Nutrient Loads Flowing into Coastal Waters from the Main Rivers of China (2006–2012)

Based on monthly monitoring data of unfiltered water, the nutrient discharges of the eight main rivers flowing into the coastal waters of China were calculated from 2006 to 2012. In 2012, the total load of NH3-N (calculated in nitrogen), total nitrogen (TN, calculated in nitrogen) and total phosphorus (TP, calculated in phosphorus) was 5.1 × 105, 3.1 × 106 and 2.8 × 105 tons, respectively, while in 2006, the nutrient load was 7.4 × 105, 2.2 × 106 and 1.6 × 105 tons, respectively. The nutrient loading from the eight major rivers into the coastal waters peaked in summer and autumn, probably due to the large water discharge in the wet season. The Yangtze River was the largest riverine nutrient source for the coastal waters, contributing 48% of the NH3-N discharges, 66% of the TN discharges and 84% of the TP discharges of the eight major rivers in 2012. The East China Sea received the majority of the nutrient discharges, i.e. 50% of NH3-N (2.7 × 105 tons), 70% of TN (2.2 × 106 tons) and 87% of TP (2.5 × 105 tons) in 2012. The riverine discharge of TN into the Yellow Sea and Bohai Sea was lower than that from the direct atmospheric deposition, while for the East China Sea, the riverine TN input was larger.

as a result of improving living standards in China, many people have shifted their dietary preferences toward animal-derived products 11,12 . These dietary changes were also associated with increased nutrient inputs for agriculture 13,14 .
Wastewater and other industrial emissions caused by the urbanization and industrialization also increased nutrient inputs to the rivers and coastal waters. From 2006 to 2012, the wastewater produced in China increased from 536.8 × 10 8 to 684.6 × 10 8 tons 15,16 . Compared with the rapid growth in wastewater amounts, wastewater treatment facilities could not keep up with the urbanization progress. From 2000 to 2005, only 30-45% of the emitted wastewater was treated before being discharged into waters 17 . After 2006, in order to control the pollution of point sources, the Chinese government began to emphasize the construction of wastewater treatment plants. The number of wastewater treatment plants increased from 1040 in 2006, to 2496 in 2010 18 , with the daily sewage treatment capacity reaching 1.25 × 10 8 m 3 in 2010. However, the growing number of wastewater treatment facilities still can't meet the demands of increasing wastewater. In 2010, the percentages of treated sewage in urban and rural in China were only about 82% and 60%, respectively 18 .
The nutrients emitted from both point and non-point sources are ultimately discharged to the coastal water through the rivers. Currently, most statistical studies for Chinese cases only focus on changes in nutrients concentrations or transport in small-scale watersheds [19][20][21][22] . For instance, Chen et al. (2000) found seasonal variances of nitrogen content varied with watersheds in the Yangtze River system, and the difference of nitrogen contamination level was related to the regional population and economic development 23 . Li et al. (2014) concluded that increased nutrient loads from the Yangtze River had led to increased Harmful Algal Blooms 3 .
The large scale nutrient export estimation to the coastal water always relied on model estimation. For example, based on the NEWS model, Qu and Kroeze (2010) calculated the export of dissolved nitrogen and phosphorus of the rivers in China, and their results suggested between 1970 and 2000 the dissolved nitrogen and phosphorus export to the coastal water increased significantly, while the export of other nutrients changed less 24 . Strokal et al. (2014) modeled the riverine inputs of nitrogen and phosphorus, and suggested dissolved nitrogen and phosphorus inputs to the Bohai Gulf, Yellow Sea and South China Sea increased by a factor of 2-5 between 1970 and 2000 14 . However, although the modeling research has been carried out recently, the knowledge about the riverine nutrient discharges into the seas in China is still limited.
In this study, monthly monitoring data (unfiltered water) of eight major Chinese rivers was used to estimate the nutrient flux into the seas of China from 2006 to 2012. The nutrients that were measured in include NH 3 -N (calculated in nitrogen), total nitrogen (TN, calculated in nitrogen) and total phosphorus (TP, calculated in phosphorus). The temporal and spatial variations of the nutrient concentrations of the eight rivers were analyzed. The monthly, seasonal and annual nutrient discharges into the coastal waters of China from 2006 to 2012 were calculated, and contributions of nutrient discharges from different rivers were compared. The impacts of construction on wastewater treatment facilities were also discussed.

Results
Concentrations of NH 3 -N, TN and TP. The yearly NH 3 -N concentrations in the eight large rivers are presented in Fig. 1. In 2012, the highest NH 3 -N concentration was found in the Haihe River, with an average of 3.7 mg/L, and the lowest was found in the Minjiang River, with an average of 0.3 mg/L. Generally, the NH 3 -N concentrations have decreased since 2006. For example, the average NH 3 -N concentration for the Haihe River in 2006 was 14.0 mg/L (the monthly values ranging from 7.0 to 24.3 mg/L), decreasing to 3.7 mg/L (ranging from 0.3 to 8.8 mg/L) in 2012. For the Liaohe River, the average NH 3 -N concentration was 3.7 mg/L (ranging from 0.4 to 9.0 mg/L) in 2006, decreasing to 1.0 mg/L (ranging from 0.2 to 2.4 mg/L) in 2012. For the Yangtze River, the average concentrations were 0.7 mg/L (ranging from 0.2 to 2.7 mg/L) in 2006 and decreased to 0.3 mg/L in 2012. Compared with the rivers in North China, the NH 3 -N concentrations were generally lower for the rivers in South China. In 2012, the average monthly NH 3 -N concentrations of the four rivers in northern China (Liaohe, Huanghe, Haihe and Huaihe Rivers) were significantly higher than those of the rivers in the south (Yangtze, Qiantangjiang, Minjiang and Zhujiang Rivers).
The yearly TN concentrations in the seven large rivers (excluding the Liaohe River) are provided in Fig. 2. During the period between 2006 and 2012, the TN concentrations were generally lower than 5 mg/L for a majority of the rivers (except the Haihe River). In 2012, the highest TN concentration was found in the Haihe River, with an average of 7.0 mg/L, and the lowest concentration was found in the Huanghe River, with an average of 0.   Table 1. In 2012, the total load of NH 3 -N, TN and TP of the selected rivers was 5.1 × 10 5 , 3.1 × 10 6 and 2.8 × 10 5 tons, respectively. In 2006, the nutrient load was 7.4 × 10 5 , 2.2 × 10 6 and 1.6 × 10 5 tons for NH 3 -N, TN and TP, respectively. The Yangtze River was the largest riverine nutrient source for the coastal waters. In 2012, Yangtze River contributed 48% of the NH 3 -N discharges, 66% of the TN discharges and 84% of the TP discharges of the eight major rivers. The Zhujiang River was the second greatest nutrient source for the coastal waters. In 2012, Zhujiang River contributed about 43% of the NH 3 -N discharges, 26% of the TN discharges and 12% of the TP discharges. The Huanghe River is the second longest river in China, but the nutrient discharge to coastal waters was much lower than that of the Yangtze and Zhujiang Rivers. The small nutrient discharge could possibly be attributed the low water discharge of the Huanghe River, which was caused by excessive water use and climate change in the Huanghe River Basin 25,26 . The nutrient loading varied a lot among different seas ( Table 2). The East China Sea received the majority of the nutrient discharges. In 2012, about 50% of the NH 3 -N (2.7 × 10 5 tons), 70% of the TN (2.2 × 10 6 tons), 87% of the TP (2.5 × 10 5 tons) loads were discharged into the East China Sea, while only 3% of the NH 3 -N loads and less than 1% of TN and TP loads flowed into the Bohai Sea ( Table 2). The South China Sea is also an important destination for nutrients. In 2012, about 2.2 × 10 5 tons of NH 3 -N, 8.0 × 10 5 tons of TN and 3.5 × 10 5 tons of TP were discharged into the South China Sea.

Discussion
The monthly variation of NH 3 -N, TN and TP concentrations in the eight rivers is provided in Figs 4, 5 and 6, respectively. For the Yangtze River, Liaohe River and Huaihe River, an annual peak in NH 3 -N concentration was observed in February-March (dry season), indicating high NH 3 -N concentrations at low flows. For these rivers, a negative relationship was found between the monthly water discharge and NH 3 -N concentrations (p = 0.002). As suggested by previous studies, the negative relationship could indicate the effects of point sources (such as discharge of sewage and industrial wastewater) on the NH 3 -N concentrations in the rivers 27 . For the Huanghe River and Minjiang River, the correlation between NH 3 -N and water discharge was not significant, and NH 3 -N concentrations showed small variations between months, implying that these rivers could have a mixture of diffuse sources, point sources and non-point sources 23 .
Monthly variation of TN concentrations in the rivers is shown in Fig. 5. For the Yangtze River, a monthly variation similar to NH 3 -N was found, and the annual peak always occurred in February-March (dry season). For the Qiantangjiang and Zhujiang Rivers, a "double-peak" of TN concentrations was observed. The peak value of TN concentrations appeared in the autumn and winter (or in the early spring), and the lowest value was usually detected in the summer. For example, in the Zhujiang River in 2007, a double-peak occurred in August and December, with a value of 4.81 and 4.53 mg/L, respectively, while the lowest concentration occurred in April, with a value of 2.07 mg/L. This phenomenon was similar to that reported in a previous study 3 , and has been attributed to the intense mixing in the water masses 3,28 . Generally, nutrient concentrations in bottom water were fairly abundant due to the    remineralization of organic matter. The vertical mixing from surface to bottom water was stronger in autumn and winter than in spring and summer. Hence, nutrient concentrations peaked during autumn and winter. In all the dissolved inorganic nitrogen (DIN) in the rivers, the nitrate is the dominant form, usually occupying for over 80% 29,30 , while in this study, a high proportion of NH 3 -N in TN was observed. The difference of nutrient concentrations between their studies and ours could possibly be attributed to the difference of water pretreatment. In the studies by Gao et al. and Liu et al. 29,30 , filtered water samples were used for the nutrient analysis, while in this study, the unfiltered water samples were analyzed for the NH 3 -N, TN and TP. Nutrients in the suspended particles have proved to account for a large proportion of nutrients in the water. It was reported that the particulate phosphorus (PP) occupied for 62-99% of the total phosphorus (TP) in the Huanghe River 31 . TN concentrations in the rivers were also much higher than DIN. In Yangtze River, DIN concentrations in the surface water were 0.53 mg/L, while the corresponding TN concentrations were 1.15 mg/L 32 . Compared with the dissolved nitrogen species such as NO 3 -N, the suspended particles in the water could adsorb the NH 3 -N much more easily 33,34 , leading to a higher proportion of NH 3 -N in the TN.
For the Yangtze, Huaihe, Qiantangjiang and Minjiang Rivers, a positive relationship was found between the TN and TP concentrations (p = 0.001, 0.02, 0.01, 0.003, respectively). This indicated that the TN and TP concentrations were affected by similar factors. However, for the Zhujiang River, no significant correlation was observed between the TP and TN concentrations (p = 0.56).
The yearly loading of NH 3 -N into coastal waters showed a decreasing trend from 2006 to 2012 (Table 1). Compared with 2006, the riverine NH 3 -N loads decreased by 31% in 2012, from 7.4 × 10 5 to 5.1 × 10 5 tons. As the largest NH 3 -N contributor to the coastal seas, the Yangtze River was showed a decrease in NH 3 -N discharge from 4.4 × 10 5 tons in 2006 to 2.4 × 10 5 tons in 2012. The decreasing trend of NH 3 -N loads was consistent with the decrease of NH 3 -N concentrations in the river. The decline of NH 3 -N discharges could possibly be attributed to the fact that more and more sewage treatment plants have been put into operation in China. According to a report from the Ministry of Housing and Urban-Rural Development, the sewage treatment capacity of cities increased from 6.5 × 10 7 m 3 /day in 2006 to 1.0 × 10 8 m 3 /day in 2010 18 . The reduction of NH 3 -N is a priority in sewage treatment plants, and the NH 3 -N abatement by the sewage treatment plants increased from 2.6 × 10 5 tons in 2006 to 7.7 × 10 5 tons in 2010 18 . In China, the total anthropogenic emission of NH 3 -N (including industrial source, domestic source, agricultural source, etc.) in 2011 and 2012 was estimated to be 2.6 × 10 6 and 2.5 × 10 6 tons, respectively 16,35 . According to our calculation, the riverine NH 3 -N discharge in 2011 and 2012 was 4.7 × 10 5 and 4.9 × 10 5 tons, respectively. This indicates that approximately 18% and 19% of the anthropogenic NH 3 -N emission was released into the coastal waters through rivers. The direct NH 3 -N discharges into the coastal waters from sewage treatment plants in coastal cities were reported to be 2.3 × 10 4 tons in 2010, 1.6 × 10 4 tons in 2011, and 1.2 × 10 4 tons in 2012, respectively 16,35,36 . Compared with the riverine NH 3 -N discharge, the direct NH 3 -N discharge from the coastal cities was minor.
For TN and TP discharges, an increase was found in 2012 compared to 2006 (Table 1). In 2006, the TN and TP discharges from the eight major rivers into the coastal waters were 2.2 × 10 6 and 1.6 × 10 5 tons, respectively, and increased to 3.1 × 10 6 and 2.8 × 10 5 tons in 2012. For the Yangtze River, the discharge of TN increased from 1.4 × 10 6 tons in 2006 to 2.0 × 10 6 tons in 2012, and the TP discharge increased from 1.2 × 10 5 in 2006 to 2.4 × 10 5 tons in 2012. It was reported that the average annual loading of TN from Mississippi-Atchafalaya River into Mexico Gulf was 1.6 × 10 6 tons 37 . In this study, the TN loading from the Yangtze River was approaching that of the Mississippi River. The direct TP discharge from the sewage plants along the coast into the coastal waters was reported to be 2901 tons in 2010, 2447 tons in 2011 and 2196 tons in 2012 16,35,36 . Compared with the riverine TP discharge, the direct TP discharge from sewage plants of the coastal cities into the coastal waters was trivial.
The nutrient loading from the eight major rivers into the coastal waters always peaked in summer and autumn (Table S1 in the Supplementary Information). For the Yangtze River, from 2006 to 2012, the average TN load was 5.0 × 10 5 tons (ranging from 3.7 × 10 5 to 6.8 × 10 5 tons) in summer and 6.2 × 10 5 tons (ranging from 3.6 × 10 5 to 7.9 × 10 5 tons) in autumn, higher than the TN load in spring (averaging 9.8 × 10 4 tons, ranging from 6.3 × 10 4 to 1.8 × 10 5 tons). The TN and TP loads in summer and autumn accounted for 64% and 67% of the annual discharge, respectively. For the Zhujiang River, the average TN load was 2.9 × 10 5 tons (ranging from 1.9 × 10 5 to 4.3 × 10 5 tons) in summer and 2.7 × 10 5 tons (ranging from 1.0 × 10 5 to 3.7 × 10 5 tons) in autumn, higher than the TN load in spring (averaging 1.0 × 10 5 tons, ranging from 0.7 × 10 5 to 1.4 × 10 5 tons). The monthly loads of NH 3 -N, TN and TP of the selected rivers have the same trend of incremental value from dry season to wet season ( Figure S1 in the Supplementary  Information). For instance, for the Yangtze River and Zhujiang River, a significant relationship between the water discharge and nutrient loads was observed (p = 0.003). Hence, the high nutrient loads in the wet season could possibly be attributed to the large water discharges.
Previous studies reported that atmospheric nitrogen deposition played an important role in the nitrogen cycles in coastal ecosystems in China 38 . Bashkin et al. (2002) estimated that the atmospheric deposition of inorganic nitrogen to the Yellow Sea and Bohai Sea was 1.06 × 10 6 tons/year 39 . In our calculation, the riverine discharge of total nitrogen into the Yellow Sea and Bohai Sea was 3.7 × 10 4 tons, much lower than those from direct atmospheric deposition. Previous studies reported that atmospheric deposition of dissolved inorganic nitrogen to the East China Sea was 1.75 × 10 6 tons/year 40  times the riverine total nitrogen inputs (2.2 × 10 6 tons) in 2012. These results indicated that direct atmospheric deposition was the dominant source of nitrogen in the Yellow Sea and Bohai Sea, while for the East China Sea, the riverine discharges had higher contributions compared with atmospheric deposition.

Methods
Study area. In this study we examined the eight main rivers flowing into the coastal waters of China include the Yangtze River, Huanghe River, Liaohe River, Haihe River, Huaihe River, Qiangtangjiang River, Minjiang River and Zhujiang River (Table 3 and Fig. 7). The total drainage area of the eight rivers was 319.6 × 10 4 km 2 (Table 3), occupying one third of the land area of China 41 . The eight rivers accounted for the majority of water discharge into the coastal waters in China. The annual water discharge of the rivers decreases from the south to the north. For example, the annual water discharge of Zhujiang River was 2833 × 10 8 m 3 , 182 times as much as the water discharge of the Haihe River (Table 3). The peak of water discharge in the selected rivers occurred in the period from April to September (wet season) 41 . The water discharge information of the rivers is presented in the Supplementary Information (Figure S2).
Due to the "west high and east low" terrains, the rivers mostly flow from the west to the east. Throughout the length of their courses, the rivers receive pollutants from point and non-point sources, and discharge into coastal waters. The rivers discharge into different seas, including the Bohai Sea, Yellow Sea, East China Sea and South China Sea, and these seas cover a total area of about 4.7 million km 2 38 . The East China Sea received 3/4 of the total water discharge. Data source and calculation. Ammonia nitrogen (NH 3 -N), total nitrogen (TN) and total phosphorus (TP) were monitored and collected every month from 2006 to 2012 by the China Ministry of Environmental Pollution, and the monitoring stations were located at the estuaries of the eight main rivers (shown in Fig. 7). The field sampling was carried out according to "Technical Specifications Requirements for Monitoring of Surface Water and Waste Water, in China (HJ/T 91-2002)" 42 . In brief, for the rivers with a width less than 50 m, a vertical sampling was set at the middle thread of channel. For the rivers with a width between 50 and 100 m, two vertical sampling were set at both left and right sides (at the place with appreciable currents), respectively. For the rivers with a width more than 100 m, three vertical sampling were set at the left side, middle thread of channel and right side, respectively. In order to avoid the impacts of point sources, the sampling was carried out at the site which could represent the general nutrient conditions, but away from the point sources such as sewage outfalls. Specifically, water mixture samples (surface: 50 cm under the surface; middle: 1/2 of the river depth; bottom: 50 cm above the riverbed) were collected using a Polyethylene bailer for each site at the ebb time, when the river is naturally flowing. The sampling equipment was cleaned thoroughly with deionized water between each site to avoid cross contamination. About 0.5~1 L water samples were collected each time. As soon as the waters samples were collected, H 2 SO 4 (GR) was added into the samples to make pH in the samples lower than 2 to avoid the impacts form the microbial processes, etc. The water samples were kept in the refrigerator before analysis (at the temperature of 4 °C). The laboratory measurement was carried out according to the "Environmental Quality Standard for Surface Water, China (GB 3838-2002)" 43 . For the riverine nutrient discharges, besides the dissolved nutrients, nutrients in the suspended particles should also be considered. Hence, in this study, the unfiltered water samples were used for analysis. The measurement was carried out within 24 hours for NH 3 -N and TP, and 7 days for TN. The limits of detection for NH 3 -N, TP and TN in the water samples were 0.01, 0.01 and 0.05 mg/L, respectively. The annual nutrient discharges were calculated by the monthly nutrient concentrations and water discharge. The following formula (E.g. 1) was used to estimate the annual nutrient fluxes (F Nutrient (tons)) from 2006 to 2012. C Nutrient (mg/L) and V Runoff (10 8 m 3 ) refer to the monthly concentrations of nutrients and water discharge in the selected rivers, respectively. 100 (tons·L/(mg·m 3 )) is the unit conversion factor. The water discharge information was collected at the hydrological stations of the river estuary, and these sites were also near the water monitoring sites in our study.