Wastewater Discharge Impact on Drinking Water Sources Along the Yangtze River (China)

Science of the Total Environment 599–600 (2017) 1399–1407

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Wastewater discharge impact on drinking water sources along the Yangtze River (China)

Zhuomin Wang a,DongguoShaoa, Paul Westerhoff b,⁎ a State Key Laboratory of Water Resources & Hydropower Engineering Science, Wuhan University, Luojia Hill, Wuhan 430072, PR China b School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ 85287-3005, United States

HIGHLIGHTS GRAPHICAL ABSTRACT

• First estimates for indirect (de facto) potable wastewater reuse potential in Yangtze River. • Levels of de facto reuse ranged from b1% to N14%. • Demonstrate strategy to predict unplanned wastewater reuse using limited datasets • Shows interconnected hydrologic na- ture of growing cites in Asia

article info abstract

Article history: Unplanned indirect (de facto) wastewater reuse occurs when wastewater is discharged into surface waters up- Received 25 March 2017 stream of potable drinking water treatment plant intakes. This paper aims to predict percentages and trends of de Received in revised form 8 May 2017 facto reuse throughout the Yangtze River watershed in order to understand the relative contribution of wastewa- Accepted 9 May 2017 ter discharges into the river and its tributaries towards averting water scarcity concerns. The Yangtze River is the Available online 17 May 2017 third longest in the world and supports more than 1/15 of the world's population, yet the importance of waste- fi Editor: D. Barcelo water on the river remains ill-de ned. Municipal wastewater produced in the Yangtze River Basin increased by 41% between 1998 and 2014, from 2580 m3/s to 3646 m3/s. Under low flow conditions in the Yangtze River near Keywords: Shanghai, treated wastewater contributions to river flows increased from 8% in 1998 to 14% in 2014. The highest Wastewater levels of de facto reuse appeared along a major tributary (Han River) of the Yangtze River, where de facto reuse Drinking water can exceed 20%. While this initial analysis of de facto reuse used water supply and wastewater data from 110 cit- Reuse ies in the basin and 11 gauging stations with N50 years of historic streamflow data, the outcome was limited by Hydrology the lack of gauging stations at more locations (i.e., data had to be predicted using digital elevation mapping) and Urban lack of precise geospatial location of drinking water intakes or wastewater discharges. This limited the predictive Pollution capability of the model relative to larger datasets available in other countries (e.g., USA). This assessment is the first analysis of de facto wastewater reuse in the Yangtze River Basin. It will help identify sections of the river at higher risk for wastewater-related pollutants due to presence of—and reliance on—wastewater discharge that could be the focus of field studies and model predictions of higher spatial and temporal resolution. © 2017 Elsevier B.V. All rights reserved.

⁎ Corresponding author. E-mail address: [email protected] (P. Westerhoff).

1. Introduction Fig. 1 showsthelocationsofriversandmajorpopulationcenterswithin the basin. The population density varies spatially, and the impact of up- Increasing economic development and human migration stress stream cities on downstream water quality is a major concern. All prov- water resources and water quality in many regions. Wastewater inces in China publish annual water resource bulletins that include discharged into surface waters represents a potentially reliable water macro data about water consumption and provide valuable information source, but it can convey chemical and microbial pollutants to cities on wastewater production within cities. Limited hydrological and drinking water intakes located downstream. Reclamation of water (streamflow) data exist along the entire length of the Yangtze and trib- after treatment in modern wastewater treatment plants (WWTPs) utaries. Consequently, the Yangtze River Basin is an important case for will likely be an important yet currently unrealized part of sustainable researching de facto reuse. The dataset for this paper included wastewa- water resource management (Bellona and Drewes, 2007). In the ter information collected from the water resources bulletins such as United States of America (USA), the National Academy of Engineering water intake and water consumption of each city and wastewater dis- published a report on wastewater reuse, and among the top ten re- charge of each province. We acknowledge that untreated wastewater search needs for human health, social, and environmental studies is flowed into the Yangtze River for centuries, but the recent advent of ad- quantifying the extent of de facto reuse (Needs, 2012). De facto reuse vanced sewage collection and centralized treatment (Wang et al., 2011) is the unplanned or incidental presence of treated wastewater in a along with major shifts in population density within urban centers in water supply source, and it occurs when treated wastewater is China is resulting in a smaller number of point sources for treated discharged from a WWTP into surface waters upstream of potable wastewater discharges. Portions of the Yangtze River are vulnerable to drinking water treatment plant (DWTP) intakes (see Fig. SI.1). The pres- water scarcity, but less so than other major rivers in Asia (Finlayson ence of treated wastewater in drinking water supplies increases the risk et al., 2013; Liu et al., 2013; Varis et al., 2012; Webber et al., 2015). Un- of water quality contamination from pharmaceuticals or other trace or- derstanding the contribution of wastewater flows into the Yangtze ganics, pathogens, and inorganic pollutants (Barnes et al., 2008; Cheng, River can contribute towards avoiding water scarcity. 2003; Duirk et al., 2011; Focazio et al., 2008; Fono and McDonald, 2008; China has the largest population in the world and has been rapidly Kaplan, 2013; Schwarzenbach et al., 2010). Estimates of de facto reuse urbanizing and investing in water infrastructure, yet many regions are exist in the USA and some European countries, but few estimates exist water stressed (Chu et al., 2004; Zhang et al., 2016). The goal of this for Asia (Kugathas et al., 2012; Williams et al., 2012). De facto reuse is paper is to estimate levels and patterns in de facto reuse within widespread and is increasing with population growth and increased China's largest river basin and, while doing so, demonstrate how de sewage collection. It is not uncommon to have a substantial portion of facto reuse can be estimated when geospatial or hydrologic national source waters originally derived from an upstream wastewater dis- datasets are limited. Assessing the extent of de facto wastewater reuse charge (Rice et al., 2013). Models capable of predicting de facto reuse in Yangtze River Basin required the following activities: (1) data com- are relevant for estimating impacts on downstream drinking water sup- pletion by reasonable assumptions and digital elevation model (DEM), plies for people. Model predictions also identify river reaches that con- (2) development of a model within the ArcGIS framework to determine tain an elevated fraction from wastewater origin, which may therefore spatial relationships between every city, (3) quantification of accumu- contain higher concentrations of chemical and microbial pollutants. lated wastewater of each city, (4) examination of wastewater percent- Wastewater discharges may pose water quality risks to downstream ages under low flow stream conditions in different years, and ecosystems and people who rely upon the river as drinking water (5) analysis about the relationship between population density and de source, but it simultaneously provides a renewable and sustainable in- facto reuse. This assessment is the first analysis of de facto wastewater stream flow that contributes towards a reliable water supply. WWTP reuse in Yangtze River Basin, which will help drinking water manage- discharges are a main source of pharmaceuticals and many other ment in the basin. In addition, it will provide a framework on how to micro pollutants in the environment, nutrients that influence stream use limited datasets to access de facto reuse in other locations. ecology, and pathogens that pose ecological and human health risks (Chen et al., 2009; Hajj-Mohamad et al., 2014; Maier et al., 2015). GIS- 2. Materials and methods based models can facilitate the assessment of potential impacts from wastewater, and several water quality models have been developed The primary datasets for this study included the quantity of water for this purpose (Kugathas et al., 2012; Williams et al., 2012; Johnson intake, water consumption, water discharge, and river streamflow in et al., 2013; Price et al., 2010; Rowney et al., 2011). For example, the the Yangtze River Basin. Data were mined from several sources: base national-scale GIS based model that includes modules for WWTP, Geographic Information System (GIS) layers were obtained from DWTP, and river reaches in the USA was useful for conducting a national China Geological Survey for topography, stream networks, and city assessment of de facto reuse occurrence on DWTPs and assessing the boundaries. Water resources bulletins were used to determine water in- potential impacts of streamflow variation (Rice and Westerhoff, 2014; take and water consumption of each of 110 cities and wastewater dis- Rice et al., 2016). There is a lack of information on de facto reuse occur- charge of each province. Each city's wastewater discharge quantity rence in many developing countries, in part due to limited locational was calculated by assuming that the amount of wastewater is propor- and design data on DWTPs and WWTPs and limited hydrological tional to the amount of water consumption in the same province. datasets. 2000 and 2010 were years of fifth and sixth census in China, and this in- The Yangtze River is the longest river in Asia and the third-longest in formation served as the basis for changes over time in population within the world. It flows for 6300 km from the glaciers on the Qinghai-Tibet the basin. Plateau in Qinghai eastward across southwest, central, and eastern There are limited numbers (n = 12) hydrologic stations along the China before emptying into the East China Sea at Shanghai. It drains 1/ Yangtze River with long datasets of historic streamflow. The historic 5th of the land area of the People's Republic of China, and its river streamflows, collected between 1954 and 2014, were obtained from basin is home to 1/3rd of the country's population (Zhang et al., Yangtze River Water Resources Committee. The DEM and hydrological 2006). More than four hundred million people live in the Yangtze analyst tool in ArcGIS were used to estimate streamflow data at sites River Basin. With its enormous population, this basin alone would be without hydrological stations. Fig. SI.2 shows the steps used to estimate the third most populated country in the world. The basin has 1/15 of streamflow using ArcGIS and the spatial analyst toolbox. the world's population and conveys 1/3 of China's water (Varis and Equations and methods to quantify de facto reuse were identical to Vakkilainen, 2001). The Yangtze River plays a large role in the culture those previously reported (Rice et al., 2013; Rice and Westerhoff, and economy of China, and the Yangtze River Basin generates as much 2015). Specifically, similar to a United States Environmental Protection as 40% of China's gross domestic product (GDP) (Chen et al., 2014). Agency (EPA) study, assumptions for calculating de facto reuse were Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407 1401

Fig. 1. (A) Yangtze River (China) drainage, major cities with streamflow gauges, and (B) population density map for different districts in the watershed. The Han River Basin is identified with a green colored river network. Water flows from the west towards the east, and the estuary drains into the East China Sea. conservative: WWTP effluent with no in-stream loss and complete 2 z g2 mixing of all water bodies (USEPA, 1980). De facto reuse is calculated K ¼ 1 þ g Ã − Ã 3−1 ð3Þ g 6 36 according to Eq. (1):

∑N−1 i¼1 Di De facto reuse ¼ ð1Þ where Q N u = mean of the logarithms (base e) of the historical annual low flows where S = standard deviation of the logarithms of the historical low flows N − 1 = number of cities in upstream reach of City N g = skewness coefficient of the logarithms of the historical low Q (m3/s) = streamflow of the river at the intake for City N N flows D (m3/s) = quantity of wastewater discharge from WWTP of up- i K = frequency factor for skewness g and return period of 10 years stream City z = Constant set to the value of −1.123 for return period of 10 years Historic streamflow data were analyzed to quantify different low De facto reuse potential was initially estimated for multiple years in flow conditions at the long-term hydrologic stations. EPA defines the 11 representative locations (Fig. 1) throughout the Yangtze River Basin. 7 day 10 year (7Q10) streamflow statistic as the “7-day, consecutive The representative cities were selected because of their varied geo- low flow with a 10-year return frequency, or the lowest streamflow graphic location, population size, and reliance upon surface water as for seven consecutive days that would be expected to occur once in their primary potable drinking water supply source. Chongqing, 10 years” (Ames, 2006). The 7Q10 low flow was computed using EPA Wuhan, Nanjing are large urban areas in upper, middle, and lower DFLOW 3.1. This method fits the historical low streamflow data to a spe- reaches of Yangtze River Basin, respectively; Shanghai is in the Yangtze cific probability density function and then uses this function to com- River estuary; and Huangzhuang, Chenglingji, and Hukou are the hydro- putes the flow whose probability of not being exceeded is 1/10. The logical stations in the Han River (which is the largest tributary of Yang- log Pearson Type III distribution is a convenient function to use because tze River), Dongting Lake, and Poyang Lake, respectively. An online it can accommodate a large variety of distributional shapes and has seen survey about DWTPs in these areas found most DWTPs in Yibin, Chong- wide-spread use in stream flow frequency analysis. The 7Q10 low qing, Yichang, Shashi, Wuhan, Jiujiang, Nanjing, and Shanghai withdraw streamflow based on the log Pearson Type III method is: water from Yangtze River, while the DWTPs in Huangzhuang, ÀÁ Chenglingji, and Hukou withdraw water from Han River, Dongting 7Q10 ¼ exp u þ K Ã S ð2Þ Lake, and Poyang Lake, respectively (Ding et al., 2013; Guo et al., 2014). 1402 Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407

3. Results and discussion (1) at each location using data from existing hydrologic gauging stations or estimated annual average streamflow conditions. Fig. 4 shows the 3.1. Wastewater and population trends of Yangtze River Basin potential for de facto reuse in the Yangtze River watershed for one representative year (2010). Surprisingly, the highest de facto reuse poten- The Yangtze River drains 1/5 of China's land area, and its river basin tials occur not only near the bottom of the watershed where the is home to 1/3 of the country's population. The river flows through the cumulative flows from wastewater are highest but also on tributaries plateau, mountains, and valleys in the upper reach, so it has characteris- in the middle reaches of the Yangtze River where there are high popu- tics of steep gradient, abundant water quantity, and high flow velocity. lations but lower streamflows. In the middle reach, the river gradient decreases and river course mean- Table 1 shows that de facto reuse increased at the 11 representative ders, causing a broader river surface and lower flow velocity. The middle sites (Fig. 1), with average de facto reuse increasing from 1.1% in 1998 to reach receives the most precipitation and accounts for 40% of the area, 2.5% in 2014. In 1998, there was a large-scale weather anomaly in China, so the volume of water and the number of tributaries and lakes increase and the entire Yangtze River Basin suffered a severe flood event. Higher dramatically. The lower reach has flat terrain, short tributaries, stable streamflow resulted in lower de facto reuse during these periods. An in- flow, and dense water network. (Ma and Li, 2008). crease in municipal wastewater fl ow and a decrease in streamflow both Fig. 2 shows the cumulated population distribution and estimated contributed to higher de facto reuse under average flow in 2014, which wastewater flows of all cities in Yangtze River Basin in 2010 as a function is the last year where a complete dataset was available. of the upstream distance from the estuary mouth of the Yangtze River into The differences in de facto reuse between 1998 and 2014 shown in the East China Sea. The shortest path method and network analysis tool- Table 1 can be attributed to differences in population (i.e., population in- box in ArcGIS were used to calculate these distances. The upper reach re- creases also increased the wastewater discharges) and variations in gion contains 40 cities and accounts for 40% of population and 37% of streamflows. For example, 1998 was the high flow year, 2003 was the wastewater. The middle reach region has 41 cities, accounting for 37% of middle flow year, and 2006 was the low flow year in the Yangtze population and 35% of wastewater. The lower reach region has 17 cities, River Basin. Fig. 5 shows the changes in cumulative upstream municipal accounting for 23% of population and 28% of wastewater. Differences be- wastewater flows and was an important step in assessing de facto reuse tween the percentage of population and percentage of wastewater pro- changes between 1998 and 2014. This period of time represents a signif- duced is related to varying levels of industrial activity within each city, icant growth in public infrastructure, including sewage systems, within which greater industrial manufacturing rather than agriculture for cities China (Kerstens et al., 2009; Sang et al., 2007). Figs. 2 and 5 show the cu- in the lower reaches of the Yangtze River. The major cities labeled in mulative wastewater of each city. The cumulative wastewater effluent Fig. 1—including Shanghai, Wuhan, Chongqing, and Chengdu—contribute contribution to 11 representative sites increased from 2580 m3/s in significantly to the cumulated wastewater and population, but many 1998 to 3646 m3/s in 2014 (41% increase). All sites received more other cities exist along the mainstream and tributaries (Figs. 2 and 3). Nu- wastewater contributions from upstream WWTPs in 2014 than in merous factors influence the extent to which wastewater contributes 1998. Chongqing, which is located in the upper reach, increased the to streamflow at any given location within the basin. For example, wastewater flow by 32% (34.6 m3/s). Wuhan, located in the middle the upper basin has low population, low precipitation rates, and low reach, increased the wastewater flow by 32% (120 m3/s). Nanjing, locat- streamflow. In the middle reach region, the population density and waste- ed in the lower reach, increased the wastewater flow by 28% (155 m3/s). water flows are high, but this is the wettest region (i.e., highest annual One of the largest changes was observed in Shanghai, the most popu- rainfall). The lower reach region is again populous, but also drier. lous city in Yangtze River Basin and located near the Yangtze River estuary, where wastewater flow increased by 65% (399 m3/s) during this 3.2. De facto reuse analysis in Yangtze River Basin period. The cumulative municipal wastewater flows grew dramatically in the Yangtze River Delta, which is the region between Nanjing and Each urban area significantly increases wastewater flows into the Shanghai and is the most affluent region in China. Between 1998 and river (Fig. 3). The potential for de facto reuse was calculated using Eq. 2014 the changes in wastewater flows were about 30% for Chenglingji

Fig. 2. The cumulated distribution of wastewater ﬂow and population of 97 cities in the Yangtze River Basin in 2014; x-axis gives distance between each city and estuary along the river. 13 additional cities (of the 110 cities studied) are located at the headwaters of rivers and tributaries and thus have zero cumulated wastewater ﬂow. Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407 1403

Fig. 3. Cumulative wastewater flows from each city into the Yangtze River and its tributaries in 2010. and 40% to 50% for Huangzhuang and Hukou. These changes were com- Because low streamflows lead to higher levels of de facto reuse, we parable to those observed at other locations on the mainstream. Among focused on the 7Q10 streamflow occurrence (Eq. (2)) to consider a the three tributary sites, Dongting Lake had largest contribution of near “worse-case” scenario for the maximum amount of wastewater wastewater, but Poyang Lake had the largest change of cumulative mu- in different river sections. The 7Q10 value roughly corresponds with nicipal wastewater discharge during the time period studied. The net ef- 5th percentile streamflow occurrence for the Yangtze River (e.g., Fig. 6 fect from wastewater infrastructure investment and increased – inset). Table 2 shows the de facto reuse at the 7Q10 streamflow con- population in urban areas was a 41% increase in wastewater treatment dition for 11 representative sites in 1998, 2000, 2003, 2006, 2010, and capacity between 1998 and 2014. 2014. Chongqing in upper reaches had de facto reuse increase from The above analysis suggests that increasing urban population and 4.2% in 1998 to 5.5% in 2014; Wuhan, which is economic center in mid- wastewater infrastructure capacity increases downstream de facto dle reaches, increased from 6.8% in 1998 to 8.9% in 2014; Nanjing, which reuse and also suggests that annual (or seasonal) streamflow variations is located in lower reaches, increased from 7.7% in 1998 to 9.8% in 2014; affect de facto reuse levels. The city of Nanjing was used as an example and Shanghai, which is in the estuary, had the highest de facto reuse to illustrate the effects of streamflow on de facto reuse (i.e., amount of among the mainstream sites and increased from 8.3% in 1998 to 14% river water dilution for discharged treated wastewater). Historical in 2014. The four cities (Wuhan, Jiujiang, Nanjing, and Shanghai) locat- stream gauge data at the Datong hydrological station were used in con- ed in the lower reach area of Yangtze River all had wastewater percent- junction with cumulative release of upstream wastewater. Fig. 6 (inset) ages (i.e., de facto reuse) of 8% and higher after 2010. Yibin, Chongqing, shows the frequency distribution of streamflows (m3/s) for the period Chenglingji, and Hukou located at upper reaches or tributary had less de of record. The streamflows varied by roughly a factor of five (8000 facto reuse. The highest de facto reuse occurred at Huangzhuang, which to 40,000 m3/s) between the 10th and 90th percentile. The is located on a major tributary (Han River) of the Yangtze River. streamflow variations result in different levels of de facto reuse (Fig. 6), assuming wastewater treatment discharges remain the same. 3.3. Case study of Han River Basin with high levels of wastewater impact These conditions showed de facto reuse would be approximately five times higher (8% versus 1.6% de facto reuse potential) at the 10th per- Because of its highest potential for de facto reuse (shown in Tables 1 centile streamflow occurrence than the 90th percentile streamflow and 2), the Han River (highlighted in Fig. 1) was selected for a more de- occurrence. tailed case study on this tributary. Strong interdependences existed

Fig. 4. Calculated de facto reuse in different districts within the Yangtze Basin for 2010. 1404 Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407

Table 1 De facto reuse under average annual streamﬂow condition of 11 sites in 1998 and 2014 (names in parentheses indicate Yangtze River tributary river names).

Year Yibin Chongqing Yichang Shashi Wuhan Jiujiang Nanjing Shanghai Huangzhuang Chenglingji Hukou (Han River) (Dongting) (Poyang)

1998 0.5% 0.7% 0.8% 0.9% 1.4% 1.3% 1.8% 1.7% 1.3% 0.9% 1.2% 2000 0.6% 0.8% 1.0% 1.2% 1.9% 1.7% 2.3% 2.4% 1.6% 1.2% 1.6% 2003 0.7% 0.9% 1.1% 1.3% 2.0% 1.8% 2.5% 2.7% 1.8% 1.2% 1.6% 2006 1.0% 1.5% 2.1% 2.2% 3.0% 2.7% 3.5% 4.0% 4.5% 1.6% 1.7% 2010 1.0% 1.2% 1.6% 1.8% 2.4% 2.2% 2.8% 3.0% 2.3% 1.1% 1.3% 2014 1.1% 1.3% 1.5% 1.7% 2.3% 2.3% 3.0% 3.5% 7.8% 1.3% 2.2%

between dense urban populations, water demand, industrial output, streamflow. As shown in Fig. 7 impacts downstream of districts with and wastewater discharges that impacted downstream communities. high population density and without sufficient streamflow available Fig. 7 shows the population density and de facto reuse potential under for dilution will have higher de facto reuse and thus higher drinking low flow (7Q10) for the Han River Basin. We estimated de facto reuse water risk. using population-based assumptions for wastewater discharges and As demonstrated above, the dense population and associated high streamflows calculated using DEM in different districts within the water demand in the Han River Basin led to potential localized stresses basin. In the upper reaches of Han River, de facto reuse did not occur. on the amount and quality of water in the river. However, the effect of As the mainstream flows through cities with large population wastewater on flow and water quality in the Han River is not confined (e.g., Hanzhong, Chenggu) and higher associated wastewater dis- locally because the middle route of the South-to-North Water Transfer charges, de facto reuse increased significantly. Negligible changes oc- Project delivers 30 million cubic meters of water from Danjiangkou res- curred downstream of smaller or less populated cities. ervoir to northern China every day (Shao et al., 2016). Thus, wastewater In the middle reaches of Han River, the high streamflow tributaries flowing in the Han River also supports water deliveries to northern that flow through cities with small population had low de facto reuse. China. The headwater source in the middle route is Danjiangkou reser- When tributaries converge into the mainstream, de facto reuse in the voir, which had an average streamflow of 1170 m3/s (Liu et al., 2012) Han River decreased. In the lower reaches near Wuhan where the Han from 1951 to 2010. The transferred water accounts for nearly 30% of River converges into the Yangtze River, the population density is high streamflow into the reservoir. This project began operating in 2015, and upstream municipal wastewater accumulates. As such, de facto and it will reduce the streamflow in the lower Han River, which will in- reuse increased dramatically and stayed high as water continued to crease de facto reuse. The de facto reuse in Danjiangkou is 5% under low flow downstream. flow condition (7Q10) (Fig. 7) and 1% under average flow condition, so The Sheqi district in the Han River, which had the highest de facto the water conveyed to northern China by this project is only partially reuse, was further analyzed. The average cumulative municipal waste- impacted by upstream municipal wastewater in the Han River basin. water discharge and drinking water intake of Sheqi are 0.89 m3/s and Thus, if this contribution of wastewater and associated pollutants 4.4 m3/s, respectively. The average streamflow and low flow of the poses human risk or perception issues, it could be more cost-effective river flowing through Sheqi are 17 m3/s and 4.0 m3/s, respectively. to improve treatment at the local wastewater treatment plants before The high de facto reuse was attributed to the large amount of wastewa- diluting the treated wastewater into the river water and exporting it ter, which was caused by higher population density, and the small to cities in northern China.

Fig. 5. Cumulative upstream municipal wastewater flows at representative sites along the Yangtze River between 1998 and 2014. Inset figure shows Yangtze River tributaries. The number adjacent to each solid black bar is the additional cumulative upstream municipal wastewater flows at each site between 1998 and 2014. Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407 1405

Fig. 6. Case study of Nanjing de facto reuse based on streamflow percentiles for the Yangtze River; x-axis gives the streamflow correlating to the Nth percentile for the frequency of occurrence of particular streamflows. Inset shows historic streamflow percentile (1954–2008) in the Yangtze River at Nanjing.

4. Summary and conclusions higher potential for de facto reuse. Large river systems in the western and desert regions of the USA can have de facto reuse levels exceeding Predictions of de facto reuse in the Yangtze River basin provide an 15% (e.g., Rio Grande River) because there is extensive water consump- upper bound on the percentage of wastewater from urban centers in tion for agriculture that removes water from the river while wastewater the Yangtze River at different locations. An increase in municipal waste- flows continue along the length of the river before it enters the Gulf of water flow and a decrease in streamflow both resulted in higher de facto Mexico. The lower portions of the Colorado River (USA) have 1 to 3% reuse under average flow, and average de facto reuse under average de facto reuse potential under average flow. Furthermore, similar to streamflow increased from 1.1% in 1998 to 2.5% in 2014. De facto our findings for Nanjing (Fig. 6), we previously found higher de facto reuse under low flow conditions increased during prolonged periods reuse under low streamflow conditions in the USA (Rice and in the Yangtze River Basin, and the percentage in Shanghai, which is Westerhoff, 2015; Rice et al., 2015b). For smaller streams, there is po- in the Yangtze River estuary, increased from 8.3% in 1998 to 14% in tential for the river to contain a majority of wastewater (i.e., de facto 2014. reuse N50%) during mild drought periods (e.g., 20th percentile stream Chemicals and pathogens in the wastewater are not only diluted, but flows) or low flow years (Rice and Westerhoff, 2015; Rice et al., some will undergo transformations, sorb and accumulate in sediments, 2015a). Overall, the trends in the Yangtze River basin were similar to or be inactivated. Thus, the predictions here identify regions of the those in the USA. Yangtze River at potential risk for impacts to both aquatic organisms This research estimated de facto reuse in the Yangtze River Basin and in the river and drinking water quality at downstream locations. In the showed the feasibility of using limited infrastructure datasets to assess USA, predictions at DWTPs of chemical pollutants associated with de facto reuse. There is a lack of information on wastewater contribu- wastewater discharges correspond well with estimates of de facto tions to the Yangtze River (i.e., de facto reuse potential) because the reuse (Rice et al., 2015a), suggesting that not all chemicals entering riv- construction of sewage infrastructure and urban population growth ers become assimilated or decay completely through natural biogeo- are relatively recent trends in China over the past two decades. While chemical processes. The regions identified as having higher levels of we used data from 110 cities in the Yangtze River basin and macroscale de facto reuse warrant more detailed modeling studies and field sam- water consumption data that was readily available in China, higher spa- pling to assess water quality and to help validate the modeling efforts. tial or temporal resolution de facto reuse model predictions would re- The de facto reuse observed in the Yangtze River in this study were quire precise locations of wastewater discharges and drinking water comparable with de facto reuse levels predicted in large river systems intakes. Furthermore, there are a limited number of streamflow gauging in USA (Rice et al., 2013; Rice and Westerhoff, 2015; Rice et al., stations within the Yangtze River basin with N50 years of historic 2015a). For example, under average flow on the Mississippi River, streamflow data, and such information is important in understanding which is the largest river in the USA, the levels of de facto reuse are low- variability in streamflow conditions. To overcome this limitation, we est in its headwaters and increase to ~1% in the lower reaches. Like the were able to use DEM and geographical analyst tools to estimate hydro- Han River tributary of the Yangtze River, it is the smaller tributaries logic streamflow conditions at multiple locations and predict de facto (i.e., lower Strahler order streams) of the Mississippi River that have reuse potential throughout the Yangtze River basin. Such approaches

Table 2 De facto reuse under annual low ﬂow condition (7Q10) at representative locations with long term historic streamﬂow data (names in parentheses indicate Yangtze River tributary river names). The location with the highest potential de facto reuse (Huangzhuang on the Han River tributary of the Yangtze River) is highlighted in bold text.

Year Yibin Chongqing Yichang Shashi Wuhan Jiujiang Nanjing Shanghai Huangzhuang Chenglingji Hukou (Han River) (Dongting) (Poyang)

1998 1.5% 4.2% 5.0% 5.3% 6.8% 7.1% 7.7% 8.3% 11% 4.5% 4.8% 2000 1.5% 4.2% 5.1% 5.8% 7.2% 7.6% 8.1% 9.4% 11% 4.5% 4.9% 2003 1.6% 4.3% 5.6% 5.8% 7.2% 7.6% 8.2% 10% 11% 4.7% 4.9% 2006 1.6% 4.6% 6.3% 6.6% 7.8% 8.1% 8.3% 12% 14% 4.8% 5.0% 2010 1.8% 5.0% 7.4% 7.7% 8.8% 9.2% 9.5% 13% 14% 5.1% 5.5% 2014 1.9% 5.5% 7.5% 7.8% 8.9% 9.5% 9.8% 14% 15% 5.8% 6.8% 1406 Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407

Fig. 7. Maps for the Han River Basin (reference Fig. 1) of the Yangtze River showing (A) population density and (B) de facto reuse levels under low ﬂow condition (7Q10). The location is marked (★) for Danjiangkou reservoir.

are probably scalable to other countries that have rapidly developing model are good locations to conduct field sampling for chemical urban population centers, investments in infrastructure, and a lack of markers of wastewater pollution and impacts of pollution on aquatic long-term historic streamflow data. life. The modeling methods presented in this paper allow identification of river reaches most impacted by wastewater, and perhaps suggest pri- Acknowledgements oritization for investments in additional wastewater treatment in reaches with high de facto reuse that could provide large benefits to- The authors thank the National Science Foundation for funding wards improving water quality in the Yangtze River and its tributaries. through the Central Arizona-Phoenix Long-Term Ecological Research Future work will include understanding and modeling agricultural, en- (DEB-1637590) and Arizona State University Decision Center for a ergy mining, and other water intensive activities on withdrawals and Desert City (Award no. 0951366). This study is supported by the return flows to the Yangtze River. Results from this work can then be National Science and Technology Major Special Project (No. used to select locations for improved hydrologic data collection that al- 2012ZX07205005) and National Natural Science Foundation of lows better characterization and assessment of river reaches most im- China (No. 51379150 and No. 51439006). Special thanks to Dr. Jacelyn pacted by wastewater (i.e., those with highest levels of de facto Rice, Longzhang Fang, and Thuy Nguyen for discussions and strategies reuse). Water treatment plants or river reaches of the Yangtze River, related to modeling de facto reuse and Laurel Passantino for technical or its tributaries, with higher levels of de facto reuse potential in our editing. Z. Wang et al. / Science of the Total Environment 599–600 (2017) 1399–1407 1407

Appendix A. Supplementary data Liu, J.G., Zang, C.F., Tian, S.Y., Liu, J.G., Yang, H., Jia, S.F., You, L.Z., Liu, B., Zhang, M., 2013. Water conservancy projects in China: achievements, challenges and way forward. Global Environmental Change-Human and Policy Dimensions 23 (3), 633–643. Supplementary data to this article can be found online at http://dx. Ma, Z.-G., Li, L.-L., 2008. Hydrology information extracting of Yang River watershed based doi.org/10.1016/j.scitotenv.2017.05.078. on GIS and DEM. Journal of Hebei North University (Natural Science Edition) 1, 020. Maier, D., Blaha, L., Giesy, J.P., Henneberg, A., Köhler, H.R., Kuch, B., Osterauer, R., Peschke, K., Richter, D., Scheurer, M., 2015. Biological plausibility as a tool to associate analyt- References ical data for micropollutants and effect potentials in wastewater, surface water, and sediments with effects in fishes. Water Res. 72, 127–144. Ames, D.P., 2006. Estimating 7Q10 confidence limits from data: a bootstrap approach. National Research Council, 2012. Water Reuse: Potential for Expanding the Nation's J. Water Resour. Plan. Manag. 132 (3), 204–208. Water Supply Through Reuse of Municipal Wastewater. National Academies Press, Barnes, K.K., Kolpin, D.W., Furlong, E.T., Zaugg, S.D., Meyer, M.T., Barber, L.B., 2008. Ana- http://dx.doi.org/10.17226/13303. tional reconnaissance of pharmaceuticals and other organic wastewater contami- Price, O.R., Williams, R.J., van Egmond, R., Wilkinson, M.J., Whelan, M.J., 2010. Predicting nants in the United States - I groundwater. Sci. Total Environ. 402 (2–3), 192–200. accurate and ecologically relevant regional scale concentrations of triclosan in rivers Bellona, C., Drewes, J.E., 2007. Viability of a low-pressure nanofilter in treating recycled for use in higher-tier aquatic risk assessments. Environ. Int. 36 (6), 521–526. water for water reuse applications: a pilot-scale study. Water Res. 41 (17), Rice, J., Westerhoff, P., 2014. Spatial and temporal variation in de facto wastewater reuse 3948–3958. in drinking water systems across the USA. Environ. Sci. Technol. 49 (2), 982–989. Chen, B., Nam, S.-N., Westerhoff, P.K., Krasner, S.W., Amy, G., 2009. Fate of effluent organic Rice, J., Westerhoff, P., 2015. Spatial and temporal variation in de facto wastewater reuse matter and DBP precursors in an effluent-dominated river: a case study of wastewa- in drinking water systems across the USA. Environ. Sci. Technol. 49 (2), 982–989. ter impact on downstream water quality. Water Res. 43 (6), 1755–1765. Rice, J., Wutich, A., Westerhoff, P., 2013. Assessment of de facto wastewater reuse across Chen, J., Wu, X., Finlayson, B.L., Webber, M., Wei, T., Li, M., Chen, Z., 2014. Variability and the US: trends between 1980 and 2008. Environ. Sci. Technol. 47 (19), 11099–11105. trend in the hydrology of the Yangtze River, China: annual precipitation and runoff. Rice, J., Via, S.H., Westerhoff, P., 2015a. Extent and impacts of unplanned wastewater J. Hydrol. 513, 403–412. reuse in US rivers. J. Am. Water Works Assoc. 107 (11), 93. Cheng, S.P., 2003. Heavy metal pollution in China: origin, pattern and control. Environ. Sci. Rice, J., Via, S., Westerhoff, P., 2015b. Extent and impacts of unplanned wastewater reuse Pollut. Res. 10 (3), 192–198. in U.S. Rivers. J. Am. Water Works Assoc. 107 (11), E571–E581. Chu, J., Chen, J., Wang, C., Fu, P., 2004. Wastewater reuse potential analysis: implications Rice, J., Wutich, A., White, D.D., Westerhoff, P., 2016. Comparing actual de facto wastewa- for China's water resources management. Water Res. 38 (11), 2746–2756. ter reuse and its public acceptability: a three city case study. Sustainable Cities and Ding, H., Meng, L., Zhang, H., Yu, J., An, W., Hu, J., Yang, M., 2013. Occurrence, profiling and Society. prioritization of halogenated disinfection by-products in drinking water of China. En- Rowney, N.C., Johnson, A.C., Williams, R.J., 2011. Erratum: cytotoxic drugs in drinking vironmental Science: Processes & Impacts 15 (7), 1424–1429. water: a prediction and risk assessment exercise for the Thames catchment in the Duirk, S.E., Lindell, C., Cornelison, C.C., Kormos, J., Ternes, T.A., Attene-Ramos, M., Osiol, J., United Kingdom. Environ. Toxicol. Chem. 30 (7), 1729. Wagner, E.D., Plewa, M.J., Richardson, S.D., 2011. Formation of toxic iodinated disin- Sang, L.-H., Huang, W., Liao, Z.-D., 2007. Present situation of urban sewage treatment and fection by-products from compounds used in medical imaging. Environ. Sci. Technol. water saving effect in Yangtze River basin. Journal of Yangtze River Scientific Re- 45 (16), 6845–6854. search Institute 4, 007. Finlayson, B.L., Barnett, J., Wei, T.Y., Webber, M., Li, M.T., Wang, M.Y., Chen, J., Xu, H., Chen, Schwarzenbach, R.P., Egli, T., Hofstetter, T.B., von Gunten, U., Wehrli, B., 2010. Global Z.Y., 2013. The drivers of risk to water security in Shanghai. Reg. Environ. Chang. 13 water pollution and human health. In: Gadgil, A., Liverman, D.M. (Eds.), Annual Re- (2), 329–340. view of Environment and Resources. Annual Reviews: Palo Alto Vol 35, Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Barber, L.B., pp. 109–136 Vol. 35. Thurman, M.E., 2008. A national reconnaissance for pharmaceuticals and other or- Shao, D., Wang, Z., Wang, B., Luo, W., 2016. A water quality model with three dimensional ganic wastewater contaminants in the United States — II untreated drinking water variational data assimilation for contaminant transport. Water Resour. Manag. 1–12. sources. Sci. Total Environ. 402 (2–3), 201–216. USEPA, 1980. Wastewater in Receiving Waters at Water Supply Abstraction Points. Fono, L.J., McDonald, H.S., 2008. Emerging compounds: a concern for water and wastewa- United States Environmental Protection Agency, Cincinatti, OH. ter utilities. J. Am. Water Works Assoc. 100 (11) 50−+. Varis, O., Vakkilainen, P., 2001. China's 8 challenges to water resources management in Guo, X., Li, J., Yang, F., Yang, J., Yin, D., 2014. Prevalence of sulfonamide and tetracycline the first quarter of the 21st century. Geomorphology 41 (2), 93–104. resistance genes in drinking water treatment plants in the Yangtze River Delta, Varis, O., Kummu, M., Salmivaara, A., 2012. Ten major rivers in monsoon Asia-Pacific: an China. Sci. Total Environ. 493, 626–631. assessment of vulnerability. Appl. Geogr. 32 (2), 441–454. Hajj-Mohamad, M., Aboulfadl, K., Darwano, H., Madoux-Humery, A.S., Guérineau, H., Wang, D., Zhang, L., Zhang, Z., Zhao, S.X., 2011. Urban infrastructure financing in reform- Sauvé, S., Prévost, M., Dorner, S., 2014. Wastewater micropollutants as tracers of sew- era China. Urban Stud. 48 (14), 2975–2998. age contamination: analysis of combined sewer overflow and stream sediments. En- Webber, M., Barnet, J., Chen, Z.Y., Finlayson, B., Wang, M., Chen, D., Chen, J., Li, M.T., Wei, vironmental Science Processes & Impacts 16 (10), 2442–2450. T.Y., Wu, S., Xu, H., 2015. Constructing water shortages on a huge river: the case of Johnson, A.C., Oldenkamp, R., Dumont, E., Sumpter, J.P., 2013. Predicting concentrations of Shanghai. Geogr. Res. 53 (4), 406–418. the cytostatic drugs cyclophosphamide, carboplatin, 5-fluorouracil, and capecitabine Williams, R.J., Churchley, J.H., Kanda, R., Johnson, A.C., 2012. Comparing predicted against throughout the sewage effluents and surface waters of europe. Environ. Toxicol. measured steroid estrogen concentrations and the associated risk in two United Chem. 32 (9), 1954–1961. Kingdom river catchments. Environ. Toxicol. Chem. 31 (4), 892–898. Kaplan, S., 2013. Review: pharmacological pollution in water. Crit. Rev. Environ. Sci. Zhang, Q., Xu, C.-Y., Becker, S., Jiang, T., 2006. Sediment and runoff changes in the Yangtze Technol. 43 (10), 1074–1116. River basin during past 50 years. J. Hydrol. 331 (3), 511–523. Kerstens, S., De Mes, T., Lue, B., 2009. Designing sustainable sanitation in urban planning Zhang, C., Zhong, L.J., Fu, X.T., Wang, J., Wu, Z.X., 2016. Revealing water stress by the ther- proposed for Changzhou, China. Water Sci. Technol. 60 (12), 3165–3172. mal power industry in China based on a high spatial resolution water withdrawal and Kugathas, S., Williams, R.J., Sumpter, J.P., 2012. Prediction of environmental concentra- consumption inventory. Environ. Sci. Technol. 50 (4), 1642–1652. tions of glucocorticoids: the river Thames, UK, as an example. Environ. Int. 40, 15–23. Liu, X., Liu, C., Luo, Y., Zhang, M., Xia, J., 2012. Dramatic decrease in streamflow from the headwater source in the central route of China's water diversion project: climatic variation or human influence? J. Geophys. Res. Atmos. 117, D6.