Kelani River is the principal consumable water source for 80% of the population in the Colombo district and an important ecosystem complex for the freshwater fish biota of Sri Lanka. However, it is the most polluted river in the country. The present study was conducted to determine the water quality parameters and pollution of the upper and lower catchments of the river and select the most suitable parameters for predicting the pollution of each catchment. Thirteen locations of each catchment were selected for the study, and 14 water quality parameters were recorded by standard techniques. Measurements were compared with the standard values permissible for drinking purposes and aquatic life and subjected to principal component analysis. The study revealed that the most polluted catchment of the Kelani River was the lower catchment, and the chemical oxygen demand (COD) and water pH were selected as the most suitable parameters to predict the pollution levels of the lower catchment. The nitrate concentration and COD were selected as the most suitable water quality parameters to predict the pollution of the upper catchment. The present study indicates an accelerating trend in water pollution of the Kelani River when compared with studies conducted two decades ago.

  • The two catchments of the Kelani River were investigated for water quality and pollution.

  • The lower catchment of the Kelani River was more polluted.

  • COD and pH of the water are suitable to predict the water quality of the lower catchment.

  • COD and nitrates are suitable to predict the water quality of the upper catchment.

  • The study indicates an accelerating trend in water pollution of the Kelani River.

Graphical Abstract

Graphical Abstract
Graphical Abstract

Water resources are the main economic background of a country. In recent years, the amount of renewable water resources in the world has decreased with increasing human population and water demand, climate change, deforestation, urbanization, and pollution (Gebeyehu et al. 2018). Pollution in aquatic environments has become a severe worldwide problem during the past few decades, and no country has succeeded in turning the trend of increasing and accelerating water pollution into a leveling out or decrease (Dybern 1974). Numerous aquatic environments in both developed and developing countries are polluted caused by either development or lack of development. In developing countries due to lack of hygienic facilities, river basins and canals are the only latrines available for the poorer part of the population, and most household wastes and waste from industries are also discharged into the water (Dybern 1974). According to Sikder et al. (2013), the rivers in developing countries are considered to be more affected with respect to dissolved metal, organic matter and fecal pollution.

Sri Lanka has 103 rivers of which 29 rivers flow directly to the sea while the rest connect to either a major river, salt marsh, lagoon, or lake (Katupotha & Gamage 2020). Many are at a low level of exploitation, except a few that are heavily regulated for domestic and irrigational water supply, and hydropower generation (Eriyagama et al. 2015). The heavily exploited rivers that flow through densely populated and intensively urbanized cities are subjected to severe pollution as can be identified in the Mahaweli River (Abeygunawardane et al. 2011; Bandara et al. 2011; Wickramasinghe et al. 2018), the Gin Ganga (Kumar et al. 2019), Walawe River (Ileperuma 2000), and Malwathu Oya (Zoysa & Weerasinghe 2016). However, the Kelani River has been identified as the most polluted river in Sri Lanka (Ileperuma 2000; Abeysinghe & Samarakoon 2017; Kumar et al. 2019).

The Kelani River is the fourth-longest river (144 km) in Sri Lanka (Kumar et al. 2019) and the second largest river in volume of discharge (Chandimala & Zubair 2007). The river originates in the central hills at Kirigalpotha mountain range, 2,420 m above sea level and discharges to the sea 144 km downstream at Colombo (Chandimala & Zubair 2007), passing four administrative districts (Nuwara Eliya, Kegalle, Gampaha, Colombo) and three provinces (Central, Sabaragamuwa, Western) in the country (Fayas et al. 2019). The upper catchment is mountainous and primarily covered with thick vegetation, including tea, rubber, grass and forest (Kumar et al. 2019), while the lower catchment is urbanized (Kumar et al. 2019) and has plain features (De Silva et al. 2012) consisting of rubber and rice cultivations (Chandimala & Zubair 2007). The Kelani River is severely polluted by natural phenomena such as saltwater intrusion (Ranmadugala et al. 2007), flood inundation due to heavy rainfalls (De Silva et al. 2012) and soil erosion (Fayas et al. 2019), and a multitude of anthropogenic activities occurring alongside the river. Many industries such as rubber industries, textile industries, breweries, oil refineries, fertilizer industries, plywood industries and leather tanning factories located along the riverbank, discharge industrial waste into the river, and household wastes are dumped directly via garbage dumping sites close to the river (Ileperuma 2000). The prevalence of Escherichia coli in many locations of the river suggests fecal pollution of the river, especially due to growing populations and poor living standards (Kumar et al. 2019). Furthermore, hydropower reservoirs and mini-hydropower plants have modified the water chemistry, especially the conductivity, dissolved oxygen (DO), and alkalinity (Surasinghe et al. 2020).

One of the major implications of such geogenic and anthropogenic contamination is groundwater pollution, and recent research has revealed that the groundwater resources of the entire Kelani River Basin are contaminated and raw water consumption is unsafe. The Water Quality Index (WQI) assessment for the groundwater of the Kelani River Basin has ranked the water as ‘poor’ for drinking purposes (Mahagamage et al. 2016), and the entire basin is known to be contaminated with total coliform bacteria (Mahagamage et al. 2020). Most of the groundwater sources are contaminated with human pathogenic bacteria such as Salmonella spp. and Campylobacter spp., and Salmonella spp. contaminations are high in groundwater during the dry season than in the wet season (Mahagamage et al. 2020). Furthermore, according to Liyanage et al. (2021), the groundwater sources of the entire lower part of the Kelani River Basin are contaminated with antibiotics and not suitable for drinking purposes.

However, the Kelani River is the principal consumable water source for 80% (over 6 million) of the human population of the Colombo district (Surasinghe et al. 2020). Furthermore, it is an important ecosystem complex for the freshwater fish biota of Sri Lanka and accounts for a total of 60 fish species of which 30 are endemic (Surasinghe et al. 2020). Therefore, it is important that the water quality parameters of the river be monitored regularly and effectively using accurate pollution predictability features of importance.

Certain water quality parameters and heavy metal concentrations of the Kelani River water have been documented by Ileperuma 2000; Mahagamage & Manage 2014; Abeysinghe & Samarakoon 2017; Kuruppuarachchi & Pathiratne 2020; Thotagamuwa & Weerasinghe 2021.

However, a methodical study assessing the water quality parameters targeting the lower and upper catchments separately has not been conducted. Furthermore, the vector features/parameters that can predict the pollution of the catchments more effectively have not been extracted. Therefore, the present study was conducted to assess the pollution of the lower and upper catchments of the Kelani River and extract the important water quality features that predict the pollution of each catchment. The results of the study will be important for establishing proper water quality management strategic plans and maintaining safe drinking water, which is a significant problem in Sri Lanka and other countries of the Eastern hemisphere which are going through rapid economic development (Li et al. 2021).

The study further intends to compare the pollution of the catchments with previous recordings and evaluate whether the trend of pollution in the river basin has accelerated, leveled out or decreased over the past few years.

Selection of locations

The river basin was considered as upper and lower catchments depending on the topography of the whole basin. The lower catchment was mostly flat and extended to an elevation of about 100 m and consisted of locations with different degrees of erosion and pollution. The upper catchment consisted of locations associated with a chain of mountains which started around 300 m elevation. Sampling locations within the two catchments were selected by considering the catchment characteristics, anthropological activities, industrial discharge, and accessibility to fish sampling sites. Twenty-six sampling locations were selected for both upper and lower catchments including 13 locations from each catchment. Certain locations were selected as reference sites which are known to have a very low pollution status according to the literature (Figure 1). Coordinates of each sampling location were recorded using a GPS (Hand-held Garmin eTrex 30 GPS receiver) in order to prepare the location map.
Figure 1

Selected sampling locations of the upper and lower catchments of the Kelani River Basin (adapted from Edussuriya & Pathirage 2016). L01: Mattakkuliya, L02: Thotalanga, L03: Wellampitiya Bridge, L04: Kolonnawa, L05: Ambathale Bridge, L06: Biyagama, L07: Kaduwela, L08: Nawagamuwa, L09: Panagoda, L10: Padukka, L11: Hanwella Bridge, L12: Wak Oya, L13: Thummodara, U01: Lahupana Ella, U02: Kotiyakumbura, U03: Bulathkohupitiya, U04: Parussalla, U05: Wee Oya, U06: Panakoora, U07: Alagal Oya, U08: Kithulgala, U09: Kalugala Bridge, U10: Koththallena, U11: Nallathanniya, U12: Goverawela B Division, U13: Bagawanthalawa.

Figure 1

Selected sampling locations of the upper and lower catchments of the Kelani River Basin (adapted from Edussuriya & Pathirage 2016). L01: Mattakkuliya, L02: Thotalanga, L03: Wellampitiya Bridge, L04: Kolonnawa, L05: Ambathale Bridge, L06: Biyagama, L07: Kaduwela, L08: Nawagamuwa, L09: Panagoda, L10: Padukka, L11: Hanwella Bridge, L12: Wak Oya, L13: Thummodara, U01: Lahupana Ella, U02: Kotiyakumbura, U03: Bulathkohupitiya, U04: Parussalla, U05: Wee Oya, U06: Panakoora, U07: Alagal Oya, U08: Kithulgala, U09: Kalugala Bridge, U10: Koththallena, U11: Nallathanniya, U12: Goverawela B Division, U13: Bagawanthalawa.

Close modal

Collection of water samples and measurement of water quality parameters

Water samples were collected once in 3 months during a 1-year period from May 2019 to May 2020 from the 26 upper and lower catchment locations. The physical and chemical parameters of the collected water samples were analyzed using standard methods according to American Public Health Association (APHA 2012), in triplicates per location for each parameter. The physical parameters, water temperature (WT) and turbidity were measured at the site itself. The HQD portable multimeter was used to measure the WT, and turbidity measurements were taken using the EUTECH TN-100 portable meter. The chemical parameters such as pH, DO, and electrical conductivity (EC) were measured at the sites using an HQD portable multimeter for pH and DO, and a conductivity meter for EC. The total suspended solids (TSS), biological oxygen demand (BOD), chemical oxygen demand (COD), nutrients such as nitrate-nitrogen (-N), nitrite-nitrogen (-N), ammoniacal-nitrogen (NH3-N), and total phosphate (T-), total hardness (TH) and total alkalinity were measured in the laboratory using collected water samples. For laboratory analysis, the water samples that were collected from the sites were filtered (using 47 mm GF/C Whatman glass microfiber filters) and placed into sealable poly-propylene bottles. Five-day BOD was analyzed using the Winkler method, and a spectrophotometer was used to get the reading for nutrients of the samples. Open reflux digestion method and titrimetric method were used to determine the COD, TH, and total alkalinity, respectively.

Statistical analysis

Multivariate analysis of variance (MANOVA) was conducted for the water quality parameters of the upper and lower catchments separately using SPSS version 25 statistical software. MANOVA was selected for the analysis rather than ANOVA as the ecological variables used in the study were not independent. Variables that are not independent have potential interactions, and the use of ANOVA will inflate the error of the test. MANOVA will not inflate error and will be more appropriate for testing the ecological variables (Parsad & Bhar 1987). MANOVA was used to determine whether multiple levels of variables on their own or in combination have an effect on the dependent variable. For these four multivariate measures: Wilks’ lambda, Pillai's trace, Hotelling-Lawley trace, and Roy's largest root were calculated to examine the variance in the data. In order to determine whether the locations selected in the study differ according to the considered water quality parameters, the significance values (p < 0.05) were considered (Mertler & Reinhart 2016). Based on the statistical significance of MANOVA multivariate measures, principal component analysis (PCA) or Factor Analysis was carried out to select the most important water quality variables that will specify the context of the locations. PCA was used as it reduces the dimensionality of the data set when it has a large proportion of interdependent variables while preserving the variability as much as possible (Jolliffe & Cadima 2016; Kalaivani et al. 2020). PCA was conducted for the locations of the upper and lower catchments separately, and eigenvalues (i.e., coefficients attached to eigenvectors) were calculated for the 14 water quality parameters. A scree plot was constructed between water quality parameter (x-axis) and eigenvalue (y-axis). Factors producing an eigenvalue of more than one were selected using the scree plot and extracted for further analysis. A rotated factor analysis was conducted to determine the correlated water quality parameters that contribute to each extracted factor. A factor analysis examines all the pairwise relationships between individual variables and seeks to extract latent factors from the measured variables. A rotated factor analysis clarifies and simplifies the results of a factor analysis and is easier to interpret (Osborne 2015).

Water quality in the Kelani River Basin

Water temperature

The WT of both the upper and lower catchment locations did not exceed 40 °C, which is the tolerance limit for the discharge of industrial waste into inland waters. However, the upper catchment of the river basin had a higher temperature variation than the lower catchment with temperatures ranging from 21.53 to 31.33 °C (Figure 2(a)). The lowest temperature was recorded in U11, Nallathanniya, which is a forested area known for water with pristine quality while the highest temperature was recorded in U2, Kotiyakumbura, a comparatively populated place (Figure 1). When considering the lower catchment, the WT varied from 26.3 °C in L7, Kaduwela to 32.9 °C in L6, Biyagama both of which are industrial areas (Figure 2(a)).
Figure 2

Variations of the following physical parameters of water in the lower and upper catchments of the Kelani River Basin: (a) water temperature, (b) turbidity, (c) total suspended solids, and (d) electrical conductivity.

Figure 2

Variations of the following physical parameters of water in the lower and upper catchments of the Kelani River Basin: (a) water temperature, (b) turbidity, (c) total suspended solids, and (d) electrical conductivity.

Close modal

Turbidity

According to the Central Environmental Authority, 5–50 NTU levels of turbidity are permissible for inland waters of the country. In the present study, all locations of the upper catchment were within the permissible levels, having the highest value recorded for U2, Kotiyakumbura (25.23 NTU) (Figure 2(b)). U2 also had the highest level of TSS (0.048 mg/l) indicating the association between turbidity and TSS and their effect on water clarity (Figure 2(c)). Turbidity levels were high in the lower catchment when compared to the upper catchment locations, with a significantly high value at L11, Hanwella Bridge (85.40 NTU) which exceeded the recommended turbidity value (Figure 2(b)).

Total suspended solids

Many locations of the lower catchment, including L1 Mattakkuliya, L4 Kolonnawa, L11 Hanwella Bridge, and L12 Wak Oya, had TSS values exceeding the permissible levels, with L5 Ambathale Bridge, the water intake point for purification, having a significantly high value of 45.8 mg/l (Figure 2(c)).

Electrical conductivity

In the present study, all selected locations of the Kelani River Basin had EC values that were not harmful to fish and other aquatic life. However, in certain locations of both the upper catchment (U2 Kotiyakumbura) and lower catchment (L2 Thotalanga, L12 Wak Oya), the EC values exceeded the levels permissible for drinking water (Figure 2(d)). The high EC values in the lower catchment can be attributed to the proximity of the lower basin to the river mouth. However, high EC values indicate high levels of salinity in the water, and the higher the salinity level, the lower the DO concentration. This was evident in U2 which had the highest EC value (452.6 μS/cm) and lowest DO concentration (5.1 mg/l) (Figure 3h).
Figure 3

Variations of the following chemical parameters of water in the lower and upper catchments of the Kelani River Basin: (a) pH, (b) alkalinity, (c) total hardness, (d) nitrate nitrogen, (e) nitrite nitrogen, (f) ammoniacal nitrogen, (g) phosphate, (h) dissolved oxygen, (i) biological oxygen demand, and (j) chemical oxygen demand.

Figure 3

Variations of the following chemical parameters of water in the lower and upper catchments of the Kelani River Basin: (a) pH, (b) alkalinity, (c) total hardness, (d) nitrate nitrogen, (e) nitrite nitrogen, (f) ammoniacal nitrogen, (g) phosphate, (h) dissolved oxygen, (i) biological oxygen demand, and (j) chemical oxygen demand.

Close modal

pH

When considering the pH values recorded for the upper catchment of the Kelani River Basin, all the values were within the prescribed limit (6.5–8.5) published by the Central Environmental Authority (CEA 1992), though U4, U9, U10, and U11 locations had values in the upper margin of the above range and U12 had a value in the lower margin (Figure 3(a)). Location U9, Kalugala Palama, which had the highest pH value (8.37), also had the highest value for alkalinity (62.8 mg/l) (Figure 3(b)). Waters with high pH and alkalinity indicate adverse effects if used for irrigational purposes. Location U12, Goverawella, with a low pH value was a residential area where many tea estates were present (Figure 1). When considering the lower catchment, the highest pH value was recorded in L7, Kaduwela (8.43), which was near the highest range of the acceptable value and the lowest pH value was recorded in L1, Mattakkuliya (5.33), which was lower to the lowest margin of the above range (Figure 3(a)).

Alkalinity and TH

The waters at L4, Kolonnawa, had high alkalinity (38.10 mg/l) when compared with the other locations of the lower catchment (Figure 3(b)). In the present study, the alkalinity of the locations of both upper and lower catchments of the river was within the permissible levels. Locations with high alkalinity displayed high TH and vice versa: U9 with the highest alkalinity (62.80 mg/l) had the highest TH (30.34 mg/l); U11 with the lowest alkalinity (6.27 mg/l) had a very low TH (3.69 mg/l); L4 with the highest alkalinity (38.1 mg/l) had the highest TH (87.4 mg/l); L13 with the lowest alkalinity (8.67 mg/l) had the lowest TH (15.17 mg/l) (Figures 3(b) and 3(c)). However, in the upper catchment locations (U9 and U11), TH was lower than the alkalinity values while in the lower catchment locations (L4 and L13), TH was higher than the alkalinity values (Figures 3(b) and 3(c)).

Forms of nitrogen (nitrates, nitrites, ammonia) and phosphates

Nitrates, nitrites, ammonia and phosphates are essential plant nutrients, which in excess can cause significant water quality problems. Excess amounts of these nutrients may accelerate eutrophication, increasing plant growth and change in plant types that will subsequently lead to low levels of DO.

In the locations of the upper catchment, nitrate, nitrite, ammonia and phosphate concentrations were well below the maximum level permissible for aquatic life, while certain locations of the lower catchment had high nitrite, ammonia and phosphate levels (Figure 3(d)–3e,f,g). L6 Biyagama, an industrialized location, and L7 Kaduwela, an industrialized and residential area, both had nitrite levels of 0.097 mg/l that were three-fold higher than the permissible level of 0.03 mg/l (Figure 3(e)). Furthermore, location L7, Kaduwela, had high ammonia (0.977 mg/l) and phosphate (0.513 mg/l) concentrations that exceeded the tolerance limits, and phosphate concentrations in L1 (0.668 mg/l) and L4 (0.442 mg/l) were also high (Figures 3(f) and 3g). The adverse effects of excessive nutrients in the waters of the lower catchment were evident when considering the DO levels. Locations with high nutrients displayed decreased levels of DO (Figure 4).
Figure 4

Fluctuations of dissolved oxygen with nutrients in locations of the lower catchment of the Kelani River Basin.

Figure 4

Fluctuations of dissolved oxygen with nutrients in locations of the lower catchment of the Kelani River Basin.

Close modal

Dissolved oxygen

DO content of the majority of the selected locations was above the proposed water quality standards for fish and aquatic life given by the CEA in which DO should be higher than 4 mg/l with a mean value of 6 mg/l. However, L4, Kolonnawa of the lower catchment had a DO of 3.65 mg/l, which is lower than the recommended level for aquatic life (Figure 3h). When considering the upper catchment, the DO content of all the locations exceeded the standard values and was suitable for fish and aquatic life (Figure 3h).

Biological oxygen demand

Low BOD content is an indicator of good quality water and high BOD values present polluted water (Ileperuma 2000). BOD levels of the upper catchment of the river were within the tolerance limits for fish and other aquatic life. However, in the lower catchment, a high BOD value of 5.6 mg/l was recorded for L3, Wellampitiya, an industrial area with an oil refinery (Figures 1 and 3(i)). L9, Panagoda, an agricultural area with paddy and rubber, had the lowest BOD value, which was 1.20 mg/l (Figures 1 and 3(i)). For the upper catchment, highest and lowest values were recorded as 3.77 mg/l for U12, Goverawella and 1.63 mg/l for U2, Kotiyakumbura, respectively (Figure 3(i)).

Chemical oxygen demand

In the upper catchment, the highest COD value was received for U6, Panakoora, which was 25.33 mg/l and lowest for the location U12, Goverawella which was 7.67 mg/l. Both locations were tea estates, and location U12 was associated with a forested area (Figures 1 and 3(j)). Within the lower catchment, the highest COD value was recorded for L4 which was 81.10 mg/l and the lowest COD value was recorded for the L8 location with a value of 14.27 mg/l. The COD levels of the lower catchment were higher than in the upper catchment, and the majority of locations had values that exceeded the maximum tolerance limits (Figure 3(j)).

The results given above showed that the water in all locations of the lower catchment with the exception of L8, Nawagamuwa was not within the permissible levels for one or more water quality parameter/s. The waters in L1 Mattakkuliya, L4 Kolonnawa, L7 Kaduwela were unsuitable for drinking purposes and aquatic life in a high number of parameters when compared with the other locations of the lower catchment. Kotiyakumbura (U2) of the upper catchment had a high EC while the COD concentrations were high in U4 Parussalla, U6 Panakoora and U8 Kitulgala.

Assessment of water quality in the lower catchment of the Kelani River Basin

When conducting MANOVA using the four multivariate measures, different values were obtained for each measure (Table 1). As the sample size increases, the value produced by the four measures are known to become similar. Therefore, it can be ascertained that the sample sizes of the current study were small. However, all four multivariate measures were statistically significant (Sig. 0.000, p < 0.05), indicating that the 13 locations of the lower catchment were significantly different from each other in terms of the selected water quality parameters (Table 1).

Table 1

Multivariate analysis of variance (MANOVA) for the lower and upper catchment locations of the Kelani River

TESTValueFError dfSig.
Lower locations Pillai's Trace 10.878 14.865 276.000 .000 
Wilks’ Lambda .000 529.194 137.416 .000 
Hotelling's Trace 66,344.083 3,747.212 122.000 .000 
Roy's Largest Root 47,129.479 72,265.202 23.000 .000 
Upper locations Pillai's Trace 10.501 12.011 288.000 .000 
Wilks’ Lambda .000 340.493 141.865 .000 
Hotelling's Trace 27,838.032 1,850.345 134.000 .000 
Roy's Largest Root 13,734.922 23,545.581 24.000 .000 
TESTValueFError dfSig.
Lower locations Pillai's Trace 10.878 14.865 276.000 .000 
Wilks’ Lambda .000 529.194 137.416 .000 
Hotelling's Trace 66,344.083 3,747.212 122.000 .000 
Roy's Largest Root 47,129.479 72,265.202 23.000 .000 
Upper locations Pillai's Trace 10.501 12.011 288.000 .000 
Wilks’ Lambda .000 340.493 141.865 .000 
Hotelling's Trace 27,838.032 1,850.345 134.000 .000 
Roy's Largest Root 13,734.922 23,545.581 24.000 .000 

As the water quality parameters of the 13 selected locations were significantly different, a scree plot was constructed (Figure 5). The scree plot denoted that six components had an eigenvalue greater than one and can be extracted to represent the total number of water quality parameters. According to Table 2, the six factors that were extracted displayed 91.5% of the total variation of the data set and thus can be accurately used to describe the data structure.
Table 2

Values for the extracted factors against analysis parameters of the lower catchment of the Kelani River

ComponentInitial Eigenvalues
Extraction sums of squared loadings
Total% of VarianceCumulative %Total% of VarianceCumulative %
4.393 29.288 29.288 3.982 26.546 26.546 
2.702 18.016 47.304 2.128 14.188 40.734 
2.546 16.972 64.276 2.106 14.039 54.773 
1.513 10.087 74.362 2.024 13.493 68.266 
1.323 8.819 83.181 1.840 12.264 80.530 
1.248 8.322 91.502 1.646 10.973 91.502 
ComponentInitial Eigenvalues
Extraction sums of squared loadings
Total% of VarianceCumulative %Total% of VarianceCumulative %
4.393 29.288 29.288 3.982 26.546 26.546 
2.702 18.016 47.304 2.128 14.188 40.734 
2.546 16.972 64.276 2.106 14.039 54.773 
1.513 10.087 74.362 2.024 13.493 68.266 
1.323 8.819 83.181 1.840 12.264 80.530 
1.248 8.322 91.502 1.646 10.973 91.502 
Figure 5

Screeplot of eigenvalues derived from the lower catchment water quality data. The plot clearly shows that six factors are above eigenvalue of 1 and beyond this point the graph levels out.

Figure 5

Screeplot of eigenvalues derived from the lower catchment water quality data. The plot clearly shows that six factors are above eigenvalue of 1 and beyond this point the graph levels out.

Close modal

The PCA/Factor analysis performed illustrated the correlated water quality parameters of each of the six extracted factors and indicated that the correlated parameters contributing to a particular factor are closely associated and contribute to other factors insignificantly (Table 3).

Table 3

PCA/factor analysis for water quality parameters of the lower catchment of the Kelani River

 
 

Note: Extraction method: principal component analysis and rotation method: varimax with Kaiser normalization.

Factor 1 accounted for 29.29% of the total variance (Table 2) and consisted of the highest positive loadings of EC, turbidity, WT, and the highest negative loadings of DO and BOD (Table 3). This factor can be considered as the water quality indicating factor which describes 29.29% of the total variance. Factor 2 had a high negative loading of alkalinity and high positive loading of TH and the amount of variation described for this factor was 18% of the total variance (Tables 2 and 3). Thus, Factor 2 can be considered as the measurement of the iron concentration of water. Similarly, phosphates, TSS and ammonia belong to Factor 3, and thus, Factor 3 can be considered as the wastewater indicating factor (Table 3). Factors 4, 5, and 6 describe nitrogen pollution, acidity and chemical pollution, respectively (Table 3). Interestingly, Factor 5 alone produces a positive loading of 0.813 for pH that accounts for a contribution of 8.82% from the total variance, and Factor 6 alone produces a positive loading of 0.921 for COD that accounts for a contribution of 8.32% variance from the total variance (Tables 2 and 3). Therefore, it can be decided that the pH and COD, which are represented by Factors 5 and 6, respectively, have a high contribution to the pollution loading in the lower catchment of the river, and these parameters are the most suitable water quality parameters to predict the water quality of the lower catchment.

Assessment of water quality in the upper catchment of the Kelani River Basin

In the upper catchment too, the four multivariate measures in MANOVA yielded different values for each measure indicating that the sample sizes of the current study were small (Table 1). However, all four multivariate measures were statistically significant (Sig. 0.000, p < 0.05), indicating that the 13 locations of the upper catchment were significantly different from each other in terms of the selected water quality parameters (Table 1).

As the water quality parameters of the 13 selected locations were significantly different, a scree plot was constructed (Figure 6). The scree plot denoted that six components had an eigenvalue larger than one and can be extracted to represent the total number of water quality parameters. According to Table 4, the six factors that were extracted displayed 91.3% of the total variation of the data set and thus can be accurately used to describe the data structure.
Table 4

Values for the extracted factors against analysis parameters of the upper catchment of the Kelani River

ComponentTotal variance explained before rotation
Total variance explained after rotation
Extraction sum of squared loadings
Rotation sum of squared loadings
Initial Eigen values
Eigen value
Total% of VarianceCumulative %Total% of VarianceCumulative %
3.873 27.665 27.665 3.873 27.665 27.665 
2.583 18.448 46.113 2.583 18.448 46.113 
2.343 16.738 62.851 2.343 16.738 62.851 
1.421 10.147 72.998 1.421 10.147 72.998 
1.320 9.429 82.427 1.320 9.429 82.427 
1.243 8.877 91.304 1.243 8.877 91.304 
ComponentTotal variance explained before rotation
Total variance explained after rotation
Extraction sum of squared loadings
Rotation sum of squared loadings
Initial Eigen values
Eigen value
Total% of VarianceCumulative %Total% of VarianceCumulative %
3.873 27.665 27.665 3.873 27.665 27.665 
2.583 18.448 46.113 2.583 18.448 46.113 
2.343 16.738 62.851 2.343 16.738 62.851 
1.421 10.147 72.998 1.421 10.147 72.998 
1.320 9.429 82.427 1.320 9.429 82.427 
1.243 8.877 91.304 1.243 8.877 91.304 
Figure 6

Screeplot of eigenvalues derived from the upper catchment water quality data. The plot clearly shows that six factors are above the eigenvalue of 1 and beyond this point the graph levels out.

Figure 6

Screeplot of eigenvalues derived from the upper catchment water quality data. The plot clearly shows that six factors are above the eigenvalue of 1 and beyond this point the graph levels out.

Close modal

The PCA/factor analysis performed illustrated the correlated water quality parameters of each of the six extracted factors and indicated that the correlated parameters contributing to a particular factor are closely associated and contribute to other factors insignificantly (Table 5).

Table 5

PCA/factor analysis for water quality parameters of the upper catchment of the Kelani River

 
 

Note: Extraction method: principal component analysis and rotation method: varimax with Kaiser normalization.

Factor 1 accounted for 27.67% of the total variance (Table 4) and consisted of the highest positive loadings of EC, turbidity, WT, nitrites and the highest negative loadings of DO (Table 5). The influence of Factor 1 was more or less similar to Factor 1 in the lower catchment. Factor 2 had a high negative loading of TSS, high positive loading of ammonia and moderate positive loading of BOD (Table 5). The amount of variation described in this factor is 18.45% of the total variance and can be considered as a minor level of organic pollution in water. Factor 3 has high positive loadings of alkalinity and TH and the amount of variation described for this factor is 16.74% of the total variance (Tables 4 and 5). Thus, Factor 3 can be considered as the measurement of the iron concentration of water. Factor 4 has the highest positive loadings of nitrates, while Factor 5 has the highest negative loadings of COD (Table 5). Factor 6 accounts for the highest positive loadings of pH and phosphates (Table 5). Interestingly, both Factors 4 and 5 represent single parameters with high variation. Factor 4 represents the nitrate concentration of water that contributes to the 10.15% variation of the data set, while Factor 5 represents the COD of water that contributes to the 9.43% variation of the data set (Tables 4 and 5). Due to representing single parameters that describe a high variation of the total data set, Factors 4 and 5 can be selected to interpret the water quality of the upper catchment. Therefore, it can be decided that the nitrate content and COD have a high contribution to the pollution in the upper catchment of the river and are suitable to predict the water quality of the upper catchment.

The quality of water in rivers is of considerable importance for the reason that they sustain various uses and processes that supports a healthy ecosystem. Riverine water quality can vary between different rivers or within and between the catchments of the same river. Spatial variability of water quality within an individual river has been attributed to the influence of landscape characteristics, climate, atmospheric deposition and topography, which are reflected in the physical and chemical features of the water (Lintern et al. 2018). With this factuality in mind, the present study was conducted to record the water quality parameters of the upper and lower catchments of the Kelani River Basin and reveal the most important parameters determining the pollution of each catchment. According to the present findings, most of the water quality parameters of the upper catchment of the river were within the standards for inland waters and were suitable for drinking purposes and aquatic life. However, most parameters of the lower catchment were not within the recommended limits permissible for aquatic life and were more polluted than the upper catchment with respect to many of the water quality parameters investigated. This finding has also been disclosed by Ileperuma (2000), who reported a regular increase in ammonia, nitrates and BOD from the origin of the river to the point of discharge into the sea and more recently by Ruvinda & Pathiratne (2020), who revealed that the physicochemical characters of the river confirm an increasing trend of pollution toward the lower reach. Kuruppuarachchi & Pathiratne (2020) also report that the lower catchment of the Kelani River is highly contaminated with toxic materials discharged by leading export processing industries located near the river bank. However, many other studies regarding the Kelani River provide different opinions on the pollution and deterioration of the two river basins and suggest various causes for the outcome. According to Mahagamage et al. (2016), the groundwater of the entire river is not suitable for drinking purposes and most of the physical, chemical and biological parameters of the water are not within the drinking water quality standards. As reported by the study, pollution is more severe in locations of the lower catchment due to increased residential and industrialized areas, and irrigational activities. Nevertheless, the upper catchment surrounded by tea estates is known to contaminate the associated groundwater aquifers with fertilizers and pesticides mixed with rainwater. Later on, Mahagamage et al. (2020) revealed that the entire Kelani River Basin is contaminated by total coliform and E. coli bacteria and confirmed that the groundwater in many locations is positive for Salmonella spp. that cause gastrointestinal diseases. Kumar et al. (2020) stated that the E. coli strains isolated from the waters of Kelani River are resistant to antibiotics such as tetracycline and sulfamethoxazole and are an emerging environmental concern owing to their potential threat to human health. Very recently, Liyanage et al. (2021) revealed that the surface and groundwater of the entire lower part of the Kelani River Basin were contaminated with total coliform and fecal coliform bacteria and that penicillin and tetracycline group antibiotics are detected at the river mouth. The present study is also in accordance with most of these previous findings and affirms that the entire Kelani River is polluted. However, the present study confirms that the lower river basin is more polluted than the upper basin and elucidates the most adequate water quality parameters that denote the pollution of each river basin.

Water quality assessment in the lower catchment

The PCA performed for the water quality parameters of the lower catchment showed that the pH of water and COD were the most appropriate physicochemical features for determining the water quality status of the catchment. The pH or hydrogen ion concentration of water can be considered as a most suitable indicator of water quality as it influences the majority of chemical reactions that takes place in aquatic medium and determines the structure of aquatic biological communities (Amić & Tadić 2018). Furthermore, the pH of natural water is closely associated with its temperature, solids in solution, and its carbondioxide tension, and a change in these parameters causes a definite and predictable change in pH of the water (Powers 1930). High pH is known to be less common in a water body than low pH, and pH values less than 6.0 or 6.5 are responsible for a number of biological effects in aquatic organisms, reduction in species numbers and replacement of acid-sensitive species with acid tolerant species (Dirisu et al. 2016). Accordingly, the pH value has been estimated for the Kelani River in many previous evaluations, and locations with unpermissible levels have been recorded. These recordings show that downstream locations have more unsuitable levels of pH when compared with upstream locations (Mahagamage et al. 2016; Abeysinghe & Samarakoon 2017; Liyanage et al. 2021), a finding that was also revealed in our study. However, when considering the specific locations with high and low levels of pH, certain conflicting observations were encountered between previous studies and the current investigation. In a previous study, Mahagamage et al. (2016) recorded high pH values for Aliwaththa, Mattakkuliya (7.89) while the present study found the lowest value at Mattakkuliya (5.33). Furthermore, Liyanage et al. (2021) recorded the lowest pH from Kaduwela sampling location (5.72), which in the present study yielded the highest pH (8.43). However, Abeysinghe & Samarakoon (2017) revealed the highest pH values for Kaduwela (6.68) with regard to other downstream locations. Regardless of these differences in the present and previous studies, it is somehow evident that most of the downstream locations have a lower pH that is unsuitable for aquatic life. Furthermore, this lower pH has been associated with the abundance and composition of freshwater fish species of the lower catchment, and a strongly positive correlation has been revealed between water pH and frequency of Pethia reval. According to the study, P. reval has been described as a pH-tolerant species, and the possibility of using the fish species as an indicator of downstream water quality has been assessed (Narangoda et al. 2021, 2022). The association of pH in this analysis signifies the importance of the feature in determining water quality and supports the appropriateness of using pH as a determinant for water quality of the lower catchment.

The COD was also selected as an appropriate measure for determining the quality of water in the lower catchment. The COD indicates the organic matter content in water and is a measurement of the oxygen required for microorganisms to carry out biological decomposition of organic matter (Abba & Elkiran 2017; Baharvand & Daneshvar 2019). Higher levels denote higher oxidation of organic compounds, which will eventually reduce the DO levels of water leading to anaerobic conditions that will be harmful to aquatic life (Abba & Elkiran 2017). This was clearly seen in the current study, where location L4 (Kolonnawa) expressed the highest COD (81.1 mg/l) and the lowest DO (3.65 mg/l) content. High COD levels at location L4 are quite apparent as Kolonnawa is known to be affected by solid waste disposal by illegal human settlements (Ranasinghe et al. 2016). In addition to location L4, most of the locations of the lower catchment had high COD levels which ranged from 14.27 to 81.10 mg/l, which is exceptionally high when considering the accepted level of 15.0 mg/l. This reveals that organic pollution is currently significantly high in the lower catchment, a fact that is also obvious when considering the COD levels of the Kelani River, which has been recorded several years ago. In the year 2000, Ileperuma recorded a COD level ranging from 1.5 to 2.0 mg/l for downstream locations and attributed the high levels in certain locations to garbage dumping and discharging of organic dye wastes by textile factories located downstream of the river. However, these COD levels were well below the maximum permissible levels, and the water was acceptable for drinking purposes and aquatic life. In 2016, levels ranging from 1.33 to 307.28 mg/l were recorded for the entire river, and the study reported more high values for the upper catchment locations (Mahagamage et al. 2016). However, by 2020, significantly high COD levels were again recorded from downstream locations such as 20–250 mg/l by Kuruppuarachchi & Pathiratne (2020), and 20–208 mg/l by Ruvinda & Pathiratne (2020). Eventhough the present study yielded a much lower value for COD (14.27–81.10 mg/l), it is evident that the waters of the lower catchment of the river contain a fairly high amount of organic waste. Therefore, the selection of COD to predict the quality of water in the lower catchment is reasonable and acceptable. Furthermore, when comparing with the BOD that also measures the organic pollution in water, COD measurements can be made in a few hours while BOD measurements usually take 5 days. The COD value is usually higher than the BOD because some organic materials in the water that are resistant to microbial oxidation and hence not involved in BOD could be easily chemically oxidized (Aniyikaiye et al. 2019).

Water quality assessment in the upper catchment

The upper catchment or headstream is the source of the river and is located the farthest distance from the river's end. Depending on the river structure between mainstreams and branches, the elevation of the upper catchment can differ, but in most cases is of high altitude. Therefore, water pollution indices for rivers in many countries are known to have more serious pollution toward the river downstream. Upstream pollution is known to affect downstream health in Indonesia mainly through upstream bathing, trashing, transportation, irrigation and industry (Garg et al. 2018). Furthermore, flood events in high mountain catchments, and land clearance and cropping on hillslopes can deteriorate the downstream water by movements and erosion of suspended sediment loads (Woodward & Foster 1997). These factors combined with increased urbanization and industrialization associated with downstream locations have in most cases contributed to lower catchments with increased pollution, as was the case for the Kelani River in the present investigation. However, assessing the water quality of the upper catchment of a river is of exceptional importance, as headwater resources of many rivers are widely used for drinking water supply and cropland irrigation, and if contaminated poses a serious risk to public health (Zhao et al. 2020). The present study showed that the nitrate content of water and COD were the most appropriate features for determining the water quality status of the upper catchment of the Kelani River.

The determination of nitrate levels in surface waters is an integral part of basic water quality assessment because its concentration is generally an indicator of the nutrient status and the degree of organic pollution of the water body. Regular monitoring of nitrates in drinking water is recommended because of the potential health risks associated with its elevated levels, especially for infants who are <6 months old and animals. The major source of accumulated nitrates in water bodies are nitrate-based fertilizers or inadequately treated or untreated sewage (Maghanga et al. 2013). Since the 1950s, the use of nitrogen fertilizers has increased by several folds (Ward 2009), and nitrate has been one of the dominant forms of increased nitrogen loading since the 1970s (Xue et al. 2016). Human exposure is mainly via ingestion of contaminated drinking water, and many health effects including cancers, diabetes, thyroid conditions and adverse reproductive outcomes have been associated with ingesting nitrate-contaminated drinking water. Nitrogen fertilizers are used widely on a large number of crops to increase productivity and are the most commonly used fertilizers in tea plantations. Nitrates within the fertilizers are very mobile and loosely bound in the soil and hence are easily leached via surface runoff into rivers passing through the tea plantations (Maghanga et al. 2013).

The upper catchment of the Kelani River consists of rural and estate communities that engage in tea, rubber and paddy cultivation. The upper catchment is rich in several other crops as well, and application of fertilizers, pesticides and herbicides is practiced (Mahagamage & Manage 2018). Nitrogen concentrations are high in fertilizers, especially those which are applied to tea cultivations, and studies have recorded various pesticides containing nitrogen from the upper catchment of the Kelani River (Mahagamage & Manage 2018). High levels of ammonia exceeding the maximum tolerance limits have been recorded for the tributaries of the Maussakelle reservoir of the upper catchment (Nandasena et al. 2019), and ammonia is known to be readily converted to nitrates by microorganisms conducting ammonia oxidation (Barth et al. 2020). Nitrates are also present in domestic sewage and urban garbage (Xue et al. 2016) and thus can accumulate in waters via improper domestic waste management practices. Therefore, it is highly possible that the nitrate concentration in water can be used to predict the quality of water in the upper catchment of the Kelani River.

Furthermore, the study reveals that the COD concentration can be used to predict the quality of water in the upper catchment as in the lower catchment. COD levels of the upper catchment are comparatively low when compared with that of the lower catchment. However, high values have been recorded in the present study (7.67–25.33 mg/l) when compared with the most recent study by Ruvinda & Pathiratne (2020) for the upper catchment of the Kelani River (1–8 mg/l).

The water quality parameters were determined for the upper and lower catchments of the Kelani River Basin, and the pollution of the river was assessed. The lower catchment of the river was more polluted than the upper catchment with respect to many water quality parameters such as pH (low), DO (low), turbidity (high), TSS (high), BOD (high), nitrite concentration (high), ammonia concentration (high), phosphate concentration (high) and COD (high). The lower catchment of the river at locations near Mattakkuliya, Kolonnawa and Kaduwela was highly polluted with respect to many water quality parameters, and most features were not within the permissible levels. However, the COD was significantly high in all locations of the lower catchment with the exception of two locations, and the pH value was low for most. Therefore, COD and pH values were selected statistically as the most suitable water quality parameters for assessing the pollution of the lower catchment. The upper catchment of the Kelani River was less polluted than the lower catchment, and two water quality parameters, COD and nitrate concentration, were statistically considered as suitable for determining its pollution status. The pollution of the upper catchment of the river is mainly attributed to nitrogen-containing fertilizers applied to tea cultivations while the pollution of the lower catchment is attributed to domestic and industrial sewage. This pollution of the Kelani River is a critical problem when considering the importance of the river to the human population, aquatic life and other biological communities. Therefore, the country has currently taken steps to maintain suitable water quality within the river by improving septic systems and sewage disposal methods, building awareness on the impact of pollution of Kelani River on the environment via workshops and training programs, monitoring and issuing licenses to industries along the banks of Kelani River by the Ministry of Environment, and strengthening the monitoring of the river by increasing the number of sampling sites and automation of the monitoring process.

The research was funded by the National Aquatic Resources Research & Development Agency (NARA), Ministry of Fisheries and Aquatic Resources Development, Sri Lanka. We are also grateful to the Environmental Studies Division, NARA for providing research facilities and support in fieldwork.

Project funding was provided by the National Aquatic Resources Research & Development Agency (NARA), Ministry of Fisheries and Aquatic Resources Development, Sri Lanka.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.

Abba
S. I.
&
Elkiran
G.
2017
Effluent prediction of chemical oxygen demand from the wastewater treatment plant using artificial neural network application
.
Procedia Computer Science
120
,
156
163
.
https://doi:10.1016/j.procs.2017.11.223
.
Abeygunawardane
A. W. G. N.
,
Dayawansa
N. D. K.
&
Pathmarajah
S.
2011
Socioeconomic implications of water pollution in an urban environment a case study in Meda Ela catchment, Kandy, Sri Lanka
.
Tropical Agricultural Research
22
(
4
),
374
383
.
Abeysinghe
N.
&
Samarakoon
M. B.
2017
Analysis of variation of water quality in Kelani River, Sri Lanka
.
International Journal of Environment, Agriculture and Biotechnology
2
(
6
),
2770
2775
.
http://dx.doi.org/10.22161/ijeab/2.6.1
.
Aniyikaiye
T. E.
,
Oluseyi
T.
,
Odiyo
J. O.
&
Edokpayi
J. N.
2019
Physico-chemical analysis of wastewater discharge from selected paint industries in Lagos, Nigeria
.
International Journal of Environmental Research and Public Health
16
,
1235
.
https://doi:10.3390/ijerph16071235
.
APHA/AWWA/WEF
.
2012
Standard Methods for the Examination of Water and Wastewater
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Baharvand
S.
&
Daneshvar
M. R. M.
2019
Impact assessment of treating wastewater on the physiochemical variables of environment: a case of Kermanshah wastewater treatment plant in Iran
.
Environmental Systems Research
8
(
18
).
https://doi.org/10.1186/s40068-019-0146-0
.
Bandara
J. M. R. S.
,
Wijewardena
H. V. P.
,
Bandara
Y. M. A. Y.
,
Jayasooriya
R. G. P. T.
&
Rajapaksha
H.
2011
Pollution of river Mahaweli and farmlands under irrigation by cadmium from agricultural inputs leading to a chronic renal failure epidemic among farmers in NCP, Sri Lanka
.
Environmental Geochemistry and Health
33
,
439
453
.
https://doi.org/10.1007/s10653-010-9344-4
.
Barth
G.
,
Otto
R.
,
Almeida
R. F.
,
Cardoso
E. J. B. N.
,
Cantarella
H.
&
Vitti
G. C.
2020
Conversion of ammonium to nitrate and abundance of ammonium-oxidizing-microorganism in tropical soils with nitrification inhibitor
.
Scientia Agricola
77
(
4
).
http://dx.doi.org/10.1590/1678-992X-2018-0370
.
Central Environmental Authority (CEA)
1992
The Proposed Water Quality Standards for Fish and Aquatic Life
. Government of Sri Lanka, Ministry of Environment and Parliamentary Affairs,
Sri Lanka
.
Chandimala
J.
&
Zubair
L.
2007
Predictability of stream flow and rainfall based on ENSO for water resources management in Sri Lanka
.
Journal of Hydrology
335
,
303
312
.
https://doi:10.1016/j.jhydrol.2006.11.024
.
De Silva
M. M. G. T.
,
Weerakoon
S. B.
,
Herath
S.
,
Ratnayake
U. R.
&
Mahanama
S.
2012
Flood inundation mapping along the lower reach of Kelani river basin under the impact of climatic change
.
Engineer
45
(
2
),
23
29
.
Dirisu
C. G.
,
Mafiana
M. O.
,
Dirisu
G. B.
&
Amodu
R.
2016
Level of pH in drinking water of an oil and gas producing community and perceived biological and health implications
.
European Journal of Basic and Applied Sciences
3
(
3
),
53
60
.
Dybern
B.
1974
Water pollution: a problem with global dimensions
.
Ambio
3
(
3/4
),
139
145
.
Edussuriya
C. H.
&
Pathirage
S.
2016
Map Compendium of the Kelani River Basin
.
Central Environment Authority and International Union for Conservation of Nature Sri Lanka Country Office
,
Colombo
,
Sri Lanka
.
Eriyagama
N.
,
Smakhtin
V.
&
Jinapala
K.
2015
The Sri Lanka environmental flow calculator: a science-based tool to support sustainable national water management
.
Water Policy
,
1
13
.
https://doi:10.2166/wp.2015.158
.
Fayas
C. M.
,
Abeysingha
N. S.
,
Nirmanee
K. G. S.
,
Samaratunga
D.
&
Mallawatantri
A.
2019
Soil loss estimation using rusle model to prioritize erosion control in Kelani river basin in Sri Lanka
.
International Soil and Water Conservation Research
7
,
130
137
.
https://doi.org/10.1016/j.iswcr.2019.01.003
.
Garg
T.
,
Hamilton
S. E.
,
Hochard
J. P.
,
Kresch
E. P.
&
Talbot
J.
2018
(Not so) gently down the stream: river pollution and health in Indonesia
.
Journal of Environmental Economics and Management
92
,
35
53
.
https://doi.org/10.1016/j.jeem.2018.08.011
.
Gebeyehu
A. E.
,
Chunju
Z.
,
Yihong
Z.
&
Pingale
S.
2018
Overview of prominent problems in huai river basin, China
.
International Journal of Hydrology
2
(
1
),
9
12
.
https://doi:10.15406/ijh.2018.02.00041
.
Ileperuma
O. A.
2000
Environmental pollution in Sri Lanka: a review
.
Journal of the National Science Foundation of Sri Lanka
28
(
4
),
301
325
.
Jolliffe
I. T.
&
Cadima
J.
2016
Principal component analysis: a review and recent developments
.
Philosophical Transactions of the Royal Society A
374
.
http://dx.doi.org/10.1098/rsta.2015.0202
.
Kalaivani
S.
,
Sivakumar
K.
&
Balamuralitharan
S.
2020
Higher order principal component analysis of eigen values with special structures covariance matrices
. In
Proceedings of the AIP Conference 2277
,
United States
.
https://doi.org/10.1063/5.0025241
.
Katupotha
J.
&
Gamage
S.
2020
Understanding the river basin classification of Sri Lanka
.
WildLanka
8
(
4
),
175
197
.
Kumar
M.
,
Chaminda
G. G. T.
&
Honda
R.
2020
Seasonality impels the antibiotic resistance in Kelani River of the emerging economy of Sri Lanka
.
npj Clean Water
3
,
12
.
https://doi.org/10.1038/s41545-020-0058-6
.
Kuruppuarachchi
H. D.
&
Pathiratne
A.
2020
Toxic hazards of industrial waste receiving canal system in the lower catchment of Kelani River basin, Sri Lanka
.
Journal of the National Science Foundation of Sri Lanka
48
(
1
),
37
47
.
http://dx.doi.org/10.4038/jnsfsr.v48i1.9932
.
Li
P.
,
Karunanidhi
D.
,
Subramani
T.
&
Srinivasamoorthy
K.
2021
Sources and consequences of groundwater contamination
.
Archives of Environmental Contamination and Toxicology
80
,
1
10
.
https://doi.org/10.1007/s00244-020-00805-z
.
Lintern
A.
,
Webb
J. A.
,
Ryu
D.
,
Liu
S.
,
Bende-Michl
U.
,
Waters
D.
,
Leahy
P.
,
Wilson
P.
&
Western
A. W.
2018
Key factors influencing differences in stream water quality across space
.
WIREs Water
5
,
1
31
.
https://doi:10.1002/wat2.1260
.
Liyanage
G. Y.
,
Illango
A.
&
Manage
P. M.
2021
Prevalence and quantitative analysis of antibiotic resistance genes (ARGs) in surface and groundwater in meandering part of the Kelani River Basin in Sri Lanka
.
Water, Air, & Soil Pollution
232
,
351
.
https://doi.org/10.1007/s11270-021-05300-2
.
Maghanga
J. K.
,
Kituyi
J. L.
,
Kisinyo
P. O.
&
Ng'etich
W. K.
2013
Impact of nitrogen fertilizer applications on surface water nitrate levels within a Kenyan tea plantation
.
Journal of Chemistry
,
1
4
.
http://dx.doi.org/10.1155/2013/196516
.
Mahagamage
M. G. Y. L.
&
Manage
P. M.
2014
Water quality index (CCME-WQI) based assessment study of water quality in Kelani river basin, Sri Lanka
. In:
Proceedings of the First Environment and Natural Resources International Conference
,
Bangkok, Thailand
.
Mahagamage
M. G. Y. L.
&
Manage
P. M.
2018
Socio-economic background of the head and transitional regions of the Kelani River Basin, Sri Lanka
.
Asian Journal of Microbiology, Biotechnology and Environmental Sciences
20
(
3
),
744
756
.
Mahagamage
M. G. Y. L.
,
Chinthaka
S. D. M.
&
Manage
P. M.
2016
Assessment of water quality index for groundwater in the Kelani River Basin, Sri Lanka
.
International Journal of Agriculture and Environmental Research
2
(
5
),
1158
1171
.
Mahagamage
M. G. Y. L.
,
Pathirage
M. V. S. C.
&
Manage
P. M.
2020
Contamination status of Salmonella spp., Shigella spp. and Campylobacter spp. in surface and groundwater of the Kelani River Basin, Sri Lanka
.
Water
12
(
8
),
2187
2204
.
https://doi:10.3390/w12082187
.
Mertler
C. A.
&
Reinhart
R. V.
2016
Advanced and Multivariate Statistical Methods
.
Routledge
,
New York, NY
.
Nandasena
W. D. T. N.
,
Amarathunga
A. A. D.
,
Narangoda
S. R. C. N. K.
&
Jayasinghe
G. Y.
2019
Pollutant levels in upper catchment of the Kelani River Basin: A case study in Maussakelle Reservoir
. In:
Proceedings of the International Symposium on Water and Air Pollution: Recent Trends in Research
.
University of Peradeniya
,
Sri Lanka
, p.
21
.
Narangoda
S. R. C. N. K.
,
Amarathunga
A. A. D.
&
Dangalle
C. D.
2021
The association of fish abundance and composition with water quality parameters in the lower Kelani river basin, Sri Lanka
. In:
Proceedings of the 41st Annual Sessions of the Institute of Biology
,
Sri Lanka
, pp.
25
.
Narangoda
C. N. K.
,
Dangalle
C. D.
&
Amarathunga
D.
2022
Selected freshwater fish species for assessing the water quality of the lower catchment of the Kelani river, Sri Lanka
.
Environmental Monitoring and Assessment
194
(
9
).
https://doi.org/10.1007/s10661-022-10319-x
.
Osborne
J. W.
2015
What is rotating in exploratory factor analysis?
Practical Assessment, Research, and Evaluation
20
(
2
),
1
7
.
https://doi.org/10.7275/hb2g-m060
.
Parsad
R.
&
Bhar
L. M.
1987
Multivariate analysis of variance
.
Journal of Counseling Psychology
34
,
1
12
.
Powers
E. B.
1930
The relation between pH and aquatic animals
.
The American Naturalist
64
(
693
),
342
366
.
Ranasinghe
G.
,
Piyadasa
R. U. K.
&
Mataraarachchi
S.
2016
A study to identify the soil characteristics of wetlands in Colombo, Sri Lanka
. In:
Proceedings of the International Research Symposium of SAITM Malabe
,
Sri Lanka
, pp.
60
63
.
Ranmadugala
B. H. S.
,
Wijeratne
E. M. S.
&
Pitawala
A.
2007
Modelling of saltwater intrusion in Kelani river estuary
. In
Proceedings of the 25th Anniversary Scientific Conference of NARA on Tropical Aquatic Research Towards Sustainable Development
,
Colombo, Sri Lanka
, pp.
55
.
Ruvinda
K. M. S.
&
Pathiratne
A.
2020
Use of liver histological alterations and erythrocytic nuclear abnormalities of two native fish species in Kelani River, Sri Lanka as biomarkers for pollution impact assessments
.
Journal of the National Science Foundation of Sri Lanka
48
(
1
),
15
26
.
http://dx.doi.org/10.4038/jnsfsr.v48i1.9937
.
Sikder
M. T.
,
Kihara
Y.
,
Yasuda
M.
,
Yustiawati
M. Y.
,
Tanaka
S.
,
Odgerel
D.
,
Mijiddorg
B.
,
Syawal
S. M.
,
Hosokawa
T.
,
Saito
T.
&
Kurasaki
M.
2013
River water pollution in developed and developing countries: judge and assessment of physicochemical characteristics and selected dissolved metal concentration
.
Clean: Soil, air, Water
41
,
60
68
.
Surasinghe
T.
,
Kariyawasam
R.
,
Sudasinghe
H.
&
Karunarathna
S.
2020
Challenges in biodiversity conservation in a highly modified tropical river basin in Sri Lanka
.
Water
12
(
26
),
1
23
.
https://doi:10.3390/w12010026
.
Thotagamuwa
H. T. B. N.
&
Weerasinghe
V. P. A.
2021
Surface water quality assessment for the management of hydrological regimes: Kalu Oya and Mudun Ela catchment in Sri Lanka
.
Environmental Nanotechnology, Monitoring & Management
15
.
https://doi.org/10.1016/j.enmm.2020.100402
.
Wickramasinghe
W. M. S.
,
Navaratne
C. M.
&
Dias
S. V.
2018
Building resilience on water quality management through grey water footprint approach: a case study from Sri Lanka
.
Procedia Engineering
212
,
752
759
.
https://doi:10.1016/j.proeng.2018.01.097
.
Woodward
J.
&
Foster
I.
1997
Erosion and suspended sediment transfer in river catchments: environmental controls, processes and problems
.
Geography
82
(
4
),
353
376
.
Xue
Y.
,
Song
J.
,
Zhang
Y.
,
Kong
F.
,
Wen
M.
&
Zhang
G.
2016
Nitrate pollution and preliminary source identification of surface water in a semi-arid river basin, using isotopic and hydrochemical approaches
.
Water
8
(
328
).
https://doi:10.3390/w8080328
.
Zhao
K.
,
Wu
H.
,
Chen
W.
,
Sun
W.
,
Zhang
H.
,
Duan
W.
,
Chen
W.
&
He
B.
2020
Impacts of landscapes on water quality in a typical headwater catchment, Southeastern China
.
Sustainability
12
,
721
.
https://doi:10.3390/su12020721
.
Zoysa
D. S. S.
&
Weerasinghe
T. K.
2016
Study on the chemical and biological status of Malwatu Oya Basin, Anuradhapura in Sri Lanka
.
British Journal of Environmental Sciences
4
(
1
),
1
10
.
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/).