In this study, a customized WQI (Seoul water quality index, S-WQI) for urban rivers that can ultimately reflect their characteristics was developed by modifying and supplementing the existing Bascarόn WQI calculation method through linkage with statistical methods such as factor analysis. We used the water quality data generated monthly at 17 water quality monitoring networks (WQMNs) in Seoul for 18 years, from 2002 to 2019. Results of the research show that the monthly S-WQI showed an average 70 out of 100, ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II),’ with an R2 value of 0.8298. The annual S-WQI was found to be 39 (bad) to 97 (very good), with an average of 72 (good). Through this study, S-WQI, a customized WQI for urban rivers, was judged to be a reasonable index that can represent the characteristics of urban river water quality. This is because it is easy to apply and is a calculation method that uses relatively fewer water quality items than the WQI calculated in the past, and it is highly likely to be linked to the currently implemented water quality grade system. In addition, to extend the application of WQI to various water quality survey points, based on the calculation methodology performed to derive the indices in this study, such as modified S-WQI (MS-WQI), by adding new water quality items and changing some items, it is also possible to develop an advanced customized WQI for urban rivers considering watershed characteristics and measurement items.

  • We developed customized water quality index (S-WQI) suitable for water quality evaluation of urban rivers.

  • S-WQI was developed by modifying and supplementing the existing WQI calculation method through linkage with statistical methods.

  • S-WQI was likely to be linked to the currently implemented water quality grade system.

  • S-WQI can be extended and applied to water quality evaluation of various urban rivers.

Urban rivers are streams that run through the centers of cities and serve as passageways for river water, forming elements of urban landscapes and serving as ecological spaces. Since urban rivers flow into the main streams, efficient management of their water quality is of utmost importance (Kim et al. 2018). National and local governments are operating a water environment monitoring network to identify the current status of water quality in public waters, including urban rivers, and to obtain basic data for establishing environmental policies. For the National Water Quality Monitoring Networks (NWQMNs) program in Seoul, 17 major rivers have been designated and are being operated as water quality monitoring networks (WQMNs).

However, since the data on individual water quality items that have been accumulated for a long time have become vast, it takes a considerable amount of time for even those in charge of water-related organizations to comprehensively interpret them. Therefore, there is a limit to the comprehensive evaluation of the constantly fluctuating water quality of rivers by just analyzing the data for each item individually (Shweta et al. 2013). In particular, in the case of urban rivers, which have water quality characteristics different from those of natural water systems, there is active river use by ordinary citizens (leisure spaces, water-friendly spaces, and urban landscapes), and interest in the water environment is increasing. Therefore, it is important to provide easy-to-understand and comprehensive water quality information. Recently, there has been an increase in the number of research cases for easy understanding of river water quality and scientific evaluation of water quality through various water quality indices (WQIs) (Gupta et al. 2015; Dutta et al. 2018). WQI can be defined as an index expressed through numerical and grade classification that integrally reflects the individual contribution of water quality evaluation criteria. The current water quality status can be comprehensively identified, and trends in river water quality can be determined based on medium- and long-term water quality analysis, systematic comparison of water quality among water systems, and so on, by converting vast amounts of water quality data into a WQI and expressing them as simple grades (scores). Thus, policy effects can be determined, and predictions related to river water quality management can be easily made (Kim et al. 2018). Furthermore, WQI becomes a tool that makes it easy for non-expert citizens to understand complex water quality conditions. Currently, the most widely used WQIs include the weight arithmetic water quality index (WAWQI) (Brown et al. 1972; Chowdhury et al. 2012), National Sanitation Foundation water quality index (NSFWQI) (Brown et al. 1970; Misaghi et al. 2017; Lee et al. 2020), Canadian Council of Ministers of the Environment water quality index (CCMEWQI) (Khan et al. 2005; Lumb et al. 2006; Kankal et al. 2012), Oregon water quality index (OWQI) (Dunnette 1979; Dinius 1987), and Bascarόn water quality index (Bascarόn WQI) (Bascarόn 1979; Debels et al. 2005; Raphael et al. 2007). They are used in various ways depending on the research subject, water quality items, and watershed characteristics (Yogendra & Puttaiah 2008; Shweta et al. 2013). In Korea, many researchers have been actively developing WQIs, such as the Korea water quality index (K-WQI) (Choi 1996; Chung & Park 2005), which considers biological indicators, and the real-time water quality index (RTWQI) (Kal et al. 2017), which improves upon CCMEWQI to consider domestic situations. However, although most WQIs target rivers and lakes, research and application cases related to the development of a WQI for urban rivers, which are currently used by many citizens, are still insufficient. In addition, since important water quality items may differ by country, local government, and water quality monitoring networks, among others, there are some water quality items that are not measured in the field. Therefore, there is a limit to applying the previously used water quality index collectively.

In this study, as one of the advanced water quality management methods applicable to urban rivers, we developed the Seoul water quality index (S-WQI), which is based on a statistical analysis method; the possibility of a linkage was examined through a comparison and analysis of the environmental standards and river water quality grade systems applied in Korea. Here, the customized WQI for urban rivers is suitable for actual site conditions and situations, and was developed using the water quality data generated at the points of the National Water Environment Monitoring Network in Seoul. Through this, it is expected that it will be possible to effectively strengthen urban river water quality management from the medium- and long- term point of view, to enhance the reliability of urban river water quality conditions, and to generate information on urban river water quality in a form that is easy to understand even for ordinary citizens.

Water quality monitoring

The WQMN in Seoul consists of 17 points in 13 rivers, namely, Anyangcheon, Jungnangcheon (3 points), Cheonggyecheon (3 points), Tancheon, Godeokcheon, Seongnaecheon, Yangjaecheon, Uicheon, Seongbukcheon, Jeongneungcheon, Hongjecheon, Dorimcheon, and Mokgamcheon (Figure 1). Nineteen water quality items per month (pH, DO (dissolved oxygen), water temperature, BOD5 (biochemical oxygen demand, 5 days), COD (chemical oxygen demand), TOC (total organic carbon), SS (suspended solids), EC (electrical conductivity), TN (total nitrogen), DTN (dissolved total nitrogen), NH3-N (ammonia nitrogen), NO3-N (nitrate nitrogen), TP (total phosphorus), DTP (dissolved total phosphorus), PO4-P (phosphate phosphorus), phenols, chlorophyll, total coliforms, and fecal coliforms) and eight water quality items per quarter (cadmium, cyanide, lead, hexavalent chromium, arsenic, mercury, ABS (alkyl benzene sulfonate), and antimony) are being analyzed according to the Water Pollution Standard Method and Quality Standard for Drinking Water (National Institute of Environmental Research 2017a, 2017b). A total of 14 water quality items, which were water temperature, pH, DO, BOD, COD, SS, EC, TN, NH3-N, NO3-N, TP, PO4-P, total coliforms, fecal coliforms, were selected to calculate the WQI, taking into account phenols, which are always undetected, TOC, which has been measured since 2011, and overlaps with other water quality items, from the 19 items generated monthly at 17 WQMNs in Seoul for 18 years, from 2002 to 2019.

Figure 1

Study area for Seoul water quality index derivation for urban rivers.

Figure 1

Study area for Seoul water quality index derivation for urban rivers.

Close modal

Water quality index development

The Bascarόn WQI calculation method widely used in South American countries and Spain was applied in this study, using R 3.6.1 (data preprocessing and statistical analysis) and Python 3.7 (WQI calculation). A customized WQI for urban rivers was developed using the procedure shown in Figure 2. This is because the Bascarόn WQI has a relatively free selection of water quality items depending on the purpose, and since several experts used reference tables based on water quality standards in the past, the water quality grade has various characteristics. In particular, the Bascarόn WQI has the advantage of improving WQI in connection with advanced statistical methodologies such as variable reduction (Kannel et al. 2007; Kocer & Sevgili 2014).

Figure 2

Flow chart for calculating water quality index.

Figure 2

Flow chart for calculating water quality index.

Close modal
The Bascarόn WQI calculation formula and each water quality item weight (1–4) and sub-index (0–100) are shown in Equation (1) and Table 1, respectively. The Bascarόn WQI is classified into five levels from very good (I) to very bad, as shown in Table 2.
formula
(1)
where n is the number of water quality items to be calculated, and Ci is the standardized value of each water quality item measured as a discrete score divided into 10 units ranging from 0 to 100 (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100). Pi is the relative weight for each water quality item and has values of 1, 2, 3, and 4; a score of 4 means that the importance of the water quality item is high.
Table 1

Weights and sub-indices for deriving Bascarón and Seoul water quality indices

Water quality itemRelative weight (Pi)Normalization factor (Ci)
1009080706050403020100
Water temperature (°C) 12/16 22/15 24/14 26/12 28/10 30/5 32/0 36/− 2 40/− 4 45/− 6 45/ <−6 
pH 7–8 7–8.5 7–9 6.5–7 6–9.5 5–10 4–11 3–12 2–13 1–14 
DO (mg/L) ≥7.5 >7 >6.5 >6 >5 >4 >3.5 >3 >2 ≤1 <1 
SS (mg/L) <20 <40 <60 <80 <100 <120 <160 <240 <320 ≤400 >400 
EC (μS/cm) <750 <1,000 <1,250 <1,500 <2,000 <2,500 <3,000 <5,000 <8,000 ≤12,000 >12,000 
BOD5 (mg/L) <0.5 <2 <3 <4 <5 <6 <8 <10 <12 ≤15 >15 
COD (mg/L) <5 <10 <20 <30 <40 <50 <60 <80 <100 ≤150 <150 
TP (mg/L) <0.2 <1.6 <3.2 <6.4 <9.6 <16 <32 <64 <96 ≤160 >160 
PO4-P (mg/L) <0.025 <0.05 <0.1 <0.2 <0.3 <0.5 <0.75 <1 <1.5 ≤2 >2 
TN (mg/L) <0.8 <3.8 <7.5 <13 <18 <27 <48 <85 <149 ≤265 >265 
NH3-N (mg/L) <0.01 <0.05 <0.1 <0.2 <0.3 <0.4 <0.5 <0.75 <1 ≤1.25 >1.25 
NO3-N (mg/L) <0.5 <2 <4 <6 <8 <10 <15 <20 <50 ≤100 >100 
Total coliforms (CFU/100 mL) <50 <500 <1,000 <2,000 <3,000 <4,000 <5,000 <7,000 <10,000 ≤14,000 >14,000 
Fecal coliforms (CFU/100 mL) <5 <50 <100 <200 <300 <400 <500 <700 <1,000 ≤1,400 >1,400 
Water quality itemRelative weight (Pi)Normalization factor (Ci)
1009080706050403020100
Water temperature (°C) 12/16 22/15 24/14 26/12 28/10 30/5 32/0 36/− 2 40/− 4 45/− 6 45/ <−6 
pH 7–8 7–8.5 7–9 6.5–7 6–9.5 5–10 4–11 3–12 2–13 1–14 
DO (mg/L) ≥7.5 >7 >6.5 >6 >5 >4 >3.5 >3 >2 ≤1 <1 
SS (mg/L) <20 <40 <60 <80 <100 <120 <160 <240 <320 ≤400 >400 
EC (μS/cm) <750 <1,000 <1,250 <1,500 <2,000 <2,500 <3,000 <5,000 <8,000 ≤12,000 >12,000 
BOD5 (mg/L) <0.5 <2 <3 <4 <5 <6 <8 <10 <12 ≤15 >15 
COD (mg/L) <5 <10 <20 <30 <40 <50 <60 <80 <100 ≤150 <150 
TP (mg/L) <0.2 <1.6 <3.2 <6.4 <9.6 <16 <32 <64 <96 ≤160 >160 
PO4-P (mg/L) <0.025 <0.05 <0.1 <0.2 <0.3 <0.5 <0.75 <1 <1.5 ≤2 >2 
TN (mg/L) <0.8 <3.8 <7.5 <13 <18 <27 <48 <85 <149 ≤265 >265 
NH3-N (mg/L) <0.01 <0.05 <0.1 <0.2 <0.3 <0.4 <0.5 <0.75 <1 ≤1.25 >1.25 
NO3-N (mg/L) <0.5 <2 <4 <6 <8 <10 <15 <20 <50 ≤100 >100 
Total coliforms (CFU/100 mL) <50 <500 <1,000 <2,000 <3,000 <4,000 <5,000 <7,000 <10,000 ≤14,000 >14,000 
Fecal coliforms (CFU/100 mL) <5 <50 <100 <200 <300 <400 <500 <700 <1,000 ≤1,400 >1,400 
Table 2

Water quality grades

GradeVery goodGoodMediumBadVery bad
WQI 91–100 71–90 51–70 26–50 0–25 
GradeVery goodGoodMediumBadVery bad
WQI 91–100 71–90 51–70 26–50 0–25 
However, in the case of the Bascarόn WQI, if many water quality items are necessary for calculating WQI, offset and interference effects between water quality items may occur, resulting in poor accuracy. Therefore, it was hypothesized that a customized WQI can be developed by advancing the methodology, which includes reducing variables, using the existing Bascarόn WQI (Liu et al. 2003; Shrestha & Kazama 2007; Sun et al. 2016; Wu et al. 2018). In this study, a customized WQI for urban rivers that can ultimately reflect their characteristics was developed by modifying and supplementing the existing Bascarόn WQI calculation method through linkage with statistical methods such as factor analysis (Kassambara 2017). This analysis was carried out in five steps: data preprocessing, factor analysis suitability test, determination of number of factors, factor extraction and rotation and result interpretation, and final essential water quality monitoring item selection. Firstly, in data preprocessing, natural logarithms were processed for water quality items other than pH and water temperature, which show a normal distribution, in order to normalize the data. In addition, the Kaiser–Meyer–Olkin (KMO) test and Bartlett's test were performed to determine whether the preprocessed data were suitable for factor analysis, and the number of factors was determined using Kaiser's rule. The maximum likelihood method was used for factor extraction, and the Varimax method among orthogonal rotations was used for factor rotation. Finally, the importance of the water quality items selected through factor analysis was equal, because we selected items with a high correlation between the potential factors representing the representative characteristics of water quality and the water quality items included in the factors. Therefore, the weight of each water quality item was set equal to 1. The calculation formula for the customized WQI (S-WQI) for urban rivers is shown in Equation (2); the sub-indices and water quality grade system shown in Tables 1 and 2 were applied as for the Bascarόn WQI.
formula
(2)

Water quality characteristics of water quality monitoring network in Seoul

Tables 3 and 4 show the average water quality concentrations for major water quality items such as BOD, TP, and fecal coliforms in 17 WQMNs in Seoul for 18 years, from 2002 to 2019.

Table 3

Average water quality results for 17 water quality monitoring networks in Seoul

RiverWater temperature (°C)pHEC (μS/cm)DO (mg/L)BOD (mg/L)COD (mg/L)SS (mg/L)
Godeokcheon 13.2 8.0 760 11.5 2.5 4.4 7.6 
Dorimcheon 14.4 8.0 378 10.8 3.4 5.7 6.1 
Mokgamcheon 15.7 7.8 557 9.6 8.0 9.3 16.1 
Seongnaecheon 15.2 7.8 276 8.2 3.3 5.0 9.6 
Seongbukcheon 16.5 7.9 349 12.6 1.4 2.7 3.7 
Anyangcheon 5 16.9 7.5 581 7.5 6.6 9.0 12.1 
Yangjaecheon 15.9 7.8 458 9.7 3.1 5.4 10.8 
Uicheon 17.4 8.2 404 12.4 2.0 4.0 6.5 
Jeongneungcheon 15.3 8.0 483 12.5 1.5 2.6 3.5 
Jungnangcheon 2 17.3 7.8 465 10.6 5.2 7.1 10.5 
Jungnangcheon 3 16.1 7.8 463 10.5 4.2 6.7 10.3 
Jungnangcheon 4 18.3 7.5 539 7.8 9.3 10.0 10.5 
Cheonggyecheon 1 14.5 7.9 192 11.3 0.8 2.2 2.2 
Cheonggyecheon 2 15.4 8.0 200 11.8 0.9 2.4 2.3 
Cheonggyecheon 3 15.6 8.1 237 11.4 1.5 2.9 4.1 
Tancheon 5 17.9 7.4 539 7.9 11.6 9.7 12.5 
Hongjecheon 15.8 8.0 510 12.0 3.1 4.3 10.4 
Average 16.0 7.9 435 10.5 4.0 5.5 8.2 
Standard deviation 8.9 0.5 213 3.2 5.3 4.0 14.5 
95% confidence interval (n) 16.0 ± 0.30 (3,406) 7.9 ± 0.02 (3,406) 435 ± 7.15 (3,401) 10.5 ± 0.11 (3,406) 4.0 ± 0.18 (3,406) 5.5 ± 0.13 (3,406) 8.2 ± 0.49 (3,406) 
RiverWater temperature (°C)pHEC (μS/cm)DO (mg/L)BOD (mg/L)COD (mg/L)SS (mg/L)
Godeokcheon 13.2 8.0 760 11.5 2.5 4.4 7.6 
Dorimcheon 14.4 8.0 378 10.8 3.4 5.7 6.1 
Mokgamcheon 15.7 7.8 557 9.6 8.0 9.3 16.1 
Seongnaecheon 15.2 7.8 276 8.2 3.3 5.0 9.6 
Seongbukcheon 16.5 7.9 349 12.6 1.4 2.7 3.7 
Anyangcheon 5 16.9 7.5 581 7.5 6.6 9.0 12.1 
Yangjaecheon 15.9 7.8 458 9.7 3.1 5.4 10.8 
Uicheon 17.4 8.2 404 12.4 2.0 4.0 6.5 
Jeongneungcheon 15.3 8.0 483 12.5 1.5 2.6 3.5 
Jungnangcheon 2 17.3 7.8 465 10.6 5.2 7.1 10.5 
Jungnangcheon 3 16.1 7.8 463 10.5 4.2 6.7 10.3 
Jungnangcheon 4 18.3 7.5 539 7.8 9.3 10.0 10.5 
Cheonggyecheon 1 14.5 7.9 192 11.3 0.8 2.2 2.2 
Cheonggyecheon 2 15.4 8.0 200 11.8 0.9 2.4 2.3 
Cheonggyecheon 3 15.6 8.1 237 11.4 1.5 2.9 4.1 
Tancheon 5 17.9 7.4 539 7.9 11.6 9.7 12.5 
Hongjecheon 15.8 8.0 510 12.0 3.1 4.3 10.4 
Average 16.0 7.9 435 10.5 4.0 5.5 8.2 
Standard deviation 8.9 0.5 213 3.2 5.3 4.0 14.5 
95% confidence interval (n) 16.0 ± 0.30 (3,406) 7.9 ± 0.02 (3,406) 435 ± 7.15 (3,401) 10.5 ± 0.11 (3,406) 4.0 ± 0.18 (3,406) 5.5 ± 0.13 (3,406) 8.2 ± 0.49 (3,406) 
Table 4

Average water quality results for 17 water quality monitoring networks in Seoul (continued)

RiverTN (mg/L)NH3-N (mg/L)NO3-N (mg/L)TP (mg/L)PO4-P (mg/L)F.Coli* (CFU/100 mL)T.Coli** (CFU/100 mL)
Godeokcheon 7.396 0.616 4.657 0.146 0.075 8,520 54,220 
Dorimcheon 6.082 0.446 3.295 0.398 0.245 8,192 1,100,502 
Mokgamcheon 10.235 1.094 5.432 0.539 0.334 6,454 1,323,703 
Seongnaecheon 3.276 0.635 1.528 0.160 0.057 6,961 210,883 
Seongbukcheon 3.443 0.128 2.823 0.033 0.010 1,216 20,727 
Anyangcheon 5 13.754 5.655 4.140 0.538 0.364 15,641 203,127 
Yangjaecheon 7.390 0.621 4.587 0.346 0.279 6,898 84,818 
Uicheon 5.158 0.239 4.113 0.073 0.024 2,841 53,452 
Jeongneungcheon 6.084 0.173 4.351 0.106 0.053 5,083 37,549 
Jungnangcheon 2 9.969 3.220 4.348 0.578 0.357 38,680 550,846 
Jungnangcheon 3 8.546 2.588 4.060 0.391 0.210 15,464 153,304 
Jungnangcheon 4 15.374 5.984 5.335 1.094 0.697 173,539 3,231,103 
Cheonggyecheon 1 2.398 0.113 1.915 0.018 0.003 115 759 
Cheonggyecheon 2 2.410 0.088 1.904 0.020 0.003 721 5,847 
Cheonggyecheon 3 2.696 0.174 2.003 0.044 0.014 3,079 32,187 
Tancheon 5 12.385 5.590 4.196 0.689 0.440 11,352 144,959 
Hongjecheon 7.240 1.410 4.929 0.111 0.042 2,420 79,072 
Average 7.285 1.692 3.742 0.311 0.189 18,069 428,650 
Standard deviation 5.439 3.582 2.253 0.458 0.377 220,405 9,385,024 
95% confidence interval (n) 7.285 ± 0.18 (3,406) 1.692 ± 0.13 (2,911) 3.742 ± 0.08 (2,911) 0.311 ± 0.02 (3,406) 0.189 ± 0.01 (3,109) 18.069 ± 8,009 (2,909) 428,650 ± 320,598 (3,292) 
RiverTN (mg/L)NH3-N (mg/L)NO3-N (mg/L)TP (mg/L)PO4-P (mg/L)F.Coli* (CFU/100 mL)T.Coli** (CFU/100 mL)
Godeokcheon 7.396 0.616 4.657 0.146 0.075 8,520 54,220 
Dorimcheon 6.082 0.446 3.295 0.398 0.245 8,192 1,100,502 
Mokgamcheon 10.235 1.094 5.432 0.539 0.334 6,454 1,323,703 
Seongnaecheon 3.276 0.635 1.528 0.160 0.057 6,961 210,883 
Seongbukcheon 3.443 0.128 2.823 0.033 0.010 1,216 20,727 
Anyangcheon 5 13.754 5.655 4.140 0.538 0.364 15,641 203,127 
Yangjaecheon 7.390 0.621 4.587 0.346 0.279 6,898 84,818 
Uicheon 5.158 0.239 4.113 0.073 0.024 2,841 53,452 
Jeongneungcheon 6.084 0.173 4.351 0.106 0.053 5,083 37,549 
Jungnangcheon 2 9.969 3.220 4.348 0.578 0.357 38,680 550,846 
Jungnangcheon 3 8.546 2.588 4.060 0.391 0.210 15,464 153,304 
Jungnangcheon 4 15.374 5.984 5.335 1.094 0.697 173,539 3,231,103 
Cheonggyecheon 1 2.398 0.113 1.915 0.018 0.003 115 759 
Cheonggyecheon 2 2.410 0.088 1.904 0.020 0.003 721 5,847 
Cheonggyecheon 3 2.696 0.174 2.003 0.044 0.014 3,079 32,187 
Tancheon 5 12.385 5.590 4.196 0.689 0.440 11,352 144,959 
Hongjecheon 7.240 1.410 4.929 0.111 0.042 2,420 79,072 
Average 7.285 1.692 3.742 0.311 0.189 18,069 428,650 
Standard deviation 5.439 3.582 2.253 0.458 0.377 220,405 9,385,024 
95% confidence interval (n) 7.285 ± 0.18 (3,406) 1.692 ± 0.13 (2,911) 3.742 ± 0.08 (2,911) 0.311 ± 0.02 (3,406) 0.189 ± 0.01 (3,109) 18.069 ± 8,009 (2,909) 428,650 ± 320,598 (3,292) 

*Fecal coliforms.

**Total coliforms.

The average water quality items of the 17 major WQMNs in Seoul were 10.5 mg/L DO (7.5–12.6), 4.0 mg/L BOD (0.8–11.6), and 5.5 mg/L COD (2.2–10.0), and the overall grade for contamination by organic matter was good. The water quality grades were average Ia grade DO (very good), average III grade BOD (normal), and average III grade COD (normal), in terms of the living environment standard of the river for water quality and aquatic ecosystems in the Enforcement Decree of the Framework Act on the Environmental Policy (Ministry of Environment 2020). TP and PO4-P, which are representative nutrient indicators, were 0.311 mg/L (0.018–1.094) and 0.189 mg/L (0.003–0.697), respectively, and fecal coliforms, a microbial contamination indicator, was 18,069 CFU/100 mL (115–173,539). With the exception of some WQMNs such as Jungnangcheon 4, Tancheon 5, and Anyangcheon 5, it was found that the water quality was maintained in good overall condition. The Jungnangcheon 4, Tancheon 5, and Anyangcheon 5 water quality monitoring networks have high population density in the watershed, and a large amount of effluent from the J sewage treatment plant flows into the corresponding river; hence, the water pollution level is relatively high.

But as indicated by the change in concentration of major water quality items shown in Figure 3, overall, from 2014 onwards, the water quality gradually improved at monitoring networks where the quality was previously poor (Anyangcheon 5, Jungnangcheon 4, and Tancheon 5) and in the remaining 14 WQMNs (other WQMNs). Overall, the reason for the improvement in the average water quality in 17 WQMNs in Seoul is that the causes (e.g. river bed structure, river maintenance work, nearby facilities) of the deterioration of water quality in the medium- and long-term were found. In addition, the improvement is the result of establishing a watershed-oriented water environment system for each subregion, such as the improvement of advanced treatment facilities to reduce the pollution load and the reduction of non-point pollution sources around the river, along with the sewage pipe maintenance project.

Figure 3

Historical changes in water quality items of three national water quality monitoring networks (NWQMNs) and other WQMNs in Seoul.

Figure 3

Historical changes in water quality items of three national water quality monitoring networks (NWQMNs) and other WQMNs in Seoul.

Close modal

Water quality index development

A customized WQI (S-WQI) optimized for urban rivers was developed by combining the Bascarόn WQI calculation method with factor analysis for 17 WQMNs in Seoul. To identify the changes in water quality for a specific period and in the medium and long term, the WQI is divided by month (the WQIs calculated based on January and July are presented as an example) and year. Thus, the trend in water quality fluctuations at a large number of WQMNs could be easily identified.

Bascarόn WQI calculation results

The Bascarόn WQIs calculated on a monthly basis to identify the changes in water quality during a specific period (January to July) are shown in Figure 4.

Figure 4

Changes in Bascarón water quality index every January (top) and July (bottom).

Figure 4

Changes in Bascarón water quality index every January (top) and July (bottom).

Close modal

The monthly WQI for January was 48 (bad) to 94 (very good), with an average of 72 (good), that for July was 45 (bad) to 97 (very good), with an average of 71 (good). Overall, the monthly WQIs for January and July showed similar ranges. Moreover, as shown in Figure 4, the water quality grades of 17 WQMNs calculated using the Bascarόn WQI were generally ‘good (II)’ to ‘normal (III)’. For January, the average water quality concentrations for the water quality items used to calculate the monthly Bascarόn WQI were 2.6 mg/L TOC (0.9–5.4), 12.9 mg/L DO (9.2–14.8), 0.138 mg/L TP (0.010–0.797), and 3,850 CFU/100 mL fecal coliforms (11–33,614), and for July, they were 2.7 mg/L TOC (1.4–4.4), 8.8 mg/L DO (5.5–11.6), 0.125 mg/L TP (0.011–0.489), and 11,399 CFU/100 mL fecal coliforms (24–108,314).

Based on the distribution of water quality grades on a monthly basis, out of a total of 201 in January, 31 (14.8%) ‘very good (I)’, 75 (35.7%) ‘good (II)’, 100 (47.6%) ‘normal (III)’, and 4 (1.9%) ‘bad (IV)’ grades were obtained, and no points corresponded to a ‘very bad (V)’ grade. In July, there was no ‘very bad (V)’ grade, and out of a total of 231, there were 14 (6.1%) ‘very good (I)’, 123 (53.2%) ‘good (II)’, 84 (36.4%) ‘normal (III)’, and 10 (4.3%) ‘bad (IV)’ grades. Overall, among the 17 WQMNs, Cheonggyecheon 1 (January: 91; July: 93), Cheonggyecheon 2 (January: 91; July: 82), and Cheonggyecheon 3 (January: 89; July: 79) were all in good condition with a grade of ‘good (II)’ or higher. On the other hand, Jungnangcheon 4 (January: 54; July: 56), Tancheon 5 (January: 55; July: 59), and Anyangcheon 5 (January: 59; July: 61) were mainly ‘normal (III)’.

Next, the Bascarόn WQI was calculated on a yearly basis to understand the water quality trend in the medium- and long- term, and the results are shown in Figure 5.

Figure 5

Changes in the middle- and long-term Bascarón water quality index (by year).

Figure 5

Changes in the middle- and long-term Bascarón water quality index (by year).

Close modal

The annual WQI was 47 (bad) to 94 (very good), with an average of 72 (good). The average water quality concentrations for the water quality items used to calculate the annual Bascarόn WQI were 2.7 mg/L TOC (1.3–4.8), 10.5 mg/L DO (7.5–13.2), 0.135 mg/L TP (0.011–0.637), and 7,257 CFU/100 mL fecal coliforms (142–55,461). In the case of water quality grades distribution, out of a total of 256, 15 (5.9%) ‘very good (I)’, 122 (47.7%) ‘good (II)’, 116 (45.3%) ‘normal (III)’, and 3 (1.2%) ‘bad (IV)’ grades were obtained, and no points corresponded to a ‘very bad (V)’ grade.

In accordance with the monthly calculation results, overall, Cheonggyecheon 1 (93), Cheonggyecheon 2 (87), and Seongbukcheon (84) were all in good condition with grades of ‘very good (I)’ and ‘good (II)’ at all times. On the other hand, Jungnangcheon 4 (55), Tancheon 5 (58), and Anyangcheon 5 (59) mainly showed ‘normal (III)’ and ‘bad (IV)’ grades. The effluent flowing into the corresponding rivers from nearby sewage treatment plants is believed to have had an impact on the grades. Therefore, it is important to manage effluent water quality from a medium- and long-term point of view and continuously manage river water quality according to the effects of various pollutants. This will have an effect on the inflow of effluent from nearby sewage treatment plants.

Furthermore, as shown in Figure 6, the possibility of a linkage between the river water quality grade system applied in Korea and the monthly Bascarόn WQI was reviewed.

Figure 6

Comparison of Bascarón water quality index and Living Environment Standard.

Figure 6

Comparison of Bascarón water quality index and Living Environment Standard.

Close modal

In the Wilcoxon–Mann–Whitney test conducted during the review, the significance probability p value was 0.4941, which was higher than the significance level of 5%; hence, there was no clear difference between the monthly Bascarόn WQI and the average water quality grade according to the environmental standards. The monthly Bascarόn WQI showed an average of 74 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’ with an R2 value of 0.7971. Therefore, overall, it was confirmed that the monthly Bascarόn WQI matched well with the average water quality grading system according to the environmental standards. Therefore, it is expected that it will be possible to intuitively and easily understand the water quality of urban rivers through water quality evaluation using the Bascarόn WQI rather than individual water quality items, with water quality grades ranging from ‘very good (Ia)’ to ‘very bad (VI)’.

Calculation of customized water quality index for urban rivers

Factor analysis was performed on 14 water quality items used in the Bascarόn WQI calculation, and thus, essential water quality monitoring items were selected and S-WQI was developed. Furthermore, the applicability of S-WQI to actual sites was reviewed by comparing the calculation results of the S-WQI and Bascarόn WQI. In factor analysis, factor 1 showed a strong correlation with BOD5, COD, and NH3-N, which is a potential factor involving organic matter. Factor 2 showed a strong correlation with total coliforms and fecal coliforms and was judged to be a potential factor involving microorganisms in the water system. Factor 3 showed a correlation between the properties provided by nitrogen compounds, such as TN and NO3-N. Factor 4 was judged to be a potential factor showing a correlation between phosphorus compounds, such as TP and PO4-P (Table 5). Thereafter, the essential water quality monitoring items were selected as water quality items representing the predetermined potential factors while avoiding multicollinearity. This is because if all water quality items exceeding the load capacity of a certain factor for each potential factor are reflected in the WQI calculation when selecting water quality items, the multicollinearity between the water quality items belonging to the same potential factor among the selected water quality items may cause negative effects such as interference with calculating a fair WQI score. Finally, as shown in Table 6, COD for Factor 1, fecal coliforms for Factor 2, TN for Factor 3, and PO4-P for Factor 4 were selected as the water quality items representing each factor. These four items were determined to be water quality items that can represent the water quality characteristics of urban rivers, and S-WQI was derived from them.

Table 5

Factor analysis results

ItemFactor 1Factor 2Factor 3Factor 4
pH −0.166 −0.071 −0.121 −0.099 
DO (mg/L) −0.305 −0.320 0.270 −0.215 
BOD5 (mg/L) 0.792 0.209 0.164 0.159 
COD (mg/L) 0.847 0.246 0.110 0.226 
SS (mg/L) 0.659 0.218 −0.086 0.100 
TN (mg/L) 0.605 0.153 0.653 0.299 
TP (mg/L) 0.657 0.345 0.103 0.573 
Water temperature (°C) 0.020 0.275 −0.373 0.136 
EC (μS/cm) 0.394 0.182 0.457 0.231 
Total coliforms (CFU/100 mL) 0.365 0.773 0.144 0.199 
NH3-N (mg/L) 0.718 0.172 0.248 0.196 
NO3-N (mg/L) 0.055 0.146 0.605 0.070 
PO4-P (mg/L) 0.528 0.363 0.131 0.730 
Fecal coliforms (CFU/100 mL) 0.280 0.896 0.165 0.146 
ItemFactor 1Factor 2Factor 3Factor 4
pH −0.166 −0.071 −0.121 −0.099 
DO (mg/L) −0.305 −0.320 0.270 −0.215 
BOD5 (mg/L) 0.792 0.209 0.164 0.159 
COD (mg/L) 0.847 0.246 0.110 0.226 
SS (mg/L) 0.659 0.218 −0.086 0.100 
TN (mg/L) 0.605 0.153 0.653 0.299 
TP (mg/L) 0.657 0.345 0.103 0.573 
Water temperature (°C) 0.020 0.275 −0.373 0.136 
EC (μS/cm) 0.394 0.182 0.457 0.231 
Total coliforms (CFU/100 mL) 0.365 0.773 0.144 0.199 
NH3-N (mg/L) 0.718 0.172 0.248 0.196 
NO3-N (mg/L) 0.055 0.146 0.605 0.070 
PO4-P (mg/L) 0.528 0.363 0.131 0.730 
Fecal coliforms (CFU/100 mL) 0.280 0.896 0.165 0.146 

Values in bold indicate strong correlation.

Table 6

Correlation coefficients of major water quality items for potential factors

CODBOD5NH3-NF.Coli*T.Coli**TNNO3-NPO4-PTP
COD  0.80 0.73 0.50 0.57 0.70 0.24 0.74 0.82 
BOD5   0.72 0.45 0.52 0.66 0.18 0.64 0.73 
NH3-N    0.43 0.47 0.70 0.16 0.63 0.68 
F.Coli    0.83 0.44 0.24 0.60 0.60 
T.Coli**      0.49 0.24 0.63 0.65 
TN       0.61 0.66 0.68 
NO3-N        0.26 0.23 
PO4-P         0.93 
TP          
CODBOD5NH3-NF.Coli*T.Coli**TNNO3-NPO4-PTP
COD  0.80 0.73 0.50 0.57 0.70 0.24 0.74 0.82 
BOD5   0.72 0.45 0.52 0.66 0.18 0.64 0.73 
NH3-N    0.43 0.47 0.70 0.16 0.63 0.68 
F.Coli    0.83 0.44 0.24 0.60 0.60 
T.Coli**      0.49 0.24 0.63 0.65 
TN       0.61 0.66 0.68 
NO3-N        0.26 0.23 
PO4-P         0.93 
TP          

*Fecal coliforms.

**Total coliforms.

Values in bold indicate the water quality items that best represent each factor.

Table 7

Main results for Bascarόn water quality index (WQI), Seoul WQI (S-WQI), and modified S-WQI (MS-WQI)

WQI (item number)ItemMean score (by year)R2*
Bascarόn WQI (14) Water temperature, pH, EC, DO, BOD5, COD, SS, TN, NH3-N, NO3-N, TP, PO4-P, Fecal coliforms, Total coliforms 72 (47–95) 0.7971 
S-WQI (4) COD, Fecal coliforms, TN, PO4-P 70 (39–97) 0.8298 
MS-WQI (4) TOC, Fecal coliforms, TN, TP 75 (60–96) 0.7647 
WQI (item number)ItemMean score (by year)R2*
Bascarόn WQI (14) Water temperature, pH, EC, DO, BOD5, COD, SS, TN, NH3-N, NO3-N, TP, PO4-P, Fecal coliforms, Total coliforms 72 (47–95) 0.7971 
S-WQI (4) COD, Fecal coliforms, TN, PO4-P 70 (39–97) 0.8298 
MS-WQI (4) TOC, Fecal coliforms, TN, TP 75 (60–96) 0.7647 

*Comparison of WQI and Living Environment Standards.

Similar to the Bascarόn WQI, the S-WQI and water quality grade were calculated on a monthly basis to understand the change in water quality during a specific period (January to July). In addition, the WQI and water quality grade were visually expressed as shown in Figure 7 for easy understanding.

Figure 7

Changes in Seoul water quality index (S-WQI) every January (top) and July (bottom).

Figure 7

Changes in Seoul water quality index (S-WQI) every January (top) and July (bottom).

Close modal

Based on the S-WQI, the monthly WQI for January was 35 (bad) to 97 (very good), with an average of 71 (good), and that for July was 37 (bad) to 97 (very good), with an average of 69 (good). Overall, the Bascarόn WQI water quality grade and monthly WQI for January and July showed similar ranges; most of the WQIs were ‘good (II)’ and ‘normal (III)’ grades. For January, the average water quality concentrations for the water quality items used to calculate the monthly S-WQI were 4.6 mg/L COD (1.9–10.2), 0.138 mg/L TP (0.010–0.797), and 3,850 CFU/100 mL fecal coliforms (11–33,614), and for July, they were 4.7 mg/L COD (2.0–7.9), 0.125 mg/L TP (0.011–0.489), and 11,399 CFU/100 mL fecal coliforms (24–108,314).

Based on the distribution of water quality grades on a monthly basis, out of a total of 211 in January, 53 (25.1%) ‘very good (I)’, 58 (27.5%) ‘good (II)’, 67 (31.8%) ‘normal (III)’, and 33 (15.6%) ‘bad (IV)’ grades were obtained, and no points corresponded to a ‘very bad (V)’ grade. In July, there were no ‘very bad (V)’ grades out of the 232, but there were 23 (9.9%) ‘very good (I)’, 77 (33.2%) ‘good (II)’, 124 (53.4%) ‘normal (III)’, and 8 (3.4%) ‘bad (IV)’ grades. Overall, among the 17 WQMNs, Cheonggyecheon 1 (January: 95; July: 94), Cheonggyecheon 2 (January: 93; July: 80), and Cheonggyecheon 3 (January: 93; July: 79) were all in good condition with a grade of ‘good (II)’ or higher. On the other hand, Jungnangcheon 4 (January: 47; July: 53), Tancheon 5 (January: 49; July: 58), and Anyangcheon 5 (January: 56; July: 60) were mainly ‘normal (III)’ or ‘bad (IV)’. However, as shown in Figure 7, it was recently confirmed that the overall water quality is stably improving, with the water quality grade being constant or upgraded in most points.

Based on the distribution of the S-WQI water quality grade compared to the Bascarόn WQI, the ratio of ‘very good (I)’ grades increased from 14.8% to 25.1% in January and from 6.1% to 9.9% in July. Moreover, ‘bad (IV)’ grades slightly decreased, from 4.3% to 3.4%, on a monthly basis in July, but increased from 1.9% to 15.6% on a monthly basis in January. S-WQI showed a tendency to give a higher score for good water quality and a lower score for bad water quality, as indicated in Figure 8. The red dot is a point where each score is 77.56, and when S-WQI is lower than 77.56, a score lower than the Bascarόn WQI is obtained, and conversely, when S-WQI is higher than 77.56, a score higher than the Bascarόn WQI is obtained.

Figure 8

Comparison of Seoul water quality index (S-WQI) and Bascarón WQI. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/ws.2022.238.

Figure 8

Comparison of Seoul water quality index (S-WQI) and Bascarón WQI. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/ws.2022.238.

Close modal

In particular, when 14 water quality items of the existing Bascarόn WQI are used, their effects on water quality conflict with each other due to the eclipsing and interference effects between the items, and similar items are reflected in the score redundantly. Hence, the ability to discriminate in score calculation may be lowered. The eclipsing and interference effects could be reduced by selecting essential items (four items) that reflect the overall characteristics of water quality, such as S-WQI. This implies that, unlike the Bascarόn WQI, which provides the WQI score in a complex manner using 14 water quality items, the S-WQI, which highly correlated with the Bascarόn WQI (R2 = 0.8232), could be used to calculate the WQI score by selecting essential water quality monitoring items that can reflect the characteristics of urban rivers, using statistical analysis techniques, and could indicate urban river water quality more judiciously.

S-WQI was calculated on a yearly basis to understand the water quality trend in the medium and long term, and the results are shown in Figure 9.

Figure 9

Changes in middle- and long-term Seoul water quality index (S-WQI) by year.

Figure 9

Changes in middle- and long-term Seoul water quality index (S-WQI) by year.

Close modal

Based on annual S-WQI calculation from 2002 to 2019, the annual WQI was found to be 39 (bad) to 97 (very good), with an average of 70 (good). As such, it was confirmed that the S-WQI had a wide range from 39 to 97 due to the difference in water quality concentration of major water quality items at each monitoring network point, such as Cheonggyecheon 1, Cheonggyecheon 2 and Cheonggyecheon 3, which maintain safe and clean water quality, or Jungnangcheon 4, which has relatively poor water quality as it is affected by effluent from nearby sewage treatment plants. The average water quality concentrations for the major water quality items effectively used in calculating the annual S-WQI were 4.6 mg/L COD (2.1–8.4), 0.135 mg/L TP (0.011–0.637), and 7,257 CFU/100 mL fecal coliforms (142–55,461), which were similar to those for the annual Bascarόn WQI.

The water quality grade distribution, overall, was in the range of ‘good (II)’ to ‘normal (III)’, and the WQI continued to increase. Out of a total of 256, 21 (8.2%) ‘very good (I)’, 110 (43.0%) ‘good (II)’, 97 (37.9%) ‘normal (III)’, and 28 (10.9%) ‘bad (IV)’ grades were obtained, and no points corresponded to a ‘very bad (V)’ grade. Similarly, compared with the Bascarόn WQI, the ‘very good (I)’ ratio slightly increased, from 5.9% to 8.2%, but ‘bad (IV)’ grades increased nearly tenfold, from 1.2% to 10.9%. In the case of bad water quality, the S-WQI was confirmed to show a slightly lower score. Overall, in accordance with the monthly calculation results, Cheonggyecheon 1 (95), Cheonggyecheon 2 (87), and Cheonggyecheon 3 (84) showed good water quality, maintaining ‘very good (I)’ and ‘good (II)’ grades at all times. On the other hand, Jungnangcheon 4 (48), Tancheon 5 (55), and Anyangcheon 5 (57) mainly showed ‘normal (III)’ and ‘bad (IV)’ grades. In addition, the possibility of a linkage between the river water quality grade system applied in Korea and the monthly S-WQI was reviewed, and the results are shown in Figure 10.

Figure 10

Comparison of Seoul water quality index (S-WQI) and Living Environment Standard.

Figure 10

Comparison of Seoul water quality index (S-WQI) and Living Environment Standard.

Close modal

The review indicated that the monthly S-WQI averaged 70 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’, similar to that based on the Bascarόn WQI, as shown in Figure 10; the correlation was confirmed to be very high, with an R2 value of 0.8298. Moreover, in the Wilcoxon–Mann–Whitney test according to non-normal distribution, the significance probability p value was 0.1712, which was higher than the significance level of 5%; hence, it was judged that there was no significant difference between S-WQI and the average water quality grade according to the environmental standards. It was concluded that S-WQI matches well with the currently implemented water quality grade systems such as the Bascarόn WQI (R2 = 0.7971).

In addition, a modified S-WQI (MS-WQI) was developed by changing some of the water quality items required for S-WQI calculation in order to expand and apply S-WQI to various water quality monitoring networks. COD and PO4-P were changed to TOC and TP, respectively. COD was changed to TOC, in consideration of the fact that TOC has been applied to the living environment standard for the river since 2013, and that TOC is used instead of COD as an indicator of recalcitrant organic matter in Seoul's water quality monitoring networks. PO4-P was changed to TP, which has a strong correlation with PO4-P (Tables 5 and 6) and is currently applied in the living environment standard for the river. In fact, there are many water quality monitoring networks where COD and PO4-P are not measured, but TOC and TP are measured instead. Therefore, the final determined MS-WQI calculation items were TOC, TN, TP, and fecal coliforms, and MS-WQI was calculated according to Equation (2). Unlike for TP, in the case of TOC, the sub-index was not determined in the Bascarόn WQI water quality category; therefore, using the water quality data from the 17 WQMNs, the relational expression between COD and TOC was derived as shown in Figure 11. The TOC sub-index was determined by applying the existing COD sub-index to the relational expression. Furthermore, since TOC data exist from 2011 onwards, MS-WQI was calculated from this year.

Figure 11

Relationship between COD and TOC.

Figure 11

Relationship between COD and TOC.

Close modal

The water quality grade distribution using MS-WQI, similar to the Bascarόn WQI and S-WQI calculation results, indicated that the water quality grades of the 17 WQMNs were generally ‘good (II)’ to ‘normal (III).’ Based on the distribution of water quality grades on a monthly basis in January, out of a total of 119, 28 (23.5%) ‘very good (I)’, 52 (43.7%) ‘good (II)’, and 39 (32.8%) ‘normal (III)’ grades were obtained, and no points corresponded to ‘bad (IV)’ and ‘very bad (V)’ grades. On a monthly basis in July, there were no ‘bad (IV)’ and ‘very bad (V)’ grades out of a total of 120, but there were 10 (8.3%) ‘very good (I)’, 83 (69.2%) ‘good (II)’, and 27 (22.5%) ‘normal (III)’ grades. In the annual distribution of water quality grades, out of a total of 151, 11 (7.3%) ‘very good (I)’, 96 (63.6%) ‘good (II)’, and 44 (29.1%) ‘normal (III)’ grades were obtained, and no points corresponded to ‘bad (IV)’ and ‘very bad (V)’ grades. Overall, the distribution showed a trend similar to that of other WQI calculation results and the distribution of water quality grades.

Finally, the possibility of a linkage between the river water quality grade system applied in Korea and monthly MS-WQI systems such as the Bascarόn WQI and S-WQI was reviewed (Figure 12). In the Wilcoxon–Mann–Whitney test according to non-normal distribution, the significance probability p value was 0.7543, which was higher than the significance level of 5%; hence, it was judged that there was no significant difference between MS-WQI and the average water quality grade according to the environmental standards. In addition, the monthly MS-WQI averaged 76 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’, showing a high level of correlation with an R2 value of 0.7647. Thus, it was confirmed that MS-WQI can be applied to the currently implemented water quality grade system.

Figure 12

Comparison of modified Seoul water quality index (MS-WQI) and Living Environment Standard.

Figure 12

Comparison of modified Seoul water quality index (MS-WQI) and Living Environment Standard.

Close modal

Table 7 summarizes the main results for the Bascarόn WQI, S-WQI, and MS-WQI calculated for 17 WQMNs in Seoul for 18 years, from 2002 to 2019.

As shown in Table 7, regardless of the calculation method, the average values of all the WQIs were generally similar, and they all showed good agreement with the average water quality grade according to the environmental standards, without any significant difference. In the case of the customized WQI for urban rivers, when selecting representative water quality items necessary for calculating the index, ambiguity, uncertainty, and eclipsing between water quality items, which have been problems in developing WQIs, were minimized by advancing the methodology based on statistical analyses such as factor analysis. The calculation items were appropriately selected according to the water quality items of the current river living environment standards. Water quality items closely related to river water quality, such as dissolved oxygen and water temperature, are not included in the process of calculating the customized WQI for urban rivers. Therefore, there may be doubts on whether the WQI developed in this study is appropriate to represent the water quality of urban rivers. However, one of the important aspects of the WQI considered in this study is to identify changes or trends in water quality for urban river water quality management from a medium- and long-term perspective. From this point of view, when dissolved oxygen and water temperature, for example, which vary greatly from day to day, are included in WQI calculation items, the WQI value changes significantly depending on the measurement time.

Therefore, considering not only the result of factor analysis but also the meaning of WQI in this study, items with severe daily changes, such as dissolved oxygen and water temperature, are not suitable as WQI calculation items for medium- and long-term urban river water quality management. Therefore, S-WQI, a customized WQI for urban rivers, is a reasonable index that can represent the characteristics of urban river water quality. The calculation method is easy to apply, since it uses relatively fewer water quality items than the WQI calculated in the past, and it is highly likely to be linked to the currently implemented water quality grade system.

In addition, to extend the application of WQI to various water quality survey points, based on the calculation methodology performed to derive the indices in this study, such as MS-WQI, by adding new water quality items and changing some items, it is also possible to develop an advanced customized WQI for urban rivers considering watershed characteristics and measurement items.

To continuously utilize rivers as water resources, it is necessary to systematically manage river water quality. To this end, a WQI can be useful because it supplements the shortcomings of river water quality evaluation using only individual water quality items; it helps improve river water quality upon selecting water quality items, and can reflect changes in water quality due to changes in the water environment, such as new pollutants. Moreover, it can be used to convert complex and diverse water quality information into simple indicators for anyone to easily understand river water quality. Therefore, in this study, in order to strengthen urban river water quality management, to enhance the reliability of urban river water quality and generate urban river water quality information in a form that is easy to understand even for ordinary citizens, a customized WQI (Seoul water quality index, S-WQI) for urban rivers was developed based on an existing widely used WQI (Bascarόn WQI). In addition, a modified S-WQI (MS-WQI) was developed in order to expand and apply the S-WQI to various water quality measurement networks. At the same time, the possibility of a linkage between the river water quality grade system applied in Korea and the three monthly water quality indices was reviewed. As a result of the review, the monthly Bascarόn WQI showed an average of 74 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’ with an R2 value of 0.7971. the monthly S-WQI averaged 70 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’, similar to that based on the Bascarόn WQI with an R2 value of 0.8298. The monthly MS-WQI averaged 76 ‘good (II)’ grades, whereas the average water quality grade according to the environmental standards was ‘slightly good (II)’, showing a high level of correlation with an R2 value of 0.7647. Therefore, it was confirmed that the three water quality indices were well matched with the average water quality grade according to the environmental standards. In particular, the S-WQI, which is calculated using only four items through factor analysis, and the MS-WQI, which has changed some water quality items, are relatively easier to calculate than other water quality indices, and can represent the water quality characteristics of urban rivers using only a few items. Therefore, the S-WQI and MS-WQI are considered to be an efficient and scientific water quality evaluation method.

Using all the WQIs calculated in this study, river water quality managers at each institution can easily understand monthly changes in water quality at 17 WQMNs; therefore, if river water quality deteriorates during a specific period, it is expected that rapid response and preventive measures for water quality abnormalities will be possible. In addition, by calculating WQI in units of one year or more, it is possible to effectively grasp river water quality in the medium and long term. Because the value of WQI (grade) based on the medium and long term does not change easily, a change implies that the WQI calculation item representing water quality has been continuously affected by some factors from the past and is changing. Therefore, by using water quality index calculation methodology based on a statistical analysis method presented in this study and a newly developed water quality index such as S-WQI or MS-WQI, water quality managers will be able to more efficiently find various factors affecting river water quality and improve it than when using the existing standardized river water quality management focusing on individual water quality items. Furthermore, since complex and diverse water quality fluctuation characteristics can be summarized at a glance, and visually excellent results that are easily understood can be obtained using WQI, it is expected that ordinary citizens will be able to improve their access to and understanding of river water quality using this index. In the future, it is necessary to identify and quantify possible sources of uncertainty in the development stage of the water quality index. In particular, since urban rivers are widely used as leisure and ecological spaces rather than simply passageways for river water, for a comprehensive aquatic ecology evaluation of urban rivers, it is judged that additional research on the connection with environmental standards and upgrading the water quality index considering both biological data and water quality data of physicochemical items will be required.

The authors would like to thank the editors and reviewers for useful comments which were helpful in improving the quality of the manuscript.

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

The authors declare there is no conflict.

Bascarόn
M.
1979
Establishment of a methodology for determination of water quality
.
Boletin Informativo del Medio Ambiente
9
,
30
51
.
Brown
R. M.
,
McClelland
N. I.
,
Deininger
R. A.
&
Tozer
R. G.
1970
A water quality index: do we dare?
Water Sewage Works
117
(
10
),
339
343
.
Brown
R. M.
,
McCleiland
N. J.
,
Deiniger
R. A.
&
O'Connor
M. F.
1972
Water quality index – crossing the physical barrier
. In:
Proceedings of the 6th International Conference on Water Pollution Research, Jerusalem
(
Jenkis
S. H.
, ed.). Vol.
6
, pp.
787
797
.
Choi
J. Y.
1996
Development of Comprehensive Water Quality Index
.
Report Korea Environmental Technology Research Institute
,
Seoul
,
Korea
.
Chowdhury
R. M.
,
Muntasir
S. Y.
&
Hossain
M. M.
2012
Water quality index of water bodies along Faridpur-Barisal road in Bangladesh
.
Global Journal of Engineering and Technology Review
2
(
3
),
1
8
.
Chung
S. W.
&
Park
J. H.
2005
Application of Korean water quality index for the assessment of river water quality in the basin of Daecheong Lake
.
Journal of Korean Society on Water Environment
21
(
5
),
470
476
.
Debels
P.
,
Figueroa
R.
,
Urrutia
R.
,
Barra
R.
&
Niell
X.
2005
Evaluation of water quality in the Chillán River (Central Chile) using physicochemical parameters and a modified water quality index
.
Environmental Monitoring and Assessment
110
(
1
),
301
322
.
Dinius
S. H.
1987
Design of an index of water quality
.
Journal of the American Water Resources Association
23
(
5
),
833
843
.
Dunnette
D. A.
1979
A geographically variable water quality index used in Oregon
.
Journal of Water Pollution Control Federation
51
(
1
),
53
61
.
Gupta
I.
,
Kumar
A.
,
Singh
C.
&
Kumar
R.
2015
Detection and mapping of water quality variation in the Godavari River using water quality index, clustering and GIS techniques
.
Journal of Geographic Information System
7
(
2
),
71
84
.
Kal
B. S.
,
Park
J. B.
,
Kim
S. H.
&
Im
T. H.
2017
Assessment of tributary water quality using integrated water quality index
.
Journal of Wetlands Research
19
(
3
),
311
317
.
Kankal
N. C.
,
Indurkar
M. M.
,
Gudadhe
S. K.
&
Wate
S. R.
2012
Water quality index of surface water bodies of Gujarat, India
.
Asian Journal of Experimental Sciences
26
(
1
),
39
48
.
Kannel
P. R.
,
Lee
S.
,
Lee
Y. S.
,
Kanel
S. R.
&
Khan
S. P.
2007
Application of water quality indices and dissolved oxygen as indicators for river water classification and urban impact assessment
.
Environmental Monitoring and Assessment
132
(
1
),
93
110
.
Kassambara
A.
2017
Practical Guide to Principal Component Methods in R. STHDA, France
.
Khan
A. A.
,
Tobin
A.
,
Paterson
R.
,
Khan
H.
&
Warren
R.
2005
Application of CCME procedures for deriving site-specific water quality guidelines for the CCME water quality index
.
Water Quality Research Journal
40
(
4
),
448
456
.
Kim
Y. J.
,
Gal
B. S.
,
Park
J. B.
,
Kim
S. H.
&
Im
T. H.
2018
Classification of Nakdong River tributaries under priority management based on their characteristics and water quality index
.
Journal of Korean Society of Environmental Engineers
40
(
2
),
73
81
.
Lee
J. H.
,
Ha
H. J.
,
Lee
M. H.
,
Lee
M. Y.
,
Kim
T. H.
,
Cha
Y. K.
&
Koo
J. Y.
2020
Assessment of water quality of major tributaries in Seoul using water quality index and cluster analysis
.
Journal of Korean Society of Environmental Engineers
42
(
10
),
452
462
.
Liu
C. W.
,
Lin
K. H.
&
Kuo
Y. M.
2003
Application of factor analysis in the assessment of groundwater quality in a Blackfoot disease area in Taiwan
.
Science of the Total Environment
313
(
1–3
),
77
89
.
Lumb
A.
,
Halliwell
D.
&
Sharma
T.
2006
Application of CCME water quality index to monitor water quality: a case of the Mackenzie river basin, Canada
.
Environmental Monitoring and Assessment
113
,
411
429
.
Ministry of Environment
2020
The Enforcement Decree of The Framework Act on The Environmental Policy
.
Ministry of Environment
,
Korea
.
Misaghi
F.
,
Delgosha
F.
,
Razzaghmanesh
M.
&
Myers
B.
2017
Introducing a water quality index for assessing water for irrigation purposes: a case study of the Ghezel Ozan River
.
Science of the Total Environment
589
,
107
116
.
National Institute of Environmental Research (NIER)
2017a
Water Pollution Standard Method
.
Ministry of Environment
,
Korea
.
NIER
2017b
Quality Standard for Drinking Water
.
Ministry of Environment
,
Korea
.
Raphael
A.
,
Monica
C.
,
Wamberto
R. S. J.
,
Taysa
T. V. M.
,
Carmem
L. M. G.
&
Malva
I. M. H.
2007
Use of index analysis to evaluate the water quality of a stream receiving industrial effluents
.
Water SA
33
(
4
),
459
466
.
Shweta
T.
,
Bhavtosh
S.
,
Prashant
S.
&
Rajendra
D.
2013
Water quality assessment in terms of water quality index
.
American Journal of Water Resources
1
(
3
),
34
38
.
Wu
Z.
,
Wang
X.
,
Chen
Y.
,
Cai
Y.
&
Deng
J.
2018
Assessing river water quality using water quality index in Lake Taihu Basin, China
.
Science of the Total Environment
612
,
914
922
.
Yogendra
K.
&
Puttaiah
E. T.
2008
Determination of water quality index and suitability of an urban waterbody in Shimoga Town, Karnataka
. In
Proceedings of Taal2007: The 12th World Lake Conference
, Vol.
342
, p.
346
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).