Spatial Heterogeneity and Formation Mechanism of Eco-Environmental Quality in the Yellow River Basin

Abstract

The Yellow River Basin (YRB) serves as a crucial ecological security barrier in China, and the scientific evaluation and understanding of the formation mechanisms of ecological environment quality (EEQ) in the YRB are of great importance for implementing dynamic regional ecosystem planning and protection. Therefore, this study aims to explore the spatiotemporal evolution characteristics and formation mechanisms of the EEQ in the YRB from 2000 to 2020, based on land use remote sensing monitoring data, utilizing research methods such as the ecological environment quality index, centroid analysis, and the geographic detector. The results are as follows: From 2000 to 2020, the overall EEQ in the YRB showed a slight increasing trend, with a significant growth trend in the upstream and middle reaches, and a decreasing trend in the downstream. The spatial heterogeneity of the EEQ in the YRB was significant, with a south—high and north—low distribution pattern overall. During the study period, the spatial distribution pattern of the EEQ was relatively stable, with a notable increase in EEQ in the northwest of the YRB. The EEQ in the study area was jointly affected by natural and socio-economic factors, with precipitation, terrain fluctuation, and temperature being the dominant factors. The upstream EEQ was most affected by temperature, the EEQ of the middle reaches was most affected by precipitation, and the downstream EEQ was more vulnerable to the interference of slope and topographic relief. The influencing factors had a synergistic effect on the EEQ, and the explanatory power of socio-economic factors for a single-factor explanation of the EEQ was relatively low, but the explanatory power of the interaction between natural and socio-economic factors was significantly enhanced. These results can provide valuable insights for guiding and informing policy-making decisions related to ecological protection in the YRB.

1. Introduction

Global eco-environmental issues, such as global warming, extreme weather events, energy consumption, and loss of biodiversity have been severely threatening human survival and development [1,2,3]. Currently, China is in the midst of accelerating urbanization and industrialization. While the Chinese economy has created a “world wonder” [4], it has also given rise to a series of eco-environmental problems [5,6]. In 2006, the China Environmental Protection White Paper pointed out that the coverage rate of China’s ecologically vulnerable areas had exceeded 60% of the national territory. As an important ecological security barrier, agricultural and animal husbandry production base, and energy base in China, the Yellow River Basin (YRB) faces obvious eco-environmental constraints in its development [7]. With the rise of the YRB’s ecological protection and high-quality development as a national strategy, the optimization and improvement of the YRB’s ecological environment has become imminent. Therefore, accurately characterizing the spatial heterogeneity and formation mechanisms of the YRB’s eco-environmental quality (EEQ) has become a scientific problem that urgently needs to be solved.

EEQ refers to the regional eco-environmental quality level, which is qualitatively analyzed and evaluated for some or all ecological factors that affect social development and human activities within a certain time frame and area [8]. Scholars have constructed indicator systems from different perspectives, such as climate change, terrestrial water storage changes, and runoff changes, to monitor and evaluate the eco-environment [9,10,11,12,13]. However, the eco-environment is controlled by complex variables, and monitoring from a single domain cannot accurately characterize the regional EEQ [14,15].

In 2013, Hanqiu Xu constructed the remote sensing ecological index (RSEI) by coupling four indicators, namely the normalized difference vegetation index, wetness, land surface temperature, and normalized difference built-up index using principal component analysis. The index has been applied in EEQ evaluations in various regions [16,17,18,19]. However, there is a high positive correlation between the evaluation factors, which leads to redundancy in ecological significance, resulting in biased results [20]. Land use/land cover change (LUCC) is one of the most prominent landscapes in the land surface system and is widely regarded as one of the main causes of global eco-environment changes [21]. LUCC primarily affects the structure and function of natural elements such as surface soil, climate, hydrology, and biodiversity through major ecological processes, including energy exchange, water cycle, soil erosion and accumulation, and crop production. This ultimately leads to positive or negative changes in the eco-environment. Therefore, LUCC is widely applied because it can comprehensively reflect the degree and evolutionary characteristics of human production activities’ impact on the EEQ [22,23,24,25,26,27]. The development and application of Geographic Information System (GIS) and Remote Sensing (RS) technologies have provided diverse data and analysis methods for land use assessment, ecological vulnerability evaluation, urban planning, and other related fields [28,29,30]. The eco-environmental quality index (EQI) is a quantitative approach that characterizes the spatiotemporal evolution of the EEQ in a region by establishing the association between LUCC and EEQ, and tracking the structural, quantitative, and spatial changes in LUCC. Previous studies have primarily focused on the expert ecological valuation of first-level land use types. However, the EQI method employed in our research involves expert ecological valuation based on second-level land use cover types, enabling a more precise characterization of regional EEQ. Additionally, existing research in this field has predominantly analyzed the EEQ of a specific study area based on quantitative characteristics, spatial distribution, and spatial variations. Nevertheless, there remains a notable gap in the analysis of the driving forces behind EEQ.

River basins, as complex systems, are more sensitive to their ecological environment [31] and exhibit significant spatiotemporal heterogeneity, spatial coupling of element endowments, and spatial constraints on internal relationships. This paper begins by reviewing relevant literature on assessing the EEQ at a regional level. Based on land use remote sensing monitoring data from 2000 to 2020 in the YRB, the EQI method, which incorporates expert ecological valuation using secondary land use cover types, is employed to enhance the effectiveness and precision of the evaluation. Subsequently, research methods such as gravity center analysis and geographical detector are utilized to analyze the spatial differentiation, temporal evolution, and formation mechanisms of EEQ in the YRB (Figure 1). The objectives of this study are as follows: (1) to explore the spatiotemporal distribution of EEQ in the YRB from 2000 to 2020, (2) to investigate the spatiotemporal evolution and formation mechanisms of EEQ in the YRB, and (3) to provide theoretical and practical references for the effective protection of the ecological environment in the YRB. This research not only enables a scientific evaluation of the effectiveness of current environmental protection policies in the YRB but also holds significant importance for policy-makers in promoting the optimization of relevant environmental policies.

2. Materials and Methods

2.1. Study Area

The YRB is located between 32° N and 42° N and 96° E and 119° E, with a length of approximately 5464 km and a basin area of 795,000 km². The climate in the YRB varies from arid and semi-arid to semi-humid and humid, with an average temperature of around 7 °C and an average precipitation of around 440 mm (Figure 2). The YRB exhibits a west-to-east gradient in topography, with higher elevations in the west and lower elevations in the east. The average elevations along the Yellow River are 2138.4 m in the upstream, 396.1 m in the middle reaches, and 44.3 m in the downstream (Figure 3). The ecological environment in the YRB is fragile, and the extensive mode of production in the last century exacerbated environmental degradation, including the reduced flow of the YRB, severe water pollution, increased soil erosion, desertification, and vegetation degradation. In order to protect biodiversity, water resources, and ecosystem services, as well as to achieve sustainable development goals, China is actively advancing ecological conservation and governance efforts in the YRB. Special emphasis is placed on addressing ecological environmental issues, which includes the establishment of multiple ecological functional zones such as the Three-River-Source Water Conservation and the Biodiversity Protection Important Area. As of 2020, the total population in the YRB was about 113.68 million people, accounting for 8.6% of the total population of China, and the GDP of the basin accounted for only 8% of the national GDP, with a per capita GDP of about 90% of the national average. Although the socio-economic development in the YRB is relatively backward, it plays an essential role in maintaining China’s ecological security and promoting high-quality economic and social development, making it one of the most important growth poles in China.

2.2. Data Sources

The administrative boundary data were obtained from the National Geomatics Center of China (http://ngcc.sbsm.gov.cn/, accessed on 1 February 2023). The LUCC data were sourced from the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (http://www.resdc.cn/, accessed on 1 February 2023), with a resolution of 1000 × 1000 m. The land use types included six primary types and twenty-five secondary types [32,33]. The spatial population density and per capita GDP of the YRB in 2000, 2010, and 2020 were obtained based on the 1 km resolution spatial distribution grid dataset from the Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences [34,35]. The data sources for the factor set used in the geographical detector analysis are shown in

Table 1. Data sources and descriptions.

First Grade Indexes	Second Index	Specific Indictors	Calculation Method	Data Source
Natural factors	Topographic factors	Mean altitude (X1)	The data are based on DEM data, extracted in ArcGIS using the Zonal Statistic tool	http://www.gscloud.cn/, accessed on 1 February 2023
		Mean slope (X2)	The data are based on DEM data extracted in ArcGIS using the Slope tool
		Topographic relief (X3)	The data are based on DEM data extracted in ArcGIS using the Block Statistics tool.
	Climatic factors	Mean annual precipitation (X4)	Extracted by using the Zonal Statistic tool in ArcGIS	http://www.resdc.cn/, accessed on 1 February 2023
		Mean annual temperatures (X5)
Socio-economic factors	Population and Economics	Population density (X6)
		Per capital GDP (X7)
	Traffic location	High grade road density (X8)	The data are obtained based on railroad, highway, national road, and provincial road data in ArcGIS using the Linear Density Analysis tool	http://ngcc.sbsm.gov.cn/, accessed on 1 February 2023
		Distance from prefectural city (X9)	Calculate the distance to the nearest prefecture-level city with the Distance tool in ArcGIS

.

2.3. Methods

2.3.1. Eco-Environmental Quality Index

The EQI characterizes the overall EEQ in a region by constructing a quantitative relationship between LUCC and EEQ [36]. The EQI was computed by assigning background values to the secondary land use types within the study area (

Table 2. Land use classification and eco-environmental quality values of different land-use types.

Level 1 Type	Level 2 Type	EQI	Level 1 Type	Level 2 Type	EQI
Cropland	Paddy Field	0.30	Impervious	Towns	0.20
	Dry Land	0.25		Rural Residential Area	0.20
Woodland	Forested Land	0.95		Other Construction Land	0.15
	Shrubbery	0.65	Barren	Sand	0.01
	Thin Stocked Land	0.45		Gobi	0.01
	Other Forest Land	0.40		Salinate Field	0.05
Grassland	High-Coverage Grassland	0.75		Moorland	0.65
	Medium-Coverage Grassland	0.45		Bare Land	0.05
	Low-Coverage Grassland	0.20		Bare Rock	0.01
Water	Rivers	0.55		Other	0.01
	Lakes	0.75
	Ponds	0.55
	Glacier	0.90
	Foreshore	0.45
	Shoaly Land	0.55

).

$$EQ{I}_t=(\sum _{i=1}^{n}LU{A}_{i,t}\times E{V}_{i,t})/\sum _{i=1}^{n}LU{A}_{i,t}$$

(1)

$$C=\left[\right(EQ{I}_{t2}-EQ{I}_{t1})/EQ{I}_t]\times [1/(t2-t1\left)\right]\times 100\%$$

(2)

EQI_t, EQI_t1, and EQI_t2 represent the EQI at time t, t1, and t2, respectively. EV_i,t represents the background value of the EEQ for land use type i at time t; LUA_i,t represents the area of land use type i at time t; n represents the number of land use types in the region; and C represents the rate of change in EQI between t1 and t2.

2.3.2. Centroid for EEQ

The concept of the centroid from physics was introduced to reveal the spatial evolution characteristics of EEQ [37]. The centroid analysis of EEQ effectively describes the spatiotemporal evolution characteristics of the EEQ. By understanding the centroid distribution of the EEQ in each study period, the spatial trends of EEQ changes can be identified. If the EEQ uniformly decreases or improves in all spatial directions, the centroid remains relatively stable. However, if there is a significant decrease or improvement in a specific direction, the centroid undergoes noticeable displacement.

$$X_t=\sum _{j=1}^{n}EQ{I}_{tj}{X}_j/\sum _{j=1}^{n}EQ{I}_{tj}\hspace{1em}{Y}_t=\sum _{j=1}^{n}EQ{I}_{ij}{Y}_j/\sum _{j=1}^{n}EQ{I}_{tj}$$

(3)

In the formula, X_j and Y_j represent the geographic center coordinates of unit j, while X_t and Y_t represent the coordinates of the centroid of EEQ in the study area at time t.

2.3.3. Standard Deviation Ellipse

The standard deviation ellipse focuses on revealing the global characteristics of the spatial distribution of geographic features [38,39].

$$P=(\sqrt{\frac{\sum _1^n{x}_i^{’}}{n}},\sqrt{\frac{\sum _1^n{y}_i^{’}}{n}})$$

(4)

$$R=\mathrm{a}\mathrm{r}\mathrm{c}\mathrm{t}\mathrm{a}\mathrm{n} \left(\frac{\sum _1^n{x}_i^{’2}-\sum _1^n{y}_i^{’2}+\sqrt{{(\sum _1^n{x}_i^{’2}-\sum _1^n{y}_i^{’2})}^2+4\sum _1^n{x}_i^{’}{y}_i^{’}}}{2\sum _1^n{x}_i^{’}{y}_i^{’}}\right)\times \pi$$

(5)

$$D=\sqrt{\frac{\sum _1^n{(x_i^{’}\mathrm{c}\mathrm{o}\mathrm{s} R-y_i^{’}\mathrm{s}\mathrm{i}\mathrm{n} R)}^2}{\sum _1^n{(x_i^{’}\mathrm{s}\mathrm{i}\mathrm{n} R-y_i^{’}\mathrm{c}\mathrm{o}\mathrm{s} R)}^2}}$$

(6)

$$S=\sqrt{\frac{\sum _1^n{(x_i^{’}\mathrm{s}\mathrm{i}\mathrm{n} R-y_i^{’}\mathrm{c}\mathrm{o}\mathrm{s} R)}^2}{n}}$$

(7)

In the equation, P represents the center coordinates of the standard deviation ellipse; $x_i^{’}$ and $y_i^{’}$ represent the longitude and latitude coordinates of the ith two-dimensional scatter plot in the study area; n represents the total number of scatter plots; R represents the orientation angle of the standard deviation ellipse; D represents the directional indicator; and S represents the dispersion indicator.

2.3.4. Geographical Detector

Geographical detectors were employed to explore the influential factors of EEQ, with the following equation:

$$P_{D,H}=1-\frac{1}{N{\sigma }^2}\sum _{h=1}^L{N}_h{\sigma }_h^2$$

(8)

where P_D,H is the detection index of the EEQ’s influencing factors; N_h is the number of sample units in the subregion; N is the number of sample units in the whole region; L is the number of subregions; σ² is the variance of EEQ in the entire region; and σ_h² is the variance of EEQ in the subregion. When P_D,H = 0, it indicates that the EEQ is randomly distributed; when P_D,H = 1, it indicates that the EEQ’s spatial heterogeneity is the strongest. When two factors interact, five situations may occur: (1) P(x ∩ y) < min(P(x), P(y)), indicating that the factors x and y show a nonlinear weakening after interaction; (2) min(P(x), P(y)) < P(x ∩ y) < max(P(x), P(y)), indicating the nonlinear weakening of a single factor after the interaction of x and y; (3) P(x ∩ y) > max(P(x), P(y)), indicating the strengthening of both factors after the interaction of x and y; (4) P(x ∩ y) > P(x) + P(y), indicating the nonlinear strengthening of x and y after interaction; and (5) P(x ∩ y) = P(x) + P(y), indicating the independence between x and y [40].

The formation of EEQ is the result of complex interactions among various variables. This study explores the spatial patterns and formation mechanisms of EEQ in the YRB by considering factors from both natural elements and socio-economic aspects. Natural elements represent key environmental components, and socio-economic indicators reflect human influences. Five factors related to terrain and climate were selected as natural elements, while population, GDP, and transportation location were chosen as indicators of socio-economic activities. Specific indicators can be found in Table 1.

3. Results

3.1. EQI Trend over Time

According to Formula (1), EQI values were calculated for different years (

Table 3. The eco-environmental quality index in 2000, 2010, and 2020 in the Yellow River Basin and its stream segments.

Year	The Eco-Environmental Quality Index
	Yellow River Basin	Upstream	Middle Reaches	Downstream
2000	0.384	0.381	0.397	0.298
2010	0.386	0.380	0.403	0.291
2020	0.387	0.385	0.398	0.293

). The EQI in the YRB experienced a slight increase from 2000 to 2020. The upstream EQI exhibited a changing trend of “decrease—rising”, dropping from 0.381 in 2000 to 0.380 in 2010, and then increasing to 0.385 in 2020. From 2000 to 2010, there was a significant increase in Other Construction Land (215.57%) and in Towns (71.82%), while High-Coverage Grassland decreased (−8.34%), leading to a decline in the EQI in the upstream. From 2010 to 2020, there was an increase in Other Construction Land (118.73%), and Dry Land resources were occupied (5.81%). However, there was an increase in Ponds (21.19%), Rivers (10.93%), High-Coverage Grassland (9.35%), and Forested Land (7.89%), resulting in an increase in EQI in the upstream.

The EQI of the middle reaches of the YRB presented a changing trend of “rising—decrease”, increasing from 0.397 in 2000 to 0.403 in 2010, and then decreasing to 0.398 in 2020. During the period of 2000–2010, there was a significant increase in the demand for Towns (80.93%), Rural Residential Areas (25.72%), and Other Construction Land (241.50%), leading to the occupation of a large amount of Dry Land (5.24%). However, due to the combined effect of the increase in High-Coverage Grassland (13.49%) and Other Forest Land (106.43%), the EQI in the middle reaches showed an upward trend. From 2010 to 2020, Other Construction Land continued to increase at a high rate (101.79%), while the areas of High-Coverage Grassland (−6.92%) and Forested Land (−6.19%) decreased, leading to a decrease in the EQI in the middle reaches. These findings indicate that the long-term implementation of ecological projects such as returning farmland to forests and grasslands, and of soil and water conservation measures in the upper and middle reaches of the YRB, have effectively improved the EEQ.

During the study period, the EQI of the downstream region of the YRB exhibited a trend of “decrease—rising”. The EQI decreased from 0.298 in 2000 to 0.291 in 2010, and then increased to 0.293 in 2020. Between 2000 and 2010, there was a significant increase in Towns (165.56%) and Other Construction Land (107.83%), while High-Coverage Grassland (48.86%), Medium-Coverage Grassland (43.45%), and Other Forest Land (44.97%) all showed a decreasing trend. From 2010 to 2020, the growth rate of Towns (15.96%) and Other Construction Land (2.57%) slowed down, while Other Forest Land (10.58%) and Shrubbery (6.52%) decreased. However, with the increase in Ponds (30.67%), Lakes (1.55%), and Rivers (5.71%), the EQI of the downstream region of the YRB showed a slight upward trend (Figure 4,

Table 4. Changes in land use types in the Yellow River Basin from 2000 to 2020.

Level 1 Type	Level 2 Type	Yellow River Basin		Upstream		Middle Reaches		Downstream
		00–10	10–20	00–10	10–20	00–10	10–20	00–10	10–20
Cropland	Paddy Field	−12.03	−3.62	−2.46	−3.08	−39.66	−3.98	−22.28	−6.24
	Dry Land	−2.51	−2.94	3.66	−5.81	−5.24	−1.72	−0.36	−2.93
Woodland	Forested Land	1.35	−2.96	1.06	7.89	1.49	−6.19	−0.69	0.92
	Shrubbery	2.61	−0.62	1.74	0.13	3.46	−1.22	−10.68	−6.52
	Thin Stocked Land	−1.10	11.09	1.46	1.72	−1.69	14.70	−13.61	−0.79
	Other Forest Land	67.97	−1.37	36.56	−6.35	106.43	1.08	−44.97	−10.58
Grassland	High-Coverage Grassland	−3.56	4.78	−8.34	9.35	13.49	−6.92	−48.86	0.53
	Medium-Coverage Grassland	1.17	−1.49	2.16	−3.95	−0.01	3.14	−43.45	4.05
	Low-Coverage Grassland	2.92	0.15	5.19	1.20	−1.49	−2.34	−53.39	−3.40
Water	Rivers	18.91	7.56	25.50	10.93	14.22	4.81	15.10	5.71
	Lakes	−3.34	4.71	−2.69	4.81	−15.85	7.25	−6.52	1.55
	Ponds	39.05	22.37	24.43	21.19	47.32	14.39	57.17	30.67
	Glacier	−4.28	0.00	−4.30	0.00	0.00	0.00	0.00	0.00
	Foreshore	480.00	−1.72	0.00	0.00	0.00	0.00	460.00	−1.79
	Shoaly Land	−13.25	−0.64	−5.50	−2.57	−26.85	4.45	−31.76	0.99
Impervious	Towns	97.40	18.77	71.82	31.97	80.93	14.72	165.56	15.96
	Rural Residential Area	11.17	7.40	−1.91	14.95	25.72	4.51	2.68	3.68
	Other Construction Land	198.46	90.32	215.57	118.73	241.50	101.79	107.83	2.57
Barren	Sand	−4.62	−0.55	−4.40	−1.55	−4.74	3.48	−98.28	0.00
	Gobi	−12.76	−7.50	−12.82	−7.54	0.00	0.00	0.00	0.00
	Salinate field	−34.44	31.35	−33.30	34.15	−9.87	13.68	−95.69	−18.18
	Moorland	12.29	−0.67	12.68	−0.90	−12.28	17.33	35.09	−10.39
	Bare Land	−30.87	8.61	−33.43	−2.78	7.80	128.95	600.00	−28.57
	Bare Rock	16.59	−14.83	16.73	−14.93	3.41	1.10	−3.70	−19.23
	Other	−87.16	−9.64	−87.17	−9.65	−80.00	0.00	0.00	0.00

).

3.2. Spatial Evolution of the EQI of YRB

Using ArcGIS, spatial visualization analysis was conducted to examine the EEQ in the study area for the years 2000, 2010, and 2020 (Figure 5). The EQI of the YRB exhibits significant spatial heterogeneity, showing a pattern of higher values in the south and lower values in the north. During the study period, the spatial distribution pattern of the EQI was stable. In the upstream of the YRB, the overall EQI was higher in the southwest than in the northeast. The high-value areas of the EEQ were primarily distributed in the southern regions of Sichuan and the southwestern part of Gansu, while the low-value areas were mainly found in the central regions of Inner Mongolia, the northern part of Gansu, and the southwestern part of Qinghai, and Ningxia. In the middle reaches, the high-value areas of EEQ were predominantly located in the southern regions of Shaanxi, the central and eastern parts of Shanxi, and the southwestern part of Henan, while the low-value areas were mainly distributed in the northern regions of Shaanxi and the western parts of Shanxi. In the downstream of the YRB, the overall EQI level was relatively low, and the level was slightly higher in the northeast than in the southwest.

To identify the spatial pattern evolution of the EEQ in the YRB, changes in the EQI were calculated for the periods of 2000–2010, 2010–2020, and 2000–2020. Using the natural fracture point method, the study area was divided into regions of significant increase, slight increase, stability, slight decrease, and significant decrease (Figure 6).

The spatial distribution pattern of the YRB’s EEQ changes underwent a significant transition process. From 2000 to 2010, the regions in the upstream of the YRB that experienced a significant improvement in EEQ were mainly located in the central part of Inner Mongolia. Areas with a slight improvement were primarily found in Ningxia and the western part of Qinghai, while regions with a slight decrease were mainly observed in Sichuan, as well as the northern and southern parts of Inner Mongolia. Regions with a significant decrease were primarily located in the northern part of Inner Mongolia. In the middle reaches of the YRB, areas with a slight improvement in EEQ were mainly found in the northern part of Shaanxi and the eastern part of Gansu, while regions with a slight decrease were primarily located in the eastern part of Shanxi and the southwestern part of Henan. The downstream area of the YRB generally showed a slight decreasing trend. From 2010 to 2020, regions with a slight improvement in EEQ were mainly located in the northern and southern parts of Inner Mongolia, while regions with a slight decrease were primarily found in the western part of Inner Mongolia, and other regions showed minor changes. In the middle reaches of the YRB, Shaanxi Province exhibited a slight increasing trend, while the central part of Shanxi and the southwestern part of Henan showed a significant decreasing trend. The changes in EEQ in the downstream area of the YRB were not significant.

Overall, from 2000 to 2020, the EEQ in the upstream area of the YRB showed a slight increasing trend, with significant improvements mainly observed in the central part of Inner Mongolia and the central part of Qinghai. Regions with a slight decrease were primarily found in Sichuan. In the middle reaches, areas with a slight improvement in EEQ were mainly located in the northern part of Shaanxi and the eastern part of Gansu, while the southwestern part of Henan and the central and southern parts of Shanxi exhibited a decreasing trend. In the downstream area of the YRB, regions with a slight improvement in EEQ were primarily located in the northeastern corner, while other regions showed a decreasing trend.

3.3. Movement Trace of EEQ Centroid in the YRB

This study analyzed the spatial distribution changes in the EQI in the YRB in 2000, 2010, and 2020 using the centroid–standard deviation ellipse method. From 2000 to 2020, the long axis of the standard deviation ellipse in the YRB was always larger than the short axis, showing a clear “southwest-northeast” distribution pattern. The long axis of the ellipse increased slightly while the short axis decreased in 2020 compared to 2000, indicating an increase in directionality. Based on the shape index, the flatness of the standard deviation ellipse in the upstream, middle reaches, and downstream of the YRB were ranked as downstream > upstream > middle reaches, indicating that the spatial distribution of the EQI in the middle reaches of the YRB was the most uniform. From 2000 to 2020, the flatness in the upstream and middle reaches showed an increasing trend, while that in the downstream showed a decreasing trend, indicating a decrease in equilibrium in the spatial distribution of the EQI in the upstream and the middle reaches, and an increase in the downstream.

Regarding the spatial distribution of the centroid, from 2000 to 2020, the centroid of the EQI in the YRB was located in the middle reaches, indicating that the EEQ in the middle reaches was higher. Based on the trajectory of centroid movement (Figure 7,

Table 5. The standard deviation elliptic parameters of eco-environmental quality in the Yellow River Basin from 2000 to 2020.

Study Area	Year	Standard Deviation Elliptic Parameters
		Center of Inertia (x, y)	Major Axis	Minor Axis	Oblateness	Distance between Weighted Centers (km)
Yellow River Basin	2000	(106.864, 36.487)	593,004	297,857	0.4977
	2010	(106.825, 36.472)	595,211	294,875	0.5046	3.780
	2020	(106.815, 36.501)	594,903	296304	0.5019	3.359
Upstream	2000	(103.851, 36.676)	580,628	181,930	0.6867
	2010	(103.769, 36.639)	579,593	182,198	0.6856	8.198
	2020	(103.828, 36.684)	582,947	182,124	0.6876	7.178
Middle reaches	2000	(109.810, 36.240)	302,649	216,672	0.2841
	2010	(109.796, 36.254)	303,483	215,964	0.2884	2.049
	2020	(109.784, 36.260)	304,087	214,392	0.2950	1.251
Downstream	2000	(116.315, 36.060)	231,294	49,846	0.7845
	2010	(116.296, 36.045)	228,428	49,526	0.7832	2.340
	2020	(116.325, 36.064)	230,154	49,997	0.7828	3.292

), from 2000 to 2010, the centroid moved towards the southwest direction with a distance of 3.78 km, indicating a significant improvement in EEQ in the southwest of the YRB. From 2010 to 2020, the centroid moved towards the northwest direction with a distance of 3.36 km. Overall, the centroid of YRB EQI moved towards the northwest direction from 2000 to 2020, showing a westward shift in the east–west direction and a southward shift followed by a northward shift in the north–south direction. This indicates that the EEQ in the northwest of the YRB improved significantly during this period, leading to the centroid moving towards the northwest. The southeast of the YRB is a population and economic agglomeration area, where a large amount of construction land has led to EEQ deterioration. In contrast, unused land with a low EEQ background index is widely distributed in the northwest, and the development of the western region has caused the unused land to be converted to other land use types with a high EEQ background index, thereby improving the regional EEQ.

The EQI centroid in the upstream of the YRB was located in the northern part of Gansu Province from 2000 to 2020. From 2000 to 2010, the centroid moved towards the southwest direction with a distance of 8.20 km, and from 2010 to 2020, it moved towards the northeast direction with a distance of 7.18 km, showing an overall movement towards the northwest direction. The centroid of the EQI in the middle reaches of the YRB moved towards the northwest direction in 20 years. The EQI centroid in the downstream was located in the middle of Shandong Province, moving towards the southwest direction from 2000 to 2010 and towards the northeast direction from 2010 to 2020. Overall, the centroid of the downstream moved towards the northeast direction.

3.4. Influencing Factors of EEQ in the YRB

The main influencing factors of the YRB were identified using the factor detection model of geographic detectors. In addition, interactive detection was used to identify the degree of explanation of the interaction effect of the factors. Using risk detection, the detection factors in 2000, 2010, and 2020 were all significant at the 0.05 level, indicating that the selected influencing factors can explain the spatiotemporal differentiation pattern of the YRB’s EEQ relatively well.

Based on the factor detection results (

Table 6. Contribution rates of influencing factors from 2000 to 2020.

Study Area	Year	Influencing Factors
		X1	X2	X3	X4	X5	X6	X7	X8	X9
Yellow River Basin	2000	0.172	0.329	0.356	0.535	0.354	0.104	0.030	0.106	0.033
	2010	0.188	0.358	0.374	0.605	0.353	0.114	0.047	0.107	0.030
	2020	0.199	0.332	0.350	0.453	0.347	0.110	0.093	0.105	0.031
Upstream	2000	0.394	0.186	0.208	0.516	0.709	0.020	0.044	0.091	0.016
	2010	0.460	0.218	0.233	0.568	0.735	0.024	0.076	0.156	0.007
	2020	0.411	0.193	0.205	0.515	0.674	0.025	0.144	0.160	0.012
Middle reaches	2000	0.122	0.585	0.641	0.952	0.111	0.210	0.067	0.336	0.062
	2010	0.136	0.575	0.618	0.939	0.146	0.219	0.049	0.146	0.071
	2020	0.155	0.566	0.619	0.599	0.146	0.194	0.081	0.088	0.060
Downstream	2000	0.042	0.893	0.603	0.727	0.525	0.420	0.095	0.051	0.014
	2010	0.047	0.937	0.640	0.686	0.597	0.390	0.022	0.065	0.024
	2020	0.045	0.877	0.594	0.247	0.546	0.460	0.165	0.096	0.019

), the influencing factors of EEQ in the upstream, middle reaches, and downstream of the YRB exhibited significant differences. Overall, natural factors such as precipitation (X₄), topographic relief (X₃), and slope (X₂) had a more significant impact on the YRB’s EEQ than socio-economic factors such as population density (X₆), GDP (X₇), and road density (X₈). In the downstream of the YRB, the topographic relief had a stronger influence on the EEQ than in the upstream and the middle reaches. This is mainly due to the downstream being an area with a high concentration of population and economy, where human activities are affected by slope and topographic relief, thus impacting the EEQ. Precipitation had the greatest impact on the EEQ in the middle reaches, followed by the downstream and the upstream, and a decreasing trend was evident in the middle reaches and the downstream during the study period. Temperature (X₅) had the smallest impact on the EEQ in the middle reaches and the largest impact on the upstream. Moreover, in terms of socio-economic factors, population density and GDP factors had a stronger impact on the EEQ in the downstream of the YRB than in the middle reaches and the upstream. The impact of GDP on EEQ continued to increase, especially in the upstream and the downstream. Road density had an increasing impact on EEQ in the upstream and the downstream of the YRB and a decreasing trend in the middle reaches.

Using geographical detector interaction analysis, it was found that the influencing factors on EEQ from 2000 to 2020 do not act alone but exhibit a synergistic effect, primarily through two modes of interaction: nonlinear enhancement and bivariate enhancement (

Table 7. The main interaction factors and associated changes.

Study Area	2000			2010			2020
	Interaction Factors	Value	Type	Interaction Factors	Value	Type	Interaction Factors	Value	Type
Yellow River Basin	X4 ∩ X5	0.722	Nonlinear enhancement	X4 ∩ X5	0.720	Nonlinear enhancement	X4 ∩ X5	0.561	Double enhancement
	X4 ∩ X6	0.601	Nonlinear enhancement	X1 ∩ X4	0.604	Nonlinear enhancement	X3 ∩ X4	0.480	Double enhancement
	X1 ∩ X4	0.595	Nonlinear enhancement	X4 ∩ X6	0.584	Nonlinear enhancement	X1 ∩ X4	0.476	Double enhancement
	X4 ∩ X8	0.557	Nonlinear enhancement	X3 ∩ X4	0.572	Double enhancement	X4 ∩ X6	0.474	Nonlinear enhancement
	X4 ∩ X7	0.541	Nonlinear enhancement	X4 ∩ X7	0.568	Nonlinear enhancement	X2 ∩ X4	0.458	Double enhancement
Upstream	X4 ∩ X5	0.635	Double enhancement	X4 ∩ X5	0.656	Double enhancement	X4 ∩ X5	0.584	Double enhancement
	X5 ∩ X9	0.581	Nonlinear enhancement	X5 ∩ X9	0.605	Nonlinear enhancement	X5 ∩ X9	0.551	Nonlinear enhancement
	X2 ∩ X5	0.568	Double enhancement	X2 ∩ X5	0.596	Double enhancement	X2 ∩ X5	0.548	Double enhancement
	X5 ∩ X8	0.568	Double enhancement	X5 ∩ X8	0.591	Double enhancement	X5 ∩ X7	0.539	Double enhancement
	X5 ∩ X6	0.565	Nonlinear enhancement	X5 ∩ X6	0.587	Nonlinear enhancement	X3 ∩ X5	0.536	Double enhancement
Middle reaches	X4 ∩ X6	0.932	Nonlinear enhancement	X4 ∩ X6	0.857	Double enhancement	X2 ∩ X4	0.616	Double enhancement
	X4 ∩ X5	0.868	Nonlinear enhancement	X2 ∩ X4	0.829	Double enhancement	X3 ∩ X4	0.612	Double enhancement
	X2 ∩ X4	0.853	Double enhancement	X3 ∩ X4	0.816	Double enhancement	X4 ∩ X6	0.597	Nonlinear enhancement
	X1 ∩ X4	0.852	Nonlinear enhancement	X4 ∩ X9	0.791	Nonlinear enhancement	X2 ∩ X3	0.576	Double enhancement
	X3 ∩ X4	0.845	Double enhancement	X4 ∩ X5	0.783	Double enhancement	X1 ∩ X3	0.575	Double enhancement
Downstream	X2 ∩ X6	0.822	Double enhancement	X2 ∩ X6	0.841	Double enhancement	X2 ∩ X6	0.861	Double enhancement
	X4 ∩ X6	0.778	Double enhancement	X2 ∩ X4	0.809	Double enhancement	X2 ∩ X7	0.770	Nonlinear enhancement
	X2 ∩ X4	0.762	Double enhancement	X2 ∩ X5	0.804	Double enhancement	X2 ∩ X5	0.743	Double enhancement
	X3 ∩ X6	0.759	Nonlinear enhancement	X4 ∩ X6	0.771	Double enhancement	X3 ∩ X6	0.737	Double enhancement
	X2 ∩ X5	0.751	Double enhancement	X2 ∩ X3	0.768	Double enhancement	X2 ∩ X4	0.736	Double enhancement

only the top four factor interaction values are listed). The YRB has a vast area with complex terrain and diverse climate types, forming a unique land use pattern and economic development pattern. The synergistic effect of precipitation and temperature factors had the greatest impact on the EEQ, followed by the combinations of precipitation and population density, precipitation and altitude, and precipitation and road density. Significant enhancement effects were also observed among the factors in each flow section, where natural and socio-economic factors jointly affected the spatial pattern of the regional EEQ. In the upstream of the YRB, the factor combination that had the greatest impact on the EEQ was temperature and precipitation, followed by temperature and distance to the city, temperature and slope, temperature and road density, temperature and population density, and temperature and per capita GDP. In the middle reaches, the synergistic effect of precipitation and population density had the greatest impact on the EEQ, followed by precipitation and temperature, precipitation and slope, precipitation and topographic relief, and precipitation and altitude. In the downstream, the synergistic effect of slope and population density had the greatest impact on the EEQ, followed by the combinations of precipitation and population density, slope and precipitation, and slope and GDP.

4. Discussion

The ecological degradation and sustainable development of the YRB have received extensive attention from scholars. In this study, the EQI and centroid analysis were used to analyze the spatiotemporal differentiation and evolution trends of the EEQ in the YRB, and geographic detectors were used to explore their influencing factors.

EQI evaluations revealed that the EEQ in the upstream and middle reaches of the YRB was higher than that in the downstream during 2000–2020. Further trend analysis showed that the EEQ in the YRB generally maintained a slightly increasing trend. The EEQ in the upstream and middle reaches of the YRB showed an increasing trend, while that in the downstream showed a decreasing trend, which is consistent with Cui’s research results [41]. The significant improvement in EEQ in the upstream and middle reaches is mainly attributed to China’s policies such as returning farmland to forest (grass), and forest protection [42,43]. However, Yin suggested [44] that the continuous expansion of vegetation due to the introduction of alien species and high-density planting may have a negative impact on EEQ. Therefore, to ensure the sustainable and high-quality development of the YRB, a balance should be struck between food supply and appropriate vegetation cover. The downstream area of the YRB has a high level of urbanization, dense population, and intensive industries, posing substantial threats to the downstream EEQ. Therefore, in the upstream and middle reaches of the YRB, it is imperative to continue intensifying efforts in vegetation restoration and ecosystem protection. In the downstream region, the focus should be on promoting green development by strengthening urban planning and land use management. It is crucial to prevent extensive occupation and degradation of forest and grassland areas in order to safeguard their ecological integrity.

In addition, there is significant spatial heterogeneity in EEQ across the YRB. The overall EEQ in the southwest of the upstream was higher than that in the northeast, and the overall EEQ in the southern part of the middle reaches was higher than that in the central and northern parts. The overall EEQ level in the downstream region was lower. Zhang’s research found that the ecologically vulnerable areas of the YRB are mainly concentrated in the northern upstream region, which is consistent with the results of this study [45]. This is mainly because the southeast of the YRB has a semi-humid climate and is suitable for vegetation growth, while the northwest is located in China’s arid and semi-arid regions, where the climate is harsh and deserts are widespread, making it unsuitable for vegetation growth [46,47]. During the study period, the center of gravity of the YRB’s EEQ moved towards the northwest, indicating that under a series of ecological protection and high-quality development strategies implemented in China, the EEQ in the northwest region of the YRB has been optimized. Hence, it is imperative for the government to further enhance efforts in ecological restoration and protection in the northwest region. This entails strengthening policies related to converting marginal farmland back to forests and grasslands, increasing vegetation coverage, improving soil quality, and reducing soil erosion. These measures will facilitate the restoration of the regional ecosystem and the improvement of the ecological environment.

EEQ is a dynamic process that undergoes continuous changes, influenced by both human activities and climate change. Overall, natural factors exert a more significant influence on EEQ than socio-economic factors. Precipitation, topographic relief, and temperature are the main driving factors of the EEQ in the YRB. Among them, temperature has the most significant impact on the upper reaches of the YRB, mainly because the Loess Plateau region in the upper reaches is situated in an arid and semi-arid area and is subject to significant wind and water erosion. Rising temperatures exacerbate aridity and soil desertification, which in turn hinders EEQ improvement [48,49]. In the middle reaches of the Yellow River, precipitation has the greatest impact, whereas in the lower reaches, slope and topographic relief are more susceptible to interference. Compared with temperature, precipitation has a more obvious effect on the EEQ in the middle reaches of the Yellow River, where abundant light resources are available, indicating that water availability is more crucial for EEQ spatial distribution. In the lower reaches, where climate conditions are more conducive to plant growth, slope and topographic relief become the key factors affecting EEQ, as they are more susceptible to human activities. The government should formulate appropriate governance measures based on the dominant factors in each region.

In addition, while socio-economic factors have weak explanatory power for the EEQ in the study area, the interaction between socio-economic and natural factors significantly enhances the explanatory power, indicating that socio-economic factors are important drivers of EEQ. This finding validates the viewpoint proposed by TIAN [50] that “the greening of more than 94% of the YRB since the 21st century is a synergistic effect of climate and human activities” and further underscores the significant impact of human activities on ecological environmental changes [51]. Therefore, it is essential for the government to incorporate the interactive effects of natural and socio-economic factors into the policy-making process and develop comprehensive ecological conservation policies.

5. Conclusions

Based on the LUCC data from 2000 to 2020, the EQI and centroid analysis were used to measure the spatiotemporal distribution pattern and evolution characteristics of EEQ in the YRB. The geographical detector model was used to explore the formation mechanism of the YRB’s EEQ by selecting natural environmental and socio-economic factors. The results of this study are summarized as follows:

(1)

From 2000 to 2020, the YRB’s EEQ showed a slight upward trend overall. The EEQ in the upper and middle reaches of the YRB was higher and increasing, while the EEQ in the lower reaches was lower and decreasing. With regard to land use change, China has long been implementing ecological engineering projects such as returning farmland to forests and grasslands and soil erosion control in the upper and middle reaches, which promoted the improvement of the EEQ in these regions. In contrast, due to recent socio-economic development and urban expansion, a large amount of forest and grassland in the downstream of the YRB has been occupied, resulting in a deterioration of EEQ. Therefore, it is imperative to further intensify efforts for vegetation restoration and ecosystem protection in the upstream and middle reaches. In the downstream area, emphasis should be placed on green development, avoiding extensive occupation and degradation of forests and grasslands, and guiding urban development towards sustainability.

(2)

The EEQ of the YRB exhibits a spatial distribution pattern with higher values in the south and lower values in the north. Specifically, the EEQ in the southwest of the YRB upstream is stronger than that in the northeast, and the EEQ in the southern part of the middle reaches is stronger than that in the central and northern parts. However, the EEQ in the downstream area is relatively low. The centroid of the standard deviation ellipse is located in the middle reaches, indicating a higher overall level of EEQ in this region. Throughout the study period, the centroid of the EEQ shifted towards the northwest, suggesting significant improvements in EEQ in the northwestern part of the YRB. This improvement can be attributed to the widespread availability of unused land in the northwest, and the development efforts in the western regions have contributed to the enhancement of the regional EEQ. Therefore, it is essential for the government to continue emphasizing ecological restoration and protection in the northwest region, aiming to facilitate the recovery of the local ecosystem and the improvement of the overall EEQ.

(3)

The natural factors had a more significant impact on the YRB’s EEQ than the socio-economic factors. Precipitation, topographic relief, and temperature were the dominant factors affecting the YRB’s EEQ. The upstream was most affected by temperature, the middle reaches were most affected by precipitation, and the downstream was more vulnerable to the interference of slope and topographic relief. Therefore, it is important for the government to develop corresponding management measures based on the dominant factors in each region. Regarding socio-economic factors, population density and GDP factors had a stronger impact on the EEQ in the downstream of the YRB than in the middle reaches and upstream. The impact of GDP on EEQ continued to increase, especially in the upstream and downstream. Road density had an increasing impact on EEQ in the upstream and downstream of the YRB and a decreasing trend in the middle reaches. Local governments should formulate sustainable population and urban planning measures to avoid excessive population concentration and resource consumption. The influencing factors on EEQ exhibit a synergistic effect; while socio-economic factors alone have weak explanatory power for the EEQ in the study area, the interaction between socio-economic and natural factors significantly enhances the explanatory power.

This study is subject to certain limitations. The spatiotemporal variations of the YRB’s EEQ exhibit multiscale characteristics, and the driving factors behind these variations are complex. The existing evaluations have primarily focused on the macroscopic analysis of the spatiotemporal trends in the YRB’s EEQ, neglecting the limiting factors imposed by the natural conditions of different regions, such as nature reserves and areas with unique topography that are designated as non-developable regions. Future research endeavors will address these limitations in two ways. First, it will delve into a comprehensive understanding of the YRB’s EEQ from a multiscale perspective. Second, it will adopt a multidimensional and integrated approach to examine the mechanisms governing the formation of EEQ in different representative regions. This approach aims to enhance the practicality and reference value of the research findings.