Spatial Heterogeneity in Economic Benefits of Water Use: Sectoral Analysis of Chinese Cities in 2017

Abstract

Spatial heterogeneity in economic benefits of water use provides crucial evidence for the evaluation of water diversion projects and the spatial equilibrium of water resource allocation. Using city-level data from 2017 on the sectoral water use and value added in 334 Chinese cities, we estimated the economic benefits of water use in the agricultural, industrial, and service sectors using the allocation coefficient method. We then revealed the spatial heterogeneity combining an exploratory spatial data analysis (ESDA) method. For the agricultural sector, the high economic benefit of water use regions are primarily concentrated on both sides of the “Hu Huanyong Line”; regions with high economic benefit of industrial water use are mainly found in the North China Plain, the middle and lower Huanghe River basin, the Yangtze River Delta, the Pearl River Delta, Chongqing and Chengdu, and the economic benefit of service water use is higher in the north than in the south. ESDA provides significant evidence for the analysis of spatial heterogeneity with regard to the economic benefits of water use in China. Based on the fundamental distribution of water resources and the spatial heterogeneity in the economic benefits of water use, potential water diversion areas can be preliminarily identified. The Haihe River Basin in the North China Plain and some areas in the southeast coastal region are potential receiving areas, and the eastern regions of Southwest China with abundant water resources and lower elevations, along with the middle and lower reaches of the Yangtze River are potential source areas. Further research about marginal benefits and water use costs, along with dynamic updates, is required for water resource allocation of China.

1. Introduction

As one of the world’s most critical challenges of the 21st century, water scarcity is intensified by the synthetic impacts of climate change and anthropogenic activities [1,2]. This challenge is mainly stemmed from the disparities between the inherent spatiotemporal heterogeneity of water resources and the geographical distribution of socio-economic activities [3]. As the world’s largest developing economy, China profoundly faces such challenges. This is mainly because the substantial economic and demographic footprint in China is concentrated on the east of the “Hu Huanyong Line”, but this pattern is fundamentally mismatched with the country’s naturally uneven water distribution between the north and the south [4]. The water governance philosophy of “spatial equilibrium” has been instituted as China’s guiding principle for a new era, seeking to foster a harmonious coexistence between water resources and the development of population, economy, and the society [5,6,7,8]. According to the general equilibrium theory and the spatial equilibrium theory, the essence of the spatial equilibrium of water resource allocation is the marginal benefit equilibrium under the conditions that consider of the cost of water transfer between two places. A critical step forward is to develop a diagnostic metric of spatial allocation-based economic benefits that is grounded in rigorous, sector-specific economic valuations of water [9]. This study focuses on the spatial heterogeneity in economic benefits of water use, which provides crucial evidence for the evaluation of water diversion projects and the spatial equilibrium of water resource allocation.

Accurately assessing the economic value of water resources is a prerequisite for conducting spatial equilibrium analysis. Methodologically, the production function approach, mathematical programming, and the benefit attribution method represent three mainstream techniques for estimating the economic value of water resources [10]. For instance, Peng, et al. [11] developed a computable general equilibrium model for water-receiving areas of water diversion projects, using Beijing as a case study to quantitatively evaluate the socio-economic impacts of such initiatives. Similarly, Wu, et al. [12] employed an improved particle swarm optimization algorithm to solve a multi-objective optimization model, examining the trade-offs between economic and ecological benefits under changing environmental conditions in the context of the Hanjiang-to-Weihe River Water Diversion Project. In a broader assessment, Yu, et al. [13] adopted an integrated analytical approach to investigate the economic, environmental, and social implications of water transfer projects across China. Notably, the benefit attribution method—also referred to as the residual value method—has been widely applied in evaluating water use benefit at regional and sectoral scales, owing to its straightforward data requirements and computational procedure [14]. Internationally, researchers have employed similar principles to assess water productivity in sectors such as agriculture and industry [15,16,17]. In China, this method has been successfully implemented both at the provincial level [18] and within river basin contexts [19], confirming its validity and applicability. For example, Wang, Jia, Bing and Liang [19] established an integrated quantitative framework incorporating the benefit attribution method, productivity change analysis, and an economic loss estimation model for water pollution, to evaluate the comprehensive social, environmental, and economic benefits of the Water Diversion from Yangtze River to the Taihu Lake Project for the year 2011.

Despite considerable progress, significant knowledge gaps persist in enabling a spatially nuanced equilibrium analysis of water use in China [20,21]. A primary limitation is the coarse resolution of existing assessments, which are largely confined to the provincial or major river-basin level. This obscures critical inter-city heterogeneity and impedes a mechanistic understanding of water-use economic benefit at the level where key policy interventions are implemented, within the prefectural city [17]. Furthermore, the prevalent application of the benefit-sharing method often overlooks a fundamental source of bias: the spatiotemporal heterogeneity of the allocation coefficient [18,22]. Compelling evidence indicates that water’s marginal contribution to economic output is not a universal constant but is intrinsically modulated by local resource endowments and climatic regimes [23]. In agriculture, for instance, this coefficient is markedly higher in arid regions or during drought years compared to humid conditions. The use of a static, uniform value risks a systematic misestimation of water use economic benefits in water-scarce contexts, thereby distorting the true landscape of regional economic benefit disparities.

This study established a sectoral-city level analytical framework to reveal the spatial heterogeneity in economic benefits of water use across China. The objectives are as follows: (1) to collect sectoral (agriculture, industry, and service) water use and value added data of 334 Chinese cities in 2017, and to calculate the economic benefit of water use in the three sectors; (2) to analyze the regional and structural differences in economic benefits of water use; and (3) to explore the reference of spatial heterogeneity in economic benefits of water use for water diversion projects and the spatial equilibrium of water resource allocation in China.

2. Materials and Methods

2.1. Dataset

Here we establish 2017 as the baseline year and select 334 cities (the scope covers mainland China, excluding cities below the prefecture level and Hainan province) in China as our research samples. The study focuses on three key sectors: agriculture (crop cultivation), industry, and services. The cities list, sector-specific value added data and the amount of sectoral water use data were obtained from China Statistical Yearbook 2018 [24]. For partially missing data, government information is available upon request through channels such as telephone and email. If data remained incomplete, interpolation was performed using data from adjacent regions or years.

2.2. Method for Calculating the Economic Benefit of Water Use

We employed the allocation coefficient method to quantify water use benefit across sectors. This approach allocates economic benefits proportionally according to the contribution of production factors, offering both conceptual clarity and practical data accessibility. For non-agricultural sectors, the allocation coefficient is typically determined by either an investment-based or fixed-asset-based method, owing to data availability. In the agricultural sector, the coefficient is derived from long-term experimental parameters. To make the calculation simple, we assumed that the water benefit allocation coefficients do not differ across cities. The economic benefits of water use can be calculated as follows (Equation (1)):

$$WUEB={\alpha }_i\times {\displaystyle \frac{y_i}{w_i}}$$

(1)

Here, $WUEB$ is the economic benefit of water use; $y$ represents the value added; $w$ is the amount of water used; $\alpha$ is the water benefit allocation coefficient; and $i$ represents the water use sectors.

This study employed the benefit allocation coefficient method. Specifically, the economic benefit of water use in a given sector is calculated as the ratio of its value added to the amount of water use (i.e., benefit per unit of water) multiplied by a corresponding water benefit allocation coefficient). The determination of allocation coefficients draws upon established research: a coefficient of 0.035 is adopted for industrial water use [25], and 0.04 for the service sector [19]. For agriculture, the range of coefficient is between 0.4 and 0.6, although Wang, Jia, Bing and Liang [19] applied a higher value of 0.53, we conservatively assign a coefficient of 0.4. This adjustment accounts for the relatively higher precipitation observed in 2017 in northern China (e.g., the Yellow River Basin), which is likely to reduce agricultural dependence on irrigation.

2.3. Exploratory Spatial Data Analysis Method

Exploratory spatial data analysis (ESDA) is an extension of exploratory data analysis (EDA) that has been developed within the fields of spatial statistics, spatial econometrics, and geostatistics [26]. It comprises interactive, visualization-driven techniques that reveal spatial dependence, outliers, and heterogeneity in geographic datasets [27]. Spatial autocorrelation analysis, which is an important part of the ESDA, is employed to assess the relationship of a certain variable at a given location with other variables at neighboring locations. Spatial autocorrelation can be performed over a complete study area using Moran’s I index and at specific locations using the Getis-Ord G_i* statistic to reflect the degree of spatial clustering of attribute variables in the whole study area [28,29]. We selected Moran’s I, Getis-Ord’s general G and Gi* to investigate the spatial heterogeneity of economic benefits of water use in China.

Moran’s I are the best-known statistics to measure global spatial autocorrelation of a quantitative variable. They correspond to the degree of linear association between the value of a variable at one location and the spatially weighted average of neighboring values. Formally, the Moran’s I is given by

$$I={\displaystyle \frac{n}{S_0}}{\displaystyle \frac{{\sum }_i{\sum }_j{w}_{ij}\left(x_i-\bar{x}\right)\left(x_j-\bar{x}\right)}{{\sum }_i{\left(x_i-\bar{x}\right)}^2}}$$

(2)

where $n$ is the number of observations, $w_{ij}$ is the degree of connection between the spatial units $i$ and $j$, and $x_i$ is the variable of interest in region $i$ and

$$S_0=\sum _i\sum _j{w}_{ij}$$

that is, the sum of all the weights.

The value of Moran’s I coefficient is [−1, 1]. When the value is greater than 0, it indicates that there is a positive spatial correlation in the study area. As numbers approach one, the degree of spatial autocorrelation is strengthened. When values are <0, values approaching −1 indicate that the spatial negative autocorrelation is stronger. A random distribution exists when the value is close to zero.

The G statistics developed by Getis and Ord provide an alternate measure of global spatial association (G) and observation specific (G_i and G*) measures of local spatial association. These last statistics detect the presence of local spatial autocorrelation: a positive value of this statistic for observation i indicates a spatial cluster of observations with high value of the variable of interest, whereas a negative value of the G_i or statistics indicates a spatial cluster of observations with low value around observation i. The difference between G_i and G* is that the latter statistics provide a measure of spatial clustering that includes the observation under consideration, while G_i does not. In other words, j = i is included in the sum (in the denominator) in G* statistics.

Formally, the G statistics can be written as follows:

$$G={\displaystyle \frac{{\sum }_{j\ne i}{w}_{ij}{x}_i{x}_j}{{\sum }_{j\ne i}{x}_i{x}_j}}$$

(3)

where $w_{ij}$ is the degree of connection between the spatial units i and j, and $x_i$ is the variable of interest in region $i$.

3. Results and Analysis

3.1. Spatial Heterogeneity of Economic Benefit of Agricultural Water Use

Regional analysis reveals that agricultural water use in China is disproportionately concentrated in several key regions, including much of Xinjiang, Jiuquan City in the Hexi Corridor, north–central Jiangsu (Figure 1 and Figure 2), Bayannur City in the Hetao Plain, and the periphery of the Sanjiang Plain in Northeast China. Significant usage is also observed in parts of Sichuan, Guangxi, Hunan, Hubei, and Jiangxi (Figure 1a). At the city level, the prefectures of Kashi, Aksu, and Jiamusi exhibit the highest use, with each exceeding 5 billion m³. Kashi’s usage (9.478 billion m³) surpasses that of Aksu (7.192 billion m³) by nearly one-third. Notably, Xinjiang accounts for eight of the top thirty cities by water use. Within the Sanjiang Plain, cities like Jiamusi, Harbin, Qiqihar, Jixi, and Suihua represent a significant proportion of high water-use areas. In Jiangsu, Yancheng and Xuzhou cities are the primary water users, with several other cities also exhibiting substantial usage. Other major consumers include Chengdu, Guilin, Nanning, Changde, Yichang, and Jiuquan, alongside multiple cities in Jiangxi province. Conversely, minimal agricultural water use is largely confined to the high-altitude terrains of the Qinghai–Tibet and Yunnan–Guizhou plateaus, the Loess Plateau, and specific areas such as the mountainous Baishan City or the archipelagic Zhoushan (Figure 2a).

As illustrated in Figure 1b, high agricultural value added is concentrated in southeastern China and parts of northwestern Xinjiang, with a distribution approximating the “Hu Huanyong Line” and the monsoon boundary. Leading regions include Heilongjiang, Hebei, Shandong, Jiangsu, Henan, Sichuan, and Hunan. A distinct northeast–southwest belt of high-value cities spans China’s second and third topographic steps, heavily represented by provinces such as Henan and Shandong (Figure 2b).

Analysis of water use benefit (Figure 1c) identifies two distinct scenarios. The highest economic benefit values are found in regions with minimal water use, such as Yushu, Alxa League, and Golog. More critically, high economic benefit driven by substantial value added is concentrated in Shandong, Henan, Anhui, Hubei, and Chongqing, which are areas that form the core of China’s productive agricultural belt. In contrast, the lowest efficiencies are prevalent in semi-arid and arid regions with underdeveloped agriculture. A notable finding is the concentration of ten low-economic benefit cities within Xinjiang, which is a region characterized by high water use and high output. This indicates a paradoxical coexistence of water profligacy and irrigation deficits, likely stemming from inefficient irrigation infrastructure, significant canal leakage, and lagging water conservancy projects. Similarly, the Northeast Plain exhibits high water use and output but suboptimal economic benefit. Despite these inefficiencies, both Xinjiang and the Northeast Plain hold strategic importance to China with regard to national food security (Figure 2c).

For agriculture, the high economic benefit of water use regions are primarily concentrated on both sides surrounding of the “Hu Huanyong Line”. Regions with high value added and high irrigation water use do not achieve high economic benefits of water use. For instance, areas such as Northeast China, Southeast China, and Xinjiang have high agricultural output but also use substantial amount of water. It is evident that China’s agricultural production relies primarily on irrigation, the area surrounding the line yields higher economic benefit of water use.

3.2. Spatial Heterogeneity of Economic Benefit of Industrial Water Use

Figure 3a illustrates the spatial distribution of industrial water use in China. Regions with high industrial water use are predominantly located along the middle and lower reaches of the Yangtze River Economic Belt, as well as in certain coastal cities of Fujian and Guangdong. Shanwei City recorded the highest industrial water use, reaching 7.40 billion m³. In southwestern China, Chongqing and Chengdu stand out with industrial water uses of 3.04 billion m³ and 1.28 billion m³, respectively. In southern China, besides of Guangzhou and Foshan in Guangdong province, Nanning also shows relatively high industrial water use, at 869 million m³. Among the cities with high industrial water use, six are from Jiangsu province, while Anhui and Hubei contribute four cities respectively. In contrast, cities with low industrial water use are mainly concentrated in northwestern and southwestern China. Qinghai, Tibet, Gansu, and Xinjiang account for five, six, five, and four such cities, respectively, all with water use below 32 million m³ (Figure 4a).

Figure 3b shows the spatial distribution of industrial value added. Most cities with high industrial output are situated at the east of the “Hu Huanyong Line”, particularly in the Pearl River Delta, Yangtze River Delta, and other coastal areas in the east and southeast area, as well as in the Yangtze River Mid-Reach and Chengdu–Chongqing urban agglomerations. It is evident that western regions characteristically demonstrate reduced levels of industrial value added. In the southwest, Chongqing and Chengdu lead with industrial value added of CNY 658.708 billion and CNY 521.720 billion, ranking fourth and eighth, respectively. Cities with high industrial output are mostly municipalities, provincial capitals, or sub-provincial cities, with notable concentrations observed in Jiangsu (six cities), Guangdong (four), and Shandong (four). In Northeast China, Changchun is the only city with relatively high industrial value added, at CNY 266.777 billion, ranking 25th. Cities with low industrial output are largely concentrated in western China, including several in Xinjiang, Tibet, Qinghai, and Gansu, each with industrial value added below CNY 5.6 billion. Yunnan also has four autonomous prefectures with low output, while Yichun and Heihe in Heilongjiang province show similarly low levels. The distribution of industrial value added exhibits a strong correlation with regional economic development (Figure 4b).

Figure 3c depicts industrial water use benefit, which shows a different spatial pattern from that of water use or industrial output, generally exhibiting higher values in northern China. Regions with high economic benefit are mainly found in the North China Plain, the middle and lower Huanghe River basin, the Yangtze River Delta, the Pearl River Delta, Chongqing and Chengdu. Some areas in northeast of China also show relatively high economic benefit. Notably, cities in Shandong—such as Yantai, Dongying, Qingdao, and Weihai—demonstrate particularly high industrial water use benefit, all exceeding 60 CNY/m³ and ranking among the top five nationally. Shenzhen, which ranks fourth in terms of industrial value added, also shows high economic benefit at 62.81 CNY/m³. Four cities in Hebei province exhibit moderate economic benefit, ranging between 30 and 35 CNY/m³. In contrast, Shanwei city, which has the highest industrial water use nationwide, shows the lowest economic benefit, at only 0.17 CNY/m³. Cities with economic benefit below 3.30 CNY/m³ are widely distributed, with six each in Tibet and Anhui, and others scattered across coastal, central, southern, southwestern, northeastern, and northwestern regions. Some regions, such as Hainan and Huangnan prefectures in Qinghai province and the Greater Khingan Range area in Heilongjiang province, exhibit high economic benefit primarily due to their low industrial water use (Figure 4c).

The southeastern coastal regions demonstrate a high level of both industrial water use and value added. Nevertheless, the economic advantage in the northern region is comparatively superior to that observed in the southern region. Industrial water use encompasses both fresh water and non-conventional sources, while internal recirculated water volumes are excluded. This may provide a partial explanation for the observed higher economic benefit of water use in northern regions compared with southern areas.

3.3. Spatial Heterogeneity of Economic Benefit of Service Water Use

As illustrated in Figure 5a, the amount of water use in China’s service sector exhibits pronounced regional disparities. High water-use regions are predominantly concentrated in the southeastern coastal areas, the Beijing–Tianjin–Hebei region, the Sichuan–Chongqing area, as well as Hunan and Hubei provinces. In contrast, water use in the service sector is considerably low in western China, particularly in the west of the “Hu huanyong Line”. Shanghai ranks the highest in service-sector water use, reaching 1.134 billion m³, which is 73% higher than that of Guangzhou, the second-ranked city. It is followed by Shenzhen, Wuhan, Hangzhou, Beijing, and Chongqing, all of which exceed 400 million m³ in water use. Hubei province contains five cities with high water use, while Guangdong and Zhejiang each contribute four cities. Cities with the lowest water use—all below 10 million m³—are largely concentrated in Xinjiang (eight cities) and Qinghai (five cities). Notably, water use in Shanghai exceeds that of Hami Prefecture, the region with the lowest consumption, by more than 1000 times. This finding highlights significant spatial inequalities in water use across China for the service sector (Figure 6a).

The spatial distribution of value added generated by the service sector (Figure 5b) largely mirrors that of water use, displaying a similar southeast–northwest gradient. High value added areas are mainly distributed along the southeastern coast, in the Beijing–Tianjin–Hebei region, and around cities such as Chongqing and Harbin. Conversely, Qinghai and Tibet have reported the lowest outputs. Beijing leads with a service-sector value added of CNY 2256.78 billion, closely followed by Shanghai at CNY 2119.15 billion. It is noteworthy that although Beijing’s service-sector output surpasses that of Shanghai, its water use is only half that of Shanghai. Other major contributors include Guangzhou, Shenzhen, Tianjin, and Chongqing, each with value added exceeding CNY 900 billion. The low added values are observed in Qinghai, Tibet, Xinjiang, Yunnan, Gansu, and Ningxia (Figure 6b).

In contrast to the distributions of water use and value added, which is higher in the southeast and lower in the northwest, the economic benefit of service water use (Figure 5c) is higher in the north than that in the south. Zhenjiang City in Jiangsu Province exhibits the highest economic benefit, at 711.16 CNY/m³ of water. Regions with high water-use economic benefit are mainly found in Xinjiang, Shandong, Jiangsu, the middle reaches of the Yellow River, northeastern China, and parts of Yunnan. Several prefectures in Xinjiang (Hami, Aksu, Altay, Turpan, and Bortala), Ordos City in Inner Mongolia, as well as Zibo, Binzhou, and Dezhou in Shandong also rank highly. Conversely, low-economic benefit areas are concentrated in Hubei, Tibet, and Guangxi (Figure 6c).

The service sector encompasses a wide variety of industries, which are characterized by relatively high economic benefit of water use and significant differences. For instance, the catering industry being a relatively low economic benefit sector within the service industry, it remains consistently above CNY 5 across the cities in southern Xinjiang. Nevertheless, the water use of service sector remains comparatively low in relation to that of other industries, resulting in its marginal consideration in the context of water resource allocation. In the context of an evolving industrial structure, there is a possibility that the service sector may be allocated increased levels of attention.

3.4. ESDA for Sectoral Economic Benefit of Water Use

As illustrated in

Table 1. Spatial autocorrelation statistics for sectoral economic benefit of water use.

Sector	Moran’s I			Getis-Ord’s General G 1
	I	Z Score	p Value	Z Score	p Value
agriculture	0.07	5.75	<0.01	1.58	0.11
industry	0.38	26.84	<0.01	7.90	<0.01
service	0.18	12.92	<0.01	1.98	<0.05

Note: ¹ The Observed General G of the three sectors are not listed, as their values are close to zero.

and Figure 7, the spatial correlation analysis results for economic benefit of water use in agriculture, industry and services across Chinese cities in 2017 are presented. According to the Moran’s I, economic benefit of water use in three sectors demonstrate statistically significant positive autocorrelation. The Getis-Ord General G statistic indicates a significant clustering of high values in the industrial sector (z = 7.90, p < 0.01), corroborating the presence of pronounced spatial agglomeration. Hotspots for agricultural economic benefit of water use are concentrated in central regions such as Henan, Anhui, and Sichuan, as well as the northeastern Qinghai–Tibet Plateau, and cold spots are primarily found in the southeastern hilly areas of Zhejiang, Fujian, Hunan, and Guangdong. The distribution of hot and cold spots for industrial and service sector water efficiency shows some similarity, with hotspots located on the North China Plain and cold spots concentrated in southern regions south of the Yangtze River. Additionally, another hotspot for the service sector is centered in the central-eastern part of Xinjiang. Specifically, coastal cities in the south are largely outside the cold spot region for the three sectors. Spatial hotspot analysis can reveal the partial distribution characteristics of the economic benefit of water use to a certain extent, such as the primary regions exhibiting high or low benefit. ESDA provides significant evidence for the analysis of spatial heterogeneity with regard to the economic benefits of water use in China.

3.5. Preliminary Identification of Potential Water Diversion Areas

Based on the fundamental distribution of water resources and the spatial heterogeneity of economic benefit of water use, potential water diversion areas can be preliminarily identified. The Haihe River Basin in the North China Plain is a potential water-receiving area, and some areas along the southeast coast can be included in the potential transfer area depending on the degree of water shortages. The eastern part of Southwest China is rich in water resources and has a low altitude, as are the middle and lower reaches of the Yangtze River (which have actually already been used as the exporting area of the East and Middle Lines of the South-to-North Water Diversion Project and the subsequent planning of the first phase of operation and subsequent planning), making them a potential exporting area. The actual situation is that the reference for water diversion areas encompasses not only economic benefits of water use but also social benefits, ecological benefits, and costs, among other factors.

4. Discussion

4.1. Spatial Heterogeneity in Economic Benefit of Water Use and Water Resource Allocation of China

According to the equilibrium conditions of the general equilibrium theory and the spatial equilibrium theory, the essence of the spatial equilibrium of water resource allocation is the marginal benefit equilibrium under the condition of considering the cost of water transfer between two places. The marginal benefit of water is determined by the inefficient water users, or in the case of administratively dominant allocation, it is determined by the main purposes of water transfer. Spatial equilibrium is dynamic, and it is necessary to evaluate the current situation and analyze the future situation. A cost–benefit analysis of the relevant areas is required for the water diversion project. The findings of this study currently indicate potential diversion regions. However, further research about marginal benefits and water use costs, along with dynamic updates, is required for water diversion projects and the spatial equilibrium of water resource allocation.

Our analysis reveals significant spatial heterogeneity in sectoral water use benefit across China, a pattern intrinsically linked to regional disparities in water resource endowment, economic development, and industrial structure. Crucially, the geographic distributions of high-economic benefit agricultural and high-economic benefit industrial and service sectors are not fully aligned. This spatial disparity underscores a central dilemma in Chinese water governance: the trade-off between efficiency and regional equity [15,16,17,30]. Within the agricultural sector, we found that regions with substantial water use, such as Xinjiang and Northeast China, exhibited relatively low water use benefit. However, the “irrigation economic benefit paradox” highlighted by Grafton, Williams, Perry, Molle, Ringler, Steduto, Udall, Wheeler, Wang, Garrick and Allen [9] serves as a critical warning. Singularly pursuing higher efficiency or economic benefit—for instance, through technological advances that reduce irrigation withdrawals—does not guarantee increased water availability at the basin scale and may instead intensify regional water stress by incentivizing the expansion of an irrigated area [21,31,32]. Consequently, water allocation policies for strategic granaries such as Xinjiang and Northeast China must consider factors beyond mere “efficiency”. The imperative of “equity”—in this context, their vital role in safeguarding national food security—must be integral to the calculus [20,33,34].

The spatial patterns of water use benefits in the industrial and service sectors are equally complex [23]. Industrial economic benefit displays a “high-North, low-South” gradient. Certain southern coastal cities (e.g., Shanwei) register considerable industrial output, but their elevated water use suppresses overall economic benefit, suggesting a history of less intensive water use in some southern developments. Notably, the highest service sector economic benefit is found not in megacities like Beijing or Shanghai, but in cities such as Zhenjiang and Ordos. This indicates that while super-sized cities generate immense service sector economic value, their water resource use may be approaching saturation, with limited gains in marginal economic benefit. This finding resonates with Wang et al. [35], who documented pronounced regional heterogeneity in the socio-economic impacts of water resource changes in China. A region cannot solely develop its advanced industries, nor can all regions develop high value-added industries. It is necessary to explore the establishment of a strategic water reserve system for the northern regions. Furthermore, the water reallocation strategy must be coupled with carefully designed compensation mechanisms to mitigate inter-regional “equity” conflicts that may arise from the redistribution of water rights [22,35,36].

4.2. Applicability of the Apportionment Coefficient Method and Insights from Cross-Sectoral Comparison

The allocation coefficient method employed in this study provides a feasible framework for the unified quantification of water use benefits across multiple sectors. However, its application inherently raises important discussions concerning methodological applicability and the interpretation of results [14,37,38]. The value assigned to the allocation coefficient is central to this method and constitutes the primary source of uncertainty. This study adopts an agricultural allocation coefficient of 0.40, based on historical calculations at the national scale [19]. Nonetheless, numerous studies highlighted significant spatiotemporal heterogeneity in the allocation coefficient of water resources to economic output. Within the agricultural sector, this coefficient is highly dependent on climatic conditions. For instance, in arid and semi-arid regions or during drought years, where water becomes the most critical limiting factor of production, the coefficient can rise substantially (e.g., to 0.5–0.6), whereas it decreases markedly in years of abundant precipitation [9,39]. For the industrial and service sectors, we applied relatively low and fixed coefficients (0.035 and 0.04, respectively), aligning with parameters used by Wang, Jia, Bing and Liang [19] in assessing the benefits of water transfer projects. In contrast, Yin, Wang, Wang, Tang, Piao, Chen, Xia, Conradt, Liu, Wada, Cai, Xie, Duan, Li, Zhou and Zhang [18] employed a coefficient of 0.05 for both industrial and domestic water use. This discrepancy in parameter selection underscores the need for considerable caution when directly comparing results across different studies, particularly in the absence of a standardized measurement protocol. In reality, the data availability in the industrial sector is limited at the city level is limited, considering that refining regional allocation coefficients would introduce more errors, fixed allocation coefficients were therefore used. Future research should focus on developing dynamic allocation coefficient systems that adapt to technological progress, climatic variations and shifts in industrial structure [14,37,38]. Furthermore, a comparative analysis may be conducted in conjunction with the results pertaining to the city-level input–output table.

Despite the aforementioned uncertainties, the trend revealed by the striking disparities in cross-sectoral water use benefit is both clear and robust. Our results unequivocally demonstrate that the economic benefit per unit of water in industrial and service sectors is typically one to two orders of magnitude higher than in the agricultural sector. This considerable discrepancy provides a powerful economic rationale for the reallocation of water resources from agriculture towards higher value-added sectors, which underpins the high expectations placed on market mechanisms for water allocation [22,40,41,42]. However, the ecological and social boundaries of such transfers must be rigorously scrutinized. Firstly, agricultural water use, particularly the use of “green water” for maintaining soil moisture and regulating regional climate, carries significant ecosystem service value. This value is not adequately captured by the allocation coefficient method based purely on economic output [43]. Secondly, the transfer of agricultural water rights must carefully consider its impacts on rural livelihoods and social stability [2,16,44,45]. Consequently, our findings should not be simplistically interpreted as advocating for “diverting water from agriculture to support industry.” Instead, they underscore the urgent need to establish a multifaceted system for evaluating water resource value [4,46,47]. In policy practice, moving beyond the economic benefit focus of this study, it is crucial to incorporate ecosystem service values. Methods such as the Contingent Valuation Method (CVM) and other non-market valuation techniques should be leveraged to quantify the benefits of ecological water use, thereby providing a more comprehensive and multi-dimensional scientific basis for complex water resource decision-making processes [5,40,48,49].

5. Conclusions

This study develops a methodological framework centered on the apportionment coefficient method and ESDA, with the aim of systematically evaluating the economic benefits of water use and its spatial heterogeneity across the agricultural, industrial, and service sectors at the city level in China for the year 2017. The sectoral water-use economic benefit in China exhibits pronounced spatial heterogeneity. Furthermore, the spatial distribution of economic benefit derived from water use does not align uniformly with the economic development level of the city.

(1)

Given that China’s agricultural production relies primarily on irrigation, the high economic benefits of water use regions are primarily concentrated on both sides of the “Hu Huanyong Line”. Regions with high economic benefits from industrial water use are mainly found in the North China Plain, the middle and lower Huanghe River basin, the Yangtze River Delta, the Pearl River Delta, Chongqing, and Chengdu. The economic benefits of service water use are higher in the northern area than in the south.

(2)

Global Moran’s I shows statistically significant positive autocorrelation for all three sectors, while the Getis-Ord General G statistic indicates pronounced high-value clustering in industry. Hotspots for agricultural economic benefits of water use are concentrated in central regions such as Henan, Anhui, and Sichuan, as well as the northeastern Qinghai–Tibet Plateau, whereas industry and service benefits are concentrated in the North China Plain, with an extra service hotspot in central-eastern Xinjiang. Cold spots are systematically located south of the Yangtze River—especially the hilly southeast (Zhejiang, Fujian, Hunan, Guangdong)—whereas most southern coastal cities escape the low-benefit zone for all sectors. Spatial hotspot analysis can reveal the partial distribution characteristics of the economic benefit of water use to a certain extent, such as the primary regions exhibiting high or low benefits. ESDA provides significant evidence for the analysis of spatial heterogeneity with regard to the economic benefits of water use in China.

(3)

The eastern part of southwestern China, distinguished by its abundant water resources and relatively low elevation, in conjunction with the middle and lower reaches of the Yangtze River, has the potential to function as water export zones. These regions have effectively been designated as transfer areas for the South-to-North Water Diversion Project, including subsequent stages of planning.

Further research about marginal benefits and water use costs, along with dynamic updates, is required for water diversion projects and the spatial equilibrium of water resource allocation.