Water quality assessment and pollution source analysis of Yaojiang River Basin: a case study of inland rivers in Yuyao City, China

As an essential part of the ecological environment, the water environment plays a critical role in human life and economic development. Therefore, water environmental assessment techniques have been continuously improved to understand the quality of the water environment more systematically and comprehensively. At present, the most extensively used water quality assessment techniques include the single factor index (SFI) method (Lu et al. 2019), Nemerow pollution index (NPI) method (Chen et al. 2016; Chen et al. 2021), the single factor water quality identification index (SFWQII) method (Xu 2005a), the comprehensive water quality identification index (CWQII) method (Xu 2005b; Miao et al. 2009; Ban et al. 2014a, 2014b), component analysis method (Ouyang 2005), and fuzzy comprehensive evaluation method (Jiang et al. 2020; Li et al. 2020a, 2020b). Among them, the NPI method is often used in the assessment of heavy metals in water bodies, highlighting the impact of the heavy metals with the highest pollution index on the quality of the water environment. By using the SFI method, a direct understanding of the relationship between assessed water quality status and the standard can be achieved. Even better, the SFWQII method and the CWQII method proposed by Xu (2005a, 2005b) can quantitatively and accurately analyze the water quality, and the evaluation results obtained can intuitively reflect whether the water quality grade conforms to the functional area standard and the number of water quality parameters that fail to meet the standard. At present, these two methods have been widely used in water environment assessment in China (Miao et al. 2009; Liu et al. 2010; Ban et al. 2014a, 2014b; Pan et al. 2020). Moreover, factor analysis has become one of the most applicable multivariate statistical methods and has been widely used in the study of water quality pollutant sources (Ouyang 2005; Ikem et al. 2013; Shirnezhad et al. 2020; Zhang et al. 2021). The basic principle of this protocol includes reduction of variable dimensions, extraction of fewer uncorrelated comprehensive parameters from multiple original variables, objective determination of the weights of these factors, and identification of the primary sources of pollutants.

Yuyao City is located in the middle of the Ningshao Plain in the south wing of the Yangtze River Delta. The entire water area in Yuyao City covers 289.26 km² (including sea area), where the inland water surface is basically covered by the Yaojiang River system. Yaojiang River, originating from Xiajialing, Dalan Town, Yuyao City, is one of the largest tributaries in the Yongjiang River system. It exits through Yuyao City in the southeast and runs into the East China Sea through the Yongjiang River downstream. The crucial tributaries along the Yaojiang River include Hutang River, Changling River, Linzhou River, Gaoqiao River, Linsi River, Xijiang River, Zhongjiang River, Dongjiang River, and Cijiang River. The mainstream of the Yaojiang River is a straight and vast water body with an optimal self-purification capacity. However, a large number of upstream wastewater pollutants are discharged into the river due to the economic activities and incomplete sewage treatment facilities, leading to an increase in the exceeding-standard water quality factors, expansion of the eutrophic area, and significant deterioration of the overall water quality environment. In the water quality surveys of the 1990s, the dissolved oxygen (DO) in Yaojiang River ranged from 0.42 to 7.46 mg/L, reflecting a generally low level of DO (Shen 2000). In contrast, the permanganate index (COD_Mn) values were at high levels (up to 59.6 mg/L) (Shen 2000; Wang 2001). In 1999, the average content of total nitrogen (TN) and total phosphorus (TP) was 2.45 mg/L and 0.2 mg/L, respectively, indicating eutrophication existed in Yaojiang River (Wang 2001). Besides, petroleum was another exceeding index, and the content in upstream of Yaojiang was higher than that in the downstream (Wang 2001). Although annual regular monitoring and studies of the water quality of Yaojiang River have been carried out, analyses mainly focused on the evaluation of the single monitoring index. Since water quality is affected by various factors, it is therefore necessary to conduct a comprehensive assessment of the degree of pollution of the Yaojiang River Basin. In this work, the SFWQII method and CWQII method were employed to assess the water quality of the Yaojiang River, as well as its upstream tributaries Changling River, Hutang River, and Linzhou River, in the wet and dry seasons from 2018 to 2019. Moreover, the factor analysis method was applied to analyze the pollution sources. To a certain extent, the conclusions of this study provide a reference to the assessment of water quality monitoring data and have a guiding significance for the proposal of watershed water pollution prevention and management policy.

A total of 10 sampling sections was laid out in Yuyao City, including S1 Xiachendu Village, S2 Pukou Village, S3 Wantou Village, S4 Xinzheng Village in Yaojiang River, S5 Nanwei Village and S6 Dongjia Village in Changling River, S7 Fangjia Village and S8 Wufu Village in Hutang River, S9 Xinxin Village and S10 Xinqiao Village in Linzhou River (Figure 1). Samples were collected during the wet season of 2018, the dry season of 2018, the dry season of 2019, and the dry season of 2019 according to the water storage capacity of the basin.

Water samples were collected in 1,000 mL glass bottles. Before sampling, the bottles were rinsed with raw water repeatedly. Sulfuric acid was added to the TN and TP samples to acidify the solution to pH < 2, and then the samples were sealed and stored at 0–4 °C.

According to the primary indexes of ‘Environmental Quality Standards for Surface Water (GB 3838–2002)’ and the existing monitoring results (Shen 2000; Wang 2001; unpublished data), six indexes including DO, COD_Mn, TN, TP, petroleum and volatile phenol (VP), were measured. Among them, DO was measured in situ by a portable dissolved oxygen analyzer (9010, JENCO, USA), while TN, TP, petroleum, COD_Mn, and VP were analyzed by using the methods proposed in ‘Environmental Quality Standards for Surface Water (GB 3838–2002)’. Briefly, TN and TP were measured by alkaline potassium persulfate digestion UV spectrophotometric method and ammonium molybdate spectrophotometric method, respectively. COD_Mn determination was achieved by the titration method. Infrared spectrophotometry was applied for petroleum measurement, while flow injection analysis and 4-aminoantipyrine spectrophotometric method were utilized for VP measurement.

Factor analysis of water quality data of 2018 and 2019 was performed by using SPSS software (v16.0) (Jung et al. 2016; Kale et al. 2020). Briefly, descriptive statistics were produced to note the missing/abnormal values. The correlation matrix test was adopted for the applicability test using Kaiser-Meye-Olkin (KMO) and Bartlett sphere methods. KMO detection of greater than 0.5 and Bartlett sphericity test probability of p < 0.05 indicate that the data are appropriate for factor analysis (Kaiser 1974). Varimax rotation (maximum variance orthogonal rotation method) was used to rotate the factor loading, select representative factors, and clarify the pollution source information of each primary component. Selection of the component number was referred to Cattel Scree criteria (1966). As cumulative variance contribution rate is more than 80%, the original data set can be replaced approximately (Chen et al. 2013; Li et al. 2020a, 2020b). Comrey and Lee (1992) elaborated that loadings in excess of 0.71 (50% overlapping variance) were considered excellent.

The assessment results of the SFWQII value reflected the level of each assessment factor in Yaojiang River, Changling River, Hutang River, and Linzhou River during the wet and dry seasons in 2018–2019 (Figure 2, Table 3). In Yaojiang River, the overall levels of TP and VP were optimal, with a slightly exceeded level of COD_Mn and petroleum, and significantly exceeded levels of DO and TN (Figure 2). In Changling River, the overall level of TP was also optimal, with slightly exceeded levels of COD_Mn and VP, and significantly exceeded levels of DO, petroleum, and TN (Figure 2). The overall levels of TP, petroleum, and VP in Hutang River were optimal, whereas the levels of DO and TN were significantly exceeded the standards (Figure 2). The assessment result of VP in Linzhou River was optimal while the levels of COD_Mn, DO, TP, TN, and petroleum severely exceeded the standards (Figure 2). All these results indicated that eutrophication pollution existed in Yuyao main inland rivers. In Yaojiang River, Changling River and Hutang River, nitrogen was the primary contributor to eutrophication pollution, while in Linzhou River both nitrogen and phosphorus were the main contributors.

During the monitoring period, the mean SFWQII value of TN of each inland river was the highest, indicating that the TN pollution was the most severe. The SFWQII values of TN of Yaojiang River, Changling River, Hutang River, and Linzhou River were in the ranges of 4.02–7.84, 4.02–6.83, 4.21–7.14 and 4.71–7.84, respectively, showing significant variation at each section (Table 3). The TN assessments of the upstream (section S1) were mostly significantly worse than that of the downstream (section S2-S4) in Yaojiang River. The TN assessments of the upstream (section S5) were slightly inferior to that of the downstream of Changling River (section S6), while the situation in Linzhou River was opposite. In Hutang River, the TN assessments of upstream (section S8) and downstream (section S7) were very close. Overall, the assessment results of TN index of each inland river in both the wet and dry seasons of 2018 were significantly better than those in 2019. As one of the important nutrient elements, the TP assessment results of each inland river were much better than those of TN. Except for Linzhou River, other rivers could meet the target water quality (Grade III). The SFWQII values of TP of Linzhou River, ranging from 3.20 to 6.23, were significantly higher than those of other rivers. When comparing the TP assessment between upstream and downstream of Yaojiang River and Linzhou River, the results were the same as that of the TN assessments. However, the opposite results were observed in Changling River. The average value of the TP index in the dry season was better than that in the wet season. The SFWQII value of another over-standard index, DO, varied considerably at different monitoring periods. The DO index of the inland rivers in 2019 exhibited the highest value in the wet season, and each river section was inferior to the target water quality level by two grades or more. In contrast, the DO index of inland rivers possessed the lowest value in the dry season of 2018, which met the target water quality. Compared with the surveys in 1990s, the levels of TN and DO were stable, but the average content of TP was reduced by one fold (Shen 2000; Wang 2001).

Although the SFWQII value of COD_Mn and petroleum varied significantly with monitoring periods, their assessment results were generally optimal. Both the COD_Mn and petroleum index of each river in the wet season of 2018 were worse than the target water quality level by one grade (Table 3). The assessment result of COD_Mn in the dry season of 2018 was the best, all of which met the target water quality. The assessment of petroleum in the dry season of 2019 was best, among which 90% of monitored sections met the target water quality. During the survey in 2019, all rivers were in line with the target water quality except the COD_Mn of Linzhou River, which was one grade lower than the target water quality. In the dry season of 2018, all others met the target water quality, except the ratio of petroleum index in Yaojiang River and Linzhou River are inferior to the target water quality was 50.0%. During the wet season of 2019, except for the Hutang River, of which all the sections met the target water quality, the ratios of petroleum indexes of Yaojiang River, Changling River and Linzhou River inferior to the target water quality were 50.0%, 100.0%, and 100.0%, respectively. The VP assessment of the SFWQII value was the best among all the determined indexes. Except that the Changling River was inferior to the target water quality by one grade during the wet season of 2018, all others met the target water quality. Comparison of DO, COD, petroleum and VP assessments of upstream with downstream of inland rivers showed inconsistent results among different monitoring periods.

The CWQII assessment of the main inland rivers in Yuyao City was conducted (Table 3). According to the classification standard of CWQII value, the water quality of the Yaojiang River was in grade III to IV. The water quality of Changling River and Hutang River was in Class II to IV, and the water quality of Linzhou River was in Class III to V (Table 4). According to the ratio of the water quality level that reached the target grade III, the water quality of Hutang River was the best while that of the Linzhou River was the worst.

From the perspective of the sampling section, the overall comprehensive water quality of the upstream of Yaojiang River (section S1) was inferior to that of the downstream (sections S2-S4). The comprehensive water quality of Xiachendu Village in Yaojiang River section S1 in 2018 was significantly inferior to its downstream (sections S2-S4). In contrast, there was no significant difference among different sampling sections of Yaojiang River in 2019 (Figure 3). Yaojiang River section S1 and section S4 are mainly urban residential areas surrounded by farmland. Their pollutants are majorly attributed to urban domestic sewage, followed by agricultural wastewater. In 2018, the second batch of ‘Zero Sewage Direct Drainage Area’ was launched in Yuyao City, including Lizhou Street located in S1 section and Hemudu Town located in S4 section, to achieve the full collection of sewage and complete coverage of the pipe network in the jurisdiction area. The comprehensive water quality assessment of these two sections during the wet and dry seasons of 2019 was significantly better than those of 2018, showing ‘Zero Sewage Direct Drainage Area’ infrastructure implemented by Yuyao government significantly improved the surface water quality (Ma et al. 2020). Section S2 around Pukou Village was mainly farmland, indicating that the pollution of this river section is primarily attributed to farming wastewater. In contrast, section S3 of Wantou Village is surrounded by aquaculture ponds, indicating that the pollution source of the section is primarily aquaculture wastewater. The comprehensive water quality of the Changling River downstream (section S6) during the wet season was better than that of the upstream (section S5), but the upstream water quality during the dry season was better or similar to the downstream water quality. In contrast to the Changling River, the comprehensive water quality of the Hutang River upstream (section S8) was better than that of the downstream (section S7) during the wet season, but the upstream water quality was worse than the downstream during the dry season. From 2018 to 2019, the comprehensive water quality of the Linzhou River upstream (section S10) was significantly better than that of the downstream (section S9) (Figure 3). According to the results of the field investigation, the Linzhou River upstream was mainly polluted by agricultural wastewater while the downstream was contaminated by urban domestic sewage and industrial wastewater.

In the two-year survey, the comprehensive water quality of each section in the dry season of 2018 was the best, all of which met the target of the water environment function zone. The CWQII value of each section in the dry season of 2019 was in the range of 3.0 to 4.2, and the ratio reaching grade III water quality standards accounted for 80% of the total, showing the second-best assessment result. The wet season of 2018 has the worst assessment result, ranging from 3.7 to 4.5, and the ratio meeting grade III water quality standard only accounts for 20%. Overall, the CWQII value in the dry season is better than that in the wet season. According to the Climate Bulletin of Yuyao City, the average annual precipitation in 2018 was about 16.9% more than the typical year. In particular, the precipitation in August was 313.0 mm, which was more than that of a normal year, whereas the precipitation in October was 51.6 mm, which was less than normal. The average annual precipitation in 2019 was the second most abundant compared with the same period in history, which is 367.5 mm more than that in 2018. Specifically, the precipitation in August is similar to the same period in 2018, while the precipitation in October is 120.5 mm more than the same period in 2018. Combined with the comprehensive water quality conditions in 2018 and 2019, it could be presumed that the surface runoff significantly impacts the pollution level. This presumption needs to be consolidated by the analyses of flow rate; however, flow rate data of each inland river are unavailable at present.

Factor analysis can quantitatively reflect the contribution of each pollution source to the overall pollution of river water bodies. Through the test, the KMO values in 2018 and 2019 were 0.652 and 0.714, respectively. The Bartlett spherical test probability p-value in both years were all less than 0.01, indicating that the data were appropriate for factor analysis.

According to the Cattel Scree criteria (1966), the first three principal components were selected for factor analysis in 2018 and 2019 (Figure 4). It was shown that cumulative variance contribution rates of 2018 and 2019 were 90.701% and 86.597%, respectively (Table 5). The result indicated that the selected three principal components contain most of the information in the original data set, which could fully reflect the characteristics of water quality.

According to the rotation load distribution of the first three principal components (Table 6), TN and TP in the first principal component (F1) in 2018 had a significant factor load, which was highly positively correlated with F1, reflecting the eutrophication of the water body. It has been reported that the TN and TP pollution could result from industrial and municipal wastewater, agricultural activities, and household wastes (Liu & Diamond 2005). In our investigation, it was found that domestic sewage in many rural areas was still discharged without regulation, except that domestic sewage in urban areas and several large towns could be effectively treated through sewage network management. In agricultural cultivation, a large amount of chemical fertilizer is used to increase soil fertility. The unabsorbed nitrogen and phosphorus in the fertilizer enter the groundwater body through surface runoff, which pollutes the river by TN and TP. In addition, since the leading agricultural industries in the surveyed area are poultry, pigs, and aquatic products, the impact of nitrogen and phosphorus contained in the farming wastewater on the eutrophication of water quality cannot be underestimated. In conclusion, the sources of F1 pollution were agricultural wastewater followed by urban domestic sewage. DO and COD_Mn had a more substantial factor load in the second principal component (F2), where COD_Mn was positively correlated with F2 but negatively correlated with DO (Figure 5). COD_Mn was an indicator of organic and reducing inorganic pollutants. The primary sources of COD_Mn were industrial sewage discharge and uncontrolled domestic sewage discharge caused by rapid urbanization. Yuyao City has a large number of dye houses and electroplating plants. Therefore, the direct discharge of industrial wastewater without treatment was one of the reasons for the higher COD_Mn factor load value in the water body. Thus, the source of F2 pollution could be defined as industrial pollution. The factor loading of the VP in the third principal component (F3) was the highest (Table 6), which further reflected industrial pollution might be caused by dye houses, paper mills, and so on.

In 2019, COD_Mn, petroleum, and TP had a significant factor loading in F1, and all exhibited a positive correlation with F1 (Figure 5), reflecting the status of organic pollution and eutrophication of water bodies. Therefore, F1 pollution was mainly attributed to industrial sewage discharge and agricultural sources. The factor loading of TN was the largest in F2 (Table 6). TN pollution generally originated from agricultural runoff, municipal sewage, and fertilizer plant wastewater. Because the survey sections S2, S3, S4, S5, S7, S8, and S10 are agricultural production areas while S1 and S9 are urban residential areas, F2 pollution is mainly caused by agricultural sources, followed by urban domestic sewage. The factor loading of VP in F3 was the highest with an industrial pollution source of dye houses and paper mills. In 2019, Ningbo Environmental Protection Bureau Yuyao Branch Bureau implemented ‘Blue Water Project’, which focused on the industries of metal surface treatment, non-ferrous metals and agricultural and sideline food processing. About 123 plants have completed the rectification. Hence, SFWQII values of VP in 2019 were significantly lower than that in 2018, reflecting a reduction of industrial wastewater pollution. Compared with the pollution sources in 2018, it was found that the discharge of urban domestic sewage and industrial wastewater significantly reduced, while that of agriculture wastewater increased due to the higher average annual precipitation in 2019.

The study on Yaojiang River Basin employing the SFWQII method showed that DO and TN of all rivers severely exceeded the standards, which was consistent with the previous studies. However, TP was remarkably improved in the two-year survey compared to the 1990s. The comprehensive water quality of each river during the wet season was inferior to the dry season, which was probably caused by the surface runoff carrying various pollutants. The surveyed watershed in 2018 was mainly polluted by agricultural wastewater followed by urban domestic sewage, while the watershed in 2019 was polluted by industrial sewage discharge and agricultural wastewater. Thus, with the combination of CWQII result of each section, this result could reflect the positive effect of the ‘Zero Sewage Direct Drainage Area’ infrastructure and the ‘Blue Water Project’ implemented by Yuyao city government.

This study was supported by the Special Scientific Research Funds for Central Non-profit Institutes, Chinese Academy of Fishery Sciences (2020TD14) and the Project of Environmental Monitoring and Fishery Resources Investigation of Main Inland Rivers in Yuyao City (1988-H-2017). We appreciate Hai Liu and Jian-Ping Wang from Ningbo Ocean and Fisheries Research Institute for providing advice on the project, as well as Yuan-Ge Chen, Wei Tian from East China Sea Fisheries Research Institute for assisting sample collection.

All relevant data are included in the paper or its Supplementary Information.