Sustainability Assessment of Regional Water Resources in China Based on DPSIR Model

Abstract

Water resources are an important aspect of China’s ecological governance. Due to the uneven spatial distribution of water resources in China, it is necessary to analyze the differences in the sustainable development level of water resources in different regions. This article combines the national policies of “promoting the coordinated development of the three waters” and “integration of people, city, industry, and economy” to construct an evaluation index system for the sustainable development of water resources in China using the water resources carrying capacity (WRCC). The spatial and temporal differentiation characteristics of water resources sustainable development in 31 provinces of China are analyzed. The results show that (1) the level of sustainable development of China’s water resources has significantly increased, with the index increasing from 6.229 in 2005 to 9.792 in 2021. (2) The spatial pattern of China’s WRCC shows an increasing trend from west to east and from north to south, and the aggregation status is obvious and fluctuates during the entire study period. (3) Currently, the strength of the expenditure of research and development funds for industrial enterprises and the sewage treatment capacity continue to increase, and the interaction between different factors shows a two-factor enhancement or nonlinear enhancement. To further improve the level of sustainable development of water resources, each province needs to formulate development plans based on local conditions and promote the coordinated development of industrial restructuring, environmental governance, and water resources development.

1. Introduction

Water resources are the material basis of human life, social development, and industrial production, and are of great significance to human society [1]. With the growth of population and the rapid development of economy, the demand for water is increasing, causing a series of negative phenomena including water shortage [2] and water pollution [3], which in turn restricts the sustainable development [4] of human society and economy. The total amount of water in China is abundant, but the per capita resource amount is only a quarter of the world’s average level [5], and there is a serious regional imbalance in distribution [6]. The Chinese government attaches great importance to the sustainable utilization of water resources and has issued a series of policy documents [7,8] and strategic plans [9] to reduce water pollution, protect water bodies, and restore natural ecology. In 2021, the National Development and Reform Commission issued the Notice on the Comprehensive Management Plan for Water Environment in Key River Basins during the 14th Five-Year Plan period and the Plan for Building a Water-Saving Society during the 14th Five-Year Plan period. In 2022, the Ministry of Water Resources issued the Opinion on Strengthening the Unified Management of Water Resources in River Basins, and the 20th National Congress of the Communist Party of China further proposed to coordinate the governance of water resources, water environment, and water ecology. This indicates that the sustainable development of water resources remains a key focus of the Chinese government’s work [10] and provides a roadmap for the next steps in water resource governance in China. Water resources carrying capacity (WRCC) is an effective indicator for evaluating water resource security. Studying the evolution trend of China’s inter-provincial WRCC and analyzing the changes in driving factors from time and space dimensions have important reference value for the government to make sustainable development decisions.

The concept of carrying capacity was first applied to ecological communities, reflecting the maximum number of biological species that an ecosystem can carry [11,12]. Today, this concept has been widely used in the field of natural sciences to represent the ability of an environment or ecosystem to sustain development and specific activities [13,14]. Clarke, A. [15], Song, X. [16], Xu Youpeng [17], Zuo Qiting [18] and others applied carrying capacity to the field of water resources, and summarized WRCC as the maximum capacity meeting the water demand of industry, agriculture, population, and ecological environment protection—that is, the maximum capacity of the water resource system to support economic and social development.

The WRCC research focuses on three aspects. The first is conceptual and theoretical research. Some scholars focus on the vulnerability of the ecological environment and carry out theoretical research on WRCC from the perspective of sustainable development [19], discussing mainly from the aspects of characteristics, connotation, and index system [20]. The second is to study the carrying capacity of water resources for industries or specific production activities. For example, He, L. et al. [21] studied the supporting role of water resources for agricultural development. Chi, M. et al. [22] studied the relationship between coal mining and carrying capacity of water resources. The third is to comprehensively evaluate regional WRCC [23,24,25] by paying attention to the coordinated development of water resources, ecological environment, economy, and society, building a WRCC evaluation index system with the help of model tools and quantitatively calculating the WRCC index, therefore providing suggestions on regional socio-economic development and water resources development and protection. Among them, the regional comprehensive evaluation of WRCC is the focus of research.

The regional differences of WRCC evaluation and research are quite large, which can be summarized into four aspects: city, inter-region, river basin, and the country. The city evaluation is mainly concentrated on urban areas or city clusters. For example, Song, XM [16] studied the carrying capacity of Tianjin water resources on population size and economic scale; Peng, T. et al. [26] built a DPESBRM model to systematically evaluate the WRCC of Guiyang City from the perspective of drive-pressure engineering–water-shortage state–ecological foundation–response management. The regional level selects regions with certain attributes according to economic, political, and climate differences. For example, Wang, G. et al. [27] conducted quantitative and qualitative evaluation of WRCC in the Changji Economic Circle under different social development plans by fuzzy comprehensive evaluation (FCE), grey correlation analysis, and multiple linear regression model. Han, Y. et al. [28] comprehensively evaluated WRCC in northern China, combining multiple factors including water resources, economy, and society. The river basin level mainly studies river basins within the watershed range. For example, Deng, L. et al. [13] used principal component analysis to study the WRCC of the Han River basin from the perspectives of water resources, society, and ecological environment. The country level is based on the national scale to study the spatial differences of WRCC in the whole country and the provinces within the country. For example, Lv, A. et al. [29] used the fuzzy comprehensive evaluation method to evaluate China’s WRCC in the process of climate change, urbanization, and industrialization. Overall, regardless of the size of the research area, it studies the coordinated relationship between society, economic conditions, ecology, and water resources as well as the carrying capacity of water resources within the area for economic and social development. It pays attention to the spatial analysis and differences on WRCC within the area, draws targeted conclusions through empirical analysis, serves for regional water resources planning, and improves water resource management efficiency.

WRCC evaluation methods mainly include principal component analysis [30,31], analytic hierarchy process [32,33], system dynamics [34,35] and fuzzy analysis [36,37]. However, traditional analysis focuses on static trend analysis, and the research perspective is single, ignoring the systematic and dynamic nature of the indicators. Therefore, some scholars try to combine multiple methods to improve the comprehensiveness of the research. For example, Wang, X. et al. [38] simulated five typical scenarios designed for different purposes in Guangzhou from 2021 to 2030 based on the constructed SD model and evaluated the coordination of the WRCC system by introducing a coupling coordination model; Wang, G. et al. [39] proposed a WRCC evaluation method that combines the improved fuzzy comprehensive evaluation (IFCE) method with the SD model. Through the feedback adjustment and system simulation of socioeconomic water resources and water environment interactions, the evaluation method calculates dynamically the carrying capacity of water resources under different social development scenarios.

Generally speaking, predecessors have conducted good studies on WRCC evaluation, which has effectively promoted the theoretical and practical progress of WRCC. However, the existing research still has the following limitations. First of all, most of the existing research focuses on specific provinces or city clusters, and there is a lack of national-level research. Secondly, most studies only analyze from a single dimension of time or space, lacking a systematic temporal and spatial differences analysis of regional WRCC. In addition, some scholars’ research involves the evaluation of the driving factors of WRCC, and few have studied the interaction intensity of multiple factors.

The contributions and innovations of this paper are shown in the following aspects: first, based on the perspective of spatial-temporal heterogeneity, WRCC in 31 provinces in China were studied, and regional WRCC temporal-spatial differences and spatial correlation were analyzed; second, the spatial-temporal differences were detected by geodetectors; the intensity of the driving factors behind them and the interaction between them were explained, providing a basis for the targeted management of “zoning-grading” water resources.

The rest of this paper is organized as follows: the second part includes method introduction and data sources; the third part is the analysis of the results of China’s inter-provincial WRCC and the analysis of the spatial effect of China’s comprehensive WRCC; the fourth part is the analysis of the factors affecting the spatial differences of China’s WRCC; and the fifth part is the conclusion and discussion.

2. Research Methods and Data Sources

2.1. DPSIR Model Construction and Index Identification

2.1.1. WRCC Logic Based on DPSIR Model

As a complex system including environment, economy, society, and other elements, ecological security needs to build a comprehensive and scientific index system for its quantitative analysis [40]. In order to comprehensively analyze and describe the relationship between environmental problems and social development, the European Environment Agency (EEA) proposed in 1997 the driving-force pressure-state impact-response (DPSIR) model [41], which provides ideas for comprehensive evaluation of ecological security.

Relevant studies have shown that [40,41,42] the DPSIR model is an evaluation model based on the causal organization correlation index, which can describe the continuous feedback mechanism between factors in detail and characterize the factors with fusion and conflict features [25]. The DPSIR model is applicable to complex systems and has excellent systematicity. However, it also has shortcomings such as subjectivity. The WRCC evaluation index system based on the DRSIR model describes a causal chain that causes environmental problems from origin to outcome: potential economic and social driving forces → pressure on water resource systems → resource and environmental states → impacts of the states → human response to the impacts. It can be expressed as follows: urbanization, population, and economy act as long-term driving forces (D) to promote the development and utilization of local water resources; the side effects of the production and life of humans on the water environment create pressure (P), resulting in changes on the state of the water environment (S); the continuous changes on the water environment state (S) in turn have an impact on human society (I), and these effects (I) prompt humans to respond (R) to the changes in the ecological environment state (S); the actions taken react against the driving force (D), the environmental pressure (pP), and the state (S).

In the model, the driving force (D) includes social, economic, and human activities; pressure (P) includes direct or indirect pollutant discharge; state (S) includes water environment, vegetation, etc.; impact (I) includes water environment-related ecosystem quality; response (R) is the policy or technical response of the decision-maker to an undesired impact. The framework of the model is shown in Figure 1.

2.1.2. Evaluation Index Selection

In 2021, the National Development and Reform Commission (NDRC) and 10 other ministries in China issued the “Guiding Opinions on Promoting the Utilization of Sewage Resources”, aimed at promoting the utilization of sewage resources in key areas such as urban, industrial, agricultural, and rural sectors. Key projects such as regional water recycling, technology innovation pilots for near-zero sewage discharge, and comprehensive pilots for sewage resource utilization were implemented to achieve the coordinated governance of the “three waters”—water resources, water environment, and water ecology—and significantly enhance the ecological and economic benefits. In 2022, the NDRC formulated the “14th Five-Year Plan for New-Type Urbanization,” proposing to create a coordinated development pattern for urbanization and strengthen the restoration of wetland ecology and water environment. Based on these national policies and using the DRSIR model and previous research [43,44,45], this article selected 21 indicators to establish the WRCC evaluation index system to objectively and scientifically assess the WRCC in 31 provinces in China. The attribute of the indicators is represented by the symbol “+” indicating that the larger the indicator is, the better, while “−” indicates that the smaller the indicator is, the better (

Table 1. China WRCC evaluation index system.

Target Layer	Standard Layer	Index Layer	Unit	Indicator Property	Comprehensive Weight
WRCC Evaluation Index System	Driving force (D)	X1: GDP per capita	yuan/person	+	0.047
		X2: Urbanization rate	%	+	0.023
		X3: Population density	person/km2	−	0.009
		X4: Proportion of tertiary industries	%	+	0.030
	Pressure (P)	X5: Discharge of industrial wastewater	kiloton	−	0.011
		X6: Fertilizer application per unit of cultivated land area	t/hm2	−	0.011
		X7: Water consumption per 10,000 yuan of industrial added value	m3	−	0.010
		X8: Water consumption per 10,000 yuan of GDP	m3	−	0.005
		X9: Domestic water consumption per capita	m3	−	0.022
	State (S)	X10: Precipitation	108 m3	+	0.035
		X11: Per capita water resources	m3/person	+	0.288
		X12: Grain production per capita	t/person	+	0.053
		X13: Water supply of surface water resources	108 m3	+	0.061
		X14: Forest coverage	%	+	0.040
	Influence (I)	X15: Water supply penetration rate	%	+	0.007
		X16: Effective irrigation area	103 hm	+	0.053
		X17: Green coverage rate of built-up areas	%	+	0.014
	Response (R)	X18: Expenditure on industrial scientific research above designated size	ten thousand yuan	+	0.119
		X19: Investment amount completed in industrial pollution control	ten thousand yuan	+	0.068
		X20: Sewage treatment rate	%	+	0.034
		X21: Sewage treatment capacity	104 m3/day	+	0.062

).

2.2. Research Methods

2.2.1. Combination Weighting of Entropy Weight and Coefficient of Variation—TOPSIS Method

Compared with subjective weighting methods such as the Delphi method and AHP, the entropy weight method is more objective and can better explain the results and use the difference between information to carry out weighting [46,47]. The specific steps are as follows.

(1)

Index normalization. Suppose the original matrix is $X={\left(x_{ij}\right)}_{m\times n}$; m is the number of provinces; n is the number of indicators; $x_{ij}$ and is the original data of the i-th province and the j-th indicator. For positive indicators and negative indicators, normalization processing is performed respectively, and a standardized matrix $Y={\left(y_{ij}\right)}_{m\times n}$ is obtained after processing. The specific formula is

$$X=\left[\begin{array}{ccc}{x}_{11}& \cdots & x_{1n}\\ ⋮& \ddots & ⋮\\ x_{m1}& \cdots & x_{mn}\end{array}\right]$$

(1)

$$y_{ij}=\frac{x_{ij}-min\left(x_{ij}\right)}{\mathit{max}\left(x_{ij}\right)-min\left(x_{ij}\right)}\mathit{(}\mathit{for\; positive\; indicators}\mathit{)}$$

(2)

$$y_{ij}=\frac{\mathit{max}\left(x_{ij}\right)-x_{ij}}{\mathit{max}\left(x_{ij}\right)-min\left(x_{ij}\right)}\mathit{(}\mathit{for\; negative\; indicators}\mathit{)}$$

(3)

$$Y=\left[\begin{array}{ccc}{y}_{11}& \cdots & y_{1n}\\ ⋮& \ddots & ⋮\\ y_{m1}& \cdots & y_{mn}\end{array}\right]$$

(4)

(2)

Determine the index weight. To use the entropy weight method to calculate the index weight, the matrix Y is first normalized to obtain the matrix $f_{ij}$. The calculation formula is as follows.

$$f_{ij}=\frac{y_{ij}}{\sum _{j=1}^m{y}_{ij}}$$

(5)

Then, calculate the information entropy $e_i$; the calculation formula is as follows.

$$e_j=-\frac{1}{lnm}{\sum }_{i=1}^m\left[f_{ij}\times ln{f}_{ij}\right]$$

(6)

Finally, calculate the index weight ${\omega }_j^{\mathrm{’}}$; the calculation formula is as follows.

$${\omega }_j^{’}=\frac{1-e_j}{{\sum }_{j=1}^n(1-e_j)}$$

(7)

The coefficient of variation method is an objective weighting method, which is not affected by the dimension of the index. It can reflect the spatial differences of WRCC in various provinces in China. The calculation formula is

$$\mathrm{C}=\frac{1}{R_0}\sqrt{\frac{1}{m}{\sum }_{i=1}^m{(R_I-R_0)}^2}$$

(8)

$${\omega }_j^{\mathrm{’}\mathrm{’}}=\frac{C}{{\sum }_{j=1}^{n}C}$$

(9)

In the formula, C is the coefficient of variation; $R_0$ is the average value of the j-th index; $R_i$ is the original value of the j-th index; and ${\omega }_j^{\mathrm{’}}$ is the weight of the j-th index. The formula for calculating the combined weight ${\omega }_j$ is:

$${\omega }_j=\frac{{\omega }_j^{\mathrm{’}}+{\omega }_j^{\mathrm{’}\mathrm{’}}}{2}$$

(10)

(3)

Construct TOPSIS model

In order to secure the objectivity, this paper creates a normalized analysis matrix $C={\left|{\omega }_j\times x_{ij}\right|}_{m\times n}$ according to the index weights ${\omega }_j$. The positive ideal solution Z⁺ and the negative ideal solution Z⁻ are the maximum and minimum values of the i-th index. The specific formula is as follows.

$$Z^{+}=max\left\{y_{ij}\right\}$$

(11)

$$Z^{-}=min\left\{y_{ij}\right\}$$

(12)

(4)

Calculation of WRCC index

In this paper, the Euclidean distance is used to calculate the specific distance between China’s WRCC evaluation index and the positive and negative ideal. D⁺ is the distance between the i-th index and $Z^{+}$; D⁻ is the distance between the i-th index and $Z^{-}$. D is the index of China’s WRCC. The larger the value is, the more secure it is. The calculation formula is as follows.

$$D_i^{+}=\sqrt{{\sum }_{j=1}^n{(Z_i^{+}-y_{ij})}^2}$$

(13)

$$D_i^{-}=\sqrt{{\sum }_{j=1}^n{(Z_i^{-}-y_{ij})}^2}$$

(14)

$$\mathrm{D}=\frac{D_j^{-}}{D_j^{+}{+D}_j^{-}}$$

(15)

2.2.2. Spatial Autocorrelation Analysis

Spatial autocorrelation analysis can reflect whether there is a correlation between a certain area and adjacent areas [46]. In this paper, the global Moran index and local Moran index methods are used to explore the horizontal spatial aggregation of WRCC in 31 provinces in China. The calculation formula is:

$$I=\frac{\sum _{i=1}^n\sum _j^n{\omega }_{ij}(x_i-\stackrel{-}{x})/(x_j-\stackrel{-}{x})}{S^2\sum _{i=1}^n\sum _{j=1}^n{\omega }_j}$$

(16)

$$I^{\mathrm{’}}=\frac{x_i-\stackrel{-}{x}}{S^2}{\sum }_j^n{\omega }_{ij}(x_i-\stackrel{-}{x})$$

(17)

In the formula, $I$ represents the global Moran index, and the value range is [−1, 1]. When I > 0, it indicates that the regional WRCC presents a phenomenon of spatial aggregation. When I < 0, it indicates that the regional WRCC tends to be dispersed, showing spatial differences. When $I^{\mathrm{’}}$ represents the local Moran index, $I^{\mathrm{’}}>0$ means that the WRCC of the province and the adjacent provinces shows a spatial aggregation of similar values—that is, “high-high aggregation” or “low-low aggregation”; $I^{\mathrm{’}}<0$ indicates that the WRCC of the province and the adjacent provinces shows non-spatial aggregation of similar values—that is, “low-high aggregation” or “high-low aggregation”. N is the number of provinces; i and j are different spatial units; x is the closeness of WRCC; $\stackrel{-}{x}$ and S² are the average value and variance of WRCC in each province; ${\omega }_{ij}$ is the spatial weight matrix. In this paper, if the spatial units are adjacent, ${\omega }_{ij}$ takes 1; otherwise, it takes 0.

2.2.3. Geodetector

A geodetector is a spatial statistical tool that detects the spatial differences of things and reveals the driving factors behind them [29]. Its core theory is that if an explanatory variable has an important impact on the spatial differences of another explanatory variable, then the spatial distribution of the tow explanatory variables should have significant similarities. This paper uses this model to detect the driving factors of the spatial-temporal differences of WRCC in 31 provinces in China. The specific formula is:

$$q=1-\frac{1}{N{\sigma }^2}{\sum }_{h=1}^{L}Nh{\sigma }_h^2$$

(18)

In the formula, q is the correlation index of the WRCC, and the value range of q is [0, 1]. The larger the q is, the greater the impact of the driving factor on the WRCC is; h is the classification number of the driving factor; N is the number of provincial units; σ² is the variance of WRCC in each province.

2.3. Data Sources

This paper selects 31 provinces in China (due to part of the data missing, the research areas do not include Hong Kong, Taiwan, and Macau) as the research areas. The data come from the “China Statistical Yearbook”, “China Water Resources Bulletin”, and “China Environmental Statistical Yearbook” from 2005 to 2021, as well as the statistical yearbooks and water resources bulletins of various provinces and cities. Part of the annual data is missing or cannot be directly obtained, and the interpolation method is used to calculate and supplement.

2.4. Study Area

China’s water resource carrying capacity is determined by a complex interplay between natural and socio-economic factors. In terms of the natural environment, China is characterized by uneven distribution of water resources, with the southern regions having higher rainfall and greater water availability than the arid north. Additionally, widespread water pollution and degradation of aquatic ecosystems pose a major threat to water security. In terms of socio-economic factors, China’s rapid economic growth has led to increased water consumption and pollution, particularly in the industrial and agricultural sectors. Population growth and urbanization have also exacerbated water scarcity and competition among different water users in certain regions. Furthermore, there are significant variations in water availability and demand among different provinces, making it necessary to develop targeted water resource management strategies and policies. The study area and provincial distribution are shown in Figure 2.

3. Spatial-Temporal Changes of Inter-Provincial WRCC in China

3.1. Time Change Analysis of WRCC in China

Using the combined weighted TOPSIS method to measure the water resource carrying capacity (WRCC) index of China and 31 provinces from 2005 to 2021, this paper selects the data of 2005, 2010, 2015, and 2021 to draw the time trend graph of WRCC index (Figure 3). As shown in Figure 2, China’s WRCC level showed a steady upward trend from 2005 to 2021, reaching the highest value of 9.79 in 2021, an increase of 57.2% compared with 2005. Through analyzing the five sub-indicators of China’s water resource carrying capacity, it is evident that the response (R) indicator has shown the highest growth rate, increasing by 279.2% since 2005, while the driving force (D) indicator has also increased by 101.3%, which is greater than the growth rate of other indicators, indicating that China’s actual driving force and basic ability to improve the ecological environment are gradually increasing and are able to implement practical measures to improve the ecological environment.

Table 2. WRCC index of each province in China in 2005, 2010, 2015, 2021.

Serial Number	Province	Sort by Position		WRCC Index
		W–E Index	S–N Index	2005	2010	2015	2021
1	Beijing	21	26	0.199	0.242	0.278	0.305
2	Tianjing	24	25	0.175	0.205	0.253	0.254
3	Hebei	19	23	0.199	0.242	0.291	0.310
4	Shanxi	14	22	0.158	0.207	0.236	0.250
5	Inner Mongolia	13	27	0.180	0.225	0.290	0.314
6	Liaoning	29	28	0.194	0.223	0.254	0.288
7	Jilin	30	29	0.192	0.215	0.259	0.297
8	Heilongjiang	31	31	0.196	0.247	0.334	0.368
9	Shanghai	28	14	0.180	0.235	0.281	0.313
10	Jiangsu	25	16	0.258	0.319	0.402	0.455
11	Zhejiang	27	11	0.236	0.283	0.345	0.383
12	Anhui	22	15	0.203	0.260	0.310	0.354
13	Fujian	26	5	0.234	0.271	0.309	0.331
14	Jiangxi	20	8	0.212	0.269	0.299	0.333
15	Shandong	23	21	0.252	0.294	0.368	0.390
16	Henan	17	18	0.198	0.241	0.287	0.333
17	Hubei	18	13	0.212	0.260	0.296	0.354
18	Hunan	15	7	0.216	0.273	0.320	0.356
19	Guangdong	16	3	0.251	0.334	0.396	0.482
20	Guangxi	10	2	0.198	0.245	0.296	0.310
21	Hainan	12	1	0.194	0.218	0.226	0.258
22	Chongqing	8	9	0.181	0.222	0.255	0.287
23	Sichuan	6	12	0.201	0.229	0.271	0.323
24	Guizhou	9	6	0.154	0.201	0.240	0.273
25	Yunnan	4	4	0.199	0.220	0.269	0.291
26	Tibet	2	10	0.352	0.367	0.328	0.351
27	Shaanxi	11	17	0.176	0.219	0.252	0.272
28	Gansu	5	19	0.142	0.164	0.207	0.232
29	Qinghai	3	20	0.137	0.158	0.185	0.219
30	Ningxia	7	24	0.135	0.170	0.197	0.213
31	Xinjiang	1	30	0.216	0.235	0.278	0.290

Note: The W–E index represents the position of Chinese provinces on the longitude dimension, with smaller values indicating closer proximity to the west; the S–N index represents the position of Chinese provinces on the latitude dimension, with smaller values indicating closer proximity to the south.

presents the WRCC Index of various provinces in China in 2005, 2010, 2015, and 2021. It can be observed that, in 2021, the provinces of Guangdong, Jiangsu, Shandong, Zhejiang, Heilongjiang, Hunan, and Anhui have higher WRCC Index compared to other provinces, while the provinces of Ningxia, Qinghai, Gansu, and Shanxi have lower WRCC Index. From 2005 to 2021, the WRCC Index of most provinces has significantly increased, with only a slight decrease observed in the WRCC Index of Tibet. Among the provinces with significant WRCC Index improvement, Guangdong, Heilongjiang, Guizhou, Jiangsu, Inner Mongolia, Anhui, and Shanghai have an increase of over 70% and rank high in China. On the other hand, the provinces of Hainan, Xinjiang, Fujian, Tianjin, Yunnan, and Liaoning have a WRCC Index increase of less than 50%, with lower growth compared to other provinces.

3.2. Spatial Change Analysis of WRCC in China

3.2.1. Spatial Pattern of WRCC in China

In order to further compare and analyze the regional differences of WRCC in various provinces in China, based on the annual WRCC index of 31 provinces from 2005 to 2021, the WRCC level is divided into Level I, Level II, Level III, and Level IV by using the standard deviation classification method [26]. The results of four grades are listed in

Table 3. WRCC Level Classification Criteria.

Grading Standards	Level I	Level II	Level III	Level IV
Division basis	(0, VB]	(VB, V]	(V, V + B]	(V + B, 1]
	(0, 0.201]	(0.201, 0.263]	(0.263, 0.326]	(0.326, 1]

Note: V in the table is the mean, and B is the standard deviation.

. A higher level indicates a better WRCC, which means better sustainability.

According to the classification principle, the WRCC calculated in 2005, 2010, 2015, and 2021 is divided into four grades. The results are shown in Figure 4, where the darker the color is, the larger the WRCC index is, and the higher the water resource carrying capacity is. As shown in Figure 3, in 2005, among the 31 provinces in China, there were 20, 10, 0, and 1 areas in level I, level II, level III, and level IV, respectively; in 2010, there were 4, 19, 6, and 2 provinces, respectively; in 2015, there were 2, 9, 14, and 6 provinces, respectively; and in 2021, there were 0, 6, 13, and 12, respectively. It can be seen that the number of provinces in China’s WRCC level I has decreased to 0, and the provinces in east China, central China, and south China have increased significantly in level IV. The per capita GDP level in east China, central China, and south China are relatively high. Wastewater discharge has dropped significantly; corporate R&D investment has increased significantly; and experience in water resources management is relatively rich [26]. With developed economic conditions, good environmental governance results, and superior natural environment, these regional WRCC levels are higher. North China and northeast China are located in high-latitude areas where the natural environment is relatively fragile. However, since 2012, the development, usage, and governance of water resources have tended to be improved [25]. WRCC has generally been upgraded from level I to level III, and water resources governance has achieved remarkable results. The northwest and southwest regions are economically underdeveloped, and the natural environment is fragile, so the increase in WRCC is relatively small. Overall, China’s WRCC has generally improved, but there is still a large gap among regions. In the future, it is necessary to strengthen overall planning and regional coordination to achieve sustainable development of water resources.

This pattern is mainly due to the differences in economic development level, natural environment conditions, and the effectiveness of water resource management in different regions. The advantages of economic conditions, environmental governance effectiveness, and natural environment in the eastern, central, and southern regions make the WRCC level in these regions higher. The relatively small increase in WRCC in the northwest and southwest regions is due to the reasons of underdeveloped economy and fragile natural environment. In addition, although the high latitude areas of north China and northeast China have relatively fragile natural environments, since 2012, the effectiveness of water resource management has been remarkable, resulting in an overall increase in WRCC from level I to level III in these regions. In general, different regions have shown different performance and effectiveness in water resource management and governance, leading to differences in WRCC patterns in different regions.

3.2.2. Spatial Change Trend of WRCC in China

First, according to the longitude and latitude of the provincial capitals, the positions of the 31 provinces are quantified and sorted from west to the east and from south to north. The results are shown in Figure 2. The W–E index indicates the results from west to east. The more eastern the province is, the larger the index is. The S–N index indicates the results from south to north. The more northern the province is, the larger the index is. Then, make a scatter plot with the position of each province as the X-axis and the WRCC index level as the Y-axis, and perform linear regression on the WRCC in 2005, 2010, 2015, and 2021, respectively. China’s 31 provinces’ WRCC level evolution trend map in the east–west direction (Figure 5A) and WRCC level evolution trend map in the north–south direction (Figure 5B) are obtained.

It can be seen from Figure 5 that in 2005, the WRCC grades of the 31 provinces in China showed a slow downward trend from west to east, while from 2010 to 2021, the WRCC grades showed an upward trend from west to east, and China’s WRCC grades have increased significantly since 2005. The average grade rose from 1.5 in 2005 to 3 in 2021. In addition, it can be seen from Figure 5A that the increase in the WRCC grade in the eastern region was significantly greater than that in the western region. Observing the evolution trend of WRCC grades in the north–south direction of China (Figure 5B), it was found that from 2005 to 2021, the WRCC grades of 31 provinces in China showed a significant increase from south to north, but the WRCC grades in the north were significantly lower than those in the south. Comparing the two graphs in Figure 5A,B, the north–south direction is relatively flat. Overall, China’s WRCC index differences between the north and the south are smaller than the those between the east and the west.

3.3. Spatial Correlation Analysis of WRCC in China

3.3.1. Global Spatial Autocorrelation Analysis on WRCC in China

Using the software Stata 17.0, the global Moran’s I index is calculated based on the adjacency spatial weight matrix for the 2005–2021 time-section data, in order to reveal the spatial agglomeration of the WRCC index China’s provinces. The results are shown in

Table 4. Global Moran’s I Index of WRCC in China from 2005 to 2021.

Year	Moran’s I	p Value	Z Value
2005	0.131	0.051	1.638
2006	0.160	0.032	1.852
2007	0.162	0.031	1.873
2008	0.160	0.032	1.853
2009	0.233	0.006	2.485
2010	0.216	0.010	2.343
2011	0.183	0.021	2.027
2012	0.249	0.004	2.618
2013	0.166	0.032	1.854
2014	0.190	0.019	2.073
2015	0.271	0.002	2.827
2016	0.291	0.001	3.018
2017	0.244	0.005	2.588
2018	0.223	0.008	2.413
2019	0.248	0.004	2.648
2020	0.269	0.002	2.841
2021	0.257	0.003	2.730

. The five-year global Moran’s I index from 2005 to 2008 and 2013 passed the 1% significance test, but the Z value was less than the critical value of 1.96, indicating that the inter-provincial WRCC distribution in China during this period is random. The global Moran’s I index for 2009–2012 and 2014–2021 all passed the 1% significance test, and the Z value was greater than the critical value of 1.96, indicating that China’s inter-provincial WRCC index has obvious spatial agglomeration characteristics. It can be seen from Table 4 that since 2009, the global Moran’s I index has generally shown a wave-like evolution trend of “declining → rising → declining → rising”, and the agglomeration state is obvious and fluctuating throughout the study period. The Moran’s I index reached a peak of 0.291 in 2016, and the spatial agglomeration phenomenon was the most obvious. In 2021, the Moran’s I index dropped to 0.257, and the spatial agglomeration effect weakened.

3.3.2. Local Spatial Autocorrelation Analysis of WRCC in China

In order to reveal the spatial agglomeration of WRCC in each province, the WRCC scatter diagrams of China in 2005, 2010, 2015, and 2021 were drawn according to the local Moran’s I index and the Moran scatter diagram (Figure 6). Figure 5 shows that in 2005, 2010, 2015, and 2021, 23, 24, 23, and 21 provinces, respectively, were in a state of positive spatial correlation (HH or LL agglomeration). The overall trend was fluctuating, but the change was not large. In 2005 and 2010, the number of HL agglomeration was the same, both in 3 provinces; in 2015 and 2021, the only HL-agglomerated provinces were 5 and 6, respectively. The number of LH agglomeration in 2010 and 2021 was the same, both in 4. In 2005, the number of LH agglomerated provinces was the largest (5), and in 2015, the number of LH agglomerated provinces was the lowest (3).

(1)

The HH agglomeration areas are mainly distributed in southeastern China, including Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Hunan, and Hubei provinces, forming areas with high WRCC index values. This area has expanded significantly after 2010, mainly because most of the provinces in this area are coastal provinces with relatively high rainfall, good economic development, and rapid water pollution control. However, since 2010, provinces in the HH region have gradually decreased. Henan, Hubei, Guangdong, and Guangxi have changed from HH agglomeration to HL agglomeration or LH agglomeration in 2021, which is mainly caused by the increasing gap between regions.

(2)

HL agglomeration areas are very scattered throughout the country, and there are very few provinces in this state. In 2005, HL agglomeration areas were Hubei, Hunan, and Tibet. In 2010, they were Hebei, Heilongjiang, and Tibet. In 2015, they were Hebei and Inner Mongolia, Heilongjiang, Hubei, and Tibet; in 2021, they are Heilongjiang, Henan, Hubei, Guangdong, Sichuan, and Tibet. Tibet has been in a state of HL agglomeration. Shanxi, Chongqing, and Yunnan adjacent to this area are in the state of LL agglomeration, with a tendency of “spreading”. Water resource management should be strengthened to improve the WRCC of the province. Hunan is in the state of HH agglomeration, and it is necessary to strengthen the effective supervision of water resources.

(3)

The spatial distribution of LH agglomeration is relatively stable, which mainly is in Shanghai, Hainan, Guangxi, and other provinces. Due to their closeness to the HH and LL agglomeration areas, they are extremely affected. Jilin LL agglomeration changed to LH agglomeration from 2015 to 2021; Yunnan and Qinghai changed from LH agglomeration to LL agglomeration from 2015 to 2021. Guangxi was LH agglomeration in 2005 and changed to HH agglomeration in 2010–2015, and it transformed into LH agglomeration in 2021. Overall, the differences in economic development levels are an important factor causing changes in LL, HH, and other indicators, reflecting the need for provinces such as Guangxi and Jilin to refer to the experience of neighboring provinces in economic development and formulate corresponding industry and economic development policies.

(4)

The LL agglomeration areas are located in the west and north of China. The number of provinces with LL agglomeration status has not changed during 2010–2015, but the regional scope has been adjusted as a whole. Inner Mongolia has changed from LL agglomeration to HL agglomeration; Jilin has changed from LL agglomeration to LH agglomeration; Yunnan and Qinghai changed from LH agglomeration to LL agglomeration; the number of provinces which are in LL agglomeration state increased by 1 during 2015–2021; Hebei and Inner Mongolia changed from HL agglomeration to LL agglomeration from a regional perspective; Sichuan has changed from LL agglomeration to HL agglomeration. In 2021, both northern China and northwestern China were in a state of LL agglomeration. It shows that the western and northern regions of China form a spatial agglomeration area with low WRCC, and the surrounding areas have a tendency to spread into the areas.

4. Analysis of Influencing Factors of Spatial Differences of WRCC

It can be seen from the above that WRCC varies greatly among provinces in China, which is mainly caused by the differences in WRCC-related indicators. Therefore, this section will use geodetectors to quantitatively identify the impact of each detection factor on the spatial differences of WRCC in China.

Considering that accidental factors in a single year will have adverse effects on the results, in order to avoid this situation, we divide the data into four groups: 2006–2009, 2010–2013, 2014–2017, and 2018–2021. First, we use the natural discontinuity point method in the software ArcGIS to grade each group of indicators. Then, we combine them with the WRCC index of different provinces and calculate the effect intensity (q value) of each indicator by using the geodetector. The results are shown in Figure 7.

The natural break method is a commonly used data classification method that determines classification boundaries based on the natural distribution characteristics of the data, avoiding errors caused by subjective classification [37,48]. In this article, we chose the natural break method as the data classification method because it can better reflect the characteristics of different indicators in the data distribution, while also reducing the impact of human intervention on the results and improving their credibility and stability. In addition, the natural break method is also one of the commonly used data classification methods in ArcGIS software, which has the advantages of simple operation and easy implementation, facilitating our research work.

4.1. Intensity of Influencing Factors of WRCC in China

From Figure 7, the main indicators affecting the spatial differences of the WRCC index have not changed significantly in different periods, but the intensity of each indicator has changed significantly. With the change of time, the effect intensity of industrial wastewater discharge showed a trend of decreasing and increasing; the action intensity of per capita domestic water consumption, surface water supply, effective irrigation area, industrial scientific research expenditure above designated size, and sewage treatment capacity showed an increasing trend.

Specifically, the most important factor causing the spatial differences of WRCC from 2006 to 2009 is the discharge of industrial wastewater, followed by sewage treatment capacity, industrial scientific research expenditure above designated size, and fertilizer application per unit of cultivated land. The main influencing factor from 2010 to 2013 is the sewage treatment capacity, followed by the expenditure on industrial scientific research above designated size and the discharge of industrial wastewater. The results from 2014 to 2017 are similar to the results from 2010 to 2014, but the effect intensity of sewage treatment capacity and industrial scientific research expenditure above designated size continues to rise, while the effect intensity of industrial wastewater discharge continues to decline. From 2018 to 2021, the most important factor causing the spatial differences of WRCC is the expenditure on industrial scientific research above designated size, followed by sewage treatment capacity and industrial wastewater discharge. For the first time, scientific and technological innovation surpasses direct sewage treatment and becomes the primary factor affecting the carrying capacity of water resources.

In summary, since 2006, the main factors affecting the spatial differences of WRCC have been transmitted along the direction of “industrial wastewater discharge → sewage treatment capacity → industrial scientific research expenditure above designated size”. This aspect reflects, on the one hand, China’s environmental protection and the emphasis on curbing sewage discharge; on the other hand, it also reflects the continuous optimization of the industrial structure and the improvement of the economic green level and technological level. At the same time, it can be seen that the intensity of the water supply of surface water resources has increased from 0.267 in 2010–2013 to 0.413 in 2018–2021, which is second only to industrial scientific research expenditures above designated size among all indicators. This is mainly due to the remarkable achievements of large-scale water conservancy projects such as China’s South-to-North Water Diversion Project.

4.2. Multi-Factor Interactive Detection Analysis

Figure 8 shows the interactive detection of driving factors of the spatial and temporal differences of WRCC in China from 2020 to 2021. It can be seen that a total of 209 factor combinations were formed after factor interaction, of which 63.2% showed nonlinear enhancement—that is, the joint effect intensity was greater than the sum of the individual intensity of the two factors; 36.8% showed double-factor enhancement—that is, the joint effect strength was greater than the maximum value in the two factors. For example, the maximum effect intensity of per capita domestic water consumption was 0.165, but the average effect intensity reached 0.425 after interacting with other driving factors, and the interaction intensity with sewage treatment capacity reached 0.724. In addition, the intensity of chemical fertilizer application per unit area of cultivated land was about 0.188, and the intensity of surface water supply was about 0.300, but the interaction intensity between the two has increased to 0.684. It shows that regional differences in WRCC level in China are the result of the joint action of multiple factors. Therefore, when improving China’s WRCC level in the future, it is necessary to comprehensively consider the individual influence and interaction of multiple factors; to ensure the driving position of economic and social development, scientific and technological progress, and industrial transformation and upgrading, on the one hand; and to strengthen the conservation and restoration of water resources to build a water resources-saving social production and consumption system, gradually improving the level of WRCC, on the other hand.

5. Conclusions and Recommendations

5.1. Conclusions

Based on the DPSIR model, this paper constructs the evaluation index system of WRCC in China and analyzes the temporal and spatial evolution characteristics and driving factors of WRCC in various provinces in China from 2005 to 2021. The conclusions are as follows.

(1)

From 2005 to 2021, China’s WRCC showed a continuous upward trend. During the study period, the number of provinces whose WRCCs are in level IV increased from 1 to 12, and the number of level I provinces decreased from 20 to 0. The WRCCs of Tibet, Guangdong, Jiangsu, and Zhejiang are firmly at the forefront of China, while the WRCCs of Ningxia, Qinghai, Gansu, and other provinces rank low.

(2)

The center of WRCC is generally moving to the southeast, indicating that the WRCC in the southeast is steadily improving. The increase is greater than that in the west and north, and the regional gap is increasing. Through the global autocorrelation analysis, it can be seen that the spatial correlation of China’s WRCC index is obvious; the positive correlation characteristics are significant; and the correlation intensity fluctuates. The local autocorrelation analysis shows that the number of cities in China where WRCCs are concentrated in HH and LL dropped from 23 in 2005 to 21 in 2021, indicating that the spatial aggregation characteristics of WRCC continue to weaken.

(3)

The degree of influence of each driving factor on the spatial differences of WRCC in different periods is different. Since 2006, the main factors affecting the spatial differences of WRCC have been along the line of “industrial wastewater discharge → sewage treatment capacity → industrial scientific research expenditure above designated size”. In the direction of industrial scientific research expenditures, the intensity of action of sewage treatment capacity and industrial scientific research expenditures above designated size continued to rise, while the intensity of action of industrial wastewater discharge continued to decline. After any factor interacted, it showed double-factor enhancement or non-linear enhancement, reflecting the complexity characteristics between compound factors and WRCC level.

(4)

The regional differences in China’s WRCC levels are the result of the combined effects of multiple factors, with the interaction between these factors having a greater impact on the overall effect than the effect of individual factors. Specifically, 63.2% of the factors exhibit non-linear enhancement, and 36.8% exhibit bi-factor enhancement. Improving China’s WRCC level requires balancing the individual effects and the interactions of multiple factors, while ensuring economic and social development, technological progress, industrial transformation, and other related factors.

5.2. Recommendations

(1)

Further narrow the development gap between regions and promote coordinated regional economic development. While optimizing regional industrial structure, strengthen cooperation between different regions; promote the flow of resources, technology, and talent; and achieve complementary development between regions.

(2)

Strengthen investment in science and technology and research and development to improve water resource utilization efficiency and reduce waste. Increase support for technological innovation in water resources; encourage innovation by enterprises and teams; strengthen the transformation and application of scientific and technological achievements; and promote the scientific management and efficient utilization of water resources.

(3)

Reduce the amount of water used per unit of industrial value added and the amount of industrial wastewater discharged. This can be achieved through optimizing industrial structure and production processes, promoting water-saving and emission reduction technologies, increasing investment in sewage treatment capacity, and reducing pollutant emissions. At the same time, strengthen supervision of industrial enterprises; strictly enforce environmental laws and standards; and achieve coordinated development of industrial production and environmental protection.

(4)

Build water-saving cities and communities by optimizing urban planning and management, strengthening the management and utilization of urban water resources, promoting green buildings and ecological urban construction, and promoting the sustainable utilization and protection of urban water resources. Encourage and support community residents to actively participate in water-saving activities and actions to jointly achieve water-saving goals.