidly (Figure 4).

In 2019, the soil erosion modulus in the water-eroded area of the Upper Yellow River was 9.37 t·hm⁻²·a⁻¹; this value was lower than the average value calculated over the past 20 years (Table 4, Figure 5a). In 2019, the total amount of soil conserved in the Upper Yellow River region was 15.02 × 108 t·a⁻¹, higher than the average value obtained for the past 20 years. From 2000 to 2019, the annual soil retention rate in the Upper Yellow River was 0.78 t·hm⁻²·a⁻¹ (Figure 5b), and the soil erosion modulus exhibited a slight increase, indicating that the soil retention service was stable and improving.

In 2019, the wind erosion modulus of the upper Yellow River was 2.42 t·hm⁻²·a⁻¹, which was lower than the average value derived over the past 20 years (Table 4, Figure 6a). The total windbreak and sand fixation amount was 6.6 × 108 t·a⁻¹. From 2000 to 2019, the average change rate of the wind erosion modulus in the Upper Yellow River region was −0.58 t·hm⁻²·a⁻¹ (Figure 6b). Among these regions, the wind erosion modulus in the western desert area on the Inner Mongolia Plateau in Zone II showed a decreasing trend, indicating that the windbreak and sand fixation amount decreased significantly in this region over the study period.

3.2. Restoration Tendency and Degree of Ecological Restoration in the Upper Yellow River Region

3.2.1. Restoration Tendency and Spatial Ecosystem Differences

(1)

Ecosystem Quality

Compared to that obtained for the 2000–2010 period, the annual mean maximum FVC in the Upper Yellow River increased by 4.5% from 2010 to 2019. The area of ‘continuously improving’ FVCs was the largest, accounting for 25.3% of the study area, while the area of ‘continuously worsening’ FVCs was the smallest, accounting for only 3.7% (

Table 5. Statistical table listing the restoration tendencies of the ecosystem quality and service factors considered herein.

Restoration Tendency	FVC Area Ratio (%)	NPP Area Ratio (%)	Water Retention Area Ratio (%)	Soil Retention Area Ratio (%)	Windbreak and Sand Fixation Area Ratio (%)
Continuously improving	25.3	55.1	28.0	10.4	2.2
Continuously worsening	3.7	0.1	2.4	0	20.9
Remaining stable	13.2	13.1	20.1	66.4	47.0
First improving and then worsening	15.2	1.7	17.6	0.89	2.0
First improving and then remaining stable	12.1	16.3	13.3	6.2	1.6
First worsening and then improving	6.4	1.2	1.4	0.3	6.7
First worsening and then remaining stable	3.8	0.3	0.7	0	10.7
First remaining stable and then improving	11.6	11.5	15.3	15.8	4.8
First remaining stable and then worsening	8.8	0.7	1.2	0.1	4.2

). The areas of continuous improvement were mainly distributed on the Loess Plateau in Zone I and on the Ordos Plateau in Zone II (Figure 7a).

Compared to that obtained for the 2000–2010 period, the annual mean NPP of vegetation in the Upper Yellow River increased by 15.6% from 2010 to 2019. In these two periods, the NPP continuously improved in 55.1% of the study area (Table 5). Over the past 20 years, the NPP of the vegetation in the Upper Yellow River region has remained stable and improved (Figure 7b).

(2)

Ecosystem Services

Compared to that obtained for the 2000–2010 period, the average amount of water conserved in the Upper Yellow River region increased by 10.5% from 2010 to 2019 (Table 4). In these two periods, the areas in which the water retention service ‘continuously improved’ accounted for 28.0% of the study area; ‘remained stable’ accounted for 20.1%; and ‘continuously worsened’ accounted for 2.4% (Table 5). Among these areas, the regions in which continuous improvements in water retention services were identified were mainly distributed in the source region of the Yellow River in Zone I, on the Hetao Plain in Zone II and in the eastern Mu Us Desert in Zone III (Figure 7c).

Compared to that obtained for the 2000–2010 period, the soil retention amount increased by 41.5% from 2010 to 2019 (Table 4). In these two periods, the areas in which ‘continuously improving’ soil retention services were found accounted for 10.4% of the study area, while ‘remained stable’ accounted for 66.4% (Table 5). The regions with ‘continuously improving’ soil retention services were mainly distributed in the northern source region of the Yellow River in Zone I and on the Loess Plateau (Figure 7d).

Compared to that obtained from 2000–2010, the wind erosion modulus in the Upper Yellow River region decreased by 47.4% from 2010 to 2019, while the windbreak and sand fixation amount decreased by 21.1% (Table 4). During these two periods, the areas in which the windbreak and sand-fixing service ‘remained stable’ accounted for 47.0% of the study area, while the areas that exhibited ‘continuous improvement’ and ‘continuously worsening’ comprised 2.2% and 20.9% (Table 5). The areas with ‘continuously worsening’ windbreak and sand-fixation services were mainly located on the northern Loess Plateau in Zone II, in the Kubuqi Desert on the Inner Mongolia Plateau and in the Mu Us Sandy Land region in Zone III (Figure 7e).

3.2.2. Spatial Differences in the Ecosystem Restoration Degree

From 2000 to 2019, the ecosystem restoration degree was strongly improved in the Upper Yellow River region, as shown in Figure 8. Regions exhibiting ‘moderately improved ecosystem restoration degrees’ covered the largest area, accounting for 32.9% of the Upper Yellow River region, and these areas were mainly distributed in the source region of the Yellow River in Zone I and on the northern Loess Plateau in Zone II. The next-largest area comprised regions with ‘strongly improved ecological restoration degrees’, accounting for 21.0% of the study area, and these areas were mainly distributed in the Kubuqi Desert on the Mongolian Plateau in Zone II and in the eastern Mu Us Desert in Zone III. The areas with ‘extremely high ecological restoration degrees’ accounted for 2.8% of the total study area, and these regions were mainly distributed in the middle Gansu region on the Loess Plateau in Zone II and in Lanzhou city of Gansu Province and its surrounding areas in Zone I. In the Upper Yellow River regions, areas characterized by ‘slightly worsened’, ‘moderately worsened’ and ‘significantly worsened’ ecological restoration degrees accounted for 4.9%, 0.6% and 0% of the study area, respectively. The areas exhibiting ecological stability accounted for 21.2% of the study area and were mainly distributed in the desert area on the western Mongolian Plateau in Zone II (

Table 6. Statistical table reflecting ecological restoration changes in the study area.

Ecological Restoration Degree Change	Area Ratio of Zone I (%)	Area Ratio of Zone II (%)	Area Ratio of Zone III (%)
Basically remained stable	24.5	16.9	16.0
Slightly worsened	6.3	3.4	1.6
Moderately worsened	0.8	0.3	0.1
Significantly worsened	0	0	0
Extremely improved ecological restoration degree	3.4	2.6	0.2
Strongly improved ecological restoration degree	20.3	21.9	22.5
Moderately improved ecological restoration degree	30.8	35.4	36.4
Some elements improved while some elements worsened	13.9	19.5	23.1

).

3.3. Contributions of Climate Factors and Human Activities to Ecosystem Quality and Service Changes

Climate factors and human activities contributed 81.6% and 18.4%, respectively, to the NPP change calculated in the Upper Yellow River region from 2000 to 2019 (

Table 7. Contribution rates of human activities and climate factors to NPP and soil erosion modulus changes from 2000 to 2019.

Zone	Changes in NPP		Changes in the Soil Erosion Modulus
	Human Activities (%)	Climate Factors (%)	Human Activities (%)	Climate Factors (%)
Upper Yellow River	18.4	81.6	22.4	77.6
Zone I	14.6	85.4	22.8	77.2
Zone II	23.7	76.3	24.0	76.0
Zone III	24.5	75.5	16.2	83.8

). Climate factors were the main drivers influencing NPP during the study period. In each region, the contribution rate of human activities to NPP was consistent with the NPP restoration tendency.

Climate factors and human activities contributed 77.6% and 22.4%, respectively, to the soil erosion modulus change derived from 2000 to 2019 (Table 7). Climate factors were the main drivers influencing changes in the soil erosion modulus. In Zone I and Zone II in the Upper Yellow River region, human activities contributed more than climate factors to the soil erosion modulus changes, mainly due to the implementation of measures aiming to return farmlands to forests and grasslands in the source region of the Yellow River, the Three Rivers region, the Qilian Mountains and southern Gansu Province, as these measures reduced the degree of soil erosion. In addition, in the inner flow area, soil erosion was less severe.

3.4. Ecological Restoration Potential of the Main Ecosystem Types in the Study Area

As shown in Figure 9, the average FVC among the forest, grassland and desert ecosystems in the Upper Yellow River region was 61.3% under the climax ecological conditions (Figure 9a). The gap between this ideal value and the actual average FVC of forest, grassland and desert ecosystems was 13.9%, and the ecological restoration potential index was 22.7% (Figure 9b).

The FVCs derived for forest, grassland and desert ecosystems under the climax ecological condition were 88.3%, 62.2% and 23.9%, respectively. The differences between these ideal values and the actual FVCs of forest, grassland and desert ecosystems were 12.7%, 14.3% and 11.5%, respectively. The FVC of the desert ecosystems exhibited the highest restoration potential, at 48.0%. The restoration potential of the forest ecosystems was the lowest, with a restoration potential index of only 14.4%. The restoration potential index of the grassland ecosystem FVC was 23.0%.

4. Discussion

4.1. Impacts of the Climate Pattern on Ecological Restoration

From 2000 to 2019, the annual mean temperature in the Upper Yellow River region gradually decreased from east to west, while the annual mean precipitation gradually decreased from south to north. The ecological restoration degree data obtained for the Upper Yellow River region overlapped with the annual mean temperature and annual mean precipitation data. The regions with relatively high and extremely high ecological restoration degrees were mainly distributed in regions with annual mean temperatures between 6 °C and 9 °C and annual mean precipitation totals between 300 mm and 400 mm (Figure 10 and Figure 11). In addition, the temperature and precipitation changes recorded over the past 20 years were not significant in the study area, exhibiting nonsignificant increases or basically remaining unchanged. These results indicate that the regional climate pattern greatly impacted ecological restoration in the study area and that poor climate conditions or severe climate change were not conducive to ecological restoration.

4.2. Impacts of Human Activities on Ecological Restoration

Human activities are an important driving factor affecting ecological restoration, agricultural production and ecological projects, as human activities influence changes in ecosystem conditions. Agricultural production is closely related to precipitation, especially in rain-fed agricultural farming areas. However, the Hetao Plain, located in Zone II of the Upper Yellow River region, is a typical irrigated agricultural area in China, and the agricultural production activities in this region are strongly dependent on irrigation water sourced from the Yellow River. The FVC and NPP of vegetation in this region both showed increasing trends from 2000 to 2019, and these changes were mainly affected by human activities. At the same time, this research showed that the cascade reservoirs along Longyang Gorge and Liujia Gorge in the Upper Yellow River region were used as means to control water and sediment fluxes in the wide desert valley reaches of the Upper Yellow River region, thus improving the relationship between water and sediment conditions [48,49], reducing soil erosion and promoting ecological restoration in this region.

Since 2000, an increased number of ecological projects have been implemented in the Upper Yellow River region, and these projects correspond to the improvements observed in the ecological restoration degree. For example, in the Longdong and Longxi regions on the Loess Plateau, ecological protection and construction measures have been conducted in consideration of the important water supply provided by the Yellow River to the Gannan ecological function zone; soil and water retention measures have been enacted on the Loess Plateau; and the Three-North Shelter Forest Program, Natural Forest Protection Project, Grain for Green Project and Grazing Withdrawal Project were implemented, thus characterizing the most significant ecological restoration areas in the Upper Yellow River region. The contribution rate of human activities to the NPP was found to be 20.7% in this study, though this contribution rate exceeded 50% in some areas. In the source region of the Yellow River and in the Kubuqi Desert, although many ecological projects have been implemented, the ecological restoration degree was not ideal over the study period and was mainly restricted by the local climate conditions.

4.3. Analysis of the Causes of Water Retention Changes

The amount of water conserved in the Upper Yellow River region increased over the past 20 years, and this increase was greatest in the source region of the Yellow River; this increase was closely related to climate warming, the humidity conditions and the implementation of ecological protection projects in the Yellow River source region. Since 2000, due to the implementation of ecological protection and restoration projects in the source region of the Yellow River and in the Gannan Tibetan Autonomous Prefecture, the grassland ecosystem areas in the source region of the Yellow River have been restored, and the restoration and reconstruction of these grassland ecosystems have improved the physical and chemical properties associated with soil retention and the water retention capacity [50]. The temperature and precipitation conditions in this area showed increasing trends over the past 20 years. These climatic condition changes extended the vegetation growing season and increased the photosynthetic capacity of the local vegetation, thus promoting vegetation growth and improving the water retention capacity [51]. At the same time, the water and wetland ecosystems in the source region of the Yellow River have gradually been effectively protected. The lake areas of Zhaling Lake and Eling Lake and the wetland areas in the Yellow River source region have grown [52], and the water-storage capacity and water-conservation capacity of wetlands have improved. The increased water retention abilities observed in the tableland and gully region of the study area over the last 20 years were related to the construction of three protection systems aimed at tablelands, slope areas and ditches in this region, as these systems improved the local water retention capacities by retaining and regulating the flow of water through measures such as slope modification, terrace construction and ditch-shelter forest construction [53]. The decreased water retention capacity observed in the Huangshui watershed was the result of long-term erosion caused by hydropower, gravity, barren soils, sparse forest and grassland vegetation and frequent natural disasters in this area, which together reduced the amount of water resources available in the Yellow River region [54].

5. Conclusions

In this paper, based on the theoretical framework of assessing ecosystem status changes through remote sensing techniques, ground–UAV–satellite-derived data and model simulations were employed to generate a dataset containing assessment index parameters characterizing the Upper Yellow River region from 2000 to 2019. The main conclusions obtained from this quantitative assessment of the ecosystem status changes that have occurred in the Upper Yellow River regions over the last 20 years are described below.

(1)

From 2000 to 2018, the areas of settlement, forest and water and wetland ecosystems increased by 2661.6 km², 1246.1 km² and 1136.2 km², respectively, while the areas of grassland, farmland and desert ecosystems decreased by 2568.4 km², 1254.5 km² and 885.8 km², respectively.

(2)

The ecosystem quality and ecosystem services in the Upper Yellow River region generally remained stable or improved from 2000 to 2019, though these factors exhibited deterioration in some regions. From 2000 to 2019, the change rates of FVC and NPP were 0.25%·a⁻¹ and 3.22 gC·m⁻¹·a⁻¹, respectively. The water retention services showed an increasing trend with a change rate of 105.46 m³·hm⁻²·a⁻¹. The soil retention service remained stable and improved, and the change rate of soil retention rate was 0.78 t·hm⁻²·a⁻¹. The windbreak and sand fixation ecosystem service showed a downward trend, and the change rate of the wind erosion modulus was −0.58 t·hm⁻²·a⁻¹.

(3)

From 2000 to 2010 and from 2010 to 2019, the ecosystem quality and ecosystem services in the Upper Yellow River region remained stable and continuously improved. From 2000 to 2019, the areas in which moderate, relatively high and extremely high ecological restoration degrees were identified in the Upper Yellow River region accounted for 32.9%, 21.0% and 2.8% of the study region, respectively. The areas with relatively high and extremely high ecological restoration degrees were mainly distributed in the gully region on the Loess Plateau in Zone I and Zone II as well as on the eastern Hetao Plain in Zone II.

(4)

From 2000 to 2019, climate factors contributed 81.6% of the calculated change in NPP and 77.6% of the derived change in the soil erosion modulus. Human activities contributed 18.4% and 22.4% to the NPP and soil erosion modulus changes, respectively.

(5)

The restoration potential index of the FVC in the Upper Yellow River region was 22.7%. The difference between the actual FVC value and that derived under ideal zonal climax ecological background values was 13.9% in the forest, grassland and desert ecosystems; among these ecosystems, the difference obtained for forest ecosystems was 12.7%, and the corresponding restoration potential index was 14.4%. The difference between the real and ideal FVC values in the grassland ecosystems was 14.3%, and the corresponding restoration potential index was 23.0%.