Integrating Water Quality Restoration Cost with Ecosystem Service Flow to Quantify an Ecological Compensation Standard: A Case Study of the Taoxi Creek Watershed

Abstract

Watershed ecological compensation is an important economic tool for solving the protection–development conflict. However, establishing a sound ecological compensation plan for rational water resource use at the watershed scale remains challenging. Monthly water samples were collected between April 2019 and July 2020 at 28 points to analyze the spatiotemporal variation in water quality in the Taoxi Creek watershed, which is important for local water security. The Soil and Water Assessment Tool was used to simulate water supply, demand, and spatial flow at the watershed scale. Water quality restoration cost was integrated with ecosystem service flow to quantify the watershed ecological compensation. The ecological compensation using water quality restoration cost based on pollutant treatment cost and water quality target level was CNY 11.9 million (USD 188 million). Taoxi Creek was identified as the major supplier of water ecosystem services for downstream residents, and the ecological compensation based on ecosystem service flow was CNY 18.9–47.2 million (USD 3.0–7.5 million). Combining both calculations, the watershed should annually receive CNY 30.8–59.1 million (USD 4.9–9.3 million) of economic compensation from downstream ecosystem beneficiaries. This study provides a scientific basis for improving the ecological compensation scheme in the Taoxi Creek watershed and a reference for ecological compensation formulation in other watersheds.

1. Introduction

Watershed ecological compensation is an important tool for measuring the value of water resources in natural–social systems [1], which can help alleviate the conflicts of interest arising between different administrative regions upstream and downstream of the watershed by considering water resource utilization [2,3,4]. From the perspective of ecosystem services, exploring the spatial and temporal flow processes of water resources in natural–social systems and establishing a basin ecological compensation mechanism based on water pollution in the basin can help alleviate the conflicts of water use within a basin and ultimately maximize the interests of all parties involved [5,6]. To realize sustainable water resource utilization in basins, a set of scientifically quantified ecological compensation schemes is of great significance.

The quantification of compensation is at the core of ecological compensation work and is key to establishing an effective watershed ecological compensation mechanism [7], which is directly related to the scientific nature and implementation effect of watershed ecological compensation. The compensation not only reasonably reflects the ecological value provided by the ecosystem service suppliers [8] but also is accepted by all parties involved to effectively correct the economic relationship between upstream and downstream stakeholders in the watershed [9]. At present, scenarios for quantitative watershed ecological compensation standards are roughly divided into three types: (1) The compensation amount is established based on input costs, which are determined by measuring the total ecological and environmental costs, including direct costs and opportunity costs [10,11], such as direct inputs of watershed water environment management, and the loss due to limitations of economic development to protect the ecological environment [12]. (2) The compensation amount is established based on environmental benefits by accounting for the value of the ecosystem services in the watershed, i.e., by monetizing the ecosystem services. Since the proposal of valuation of ecosystem services [13], there have been numerous cases of ecological compensation based on the value of ecosystem services [14,15,16]. (3) The compensation amount is established based on willingness to pay, and the maximum or minimum willingness to pay is determined by resident surveys using a questionnaire [17,18]. The conditional valuation method is also this method [19]. Thus, there are various methods, with different theoretical bases, to quantify ecological compensation in watersheds. However, ecological compensation based on input cost or ecosystem service values cannot reflect the transregional ecological benefits [20]. The willingness to pay relies too much on respondents’ subjective perceptions ignoring hydrological processes [8]. Overall, most compensation criteria do not integrate ecological processes and socioeconomic impacts to establish direct links with transregional water consumers [21].

Ecosystem service flow reveals the transregional benefits transferred from service supply areas to benefit areas, corresponding to the sellers and buyers of interregional ecological compensation [5,22,23]. However, ecological compensation usually ignores this close connection.

Water supply service was often used as a case for ecological compensation with typical transregional flow properties [5]. Water value assessment methods are also used to comprehensively and objectively assess the value of water resources from different dimensions and perspectives, and assessment outcomes objectively reflect the value of water resources to a certain extent [24]. Only water resources that meet water quality standards have ecosystem service values. Few studies have integrated water quality restoration costs from the perspective of the spatial flow of water supply services to calculate the amount of ecological compensation [25]. The water environment varies among different watersheds, and it is necessary to assess the value of ecosystem services considering water pollution. Furthermore, ecological compensation schemes are formulated based on ecosystem service flow and water pollution.

Therefore, in this study, we monitored water quality for 16 months, performed ecosystem service flow analyses, and quantified the ecological compensation for Taoxi Creek. This creek is located in the upper reaches of the Jinjiang watershed and is an important water source; however, water pollution can affect downstream water resources. We used the Soil and Water Assessment Tool (SWAT) model [26,27] and geographic information system (GIS) software to visualize and simulate the supply and demand balance of ecosystem service flows in the watershed and quantified the ecological compensation based on water quality restoration cost and ecosystem service flows. Finally, we integrated the water quality restoration cost with the ecosystem service flow to quantify the ecological compensation standard for the Taoxi Creek watershed. Our aim was to provide a reference for the establishment of an ecological compensation system in Taoxi Creek and the management of water resources in the watershed by the local government.

2. Methods

2.1. Study Area and Data

Taoxi Creek is located in the upper reaches of the Jinjiang watershed in Fujian Province (24°31′–25°32′ N, 117°44′–118°47′ E). It has a main channel length of 61.75 km and a creek area of 476 km². The creek flows through 11 townships in Yongchun County, Quanzhou City, with a total population of 319,000 living in the basin. Taoxi Creek is surrounded by mountains on three sides, with a basin topography of high all around and low in the middle. The elevation ranges between 56 m and 1365 m. The basin comprises two main streams, the Taoxi and Huyang creeks, which converge in Dongguan Town and flow toward the Shanmei Reservoir, providing water to more than three million people downstream of the Jinjiang watershed and representing the main water source for the downstream counties and cities (Figure 1). Nitrogen and phosphorus nutrient pollution in the Taoxi Creek watershed is severe, and the sources of pollution are diverse [28]. With the development of the livestock farming industry, nutrients from livestock manure and feed enter the river with the wastewater. Excessive fertilizer and pesticide applications lead to non-point-source pollution. There are many dry toilets but no adequate sewage treatment facilities. The low sewage treatment capacity leads to a large amount of domestic sewage discharged into rivers without treatment.

As the water source in the upper reaches of the Jinjiang watershed, the quality of water in the Taoxi Creek watershed determines the safety of drinking water in downstream areas. Therefore, it is necessary to set ecological compensation standards by taking water quality into consideration and to use economic means to achieve the purpose of watershed environmental protection.

The data used in this study mainly included water system data, administrative division data, digital elevation model (DEM) data, meteorological data, land use data, soil data, hydrological data, and socioeconomic statistics of Taoxi Creek. The resolution of the land use data and DEM was 30 m. The meteorological data from January 2005 to December 2019, including precipitation, temperature, wind speed, relative humidity, and solar radiation, were obtained from the China Meteorological Assimilation Driving Datasets (http://data.cma.cn/, accessed on 31 December 2019) for the SWAT model v1.1. The soil data were obtained from the Chinese soil data set of the Harmonized World Soil Database (https://previous.iiasa.ac.at/web/home/research/researchPrograms/water/HWSD.html, accessed on 31 December 2019). The hydrological data were the monthly runoff from January 2005 to December 2019 measured at three hydrological stations in the Jinjiang watershed. In addition, we used the water resource bulletins (http://slj.quanzhou.gov.cn/zwgk/tzgg/202111/t20211119_2655072.htm, accessed on 31 December 2019) provided by the Yongchun County Water Resources Bureau and the Environmental Protection Bureau and the statistical yearbooks of the governments of Quanzhou City and Yongchun County. For the socioeconomic statistics, we mainly used data on population size, total economic value, and water consumption by sector (http://www.quanzhou.gov.cn/zfb/xxgk/zfxxgkzl/tjxx/ndtjsc/, accessed on 31 December 2019). The unit price data of water resources was provided by Fujian Water Resources Department (http://zfgb.fujian.gov.cn/5583, accessed on 31 December 2019). The water quality restoration cost was derived from the Yongchun County Rural Domestic Sewage Treatment Special Plan (2019–2030), which is available on request from the Yongchun County government (http://www.fjyc.gov.cn/zwgk/zfxxgkzl/ndbg/, accessed on 31 December 2019).

2.2. Water Quality Monitoring

Twenty-eight water quality monitoring points were set up along Taoxi Creek; 14 were township cross-sectional points, and the other 14 were tributaries and main-stream points. The sampling period was April 2019 to July 2020, and sampling was performed once at the end of the month. The factors monitored were NH₄⁺-N, NO₃^–-N, and total phosphorus (TP). The sampling point locations and names are shown in Figure 2.

After obtaining the water quality pollutant data, Class II water quality was set as the target water quality level for Taoxi Creek according to the Environmental Quality Standards for Surface Water (GB 3838-2002) [29]. Based on the cost of pollutant treatment at the Yongchun County wastewater treatment plant and the water quality of the polluted water, the cost of restoring the water quality to the target level was calculated and was considered in the ecological compensation.

We used the water quality restoration cost to calculate the water quality compensation [30]. We calculated the amount of pollution in reference to the target water quality and multiplied it by the section flow rate and the pollutant restoration cost to derive the total compensation amount. This method considers the migration law of pollutants from upstream and downstream cross-sections of the river from a macroscopic perspective. The following formula was used to calculate the compensation amount:

$$P_{WQEC}=W_{over} R=\left[\left({\rho }_{actual}-{\rho }_{target}\right)Q\right] R,$$

(1)

where $P_{WQEC}$ is the compensation amount of a single pollution indicator, $W_{over}$ is the actual amount of water with a quality higher than the target water, ${\rho }_{actual}$ is the water quality (concentration) measured at the point, ${\rho }_{target}$ is the target water quality (concentration) at the point, $Q$is the flow rate at the point, and$R$ is the unit pollutant reduction cost. The flow rate was obtained from the Ecological Environmental Bureau of Yongchun County and was measured in March and September 2019. The compensation amount was calculated for wet and dry seasons.

2.3. Supply and Demand Models of Ecosystem Service Flows

Water supply service (WSS) is critical for regional water balance and water circulation; however, the relationship between WSSs and human well-being has not been established [31].

2.3.1. Water Supply

Calculating the water yield in the Jinjiang watershed is the basis for conducting service flow studies. We used the SWAT model [26] to calculate the direction of water flow and confluence accumulation based on the DEM data of the filled pits, generate the river channels, and divide the entire watershed into 33 sub-basins. The water yield in each sub-basin was calculated based on the water balance method, i.e., the total amount of water injected into the river channel from a unit of basin area in a given period. Land use data from 2013, 2016, and 2019 were used to analyze the spatial and temporal variations in WSSs in the Jinjiang watershed and were input into the SWAT model to simulate the water supply under different land use patterns in different periods.

(1) SWAT model

The SWAT hydrologic model is a watershed-scale model developed by the U.S. Department of Agriculture that combines several hydrologic models [26]. It has a strong hydrologic–physical mechanism, is suitable for simulating watershed runoff over a long time series, and is globally the most widely used hydrologic model [32,33]. The SWAT model simulates hydrological processes based on the following water balance equation:

$$Q_{WYLD}=Q_{Surf}+Q_{Lat}+Q_{GW}+Q_{TLoss}$$

(2)

where $Q_{WYLD}$ is the water yield (mm), $Q_{Surf}$ is the surface runoff (mm), $Q_{Lat}$ is the lateral flow (mm) within the HRU soil profile entering the main river channel within the time step, $Q_{GW}$ is the time step (mm) of water from shallow aquifers entering the main channel, and $Q_{TLoss}$ is the water loss (mm) of tributaries transported through beds in HRU.

(2) Model calibration and validation

Because of the large number of parameters included in the SWAT model, proper screening is required prior to parameter calibration and validation. Sensitivity analysis can be used to derive the magnitude of the effect of each small parameter change on the model output, which can be used to determine the parameters that have a greater impact on runoff and thus form the optimal combination of model parameters [33].

After the determination of the sensitive parameters, runoff data obtained at three hydrological stations in the Jinjiang basin (Shilong, Shanmei Reservoir, and Anxi) were used for calibration and validation. The determination coefficient R², Nash–Sutcliffe efficiency coefficient (NSE), and percent deviation (PBIAS) of runoff simulations for the hydrological stations were used to comprehensively evaluate the simulation results. The NSE reflects the magnitude of the fit between the simulated and measured values, and its value range is from –∞ to 1. The decision coefficient R² reflects the consistency of the trends between the simulated and measured values, and its value ranges from 0 to 1; the closer to 1, the better the trend correlation. According to previous studies [34,35,36], the model results are acceptable when the three indicators simultaneously satisfy R² > 0.6, NSE > 0.5, and PBIAS < 25%.

2.3.2. Water Demand

Water demand considers the demand and consumption of water resources for human activities, including living and production, and reveals the spatial distribution characteristics of water demand, excluding the loss of surface water due to natural processes, such as vegetation absorption and utilization, river retention, and infiltration [37]. The water demand model used in this study included three main categories: agricultural water, industrial water, and residential water [38] (subdivided for rural and urban residents). Water consumption was calculated as follows:

$$\begin{gathered} W_{\left(x\right)}=P{W}_{\left(x\right)}+GDP{W}_{\left(x\right)}+AG{W}_{\left(x\right)} \\ =P_{\left(x\right)}\times P{W}_{\left(xper\right)}+GD{P}_{\left(x\right)}\times GDP{W}_{\left(xper\right)}+A{G}_{\left(x\right)}\times AG{W}_{\left(xper\right)}, \end{gathered}$$

(3)

where $P_{\left(x\right)}$ indicates the population, $P{W}_{\left(xper\right)}$ indicates the per capita domestic water consumption of residents, $GD{P}_{\left(x\right)}$indicates the output value of CNY 10,000 gross domestic product, $GDP{W}_{\left(xper\right)}$ indicates the water consumption per CNY 10,000 gross domestic product, $A{G}_{\left(x\right)}$ indicates the area of irrigated farmland, and $AG{W}_{\left(xper\right)}$ indicates the average water consumption of irrigated farmland.

To match the scale of water supply and obtain the water demand at the sub-basin scale, we allocated the water demand calculated by county administrative district to each sub-basin in ArcGIS by means of the zonal statistics tool.

2.3.3. Supply and Demand Balance and Ecosystem Service Flows

The spatial flow of WSSs is a directional ecosystem service flow with water flow as the carrier and the river as the flow path. To quantitatively simulate the spatial flow process and path of WSSs in the Jinjiang watershed, we used the sub-basins as the connecting scale and the Jinjiang watershed system network as the connecting object to visualize the spatial flow and spatial pattern of the water supply and demand balance in the study area [38,39].

The water resource supply:demand (S:D) ratio in a watershed is an important indicator for measuring the regional water supply–demand balance when analyzing the spatial relationship between the supply and demand of ecosystem services [40,41]. In this study, because of the great variability of water resources in different river sections and the fact that the amount of water produced varies greatly with precipitation from year to year, we used the freshwater security index (FSI) [38], i.e., the logarithm of the S:D ratio, to improve the spatial visibility and comparability of the differences in supply and demand.

$$FS{I}_i=lg\left(\frac{S_i}{D_i}\right),$$

(4)

where$i$ is the number of the sub-basin, $S_i$ is the water supply of sub-basin $i$, and $D_i$ is the water demand of sub-basin $i$. An FSI value > 0 indicates that the supply of water resources in the basin is greater than the demand and represents a supply area, i.e., an area that provides a range of ecosystem goods and services at a given time; an FSI value < 0 indicates that the supply of water resources in the basin is less than the demand and represents a benefit area, i.e., an area in which all ecosystem goods and services are consumed and used at a given time.

2.3.4. Cost Estimation of Water Supply Service Flow

We used the unit input cost for water maintenance as the unit compensation cost for water quantity. The unit input cost for water maintenance was calculated based on the total amount of water supply service flow, which was used as the lower limit of the water compensation standard.

The market price of water resources reflects the end price of water in the entire natural system, which eventually enters the social system and is used by people. We considered the unit market price of water resources as the upper limit of the water compensation standard. Referring to the Fujian Province Water Resources Levy and Use Management Measures issued by the Fujian Water Resources Department in 2014, the unit price of water resources in Quanzhou City, where Taoxi Creek is located, is 0.06 CNY/m³ (0.009 USD/m³), and the unit input cost of maintaining water quantity was 0.024 CNY/m³ (0.004 USD/m³). Thus, we used the water maintenance cost and the unit price of water resources as the upper and lower limits for calculating the water quantity compensation standard.

$$P_{WSSF-upper}={\mathrm{ESF}}_{WS} R_{upper}=\left(E{S}_{supply}-E{S}_{demand}\right) R_{upper}$$

(5)

$$P_{WSSF-lower}={\mathrm{ESF}}_{WS} R_{lower}=\left(E{S}_{supply}-E{S}_{demand}\right) R_{lower}$$

(6)

where $P_{WSSF-upper}$ is the upper compensation amount of WSS flow; $P_{WSSF-lower}$ is the lower compensation amount of WSS flow; $ES{F}_{WS}$ is the water supply service flow; $E{S}_{supply}$ is the supply of water supply service; $E{S}_{demand}$ is the demand for water supply service; $R_{upper}$is the unit price of water resources, which is 0.06 CNY/m³ (0.009 USD/m³); and$R_{lower}$ is the unit price of maintaining water quantity, which is 0.024 CNY/m³ (0.004 USD/m³).

$$P_{upper}=P_{WSSF-upper}+P_{WQEC}$$

(7)

$$P_{lower}=P_{WSSF-lower}+P_{WQEC}$$

(8)

where $P_{upper}$ is the upper limit of the total compensation amount and $P_{lower}$ is the lower limit of the total compensation amount.

Through combination with Formula (1) of Section 2.2, we integrated water quality restoration cost and compensation of WSS flow. Formulas (7) and (8) calculate the upper and lower limits of compensation amount in the Taoxi Creek watershed, respectively.

3. Results

3.1. Spatial and Temporal Variations in Water Quality

Analysis of the nitrogen pollution in Taoxi Creek revealed that the main nitrogen pollution compounds in Taoxi Creek were NO₃^–-N and NH₄⁺-N (Figure 3). The spatial distributions of the nitrogen compounds were similar in the upper and lower reaches of the watershed. NO₃^–-N accounted for more than 50% and NH₄⁺-N contributed less than 50% of total pollution; however, in the lower reaches of the watershed, NH₄⁺-N concentrations at the sampling points J06, J08, J09, J10, J11, T12, T13, and T14 were higher than those at other sites.

The intermonth variation in the three pollution indices presented obvious characteristics (Figure 4). The seasonal variation in the TP concentration was in the order winter > spring > summer > autumn, with the highest average concentration in winter (0.19 mg/L) and the lowest in autumn (0.13 mg/L). The seasonal variation in the NH₄⁺-N concentration was in the order winter > autumn > spring > summer, with the highest average concentration in winter (1.17 mg/L) and the lowest in summer (0.47 mg/L). The seasonal variation in the NO₃^–-N concentration was in the order summer > spring > autumn > winter, with the highest average concentration in summer (1.92 mg/L) and the lowest in winter (1.51 mg/L).

3.2. Ecological Compensation Considering the Water Quality Restoration Cost

According to the Class II water quality standard, the concentration standards of NH₄⁺-N and TP are 0.5 (mg/L) and 0.1 (mg/L), respectively. We counted 28 points of water quality standards; 20 of 28 exceeded the NH₄⁺-N standard, and 18 of 28 exceeded the TP standard (Figure 5).

According to inlet and outlet water concentration data of the Yongchun County wastewater treatment plant provided by the Yongchun County Environmental Protection Bureau, in 2014, the unit treatment costs of NH₄⁺-N and TP at the plant were 31,873.36 CNY/ton (5039.19 USD/ton) and 275,310.60 CNY/ton (43,526.68 USD/ton), respectively, under the secondary discharge standard.

Based on pollutant exceedance and the treatment cost, the amounts of water quality ecological compensation in Taoxi Creek during the flood season and the dry season were calculated to be CNY 7,867,801 (USD 1,243,901) and CNY 15,256,231 (USD 2,412,013), respectively. The average cost of water quality ecological compensation was calculated as the average of the two seasons and was approximately CNY 11.9 million (USD 1.88 million).

3.3. Ecological Compensation Based on Ecosystem Service Flows

3.3.1. SWAT Model Calibration and Verification

We conducted sensitivity analysis using the SWAT Calibration and Uncertainty Programs tool [42]. The model rate determination results were derived from the sensitivity analysis results based on the applicable conditions and significance of each parameter (

Table 1. Model calibration parameters.

Parameter Name	Parameter	Final Value
CN2	Runoff curve coefficient	1.5
CANMX	Maximum canopy interception	0.3
ESCO	Soil evaporation compensation factor	1.26
CH_K2	River hydraulic conductivity (mm/h)	0.25
ALPHA_BF	Baseflow regression coefficient (days)	0.53
GW_DELAY	Groundwater delay (days)	0.27
CH_N2	Main channel Manning’s coefficient	0.36
SOL_K	Soil saturation hydraulic conductivity (mm/h)	0.0682
SOL_AWC	Soil effective water holding	0.38
GW_REVAP	Groundwater re-evaporation coefficient	0.05
RCHRG_DP	Deep groundwater infiltration coefficient	0.41

).

Based on runoff data from 2005 to 2019 for the Jinjiang watershed, 2005 was considered as the model warm-up period, 2006–2012 as the model calibration period, and 2013–2019 as the model validation period. Calibration and validation were carried out in monthly time steps.

The simulated value curves of runoff in the Jinjiang watershed fit well with the observed value curves in most periods, and the overall fluctuation trend was generally consistent, with good peak fit performance (Figure 6, Figure 7 and Figure 8). We evaluated the accuracy of the simulation results for the three hydrological stations (

Table 2. Simulation results for the calibration and validation periods for three hydrological stations in the Jinjiang watershed.

Station	Calibration Period (2006–2012)			Validation Period (2013–2019)
	NSE	R²	PBIAS	NSE	R²	PBIAS
Shilong	0.86	0.83	8.81	0.87	0.90	7.62
Shanmei	0.84	0.83	7.17	0.84	0.85	8.25
Anxi	0.72	0.77	10.82	0.81	0.83	9.76

). All three hydrological stations met the criteria of R² > 0.6, NSE > 0.5, and PBIAS < 25%, indicating that the model simulation results were good and realistically reflected the monthly runoff in the Jinjiang watershed; thus, the simulation results could be used for subsequent ecosystem service flow determination.

3.3.2. Water Supply Simulation

We constructed a SWAT model for the Jinjiang watershed, considering the data availability and the feasibility of model verification for the Taoxi Creek watershed.

The water supply in 2013, 2016, and 2019 was 3.19 billion m³, 7.16 billion m³, and 5.38 billion m³, respectively (Figure 9). The water supply was the highest in 2016, with an average water supply of 216 million m³ per sub-basin. The average water supply per sub-basin in 2013 and 2019 was 96 million m³ and 163 million m³, respectively. Climate change is an important factor affecting the water supply in the basin. The trends in precipitation and basin water supply in the same period were consistent. Average precipitation in the basin reached 2522.6 mm in 2016, which was 47.6% more than that in 2013 and 54.5% more than the average over all 3 years, which is typical of a water-abundant year. Accordingly, the water supply reached 7.16 billion m³, which was more than double that in 2013.

3.3.3. Water Demand Simulation

To match the calculation scale of water supply with the water demand, spatial statistical analysis of water consumption at the district and county scales was performed using GIS to obtain water demand at the sub-basin scale (Figure 9).

The water demand downstream of the watershed was substantially greater than that in the upstream areas because of the dense distribution of towns and industrial production. Water consumption was mainly concentrated in the downstream sub-basins 27, 28, 29, and 31. In the upper part of the basin, sub-basins 13, 14, 15, and 11 also exhibited high water consumption as they comprised the urban areas of Yongchun and Anxi counties. Sub-basins 1–8, 22, 26, and 32 in the upper part of the watershed had lower water consumption, mainly because these areas are dominated by forest and grass, with some unused land, and are sparsely populated (Figure 10).

To determine the dominant factors of water demand and to identify the main beneficiaries of ecological compensation based on WSS flows, we performed Spearman correlation analysis of water consumption with urban population, rural population, irrigated farmland area, and industrial GDP output. The water consumption showed a highly significant positive correlation with the irrigated farmland area (

Table 3. Spearman correlation of influencing factors of water consumption simulation in the Jinjiang watershed.

Type	Parameter	Urban Population	Rural Population	Area of Irrigated Farmland	Industrial GDP
Water consumption	Spearman correlation	0.35	0.617	0.833 **	0.467
	Significance	0.356	0.077	0.005	0.205

** 99% confidence interval.

), indicating that irrigated farmland water is the main beneficiary.

3.3.4. S:D Ratio and Service Flow Compensation

Once the spatial distributions of water supply and water demand in the Jinjiang watershed are determined, the water resource supply and demand balance can be expressed in the form of the S:D ratio. To explore the water resource surplus and deficit situation, we mapped benefit and supply areas based on the S:D ratios for 2013, 2016, and 2019 (Figure 10).

The upstream and most midstream areas of the basin are the water supply areas, whereas sub-basins 29, 31, and 27, which are located in the downstream areas, are all beneficiary areas (Figure 11). Although the water supply in the latter sub-basins was high, their water demand was also high. Sub-basins 13 and 15, which are in the western stream area of Jinjiang, showed similar conditions in three years because these sub-basins are in the vicinity of Anxi County, where domestic water consumption is high, resulting in insufficient water supply in some years.

Taoxi Creek showed higher supply than demand in all three years, which indicates that it is a continuous supplier of ecosystem services in the entire Jinjiang watershed. The total service flow provided by Taoxi Creek to the downstream areas was calculated to be 787,180,000 m³. The Jinjiang watershed is a Class I area as specified in the Fujian Province Water Resources Levy and Use Management Measures [43], with an urban domestic water unit price of 0.06/m³, and the upper limit of compensation for the calculated available water was determined to be CNY 47,230,800 (USD 7,467,201). The unit input cost of maintaining water quantity was 0.024 CNY/m³ (0.004 USD/m³), and the lower limit amount of compensation for water quantity in Taoxi Creek was calculated to be CNY 18,892,320 (USD 2,986,880). Therefore, the net benefit of the water compensation for the Taoxi Creek watershed should range between CNY 18.9 million and 47.2 million (USD 3.0 million and 7.5 million).

3.4. Integration of Water Quality Restoration Cost with Water Supply Service Flows

When considering both water quality restoration cost and ecological service flows, Taoxi Creek watershed should receive CNY 30.8–59.1 million (USD 4.9–9.3 million) of gross economic compensation from downstream ecosystem beneficiaries annually, using it for water quality improvement and ecological conservation.

In 2019, the Quanzhou municipal government issued the Management of Special Funds for the Protection of Water Resources in the Upper Jinjiang and Luoyang Rivers [44]. According to this document, Jinjiang City and Nan′an City, as the main beneficiaries of water resources in the Jinjiang watershed, should provide ecological compensation funds of approximately CNY 120 million (USD 19.0 million). This compensation amount considers water quality, water quantity, and ecological protection factors. In view of the current problems of compensation allocation and compensation basis, the compensation funds allocated to Taoxi Creek should be within the range of CNY 30.8–59.1 million (USD 4.9–9.3 million) as estimated based on water quality restoration cost and WSS flow.

4. Discussion

4.1. Relationship between Water Quality Changes and Compensation

The annual ecological compensation amount based on water quality restoration cost is CNY 11.9 million (USD 1.88 million), which was calculated as the average of the flood and dry seasons. Figure 3 implies that an increase in the concentration of NH₄⁺-N downstream increases the amounts to be compensated. Downstream of the watershed, especially near the county, the concentrations of pollutants, especially NH₄⁺-N, are dramatically increased because of the accumulation of domestic sewage discharge, increasing the cost of restoration.

The concentrations of NH₄⁺-N and TP in Taoxi Creek were found to be high in the dry and transition seasons and low in the flood season, which is mainly because of the different river water volumes in different seasons. The same finding was reported in a monitoring study by our research group in the Jiulong River watershed [45,46] and in previous studies on NH₄⁺-N in the Jiulong River watershed [47,48]. The pollution concentrations of NH₄⁺-N ranged from high to low in the dry, flat, and flood seasons. Further, we found that the concentrations of NH₄⁺-N and TP were the lowest in forest areas and gradually worsened near urban areas and agricultural lands. The reason for this phenomenon may be that the accumulated sewage water and water from agricultural lands and fruit orchards near the watershed are not treated, but discharged directly into the river [48].

In this study, the indicators selected to calculate the compensation based on water quality were NH₄⁺-N and TP. The choice of indicators affects the final compensation amount. Under the condition of limited human and material resources, the results will be more accurate if the three indicators of chemical oxygen demand, NH₄⁺-N, and TP are included as assessment factors in watershed ecological compensation calculation [49]. In this study, because of the long monitoring period and numerous experimental factors, only NH₄⁺-N and TP were monitored to ensure the accuracy of the water quality experimental data. In future studies, chemical oxygen demand can be included to render a more accurate and objective compensation amount [8,50].

4.2. Ecosystem Service Flows and Determination of Unit Costs

The water supply capacity was found to be low in the upper reaches of the Jinjiang River basin and significantly high in the downstream sub-basins. The catchment area in each sub-basin is an important factor contributing to the water yield. For example, the high water supply capacity in sub-basins 11 and 27 can be explained by the fact that these basins had the highest catchment areas, in addition to factors such as precipitation and land use type. In addition to the catchment area, the water yield in a sub-basin is significantly related to the structural composition and physicochemical properties of the soil and its coverage. For example, the water supply in sub-basin 22 is significantly smaller than that in sub-basins 11 and 27, which have similar catchment areas and precipitation, mainly because sub-basin 22 comprises mostly agricultural land and only 0.1% of urban construction land. In contrast, the grassland and farmland areas in sub-basins 11 and 27 are smaller than those in sub-basin 22, whereas the area of urban construction land is substantially larger than that in sub-basin 22, which indicates that the flow production capacity of agricultural land is lower than that of urban construction land [51]. Although the water yield on urban construction land is higher because its large impervious area prevents water infiltration, most of the precipitation that falls to the ground flows into underground drainage pipes, making it difficult for water resources to be used [52].

The sub-basins 9, 10, 11, and 12 located along Taoxi Creek were consistent ecosystem service suppliers in the 3 years investigated and were natural resource providers for the entire Jinjiang watershed, whereas the beneficiary areas were mainly concentrated in the downstream part of the Jinjiang watershed and sub-basins 13, 15, and 23 in Anxi County in the western Jinjiang stream. This implies that the ecosystem services in the entire Jinjiang watershed flow from upstream to downstream and from areas with abundant water resources to areas with high water demand. Through data statistics analysis, we found that the total water consumption in the three major cities in the Jinjiang watershed slowly decreased from 2013 to 2019. This may reflect current good practices of water reuse and water conservation, which are beneficial to sustainable water resource development. The positive correlation between water consumption and irrigated farmland area indicated that agricultural water demand is the main beneficiary of the Jinjiang watershed, which is a reference for the allocation of compensation funds based on WSS flow.

The upstream region provides ecological services for the downstream and offers a certain amount of development authority. In practice, the amount of ecological compensation is affected by factors such as changes in water quality and fluctuations in water quantity. The ecological compensation has to be determined considering a reasonable interval to ensure reasonable upper and lower limits for the compensation [24]. In the Jinjiang watershed, the most direct factor affecting water quantity is precipitation, which is a natural factor. To determine the cost of water maintenance, one can refer to the governmental input for various types of erosion control projects [8].To determine the unit cost of service flow, the upper and lower bounds were considered the unit market price of water resources and the unit cost of water conservation, respectively. Many factors have to be considered when determining the unit cost of water conservation as well as soil and water conservation inputs [5,53,54]. The construction of protective forests and water conservation facilities is part of the cost of water conservation [18]. The cost of water conservation involves numerous subjective factors and is difficult to measure objectively [55]. The determination of the unit cost has a direct impact on the compensation amount; thus, a more objective and comprehensive unit cost of water is very important for calculating a compensation amount and should be researched and established in future studies [16,56].

4.3. Methodological Innovations and Limitations

Previous studies considered service flow or water quality compensation separately [5,8]; the novelty of this study contrastingly lies in the integration of water quality compensation and WSS flow compensation and the comprehensive consideration of both the quantity and quality of water resources.

The local polluters in the Taoxi Creek watershed have two main components. Firstly, due to livestock and poultry farming and excessive application of chemical fertilizers and pesticides, non-point-source pollution is serious. Secondly, the relatively few sewage treatment plants and poor sewage treatment capacity in the Taoxi Creek watershed lead to point-source pollution as domestic sewage is often discharged directly into the river without treatment [28]. Information on local polluters can inform the allocation and use of compensation funds [8,28].

The amount of water pollution compensation is greatly affected by the selected pollutant indicators. Based on combined monitoring data and local pollutant characteristics, NH₄⁺-N and TP were selected as compensation indicators in this study. However, pollutant concentration is not the only measure of water quality. Water bodies have the ecological function of providing a habitat environment and nutrients for aquatic animals and plants, and the value thereof is difficult to estimate [50]. Moreover, we considered only the cost of pollutant treatment and ignored the cost of wastewater transportation and sewage treatment plant construction. Thus, our calculations are an underestimation [57,58].

The SWAT model has been widely used in hydrological applications, such as water quality simulation and sediment erosion prediction at the watershed scale. However, in practical modeling, a wide range of geographical data requiring high accuracy (topography, landform, climate, and precipitation) are involved. In this study, the SWAT model was used to describe the ecological compensation based on the WSS flow. Compared with the widely used InVEST model, the SWAT model can simulate water yield for a longer time [59]. Under the condition that all data requirements are satisfied, more accurate simulation results can be obtained at the watershed scale [60]. However, one limitation of this study was that the water yield could not be analyzed at the pixel scale, which simplifies the loss in the water supply service flow process [61]. Consistent with the results of Wang et al. [25], the WSS flow representing water quantity is more dominant than water quality in ecological compensation, which government departments and stakeholders need to focus on. This study did not consider the impact of some socioeconomic activities (e.g., artificial withdrawals and government measures), as well as groundwater and multistakeholder involvement in the payment process, which should be further studied in the future.

5. Conclusions

We proposed a new basis for water-related ecological compensation integrating WSS flow and water quality restoration costs, based on which we calculated the ecological compensation amount for the Taoxi Creek watershed to be from CNY 30.8 million to 59.1 million (USD 4.9–9.3 million). On the one hand, replacing traditional water quantity compensation with WSS flow takes into account the human demand for ecosystem services. On the other hand, the sampling and monitoring of each sub-basin cross-section make our water quality monitoring more extensive and more representative in space compared to a single cross-section.

The calculated water quality restoration cost was approximately CNY 11.9 million (USD 1.88 million), and the compensation for water supply service flow should range between CNY 18.9 million and 47.2 million (USD 3.0 million and 7.5 million). The total ecological compensation for Taoxi Creek was calculated to be CNY 30.8–59.1 million (USD 4.9–9.3 million). This study provides not only a scientific basis for improving the ecological compensation scheme in the Taoxi Creek watershed but also a reference for the formulation of ecological compensation in other watersheds.