Per fl uoroalkyl acids (PFAAs) have unique properties including surface activity, repellency of water and oil, and resistance of acid and heat (Giesy and Kannan, 2002). In industry, they are widely used in manufacturing processes and products (Giesy et al., 2006), but when dispersed into the environment, they can transport over long distances and accumulate to toxic concentrations, due to their persistence (Giesy et al., 2010). Ionic PFAAs, mostly known as per- fl uoroalkyl carboxylic acids (PFCAs) and per fl uoroalkyl sulfonic acids (PFSAs), are relatively soluble compared to many of the organochlorine compounds with similar molecular size such as DDT and Hexachlorocyclohexane. Thus, water is the primary reservoir of PFAAs and the major medium for their transportation (Prevedouros et al., 2005). Concerns over sources, transport and fate of PFAAs in the aquatic environment have grown rapidly in recent years. Per fl uorooctane sulfonic acid (PFOS) is the most frequently detected PFAA that is released from a variety of diffuse sources, and is usually the predominant PFAAs detected in aquatic biota (Giesy et al., 2010). Per fl uorooctanoic acid (PFOA) is discharged primarily from point-sources in industrial regions, especially from manufacturing facilities (Pistocchi and Loos, 2009). Wastewater treatment plants (WWTPs) are also important point sources of releasing PFAAs, along with more diffuse inputs including rain, dry deposition and release during the use of products (Muller et al., 2011). The majority of PFAAs reach the coastal marine environment dissolved in water, while for long chain PFCAs (C12 e C15), about a half of the load was absorbed to particles during transportation (Zushi et al., 2012). Higher af fi nity for organic carbon leads to enhanced sorption of longer chain PFCAs to particles and solids in sediment and sludge (Armitage et al., 2009). The transportation of PFAAs in water indicated that rivers would be the main source of PFAAs for coastal water. Due to the re- striction on the production and use of C8 PFOS and PFOA, the C4 and C6 chemicals have been developed well to adequately replace most current C8 and higher homologues. However, this has also led to the emergence of short chain PFAAs in the environment (Moller et al., 2010; Oliaei et al., 2013). This suggested the importance of the measurement of PFAAs with different carbon chain lengths with different properties, in order to trace the trend of PFAAs contamination in the environment. The Bohai-Rim Economic Circle is a highly urbanized and industrialized region in Northern China (Fig. S1). There are more than 40 rivers fl owing into the Bohai Sea, a semi-enclosed sea. Based on estimation of mass fl uxes of several chemical pollutants (i.e. petroleum hydrocarbons, heavy metals) to the Bohai Sea, the rivers contributed 50% e 70% of the total inventory among the fi ve sources: rivers, drains, atmospheric deposition, cultivation and non-point sources (Wang and Li, 2006). The authors have conducted systematic studies on PFAAs in the northern part of the Bohai coastal region since 2008, and found that PFAAs were widely distributed in the environmental matrices with PFOA dominant in the Northern Bohai coastal region (Wang et al., 2011b), PFOS was dominant in the aquatic products in Tianjin (Chen et al., 2011), PFAAs concentrations in surface water were correlated with the level of industrialization in Northern China (Wang et al., 2011a), Fluoropolymer production in the northern Bohai coastal region had posed potential impacts to local soils (Wang et al., 2013b), and the PFOS emissions from industrial and domestic sources in the eastern coastal region of China were identi fi ed and estimated (Chen et al., 2009; Xie et al., 2013a, 2013b). Results of the studies conducted by other researchers in the northern part also indicated the presence of great concentrations of PFAAs in river water, sediment, soil, precipitation, organisms and human blood from various sources including fl uorine industry parks (Jin et al., 2007; Bao et al., 2009, 2010; Liu et al., 2009; Li et al., 2011; Pan et al., 2011). In recent years, urbanization has also sped up in the southern part of the Bohai coastal region, while information on the concentrations of PFAAs in the rapidly urbanized southern portion of Bohai-Rim is still limited. This study is an extension of our research on tracing the source and fate of PFAAs from adjacent riverine and estuarine areas of the Bohai Sea. The major aim was to investigate the key species of PFAAs in southern Bohai coastal region, and to identify their potential sources and fate. With less usage and strict control over emissions of PFOS in the world, it is still necessary to measure the whole series of known PFCAs and PFSAs in surface water for understanding the status and trends of their production and presence in the environment. Spatial analysis of inte- grated geographic information will help to discover potential ef fl uents and understand fate and transport from sources to surface water and from rivers to the sea. The hydrological cycle between land and ocean brings PFAAs from the land to the sea through river fl ow. Thus, the quanti fi cation of PFAAs loading in each river to the southern Bohai Sea will contribute information not only to understanding the behavior of PFAAs, but also to more accurate modeling of trace contaminants in these systems for estimating potential risks. Among the 17 PFAAs quanti fi ed, the concentrations of per- fl uorododecanoic acid (PFDoA), per fl uorotridecanoic acid (PFTrDA), per fl uorotetradecanoic acid (PFTeDA), per fl uorohexadecanoic acid (PFHxDA), Per fl uorooctadecanoic acid (PFODA) and per- fl uorodecanesulfonate (PFDS) were less than the LOQ in all samples, therefore they were not discussed further in this or the following sections. Concentrations of the remaining PFAAs were listed in Fig. 1 and Table S5. PFAAs were detected in all the rivers with concentrations of sum P PFAAs ( PFAAs) ranging from 2.21 to 5068.97 ng/L. In the Xiaoqing River, PFOA was the dominant PFAA with a mean concentration of P 3112.28 ng/L, which contributed 90.1% of the PFAAs, and was followed by short chain PFCAs, including PFBA (mean concentration of 49.80 ng/L, 1.4%), PFPeA (mean concentration of 70.97 ng/L, 2.0%), PFHxA (mean concentration of 123.34 ng/L, 3.5%) and PFHpA (mean concentration of 91.72 ng/L, 2.6%). For concentrations of the other long chain PFCAs and all PFSAs, the total contribution was less than 1%. In the remaining 11 rivers, the mean concentration of P PFAAs was 25.78 ng/L with the average contribution of individual PFAA in decreasing order of percentage: PFOA (38.2%) > PFBA (19.5%) > PFOS (14.3%) > PFHxA (6.7%) > PFPeA (6%) > PFBS (3.9%) > PFNA (4.6%) > PFHpA (4.3%) > PFDA. (1.1%) > PFHxS (1%) > PFUdA (0.2%) (Fig. 2a). PCA analysis on the 11 PFAAs and 9 water parameters showed that PFBA, PFPeA, PFHxA, and PFOA were associated when the Xiaoqing River was excluded due to the extremely high levels of PFOA (Fig. 2b). When the Xiaoqing River was included, the association among the four PFCAs became much stronger (Fig. 2c). This indicated that these compounds might come from similar sources. However, there were still differences in the two scenarios, which might explained by the different weights of various sources to the rivers. Concentrations of PFOS ranged from 0.40 to 12.78 ng/L with a mean concentration of 3.09 ng/L for these rivers. The highest concentration of PFOS was in SR-1, where concentrations of PFBS (24.19 ng/L) and PFHxS (0.59 ng/L) were also highest. SR-1 was located in an estuary where there might be local releases of PFSAs (Table S1). Furthermore, concentrations of PFOS and PFBS were always correlated in the two scenarios of PCA (Fig. 2b and c), which indicated a probable co- emission of these two PFSAs in the study area. The pro fi le of PFAAs showed that the south Bohai coastal rivers were mainly contaminated by PFOA followed by shorter chain PFCAs and PFOS. In recent studies, PFOA has been found to be the predominant PFAA in North Bohai coastal rivers (Wang et al., 2011b), rivers in Tianjin (Pan et al., 2011), Dianchi Lake (Zhang et al., 2012), Hanjiang River (Wang et al., 2013a), the Huaihe River Basin and Taihu Lake in China (Yu et al., 2012); Water samples from Hanoi city and its surrounding areas in Vietnam (Kim et al., 2013); Yodo River basin in Japan (Lien et al., 2008); the watershed of River Po in Northern Italy (Loos et al., 2008); and Mediterranean coastal rivers in Spain (Sánchez-Avila et al., 2010). However, only water in the Yodo and Po Rivers contained concentrations of PFOA that exceeded 1000 ng/L, which was comparable to this study (Table S7). As far as we know, this is the highest concentration of PFOA in river water of China that has ever been reported. High levels of PFOA implied local point sources in this area. As a result, the fl uoropolymer industry in Shandong Province was investigated and the major manufacturing facilities were found located along the Xiaoqing River (Fig. 3), with production begun in 2001. The manufacturing history of the pro fi le of PFAAs and related products in these facilities is unknown, but until now, fl uorinated refrigerants, intermediates for production of pesticides and medicine, polytetra fl uoroethylene (PTFE) and tetra- fl uoroethylene (TFE) have been the main products of these facilities. Facility 1 is the largest with an annual capacity of 37,000 tons of PTFE, 50,000 tons of TFE, 10,000 tons of hexa- fl uoropropylene (HFP), and more than 200,000 tons of different types of fl uorinated refrigerants by the end of 2012 (Fig. 4) (Dongyue Group Limited, 2012). The fl uoropolymers production capacities of the other facilities ranged from hundreds to thou- sands of tons. PFOA is mainly produced and used as ammonium per fl uorooctanoate (APFO) and further used as important pro- cessing additives for production of fl uoropolymers and fl uo- roelastomers. For example, high purity APFO is used primarily in the dispersion polymerization process to produce PTFE. PTFE has unique properties like repellence to acid and alkali, thermal resistance, almost insoluble in solvents etc. Thus it has been used in many industrial and consumer products, including soil, stain, grease, and water resistant coatings on textiles and carpet; uses in the automotive, mechanical, aerospace, chemical, electrical, medical, and building/construction industries; personal care products; and non-stick coatings on cookware (European Commission, 2010). The emission of PFOA came from both the PTFE production discharge and applications of PTFE products (Fig. 4). Further study is needed to estimate the weights of the two ways for PFOA emissions. The fl uorinated refrigerants (like R22), TFE and HFP are all important intermediates to produce PTFE and other fl uoropolymers in different processes, with limited information on their emission of related PFAAs. While fl uorinated ethylene propylene (FEP) is a modi fi ed material to PTFE and the production of FEP is still at the primary stage. Studies on these intermediates and new materials are also important not only to quantify known per fl uoroalkyl and poly- fl uoroalkyl substances (PFASs), but also to qualify unknown species. Spatial analysis of PFAAs levels, rivers, and production facilities together with the results obtained by other researchers indicated that facilities along the Xiaoqing River and its tributaries exhibited the greatest emissions of wastes, including sewage discharged directly to the river, and consequently resulted in the highest concentrations of PFOA measured in this study (Fig. 3). In addition, atmospheric transport and subsequent degradation of PFOA precursors including fl uorotelomer alcohols (FTOHs) and per- fl uorooctane sulfonyl fl uoride (POSF)-based chemicals (e.g., per fl uorooctyl sulfonamidoethanols) could account for as much as 10% of PFOA emissions (Pistocchi and Loos, 2009). In a study conducted at one of 3M ’ s largest fl uoropolymer facilities in Minnesota (USA), WWTPs from both industrial and domestic discharges played a key role in the PFAAs releases to surface waters, and stormwater runoff from PFAAs-related commercial and industrial releases might also be a signi fi cant source of PFAAs to the surface water (Oliaei et al., 2013). So the signi fi cant correlations among PFCAs with carbon chain lengths from 4 to 8 in all rivers could be explained by the direct emission of waste from manufacturing facilities, atmospheric transport and degradation of precursors, input from WWTPs, or stormwater runoff, where the weights of these pathways might vary for different rivers. For the Yellow River, there are two main reasons for less concentrations of PFAAs, especially that of PFOA. One is that the riverbed of the section in Shandong Province has an average height of 4 e 6 m above the land surface, there is almost no possibility for it to receive waste from local facilities, and contribution from the upstream was also limited. The other is that the relatively greater water discharge compared to other rivers in this study would lead to more dilution (Table S1). Although in 2006, the eight major fl uoropolymer and fl uotelomer manufactures joined the US EPA 2010/2015 PFOA Stewardship Program working toward elimination of PFOA, its precursors and related higher homologue chemicals from emissions and their products by 2015, facilities in this region as well as other facilities in China are still scaling up production to meet domestic and inter- national demand without suf fi cient regulations on PFAAs emission (Wang et al., 2013b). The trend that PFOA levels were the highest among all PFAAs detected in this study was consistent with the results of other studies conducted in this area. Concentrations of PFOA in Mollusks were predominant with the highest concentration of 126 ng/g dry weight observed in Laizhou Bay, the estuary of Xiaoqing River, which was almost 40 times higher than that of PFOS (Pan et al., 2010). In the region of Zouping County where facilities 1, 2 and 3-1 were located, the median concentration of PFOA in the whole human blood was 3.26 ng/mL, while the median concentration of PFOS was 2.19 ng/mL, which were the highest and the lowest among all cities investigated in Bohai Rim, respectively (Guo et al., 2011). PFOA was the dominant PFAA in precipitation across eastern and central China with a maximum concentration of 88 ng/ L observed in the city of Weifang (Zhao et al., 2013), where there is only small scale fl uorinated chemical manufacturing compared to those in the Xiaoqing River basin. However, Tai ’ an city was also on the list of investigation in the precipitation study and had a much lower concentration of PFOA, but it is closer to the fl uoropolymer manufacturing facilities than Weifang City. In this study, geomor- phic analysis indicated that Mount Tai (the peak in Shandong Province) might be a natural block for the transport of volatile PFAAs in the atmosphere (Fig. S1). This result demonstrated that the mass loading of PFAAs for atmospheric deposition might be determined more by regional conditions than by local conditions. The major emission of PFOS in Shandong Province came from textile treatment and metal plating (Xie et al., 2013b). Unlike PFOA, release of PFOS from these kinds of manufactures is distributed more like non-point sources (Pistocchi and Loos, 2009), and no direct-emissions from these industries were observed in this study. Furthermore, PFOS and PFOA chemicals are still used in some pesticides with exceptions to the phase-out action. For example, sul fl uramid will be phased-out by the year 2016 (Fluoride Action Network Pesticide project). In Shouguang County, which is the most famous production base of vegetables in China, although pesticide-free vegetables are dominating nowadays, the residues of PFAAs and their precursors that will degrade to PFAAs could be an issue for public health (Houde et al., 2011). PFAAs undergo a mixing process and are dissipated by waves and currents when they move from the rivers to the sea. The process of dissipation in saline waters of estuaries when calculating the mass fl ux of PFAAs has been discussed previously (McLachlan et al., 2007). In this study, on a smaller scale, the effect of the estuarine drainage areas (EDA) showed more interesting characteristics of pollution dispersion within proximate rivers. Salinity was used to explain the extent to which water mixes (Fig. S3). However, it must be illustrated that the water was sampled in different times of the day while the tidal time varied daily, so the salinity only repre- sented the status at the moment of sampling. The results showed that in all the rivers salinity in the EDA were greater than those at upstream locations despite the difference at the time of sampling. As an almost enclosed sea, the Bohai Sea has an estimated mixing time of approximately 30 years. For some of the small bays within the Bohai Sea, the turnover time could be even longer. It has been reported that poor exchange of seawater has led to accumulation of containments (i.e. chemical oxygen demand, heavy metals, nutrient salts) in these bays (Wang and Li, 2006), which indicated that in addition to the dilution process, part of the concentration of PFAAs in the EDA water might be contributed by intrusions of salty water, especially from adjacent rivers. The Yellow River represents a special case, in which the large discharge would make the estuary less in fl uenced by seawater. When a numerical model was used to simulate wave-induced, near-shore currents and transport of pollutants in Bohai Bay, it was concluded that due to the action of waves in the near-shore zone with shallow water, pollutants were transported parallel to the shoreline (Sun and Tao, 2006). For the three rivers discharging into Bohai Bay, little in fl uence on concentrations of PFOA was observed between the upstream and the EDA sites, so the in fl uence of waves on these three rivers was not obvious. However in Laizhou Bay, in addition to greater salinity, concentrations of PFOA in EDA of the Mi, Sha, Wang and Huangshui Rivers were signi fi cantly higher than those at the upstream locations. In the Yellow, Xiaoqing, Wei, Jiaolai and Jia Rivers there was no signi fi cant difference in concentrations of PFOA as a function of salinity. This indicated that the majority of the PFOA at MR-1, SR-1, WR-1 and HS-1 might come from XQ-1 with the waves and currents along the shoreline. Although concentrations of PFOA were less at other EDA sites, they might increase notably with the rising tide. The ebb-tide (recession of sea level) in Laizhou Bay would lead to a current from the top of the bay eastward to HS-1 (Zhang, 2007), which is in good agreement with the results of the present study (Fig. 3). More attention should be paid to the in fl uence of salinity changes arisen by waves and currents in the EDA as it could in- fl uence the property of PFAAs and also the physiology of organisms, which would consequently contribute to sorption and bio- accumulation of PFAAs in these organisms (Houde et al., 2011). Mass fl ux will provide information on the environmental inventory of PFAAs. In order to further eliminate the in fl uence of residual seawater, sites used to calculate the mass fl ux of PFAAs were chosen based on salinity. Although the salinity of some EDA water was less than the freshwater threshold during ebb-tide, considering the frequent mixing, upstream sites of EDA were used to estimate loadings of PFAAs from rivers to the Bohai Sea. The mass fl ux was calculated based on instantaneous concentrations of PFAAs multiplied by the average annual water discharge data to give a rough yet valuable approximation (Table 1) (McLachlan et al., P 2007; Filipovic et al., 2013). The mass fl ux of PFOA and PFAAs in the Xiaoqing River were 3.6 tons and 4 tons per year, which accounted for 90% and 80% of those among all the rivers, respectively. Even though the concentrations were less, there was the largest discharge in this study, and thus the Yellow River accounted P for the second largest mass of PFAAs and PFOS, which were about 0.4 ton and 0.07 ton per year, respectively. Excluding the extensive production capacity along the Xiaoqing River and the huge discharge of the Yellow River, mass fl ux of PFOS and PFOA into the Bohai Sea from the southern coastal rivers in this study was larger but still comparable to that of northern coastal rivers, which were calculated to be 0.02 and 0.2 ton per year for PFOS and PFOA, respectively. (Wang et al., 2011a) In comparison, the mass fl ux of PFOA from main rivers in the European Continent was estimated to be 14.3 tons per year (McLachlan et al., 2007). However, when averaged by area and population, the emission of PFOA in the South Bohai area was about 20 times greater than that in the European Continent, respectively, which posed a much heavier burden to the local environment. Concentrations of PFOA in the Xiaoqing River exceeded several drinking water criteria including the New Jersey guidance for PFOA in drinking water (40 ng/L), the US EPA provisional health advi- sories for PFOA (400 ng/L), and the Health Canada drinking water guidance value for PFOA (700 ng/L) (New Jersey Department of Environmental Protection (NJDEP), 2007; USEPA, 2009; Paterson et al., 2012). The residents in the study area have not used the river water as drinking water for a long time. However, according to a study conducted on transport of PFOA near a fl uoropolymer manufacturing facility, the atmospheric deposition would not only in fl uence concentrations in surface waters, but also the underlying aquifer by migration downward with precipitation and river recharge (Davis et al., 2007). So there is still a large potential risk to the local drinking water system. Meanwhile, river water is used for irrigation, which might pose risks due to PFAAs in soils and subsequent accumulation into crops and vegetables and eventually accumulation in humans. None of the concentrations of PFOS or PFOA measured in this study exceeded any water quality criteria for the protection of freshwater aquatic organisms including criteria maximum concentration (CMC) and criteria continuous concentration (CCC). For example, the CMC was calculated to be 3.78 mg/L for PFOS and 45.54 mg/L for PFOA, while CCC was 0.25 mg/L for PFOS and 3.52 mg/L for PFOA that were derived for the protection of freshwater aquatic life in China (Yang et al., 2014). The CMC for PFOS (21 m g/L) and PFOA (25 mg/L), CCC for PFOS (5.1 m g/L) and PFOA (2.9 mg/L) and avian wildlife value (AWV) for PFOS (47 ng/L) were also derived based on toxicology data of organisms resident in North America (Giesy et al., 2010). When biodiversity in the main EDAs in China was investigated by use of the Shannon e Wiener index, Laizhou Bay exhibited relatively greater indices, indicative of good biodiversity and a constant environment (Huang et al., 2012), and was also consistent with the in-situ fi ndings of this study. However, considering the almost lacking degradation property of PFOA, the scaling-up of its production, less water-mobility in Laizhou Bay and intensive fi shery in this area, especially the aquaculture and salt fi elds in and around the coastal mud fl at created by the rising tide (Fig. 3), further research is needed to evaluate the health risk of fi sh consumption for local residents to make sure that the pollution is controllable. Furthermore, the wave-induced near-shore current is usually considered to be the reason for sediment suspension in the near-shore zone (Sun and Tao, 2006), and many aquatic organisms ingest particles in water (Jeon et al., 2010). This would increase the bioconcentration factors of the aquatic ecosystem and also the overall risk. The present study gave a general characterization of the PFAAs pollution in the main rivers of the rapidly urbanized South Bohai coastal region with PFOA dominant in pretty high concentration. Previous studies on source identi fi cation would use a combination of geographic information, such as population density and land use as indicators for the in fl uence of urbanization on PFAAs emissions, especially for PFOS (Murakami et al., 2008; Pistocchi and Loos, 2009; Zushi and Masunaga, 2009). Because the majority of the PFOA loads are emitted from industries, distributions of fl uoropolymer industry in Shandong Province were investigated and the main facilities were identi fi ed so that monitoring could be conducted to establish the distribution, status and trend in concentrations of PFAAs in this region. The data on the concentrations of PFAAs was also combined with GIS data including land and sea DEM, and vegetation data to provide a visual description of contamination in the region. For China, the scaling-up of fl uoropolymer production would predictably bring actions on risk assessment and regulation in the future, and results in this study would provide valuable information. It may also provide hint for other countries with rapid urbanization to take precautionary approach to tackling the emerging pollution. This study was supported by the National Natural Science Foundation of China under Grant No. 41371488 and 41171394, the International Scienti fi c Cooperation Program with Grant No. 2012DFA91150, and the Key Project of the Chinese Academy of Sciences under Grant No. KZZD-EW-TZ-12. We would like to thank the editors and reviewers for their valuable comments and sug- gestions. Prof. Giesy was supported by the Einstein Professor Program of the Chinese Academy of Sciences and the Canada Research Chair program. Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2014.03.030.