Spatio-Temporal Evolution, Factors, and Enhancement Paths of Ecological Civilization Construction Effectiveness: Empirical Evidence Based on 48 Cities in the Yellow River Basin of China

Abstract

Climate change, resource scarcity, and ecological degradation have become critical bottlenecks constraining socio-economic development. Basin cities serve as key nodes in China’s ecological security pattern, playing indispensable roles in ecological civilization construction. This study established an evaluation index system spanning five dimensions to assess the effectiveness of ecological civilization construction. This study employs the entropy-weighted Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) and Back-Propagation (BP) neural network methods to evaluate the level of ecological civilization construction in the Yellow River Basin from 2010 to 2022, to analyze its indicator weights, and to explore the spatio-temporal evolution characteristics of each city. The results demonstrate the following: (1) Although the ecological civilization construction level of cities in the Yellow River Basin shows a steady improvement, significant regional development disparities persist. (2) The upper reaches are primarily constrained by ecological fragility and economic underdevelopment. The middle reaches exhibit significant internal divergence, with provincial capitals leading yet demonstrating limited spillover effects on neighboring areas. The lower reaches face intense anthropogenic pressures, necessitating greater economic–ecological coordination. (3) Among the dimensions considered, Territorial Space and Eco-environmental Protection emerged as the two most influential dimensions contributing to performance differences. According to the ecological civilization construction performance and changing characteristics of the 48 cities, this study proposes differentiated optimization measures and coordinated development pathways to advance the implementation of the national strategy for ecological protection and high-quality development in the Yellow River Basin.

1. Introduction

As climate change, resource scarcity, ecological degradation, and other eco-environmental problems intensify [1], they have become significant bottlenecks constraining socio-economic development [2,3]. Against this backdrop, ecological civilization, as an advanced form of civilization following industrial civilization, aims to achieve harmonious coexistence between humans and nature, becoming an inevitable requirement for sustainable development and an important direction for human societal progress [4]. The ecological civilization construction is a crucial task for China’s sustainable development. It constitutes a key component of the “Five-sphere Integrated Plan”, encompassing economic, political, cultural, social, and ecological civilization construction, within the overall architecture of socialism with Chinese characteristics, permeating all aspects of China’s socialist modernization drive. Its emergence stems from a critical reflection on traditional development models, shifting from purely economic growth-oriented aims to a new development philosophy centered on the harmonious coexistence between humanity and nature [5]. Its core concept aligns with the principle of strong sustainability, which maintains that natural resources cannot be fully substituted by other forms of capital. It asserts that economic development must respect ecological limits to achieve sustainable progress in the economic, social, and environmental spheres [6,7]. In recent years, ecological civilization construction has become a policy orientation in multiple countries, and its effectiveness evaluation has gradually become a core component in related research [8]. Through scientific and objective evaluation, the actual progress of construction can be quantified and reflected, providing critical references for decision-making. It is not only an important tool for measuring the progress of environmental protection and sustainable development but also a key means to enhance the effectiveness of ecological governance [9]. Therefore, systematically evaluating the effectiveness of ecological civilization construction is of great significance for global ecological civilization building and sustainable development.

To explore the path of ecological civilization construction, in December 2013, China issued the Notice on Printing and Distributing the Construction Plan for National Ecological Civilization Pilot Demonstration Zones (Trial) [10], planning to select representative regions to carry out the construction of national ecological civilization pilot demonstration zones. In 2015, the Guiding Opinions on Accelerating the Construction of Ecological Civilization was promulgated [11]. This document emphasized the need to strengthen ecological protection work in key river basins, improve the ecological protection compensation mechanism, and establish a fair ecological environment assessment system. As an important strategic area for China’s ecological civilization construction, the Yellow River Basin plays a crucial role in social economy, culture, and ecological security [12]. However, the Yellow River Basin is characterized by an inherently fragile eco-environmental baseline [13]. Meanwhile, rapid socio-economic development has exacerbated the degradation of the basin’s ecosystems and increased socio-ecological risks [14,15], with issues like soil erosion and desertification continuously threatening food security and resource sustainability [16,17]. Due to the basin’s large regional span, the significant geographical, climatic, and economic differences between its upper, middle, and lower reaches present challenges of complex spatio-temporal distribution disparities and uneven resource allocation [18,19] (for a detailed introduction to the Yellow River Basin, please refer to Section 3.1). Against this backdrop, scientifically evaluating the effectiveness of ecological civilization construction in the Yellow River Basin not only concerns regional ecological security, residents’ well-being, and coordinated development but also bears the significant responsibility of advancing the implementation of major national strategies and economic transformation, holding special and urgent significance.

The remaining part of this study is organized as follows: Section 2 reviews the relevant research on the assessment of the level of ecological civilization construction. Section 3 describes the methods and data of this study, including the index system for assessing the level of ecological civilization construction, the assessment methods, data sources, and data preprocessing methods. Section 4 analyzes the empirical results and the spatio-temporal evolution of the level of ecological civilization construction in the Yellow River Basin. Section 5 discusses and compares the results of this study with those of others and analyzes the shortcomings and prospects of this study. Section 6 summarizes this study.

2. Literature Review

In recent years, research on the measurement and evaluation of ecological civilization has yielded substantial results [20]. Regarding the construction of evaluation indicators, there are two main approaches. One is applying theoretical models to build indicator systems. Based on the understanding that ecosystem health is affected by socio-economic activities, Sun et al. used the Pressure-State-Response (PSR) model to reveal the relationships between ecology, society, and economy [21]. This model provides a comprehensive framework for assessing systems, helping to identify key driving factors [22]. Chen et al. further expanded it into the Driving Force-Pressure-State-Impact-Response (DPSIR) model to better explain the dynamic feedback mechanisms between ecological systems and socio-economic systems [23]. Additionally, believing that governments bear significant responsibility in building ecological civilization cities, Meng et al. specifically considered the government’s impact on ecological civilization construction and built an indicator system covering five aspects: resource utilization, ecological culture, ecological development, policy, and economy [24]. From the perspective of respecting nature, conforming to nature, and protecting nature, Zhang et al. constructed an indicator system encompassing five major areas: water, land, atmosphere, energy, and environment [25]. The other approach involves screening key indicators by reviewing the relevant literature. Wang et al., when determining their indicator system, employed a literature statistical method, compiling the literature on ecological civilization indicator systems, and comprehensively considered indicator selection based on the balance and linkage between the eco-environment and the socio-economy [26]. Taking Guiyang City in China as the research object, Peng et al. identified a series of factors with significant influence on ecological civilization construction and established an indicator system combined with local conditions [27]. Extensively referencing existing indicator documents published internationally, Ye et al. established an indicator system comprising five dimensions: society, economy, institutions, culture, and ecology [28].

In terms of the spatial scale of evaluation, existing studies cover a broad spectrum ranging from macro-provincial levels to micro-county levels and even specific functional zones. The provincial level is a common evaluation scale, facilitating comparative analysis and policy assessment across large regions. For example, Duan and Wang constructed a provincial evaluation system focusing on three aspects: resource reserves, resource utilization efficiency, and environmental governance investment [29]. Guo et al. selected 46 indicators spanning economic development, pollution emissions, environmental quality, health status, and response measures to evaluate the ecological civilization construction level at China’s provincial level [30]. Dong et al. developed a provincial ecological civilization evaluation system comprising 37 indicators based on China’s officially released Green Development Indicator System [31]. Studies at the city/county level focus on more specific urban or county units, offering finer-grained insights into local practical differences and implementation outcomes. Zhang et al. established indicator systems covering green environment, green production, green living, and green infrastructure to evaluate the ecological civilization levels of cities in China, respectively [32,33]. Deng and Hu examined the coupling coordination relationship between tourism development and ecological civilization construction across four types of resource-based cities: developing, mature, declining, and regenerative [34]. Tian et al. focused on urban ecological security and water ecological civilization, evaluating the construction levels of three urban agglomerations in the Yangtze River Economic Belt [35]. Additionally, some studies target smaller scales, such as counties or specific functional zones. Zhang et al. assessed the ecological civilization level of China’s island counties using the PSR model [36]. Liu et al. investigated factors influencing sustainable development in the Beijing Economic-Technological Development Area, an industrial park [37].

Although research on the measurement of ecological civilization construction has made significant progress, notable deficiencies remain. (1) Although existing research has provided rich ideas and methods for the assessment of ecological civilization construction levels in river basin areas, the developed indicator systems still have deficiencies. Research on the assessment of ecological civilization construction levels in river basin areas with special characteristics still needs to be strengthened. (2) Currently, research on river basin areas lacks in-depth spatio-temporal evolution analysis and collaborative path research. Most existing studies remain at the provincial level and fail to fully analyze the development differences among different cities, thus being unable to accurately identify key shortcomings.

This study systematically evaluated the effectiveness of ecological civilization construction in 48 cities along the Yellow River Basin from 2010 to 2022. Considering the specific development needs, ecological security, and strategic positioning of the basin, we developed a customized evaluation index system. The Entropy TOPSIS method was used for objective scoring, and the BP neural network was employed for adaptive weight analysis, revealing the evolution patterns and regional differences in ecological civilization construction. This approach scientifically identified key deficiencies and proposed a suggested path for the coordinated advancement of ecological civilization construction within the basin.

3. Materials and Methods

3.1. Study Area

The Yellow River is the second-longest river in China and is known as the “Mother River” of China. It originates from the Bayan Har Mountains in Qinghai Province, flows through nine provinces, and finally empties into the Bohai Sea. The main stream is approximately 5464 km long, and its drainage area covers 795,000 square kilometers [38]. According to the division plan of the Yellow River Conservancy Commission of the Ministry of Water Resources, the Yellow River can be divided into the upper, middle, and lower reaches [39]. The upper reaches flow through four provinces: Qinghai, Sichuan, Gansu, and Ningxia. The middle reaches flow through Shaanxi and Shanxi provinces, and the lower reaches flow through Henan and Shandong provinces. The Yellow River Basin has diverse ecosystem types, including forests, grasslands, and wetlands, and is an important grain production base in China [40]. In 2020, the total population in the Yellow River Basin was 422 million, accounting for 30% of the national population, and the regional GDP was CNY 25.39 trillion, accounting for 25% of the national GDP [41]. As an important base for energy, industry, and agriculture, as well as an ecological barrier in China [42,43,44], the Yellow River Basin has an irreplaceable strategic position in maintaining national ecological security, ensuring grain supply, and promoting regional coordinated development. Therefore, based on the consistent statistical standards and definitions of the indicator data for each city in the Yellow River Basin, this study selected 48 cities covering different economic development levels, ecological characteristics, and resource endowments in the upper, middle, and lower reaches of the Yellow River to ensure that the research results can reflect the overall situation of ecological civilization construction within the Yellow River Basin (Figure 1).

3.2. Evaluation Index System

Ecological civilization construction is a comprehensive and systematic project, aiming to promote the coordinated and sustainable development of the economy, politics, culture, society, and ecological environment, and ultimately achieve the goal of harmonious and sustainable economic and social development [45]. Given its systematic and long-term nature, scientifically and objectively measuring and evaluating the progress and effectiveness of ecological civilization construction is a key link in promoting its in-depth development. Therefore, establishing a scientific evaluation index system is the foundation for assessing the effectiveness. The selected indicators directly affect the representativeness of the evaluation results and are decisive for determining the countermeasures and priorities of ecological civilization construction [46]. This study adheres to the principles of scientificity, regionality, and reliability. It ensures that these indicators are consistent with the country’s major development strategies to reflect the actual situation of ecological civilization construction. Considering that the research scope covers multiple geographical levels, the evaluation system needs to reflect regional geographical characteristics while ensuring that the ecological civilization tasks of each city within the basin are implemented, to guarantee the reliability and validity of the indicators.

Research on ecological civilization construction is increasingly abundant, mainly focusing on aspects such as urban development, economic development, resource and energy consumption, living environment, ecological environment, and investment and construction intensity [8,47]. Drawing on the existing literature [48,49,50] and the China Ecological Civilization Construction Assessment Target System, this study constructed a 38-indicator system across five dimensions: Territorial Space, Socio-Economic Development, Resource and Energy Utilization, Eco-Environmental Protection, and Policy Inputs. Specifically, the Territorial Space dimension reflects urban spatial development intensity and land use efficiency; for instance, GDP per Unit Land Area (X1) effectively measures the economic output intensity of urban land, while the Ratio of Built-up Area to Urban Area (X2) indicates the extent of urban spatial expansion and development density. The Socio-Economic Development dimension covers both the scale of the urban economy and individual residents’ economic level; GDP per Capita (X14) serves as a core indicator for assessing the overall regional economic development level, and Per Capita Disposable Income of Urban Households (X15) directly reflects the actual economic well-being of individual residents. The Resource and Energy Utilization dimension measures energy consumption intensity and resource utilization efficiency, revealing the degree of urban development’s reliance on natural resources. Energy Consumption per Unit GDP (X18) and Water Consumption per Unit GDP (X20) quantify the consumption intensity of energy and water resources per unit of economic growth, respectively. The Eco-Environmental Protection dimension represents achievements in pollution control and environmental quality; the Proportion of Days with Good Air Quality (X29) provides a direct measure of air pollution control effectiveness, and the Harmless Treatment Rate of Domestic Waste (X31) signifies the level of solid waste pollution control. The Policy Inputs dimension focuses on assessing the government’s financial commitment to ecological civilization construction; R&D Expenditure as % of GDP (X33) demonstrates the intensity of investment in scientific and technological innovation, while Energy Conservation and Environmental Protection Expenditure as % of Public Finance (X36) directly reflects the degree of fiscal support for environmental protection efforts.

Table 1. Evaluation index system for ecological civilization construction.

Dimension	Indicator Name	Unit	Code	Attributes
Territorial Space	GDP per Unit Land Area	CNY 10,000 per km2	X1	Positive
	Ratio of Built-up Area to Urban Area	%	X2	Positive
	Number of Regular Higher Education Institutions	Count	X3	Positive
	Road Area per Capita	m2	X4	Positive
	Per Capita Housing Floor Area	m2 per person	X5	Positive
	Population Density	Persons per km2	X6	Positive
	Afforested Area as % of Administrative Area	%	X7	Positive
	Per Capita Public Green Space	m2	X8	Positive
Socio-Economic Development	Urbanization Rate of Permanent Residents	%	X9	Positive
	Share of Tertiary Industry in GDP	%	X10	Positive
	Engel’s Coefficient	%	X11	Negative
	Education Investment	CNY 10,000	X12	Positive
	Overall Labor Productivity	Yuan per person	X13	Positive
	GDP per Capita	Yuan	X14	Positive
	Per Capita Disposable Income of Urban Households	Yuan	X15	Positive
	Surveyed Urban Unemployment Rate	%	X16	Negative
Resource and Energy Utilization	Construction Land Use per Unit GDP	m2 per CNY 10,000	X17	Negative
	Energy Consumption per Unit GDP	Tons of standard coal per CNY 10,000	X18	Negative
	Electricity Consumption per Unit GDP	kWh per CNY 10,000	X19	Negative
	Water Consumption per Unit GDP	Tons per CNY 10,000	X20	Negative
	Comprehensive Utilization Rate of Solid Waste	%	X21	Positive
	Per Capita Cultivated Land Area	ha	X22	Positive
	Pesticide Usage per Unit of Cultivated Land	kg per ha	X23	Negative
	Sewage Treatment Rate	%	X24	Positive
Eco-Environmental Protection	Nitrogen Oxides Emissions per Unit GDP	kg per CNY 10,000	X25	Negative
	Sulfur Dioxide Emissions per Unit GDP	kg per CNY 10,000	X26	Negative
	Wastewater Discharge per Unit GDP	Tons per CNY 10,000	X27	Negative
	Green Coverage Rate of Built-up Areas	%	X28	Positive
	Proportion of Days with Good Air Quality	%	X29	Positive
	Forest Coverage Rate	%	X30	Positive
	Harmless Treatment Rate of Domestic Waste	%	X31	Positive
	Cumulative Treated Area of Soil Erosion	km2	X32	Positive
Policy Inputs	R&D Expenditure as % of GDP	%	X33	Positive
	Industrial Pollution Control Investment	CNY 10,000	X34	Positive
	Environmental Pollution Control Investment	CNY 10,000	X35	Positive
	Energy Conservation and Environmental Protection Expenditure as % of Public Finance	%	X36	Positive
	Green Travel Volume	10,000 person-times per 10,000 people	X37	Positive
	Nature Reserve Area as % of National Land Area	%	X38	Positive

details the specific indicators within each dimension, summarizing their classification, units, and codes.

3.3. Evaluation System

3.3.1. The Introduction of Entropy TOPSIS

This study employs the Entropy TOPSIS method to quantify the ecological civilization construction level of 48 cities in the Yellow River Basin from 2010 to 2022. This approach integrates the entropy weight method’s objectivity in determining indicator weights with the TOPSIS model’s capability in spatial distance quantification. Specifically, the entropy weight module dynamically assigns weights based on information entropy, mitigating subjective bias inherent in traditional methods, e.g., AHP, expert scoring [51]. Subsequently, the TOPSIS model ranks the cities by calculating their relative closeness to both the positive ideal solution and the negative ideal solution. The combination of these methods significantly enhances the objectivity and robustness of the evaluation for this complex system. According to the modeling steps of Entropy TOPSIS [52,53,54], the computational procedure is as follows:

Heterogeneous indicators, e.g., GDP per Unit Land Area measured in CNY 10,000 per km² vs. Ratio of Built-up Area to Urban Area in %, introduce dimensional conflicts, while their directional impacts on the system may oppose, e.g., Green Coverage Rate of Built-up Areas is positive, whereas Nitrogen Oxides Emissions per Unit GDP is negative. To eliminate dimensional bias and unify evaluation criteria, raw data $X_{ij}$ are normalized:

For positive indicators:

$$Y_{ij}={\displaystyle \frac{X_{ij}-min \left(X_{ij}\right)}{max \left(X_{ij}\right)-min \left(X_{ij}\right)}}$$

(1)

For negative indicators:

$$Y_{ij}={\displaystyle \frac{\mathit{max}\left(X_{ij}\right){-X}_{ij}}{max \left(X_{ij}\right)-min \left(X_{ij}\right)}}$$

(2)

where $X_{ij}$ is the value of the $j$th dimension indicator in the $i$th year, and $Y_{ij}$ is the standardized data of the $X_{ij}$.

The specific gravity is calculated as follows:

$$P_{ij}={\displaystyle \frac{Y_{ij}+d}{{\sum }_{i=1}^n(Y_{ij}+d)}}$$

(3)

where $d$ is a very small constant added to all standardized values for the convenience of calculating entropy. In this calculation, $d=0.001$.

Entropy is computed using the following formula:

$$E_j=-{\displaystyle \frac{1}{\mathit{ln}n}}\sum _{i=1}^n{P}_{ij}\mathit{ln}(P_{ij})$$

(4)

where ${\displaystyle \frac{1}{\mathit{ln}n}}$ is the normalization factor (entropy coefficient).

Weights are derived as follows:

$$W_{ej}={\displaystyle \frac{{1-E}_j}{{\sum }_{j=1}^n{(1-E}_j)}}$$

(5)

Construct the weighted standardized matrix:

$$Z={\left(Z_{ij}\right)}_{m\times n}=\left[\begin{array}{c}\begin{array}{c}{Z}_1\\ Z_2\end{array}\\ ⋮\\ Z_j\end{array}\right]=\left[\begin{array}{cccc}{W}_{e1}·Y_{11}& W_{e1}·Y_{21}& \cdots & W_{e1}·Y_{i1}\\ W_{e2}·Y_{12}& W_{e2}·Y_{22}& \cdots & W_{e2}·Y_{i2}\\ ⋮& \hspace{1em}⋮& & ⋮\\ W_{ej}·Y_{1j}& W_{ej}·Y_{2j}& \cdots & W_{ej}·Y_{ij}\end{array}\right]$$

(6)

Identify ideal solutions:

$$\begin{gathered} Z_j^{+}=max\left(Z_j\right) \\ {Z}_j^{-}=min\left(Z_j\right) \end{gathered}$$

(7)

Distance to positive/negative ideal solutions:

$$\begin{gathered} d_j^{+}=\sqrt{\sum _{j=1}^m{(Z_{ij}-Z_j^{+})}^2} \\ {d}_j^{-}=\sqrt{\sum _{j=1}^m{(Z_j^{-}-Z_{ij})}^2} \end{gathered}$$

(8)

Expected ecological civilization construction level:

$$Y_{expected}={\displaystyle \frac{d_j^{-}}{d_j^{+}+d_j^{-}}}$$

(9)

3.3.2. BP Neural Network

BP neural networks have gradually been introduced into the assessment of ecological civilization construction due to their linear mapping and fault-tolerant capabilities [36]. In this study, to capture the mapping relationship between the ecological civilization construction indicators and the comprehensive score in the Yellow River Basin, a three-layer BP neural network model was applied. This overcomes the limitations of subjective weight setting in traditional assessment methods and achieves objective evaluation through an adaptive learning mechanism (Figure 2).

Here,$X_n$ represents the input signals (standardized data of the indicators), and $n$, $l$, and $m$ represent the number of neurons in the input layer, hidden layer, and output layer, respectively. In the input layer, based on the indicator system for ecological civilization construction in the Yellow River Basin (which consists of 38 indicators), the number of input neurons is 38 ($n=38$). In the output layer, this study sets the number of neurons as 1 ($m=1$). There is no unified conclusion regarding the number of neurons in the hidden layer. It can be set using an empirical formula:

$$l=\sqrt{n+m}+a$$

(10)

where $a$ is the adjustment parameter ranging from 0 to 10. Based on model performance and sample size, this study ultimately sets $a=10$.

In the BP neural network model, the weights of the 38 indicators equal the product of the input-to-hidden layer weight matrix and the hidden-to-output layer weight matrix [55]. Therefore, the indicator weights can be calculated using the following formula:

$$W_{Bj}=\sum _{k=1}^n\left|{\left(W_{ih}\right)}_{jk}\times {\left(W_{ho}\right)}_k\right|$$

(11)

where $n$ is the number of hidden layer neurons, ${\left(W_{ih}\right)}_{jk}$ is the weight connecting input feature $j$ to hidden neuron $k$, ${\left(W_{ho}\right)}_k$ is the weight linking the hidden neuron $k$ to the output neuron, and $W_{Bj}$ is the feature weight of the feature $j$.

3.3.3. Classification of Levels for Ecological Civilization Construction

To visually present the spatial differentiation characteristics and hierarchical structure of the ecological civilization construction levels across 48 cities in the Yellow River Basin, this study, drawing on practices from similar research [28,56], employs the Natural Breaks Classification method (Jenks) to categorize the cities into hierarchical levels (divided into four levels; see

Table 2. Comprehensive score hierarchy division.

Comprehensive Scoring Range	Level of Construction
0.317–0.515	Ideal
0.212–0.316	Good
0.154–0.211	Medium
0.068–0.153	Poor

). The core principle of this method is to identify inherent natural breakpoints within the data through algorithms, based on the data’s own statistical distribution characteristics. It aims to maximize the homogeneity of city scores within the same level while simultaneously maximizing the heterogeneity of scores between different levels. Compared to the equal interval or quantile classification methods, the Natural Breaks method can more effectively reveal the intrinsic natural grouping structure within the data, reducing the subjectivity associated with manually setting thresholds. Consequently, this enables the classification results to more objectively reflect the actual spatial distribution pattern of ecological civilization construction levels within the study region.

3.4. Data Sources

The data used in this study is a panel dataset covering 48 cities in the Yellow River Basin from 2010 to 2022. All data originates from statistical sources. Among these, Construction Land Use per Unit GDP is calculated as built-up area/GDP, and Green Travel Volume is calculated as public transport passenger volume/urban population. The remaining data was primarily obtained from the China Urban Construction Statistical Yearbook, provincial statistical yearbooks (Qinghai, Inner Mongolia, Gansu, Shaanxi, Shanxi, Henan, Shandong), the official websites of the 48 city governments, municipal statistics bureau websites, municipal statistical yearbooks, and the National Economic and Social Development Statistical Bulletins. During data collection, there were instances where some indicators were missing for certain years (accounting for less than 1% of the total sample size). To ensure scientific rigor and objectivity of the results, this study supplemented the missing data by referencing the interpolation and trend analysis method [57].

4. Results

4.1. Evaluation Results of Ecological Civilization Construction

This study employed a BP neural network model to evaluate the ecological civilization construction level of 48 cities in the Yellow River Basin from 2010 to 2022. The R², MSE, MAE, and RMSE metrics were adopted as validation indicators to confirm that the BP neural network model can effectively assess ecological civilization construction levels (

Table 3. BP neural network verification indicators.

	R2	MSE	MAE	RMSE
Train	0.9376	0.0004	0.0151	0.0197
Test	0.9477	0.0004	0.0150	0.0198

).

Both the training set and test set achieved R² values exceeding 0.93, indicating that the model has an excellent goodness of fit and can effectively capture the level of ecological civilization construction. Furthermore, the values of all three error metrics (MSE, MAE, and RMSE) were very low (each below 0.02). The highly consistent results between the training set and test set further demonstrate the model’s outstanding prediction accuracy and robust generalization capability. When comparing true values and predicted values in the test set (Figure 3), minimal discrepancies were found. This strongly confirms that the BP neural network model developed in this study can accurately and reliably evaluate the ecological civilization development level of cities in the Yellow River Basin.

The Formula (10) was used to calculate the index weights in the BP neural network model (Figure 4). The analysis demonstrates significant variation in the relative contribution of individual indicators, with GDP per Unit Land Area (X1) possessing the highest weight of 0.492, closely followed by Green Coverage Rate of Built-up Areas (X28) at 0.490, underscoring the paramount importance of economic density and urban greening in assessing ecological civilization levels in this specific regional context. Other indicators within the Territorial Space layer, namely Per Capita Housing Floor Area (X5, 0.468), Road Area per Capita (X4, 0.449), and Number of Regular Higher Education Institutions (X3, 0.438), also carry substantial weights, collectively highlighting the critical role of spatial development quality, infrastructure, and educational resources.

The Socio-Economic Development layer shows considerable influence, particularly through Overall Labor Productivity (X13, 0.432) and Share of Tertiary Industry in GDP (X10, 0.405), emphasizing economic structure and efficiency. Resource and Energy Utilization indicators generally carry lower weights, though Comprehensive Utilization Rate of Solid Waste (X21, 0.386) and Sewage Treatment Rate (X24, 0.365) are relatively prominent, pointing to the significance of waste management efficiency. The Eco-Environmental Protection layer is strongly represented by X28 and also features Proportion of Days with Good Air Quality (X29, 0.313) and Forest Coverage Rate (X30, 0.359), reinforcing the centrality of environmental quality metrics. Policy Input indicators, such as R&D Expenditure as % of GDP (X33, 0.312) and Industrial Pollution Control Investment (X34, 0.357), exhibit moderate importance, while Energy Conservation and Environmental Protection Expenditure as % of Public Finance (X36, 0.206) and Nature Reserve Area Coverage (X38, 0.309) show comparatively lower weights, potentially indicating a lag between policy input and measurable outcomes.

The distribution of weights clearly shows that the Territorial Space and direct Eco-Environmental Protection dimensions are the most significant in this assessment model for the Yellow River Basin cities. The high weighting of negative indicators like Surveyed Urban Unemployment Rate (X16, 0.447) and Energy Consumption per Unit GDP (X18, 0.313) confirms that social challenges and inefficiencies act as significant constraints on ecological civilization progress. This weighting structure reflects the complex interplay between economic development intensity, environmental quality, resource efficiency, and social well-being inherent to evaluating ecological civilization in this strategically vital and ecologically sensitive basin. The results provide a quantitative foundation for identifying key leverage points for policy intervention and urban management aimed at enhancing ecological civilization levels across the region.

4.2. Changes in Level of Ecological Civilization Construction

To reveal the dynamic characteristics of the ecological civilization construction level in cities along the Yellow River Basin during the study period and to analyze the phased features and trends in its development process, this study classified the 48 cities into different tiers based on their annually calculated comprehensive scores for ecological civilization construction from 2010 to 2022. By systematically analyzing the distribution changes in cities across tiers over these 13 years, the temporal evolution of the percentage distribution of cities at each tier was summarized and presented (Figure 5).

Between 2010 and 2022, the level of ecological civilization construction in 48 cities along the Yellow River Basin exhibited significant dynamic evolution characteristics. The analysis shows that the overall level of ecological civilization construction gradually increased, but with periodic fluctuations. The number of cities at the Poor level generally decreased from 29 in 2010 to 17 in 2022, but there was a temporary increase in 2015. The number of cities at the Medium level significantly increased from 9 in 2010 to 19 in 2022, becoming the most numerous category. This number also experienced a decline in 2015. In combination with the changes in the Poor level, it indicates that some cities’ ecological civilization construction levels, after reaching the Medium level, show instability and may fall back to the Poor level. The number of cities at the Good level fluctuated between eight and nine, without showing a clear upward or downward trend. The number of cities at the Ideal level remained at a relatively low level, varying between two and five, with relatively stable annual changes and limited growth.

4.3. The Spatio-Temporal Distribution of the Level of Ecological Civilization Construction

The ecological civilization construction level in the Yellow River Basin demonstrated a steady upward trend from 2010 to 2022. This study selected four years, specifically 2010, 2014, 2018, and 2022, for visualization analysis (Figure 6), clearly illustrating the grade transitions of cities across the basin. The results indicate that three core cities, namely Zhengzhou, Taiyuan, and Xi’an, consistently maintained leading positions throughout the study period; yet, surrounding cities showed limited improvement. Although no cities in the upstream attained the Ideal level, they featured a higher concentration of Good level cities, with their overall average level surpassing those in the midstream and downstream. While the midstream cities exhibited overall progress, significant shortcomings persisted, exemplified by cities like Linfen, Weinan, and Yuncheng remaining at the Poor level. The downstream witnessed notable cases of rapid advancement, such as Anyang and Jining. In conclusion, despite overall improvements in ecological civilization construction across the basin, regional development disparities remain pronounced: the upstream sustains favorable conditions leveraging their ecological endowment; core cities in the midstream maintain high standards through inherent advantages but demonstrate limited spillover effects to neighboring areas; and although not yet outstanding overall, the downstream displays consistent and steady growth momentum.

5. Discussion

5.1. Spatio-Temporal Evolution of Ecological Civilization Levels in 48 Cities of Yellow River Basin

In the Yellow River Basin, except Xining and Lanzhou, the ecological civilization construction in other provincial capital cities generally achieves a relatively high level. This phenomenon may be related to factors such as the fragile ecological environment and lower economic development levels in the northwestern region [58]. Among them, provincial capital cities like Zhengzhou and Xi’an consistently maintain a high level, benefiting from easier access to policy support and financial resources, which enables them to attract more high-quality resources and talent. Taking Zhengzhou as an example, after being designated as a National Central City in 2017, it achieved significant progress across economic, ecological, and social welfare fields by leveraging its powerful urban siphon effect [39]. However, the improvement in ecological civilization construction levels in surrounding cities has been relatively limited, indicating that the central cities have not yet produced a significant positive spillover effect on their neighboring areas [33]. This also points to deficiencies in regional coordinated development mechanisms. The lack of effective coordination mechanisms and information-sharing platforms often prevents provincial capitals and their surrounding cities from making joint efforts in industrial development and infrastructure construction, hindering coordinated progress.

5.2. Future Directions for Building Ecological Civilization

Enhancing ecological civilization in the Yellow River Basin necessitates integrated strategies prioritizing regional differentiation, cross-jurisdictional collaboration, and institutional reinforcement. Region-specific interventions must address critical vulnerabilities. Upper reaches require establishing robust ecological barriers through natural forest conservation and wetland restoration, with particular emphasis on headwater protection [59]. Middle reaches demand comprehensive soil erosion control via integrated watershed management while strengthening ecological risk mitigation capabilities [60]. Lower reaches facing significant anthropogenic pressures should intensify multipollutant abatement strategies targeting industrial, agricultural, and municipal sources, implementing strict effluent controls to rehabilitate degraded ecosystems [44].

Effective transregional coordination constitutes a fundamental requirement. Establishing basin-wide governance frameworks facilitates interjurisdictional infrastructure connectivity across transportation, energy, and digital networks, thereby optimizing resource mobility. Core cities should catalyze knowledge spillovers and resource redistribution toward peripheral areas through joint industrial parks, technology transfer programs, and skilled workforce development initiatives [19]. This approach fosters synergistic regional advancement. Implementing cross-boundary ecological compensation mechanisms remains essential to align upstream conservation incentives with downstream beneficiary responsibilities.

Systematic institutional strengthening underpins sustainable outcomes. Integrating resource efficiency, environmental impact, and ecological performance metrics into subnational government evaluation systems incentivizes sustainability-oriented governance [43]. Concurrently, expanding stakeholder engagement channels ensures information transparency, participatory decision-making, and public oversight. Targeted environmental awareness campaigns cultivate low-carbon lifestyles while consolidating society-wide stewardship of ecological commons. This institutional ecosystem enables sustained policy implementation beyond political cycles.

6. Conclusions

This study selected 48 cities in the Yellow River Basin as research objects. An evaluation system comprising 38 indicators across five dimensions—Territorial Space, Socio-Economic Development, Resource and Energy Utilization, Eco-Environmental Protection, and Policy Inputs—was constructed. By comprehensively applying the Entropy TOPSIS and BP neural network methods, the ecological civilization construction levels of cities within the basin from 2010 to 2022 were assessed and their spatio-temporal evolution characteristics and key influencing factors were analyzed.

The findings reveal that from 2010 to 2022, the ecological civilization construction level of the 48 cities in the Yellow River Basin generally exhibited a steady upward trend, confirming the overall effectiveness of ecological civilization initiatives. However, regional development imbalances remain prominent. Provincial capitals generally demonstrated strong performance, yet their spillover effects on surrounding cities were limited. The ecological civilization development in the upper reach is primarily constrained by economic development levels and financial capacity. The middle reach displayed significant disparities in ecological civilization levels, where provincial capitals maintained leading positions while exhibiting limited catalytic effects on neighboring cities. The lower reach faces substantial anthropogenic pressures and requires coordinated development between economic growth and ecological conservation. The indicator weights analysis indicates that enhancing land-intensive use efficiency and increasing urban green space coverage are effective measures for improving ecological civilization construction. High-quality urban infrastructure and educational resources can elevate ecological civilization levels through eco-innovation and heightened public environmental awareness. Due to the lagged implementation effects, policy inputs demonstrate limited effectiveness in promoting ecological civilization construction.

By integrating the objective weighting advantages of Entropy TOPSIS and the nonlinear mapping capabilities of BP neural networks, this study established a more reliable evaluation paradigm for ecological civilization construction, providing methodological innovation for basin-scale assessment. Tailored optimization measures and regional coordination pathways were proposed according to the distinct ecological contexts, challenges, and development shortcomings across the upper, middle, and lower reaches. This research holds significant importance for advancing the in-depth implementation of national ecological protection and high-quality development strategies in the Yellow River Basin.

Despite these contributions, several limitations persist. The study primarily relied on statistical data, and due to data availability constraints, not all cities in the Yellow River Basin were included. Evaluation results may also be influenced by statistical calibers or data quality issues. Future research could incorporate multi-source data to enrich evaluation dimensions.