Environ Sci Pollut Res (2013) 20:5630–5644 DOI 10.1007/s11356-013-1542-z

RESEARCH ARTICLE

River water quality assessment using environmentric techniques: case study of Jakara River Basin Adamu Mustapha & Ahmad Zaharin Aris & Hafizan Juahir & Mohammad Firuz Ramli & Nura Umar Kura

Received: 3 November 2012 / Accepted: 3 February 2013 / Published online: 27 February 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract Jakara River Basin has been extensively studied to assess the overall water quality and to identify the major variables responsible for water quality variations in the basin. A total of 27 sampling points were selected in the riverine network of the Upper Jakara River Basin. Water samples were collected in triplicate and analyzed for physicochemical variables. Pearson product-moment correlation analysis was conducted to evaluate the relationship of water quality parameters and revealed a significant relationship between salinity, conductivity with dissolved solids (DS) and 5-day biochemical oxygen demand (BOD5), chemical oxygen demand (COD), and nitrogen in form of ammonia (NH4). Partial correlation analysis (rp) results showed that there is a strong relationship between salinity and turbidity (rp =0.930, p=0.001) and BOD5 and COD (rp =0.839, p= 0.001) controlling for the linear effects of conductivity and NH4, respectively. Principal component analysis and or factor analysis was used to investigate the origin of each water quality parameter in the Jakara Basin and identified three major factors explaining 68.11 % of the total variance in water quality. The major variations are related to anthropogenic activities (irrigation agricultural, construction activities, clearing of land, and domestic waste disposal) and natural processes (erosion of river bank and runoff). Discriminant analysis (DA) was applied on the dataset to Responsible editor: Michael Matthies Highlights: ►Surface water samples were collected and analyzed in the laboratory. ►Environmentric tools were successfully applied on the datasets. ►Physicochemical parameters’ relations were revealed. ►Water pollution source’s apportionments were identified.

A. Mustapha : A. Z. Aris (*) : H. Juahir : M. F. Ramli : N. U. Kura Environmental Forensics Research Centre, Faculty of Environmental Studies, Universiti Putra Malaysia, 43400, UPM Serdang, Selangor, Malaysia e-mail: [email protected]

maximize the similarities between group relative to withingroup variance of the parameters. DA provided better results with great discriminatory ability using eight variables (DO, BOD5, COD, SS, NH4, conductivity, salinity, and DS) as the most statistically significantly responsible for surface water quality variation in the area. The present study, however, makes several noteworthy contributions to the existing knowledge on the spatial variations of surface water quality and is believed to serve as a baseline data for further studies. Future research should therefore concentrate on the investigation of temporal variations of water quality in the basin. Keywords Discriminant analysis . Jakara Basin . Partial correlation . Principal component analysis . Water quality variation

Introduction River basin has been a major source of water supply for many purposes and provides fertile lands, which support the development of highly populated residential areas due to its favorable conditions (Mouri et al. 2011). Human settlements and industries have long been concentrated along rivers, estuaries, and coastal zones owing to the predominance of water-borne trade. A river’s water quality is the composite of several interrelated compounds, which are subjected to local and temporal variations and also affected by the volume of water flow (Mandal et al. 2010). Rivers constitute the main inland water body for domestic, industrial, and agricultural activities and often carry large municipal sewage, industrial wastewater discharges, and seasonal runoff from an agricultural field (Singh et al. 2004; Pradhan et al. 2009; Hu et al. 2011). The river waters have been contaminated as a result of the discharges of wastewater containing degradable organics, nutrients, domestic effluent, and agricultural waste (Dimitrovska et al. 2012).

Environ Sci Pollut Res (2013) 20:5630–5644

River water pollution can be linked to the type of wastewater produced by urban, industrial, and agricultural activities that flows into surface and subsurface waters (Vittori et al. 2010). The increase in human population and economic activities has grown in scale; the demands for large-scale suppliers of fresh water from various competing end users have increased tremendously. The decline in the quality and quantity of surface water resources can be attributed to water pollution and the improper management of the resource (Mustapha and Nabegu 2011). Many regions around the world are simultaneously impacted by urbanization processes and industrial and agricultural activities, and many cities in developing countries have been developed without adequate and proper planning. This has led to indiscriminate actions, including dumping of wastes into the water and washing and bathing in open surface water bodies (Cukrov et al. 2012). The deteriorating water quality affects man, animal, and plant life with far-reaching consequences. From the environmental, economical, and/or social point of view, it is important to identify these sources and their contribution to the total contamination of an area (Tobiszewski et al. 2010). In recent years, there has been increasing awareness of, and concern about, surface water pollution all over the world, and new approaches toward the sources of pollutants and achieving sustainable exploitation of water resources have been developed. The combined use of environmentric tools such as multivariate statistical techniques enables the classification of water samples into distinct groups, source apportionments, relationship, and differences in the parameters used based on hydrochemical characteristics (Shrestha et al. 2008). They reflect more accurately the multivariate nature of the natural ecosystem, which provides a way to handle large datasets with a large number of parameters by summarizing the redundancy and provides a means of detecting and quantifying truly multivariate patterns of the datasets (McGarigal et al. 2000). The use of conventional techniques of descriptive analysis to interpret surface water quality has several limitations of not detecting the long-term correlation between variables and poor delineation in the source apportionments of the surface water quality variation. The use of environmetric techniques has several advantages to overcome these limitations. Principal component analysis (PCA) and factor analysis (FA) have been utilized by various researchers to explore the pollution sources of river water: for example, PCA/FA techniques have been applied by Han et al. (2009) in their studies on the Nakdong River watershed; they used PCA/FA to identify pollution sources in the study area and discovered that anthropogenic pollutants are responsible for the high variation in the water quality of the studied area. Equally, Wong (2005) studied the spatial variability of physiochemical elements in Hong Kong River. PCA/FA was used to identify latent factors or pollution sources; it

5631

yielded four components in which the different watercourses studied fall under nutrient and organic pollutants and heavy metal contamination; other water courses suffered multiple types of pollution or have a low pollution problem. The regional distributions of physiochemical determinants verified that marine and anthropogenic sources have a strong influence to rivers/streams in Hong Kong. Similarly, Tanriverdi et al. (2010) applied PCA/FA in the surface water quality data to assess and examine the water quality of Ceyhan River. Three PCs were significantly identified corresponding to areas close to cities, which presented low dissolved oxygen contents and high concentrations of physiochemical parameters, suggesting anthropogenic inputs. The stations in the vicinity of industries have higher pollution due to the discharge of wastewater from industries and domestic activities. Various studies conducted used discriminant analysis (DA) to bring out the most significant variables that result in water quality variation and to optimize the monitoring program in the future by decreasing the monitoring frequency and the number of parameters monitored and, thus, the subsequent cost (Li et al. 2009; Koklu et al. 2010; Schaefer and Einax 2010). For example, Singh et al. (2004) studied wastewater pollutants in Lucknow, Uttar Pradesh, India. Spatial DA was performed with a 2-year raw dataset of wastewater which comprises 29 parameters. The standard DA mode constructs discriminant functions including all the parameters under study, while forward stepwise reduced the parameters to six discriminant variables with 70 % cases correctly. Similarly, Singh et al. (2005) reported that DA showed the best results for data reduction and pattern recognition for both temporal and spatial evaluations in Gomti River, India. It revealed five parameters affording more than 94 % for temporal variation and ten parameters affording 97 % correct assignation in the spatial evaluation of three different regions of Gomti River, India. Zhou et al. (2007) revealed that DA provided better results with great discriminatory ability for both temporal and spatial analyses, and it reduced the parameters under study by using only six parameters affording about 84 % correct assignation and seven parameters with more than 90 % correct assignation in temporal and spatial analyses. Moreover, Papaioannou et al. (2010) used DA to construct the best discriminant functions and to confirm the clusters determined by cluster analysis. The standard and forward stepwise modes of discriminant functions used 17 and 10 parameters with a classification matrix correctly assigning 96.97 and 96.36 % of the cases, respectively. This study concluded that, DA proven to be a useful tool in recognizing the discriminant parameters in spatial variation of portable water quality. The relationship between the sampling stations and water quality variables was studied by Varol et al. (2012) in their

5632

Environ Sci Pollut Res (2013) 20:5630–5644

studies of surface water quality in Tigris River Basin, Turkey. A high and positive correlation was observed between various parameters responsible for water mineralization and phytoplankton growth. Singh et al. (2012) reported a high positive correlation among the cations and a moderate relationship between the nutrients, which reveals the influence of silicate lithology and a similar source and/or geochemical behavior during ionic mobilization. The impact of non-point sources of pollution from surface water resource quality in Tehran province, Iran, was studied by Gholikandi et al. (2012). The researchers observed a significant association of the physicochemical parameters with a negative correlation of dissolved oxygen and temperature. It is well known that oxygen depletion in water bodies signifies a high load of organic matter. In Jakara Basin, however, the use of environmetric tools is rather an emergent; consequently, not much work has been done to evaluate the relationship of physiochemical parameters and water pollution source apportionments, and very little work has been reported about the quality of the Jakara Basin. It is against this background that this study was carried out to provide an overview of the relationship between physiochemical parameters and the possible sources of water pollution in the basin.

from the southern high-pressure belt is humid in nature and is attended with on-shore southwestern winds off and the Gulf of Guinea. The tropical continental air mass on the other hand is dry and accompanied by northeast trade winds. These two air masses control the climate of the Jakara Basin, making it the wet and dry type. The basin is typically very hot throughout the year, though in December through February is noticeably cooler. Nighttime temperatures are cool during the months of December, January, and February, with average low temperatures ranging from 11 to 14 °C (Mustapha and Aris 2012a). Relative humidity in the basin is high in August, up to 80 %, and lowest in January, November, and December, with 23 %. The rainy season has a moderate effect on temperature, which falls to the lowest in August with a mean monthly value of 24.5 °C. The monthly evaporation rate in the Jakara Basin varied from 55 % in March and April at the beginning of the wet season to 78 % in September during the end of the rainy season.

Description of the study area

A total of 27 sampling sites were selected in the riverine network of the Upper Jakara Basin (Fig. 1). Surface water samples were collected in triplicate at a depth of 10–15 cm with 1-L plastic containers, which were pre-rinsed with trioxonitrate(v) and soaked overnight with distilled water to avoid an unpredicted change in the characteristic of the water samples. The samples were then placed in a box containing ice packs and transported to the Soil and Water Laboratory of the Kano State Ministry of Environment and kept at a temperature of about 4 °C prior to the analyses. In order to provide greater data confidence from the analytical procedure regarding bias and variability, appropriate quality assurance and quality control (QA/QC) on water samples were ensured. The QA/QC were followed to ensure that data products are of documented high quality and are reproducible. The overall data quality is assessed through precision, accuracy, representativeness, completeness, sensitivity, and comparability (Aris et al. 2012). Analytical precision was assessed by the use of a control chart and blind samples. The same laboratory equipment were used for all the samples collected in order to control variability from sampling irregularities. Equipment blank was used to test for bias from possible contamination of blank water which consists of distilled water. Internal blind samples were evaluated to describe differences between the filtered and unfiltered samples. The relative standard deviation and percent recovery were calculated to demonstrate precision and accuracy.

The study area (Jakara River Basin) extends from longitude 8°31′ E to 8°45′ E and latitude 12°10′ N and 12°13′ N. It is located about 481 m (1,580 ft) above mean sea level, covering an area of about 150 km2 (Mustapha et al. 2012b). The basin is one of the most significant in supporting residential, industrial, and agricultural activities in Kano metropolitan with a population of more than 500,000 inhabitants. The climate of the basin is strongly influenced by the tropical continental and maritime air masses. The two air masses control the climate, making it dry and of wet type (Mustapha et al. 2012b). The pre-Cambrian rock of the basement complex which comprises gneisses, amphibolites, marble, and older granite underlies a larger part of Nigeria, including 80 % of the study area. Jakara River Basin is underlain by a crystalline Basement complex of pre-Cambrian origin which loses its identity by disappearing into the Chad Formation. The Basement complex consists of granite rocks, which are generally gneissic and commonly developed in a mixture of pegmatite of schist granite, gneiss, and irregular masses of pegmatite. Aeolian sand derived from wind deposits covers most part of the area, with thickness of about 5 m in the upland and 10 m along the lowland plains of the basin (Mustapha and Aris 2012a). The climate of the area is strongly influenced by the tropical maritime air mass and the tropical continental air mass, like in most parts of West Africa. The tropical maritime air mass which originates

Methods Sampling and analytical techniques

Environ Sci Pollut Res (2013) 20:5630–5644

5633

Fig. 1 Jakara Basin and its land uses

Water temperature, conductivity, dissolved oxygen (DO), and pH of the river water samples were measured in situ using a multi-parameter monitoring instrument (YSI incorporated, Yellow Spring®, Ohio, USA). Salinity was measured using Eutech salinity pocket tester SaltTestr®. Five-day biochemical oxygen demand (BOD5) was measured using Winkler’s azide methods and chemical oxygen demand (COD) using a dichromate reflex technique. The instruments used in situ were calibrated using a specific calibration solution before each measurement (APHA 2005). Suspended solids (SS) and dissolved solids (DS) were separated gravimetrically: filtering the water through a 0.45-μm filter paper and determined

according to a standard procedure (APHA 2005). NH4 was determined using a molecular absorption spectrophotometer. Turbidity was directly measured with a turbidimeter (Hach 2100 AN). Trace elements (Cr, Cd, Pb) were analyzed using atomic absorption spectrometry (AAS, Perkin Elmer® 4100) with air/acetylene at a specific wavelength. In order to maintain detection precision, internal standard reference materials (SRM) were used for every ten samples. The water quality parameters, abbreviation, and analytical methods used in this study were summarized in Table 1. The absorption wavelength and conditions for the detection of trace elements used for this study are shown in Table 2.

5634

Environ Sci Pollut Res (2013) 20:5630–5644

Table 1 Water quality parameters, abbreviation and analytical methods

Parameters

Abbreviation

Unit

Analytical techniques

Temperature Conductivity pH Salinity BOD COD Dissolved Suspended solids Dissolved solids Ammonia nitrogen Turbidity Chromium Cadmium Lead

Temp Cond pH Sal BOD5 COD DS SS DS NH4 Tur Cr Cd Pb

°C μS/cm

Handheld multiparameter (YSI, 6820) Handheld multiparameter (YSI, 6820) Handheld multiparameter (YSI, 6820) Handheld multiparameter (YSI, 6820) 5-day incubation at 20 °C Potassium dichromate oxidation reflux Filtration and gravimetric Filtration and gravimetric Filtration and gravimetric Nesslerization, UV spectrophotometer Handheld turbidimeter (HACH, 2100 AN) Instrumental, digestion (AAS Perkin Elmer) Instrumental, digestion (AAS Perkin Elmer) Instrumental, digestion (AAS Perkin Elmer)

mg/L mg/L mg/L mg/L mg/L mg/L mg/L NTU mg/L mg/L mg/L

Pearson product-moment correlation analysis

Partial correlation

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.

Partial correlation (rp) is a relationship that measures the degree of association between two random variables, with the effect of a set of controlling variables. Partial correlation procedure computes a coefficient that describes the linear relationship between the bivariate while controlling for the effects of one or more additional variables. The correlation between X and Y can arise from the fact that both X and Y show a correlation with a third variable, Z; that is, after removing variance that the criterion and the predictor have in common with other predictors, the partial expresses the correlation between the residual predictor and the residual criterion. The partial correlation of X and Y, with the effects of Z removed (or held constant), is given in Eq. 2.

n P

r¼

i¼1

rp XY  Z ¼

ðxi  xÞðy  yÞ ð1Þ

ðn  1Þsx sy

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.

Table 2 Condition for detecting trace element Ion

DL

MS

SRM (μg/kg)

SR (%)

WC (r)

Cd Cr Pb

0.5 0.9 0.6

21.3 3791 2767

22.79±0.96 38.60±1.60 27.89±0.14

93.46 98.21 96.02

0.999 0.99 0.998

DL detection level, MS measurement standard, SRM standard reference materials, SR sample recovery, WC working curve

rXY  ðrXZ ÞðrYZ Þ Sqrtð1  r2 XZ Þ Sqrtð1  r2 YZ Þ

ð2Þ

The same general structure would apply for calculating the partial correlation of X and Z, and Y and Z with the effects of Y and X being removed, as in Eqs. 3 and 4, respectively. rp XZ  Y ¼

rXZ  ðrXY ÞðrYZ Þ Sqrtð1  r2 XY Þ Sqrtð1  r2 YZ Þ

ð3Þ

rp YZ  X ¼

rYZ  ðrXY ÞðrXZÞ Sqrtð1  r2 XY Þ Sqrtð1  r2 XZÞ

ð4Þ

Principal component analysis and factor analysis PCA and FA are statistical approaches that can be used to analyze interrelationships among a large number of variables and to explain these variables in terms of their common underlying dimension by providing empirical estimates of

Environ Sci Pollut Res (2013) 20:5630–5644

5635

the structure of the variables (Hair et al. 1995). This reduces a relatively large number of variables into a smaller set of variables that still captures the same information. PCA is about extracting a set of independent linear combination of the parameters of the study so as to capture the maximum amount of variability of a given dataset (Panigrahi et al. 2007). PCA/FA can be calculated using Eq. 5. fij þ fj1zi1 þ fj2zi2 þ . . . fjmzm þ eij

ð5Þ

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. FA collapses the column of the dataset to construct a smaller number of new factor or indices that are linear combinations of the original variables (Rogerson 2006). FA focuses on reducing the contribution of less significant variables to simplify even more the data structure coming from PCA (Liu et al. 2011). It is used to describe the variability among the observed variables in terms of fewer unobserved variables called factors (Esmaeili and Moore 2012). This can be achieved through the use of factor rotation by rotating the axis defined by PCA according to well-established rules. This study applies varimax rotation in which the value of PCA can be cleaned up by means of a rotation procedure of the eigenvalue. By this method, variables are obtained in which original variables participate more clearly (Helena et al. 2000). The basic motivation for using any rotational methods is to achieve a simpler and more meaningful representation of the underlying factors, producing a new group of variables known as varifactors (VFs; Cho et al. 2009). Discriminant analysis DA is a supervised pattern recognition that can be utilized for the classification of objects or cases into exhaustive and mutually exclusive groups based on a set of independent variables. It is an appropriate statistical technique when the dependent variable is a categorical variable and the independent variables are metric. The objective of DA is to maximize the similarities of the between-group relative to the within-group variance (Singh et al. 2005; Kowalkowski et al. 2006; Shrestha and Kazama 2007; Koklu et al. 2010). The classification table known as confusion or prediction matrix is used to assess the performance of the DA. This is simply a table in which the rows are the observed categories of the dependent variable; when the prediction is perfect, all cases will lie on the diagonal side of the table. The percentage of cases on the diagonal side is the percentage of correct classification. The model parameters were Wilks’ lambda, an index of the discriminating power ranging between 0 and 1 (the lower the value, the higher its discriminating power);

eigenvalue, a measure of variance in the dependent variable for each function; and canonical correlation, a measure of association between the groups formed by the dependent and the given discriminant function (Rani et al. 2011). In DA, the results of classification and their accuracy in the error matrix are determined by kappa values. The kappa coefficient (k) is calculated from the error matrix. Kappa indicates to what extent classification accuracy is due to true agreement of the field data and the classified data and to what extent it could have been achieved by chance (Manly 2005). DA is achieved by calculating the variate weight for each independent variable; the variate for the DA is known as discriminant function and is derived from the equation below. zjk ¼ a þ w1 x1k þ w2 x2k þ . . . þ wn xnk

ð6Þ

where Zjk is the Z score of the discriminant function j for object k, a is an intercept, w1 is the discriminant weight for independent variable 1, and x1k is the independent variable 1 for object k.

Results and discussion Descriptive statistics The statistical summary (mean, standard error of the mean, and standard deviation) of the selected parameters for the water samples is presented in Table 3. A total of 14 physicochemical variables were analyzed from 27 sampling points in the Jakara Basin. Water temperature varied from 25.3 °C in sampling point 1 to 31.6 °C in sampling point 9, which is within the portable range of 25–32 °C by the World Health Organization. The pH value is within the acceptable limit of 6.5–8.5, varying between 6.2 and 7.9, with the maximum limit of 7.9 at sampling point 11. The pH affects chemical and biological processes and temperature affects the availability of oxygen concentration in the water (Kowalkowski et al. 2006). BOD5, COD, and NH4 of the water samples varied from 4.5 to 8.1 mg/L, from 1.0 to 1.4 mg/L, and from 3.8 to 9.1 mg/L, respectively. These concentrations must reflect anthropogenic influences since majority of the sampling points are located in the most vigorous urban area of Kano City, where the rivers are seriously polluted by residential wastewaters (Fig. 1) BOD is a measure of the quantity of oxygen consumed by microorganisms during the decomposition of organic matter. BOD is the most commonly used parameter for determining the oxygen demand on the receiving water of a municipal or industrial discharge. These organic compounds are indicators of organic pollution; unpolluted natural water has a BOD value of <5 mg/L. COD concentrations in all the sampling points show that the value

0.74 0.90 0.01 0.73 0.10

0.00 0.62 0.10 0.03 0.86 0.91 0.01 0.85 0.11 0.01 0.76 0.01 0.02 0.81 0.01 0.02 0.91 0.01

0.03 0.89 0.01 0.04 0.50 0.01 0.50 0.57 0.01 0.03 0.62

Mean SE mean SD Mean SE mean

SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean

SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean

1

12

11

10

9

8

7

6

5

4

3

2

DO (mg/L)

Sampling point

0.25 7.09 0.02 0.28 7.80 0.02 0.29 7.60 0.02 0.43 6.80

0.00 7.99 0.20 0.43 6.20 0.21 0.43 6.23 0.19 0.02 5.90 0.02 0.06 7.88 0.02 0.25 7.22 0.02

7.70 0.17 0.02 7.80 0.17

BOD5 (mg/L)

0.03 1.30 0.06 0.04 1.34 0.06 0.06 1.40 0.06 0.03 1.34

0.07 1.20 0.06 0.02 1.00 0.06 0.03 1.28 0.06 0.03 1.32 0.06 0.02 1.08 0.06 0.01 1.28 0.06

1.48 0.06 0.02 1.28 0.06

COD (mg/L)

0.07 0.70 0.43 0.07 1.23 0.44 0.07 2.02 0.55 0.07 2.12

0.07 1.28 0.65 0.07 1.59 0.55 0.07 1.40 0.52 0.07 1.43 0.55 0.07 −0.30 0.66 0.07 0.60 0.53

2.40 0.66 0.07 1.04 0.67

SS (mg/L)

0.09 6.98 0.59 0.11 7.33 0.23 0.00 7.96 0.47 0.01 7.93

0.01 7.03 0.32 0.03 6. 79 0.34 0.02 6.80 0.45 0.01 6.69 0.36 0.09 6.49 0.44 0.00 6.90 0.20

7.08 0.35 0.04 7.05 0.23

pH

0.02 6.30 0.55 0.20 7.20 0.76 0.06 8.00 0.98 0.10 9.10

0.01 5.70 0.54 0.02 8.20 0.75 0.02 9.10 0.61 0.04 6.80 0.54 0.23 7.50 0.72 0.01 7.66 0.77

3.80 0.63 0.07 5.10 0.66

NH4 (mg/L)

0.33 31.67 2.41 0.17 30.21 2.88 0.45 27.31 2.99 0.13 27.22

0.11 26.88 2.10 0.40 28.61 2.10 0.60 28.59 2.70 0.13 30.58 2.22 0.11 30.16 2.17 0.66 29.45 2.15

25.37 2.04 0.23 26.19 2.11

Temperature (°C)

Table 3 Univariate statistical summary of physicochemical parameters of the study area

0.07 1160 0.56 0.02 1300 0.34 0.03 1200 0.27 0.02 1209

0.01 1167 0.65 0.12 1340 0.21 0.03 1230 0.26 0.21 1189 0.32 0.02 1198 0.36 0.05 1201 0.52

1200 0.34 0.04 1260 0.56

Conductivity (μS/cm)

0.02 690 0.34 0.03 830 0.20 0.04 850 0.56 0.07 760

0.01 740 0.32 0.01 780 0.20 0.01 790 0.21 0.01 740 0.26 0.03 680 0.35 0.02 740 0.23

680 0.10 0.01 770 0.20

Salinity (mg/L)

0.05 9.80 0.43 0.02 11.90 0.55 0.03 12.90 0.55 0.01 9.90

0.03 11.70 0.61 0.04 9.90 0.45 0.02 11.30 0.65 0.04 11.00 0.54 0.03 12.10 0.55 0.02 9.80 0.57

9.10 0.58 0.06 9.50 0.52

Turbidity (NTU)

0.03 987 0.60 0.03 887 0.24 0.01 934 0.56 0.03 832

0.30 892 0.41 0.01 923 0.65 0.02 944 0.54 0.03 888 0.34 0.02 892 0.41 0.01 898 0.51

822 0.44 0.04 875 0.50

DS (mg/L)

0.02 1.00 0.15 0.01 0.01 0.13 0.01 0.01 0.09 0.02 0.02

0.01 0.01 0.13 0.02 0.02 0.14 0.01 0.02 0.11 0.02 0.04 0.13 0.01 0.02 0.11 0.02 0.02 0.12

0.01 0.12 0.01 0.02 0.11

Cd (mg/L)

0.02 0.04 0.02 0.01 0.01 0.01 0.02 0.01 0.03 0.03 0.02

0.02 0.01 0.01 0.01 0.02 0.02 0.01 0.03 0.04 0.02 0.01 0.01 0.02 0.03 0.02 0.01 0.01 0.01

0.02 0.01 0.01 0.02 0.04

Cr (mg/L)

0.01 0.01 0.02 0.01 0.02 0.02 0.02 0.02 0.01 0.02 0.01

0.02 0.01 0.01 0.01 0.03 0.05 0.02 0.02 0.02 0.01 0.03 0.03 0.01 0.02 0.01 0.00 0.02 0.01

0.01 0.01 0.01 0.01 0.02

Pb (mg/L)

5636 Environ Sci Pollut Res (2013) 20:5630–5644

23

22

21

20

19

18

17

16

15

14

13

Sampling point

0.84 0.23 0.60 0.84 0.01 0.02 0.95 0.01 0.02 0.90 0.01 0.04 0.80 0.01 0.44 0.91 0.01 0.32 0.87

0.01 0.21 0.66 0.01 0.61 0.71 0.01

SE mean SD Mean SE mean SD Mean SE mean

0.01 0.20 0.66 0.01 0.23 0.70 0.01 0.22

DO (mg/L)

Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean SE mean SD Mean

SE mean SD Mean SE mean SD Mean SE mean SD

Table 3 (continued)

0.02 0.01 8.11 0.02 0.03 6.98 0.02

4.90 0.02 0.27 6.80 0.02 0.34 8.01 0.02 0.33 7.77 0.02 0.22 8.00 0.02 0.23 7.00 0.02 0.01 7.32

0.02 0.26 4.50 0.02 0.23 6.90 0.02 0.26

BOD5 (mg/L)

0.06 0.03 1.28 0.06 0.03 1.30 0.06

1.26 0.06 0.03 1.28 0.06 0.03 1.01 0.06 0.06 1.23 0.06 0.05 1.26 0.06 0.00 1.30 0.06 0.07 1.23

0.06 0.02 1.26 0.06 0.03 1.30 0.06 0.09

COD (mg/L)

0.52 0.07 2.36 0.45 0.07 1.67 0.60

1.83 0.32 0.07 0.78 0.45 0.07 1.72 0.87 0.07 1.30 0.12 0.07 0.00 0.33 0.07 0.60 0.43 0.07 1.54

0.76 0.07 1.79 0.81 0.07 2.21 0.01 0.07

SS (mg/L)

0.66 0.03 6.55 0.34 0.08 7.95 0.81

6.69 0.76 0.01 6.50 0.44 0.04 6.84 0.26 0.02 6.44 0.31 0.07 6.57 0.49 0.01 6.52 0.37 0.02 6.66

0.43 0.20 6.25 0.54 0.01 7.98 0.02 0.01

pH

0.34 0.06 7.10 0.67 0.03 6.50 0.65

7.90 0.50 0.34 7.40 0.66 0.02 7.60 0.73 0.06 8.20 0.88 0.04 8.80 0.81 0.02 7.90 0.66 0.04 7.90

0.56 0.05 8.60 0.45 0.01 6.90 0.66 0.01

NH4 (mg/L)

2.54 0.34 30.52 2.66 0.25 28.31 2.66

26.90 2.41 0.33 27.00 2.65 0.32 29.00 2.97 0.29 28.00 2.33 0.65 26.94 2.07 0.29 30.17 2.66 0.75 27.00

2.10 0.35 27.20 2.97 0.19 27.80 2.91 0.87

Temperature (°C)

0.54 0.01 1200 0.27 0.02 1401 0.43

1278 0.26 0.01 1300 0.56 0.03 1200 0.28 0.01 1340 0.54 0.01 1490 0.28 0.02 1301 0.44 0.07 1201

0.38 0.03 1300 0.49 0.03 1350 0.65 0.04

Conductivity (μS/cm)

0.34 0.03 840 0.11 0.01 745 0.18

670 0.15 0.01 866 0.17 0.01 734 0.12 0.01 745 0.13 0.01 723 0.11 0.01 744 0.09 0.01 787

0.65 0.03 733 0.20 0.02 734 0.23 0.03

Salinity (mg/L)

0.49 0.03 9.50 0.49 0.03 9.90 0.50

11.30 0.64 0.07 8.90 0.50 0.05 8.60 0.55 0.03 9.60 0.56 0.05 9.90 0.73 0.08 6.20 0.56 0.04 7.80

0.55 0.03 9.60 0.44 0.02 12.80 0.52 0.04

Turbidity (NTU)

0.40 0.05 898 0.58 0.05 860 0.45

756 0.43 0.03 698 0.28 0.01 967 0.45 0.04 866 0.60 0.06 856 0.28 0.02 745 0.39 0.04 687

0.48 0.04 756 0.67 0.06 888 0.38 0.04

DS (mg/L)

0.00 0.01 0.04 0.03 0.01 0.03 0.01

0.03 0.11 0.01 0.03 0.00 0.02 0.01 0.01 0.02 0.02 0.02 0.01 0.05 0.03 0.02 0.01 0.00 0.01 0.02

0.00 0.01 0.01 0.00 0.02 0.01 0.00 0.01

Cd (mg/L)

0.03 0.01 0.02 0.02 0.01 0.03 0.01

0.02 0.01 0.01 0.01 0.03 0.01 0.03 0.02 0.01 0.01 0.01 0.01 0.04 0.04 0.02 0.01 0.01 0.03 0.01

0.01 0.01 0.01 0.03 0.01 0.03 0.02 0.02

Cr (mg/L)

0.01 0.02 0.01 0.01 0.01 0.01 0.03

0.01 0.03 0.03 0.03 0.02 0.02 0.01 0.03 0.01 0.03 0.02 0.02 0.04 0.01 0.01 0.01 0.03 0.01 0.02

0.03 0.01 0.04 0.01 0.01 0.04 0.02 0.01

Pb (mg/L)

Environ Sci Pollut Res (2013) 20:5630–5644 5637

0.02 0.02 0.03 0.02 0.01 0.02 0.01 0.02 0.01 0.01 0.02 0.01 0.01 0.00 0.01 0.76 0.07 870 0.65 0.05 0.15 0.02 740 0.13 0.01 27

0.01 0.32 0.87 0.00 0.22

26

25

SE mean SD Mean SE mean SD

0.02 0.02 7.10 0.02 0.01

0.06 0.00 1.30 0.05 0.02

0.50 0.07 1.57 0.06 0.05

0.49 0.03 6.48 0.07 0.05

0.55 0.02 8.10 0.74 0.04

2.44 0.34 28.00 2.50 0.45

0.51 0.04 1452 0.29 0.01

0.52 0.02 8.90 0.58 0.06

0.01 0.02 0.02 0.02 0.03 0.03 0.02 0.01 0.03 0.01 0.02 0.02 0.01 0.01 0.01 0.02 0.02 780 0.49 0.03 990 0.27 0.01 950 0.02 777 0.16 0.02 790 0.16 0.02 810 0.32 0.78 0.01 0.20 0.81 0.01 0.11 0.85 24

SD Mean SE mean SD Mean SE mean SD Mean

0.05 6.91 0.02 0.23 7.88 0.02 0.13 7.20

0.02 1.40 0.06 0.05 1.34 0.06 0.03 1.23

0.07 1.91 0.71 0.07 1.91 0.43 0.07 1.92

0.09 7.05 0.77 0.06 6.27 0.66 0.07 6.46

0.08 6.88 0.60 0.01 7.43 0.58 0.03 7.00

0.43 27.81 2.43 0.11 27.63 2.88 0.56 26.77

0.02 1400 0.28 0.06 1399 0.66 0.05 1299

0.04 9.10 0.61 0.03 9.30 0.48 0.03 8.70

Cr (mg/L) Cd (mg/L) DS (mg/L) Turbidity (NTU) Salinity (mg/L) Conductivity (μS/cm) Temperature (°C) NH4 (mg/L) pH SS (mg/L) COD (mg/L) BOD5 (mg/L) DO (mg/L) Sampling point

Table 3 (continued)

0.02 0.01 0.03 0.01 0.02 0.02 0.01 0.01

Environ Sci Pollut Res (2013) 20:5630–5644 Pb (mg/L)

5638

is within the acceptable limit. The presence of organic compound in water under normal conditions supports the growth of bacteria and other microorganisms, which may enhance the concentration of BOD5, COD, and NH3 and the very low available DO in water. The DO value for the riverine network ranged from 0.50 to 0.95 mg/L. The concentration of DO is not within the WHO acceptable limit of 4–10 mg/L. The DO in water can be depleted as it is used in the oxidation of organic matter (Onojake et al. 2011). The distributions of BOD5, COD, NH3, and DO in the sampling points are shown in Fig. 2. The concentrations of conductivity ranged from 1,167 to 1,490 μS/cm. Salinity varied from 670 to 866 mg/L, DS varied from 687 to 866 mg/L, and turbidity ranged from 6.20 to 12.90 mg/L. The concentrations of these parameters were above the threshold limits of the WHO. The higher concentrations of conductivity, salinity, turbidity, and DS indicate that these parameters were from a common source of origin (Onojake et al. 2011) and might be due to a high amount of dissolved ions in the Jakara Basin. The distributions of conductivity, salinity, turbidity, and DS in the Jakara Basin are shown in Fig. 3. Physicochemical parameter relationships Pearson’s correlation matrix is presented in Table 4. The correlation between the physiochemical parameters under study showed a significant positive relationship between salinity and DS, and conductivity with DS (r=0.987 and 0.992, p<0.01). This is evident because solids that dissolve in water break into positively and negatively charged ions, which increases the concentration of conductivity (Mustapha et al. 2012). Conductivity is the ability of water to conduct an electric current, and the dissolved ions are the conductors. Dissolved ions equally increase the salinity level. A significant relationship between SS and turbidity (r= 0.889, p<0.01) was observed. Turbidity values are associated with the cloudiness and color of the water being sampled. A high value of turbidity level in water can be caused either by particulates or inorganic matter in the water which may enhance the growth of microorganisms and lower the effectiveness of disinfection processes (Calijuri et al. 2012). Turbidity of water depends on the suspended particles, shape, size, and amount of the surface area which can cause variations in reflection and absorption of light in the water. Adequate relationships between turbidity and SS have been determined and reported in many surface water bodies in regions around the world: for example, a study conducted in Puget Lowland in the USA showed a positive relationship between turbidity and SS in all the studied water bodies regardless of differences in lithology, drainage area, and land use pattern (Packman 1999). Similarly, Acheampong et al. (2012) have reported that

Environ Sci Pollut Res (2013) 20:5630–5644

5639

Fig. 2 Distribution of biodegradable organic compound in the Jakara Basin

10.00 9.00

Concentration in mg/L

8.00 7.00 6.00

DO

5.00

BOD

4.00

COD

3.00

NH4

2.00 1.00 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Sampling points

turbidity is a surrogate measure of SS and that turbidity level is influenced by the concentration of solids in water bodies. Excessive levels of SS in water bodies may have a significant deleterious impact on the physical, chemical, and biological properties of the water bodies. A significant relationship was revealed by a bivariate correlation between BOD5 and COD (r=0.899, p<0.05), COD and NH4 (r=0.532, p<0.05), and BOD5 and NH4 (r=0.841, p<0.05). This indicates the presence of biodegradable organic matter in the sampled water of Jakara River (Mustapha et al. 2012a, b, 2012). BOD5 is a measure of the amount of oxygen that is consumed by bacteria during the decomposition of organic matter

1600

Concentration in mg/L except conductivity in µS/cm

Fig. 3 Distribution of conductivity, salinity, turbidity, and dissolved solids in the Jakara Basin

under aerobic conditions, whereas COD is a measure of the total quantity of oxygen required to oxidize organic materials into carbon dioxide and water under strong oxidants (Mandal et al. 2010). The degradation of organic matter in the water consumes the available DO, leading to the rapid depletion of available DO in water, resulting in high BOD5, COD, and NH4. There is a nonlinear relationship between BOD5 and DO (r=−0.626, p<0.05), COD and DO (r=−0.594, p<0.05), and NH4 and DO (r = −0.532). This is an indication of the utilization of organic compounds from municipal waste (Onojake et al. 2011). Significantly, the lowest DO content in the water suggested that the discharge of domestic

1400

1200

1000

Conductivity 800

Salinity Turbidity

600

DS

400

200

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Sampling points

5640

Environ Sci Pollut Res (2013) 20:5630–5644

Table 4 Pearson product-moment correlation coefficient of physiochemical parameters under study DO

BOD5

COD

SS

pH

NH4

DO BOD5 COD SS pH NH4 Temperature Conductivity

1.00 −0.626 1.00 −0.594 0.899 1.00 0.286 0.023 0.374 1.00 0.429 0.197 0.227 0.644 1.00 −0.532 0.841 0.532 −0.058 0.242 1.00 −0.278 0.119 −0.099 −0.584 −0.260 0.157 0.294 −0.073 −0.126 −0.118 0.159 −0.128

Salinity Turbidity DS Cd Cr Pb

0.296 0.324 0.292 0.061 0.039 −0.094

−0.075 −0.083 −0.068 0.298 0.193 −0.338

−0.126 −0.117 0.160 0.194 0.889 0.557 −0.126 −0.118 0.161 −0.213 0.040 0.267 −0.013 0.041 0.246 −0.208 −0.164 −0.278

Temperature Conductivity Salinity Turbidity DS

1.00 0.408

−0.130 0.408 −0.078 −0.625 −0.123 0.408 0.413 0.166 0.375 −0.065 −0.189 0.136

wastewater from the metropolitan area close to the basin has induced serious organic pollution in the Jakara River since DO decrease is normally caused by organic compounds (Mustapha et al. 2012a). Human activities have negatively influenced water quality and aquatic ecosystem functions, especially around urban areas, since rivers passing through cities receive multiple contaminants released from domestic sewage and agricultural activities. This has generated great pressure on the ecosystem, resulting in a decrease of water quality and biodiversity (Wang et al. 2012). Different pollutants are being released from the households, ranging from detergents, oil and grease, and solid wastes. These pollutants have negative effects on the surface water; they pollute the water and thus cause undesirable smells. Jakara Basin, being in the heart of Kano metropolitan, receives domestic wastewater generated from the urban residential areas. Jakara Basin has high population growth coupled with high industrial and commercial activities, which may result in the disposal of domestic wastewater. Surface waters are facing a variety of pressures affecting both the ecosystem and human health through municipal wastewater discharge and disposal practices that lead to the introduction of high nutrient loads, hazardous chemicals, and pathogens, causing diseases. Singh et al. (2005) have reported that domestic sewage has constant strength and pollution ability and that it is the combination of gray water and municipal waste which may contain pathogens. Similarly, a considerable amount of recent literature published showed that domestic sewage constitutes suspended solids (such as partially disintegrated particles) and very small solids in colloidal forms, large floating particles such as rags, plastic, and clothing, among others (Singh et al. 2004; Shrestha and Kazama 2007; Panigrahi et al. 2007).

Cd

Cr

Pb

1.00 0.988 0.766 0.992 0.213 −0.141 −0.166

1.00 −0.161 1.00 0.987 −0.163 0.213 0.222 −0.142 0.169 −0.166 −0.037

1.00 0.216 1.00 −0.140 0.354 1.00 −0.170 −0.086 0.183 1.00

Identification of the suppressor variable using partial correlation techniques Partial correlation (rp) was used to further explore the relationship between salinity and turbidity, BOD5, and COD controlling for the linear effects of conductivity and NH4, respectively. Based on the zero-order partial correlation output obtained in Table 5, there was a strong partial correlation (rp =0.930, p=0.001) with the high level of conductivity being associated with the high levels of salinity and turbidity (rp =0.839, p=0.001) and the high levels of BOD5 and COD being associated with NH4. An inspection of zeroorder Pearson product-moment correlations between salinity, turbidity, and conductivity (r = 0.988, r = 0.766) and BOD5, COD, and NH4 (r=0.841, r=0.532), respectively, shows that conductivity and NH4 are the suppressor variables on the assumption that they are suppressing the larger correlation appearing between salinity and turbidity, and BOD5 and COD, respectively. Water pollution source identification using PCA/FA Preliminary analysis prior to factor analysis was conducted; a quick check at the Kaiser–Meyer–Olkin (KMO) test and the Bartlett test of sphericity table was done to ensure no violation of the assumption of factor analysis. KMO measures sampling adequacy and Bartlett’s test of sphericity investigates the relationship within the variables under study (Hinton and Brownlow 2004). The KMO result was 0.82 and Bartlett’s sphericity test was significant (0.001, p<0.05), showing that PCA/FA could be considered appropriate and useful to provide significant reduction in data dimensionality. To obtain

Environ Sci Pollut Res (2013) 20:5630–5644 Table 5 Partial correlation controlling for the linear effects of conductivity and NH4

5641

Controlling for conductivity Control variables: Conductivity

Salinity Turbidity

more reliable information about the relationships among the variables, PCA/FA was applied to the datasets to explore the extent of the physiochemical relationship and water pollution source identification. Varimax rotation method was used to maximize the sum of the variance of the factor coefficient, which better explained the possible group/sources that influenced the water chemistry in the Jakara River. The factor loadings were ranked following the correlation coefficient matrix between the variables. Table 6 summarized the PCA/FA results including the loadings, eigenvalues, variance of each factor, and the overall cumulative variance of the variables. In this study, a factor with an eigenvalue >1 was considered for subsequent discussion. Following the rule of Kaiser one criterion, three independent varimax factors (VF) were extracted, which explained 68.11 % of the total variation of water quality in the Jakara River. The first VF explained 26.16 % of the total variance and was best represented by conductivity, salinity, and DS. This factor represents a pollution source from irrigation agricultural activities along the Jakara River. This is evident as farmers are practicing irrigation activities around the

Table 6 Varimax-rotated component matrix Variables

VF1

VF2

VF3

DO BOD5 COD SS pH

0.369 −0.022 −0.156 −0.075 0.295

0.566 −0.137 0.216 0.908 0.742

−0.541 0.944 0.662 0.114 0.343

NH4 Temperature Conductivity Salinity Turbidity DS Cd Cr Pb Eigenvalue Variance CV (%)

−0.042 0.421 0.972 0.973 −0.11 0.973 0.364 −0.076 −0.261 3.66 26.16 26.16

−0.147 0.704 −0.078 −0.076 0.902 −0.078 0.123 0.160 −0.218 3.10 22.16 48.32

0.908 0.115 −0.091 −0.093 0.062 −0.086 0.452 0.377 −0.329 2.77 19.79 68.11

CV cumulative variance

Controlling for NH4 Salinity

Turbidity

1 0.930

1

Control variables: NH4

BOD5 COD

BOD5

COD

1 0.839

1

area. The most common form of stream pollution associated with agricultural activities is the increased concentrations of soil particles washed into the stream by land clearing and farming activities (Mustapha et al. 2012). Water with a high concentration of DS is salty, and salinity is the total of all dissolved solids in water. Leaching of salts from the irrigation field and the return flow from irrigation water are the possible principal contributors that control the largest variation of water in the Jakara River. Tlili-Zrelli et al. (2012) revealed that an elevated salinity value would be attributed to the leaching of salty water, return flow of irrigation water, and use of fertilizer. This may have contributed to the higher concentration of DS and the subsequent higher values of salinity and conductivity. VF2 had a strong loading on pH, temperature, and turbidity and explained 22.6 % of the total variance. Turbidity is the condition resulting from suspended solids in the water. Particles suspended in water may absorb heat in the sunlight, hence raising water temperature. The strong loading on these parameters could have been due to anthropogenic activities through road construction, clearing of lands, runoff, and erosive processes taking place near the study area (Mustapha and Aris 2012b). VF3 explained 19.79 % of the total variation and has a positive loading on BOD5, COD, and NH4 and a negative loading on DO; these parameters are indicators of organic pollution (Mandal et al. 2010). This factor explains the biological processes due to phytoplankton productivity and more production of organic matter resulting in more microorganism activity, which in turn increases the concentration of BOD. A high concentration of organic matter in water may consume large amounts of available DO which undergoes anaerobic fermentation processes, leading to the formation of NH4 and organic acid. High loading on organic compounds in the water body indicates that the river is heavily polluted with both oxidizable organic and inorganic pollutants (Otokunefor and Obiukwu 2005). The direct dumping of waste and the discharge of sewage effluent into the river have been identified as the main contributing factors enhancing BOD5, COD, and NH4 (Mustapha et al. 2012a). DO may be consumed by the bioxidation of nitrogenous materials in water; the continual depletion of DO in surface water can encourage microbial depletion. The positive correlations of BOD5, COD, and NH4

5642

Environ Sci Pollut Res (2013) 20:5630–5644

with the negative correlation of DO affirm this claim (Mustapha et al. 2012b). Various studies conducted attributed the BOD5, COD, and NH4 with lower DO in sewage water to the presence of biodegradable organic matter and the utilization of DO by a microorganism in the water (Mandal et al. 2010; Mustapha et al. 2012). Discriminant analysis DA is used in this study to predict the variables which discriminate between two natural groupings in river water quality. The predictors used were the parameters under study. Preliminary checking on the results indicated that a descriptive univariate ANOVA Box’s M test of chi-square asymptotic approximation (χ2observed =321.229, p<0.0001) shows that the assumption of equality of covariance within the group is achieved. The discriminant function revealed a significant association between groups and all the predictor variables accounting for more than 85 % of the between-group variability. The closely validated classification matrices revealed that overall, 97.22 % were correctly classified. Table 7 presented the standard mode of discriminant analysis. Stepwise discriminant analysis was performed as an exploratory analysis to identify the most significant variables among the predictors. In stepwise DA, the most correlated independent variables are entered first by the stepwise analysis, then the second most significant variable until an additional dependent variable adds no significant differences. Table 8 presents the stepwise forward and back discriminant analysis results. The stepwise forward and backward statistics table shows that eight variables were successfully discriminated as the most significant among the predictors. DA reveals

Table 7 Standard mode of discriminant variables Variable

Lambda

F statistics

p value

DO BOD5 COD SS

0.467 0.695 0.670 0.677

38.782 14.915 16.781 16.230

<0.0001 0.000 0.000 0.000

pH NH4 Temperature Conductivity Salinity Turbidity DS Cd Cr Pb

0.786 0.483 0.997 0.990 0.992 0.557 0.486 0.980 0.990 0.960

9.280 11.597 0.107 0.342 0.279 1.516 7.472 0.327 0.324 0.344

0.004 0.000 0.746 0.000 0.000 0.077 0.002 0.688 0.721 0.711

Table 8 Stepwise forward and backward modes of discriminant variables Variable

Lambda

F statistics

p value

DO BOD5 COD SS NH4 Conductivity Salinity DS

0.467 0.622 0.753 0.677 0.483 0.399 0.492 0.486

38.782 14.915 16.781 16.230 11.597 6.322 8.279 7.472

<0.0001 0.000 0.001 0.000 0.000 0.003 0.001 0.002

that DO, BOD5, COD, SS, NH4, DS, conductivity, and salinity add some predictive power to the discriminant function as all are significant with p<0.01. As depicted in Table 8, the strongest variable with largest F statistics was DO (F=38.782, p<0.0001), with Wilk’s lambda value of 0.467; this means that DO makes the strongest unique contributions in explaining the variation of water quality in Jakara River when the variance explained by other predictors in the stepwise model is controlled. COD (F=16.781, p=0.001) and SS (F=16.230, p=0.0001) were the second and third contributor variables, respectively. The F statistics for conductivity was the smallest among the stepwise forward and backward DA and indicates that it made the least contribution in the water quality variation in the Jakara River.

Conclusions The results of the environmetric techniques used in this study seem to give evidence on the reasons behind the water quality variations in the Jakara River Basin. The application of PCA coupled with FA on the available data indicated that the water quality variations are mainly due to anthropogenic (irrigation agriculture, construction, clearing of land, and domestic waste disposal) and natural processes (erosion and runoff). DA rendered an important data reduction as it uses only eight parameters (DO, BOD5, COD, SS, NH4, conductivity, salinity, and DS), affording more than 90 % correct assignation. Furthermore, partial correlation revealed a strong partial relationship between salinity and turbidity, and BOD5 and COD controlling for the linear effects of conductivity and NH4, respectively. These environmental tools provided a more objective interpretation of surface water physicochemical parameters and identification of water pollution source apportionment as part of the effort toward the management of a sustainable river basin.

Environ Sci Pollut Res (2013) 20:5630–5644

References Acheampong MA, Paksirajan K, Lens PN (2012) Assessment of the effluent quality from a gold mining industry in Ghana. Environ Sci Pollut Res. doi:10.1007/s11356-012-1312-3 APHA (2005) Standard methods for the examination of water and wastewater. American Water Works Association, Environment Federation, Washington, DC Aris AZ, Praveena SM, Abdullah MH, Radojevic M (2012) Statistical approaches and hydrochemical modeling of groundwater system in a small tropical island. J Hydroinform 14:206–220 Calijuri ML, Couto EA, Santiago AF, Camargo AR, Silva MD (2012) Evaluation of the influence of natural and anthropogenic processes on water quality in Karstic Region. Water Air Soil Pollut 223:2157–2168. doi:10.1007/s11270-011-1012-5 Cho KH, Park Y, Kang J, Ki SJ, Cha S, Lee SW et al (2009) Interpretation of seasonal water quality variation in the Yeongsan Reservoir, Korea using multivariate statistical analyses. J Hydroinform 59(11):2219–2226 Cukrov N, Tepic N, Omanović D, Logen S, Bura-Nakić S, Vojvodic E et al (2012) Qualitative interpretation of physico-chemical and isotopic parameters in the Krka River (Croatia) assessed by multivariate statistical analysis. Int J Environ Anal Chem 92(10):1187–1199 Dimitrovska O, Markoski B, Toshevska BA, Milevski I, Gorin S (2012) Surface water pollution of major rivers in the Republic of Macedonia. Procedia Environ Sci 14:32–40. doi:10.1016/j.proenv.2012.03.004 Esmaeili A, Moore F (2012) Hydrogeochemical assessment of groundwater in Isfahan province, Iran. Environ Earth Sci 67:107–120. doi:10.1007/s12665-011-1484-z Gholikandi GB, Haddadi S, Dehghanifard E, Tashayouie HR (2012) Assessment of surface water resources quality in Tehran province, Iran. Desalin Water Treat 37(1–3):8–20 Hair JF, Anderson RE, Tatham RL, William C (1995) Multivariate data analysis with readings. Prentice Hall, Englewood Cliffs Han S, Kim E, Kim S (2009) The water quality management in the Nakdong River watershed using multivariate statistical techniques. Korean J Civ Eng 13(2):97–105 Helena B, Pardo R, Vega M, Barrado E, Fernandez J, Fernandez L (2000) Temporal evolution of groundwater composition in an alluvial aquifer (Pisuerga River, Spain) by principal component analysis. Water Res 34(3):807–816 Hinton PR, Brownlow C (2004) SPSS explained. Routledge, New York Hu J, Qiao Y, Zhou L, Li S (2011) Spatiotemporal distributions of nutrients in the downstream from Gezhouba Dam in Yangtze River, China. Environ Sci Pollut Res 19:2849–2859. doi:10.1007/ s11356-012-0791-6 Koklu R, Sengorur B, Topal B (2010) Water quality assessment using multivariate statistical methods: a case study of Melen River system (Turkey). Water Resour Manage 24(5):959–978 Kowalkowski T, Zbytniewski R, Szpejna J, Buszewski B (2006) Application of chemometrics in river water classification. Water Res 40(4):744–752 Li Y, Xu L, Li S (2009) Water quality analysis of the Songhua River Basin using multivariate techniques. J Water Resour Prot 1(2):110–121 Liu WC, Yu HL, Chung CE (2011) Assessment of water quality in a subtropical Alpine Lake using multivariate statistical techniques and geostatistical mapping: a case study. Int J Environ Res Public Health 8(4):1126–1140 Mandal P, Upadhyay R, Hasan A (2010) Seasonal and spatial variation of Yamuna River water quality in Delhi, India. Environ Monit Assess 170(1):661–670 Manly BF (2005) Multivariate statistical methods. Chapman & Hall/CRC McGarigal K, Cushman S, Stafford SG (2000) Multivariate statistics for wildlife and ecology research. Springer, New York

5643 Mouri G, Takizawa S, Oki T (2011) Spatial and temporal variation in nutrient parameters in stream water in a rural–urban catchment, Shikoku, Japan: effects of land cover and human impact. J Environ Manage 92(7):1837–1848 Mustapha A, Aris AZ (2012a) Spatial aspects of surface water quality in the Jakara Basin, Nigeria using chemometric analysis. J Environ Sci Health Part A 47:1455–1465 Mustapha A, Aris AZ (2012b) Multivariate statistical analysis and environmental modeling of heavy metals pollution by industries. Pol J Environ Stud 21:1359–1367 Mustapha A, Nabegu AB (2011) Surface water pollution source identification using principal component analysis and factor analysis in Getsi River, Kano, Nigeria. Austr J Basic Appl Sci 5:1507–1512 Mustapha A, Aris AZ, Juahir H, Ramli MF (2012) Surface water quality contamination source apportionment and physicochemical characterization at the upper section of the Jakara Basin, Nigeria. Arab J Geosci. doi:10.1007/s12517-012-0731-2 Mustapha A, Aris AZ, Ramli MF, Juahir H (2012a) Spatial-temporal variation of surface water quality in the downstream region of the Jakara River, Northwestern Nigeria: a statistical approach. J Environ Sci Health Part A 47:1551–1560 Mustapha A, Aris AZ, Ramli MF, Juahir H (2012b) Temporal aspects of surface water quality variation using robust statistical tools. The Scientific World Journal 2012:294540. doi:10.1100/2012/294540 Onojake MC, Ukerun SO, Iwuoha G (2011) A statistical approach for evaluation of the effects of industrial and municipal wastes on Warri Rivers, Niger Delta, Nigeria. Water Qual Expo Health 3:91–99 Otokunefor TV, Obiukwu C (2005) Impact of refinery effluent on the physicochemical properties of a water body in the Niger delta. Appl Ecol Environ Res 3(1):61–72 Packman JJ, Comings KJ, Booth DB (1999) Using turbidity to determine total suspended solids in urbanizing streams in the Puget lowlands. In: Confronting Uncertainty: Managing Change in Water Resources and the Environment; Canadian Water Resources Association Annual Meeting, Vancouver, BC Panigrahi S, Acharya BC, Panigrahy RC, Nayak BK, Banarjee K, Sarkar SK (2007) Anthropogenic impact on water quality of Chilika lagoon RAMSAR site: a statistical approach. Wetlands Ecol Manage 15(2):113–126 Papaioannou A, Mavridau A, Hadjichristodoulou C, Papastergiou P, Pappa O, Dovriki E, Rigas I (2010) Application of multivariate statistical methods for groundwater physicochemical and biological quality assessment in the context of public health. Environ Monit Assess 170(1–4):87–97 Pradhan UK, Shirodkar PV, Sahu BK (2009) Physico-chemical characteristics of the coastal water off Devi estuary, Orissa and evaluation of its seasonal changes using chemometric techniques. Curr Sci 96(9):1203–1209 Rani N, Sinha RK, Prasad K, Kedia DK (2011) Assessment of temporal variation in water quality of some important rivers in middle Gangetic Plains, India. Environ Monit Assess 174(1–4):401–415 Rogerson PA (2006) Statistical methods for geography: a student’s guide. Sage, London Schaefer K, Einax JW (2010) Analytical and chemometric characterization of the Cruces River in South Chile. Environ Sci Pollut Res 17(1):115–123 Shrestha S, Kazama F (2007) Assessment of surface water quality using multivariate statistical techniques: a case study of the Fuji River Basin, Japan. Environ Modell Softw 22:464–475 Shrestha S, Kazama F, Nakamura T (2008) Use of principal component analysis, factor analysis and discriminant analysis to evaluate spatial and temporal variations in water quality of the Mekong River. J Hydroinform 10(1):43–56 Singh KP, Malik A, Mohan D, Sinha S (2004) Multivariate statistical techniques for the evaluation of spatial and temporal

5644 variations in water quality of Gomti River (India)—a case study. Water Res 38:3980–3992 Singh KP, Malik A, Mohan D, Sinha S, Singh VK (2005) Chemometric data analysis of pollutants in wastewater: a case study. Anal Chim Acta 532(1):15–25 Singh AK, Mondal GC, Singh TB, Singh S, Tewary BK, Sinha A (2012) Hydrogeochemical processes and quality assessment of groundwater in Dumka and Jamtara districts, Jharkhand, India. Environ Earth Sci 67:2175. doi:10.1007/s12665-0121658-3 Tanrıverdi Ç, Alp A, Demirkıran AR, Üçkardes F (2010) Assessment of surface water quality of the Ceyhan River Basin, Turkey. Environ Monit Assess 167(1):175–184 Tlili-Zrelli B, Hamzaoui-Azaza F, Gueddari M, Bouhlila R (2012) Geochemistry and quality assessment of groundwater using graphical and multivariate statistical methods. A case study: Grombalia phreatic aquifer (Northeastern Tunisia). Arab J Geosci. doi:10.1007/s12517-012-0617-3 Tobiszewski M, Tsakovski S, Simeonov V, Namiesnik J (2010) Surface water quality assessment by the use of combination of multivariate

Environ Sci Pollut Res (2013) 20:5630–5644 statistical classification and expert information. Chemosphere 80 (7):740–746 Varol M, Gökot B, Bekleyen A, Sen B (2012) Spatial and temporal variations in surface water quality of the dam reservoirs in the Tigris River Basin, Turkey. Catena 92:11–21 Vittori AL, Trivisano C, Gessa C, Gherardi M, Simoni A, Vianello G et al (2010) Quality of municipal wastewater compared to surface waters of the river and crtificial canal network in different areas of the eastern Po Valley (Italy). Water Qual Expo Health 2(1):1–13 Wang Y, Wang P, Bai Y, Tian Z, Li J, Shao X et al (2012) Assessment of surface water quality via multivariate statistical techniques: a case study of the Songhua River Harbin region, China. J Hydroenviron Res. doi:org/10.1016/j.jher.2012.10.003 Wong WS (2005) Using multivariate statistical techniques to examine the spatial variability of physiochemical elements and water quality in Hong Kong’s Rivers. Asian Geographer 24(1):129–150 Zhou F, Liu Y, Guo H (2007) Application of multivariate statistical methods to water quality assessment of the watercourses in Northwestern New Territories, Hong Kong. Environ Monit Assess 132(1–3):1–13