Introduction

A vital resource for human survival is water. Although the earth is divided into three parts land and seven parts water, the fresh water available for human production and life makes up only 105 trillion cubic meters. Since the 1990s, global floods and droughts have occurred frequently, and the world’s freshwater resources are becoming increasingly scarce. Water pollution is becoming increasingly serious, and the stability of the earth’s ecosystem has been seriously damaged. Population increase and the acceleration of economic growth are increasing the demand for water resources, which is causing a water crisis in many nations and limiting economic growth.

China’s unique geographical makeup gave rise to substantially imbalanced regional water distribution. China’s water supplies are distributed unevenly by both space and time due to the country’s monsoon climate, which makes it harder to wisely develop and use the country’s water resources. China has the sixth-largest water resources in the world, yet due to its massive population, it only uses 2300 cubic meters of water per person per year, or less than a fifth of the global average. China is one of 13 countries with limited access to water, according to the United Nations. The United Nations Environment Programme currently identifies China as a nation with “fragile water resources” (Song et al. 2022). With a 2.35% average annual increase rate from 1987 to 2017, the total amount of wastewater released in China increased from 3.49 1010 tons in 1987 to 7.00 1010 million tons in 2017 (Zhang et al. 2021). In comparison with other industrialized nations, China has made significant efforts to conserve water and treat sewage, but it has not yet reached per capita water savings or lowered per capita sewage output, and the country’s water scarcity and pollution issues remain very serious. Therefore, it is inevitable that the many issues with water resources in China will be further scientifically examined and the best utilization will be realized (Zhang et al. 2021).

In its new stage of development, China has recently realized that, as a result of the country’s fast urbanization and expanding gap link water pollution, water scarcity, and economic development, using water resources sustainably is a crucial component of green and high-quality development (Song et al. 2022). The State Council has recently passed a number of measures in reaction to the serious pollution and lack of water resources. The “Views on Implementing the Strictest Water Resources Management System” and “Decision on Advancing the Reform and Development of Water Resources” were published in 2011 and 2012. They made it clear that focus should be placed on enhancing water usage efficiency, on hastening the development of a society that conserves water, and on preventing and controlling water pollution. In addition, it was noted in the Water Law and the Water Pollution Prevention Law, which underwent successive revisions in 2016 and 2017, that water resources should be used judiciously in order to ensure the control and prevention of water pollution (Song et al. 2022).

With a population making up 43%, and a GDP that accounts for 46.2%, of China, the Yangtze River Economic Belt (YREB) covers various regions of central, eastern, and western China; it includes nine provinces and two cities. It is frequently referred to as China’s golden waterway, and it significantly influences the country’s plan for economic development (Shi et al. 2022). In 2017, the Protection Strategy for the Ecological Environment of the YREB proposed that the discharge of pollutants in the YREB is large, and the discharge of wastewater accounts for more than 40% of the country, with many potential risks. Sulfur dioxide, chemical oxygen requirement, nitrogen, ammonia, volatile organic compound, and nitrogen oxide discharge intensities are 1.5 to 2.0 times higher per unit area than the national average.

The issues of relative scarcity, unequal distribution, and low water resource usage efficiency in the YREB have not been fundamentally analyzed for a very long time. The goal of achieving high-quality development of the YREB is still far off. Due to the acceleration of urbanization, economic and social activities have generated a significant amount of wastewater that has not been effectively prevented and treated and is being discharged in excess of the current water environment carrying capacity. Therefore, this paper will explore the temporal and geographic evolution law of CWRGE in the YREB and its influencing factors in order to increase the effectiveness of water resource usage in the YREB, realize sustainable green development of water resources, and enhance the capacity for environmental water pollution prevention and management.

Literature review

The primary resource on which humans rely for survival and growth is water. Understanding how to utilize water resources effectively significantly enhances people’s lives and fosters economic development. Encouraging the efficient use of water resources and achieving the United Nations’ 2030 Sustainable Development Goal depend heavily on the efficiency of water resource utilization as evaluated by science (Shi et al. 2022). There are two categories of measuring how effectively water resources are being used nowadays: the single-factor index and the full-factor index. For example, Zhao et al. (2016) calculate water resource efficiency as water use per 10,000 yuan of GDP using the single-factor index. Although this method is simple and convenient, it lacks integrity and ignores the joint action of multiple factors involved in the water efficiency calculation process. It is challenging to draw conclusions accurately depicting water resource efficiency. In this paper, the all-cause self-index, which uses comprehensive water use efficiency after considering various production variables’ contributions to the final output and can more accurately reflect the actual situation and regional differences of city water resource green efficiency (CWRGE), is used to measure water resource efficiency.

The following are the primary components of current academic studies on water resource efficiency: the first is the research scope, which focuses mostly on agricultural water resource utilization efficiency (Shi et al. 2022), the effectiveness of domestic urban water resources (Liang et al. 2022), efficiency of industrial water resource use (Zou and Cong 2021), and comprehensive water resource usage efficiency.

The second is calculation. Some scholars have used water footprint (Cao et al. 2021; Sun et al. 2014), the direction distance function model (Zou and Cong 2021), the material flow analysis method (Meng et al. 2020), and the analytic hierarchy process (AHP) (Zhang et al. 2023), which Yang et al. (2019) combine with the Systems Dynamics (SD) model, and a qualitative and quantitative assessment of water resources efficiency is carried out. In addition, the AHP and entropy weight method (EWM) are also utilized to calculate the weights for each metric of water resource usage effectiveness and economic development level (Zhang et al. 2022). Moreover, the calculation methods previously used also include the principal component analysis method (Li et al. 2019). Methods for stochastic frontier analysis, projection pursuit, and data envelopment analysis (DEA) are also used (Chiu and Tang 2022; Morán-Valencia et al. 2023) etc.

DEA is a powerful nonparametric statistical technique for evaluating issues with numerous inputs and outputs (Le et al. 2021). The ability to cope with many index inputs and outputs is a significant advantage over traditional measurement methods, which are limited by complicated cross-relationships between different components and the operational process’ subjective weight assumption (Shi et al. 2022). However, the basic DEA model cannot further compare the effective DUM with an efficiency value of 1, and the consideration of slack variables is insufficient. Additionally, it fails to address the issue of unintended emissions of environmental pollutants. Environmental contaminants are a bad output and go against the DEA’s “maximum output” theory (Ma et al. 2022).

Tone and Tsutsui (2009) put forward the Super-SBM model, which is constructed by combining the traditional DEA model with the SBM model. The model has nonradial and nonangular parameters and takes into account slack variables. This model is applied to calculate efficiency values in various fields. Scholars have previously used this method to measure efficiency in different aspects, such as green energy efficiency (Meng and Qu 2022), scores for each nation’s and region’s efficiency from 2011 to 2015 (Gökgöz and Erkul 2019), and efficiencies in health expenditures that are both static and dynamic (Liu et al. 2019). The use of Super-SBM also provides a new measurement for evaluating the effectiveness of water resources. Previous scholars have used the Super-SBM method to assess the effectiveness of water resources from several perspectives, for example, industrial water resources’ green total factor efficiency (Jin et al. 2019) and comprehensive agricultural water resource use efficiency (Huang et al. 2021b).

The third is the study of the influencing factors of efficiency in water resource use. Previously, researchers from both home and abroad have examined a variety of factors that can affect how effectively water resources are used, including technical innovation (Jin et al. 2019), environmental supervision (Jin et al. 2019; Wang and Wang 2021), foreign investment (Song et al. 2022), industrial development level (Song et al. 2022), population density (Song et al. 2022), water structure (Song et al. 2022), natural resource endowment (Zou and Cong 2021), level of economic development (Zou and Cong 2021), property right structure (Ma et al. 2022), financial agglomeration (Qu et al. 2020), water rights trading policy (Chen et al. 2021a), digital economy (Fu 2022), marketization (Chen et al. 2019), and urbanization (Wang et al. 2022). It is worth mentioning that Jin et al. (2019) utilized education level, progress in the industry, and openness as control variables, while regulation of environment, technological innovation, and their interplay were specified as fundamental variables. How these two variables affect China’s 30 provinces and cities’ green total factor efficiency (GTFE) of industrial water supplies was investigated.

The fourth is the investigation of how effectively water resources are used in diverse spatial and temporal contexts, which academics have examined from the national, provincial (Li et al. 2022; Wang et al. 2019), river basin (Guan et al. 2016; Jing et al. 2022; Yang and Liu 2014), city (Song et al. 2022; Xie et al. 2021), and regional (Bao et al. 2023) levels. At the same time, relevant research has been carried out at different levels. The main methods used include the exploring spatial data analysis (ESDA) model (Wang et al. 2019), spatial error model (Chen et al. 2019), the social network analysis method (Shi et al. 2022), the Integrated Water Ecological Footprint Model (Jing et al. 2022), and the classification DEA method (He et al. 2020).

At present, Moran’s I is widely used in the spatial analysis of data. Moran’s I, which includes the global autocorrelation index and the local autocorrelation index, is a crucial research indicator to examine the potential dependency between the observed data of variables in the same distribution area. This approach has been used in numerous fields, including to analyze the spatial effect of different population sizes on prevalence rate (Assuncao and Reis 1999) and to analyze the spatial dependence of household solid waste (HSW) components (Cheniti et al. 2021). The spatial impact of regional industrial water resource utilization efficiency (IWRUE) is tested using global Moran’s I by Zou and Cong (2021). The overall Moran’s I value shows a strong spatial correlation between provincial IWRUEs, with neighboring provinces frequently having comparable IWRUEs. The worldwide Moran index and the Gini coefficient are used to calculate the spatial distribution and water use efficiency of cities in the YREB, and they are classified according to unequal spatial distribution (Babuna et al. 2020).

While the technology for using the DEA model to study efficiency in water resource use is currently fairly developed, the majority of the research is mainly concerned with the national and provincial levels, but the research on efficiency of water resource use in particular regions, such as the YREB, is insufficient. This can easily be seen by searching through the domestic and foreign literature. In addition, at present, most studies only consider input and output. Few studies relate sewage discharge, wastewater discharge volume, chemical oxygen demand, etc. as undesired outputs of the efficiency of water resource use. In addition, most of the literature does not consider the spatial evolution of the efficiency water resource use in the YREB. Based on this, the YREB is the subject of this essay’s research, based on panel data from 2006 to 2021, as well as on the Super-SBM model, a time series analysis, and a spatial correlation evaluation of the CWRGE of 108 cities in the YREB. Then, a Tobit model is used to empirically test the influencing factors of CWRGE.

Compared with previous studies, the innovation of this study is mainly reflected in three aspects: first, existing studies mainly focus on the efficiency of water resource use, but relatively few involve the green efficiency of water resources. Therefore, this paper constructs a CWRGE index system and uses industrial wastewater discharge as an undesirable output, making the CWRGE measurement more accurate. Second, the existing research mainly studies water resource efficiency from the time dimension, and the discussion of its spatial differentiation characteristics is insufficient. In this paper, ArcGIS is used to discuss the spatial differentiation characteristics of CWRGE in the YREB, which helps to visualize the results of the CWRGE. Third, the previous literature mainly focuses on the provincial or national levels, and there are few studies on prefectures and levels above cities, especially river basin cities. This paper takes 108 cities in the YREB as its research object, which is a breakthrough in research objects.

How to improve the green efficiency of urban water resources has become the key to solving the urban water shortage. Based on this, the main contribution of this paper is the measurement of CWRGE in the YREB using the Super-SBM model and the exploration of its influencing factors using the Tobit model. On this basis, this paper puts forward different water-saving and water pollution control plans for cities in the upper, middle, and lower reaches of basins, as well as cities of different sizes, including large and medium. At the same time, for the different subjects, countermeasures and suggestions for improving CWRGE are put forward from the perspectives of accelerating industrial transformation, strengthening the application of water-saving technology, improving the standard water-saving system, popularizing citizens’ environmental protection and water-saving education, etc., which provides a certain extent reference for putting an end to the water inefficiency of watershed cities and strengthening environmental water protection.

Model and data sources

Super-SBM model

At present, the DEA model is frequently employed to assess the effectiveness of water resources. The basic DEA model cannot compare the effective DMU with an efficiency value of 1, and it does not sufficiently consider the slack variables. Moreover, it cannot solve the problem of environmental pollutants as an accidental output, and its measurement results only come from radials and angles with low accuracy. Tone and Tsutsui (2009) combine the traditional DEA model and SBM model to construct the Super-SBM model.

$$\left\{\begin{array}{c}\min \rho =\frac{1+\frac{1}{h}\sum\limits_{i=1}^h{s}_i^{-}/{x}_{ik}}{1-\frac{1}{s_1+{s}_2}\left(\sum\limits_{r=1}^{s_1}{s}_r^{+}/{y}_{rk}+\sum\limits_{t=1}^{s_2}{s}_t^{b^{-}}/{b}_{rk}\right)}\\ {}\sum\limits_{j=1,j\ne k}^n{x}_{ij}{\lambda}_i-{s}_i^{-}\le {x}_{ik}\\ {}\sum\limits_{j=1,j\ne k}^n{y}_{rj}{\lambda}_i+{s}_r^{+}\ge {y}_{rk}\\ {}\sum\limits_{j=1,j\ne k}^n{b}_{tj}{\lambda}_j-{\textrm{s}}_t^{b^{-}}\ge {b}_{tk}\\ {}{\lambda}_j,{s}^{+},{s}^{-}\ge 0\\ {}i=1,2,\dots, h;r=1,2,\dots, s;j=1,2,\dots, n\left(j\ne k\right)\end{array}\right.$$
(1)

The objective function value of ρ is the efficiency value of DMU; h represents the number of input indicators; λj represents the jth constraint; s represents the number of constraints; n represents the number of decision units; \({s}_i^{-}\) and \({s}_r^{+}\) represent the relaxation variable of input and output; xikrepresents the i input of the kth decision unit; and yrk represents the r output of the kth unit. \({s}_i^{-}/{x}_{ik}\) and \({s}_r^{+}/{y}_{rk}\) respectively represent the inefficiency of the i-type input and the r-type output of the kth decision-making unit; \(\frac{1}{h}\sum\limits_{i=1}^h{s}_i^{-}/{x}_{ik}\) is the inefficiency of this DUM’s input; \({s}_1^{-}\) indicates the desired number of output indicators; \({s}_2^{-}\)represents the number of undesirable output indicators; and \({s}_t^{b^{-}}\) represents the relaxation variable of the undesired output.

This nonradial and nonangular Super-SBM model based on undesired outputs has many advantages. First, it effectively solves the problem of input and output relaxation caused by radial and angle selection (Shah et al. 2022). Second, it solves the problem of the SBM model not being able to distinguish decision units. It can further measure the efficiency of the SBM-effective DMU, realize the complete ranking of DMU, and produce more objective and accurate efficiency measurement results (Li and Zhang 2022). Third, the model has controllable and uncontrollable undesirable outputs for each production unit (Shah et al. 2022). It solves the problem of environmental pollutants being an accidental output, and it is especially suitable for dealing with the efficiency measure considering the undesired output. Therefore, compared with other DEA models, the Super-SBM model with unexpected output can better represent the nature of regional water resource efficiency evaluation.

Global spatial autocorrelation

A spatial statistical technique that can describe the regional structure of space is called spatial autocorrelation analysis. Global spatial autocorrelation represents spatial dependence within a certain total spatial range. The global Moran index can be used to determine whether there are spatial aggregation characteristics or abnormal values in a certain spatial range. The most often employed correlated index is Moran’s I, so we use Moran’s I to quantitatively analyze the spatial distribution characteristics of the YREB.

$$I=\frac{M}{S_0}\times \frac{\sum\limits_{i=1}^m\sum\limits_{j=1}^m{w}_{ij}\left({x}_i-\overline{x}\right)\left({x}_j-\overline{x}\right)}{\sum\limits_{i=1}^m{\left({x}_i-\overline{x}\right)}^2}$$
(2)

where I stands for Moran index; m represents the cities there are, and also m research units; xi and xjrepresent i and j research units’ attribute values, respectively;\(\overline{x}\) represents the average value of m spatial unit sample attributes; S0 represents the total of the spatial weight matrix’s elements; and wij is research units i and j’s spatial weight matrix. Moran’s I ranges from −1 to 1, and when it is between 0 and 1, it denotes a spatially positive correlation; when it is between −1 and 0, it denotes a spatially negative correlation; and when it is 0, there is not a connection. All computations and maps are generated using ArcGIS 10.2.

Local spatial autocorrelation

In order to express spatial heterogeneity, local spatial autocorrelation is used to quantify how closely each spatial unit adheres to the overall general pattern. Compared with global autocorrelation, it can explore the specific location, internal agglomeration, and connection of its spatial agglomeration. Local autocorrelation is defined as follows:

$$LocalI=\frac{{\left({x}_i-\overline{x}\right)}^2\sum\limits_{j=1}^m{w}_{ij}{\left({x}_j-\overline{x}\right)}^2}{S^2}$$
(3)

where I represents Moran index, and m is the quantity of cities; xiand xj represent i and j research units’ attribute values, respectively; \(\overline{x}\) represents the average value of m spatial unit sample attributes; wij is the spatial weight matrix of research units i and j; and S2 represents the standard deviation, which explains how a community and its neighboring cities are related. Positive local I denotes either high-high (H-H) or low-low (L-L) polymerization types; negative local I designates either a low-high (L-H) or a high-low (H-L) polymerization type. ArcGIS 10.2 is used for all computations and images.

Tobit regression model

Some factors may impact the CWRGE in the YREB. A model with dependent variables that satisfy certain requirements is called a Tobit model. Tobit model, also known as sample selection model and finite dependent variable model, is a kind of finite dependent variable regression. It is good at dealing with constraints within a certain range or with truncated dependent variables (Wang et al. 2018). In this paper, the efficiency value belongs to the typical broken tail. If the regression result of traditional least squares is biased to 0, the error is large. In this case, Tobit is the most suitable choice (Xue et al. 2020). So, this study makes use of the Tobit model to investigate how various factors in the YREB affect the CWRGE. The Tobit model’s formula is as follows:

$$y=\left\{\begin{array}{c} yi-\beta 0+\sum\limits_{i=1}^{\textrm{m}}\beta iXi+\varepsilon, yi>0\\ {}0,y\le 0\end{array}\right.$$
(4)

where yi represents the ith decision-making unit’s efficiency value; β0 is the intercept; βireflects each impacting factor’s coefficient; m represents the number of CWRGE influencing factors; Xi stands for various influencing factors; and ε is a random error direction that is consistent with normal distribution.

Selection of indicators and data

The time range of each index in this paper is from 2006 to 2021. All the data come from the China Statistical Yearbook, the China Urban Construction Yearbook, and the local statistical yearbooks and local statistical bulletins of cities along the YREB. A small number of missing data was found using the interpolation method.

Input-output index system

The indicators selected in the empirical study mainly include input indicators, output indicators, and unexpected output indicators. Water is a natural resource that must be supplemented with other manufacturing elements because it seldom ever produces anything on its own (Song et al. 2022). The input indicators are primarily selected from four aspects:water resources, capital, labor, and municipal construction. The output index is indicated by the GDP of the municipal district. Whether it is in daily life, agricultural work, or industrial life, wastewater will inevitably be produced, so the amount of industrial wastewater is selected for the unexpected output. Table 1 lists the particular indicators.

Table 1 Input and output index system
Full size table

At the same time, based on the benchmark price from 2006, this study adjusts the investment in assets fixed in the input indicators and the GDP of municipal districts in the output indicators by using a logarithm in order to avoid the impact of price fluctuations and ensure comparability and consistency across indicators throughout time.

Water resources

The total water supply directly affects the efficiency of water utilization (Cheng et al. 2023). Insufficient water supply limits industrial and agricultural production and economic and social development (Bell et al. 2016). This paper selects the water supply of municipal districts as one of the input indicators. Thus, the total amount of water given to municipal districts per year is referred to as the water supply of municipal districts. Effective water supply and water leakage are among them and can comprehensively reflect CWRGE.

Capital

The investment in fixed assets largely represents the demand for water resources as a result of economic development and urbanization (Balha et al. 2020). This paper selects municipal districts’ investment into fixed assets to describe the capital investment index. The amount of investment into fixed assets represents the workload of building and purchasing fixed assets and the related expenses, which can represent the capital investment of cities to a certain extent.

Labor force

The economic development of a city cannot be separated from its labor force, which is an important driving factor for economic development. Labor force plays an important role in the management of water resource utilization (Cheng et al. 2023). The number of employed citizens is an important input factor for evaluating the efficiency of water resource use (Sheng and Qiu 2022). In this study, the number of employed people in municipal districts is defined as labor input.

Municipal construction

This study introduces municipal construction into the index system, which represents government management. The construction of water supply and drainage projects directly affects the quality of people’s lives and the normal operation of urban functions, and it has a certain degree of influence on urban environmental protection (Li and Li 2020). Using the density of drainage pipes in municipal districts to describe municipal construction also reflects the government’s investment in water resource construction to some extent.

Economic development (output indicator)

The purpose of improving the efficiency of water use is to promote sustainable development of society and the economy, and GDP is the key index to measure the sustainable development of the economy (Cheng et al. 2023). The result of a city’s production activities throughout the course of a given period is represented by the total output value of its municipal districts, which is a more precise way to gauge the state of the national economy. To describe the urban economic development (EDL) in this study, the total production value of municipal districts is used. Introducing the unexpected output index into the input-output index system can ensure more accurate and reasonable CWRGE results.

Environmental pollution (unexpected output)

When cities use water resources for industrial activities, wastewater will inevitably be produced, and most of the wastewater discharge in urban life comes from industrial activities. China’s 12th Five-Year Plan for energy conservation and emission reduction includes sulfur dioxide and wastewater emissions as key indicators for environmental regulation (Meng and Qu 2022). Therefore, this study selects the wastewater discharge of municipal districts as a measure of unanticipated output.

Influencing factors and indicators

In previous studies, scholars believed that the utilization of water resources was a function of many factors (Song et al. 2022). On the effectiveness of using water resources, it is thought that economic growth, foreign direct investment, industrial development level, the intensity of government environmental control, population density, and water usage structure all have substantial and stable influences. Some scholars believe that financial aggregation (Qu et al. 2020), water resource endowment (Zou and Cong 2021), degree of marketization (Chen et al. 2019), and so on will also affect the utilization efficiency of water resources, considering the principles of data availability, comparability, and scientificity. Various levels of influencing factors were selected from the following aspects to explore their impact on CWRGE: environmental factor, scientific and technological factor, industrial factor, water consumption factor, economic factor, and open factor. Specific indicators are described in Table 2. It is worth noting that the indicators used throughout, such as GDP, investment in science and technology, and foreign direct investment, are all reduced based on 2006 to eliminate the influence of price factors. The per capita GDP, the number of industrial enterprises, and the investment in science and technology all take natural logarithm, which brings the curve of discrete data closer to smooth curve so as to solve the problem that the original value coefficient is too small to display.

Table 2 Influencing factors of CWRGE
Full size table

Environmental regulation

Water resources, a public resource on which human beings depend for survival and development, inevitably produce sewage in the course of use, among which domestic sewage is an important part. With the rapid development of urbanization and industrialization, water resources are becoming increasingly scarce, but the construction of sewage treatment plants to treat domestic sewage can ensure the sustainable development of water resources (Wu et al. 2023). Sewage treatment rate is a key measure of ecological sustainable development, covering urban health, people’s living welfare, water resource recycling, and so on (Cheng et al. 2023). As the main body of macrocontrol, the government plays a great role in environmental regulation (ER). This paper uses the treatment rate of urban domestic sewage to describe ER, which can reflect the government’s ER to some extent.

Economic factor: economic development

This study describes the state of EDL according to GDP per capita. IWRUE may be significantly impacted by government funding, numerous infrastructural investments, and technological advancements that are intimately tied to the size of regional economic growth (Ma et al. 2022). In the early stage of economic development, people often pursue rapid growth and adopt extensive development methods, and the efficiency of water resource utilization declines. When the economy develops to a certain extent, resource shortage and environmental pollution hinder further development of the economy, and people have to consider the use of resource-saving and environment-friendly production modes to improve the efficiency of resource utilization (Song et al. 2018). On the one hand, enhancement of the degree of economic growth is favorable to the advancement of technology and industry, thus increasing the effectiveness of water resource use. On the other hand, after the economic development level reaches a certain level, there will be problems such as imbalance between supply and demand and uneven market allocation of water resources that will prevent the CWRGE from improving. Thus, how the EDL level affects the sustainability of the YREB’s CWRGE needs to be tested.

Industrial scale

Number of industrial firms is used in this study to describe the size of industry. The quantity of industrial businesses in a city can be a good indicator of how industrialized a region is. On the one hand, the growth of mechanized scale can encourage the use of water resources, boosting economic production and enhancing the environment-friendly effectiveness of water resources. On the other hand, industrial development hinders improvement of the environmental performance of water (Zhang et al. 2020). The growth in industrial enterprises and the expansion of industrial scale (IS) may cause more water resources to be wasted and more industrial wastewater and sewage to be produced, which may cause environmental pollution, decreasing the water resources’ ability to be used sustainably. Therefore, this study takes IS as an influencing factor and explores its influence on CWRGE.

Water consumption structure (WS)

A rational water use structure is very important to promoting regional economic development (Cheng et al. 2023). Water resources are used in many fields, among which residential water use is an important part. This paper describes WS in terms of the ratio of residential water use to total water supply, so this study explores the influence of residential water use on CWRGE.

Technology investment

This paper studies the financial expenditure of science and technology to describe the input of science and technology. The damaging effects of economic growth on the environment might eventually lessen if industries embrace cleaner and more energy-efficient industrial technology (Sharma et al. 2021). Industrial water-saving technology may simultaneously increase the pace of water resource circulation and the effectiveness of the entire industrial water circulation system (Ma et al. 2022). On the one hand, scientific and technological innovation provides more effective water-saving equipment to improve water efficiency. On the other hand, technological innovation in production increases water pollution emissions, which is not conducive to improving water use efficiency (Cheng et al. 2023). Therefore, it is necessary to further explore the impact of science and technology expenditure on CWRGE.

Opening up

In this study, opening up (OU) is defined as the actual amount of foreign direct investment. This paper uses the treatment method of Liang et al. (2021) for reference and chooses the proportion of foreign direct investment to GDP to represent OU. With continuous promotion of the reform and OU policy, China is constantly absorbing foreign investment. On the one hand, more and more foreign investment can enliven China’s market economy, promote regional economic development, and introduce advanced green water resource technology. On the other side, it might also add to environmental load, raising the amount of effluent and sewage discharged from industry, leading to the issue of “polluting paradise.”

Empirical analysis and discussion

Based on the analysis approach, this research uses MaxDEA8.0 to determine 108 cities’ CWRGE in the YREB from 2006 to 2021. Table 3 displays the findings after analysis of the entire region and all the cities based on the calculation results. The YREB’s water resources have an overall green efficiency of 0.7271, which is lower than 1 and not as effective as DEA. This conclusion is consistent with Huang et al. (2021a).

Table 3 Measurement results of CWRGE in YREB
Full size table

Characteristics of evolution over time

Analysis of overall evolution trend

The CWRGE in the YREB generally shows a trend of “first decreasing and then increasing, then decreasing and then increasing” and shows a “W”-shaped evolution law. Overall, there is a rising trend of volatility, which is consistent with Xiang et al. (2023). However, Qing and Nie (2022) show that YREB water resource efficiency presents an increasing trend year by year, which may be due to the fact that their research object is provincial water resource efficiency, and the differences in index system and time span may have also led to the differences in the research results.

The highest point is 0.7937 in 2021, indicating that, from 2006 to 2021, the overall CWRGE in the YREB showed a fluctuating upward tendency. As shown in Fig. 1a and b, we can divide its evolutionary trend into four stages. The initial period runs from 2006 to 2011, during which the CWRGE in the YREB decreased from 0.7165 to 0.6922, a decrease of 3.39%, indicating that the water resource efficiency declined to some extent in this stage.

Fig. 1
figure 1

a Changes in CWRGE in the different scales of the YREB, b distribution of CWRGE in the YREB from 2006 to 2021, c changes of CWRGE in different regions of the YREB, d changes in the CWRGE level in the YREB, e changes in CWRGE in different regions of the YREB

Full size image

The second phase is 2011 to 2012. At this stage, the CWRGE in the YREB increased from 0.6922 to 0.7238, an increase of 4.56%. Compared with the previous stage, the CWRGE greatly increased in this year. The reason may be that, in 2011, the State Council issued the No. 1 Document of the Central Committee, which proposed implementation of the strictest water resource management system and clearly pointed out the new concept of “building a water-saving society with water as the target.”

However, due to the economic slowdown caused by water-saving technological transformation and pollution restriction in various places, in the third stage, from 2012 to 2016, the CWRGE in the YREB decreased from 0.7238 to 0.7166, a decrease of 0.99%. The decline in this phase was smaller than the previous increase, demonstrating that the strictest plan for managing water resources, put forward in 2011, produced outstanding results.

The fourth stage spans the years 2016 through 2021, and during this time, the CWRGE in the YREB attained a steady growth trend. This trend reflected the new idea of “looking at the connection between environmental protection and significant development from the viewpoint of ecological civilization, as well as the impact of governmental environmental regulatory policies like “to step up conservation of the Yangtze River and stop its over development.”

Time evolution trend analysis of different city scales

There may be a discrepancy in the CWRGE of various cities as a result of various economic and environmental and other developments. Due to this, according to the Notice on Adjusting the Criteria for Classifying the Size of Cities, cities in the YREB are split into three groups based on the number of permanent residents, namely small- and medium-sized cities (populations less than one million), big cities (populations of one million to five million), and mega cities (populations greater than five million). According to the populations in 2021, at the conclusion of the study’s time period, the YREB consisted of 6 megacities, 63 bigcities, and 39 small- and medium-sized cities. The CWRGE in megacities is primarily centered around 1, as seen in Fig. 1c, with a highest average value of 0.968, followed by big cities and small- and medium-sized cities. This demonstrates that the degree of urban development in megacities relatively high and that economy and technology have reached their full potential. This encourages the equitable distribution of water resources and raises the level of water resources’ environmental efficiency. However, the growth of small- and medium-sized cities is generally behind the other two types of cities, and these cities are impacted by things including bad geographic position, large resource consumption, backward production technology, serious pollution discharge, etc., as well as inadequate ecological development, poor urban economic and social growth, and inadequate water resource utilization, all of which contribute to these cities’ low efficiency ratings.

The average level is generally W-shaped from the standpoint of time evolution, with the time development law of “first dropping and then rising, then falling and then increasing” (as illustrated in Fig. 1a, c, and e). Figure 1e shows that the CWRGE of megacities far outpaces the level of the overall level, while the difference between the CWRGE of big cities and the overall level is small and the CWRGE of small- and medium-sized cities is lower than the overall level. We refer to the years 2006 to 2011 as the “shaking period” from the perspective of temporal progression. During this period, the CWRGEs of megacities, big cities, and small- and medium-sized cities all showed an unstable downward trend. The years 2011 to 2012 are referred to as the “fast growth period” because, during this time, the overall CWRGE rapidly rose. This may have been a positive result of the enactment of the strictest water management system. However, CWRGE in big cities also reached its peak in 2012.

Big cities can better implement the relevant policies for water resources by relying on advanced technology and scientific management. The period of 2012–2021 is referred to as the “fluctuating rise period.” During this period, megacities, big cities, and small- and medium-sized cities achieved different degrees of growth. The green water resource efficiency of megacities increased from 0.9375 to 1.0225, a growth rate of 9.01%, and big cities’ growth rate went up from 0.7443 to 0.8157, a growth rate of 9.59%. For small- and medium-sized cities, CWRGE fluctuated from the highest point of 0.6577 to 0.7230, a growth rate of 9.93%. Therefore, from 2012 to 2021, the overall efficiency of green water resources increased. This shows that the efficiency of small- and medium-sized cities in this period of time greatly improved.

Time evolution analysis of different regions

In order to further study the regional evolution characteristics of CWRGE, this paper divides the YREB into upstream, midstream, and downstream sections by provinces. Upstream includes 31 cities in Chongqing, Sichuan, Guizhou, and Yunnan; midstream includes 36 cities in Jiangxi, Hubei, and Hunan; and downstream includes 41 cities in Shanghai, Jiangsu, Zhejiang, and Anhui. Figure 1c illustrates the geographical pattern of the highest CWRGE in the lower reaches, followed by the upper reaches, and the lowest in the middle reaches; this is consistent with Huang et al. (2021a), Xue et al. (2020), and Li et al. (2023).

The average CWRGE of the upper, middle, and lower reaches was 0.736, 0.691, and 0.752, respectively. Many regions exhibited various trends in CWRGE due to variances in geographical orientation, industrial structure, and economic foundation. The CWRGE was better in the lower reaches of the YREB than in the middle and upper reaches due to a strong financial base, cutting-edge science and technology, and a high degree of openness to the outside world. Even though the upper reaches of the YREB do not have the same economic strength as the lower reaches, they actively implement the state’s green development policy, and there are not many polluting industries there; this prevents upstream cities from achieving economic development at the expense of the environment.

Especially since the implementation of the “western development” strategy, cities in the interior reaches have vigorously developed environmental protection and energy-saving industries, and CWRGE has been significantly improved. The midstream is the main rice-growing area in China, and its agricultural freshwater consumption is very large. The middle reaches are also the most densely populated areas in China, so domestic freshwater consumption and urban domestic sewage discharge are also significantly higher than the other two major areas of the YREB (Xue et al. 2020). Lastly, the middle reaches of the YREB are traditional industrial bases, and most of these industries include steel, chemicals, nonferrous metals, etc., which produce a lot of environmental pollutants. Therefore, the CWRGE in the middle reaches is the lowest.

The average level is generally W-shaped from the standpoint of time evolution, with the time development law of “first dropping and then rising, then falling and then increasing” (as illustrated in Fig. 1a, c). In particular, the “fluctuation drop phase” runs from 2006 to 2011, during which time the overall efficiency of green water resources fell from 0.7165 to 0.6922, with a decrease of 0.0243, or 3.39%, of which the upstream and downstream decreased by 3.29% and 6.98%, respectively. However, in this stage, the midstream increased slightly by 1.19%. The middle reaches showed a downward trend from 2006 to 2009, which was consistent with the overall trend, and an upward trend from 2009 to 2011. The main reason for this change was that Huaihua, Loudi, and other cities paid attention to water saving and pollution prevention, which led to the rapid growth of CWRGE in these cities at this stage, driving the overall improvement of CWRGE in the middle reaches. But, compared with other regions, this growth rate is very small, so it did not change the overall downward trend.

The period from 2011 to 2012 was a “rapid rising period.” The overall average value increased from 0.6922 to 0.7238, an increase of 4.57%. The increases in the upper, middle, and lower reaches were 6.69%, 0.003%, and 6.77%, respectively. This increase may be attributed to the adoption of the strictest water resource management system in 2011, which established three systems: the red line of water efficiency control, the red line of water function assessment, and the red line of water quality assessment. The system had a favorable response according to the measurements of water resources. In contrast, the middle reaches’ CWRGE slowly grew, which shows that the real time and promotion of the system in the middle reaches still needs to be improved.

The period from 2012 to 2016 is the third stage. It is called the “fluctuating decline period.” In this stage, there was generally a downward fluctuating trend. At this stage, the upstream is relatively stable, the middle reaches rise slightly, and the downstream shows a significant declining trend. The decrease rate is 5.61%.

The period from 2016 to 2021 is called the fluctuating rise period, during which the overall trend of fluctuation rises with an increase of 10.76%. Upstream, downstream, and midstream all show varying degrees of growth during this period at rates of 6.53%, 15.69%, and 9.88%, respectively. Among them, the midstream shows the largest increase, and the change range is more obvious than that in the other two regions. Figure 1d shows that the middle reaches of the Yangtze River experienced a strong upward trend in CWRGE that mostly began in 2017. By analyzing the expected output and the unexpected output indicators, industrial wastewater shows a substantial decline between 2017 and 2021. At the same time, the GDP in the middle reaches of the YREB also increases sharply. This demonstrates that the economy in the middle reaches experienced rapid development without considering environmental contamination as a cost, which considerably increased CWRGE.

Analysis of CWRGE in typical cities

From the vantage point of cities, there are 45 with a CWRGE exceeding the average value of the YREB, and only seven with effective DEA, namely Shanghai, Yuxi, Changsha, Changde, Chnagzhou, Hangzhou, and Xuzhou (see Table 4). Shanghai has the highest average CWRGE, a value of 1.3562; this is consistent with Xue et al. (2020). As can be seen from Table 4, the cities that have achieved DEA efficiency are mainly gathered downstream, and most of them are megacities or large cities.

Table 4 Data envelopment analysis effective region of CWRGE in the YREB
Full size table

These cities have superior geographical location, relatively high level of economic development, more advanced water-saving technology and sewage treatment technology, and government support for scientific management and utilization of water resources, all of which lead to higher CWRGE.

Yuxi City ranks second in green water resource efficiency; it is worth mentioning that Yuxi is the only small- and medium-sized city with DEA efficiency exceeding 1. Moreover, it is the only city on the upper Yangtze River. Yuxi, with a GDP ranking second in Yunnan Province, is a crucial gateway to South Asia and Southeast Asia in the “belt and road initiative” strategy. It is also a crucial area for economic development in Yunnan Province. Not only is its economic progress more obvious but so is its high concentration of water industries, its ongoing improvement of reuse rates, and its ongoing expansion of sewage treatment capacity, which all contribute to the higher CWRGE.

Table 5 shows 10 cities with low CWRGE, namely Puer, Huanggang, Lincang, Xiaogan, Yichun, Shangrao, Ji’an, Ganzhou, Shaoyang, and Jingzhou. Most of them are distributed in the middle and upper reaches of the YREB; from the viewpoint of urban types, most are small- and medium-sized cities, a few are big cities, and there are no megacities. These cities are affected by factors such as poor geographical location, inadequate ecological environment construction, slow urban economic and social development, backward production capacity technology, large resource consumption, and high pollution degree; the overall CWRGE is low.

Table 5 Cities with low CWRGE in the YREB
Full size table

Spatial evolution analysis

Spatial correlation analysis of CWRGE

In order to further explore the spatial differentiation characteristics of CWRGE in the YREB, geographical correlation analysis was performed using GeoDa software. In order to show the stage change of spatial correlation, this study selects the data from 2006, 2011, 2016, and 2021 for analysis and uses the global autocorrelation formula to calculate the global Moran index of CWRGE in the YREB in typical years (Fig. 2).

Fig. 2
figure 2

Global Moran scatterplot of CWRGE in the YREB. Note: y—CWRGE; lagged y—lagging behind CWRGE

Full size image

In Fig. 2, it can be seen that the numerical distribution of the global Moran index of CWRGE in the YREB is between −0.018 and 0.198, and most numbers are greater than 0, which shows that the global spatial correlation of CWRGE in the YREB is mainly positively correlated in space; this is basically consistent with Deng and Zhang et al. (2022). It shows a fluctuating downward trend and tends toward 0, which shows that the spatial correlation of CWRGE in the YREB is weakening with time, and it will be −0.018 in 2021, a weak negative correlation; there are some differences with (Deng and Zhang 2022), which may be due to the different selected index system and time span.

The global Moran index scatterplot is divided into four quadrants. Quadrants I–IV represent H-H aggregation, H-L outliers, L-L aggregation, and L-H outliers, respectively. In Fig. 2, more points are distributed in Quadrants I and III, indicating that the low-value and high-value cities with efficiency have an agglomeration effect on the adjacent low-value and high-value cities. However, the number of points being divided into the second and fourth quadrants is relatively small; this is consistent with Deng and Zhang et al. (2022).

The global Moran index measures the spatial correlation as a whole and can only discern whether there is agglomeration or abnormal value in the space, which only reflects the average level of spatial correlation between the city and surrounding cities. The LISA spatial form is drawn in this study using local spatial autocorrelation analysis at the level of 0.05 significance in order to better identify the precise site of agglomeration or abnormal characteristics of the CWRGE (as shown in Fig. 3). The properties of local spatial correlation show that the green efficiency of water resources in the YREB mainly has two forms of positive spatial agglomeration (H-H type and L-L type); that is, the areas with higher (or lower) green efficiency of water resources tend to be connected with the surrounding areas in space.

Fig. 3
figure 3

Local spatial autocorrelation diagram of CWRGE in YREB

Full size image

As shown in Fig. 3, L-L aggregation experienced a trend of falling fluctuation, and its spatial distribution is significantly different. L-L aggregation basically disappeared in 2021. In 2006, it was mainly distributed in Yunnan Province and Jiangxi Province. In 2011, it was mainly allocated in Jiangxi Province, Hunan Province, and Yunnan Province, and in 2016, it was distributed in Anhui Province, Jiangxi Province, Hunan Province, and Yunnan Province and mainly in the middle reaches of the YREB, indicating that the overall CWRGE in the middle reaches showed low-value aggregation because most cities in the middle reaches had poor geographical location.

Moreover, many cities mainly produce traditional steel, chemicals, nonferrous metals, etc., and their low investment in scientific and technological innovation, backward production management technology, and serious industrial pollution lead to low CWRGE. By 2021, the number of cities with L-L aggregation greatly reduce and exist mainly in Yunnan Province, which shows that the overall CWRGE in the middle reaches improves. From 2006 to 2021, L-L clusters exist in Yunnan Province, which may be due to the fact that most areas of Yunnan are mountainous, with severe soil erosion and frequent natural disasters, and agriculture is the main economic source. Its economic development is slow, which leads to low CWRGE.

However, the number of cities with H-H clusters is relatively small. The cities with H-H concentration show an increasing trend at first, then decreasing (Fig. 3). In 2006, there were two cities with H-H concentration, Suzhou and Kunming, among which Kunming, the capital city of Yunnan, had a useful economic development—the province invested in more technology and resources—so it showed a positive spatial correlation with high efficiency. However, Suzhou enjoys a superior geographical position. Its labor, water resources, capital, and other resources are fairly allocated, and it places great importance on ecology and the environment, which results in higher efficiency of green water resources. Its economic development and technological development are at an advanced level throughout the entire city.

In 2011, there were two cities showing H-H agglomeration, namely Nantong and Zhangjiajie. By 2016, only Nantong presented H-H aggregation. By 2021, there were two cities showing H-H agglomeration, namely Nantong and Ziyang. Among them, Nantong is in the H-H cluster for most of the time because it has a reasonable industrial structure, and attaches importance to environmental protection, so its efficiency value is relatively high. Moreover, it is adjacent to cities with high efficiency values, such as Jiangsu and Shanghai, so it presents H-H aggregation in space.

The number of L-H aggregating cities shows a trend of declining first and then rising, but the regional distribution changes greatly. L-H agglomeration gradually expands from downstream to upstream and midstream, and finally, a majority of L-H aggregating cities move to the middle reaches. In 2006, the provinces where L-H agglomeration appeared were Jiangsu and Zhejiang (Fig. 3). These cities had low CWRGE, but they were clustered in the lower reaches of the YREB, and the efficiency values of their surrounding cities were generally high, thus showing L-H agglomeration.

In 2011 and 2016, the number of L-H gathering cities has increased, and in 2016, the cities with L-H agglomeration were mainly distributed in the upper and lower reaches of the YREB. By 2021, the cities with L-H agglomeration were located in the middle reaches of the Yangtze River. Among them, Yichun is adjacent to provincial capitals such as Changsha and Nanchang. Several cities experienced significant economic growth, so the areas surrounding Yichun City showed a high efficiency value, while its own efficiency value was low; that is, L-H aggregation existed.

The number of cities with H-L concentration shows an increasing trend. In 2011 and 2016, they were mainly distributed in the middle reaches of the YREB. By 2021, they moved to the middle and lower reaches of the Yangtze River (Fig. 3).

Spatiotemporal evolution analysis of CWRGE

In this study, ArcGIS 10.8.1 is used to classify and graphically illustrate the CWRGE of 108 cities in the YREB (Fig. 4). It is separated into four categories to more intuitively describe the spatial and temporal evolution characteristics of CWRGE; the darker the hue, the better the efficiency value. Efficiency values above 1 make up the first category, and the other three categories are classified according to the natural discontinuity classification method. These four categories are labeled high, middle and high, middle and low, and low.

Fig. 4
figure 4

Temporal and spatial distribution of CWRGE in YREB

Full size image

Figure 4a to e depict the distribution of efficiency in 2006, 2011, 2016, and 2021 respectively; there were 15, 13, 12, and 21 cities with high-efficiency values, respectively. 2016 had the least cities with high-efficiency levels, showing a trend of first decline and then increase. However, the number of cities with middle- and high-efficiency values was 19, 17, 18, and 12, showing a downward trend. While the number of cities with middle- and low-efficiency values was 54, 47, 58, and 34, showing a fluctuating downward trend, the number of cities with low-efficiency values was 20, 31, 20, and 41, respectively, showing a fluctuating upward trend. Although the number of cities with lower efficiency values increased, the classification of these cities increased from 0.33–0.58 to 0.45–0.66. The number of middle- and low-efficiency cities decreased significantly while the number of middle and high efficiency cities decreased, but the classification of middle and high efficiency cities increased from 0.73–1.00 to 0.89–1.00. The number of high-efficiency cities is rising; the green efficiency of urban water resources from 2006 to 2021 overall showed an upward trend.

However, the growth rate of CWRGE in the YREB is slow. During the sample period, the CWRGE in many cities was not effective. Spatially, it is characterized by “small dispersion and large agglomeration”; this is consistent with Song and Du (2021). And the CWRGE in each city fluctuated in different periods, which is consistent with Wang (2021). Among them, only Changzhou, Changde, Yuxi, Changsha, and Shanghai were always highly efficient cities. These cities place a high value on environmental protection, rational industrial organization, and scientific and technical innovation, so their efficiency value is always high. The cities with low efficiency were Puer, Huanggang, Lincang, Ji’an, Ganzhou, and Xiaogan, which are mainly distributed in the middle and upper reaches of the YREB. This is basically consistent with the results of Guo and Tang (2021).

Analysis of influencing factors

This paper uses Stata12.0 to examine the relevance of each influencing element on CWRGE. The estimated findings of each independent variable’s coefficient are provided in Table 6 and are based on the Tobit model.

Table 6 Regression results of Tobit model
Full size table

As shown in Table 6, the EDL-level coefficient is 0.0789, indicating that the level of economic development favorably affects the sustainability of the YREB’s urban water resource use. The level of EDL significantly affects efficiency at 1%. This is consistent with Chen et al. (2021b), who argue that economic development promotes the upgrading of water equipment, the development and utilization of unconventional water sources, and the efficiency of water utilization. The research period of this paper is 2006–2021, during which the slogan of “To step up conservation of the Yangtze River and stop its over development” was put forward. The early “catch-up-oriented” economic development model, which only paid attention to economic development and ignored environmental protection, has been abandoned. Instead, high-quality, all-around balanced economic development is being pursued.

The coefficient of ER is −0.0005, which shows that ER negatively affects CWRGE in the YREB with a significance level of 1%, and the severity of environmental control has caused “crowding out” and “restraining” consequences. This aligns with Jin et al. (2019) and Li and Xian (2022), who believe that this can be given in two stages. First, insufficient enforcement of environmental rules and regulations has little impact on promoting CWRGE because of the pursuit of national economic growth and economic benefits in various regions. Second, environmental oversight was stepped up after China’s economy reached a stage of high-quality development. The cost of sewage generated by industrial development increased, but the manner of industrial development was not altered in many locations. As a result, businesses are less enthusiastic about investing in R&D and green production, which has a detrimental impact on environmental legislation (Jin et al. 2019). Compared with ordinary private goods, the threshold of water resource utilization is lower, which causes water resource waste and pollution (Zou and Cong 2021). However, this is inconsistent with Huang et al. (2021a). They believe that ER has a negative effect, which may be due to the fact that their study measures provincial water resource efficiency, while this paper measures urban water resource efficiency. There are also gaps in the index system and measurement model used. Although strict ER may reduce sewage discharge and environmental pollution, the blind pursuit of environmental protection may consume a lot of labor power, material resources, and financial resources, ultimately leading to low CWRGE. The government can employ ER as a primary mechanism of market macrocontrol to impose significant restrictions on the use of water resources in industrial output. Yet, if ER is implemented too strictly, it may restrict businesses’ capacity for innovation (Matthews 1981). Therefore, the government should reasonably consider ER when formulating water resource policies and not be too austere, which may negatively impact CWRGE.

Although the impact of technology investment (TI) on CWRGE in the YREB is negative, it did not pass the significance test, indicating that its impact on CWRGE is not substantial; this came to the same conclusion of Huang et al. (2021a). The reason may be that the level of science and technology in the provinces of the YREB is low, the costs of the science and technology input and output are high, the contribution rate is low, and the science and technology cannot be applied in practice. Scientific and technological innovation plays a significant role in promoting CWRGE. The interaction between technological innovation and ER prevents improvement of comprehensive utilization efficiency of industrial water resources, demonstrating that technological innovation may harm comprehensive utilization efficiency of industrial water resources in unfavorable environmental conditions (Jin et al. 2019). In this study, wastewater discharge is regarded as an unexpected output, and the environmental impact is considered to a great extent. There is a contradiction between the ecological environment and the economic development of the YREB, making the effect of scientific and technological investment on CWRGE unstable and unable to significantly contribute to CWRGE.

The coefficient of IS in the YREB is 0.0534, indicating that IS favorably affects CWRGE and that IS will significantly affect efficiency at 1%, assuming all other factors are held constant. This largely agrees with Jin et al. (2019) and Li and Xian (2022). Since the reform and OU, although there has been the idea of “pollution first and then treatment” in areas with rapid economic development, industrial enterprises and governments in industrialized areas have paid more attention to environmental protection than in other areas. They have also had an earlier understanding of industrial restructuring and environmental protection than people in other regions (Jin et al. 2019). Also, places with high efficiency frequently have an industrial infrastructure that is more complete, including cutting-edge water and energy-saving technologies, and a stronger sense of environmental protection. So, “clean” technologies are preferred in the process of technology development and equipment upgrading. This means that large-scale industrial cities can create more economic value and reduce water resource waste. Therefore, reasonable expansion of IS is beneficial to the improvement of CWRGE.

WS favorably impacts CWRGE in the YREB, as evidenced by the coefficient of WS, which passes the 1% significance test with a value of 0.2235. The term “WS” here refers to the ratio of domestic water use by residents to total water consumption, which means that increasing the ratio of domestic water use by residents can increase the rate at which urban water resources are used in a green manner. This is inconsistent with Su (2020), which may be caused by the inconsistency between the two research areas. This could also be due to urban water supply being primarily made up of agricultural water, industrial water, and residents’ domestic water. Agricultural production uses a lot of water resources, but its economic output is very low, so water resources cannot be well-utilized. As industrial water consumption may lead to more sewage and wastewater discharge, it is detrimental to the environment. With the acceleration of urbanization, more and more people move from rural areas to cities, which increases the number of urban residents and their water consumption, as well as, to a certain extent, economic output value and the urban labor force, and fosters economic development, all of which positively affect CWRGE in the YREB.

OU negatively affects the CWRGE in the YREB, as evidenced by the coefficient of foreign direct investment, −0.5476. This conclusion is inconsistent with Guo and Tang (2021). This is due to the fact that, despite China experiencing a significant inflow of foreign capital, it has also brought with it cutting-edge enterprise management concepts and environmental technologies, both of which raise the technical standards of local enterprises’ environmental performance (Ma et al. 2022). However, foreign direct investment is a double-edged sword, which may have a “pollution haven” effect (Manderson and Kneller 2011). Most enterprises invest in developing countries like China, which mainly produce traditional steel, chemicals, nonferrous metals, etc. These enterprises have low investment in scientific and technological innovation, backward production management technology, and serious industrial pollution. Because potential pollution havens are cheaper to enter, the pollution paradise effect encourages a larger proportion of multinational companies to locate high-pollution production sites there in order to reduce the environmental burden of their own countries(Manderson and Kneller 2011). Such businesses release a significant volume of industrial sewage, which will greatly harm China’s environment and reduce CWRGE in the YREB.

Conclusions and policy recommendations

The key to achieving healthy and sustainable use of water resources in the YREB is to optimize the efficiency system of urban green water resources. In this study, the CWRGE in the YREB from 2006 to 2021 is determined using the Super-SBM model. It makes more sense to use this model than the traditional water resource efficiency calculation method. ArcGIS software was used to examine the geographic and temporal evolution of CWRGE in the YREB, and a spatial autocorrelation model was used to analyze the Moran’s I of CWRGE. Further, a Tobit regression model was employed to conduct an investigation into the variables influencing CWRGE, and the following findings were discovered:

  1. (1)

    On the whole, the overall efficiency of the YREB’s green water resources is 0.7271, which is not as effective as DEA. The CWRGE in the YREB generally shows a trend of “first decreasing and then increasing, then decreasing and then increasing” and generally shows a “W”-shaped evolution law. Urban water resources in the YREB are now being used more sustainably than before the sample period. But, there is still potential for improvement, as the overall efficiency is not yet at its best.

  2. (2)

    There was a difference in CWRGE across different city scales during the reporting period: megacities > big cities > small- and medium-sized cities. This further demonstrates that small- and medium-sized cities have greater room to enhance CWRGE than megacities and large cities. From a regional standpoint, the YREB’s CWRGE was highest downstream, followed by upstream, and the lowest was the middle. From the scope of cities, there were 45 cities with CWRGE exceeding the average value of the YREB, and only seven cities with effective DEA, namely Changzhou, Hangzhou, Shanghai, Xuzhou, Changde, Changsha, and Yuxi.

  3. (3)

    The results of spatial correlation show that the CWRGE in the YREB is mainly positively correlated from 2006 to 2021, and the global Moran index shows a fluctuating downward trend over time. Both the aggregation range and the outlier range are distributed, there are mainly two positive spatial aggregation forms (H-H type and L-L type), and spatial distribution changes.

  4. (4)

    The findings of the spatiotemporal evolution demonstrate that there are more and more cities with high efficiency and fewer cities with low efficiency, and the spatial distribution difference is significant. However, the efficiency value range of inefficient cities has increased from 0.35–0.58 to 0.35–0.66. Yet, the growth rate of water resource green efficiency utilization in cities in the YREB is slow during the sample period, and there are still many cities whose CWRGE is not effective.

  5. (5)

    According to the Tobit regression model’s findings, the degree of urbanization, IS, and WS all significantly improve the CWRGE utilized in the YREB, whereas the effects of ER and foreign direct investment have a significant negative impact, and scientific and technological investment are not statistically significant. The research findings conclusively demonstrate that it is still inefficient and unfriendly to the environment for cities in the YREB to use their green water resources and that this efficiency needs to be further increased. As a result, this report offers the following detailed policy recommendations:

  1. i.

    CWRGE in the YREB clearly displays regional differentiation characteristics. Thus, the government must develop water-saving and water pollution control plans in different regions in accordance with local circumstances. The downstream of the YREB have the highest water resource efficiency, but not all cities have achieved high efficiency. For high-efficiency areas such as Shanghai and Jiangsu Province, they should promote the development of the surrounding cities. Advanced water resource management models and science and technology should be actively studied at home and abroad, the industrial structure in the industry should be adjusted, businesses with high water usage and wastewater emissions should be transformed and upgraded, and a high agglomeration space for CWRGE should be gradually formed. In upstream cities, CWRGE is in the middle and the geographical position of upstream cities is poor, but coordinated development can be promoted by strengthening ties with midstream and downstream cities. As an economically backward area, the upstream region should take full advantage of ecological conditions and resource advantages. The protection of the environment should not be compromised while the local economy is quickly being developed. Investment in R&D ought to rise as a result, relevant policies on water resources ought to be formulated, and enterprises and society ought to be encouraged and guided to boost spending on R&D in water resource development and utilization, saving water, reducing emissions, sewage treatment technology, etc., as well as optimizing the resource allocation of coordinated economic growth and water resource utilization. The red line of water use should be strictly controlled, the green economy vigorously developed, and the allocation and management capacity of water resources improved. The middle reaches of urban water supplies have the lowest rates of green use. First, cities in the middle reaches should reduce traditional heavy industries such as steel, cement, heavy metals, and chemicals, which use a lot of water and hurt the environment, and focus on developing advanced manufacturing industries, such as biotechnology, new materials, new energy sources, information technology, and environmental protection. Second, the middle reaches, as the node connecting the upstream and downstream areas, should take full advantage of its location, enhance internal and external connectivity, actively take in overflow resources and technology from the lower regions, and promote green efficiency of local water resources.

  2. ii.

    Megacities and big cities have a high level of urban development, but as the development of small- and medium-sized cities is relatively sluggish, these cities have more room for improvement. The methods of speeding up industrial transformation, reducing water consumption, increasing investment in research and development, introducing advanced technologies, strengthening ecological environment construction, encouraging the development and spread of technology that conserves water and reduces emissions, and focusing on water conservation and utilization should be applied.

  3. iii.

    Based on the relationship between the YREB’s CWRGE and its contributing factors, promoting the level of urban development should be considering firstly. China has reached the stage of high-quality economic development, and innovation should be the main driver of this stage. In addition, environmental protection and achieving economic growth with significant environmental and energy-related added values should be focused. Cities should support the growth of emerging sectors in accordance with their local conditions, bring into full play their advantages, and actively encourage the free flow of people, ideas, capital, and other resources between regions. Second, increasing the size of high-quality industrial companies in cities is vital, and a reasonable expansion of IS is conducive to promoting regional economic growth, thus enhancing CWRGE. To realize greener urban industrial development, “clean” technology should be prioritized in the process of technology development and equipment upgrading. Third, WS should be adjusted. The rising percentage of residents’ water consumption is helpful for increasing CWRGE. Therefore, with the acceleration of urbanization, the policy of “population mobility” should be formulated, as it is favorable to the development of the YREB. Urban development cannot be separated from labor, and the attraction of cities to labor comes from the development of cities. Thus, it is necessary to improve the construction of public infrastructure, as well as medical insurance and education systems.

Although the improvement of CWRGE is inseparable from ER, the extent of environmental control, which may lead to a lot of waste of financial and material resources, is something else to be considered. Therefore, the government should reasonably consider ER factors when formulating water resource policies without being too strict, which negatively impacts CWRGE. In tandem, regardless of the fact that foreign investment should be encouraged, the government should attract more investment in energy-saving and environmental protection technologies, but not at the expense of the environment.

  1. iv.

    ER should be strengthened using different subjects. First, enterprises should hasten water saving. Businesses with excessive water consumption should be encouraged to develop intelligent monitoring systems and improve their current technologies. Pipeline leakage and sewage prevention systems should be implemented for precise management and optimal control of industrial water. Second, the government should actively formulate a standard water-saving system in line with its own urban development, abide scrupulously by the central environmental protection requirements, and share extensive public guidance to let all sectors of society know about relevant policies. The formulation and revision of water-saving practices, water quotas, water balance tests, and water-saving technical standards should be accelerated. Third, citizens’ education on environmental preservation and water conservation should be strengthened. Citizens should constantly strengthen their awareness of water conservation and environmental preservation, consciously assume responsibility of environmental preservation, and adopt a low-carbon and environmentally friendly lifestyle. This raises the rate at which urban water resources are used in a green manner.

Research Deficiency and Prospects

The following limitations of this study suggest directions for future investigation: (1) Industrial wastewater discharge is employed as an unexpected output in this work, but due to the lack of data, other pollutants, such as industrial NH4+–N and industrial COD are not added to the account. The index system will need to be updated in the future in order to increase the precision of efficiency measurement. (2) This paper’s research focuses on CWRGE in the YREB, which covers only the cities around the YREB. In the future, the research area should be expanded so as to put forward more comprehensive conclusions on CWRGE. In addition, the research units can be reduced to county-level areas. This will provide more accurate information for efficient water resource use in China or globally. (3) CWRGE is influenced by many factors, and the factors that can be discussed are also very extensive. Owing to data limitations, the influencing factors addressed in this work are very constrained. Further elements that affect CWRGE can be researched in future study.