Regional Synthetic Water Pollutants Embodied in Trade and Policy Simulations for Mitigating Pollutant Discharge in China

Abstract

Inter-regional trade in commodities causes the flow of water pollutants, referred to as virtual pollutant transfer. However, existing studies usually focus on a single water pollutant and cannot characterize the integrated discharge of multiple ones. As a result, it is impossible to analyze the integrated virtual flow of multiple water pollutants among regions, much less simulate the effects of possible water pollutant reduction scenarios. To this end, we empirically synthesize several water pollutant indicators as a whole and then make it the occupancy in the framework of input–output analysis, which helps us to quantify the virtual transfer of water pollutants and simulate scenarios’ mitigating effects. The constructed indicator is called the synthetic water pollutant (SWP) discharge index. By accounting for SWP and then its virtual flows based on the compiled multi-regional input–output tables, we analyze the temporal and spatial differences in synthetic net virtual transfer of regional multiple water pollutants occurring with inter-regional trade. The results show that the national SWP discharge scale of six water pollutants (chemical oxygen demand, ammonia nitrogen, total nitrogen, total phosphorus, petroleum, and volatile phenol) is falling from 2012 to 2020. The net intake of virtual pollutants has become more concentrated. Central (e.g., Shanxi and Hunan) and western (Xinjiang, Inner Mongolia) China are the central regions of net virtual receiving. The simulation results show that reducing 10% of importing regions’ inputs while cutting 10% of exporting regions’ consumption mitigates the SWP discharge of the entire economic system by 3.45%. The decrease rate is 3.02%, increasing international imports by 10% in all regions. An incremental SWP reduction of 2.75% by reducing SWP discharge per output unit by 10% in the top 10 regions of discharge intensity indicates reducing the SWP discharge intensity is the most direct and effective approach. However, the growth of fixed asset investment in wastewater treatment and its recycling seems to contribute little to achieving China’s policy target of wastewater treatment capacity increase by 2025. This study provides regional results for managing water pollutants in China and a basis for future policymaking.

1. Introduction

Reducing water pollutants is essential to relieve the pressure of wastewater purification and reduce the cost of wastewater treatment. The leading pollutant indicators in wastewater are chemical oxygen demand (COD), ammonia nitrogen (NH₃-N), total nitrogen (TN), total phosphorus (TP), petroleum (P), and volatile phenol (VP). The discharge of large amounts of pollutants into water bodies leads to the deterioration of water quality. It increases the gap between the evaluated gray water footprint and to water available, thus triggering more severe water scarcity [1]. In addition, due to the differences in socioeconomic, resource endowment, productivity, and lifestyle among regions in China, the content of different pollutants in wastewater discharged regions varies (Figure A1). According to the data published by the National Bureau of Statistics, the COD discharge in wastewater increased by 94.42% from 2003 to 2020, while the NH₃-N discharge decreased by 24.06%. By contrast, from 2011 to 2020, the discharge of TN, TP, P, and VP pollutants decreased by 27.91%, 39.14%, 99.82%, and 97.54%, respectively. For example, a considerable reduction in VP comes primarily from high discharge areas (Figure A1f). In the context of deepening green transformation, it is essential to study the flow of water pollutants implied by inter-regional trade in commodities to provide regional and sectoral results for the sustained reduction of water pollutants and promote sustainable wastewater management in China [2].

Inter-regional commodity trade causes the transfer of water pollutants among different regions. The stronger the inter-regional economic ties and more frequent commodity trade, the greater the quantity of inter-regional water pollutant transfer. Inter-regional trade makes water pollutants leak and transfer from developed to developing regions, alleviating or exacerbating regional imbalances in water pollutant discharges. Commodity trade implies adverse water pollutant flows, which affect regional freshwater resource use. Scholars have explored the impact of commodity trade on regional water quantity [3,4,5,6,7], including the amount of freshwater needed to assimilate water pollutants, i.e., the gray water footprint [6,8], and the flows of gray water across administrative boundaries, i.e., virtual gray water [6,9]. The findings suggest that economically developed regions outsource their gray water footprint to developing regions through inter-regional trade, which further exacerbates the imbalance in regional water distribution [5,6].

Some scholars have studied the water problems caused by water pollutant discharges from inter-regional trade. For example, by quantifying the nitrogen footprint of urban areas in China, it was shown that the inter-regional supply of food products leads to significant nitrogen discharge [10]. Likewise, commodity trade has an important impact on unbalanced phosphorus flows [11]. Other scholars have studied the discharge of certain water pollutants on a regional scale [12,13]. Among them, Yang et al. [12] studied the critical sectors of COD, TN, and TP discharges using pollutant association analysis based on a non-competitive input–output table for the Liaohe River region. Wang et al. [13] quantified the domestic COD footprints and their interactions for each region in 2010 using a multi-regional input–output model.

Scholars have pointed out that water-scarce regions can import water-intensive products and thus conserve local water resources [14,15,16,17]. However, the production of high water-intensive products exacerbates water pollutant discharges in the regions where they are produced. This reflects the water needed to make goods and services along their supply chains through virtual water [18,19,20]. Zhao et al. [14] show that virtual water transfers increase water stress in exporting locations instead of playing a significant role in relieving water stress in receiving regions. The deep-seated reason may attribute to micro-level aspects, especially enterprises. An enterprise with a traditional business model mainly pursues economic value maximization [21]. Goods or services flowing in the supply chains essentially mainly transfer economic value. The environmental value is secondary to those poorly sustainable business models, which implies the virtual water transfer conclusions of Zhao et al. [14]. However, this may have regional differences. For example, Ferng [7] estimated that Taiwanese economic production relies on imports of international goods and services for a quarter of its total freshwater demand. Here, we reflect on the water pollutants discharged during the production of goods and services through virtual water pollutants. Important questions include: Since water reallocation is not effective in relieving regional water stress in China, what are the flows of virtual water pollutants? What can be done to reduce the discharge of water pollutants? It is crucial to explore reduction options for wastewater pollutants based on analyzing water pollutant transfer relationships and promoting sustainable wastewater management. Yet, there is a relative lack of studies on the significant water pollutant discharge transfers in response to commodity production exchange activities. Most related studies focus on only a single water pollutant (e.g., COD or nitrogen), and do not address water pollutant net virtual transfer due to inter-regional trade.

Based on water pollutant accounting indicators, this paper analyzes the regional water pollutants’ net virtual transfer with the inter-regional flow of commodity trade. The development of a framework for regional transfer analysis of water pollutants from the level of the macroeconomic system has theoretical and applied implications. First, from a theoretical point of view, the study of the regional transfer of water pollutants can increase the knowledge of the regularity of the correlated discharge of multiple regions in the macroeconomic system and identify the critical elements for promoting sustainable water pollutant management. Second, from a methodological perspective, the comprehensive analysis of the regional transfer of multiple water pollutants complements existing studies and provides a methodological reference for dealing with similar problems. Third, as for applications, the measurement of social, economic, energy, and environmental effects based on quantitative transfer analysis, together with rationally designed scenario simulations, can guide policy formulation and implementation of water pollution management.

Our study quantifies regions’ net virtual flow of synthetic water pollutants (SWPs) and the transfer amount. It analyzes the evolutionary characteristics of the net virtual transfer of SWPs and it captures the integrated discharge relationships among regions by an SWP index. The scenario simulation on the compiled 2017 multi-regional input–output (MRIO) table shows that wastewater treatment and its recycling policy have significant social, economic, energy, and environmental benefits. Our work’s innovations are summarized as (1) innovatively proposing a synthetic water pollutants (SWPs) index while existing studies focus on a single pollutant only, and (2) providing a framework of “linkage quantification-scenario simulation” that can be applied to the study of related environmental management issues. This framework complements the existing policy effect simulation studies from knowledge discovery to knowledge mobilization (application). The contributions of our work are as follows: (1) the integrated measurement of water pollutant discharges portrays the associations (Appendix B) formed by synthetic regional discharges of multiple water pollutants, synthetic inter-regional transfers, and their impacts, and (2) the integrated measurement of water pollutant discharges can be used in the assessment of sustainable, high-quality green development and improve coordinated development in an economic system.

2. Data and Methods

2.1. Multi-Regional Input-Output (MRIO) Tables

This paper sets an analysis framework based on inter-regional input–output tables for 30 sectors in 30 provincial administrative regions (all provincial administrative regions except Tibet, Hong Kong, Macau, and Taiwan region, see

Table A1. The names and abbreviations of the 30 provincial administrations.

Region	Abbreviation	Region	Abbreviation	Region	Abbreviation	Region	Abbreviation
Beijing	BJ	Shanghai	SH	Hubei	HuB	Yunnan	YN
Tianjin	TJ	Jiangsu	JS	Hunan	HuN	Shaanxi	SAX
Hebei	HB	Zhejiang	ZJ	Guangdong	GD	Gansu	GS
Shanxi	SX	Anhui	AH	Guangxi	GX	Qinghai	QH
Inner Mongolia	IM	Fujian	FJ	Hainan	HN	Ningxia	NX
Liaoning	LN	Jiangxi	JX	Chongqing	CQ	Xinjiang	XJ
Jilin	JL	Shandong	SD	Sichuan	SC
Heilongjiang	HLJ	Henan	HN	Guizhou	GZ

) in 2012, given in Mi et al. [22]. The sectoral data are combined according to the characteristics of the inter-regional input–output tables [23]. Each region’s intermediate inputs and final demand are the combined data of 30 intermediate sectors and five final demand sectors, respectively.

Using the 2017 China regional input–output tables released by the National Bureau of Statistics, we compiled input–output tables for the 30 provincial administrative regions of China in 2017. The MRIO table reflects inter-regional intermediate inputs and end-use, a non-competitive input–output model. The social (employment), economic (value added), and energy (electricity consumption) data for each administrative region in the table are obtained from the 2017 National Statistical Yearbook published in 2018. The 2017 China Urban and Rural Construction Statistical Yearbook [24] announced each administrative region’s environmental data (wastewater treatment capacity, total wastewater treatment, total sludge treatment, recycled water utilization capacity, and recycled water consumption). The water pollutants are COD, NH₃-N, TN, TP, P, and VP. Their annual discharge data are from the National Bureau of Statistics [25,26]. The multi-regional input–occupancy–output tables provide a basis for studying the social, economic, energy, and environmental effects of the virtual transfer of water pollutants across regions and the changes in final regional demand in China. We refer to the structure of the combined generated MRIO tables for 2012 and the compiled 30 administrative regions for 2017 in

Table A3. Format of the input–output table among 30 provincial administrative regions (Tan et al., 2021) (TIU: total intermediate use; TII: total intermediate input).

	Intermediate Demand				Final Demand			Export	Others	Total Output
Region	1BJ	…	30XJ	TIU	1BJ	…	30XJ
1BJ	z11	…	z1n	u1	f11	…	f1n	ex1	err1	x1
…	…	…	…	…	…	…	…	…	…	……
30XJ	zn1	…	znn	un	fn1	…	fnn	exn	errn	xn
TII	v1	…	vn
Import	im1	…	imn
Pollutant Discharge Quantity	p1	…	pn
Value Added	va1	…	van
Total input	x1	…	xn

.

2.2. Construction of the Regional SWP Discharge Index

Different pollutants in discharged wastewater vary significantly among regions (Figure A1), with various environmental and human health hazards. The U.S. Agency for Toxic Substances and Disease Registry (ATSDR) published the 2019 Chemical Substance Priority List [27], which scores the hazards of 275 chemical substances by combining their frequency of occurrence, toxicity, and exposure risk. This paper ranks the environmental hazards and human health hazards of six water pollutants (

Table 1. Ranking of environmental and human health hazards of 6 water pollutants.

Hazard	Rank
Environmental hazards	I: volatile phenol
	II: chemical oxygen demand
	III: ammonia nitrogen
	IV: petroleum, total nitrogen, total phosphorus
Human health hazards	I: volatile phenol
	II: chemical oxygen demand, ammonia nitrogen
	III: total nitrogen, total phosphorus
	IV: petroleum

) concerning their published chemical EPA (U.S. Environmental Protection Agency) reportable quantities (RQ values) and ATSDR toxicity/environmental scores (TES values) (

Table 2. RQ and TES values of wastewater pollutants and structurally related chemicals published by ATSDR.

Chemical Substance	RQ	TES
Ammonia	100
Azinphos-methyl	1
Phenol	1000
Polycyclic Aromatic Hydrocarbons		1

). The structurally related substances determine the ranking of P, TN, TP, and NH₃-N in Table 2. The “Classification Information Sheet of Hazardous Chemicals” [28], “List of Hazardous Chemicals (2018 Edition)” [29], “List of Priority Control Chemicals (First Batch)” [30], and “List of Priority Control Chemicals (Second Batch)” [31] published by China support the ranking given in Table 1, where rank I indicates the least hazardous, and so on. The hierarchy in Table 1 incorporates the evaluation of environmental hazards and health hazards of structure-related substances from the “Classification Information Sheet of Hazardous Chemicals” [28].

As it is difficult to reflect the regional and inter-regional differences by examining a single pollutant discharge, this paper integrates two dimensions of environmental and human health hazards. It assigns weights to different pollutants by the analytic hierarchical process [32]. Specifically, environmental hazards and human health hazards are considered two criteria for judging the degree of pollutant risks. Since exposure to environmental hazards can threaten human health, human health hazards are essentially or strongly important to environmental hazards. Referring to the nine importance levels and their assigned values given by Saaty [32], the quantitative importance level of human health hazards relative to environmental hazards is 5. The judgment matrix set to the criterion level is as follows:

$$C=\left(\begin{array}{cc}1& \frac{1}{5}\\ 5& 1\end{array}\right)$$

We treat the six pollutants as different scenarios. The importance quantification values start from 1, which are all 1 among pollutants of the same hazard rank. Here we follow the nine importance levels and their assigned values of Saaty [32] (

Table A2. Saaty’s (1977) nine levels of importance and their assigned values.

Factor i Over Factor j	Explanation
1	Factor i and factor j contribute equally to the objective.
3	Factor i is slightly more important than factor j.
5	Factor i is essentially or strongly important relative to factor j.
7	Factor i is strongly more important than factor j.
9	Factor i is extremely more important than factor j.
2, 4, 6, 8	An intermediate value of two adjacent judgments.

). The importance quantification values for ranks II, III, and IV for environmental hazards over rank I are 3, 5, and 7, respectively. The importance quantification values for ranks III and IV over rank II are 3 and 5, respectively, and 5 for class IV over class III. For human health hazards, the importance quantification values are set slightly differently. The importance quantification values for ranks II, III, and IV concerning rank I are 3, 5, and 9, respectively. According to the hazard ranks given in Table 2, the human health hazards for both COD and NH3-N are rank II, and those for TN and TP are rank III. Considering that both NH3-N and TN contain nitrogen, the importance quantification values for TN and TP to COD and NH3-N are set to 2, and that for rank IV relative to both ranks II and III are 5. Finally, in the order of COD, NH3-N, TN, TP, P, and VP, the environmental hazard pairwise comparison matrix and human health hazard pairwise comparison matrix of pollutants are obtained:

$$S_1=\left(\begin{array}{cccccc}1& \frac{1}{3}& \frac{1}{7}& \frac{1}{7}& \frac{1}{7}& 3\\ 3& 1& \frac{1}{5}& \frac{1}{5}& \frac{1}{5}& 5\\ 7& 5& 1& 1& 1& 5\\ 7& 5& 1& 1& 1& 7\\ 7& 5& 1& 1& 1& 7\\ \frac{1}{3}& \frac{1}{5}& \frac{1}{5}& \frac{1}{7}& \frac{1}{7}& 1\end{array}\right)\hspace{1em}\hspace{1em}{S}_2=\left(\begin{array}{cccccc}1& 1& \frac{1}{2}& \frac{1}{2}& \frac{1}{5}& 3\\ 1& 1& \frac{1}{2}& \frac{1}{2}& \frac{1}{5}& 3\\ 2& 2& 1& 1& \frac{1}{5}& 5\\ 2& 2& 1& 1& \frac{1}{5}& 5\\ 5& 5& 5& 5& 1& 9\\ \frac{1}{3}& \frac{1}{3}& \frac{1}{5}& \frac{1}{5}& \frac{1}{9}& 1\end{array}\right)$$

Under the above judgment matrix and pairwise comparison matrices, the hierarchical total ranking weights of 6 pollutants are calculated according to the calculation method of hierarchical analysis, as shown in

Table 3. Weights of 6 water pollutants.

	COD	NH3-N	TN	TP	P	VP
Weight	0.0800	0.0867	0.1690	0.1709	0.4595	0.0339

.

Due to the significant difference in the order of magnitude of different pollutant discharges, and to avoid the SWP discharge index being dominated by a single pollutant, we use the maximum–minimum criterion to normalize all regional discharges of each pollutant. Finally, the normalized discharge is then weighted to obtain the SWP discharge index (therefore, the SWP discharge index does not have units), namely:

$$p_r={\displaystyle \sum }_{\mathrm{j}=1}^{12}\frac{p_{rj}-p_{j,min}}{p_{j,max}-p_{j,min}}{w}_j$$

(1)

where $p_r$, $p_{rj}$, $p_{j,max}$, $p_{j,min}$, and $w_j$ denote the SWP discharge index of region r, the jth water pollutant’s actual discharge in region r, the maximum value of the jth water pollutant discharge in all regions, the minimum value of the jth water pollutant discharge in all regions, and the total ranking weight of the jth water pollutant in the hierarchy, respectively.

2.3. Net Virtual Transfer Accounting

Based on the characteristics of a non-competitive, inter-provincial, input–occupancy–output model, Tan et al. [23] multiplied the diagonal matrix of direct water use coefficients with the combined matrix of intermediate inputs and final demand to obtain the total virtual flow matrix of water. The difference between this matrix and its transpose represents the difference between the two-way virtual flows. Thus, the negative numbers are set to zero to obtain an allocation matrix representing the one-way net virtual transfers. Ultimately, the water net virtual flow of pollutant-receiving and pollutant-exporting regions are the column sum and row sum of the distribution matrix, respectively. The difference between the corresponding elements of the two is the net virtual transfer of inter-provincial water consumption. This paper draws on this calculation method to further account for the SWP regional net virtual transfer.

First, we define the direct SWP discharge intensity to calculate the virtual SWP trade for region r:

$$d_r=\frac{p_r}{ x_r}$$

(2)

where $d_r$ is a direct SWP intensity vector, representing the direct SWP discharge per unit of output for each region r; $p_r$ is a vector of SWP discharge for region r and r is the total output in region r.

Second, referring to the approaches of “emissions embodied in trade” (EET) [33] and “water use embodied in trade” (WET) [14], we combine the method of Tan et al. (2021) to illustrate the SWP impact of production and consumption in the region, and inter-regional product transfers separately. Here, $a_r$ denotes the ith element of the main diagonal of the Leontief technical coefficient matrix $A$ (Appendix B), i.e., $a_r=A_{rr}$, and $t_r$ is the total SWP discharge intensity, i.e., the total SWP discharge in the region r production of a unit of product. Then, the total SWP discharge intensity vector is:

$$t=\left[d_1{\left(1-a_1\right)}^{-1},\cdots ,d_r{\left(1-a_r\right)}^{-1},\cdots ,d_n{\left(1-a_n\right)}^{-1}\right]$$

(3)

The virtual SWP due to intermediate and final demand can be calculated as follows:

$$vpi=\mathrm{diag}\left(t\right)z$$

(4)

$$vpf=\mathrm{diag}\left(t\right)f$$

(5)

where the $vpi$ and $vpf$ are the vectors of virtual SWPs resulting from the intermediate and final demand, respectively; $\mathrm{diag}\left(·\right)$ denotes the diagonalization operator (If it is used for a matrix, it represents the matrix’s primary diagonal element row vector). Thus, the virtual SWPs due to the regions’ production and consumption, inter-regional demand-induced imports are:

$$SVPI=\mathrm{diag}\left(vpi\right)$$

(6)

$$IVPI=\left(vpi-\mathrm{diag}\left(vpi\right)\right) \mathrm{i}$$

(7)

$$SVPF=\mathrm{diag}\left(vpf\right)$$

(8)

$$IVPF=\left(vpf-\mathrm{diag}\left(vpf\right)\right)\mathrm{i}$$

(9)

where $SVPI$, $IVPI$, $SVPF$, and $IVPF$ denote the virtual SWP transferred from the region’s production, inter-regional product production, the region’s final demand, and inter-regional final demand, respectively; $\mathrm{i}$ is a column vector with all elements of 1.

We express the unidirectional net flow of virtual SWP as:

$$unf=max\left(0, \left(vpi+vpf\right)-{\left(vpi+vpf\right)}^{\prime }\right)$$

(10)

Ultimately, a province’s net virtual flow quantity (NVFQ) is the difference between the receiving net virtual SWP flow and the exporting net virtual SWP flow:

$$NVFQ=unf·i-i^{\prime }unf$$

(11)

2.4. The Policy Simulation Approach Using an MRIO Table

The accounting of regional net virtual transfers in Section 2.3 categorizes all regions into net virtual receiving and net virtual exporting regions. Due to internal and external production and consumption, regions with high net virtual receiving are affected by the “SWP embodied in trade” (SWPET). Measures such as restructuring product import and export among regions, expanding international imports, and reducing discharge in provinces with high SWP per product unit are likely to offer solutions for reducing SWP discharge that are conducive to mitigating this impact. Hence, we propose four simulation scenarios (

Table 4. The proposed simulation scenarios.

Scenario	Description
${\mathrm{S}}_1$	The regions with net virtual receiving of SWPs greater than zero reduce the intermediate demand by 10% (i.e., the element u of the corresponding province in Table A3 is reduced by 10%). By contrast, regions with a net virtual export of SWP less than zero reduce the primary input by 10% (i.e., the element v of the corresponding province in Table A3 is reduced by 10%).
${\mathrm{S}}_2$	We add a 10% increase in international imports ($\widetilde{im}$) to all regions under the settings of the scenario ${\mathrm{S}}_1$.
${\mathrm{S}}_3$	We assume that the top ten regions in terms of SWP discharge intensity have a 10% reduction in SWP emission intensity based on the total output $\tilde{x}$ of scenario ${\mathrm{S}}_2$.
${\mathrm{S}}_4$	The annual growth rates of investment in wastewater treatment and recycling in cities, counties, established towns, townships, and villages in each region of China for the 14th Five-Year Plan period (2021–2025) are 15.27%, 21.99%, 24.22%, 27.29%, and 30.53%, respectively. These figures represent the compound annual growth rate (CAGR) for China’s 13th Five-Year Plan period (2016–2020).

).

Under the MRIO structure of Table A3, the following balance equation holds:

$$x=i^{\prime }·z+va+im$$

(12)

See Table A3 for the meaning of other symbols. Thus, to investigate whether changing the product production structure of regions can reduce the discharge of SWP, we assume that regions with different net virtual transfer types of integrated water pollutants adjust their production structure according to different directions. This is the simulation scenario ${\mathrm{S}}_1$ (Table 4).

We change u and v technical and structural changes, thus requiring re-estimating the intermediate matrix using the GRAS method (Appendix B). It was initially proposed by Günlük-Senesen et al. [34] to generalize the RAS method [35]. We use the 2017 intermediate demand as the benchmark (base) matrix in the GRAS method. Under the new production structure, the total SWP discharge is:

$$p_{S_1}=d \tilde{x}=d\left(i^{\prime }·\tilde{z}+va+im\right)$$

(13)

where $\tilde{x}$ and $\tilde{z}$ denote the estimated total output vector and the intermediate input matrix obtained by GRAS under scenario ${\mathrm{S}}_1$, respectively.

In scenario ${\mathrm{S}}_2$, we add international imports on top of the scenario ${\mathrm{S}}_1$. The total SWP will be:

$$p_{S_2}=d \tilde{x}=d\left(i^{\prime }·\tilde{z}+va+\widetilde{im}\right)$$

(14)

For scenario ${\mathrm{S}}_3$, $p_{S_3}=\tilde{p}·i$, where for region r, there is:

$$\widetilde{p_r}=0.9\widetilde{x_r}\frac{p_r}{x_r}$$

(15)

Finally, we assess whether the country can achieve the objectives of the environmental infrastructure upgrading project proposed in the comprehensive work plan of the 14th Five-Year Plan for energy conservation and emission reduction. In other words, by 2025, we will add 20 million cubic m/day of wastewater treatment capacity.

The compilation of the MRIO for 2017 allows considering the impact of changes in exogenous variable final demand on the compensation of employees, employment, output, value added and GDP, tax revenue, electricity use, CO₂ emissions, wastewater treatment, and recycled water production in each region. To this end, we assume that the technical coefficients between regions are the same as those in 2017. The wastewater treatment and recycling investments in cities, counties, established towns, townships, and villages grow at annual compound growth rates (

Table A4. The compound growth rate of investment in water supply, drainage, wastewater treatment, and recycling in cities, counties, established towns, townships, and villages from 2015 to 2020.

	City	County	Established Town	Township	Village
Water Supply (%)	3.87	8.23	2.19	1.98	9.21
Drainage (%)	16.57	16.11	17.66	18.45	20.33
Wastewater Treatment and Its Recycling (%)	15.27	21.99	24.22	27.29	30.53

). It then accounts for the effects on each region’s employed population, value added, electricity use, and wastewater treatment and recycling capacity. Namely,

$${\Delta }_m=A_m{\left(I-A\right)}^{-1}\Delta f$$

(16)

where m denotes the indicators of employment, value added, electricity consumption, and wastewater treatment and recycling capacity; ${\Delta }_m$ is the amount of change in indicator m. $A_m$ represents the coefficient row vector of indicator m, i.e., the ratio of the value of regional indicator m to total output; $\Delta f$ represents the matrix of changes in final demand. Here is the matrix of changes in regional wastewater treatment and recycling after the increase in investment.

The main reason why Equation (16) holds is that there is ${\Delta }_m=A_m{\left(I-A\right)}^{-1}{f}^{sim}-A_m{\left(I-A\right)}^{-1}f=A_m{\left(I-A\right)}^{-1}\Delta f$ according to Equations (4) and (5). This is the procedure of simulation scenario ${\mathrm{S}}_4$.

3. Results

3.1. Water Pollutants Discharge Status

According to the data published by Tian et al. [36], livestock farming and the livestock product industry is the largest source of COD, NH₃-N, and phosphorus discharge in China. Crop farming is the second largest source of NH₃-N and phosphorus discharge. Fisheries account for the second and third largest agricultural discharge sources of COD and phosphorus, respectively. In China, organic matter and more minor frequently contains heavy metals dominate livestock and fish farming tailwater. The tailwater discharge without adequate treatment is perhaps the root cause of large quantities of COD, NH₃-N, and phosphorus discharges. Coupled with the fact that nitrogen and phosphorus can form a surface diffusion source through the ground and soil and are not easily collected, the large amount of nitrogen and phosphorus fertilizers used in the process of crop cultivation has resulted in NH₃-N and phosphorus pollution of water. The data published in the Annual Report on Ecological Environment Statistics of China in previous years show that China’s primary industrial discharge sources of COD, NH₃-N, TN, and TP are agricultural and food processing [37], chemical raw materials and chemical product manufacturing, the textile industry, paper product industry, and food manufacturing. For example, in 2020, among the 42 industrial industries surveyed by the Ministry of Ecology and Environment, the top three industries in terms of COD, NH₃-N, TN, and TP discharge were chemical raw materials and chemical product manufacturing, agro-food processing, and the textile industry. In 2020, they together discharged 172.00 thousand tons of COD, 8.00 thousand tons of NH₃-N, 8.00 thousand tons of TN, and 1.00 thousand tons of TP, accounting for 39.70%, 42.90%, 43.7%, and 46.50%, respectively, of the discharge. Among them, chemical raw materials and chemical product manufacturing industry’s NH₃-N and TN discharge accounted for 22.50% and 20.90%, respectively, both ranked first. The textile industry discharged 14.00% of the country’s COD, followed by chemical raw materials and chemical product manufacturing, and agricultural and food processing industries. The agricultural and food processing industry discharged the most considerable amount of total phosphorus, accounting for 26.20%. These results are consistent with the “Second National Pollution Source Census Bulletin” data [38]. Tian et al. [36] provides a more detailed distribution of COD, NH₃-N, and phosphorus discharge by industry.

As stated above, the chemical raw materials and chemical product manufacturing industry are important contributors to water pollutant discharge. In China’s gradually aging society, the growth of demand for medical treatment will inevitably drive the development of chemical raw materials and chemical product manufacturing, which will further increase discharge. In the future, more attention should be paid to the effective treatment of wastewater from these sectors.

3.2. SWP Discharge Index

Based on the calculation of the SWP discharge index (Equation (1)) and the method of net virtual transfer accounting (Equations (2)–(11)), the spatial–temporal synthetic evolution pattern of the overall discharge and transfer of multiple pollutants caused by inter-regional trade can be measured using the prepared MRIO tables for 2012 and 2017. Combining the weights of the six water pollutants in Table 3 and the calculation of Equation (1), the SWP discharge index in 2012, 2017, and 2020 were 11.18, 9.62, and 7.32, respectively. Compared with 2012, the decrease in 2020 was 52.73%. This indicates that the scale of the synthetic discharge index of six water pollutants is declining nationwide (Figure 1). From a regional point of view, Shandong, as one of China’s major agricultural cultivation provinces, has been at the top of the SWP discharge index. Developed agriculture is perhaps the main cause of this phenomenon, as Wang et al. [13] and Tang et al. [39] point out that the agricultural sector is mainly responsible for COD emissions. Meanwhile, the weight of TP in SWP is as high as about 0.46. Furthermore, Shandong is a high TP emitting province, which may be related to the high use of phosphorus fertilizer in agriculture. Heilongjiang, which is also a large agricultural province, had a discharge index of 0.24 in 2020, only 58.54% of that of Shandong Province. In 2017 and 2020, Guangdong Province ranked second in the country on the SWP discharge index. Jiangsu Province, in a coastal area, had a 47.37% lower discharge index in 2020. China’s coastal regions have a developed economy, large population, high quantities of residential water consumption, and fishing is more developed, which may be the reason for the high discharge of SWP in Guangdong Province. In 2020, the Hubei Province SWP discharge index rose to first place, reaching 0.83, an increase of 102.44% from 2017. At this time, at a critical point in the prevention and control of the epidemic in Hubei, residents were in home isolation and increased domestic water consumption. The most stringent measures to contain the pandemic development in that year increased water use and the water pollutants discharged.

3.3. Temporal and Spatial Analysis of SWP Virtual Transfer

3.3.1. SWP Virtual Transfer Due to Production and Consumption

The multi-region input–output table separates primary inputs from different regions and is a non-competitive table. We use the products traded across regions in the table to partly satisfy intermediate sector production and partly satisfy end-user demand. Hence, we can follow Equation (6) in the Tan et al. [23] paper to measure the discharges generated by each intermediate and final demand, production, and consumption in the region, and inter-regional transfers in each of the SWP total regional virtual flows. The results are shown in Figure 2.

3.3.2. The SWP Direct Discharge and Net Virtual Transfer

Figure 3 shows the spatial and temporal evolution patterns of the net virtual transfer of SWP from inter-regional trade. Among them, the results for 2020 were accounted for under the same input–output structure as in 2017, using the actual regional discharge index of six water pollutants in 2020. In 2012, the net virtual receiving of SWP discharge in China was concentrated in western China. Qinghai, Xinjiang, and Guangxi ranked in the top three, with net virtual receiving at 0.39, 0.35, and 0.15, respectively (Figure 2a). The Beijing–Tianjin region, Hainan, Jiangxi, Guangxi, and Jilin were the main exporting regions to Qinghai and Xinjiang. Guangxi mainly received pollutants from Sichuan, the Beijing–Tianjin region, and Zhejiang (Figure 4a).

By 2017, the inland central regions, such as Anhui and Shanxi, and the western regions, such as Guizhou, Xinjiang, and Shaanxi, became the leading net virtual receiving regions of SWP. By comparison, Sichuan and Jiangxi maintained high net virtual exporting (Figure 4b). The Beijing–Tianjin region’s net virtual exporting scale slightly increased. From 2012 to 2020, the net virtual receiving became more concentrated. For example, Anhui and Jiangxi were the regions with the largest net virtual receiving in 2017 and 2020 at 0.71 and 1.01, respectively (Figure 4).

3.4. Simulating the SWP Reduction Scenarios

Table 5. Synthetic water pollutant reduction effects on all regions under simulation scenarios ${\mathrm{S}}_1$, ${\mathrm{S}}_2$, and ${\mathrm{S}}_3$.

Scenario	SWP Discharge Quantity	Decrease Quantity	Decrease Rate (%)
${\mathrm{S}}_1$	9.30	0.33	3.45
${\mathrm{S}}_2$	9.34	0.29	3.02
${\mathrm{S}}_3$	9.07	0.56	5.77

shows all regions’ SWP discharge, the reductions relative to 2017, and the reductions as a percentage of the 2017 discharge under scenarios ${\mathrm{S}}_1$ (Equation (13) based on Equation (12)), ${\mathrm{S}}_2$ (Equation (14)), and ${\mathrm{S}}_3$ (Equation (15)). The simulation results show that by reducing the intermediate demand for the net virtual receiving regions while reducing the inputs to the net virtual exporting regions, it is possible to mitigate all regions’ SWP discharge by 0.33, or 3.45% of the 2017 emissions. The reduction, however, still amounts to 0.29, or 3.02%, when superimposed on the 10% international imports of all regions. It is only a 0.04 decrease relative to the discharge reduction of scenario ${\mathrm{S}}_1$. The results of scenario ${\mathrm{S}}_3$ suggest that the most effective measure to reduce SWP discharge may be to reduce discharge from provinces with high SWP discharge per unit of output. According to our assessment, reducing discharge per unit of a product by 10% in the top ten regions in discharge intensity achieves a stacked reduction rate of 2.75% relative to scenario ${\mathrm{S}}_2$.

In China, the investment in fixed assets for wastewater treatment and recycling has maintained a high growth rate over the years. Especially in rural areas, the compound annual growth rate of investment in fixed assets for wastewater treatment and recycling was as high as 30.53% during the 13th Five-Year Plan. In 2020, the total wastewater treatment capacity of cities, counties, established towns, townships, and villages reached 293 million cubic meters per day [40]. Although guided by the policy, the potential social, economic, energy, and environmental effects of the high rate of fixed asset investment growth in wastewater treatment and its recycling need to be further quantified and analyzed. In particular, whether the country will be able to add 20 million cubic meters/day of wastewater treatment capacity by 2025.

According to the data from the Process Industry Website in 2022, the proportion of electricity consumption of wastewater treatment plants, industrial wastewater treatment, and sludge treatment in China is more than 2% of the total electricity consumption of the country. Therefore, 2% of the total electricity use in each region is incorporated here into the electricity consumption for wastewater treatment. Under scenario ${\mathrm{S}}_4$, the national investment in fixed assets for wastewater treatment and its recycling will increase by 97.69 billion Chinese yuan (CNY) by the end of 2025, which is about 45.45% of the total national investment in fixed assets for wastewater treatment and its recycling in 2020.

Table 6. Aggregate simulation results of the social, economic, energy, and environmental effects of scenario ${\mathrm{S}}_4$.

	Employment Increase (in 10 Thousand)	Value Added Increase (100 Million Yuan)	Electricity Consumption Increase (10 Thousand kWh)	Wastewater Treatment Capacity Increase (104 m3/day)
Total	74.90	825.44	14,359.60	29.24
	Total Wastewater Treatment Increase (104 m3)	Total Sludge Treatment Increase (tons)	Recycled Water Production Capacity Increase (104 m3/day)	Recycled Water Usage Increase (104 m3)
Total	6500.97	11,473.16	6.11	1427.85

shows that a 45.45% increase in national investment in fixed assets for wastewater treatment and recycling can create new employment for 748,000 people. It also pulls GDP growth of 82.54 billion CNY and increases electricity consumption by 143.59 million kWh. The country will increase 19 wastewater treatment plants, adding 292.41 thousand cubic meters per day of wastewater treatment capacity, putting the total amount of treated wastewater and sludge increase by 65.01 million cubic meters and 11,473.16 tons, respectively. The recycled water production capacity will increase by 61.11 thousand cubic meters per day, increasing to 14.28 million cubic meters of recycled water utilization.

Table A5. Simulation results of the social, economic, and energy effects of the growth of wastewater treatment and recycling investment in cities, counties, towns, townships, and villages in each region at a compound annual growth rate of 15.27%, 21.99%, 24.22%, 27.29%, and 30.53%, respectively.

Region	Employment (in 10 Thousand)	Value Added (100 Million Yuan)	Electricity Consumption (10 Thousand kWh)
BJ	0.43	10.31	83.87
TJ	0.25	7.31	68.96
HeB	2.61	24.17	559.14
SX	1.10	9.82	296.19
IM	0.81	9.69	510.01
LN	0.74	7.74	160.28
JL	0.60	7.07	76.84
HLJ	1.05	11.32	144.36
SH	1.41	31.51	324.18
JS	4.56	80.07	1188.68
ZJ	4.09	55.63	1024.79
AH	3.89	35.09	582.03
FJ	2.60	38.09	586.04
JX	2.82	25.25	405.13
SD	3.44	45.13	866.56
HN	4.72	43.08	656.00
HuB	3.36	36.59	442.24
HuN	3.87	39.99	455.11
GD	11.74	147.90	2311.08
GX	1.56	11.27	247.35
HaN	0.29	2.37	38.57
CQ	1.41	16.37	199.94
SC	7.44	58.11	898.26
GZ	2.21	15.84	371.06
YN	2.84	16.55	409.26
SAX	1.35	13.78	223.11
GS	2.00	11.04	414.36
QH	0.08	0.77	43.27
NX	0.25	2.37	151.89
XJ	1.36	11.21	621.02

and

Table A6. Simulation results of the environmental effects of the growth of wastewater treatment and recycling investment in cities, counties, towns, townships, and villages in each region at a compound annual growth rate of 15.27%, 21.99%, 24.22%, 27.29%, and 30.53%, respectively.

Region	Wastewater Treatment Capacity (104 m3/day)	Total Wastewater Treatment (104 m3)	Total Sludge Treatment (Tons)	Recycled Water Production Capacity (104 m3/day)	Recycled Water Usage (104 m3)
BJ	0.29	67.94	608.90	0.25	44.19
TJ	0.15	42.83	58.45	0.07	13.98
HeB	0.85	172.74	382.56	0.47	69.79
SX	0.34	81.36	217.01	0.19	19.90
IM	0.24	60.38	193.58	0.15	22.71
LN	0.38	108.44	184.62	0.08	11.57
JL	0.27	66.68	102.92	0.04	8.89
HLJ	0.40	97.54	130.59	0.03	19.14
SH	0.93	220.25	497.88	0.00	0.00
JS	2.41	470.27	976.54	0.53	124.56
ZJ	1.79	406.50	894.07	0.23	48.02
AH	1.44	320.66	392.66	0.39	101.80
FJ	1.18	212.41	301.01	0.20	34.20
JX	0.71	189.17	186.99	0.00	0.50
SD	1.44	258.03	569.65	0.49	111.11
HN	1.47	263.11	503.07	0.37	79.69
HuB	1.34	322.79	379.74	0.19	49.69
HuN	1.44	386.23	857.92	0.11	24.50
GD	6.33	1409.67	1956.53	1.44	467.89
GX	0.64	115.39	90.76	0.03	10.20
HaN	0.08	21.79	40.45	0.01	1.36
CQ	0.50	129.26	3.58	0.02	1.49
SC	2.18	482.08	736.42	0.23	53.98
GZ	0.64	136.90	145.56	0.04	7.33
YN	0.50	140.51	186.72	0.05	35.10
SAX	0.37	97.89	310.44	0.15	18.30
GS	0.38	92.86	204.98	0.11	11.72
QH	0.03	6.61	8.52	0.01	1.07
NX	0.12	25.38	53.24	0.04	5.40
XJ	0.42	95.30	297.80	0.19	29.75

list the impacts on employment, value added, electricity, and wastewater treatment and its recycled water use capacity for each region under the simulation scenario ${\mathrm{S}}_4$ (Equation (16)).

4. Discussion

This paper provides an SWP discharge index to measure the regional integrated discharge of multiple water pollutants. Under the input–occupancy–output framework, the associations formed by the synthetic transfer among regions and its impact is portrayed through this index. It also offers an approach to analyzing the synthetic virtual flow of multiple water pollutants among regions and the effects of implementing potential reduction scenarios nationally. The calculation of the SWP discharge index relies on the setting of reasonable water pollutant weights. Different weighting calculation methods have dramatically different measurements. Here, we refer to the identification of environmental hazards and human health hazards of water pollutants or structure-related substances at home and abroad, then give a more objective weighting. The results of our research provide SWP transfer relationships among regions. We also analyzed the carrier sectors of the water pollutant transfer by analytical means.

The results of other studies are reasonably comparable with the findings of this study. For example, agriculture is one of the significant sources of COD and ammonia discharges in China [37]. The study by Wang et al. [13] showed that China’s COD discharge was mainly from the agricultural sector, followed by the household sector. Tang et al. [39] considered the marginal abatement costs of COD discharge from agricultural production under three alternative scenarios using a directional distance function approach. Their analysis showed that China has room for simultaneous improvement in agricultural production, pollution reduction, and resource conservation. Agriculture also has more benefits in COD discharge reduction compared to other sectors. Liu et al. [41] analyzed the spatial characteristics and drivers of anthropogenic phosphorus discharge in China’s Yangtze River Economic Zone, with agriculture being the largest source of total phosphorus. The high-value regions of total phosphorus discharge are Hunan, Hubei, Jiangsu, Anhui, and Jiangxi. Our study also points out that regions (i.e., Hubei, Hunan, Shandong, and Guangdong) with high SWP discharge should pay more attention to the reduction of agricultural and domestic sources.

Researchers can use the ideas provided in this paper to analyze non-water pollutants, for example, air pollutants, synthetically. First, it is necessary to compare the extent of different hazards of pollutants. Based on this information, we give a reasonable ranking and calculate the weights of pollutants according to an appropriate method (e.g., the analytic hierarchical process). In this case, the synthetic pollutant index is the weighted value of regional pollutant emission data after normalization. The synthetic pollutant index is used as the occupancy vector of the MRIO table to account for the synthetic net virtual transfer. Although limited by the availability of data and the potential subjectivity and uncertainty associated with the process of calculating pollutant weights, we have the flexibility to cheaply calculate the weights based on more scientific criteria and well-developed data. The ideas provided in this paper have the value of wide application and expansion. It is a quantitative complement for the assessment of sustainable, high-quality green development to improve the understanding of green and coordinated development in an economic system.

5. Conclusions

The main findings of the study are as follows.

(1)

The scale of synthetic discharge of six water pollutants, namely, COD, NH₃-N, TN, TP, P, and VP, is declining nationwide. Regionally, Shandong Province has been at the forefront of the country in terms of the SWP discharge index. Heilongjiang and Hebei, which are also large agricultural provinces, have significantly lower SWP discharge indices than Shandong Province. From 2017 to 2020, the coastal Guangdong Province had the second-highest SWP discharge index. Jiangsu Province showed a significant reduction in its discharge index. For different regions with similar economic structures and geographic locations, high-discharge regions can learn from the advantageous experience of low-discharge regions to reduce synthetic discharge. China’s environmental regulation is adept at summarizing good experiences, making it likely that similar economic structures and geographic locations are intrinsic to making this solution easier to implement.

(2)

From 2012 to 2020, the net virtual pollutant-receiving became more concentrated. The Anhui and Jiangxi were the regions with the largest net virtual receiving in 2017 and 2020, respectively. Central (e.g., Shanxi and Hunan) and western (Xinjiang, Inner Mongolia) areas are China’s main regions of net virtual receiving. The SWPET intensifies the stress of water pollutant discharge in these regions. By contrast, it alleviates the water pollutant discharges in the net virtual exporting regions, as shown in the Beijing–Tianjin region and Hainan. The net virtual receiving and net virtual exporting concentrations can strengthen cooperation to enhance the responsibility of water pollutant reduction from the supply chain source. Regions with strong supply chain dependencies have a common demand for centralized management and lower long-term costs. Strengthening the responsibility for water pollution reduction among strongly dependent regions may reduce short-term gains, but it can enhance social value and bring more long-term value.

(3)

Restructuring inter-regional product production, expanding international imports, and reducing discharge in regions with high pollutant discharge per product unit have the potential effects to mitigate SWP discharge in all regions. Among them, reducing 10% of inputs from net virtual importing regions to the other regions while reducing 10% consumption from net virtual exporting regions to the other regions can reduce the SWP discharge of the economic system by 3.45%. This reduction rate is 3.02% under the scenario of increasing 10% international imports in all regions. Reducing the intensity of SWP discharge is the most direct and effective approach. For example, a 10% reduction in SWP discharge per output unit in the top 10 regions in discharge intensity would result in an incremental SWP reduction of 2.75%. However, the growth of fixed asset investment in wastewater treatment and its recycling has limited contribution to achieving China’s policy target of adding 20 million m³/day of wastewater treatment capacity by 2025, with only 292,400 m³/day, or 1.46%, of new wastewater treatment capacity. To achieve the policy target by 2025, improvement is needed in wastewater treatment capacity from raw materials, as well as a move toward more advanced technologies for wastewater treatment. The biggest challenge in achieving this improvement may be the high cost of raw materials and advanced technologies. Their applicability also hinders their wider diffusion. It is believed that these problems will be gradually alleviated as investments in water pollutant treatment increase and older equipment are replaced.