Water Quality Assessment using Environmetrics and Pollution Indices in A Tropical River, Kerala, SW Coast of India

Envirometrics and pollution indices are proxies to assess water quality of a wetland ecosystem. Hence, the present study is focused on establishing water quality and elucidating the pollution status of Karamana River (KR) in Kerala, SW coast of India. The Karamana River Basin – KRB (n=6th; L= 68 km, A=695 km2), is the main source of water for domestic and drinking purpose in Thiruvananthapuram city. The Killi River (n= 5th; L= 24 km, A= 102 km2), the largest tributary of KR, carry heavy load of pollutants mainly from city and joins KR towards its downstream side. For this study, about 12 sampling stations were selected along the KR from upstream to downstream (interval= ~3km), and water samples (n=12x2= 24) were collected during non-monsoon (NON) and monsoon (MON) of 2015 to assess the variability and sourcing of key hydrochemical variables. Environmetric methods, viz., Pearson Correlation and Principal Component Analysis (PCA) were applied for apportionment of pollution sources significantly responsible for the surface water quality. It was found that sewage effluents and seawater intrusion were the primary factors deteriorating water quality in downstream. Further, the results of water quality analyses and Pollution Indices, viz., Organic Pollution Index (OPI), Eutrophication Index (EI) and Comprehensive Pollution Index (CPI) indicate that lower reaches (L= ~4 km) of KR is seriously polluted. A distinct Zone of Pollution Influence (ZPI) has been delineated based on the indices and this attempt is first of its kind in KR. The present study provides several noteworthy contributions to the existing knowledge on the factors influencing surface water quality and serves as a baseline data for watershed managers and administrators. Current World Environment www.cwejournal.org ISSN: 0973-4929, Vol. 15, No. (1) 2020, Pg. 11-23


Introduction
Over the years, rivers are progressively subjected to environmental stress caused by human intervention. 1 Water quality of a river is being carved up by several interrelated factors, which are subjected local and temporal variations in addition to discharge. 2 As a result of increased anthropogenic pressure including pollution, large-scale damming, sand mining etc., the natural self purification capacity of rivers becomes restricted to a minimal level. 3 Excessive loading of nutrients into the riverine environment in the form of eutrophication turns out to be a threat to human health as well as other biota in freshwater ecosystem. 28,29 Consequently, understanding the extent of water quality degradation and sources of pollution is important for an effective management of water resources.
Conventional techniques of descriptive analysis to interpret river water quality have several limitations 4 of not detecting correlation between variables and poor delineation in the pollution source apportionments influencing the quality status. The use of environmetric techniques has potentialities to surpass these limitations.
Environmetrics are Multivariate statistical techniques applied in environmental problems. Application of environmetrics, i.e., multivariate statistical techniques, is an important tool for environmental decision making in water quality problems.4 Correlation Analysis and Principal Component Analysis (PA) have been applied successfully in scrutinizing latent factors and mechanisms influencing water quality worldwide. 5,6,7,8,9,10,11 Further, pollution indices provide supplementary information on water quality in a single value by comparing different variables as per the standards. A plethora of various water quality indices is available in literature 12 and based on the purpose of assessment, water quality indices were formulated worldwide viz., Harkins Index, 13 Horton Quality Index, 14 Water Quality Index, 15 Agricultural Water Quality Index 16 etc. Again, to assess the grade of pollution, certain specific pollution indices, viz., Carlson's Trophic State Index, 17 Organic Pollution Index, 18 Comprehensive Pollution Index, 19 Eutrophication Index 20 have been commonly applied in numerous studies. 21,22,23,24,25 Karamana River (KR) is the main resource to meet the domestic and drinking purpose of Thiruvananthapuram city. This tropical river is suffering from invariable pressure of fast growing urbanization analogous to other overpopulated river basins globally 26,27,28 and the available data on water quality of this river is very sparse. Hence, monitoring the water quality of this river is very relevant in the current scenario. It is against this backdrop, the present study was carried out to provide an overview of pollution status and its underlying sources in KR using environmetric techniques viz., Pearson Correlation analysis and Principal Component Analysis coupled with pollution indices viz., Organic Pollution Index (OPI), Eutrophication Index (EI) and Comprehensive Pollution Index (CPI).

Study Area
The Karamana river basin, KRB (n= 6 th ; L= 68 km; A = 702 km 2 ), is one of the prominent river basins in southern Kerala, SW India. The basin lies between latitudes 8°27'36"N to 8°38'24"N and longitudes 76°54'0" to 77°15'0"E. The Karamana river (KR) originates from Chemmunji Motta and Aathiramala peaks of Western Ghats, and flows westward and debouch into the Arabian sea at Poonthura (estuary), SW of Thiruvananthapuram. The Killiyar (n= 5 th ; L= 24 km, A= 102 km 2 ) merges with KR ~3 km upstream of Poonthura estuary (Fig.1). TS canal (Parvathy Puthanar Canal), running parallel to the coast with untreated sewage effluents rushes in, further pollutes the lower reaches of KR. The average annual stream flow of KR is calculated to be 836Mm. 3.30

Methodology
The sampling was carried out during non monsoon-NON (March) and monsoon-MON (June) seasons in 2015. A total of 12 sampling stations were identified from upstream to downstream (interval = 3 km) and physico-chemical parameters were analyzed using standard procedures. 31,32 Correlation and factor analyses were done using SPSS 17 software. The Spearman correlation was calculated by applying the Pearson correlation formula to the ranks of the data.
Pearson's correlation analysis (r) is a measure of the extent to which two quantitative variables are linearly related. It summarizes the magnitude of a linear relationship between pairs of variables. The value of relationship takes values ranging from -1 to +1, where +1 represents an absolute perfect positive linear relationship, 0 represents no linear relationship, whereas -1 represents an absolute inverse relationship between the bivariates. The sign in front of the correlation coefficient value determines the direction of the relationship. A plus sign denotes a positive relationship and a minus sign denotes negative correlation. The correlation (r) provides a standardized measure of the linear association between two variables, as given in Eq.1.
where x and y are the bivariates to be correlated and Sx and Sy are the sample standard deviations of variables x and y, respectively.
PCA reduces a relatively large number of variables into a smaller set of variables that still captures the same information. 33 PCA is about extracting a set of independent linear combination of parameters of the study so as to capture the maximum amount of variability of a given dataset. PCA can be calculated using Eq.2.

Fij+fjlzi1+fj2zi2+...fjmzm+eij
Where j is the measured variable, f is the factor loading, z is the factor score, e is the residual term accounting for errors, i is the sample number, and m is the total number of factors.
Varimax rotation method was applied in factor analysis by rotating the axis defined by PCA according to well-established rules to find a simple structure of datasets. By this method, variables are obtained in which original variables are demonstrated more clearly 33 thus by achieving a simpler and meaningful representation of the underlying factors. 34 Eutrophication Index (EI) was used for evaluating the trophic condition of water body.

Results and Discussion
The results of surface water chemistry variables for Non Monsoon (NON) and Monsoon (MON) are summarized in Table 2 and 3. Spatio-temporally, pH ranged from slightly acidic to alkaline.
DO in KR were found to be less than the standard limit (6 mg/L) 35 in downstream during both seasons, which is due to the input from nutrient rich Killi river and TS Canal (Fig.1). BOD values exceeded the standard limit (2 mg/L) 35 from middle stream towards downstream (i.e., S6 to S12), which shows a strong signature of anthropogenic influence. The BOD signature reflects high loading of organic compounds enhancing microbial growth, thus by reducing the level of DO in the water. The Cl level also showed an increasing trend from station 8 downwards, exceeding permissible limit (250 mg/L) 35 for both seasons which shows marine influence. NO 3 also exceeded the standard limit (45 mg/L) 35 in the last three stations (S10-S12) in the downstream during both seasons (  3 and DIP (Fig. 2) showed an erratic trend, the DO showed a declining trend from upstream to downstream for both seasons (Fig.2).

Environmetrics
Environmetrics, also known as multivariate statistical analysing techniques, viz.,Correlation and Principal Component analyses were carried out to identify the factors influencing water quality of KR.

Correlation Analysis
Pearson correlation was examined among the major ions and other physical parameters measured.
During NON, the SO 4 showed weak correlation (0.408) and insignificant (p>> 0.05) correlation with NH 3 , significant (p < 0.05) and moderate correlation (r ≥ 0.60 to < 0.80) with NO 3 , DIN and DIP, which suggests that the source of these nutrients might be wastewater effluents discharged into the river, as leaching is not favoured during this dry period. Ca revealed strong significant relationship with Mg, Na, SO 4 and Cl, which is an indication of seawater intrusion. During MON, the DO and DIN were significantly associated with strong negative correlation (-0.920), and DIN due to a permutation of both anthropogenic as well as atmospheric inputs. NH3 was in moderate correlation (0.643) and insignificant (p >> 0.05). As Mg significantly (p < 0.05) and strongly (r ≥ 0.80) correlated with Na, K and Cl, while K was found to be correlated with Cl and SO 4 , all implies that the origin of these nutrients are most

Principal Component Analysis (PCA)
The results of PCA during NON and MON are shown in Table 4 and 5.
In order to determine the factors affecting hydrochemical regime of study area, Principal Component Analysis (PCA) was applied and the analysis during Non monsoon revealed three Principal Components (PCs) effective in explaining the variations in water quality ( The results also indicated that there is a continuum mixing between nutrient-rich sewage effluent and seawater. PC II accounts for 6.40% of total variance (Table 4). SO 4 , K and Na were found to have strong positive loading and this reflects weathering and saline water ingression. The sum of absolute contributions of parameters on PC III yields only 3.57%. Among those factors, NH 3 had the strongest loading (0.902), whereas BOD, DO and pH exhibited moderate loading, highlighting anthropogenic pollution. BOD and DO are in a moderate negative relationship.
Principal Component Analysis (PCA) during monsoon revealed three PCs effective in explaining the variations in the water quality, and these factors explained 95.56% of total variance (Table 5). PC I accounted for 85.51% of total variance (Table 5) with a strong positive loading of NH 3 , Na, K, BOD, SO 4 , COD and Cl. While DO displayed moderate negative loading; Ca, TH and Mg showed moderate positive loading. This component evinces an influence of marine spray, leaching of secondary salts and surface runoff. The sum of absolute contributions of parameters on PC II yielded only 7.24%. NO 2 and pH revealed strong positive loading, whereas; NO 3 and DIN had moderate positive loading. The relationship between nitrification and pH has been studied by many researchers. 38,39,40,41 PC III explained 5.47% of total variance. Temperature and DIP were the parameters having strong positive loading in this component (Table 5). This could have been due to the facts that, phosphate release from sediment increases with increase in temperature, as a result of mineralization. 42,43,44 The release of phosphorus at the sediment-water interface results in an increase in dissolved inorganic phosphorus in the overlying water. 43,45,46,43

Pollution Indices
The results of Organic Pollution Index (OPI), Eutrophication Index (EI) and Comprehensive Pollution Index (CPI) of Karamana River for non monsoon and monsoon 2015 are shown in Table 6 and 7 respectively. This range was found to be higher than OPI obtained in few other parts of the world. This high level of OPI in the downstream (L= 4km) is a clear indication of untreated sewage input and poor dilution capacity of river. 3 Surface water samples from the locations S1-S8 of KR ( fig. 1) were representatives of excellent to good classification based on OPI results. Compiling the results obtained from pollution indices and water quality assessment, it is evident that the final 4 km stretch of KR is severely polluted. Intriguingly, this stretch coincides with the zone identified as one of the most critically polluted Indian River stretches based on Criteria-I by CPCB. 52 From these results, a distinct stretch of ~4km in the downstream up to estuary is identified as the "Zone of Pollution Influence (ZPI)" (Fig. 3).

Conclusion
The results of study seem to provide evidence on the water quality variations in Karamana River-KR (Kerala, India). From the physico-chemical analyses of surface water samples for non-monsoon (NON) and monsoon (MON) seasons, it is interpreted that water quality is good from upstream to middle stream (L=45 km). Most of the parameters (pH, Cl, DO, BOD, NO 3 etc.) exceeded CPCB 1995 desirable limits 35 in the hindmost stretch of downstream region (L=4 km). The application of Environmetric analysis viz. Pearson Correlation Analysis and Principal Component Analysis (PCA) connote that hydrochemical attributes of KR is primarily governed by natural (weathering, atmospheric deposition, seawater intrusion) as well as anthropogenic (sewage inflow) perturbation. The above findings are corroborated by the Eutrophication Index (EI) values, and based on this; river is affected by eutrophication in the last ~ 3 km of lower reaches. Again, the results of Comprehensive Pollution Index (CPI) and Organic Pollution Index (OPI) indicate that the downstream ~ 4 km fluvial stretch, up to estuary is severely polluted. The identified zone of pollution influence (ZPI) in this river needs utmost attention by stakeholders and administrators for pollutant mitigation programmes. Application of environmetric and pollution index tools is an emergent technique in this river; consequently, very little work has been reported on relationship of physicochemical parameters and water pollution source apportionments in KR. These environmental tools provided a more objective interpretation of surface water physicochemical parameters and identification of pollution source as part of the effort toward sustainable management of this river basin.

Acknowledgement
This work is the first part of research work done by the first author during 2015 and thank university of Kerala for providing fellowship to carry out this work.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.