American International Journal of Contemporary Research Vol. 3 No.

3; March 2013

Application of Sequential Analysis and Geographic Information Systems for

Hydrochemical Evolution Survey, Shagari Environ, Southwestern Nigeria

Joshua Oluwasanmi Owoseni

Department of Applied Geology
Federal University of Technology
PMB 704, Akure. Nigeria.
Elisama Ugeh Tamarautobou
Department of Applied Geology
Federal University of Technology
PMB 704, Akure. Nigeria.

Yinusa Ayodele Asiwaju-Bello

Department of Applied Geology
Federal University of Technology
PMB 704, Akure. Nigeria.

Abstract
Integrated multivariate statistical, spatial, and graphical methods were applied to some hydrochemical data in
Shagari area, Akure, southwestern Nigeria with a view to elucidating groundwater hydrochemical evolution.
Water samples from 23 dug wells were clustered into distinct groups by hierarchical cluster analysis to depict
different hydrochemical facies. Factor analysis reduced bulk hydrochemical data to principal components
explaining possible dominant processes controlling water chemistry. Three factors which together explain 83.48
% of the total variance in the dataset were retained and interpreted. Factor 1 explains 35.82 % of the total
variance and indicates atmospheric controls and silicate mineral weathering processes. Factor 2 explains 35.41
% of the total variance and reflects atmospheric controls, acid rain and speciation reactions producing inorganic
carbon ions in solution. Factor 3 explains 12.24 % of the total variance, indicating silicate mineral weathering
processes resulting in elevated pH. Generally, water types tend towards sodium chloride bicarbonate.

Key-words: Geographic information systems; groundwater quality; hydrchemical evolution; sequential

analysis.

1.0 Introduction
Shallow weathered Basement rock aquifers are the major sources of water for domestic purposes in the sudy area.
Groundwater quality in many parts of the region is often questionable due to anthropogenic impacts from sewage,
fertlizers and refuse dumps. Mineral enrichment from host rocks also alter groundwater chemistry appreciably in
some parts, thereby making its usability objectionable. Consequently, groundwater characterization is inevitable
in most parts of the region. Several wokers (like Elueze et al., 2001 and Ehinola, 2002) had used the traditional
graphical approach to characterize groundwater geochemistry in parts of southwestern Nigeria. The absence or
inadequacy of input data is a major limitation in groundwater models application in developing countries like
Nigeria.
However, evolving emperical assessment methodologies (EAM) are common in recent years in groudwater
hydrochemical studies. Multivariate statistical methods have been employed to extract critical information from
hydrochemical datasets with respect to groundwater quality assessment in several places (Olobaniyi and
Owoyemi, 2006; Shiab and Hashim, 2006; Petalas et al., 2006; Papatheodorou et al., 2007; Belkhiri et al., 2010).

38
© Centre for Promoting Ideas, USA www.aijcrnet.com

Locsey and Cox (2003) differentiated groundwaters sourced from diferent lithological formations by multivariate
analysis of physico-chemical data for easy identification of the likely host rocks of groundwaters from
unidentified lithological units, and better understanding of groundwater resource. Yidana (2011) applied
multivariate statistics and GIS in a sequential manner to understand hydrochemical processes in a tropical setting
on a regional scale. An integrated approach that uses hydrochemical data to improve characterization of watershed
hydrology by standard multivariate statistical techniques, spatial analysis and inverse geochemical modeling had
also been reported (Thyne et al., 2004). The current study explores an integration of the strengths of multivariate
statistical, spatial, and graphical methods in a sequential manner for a robust interpretation of hydrochemical
dataset. This is with a view to elucidating possible underlying hydrochemical processes responsible for variability
of groundwater composition, and hydrochemical evolution of water in Shagari area, Akure, southwestern Nigeria.
Multivariate factor analysis (FA) by principal component extraction method with varimax rotation is used to
characterize groundwater hydrochemical facies in the area.
Factor analysis is often used in hydrochemical studies to interprete the structure within the variance-covariance
matrix obtained from a collection of multivariate hydrogeochemical observations by extracting the eigenvalues
and eigenvectors from the matrix of correlations or covariances of measured standardized variables (Davis, 1973).
Bulk hydrochemical data reduction gives factor score groups that correspond to distinguishable underlying
controlling environmental factors. Determination of the number of components to retain and interprete is often
based on some criteria, viz: components with eigenvalue > 1, components accounting for at least 70% total
variability, and components within the sharp descent before eigenvalues level off on the scree plot.
1.1 Description of the study area
The study area lies within latitudes 7 o 16l N and 7o 18l N of the equator and longitudes 5o 10l E and 5o 13l E of the
Greenwich meridian. It is accessible by footpaths and secondary roads. The topography is generally characterized
by hills of varying heights ranging from about 300m to 400m (Fig. 1). The climate is of the West African
monsoonal type, characterized by distinct wet and dry seasons typical of West African tropical regions. Average
temperatures reach a peak of about 32o C around February and a threshold of about 21o C around August. Relative
humidity ranges from about 70% around January to about 90% in July. Typically, the vegetation of the study area
is the tropical rain forest with thick undergrowth but had in many parts been modified by human activities such as
urbanisation, construction, land cultivation and deforestation.
1.2 Geological and hydrogeological setting
The study area is underlain by the Precambrian Basement Complex of southwestern Nigeria. The assemblages
within the Basement Complex had been grouped into four, viz: minor intrusives, Older granites, schist belts and
migmatite-gneiss-quartzite complex (Adekoya et al., 2003). The lithologic units encountered in the study area
include granite gneiss, porphyritic gneiss, pelitic schist, charnockite, and migmatite gneiss complex (Fig. 2). They
vary in texture, and range from fine to medium to coarse porphyritic grains. Structural features occurring on the
lithologic units include veins, veinlets, foliations, intrusions and rock-to-rock contacts. Other structures include
joints, faults and folds. Generally, the fissure zones trend north-south while the foliation trends are NNW-SSE and
NNE-SSW (Olorunfemi et al., 1999). Availability of groundwater in the area often depends on the thickness of
the weathered overburden and the frequency of fractures within the crystalline rocks. Depth to static water level in
the area range between 1.47m and 8.74m. Groundwater recharge is primarily through infiltration by direct
precipitation which reaches mean annual value of about 1350mm. The wet season spans between April to
September while the dry season period is between October and March. Secondary recharge occurs by influent
from rivers and their tributaries. Discharge of groundwater occurs through seepages, springs, well water
abstraction, and flow into rivers and streams. The area is drained by streams and rivers displaying dendritic
pattern and maintaining a meandering pattern of flow in south-eastern and southwestern directions.
2.0 Methodology
2.1 Sample collection and laboratory analysis
A total of twenty-three (23) groundwater samples were collected from hand dug wells in Shagari area for
chemical analysis. Samples were taken from wells with a rubber container to avoid metallic contamination, and
storage was by the use of plastic containers.
39
American International Journal of Contemporary Research Vol. 3 No. 3; March 2013

Total dissolved solids (TDS), pH, Temperature and electrical conductivity (EC) were measured in situ using the
HANNA COMBO meter (HI 98129 model). Samples were analysed in the laboratory using JENWAY Flame
photometer (PFP 7 model) for major cations (Na + and K), while JENWAY Colorimeter (6015 model) was used
for analysing SO42-, and EDTA Titrimetric method for Ca2+, HCO3- and Cl- as well as total hardness and alkalinity.
The concentrations of the major ions and the ionic balance of the chemical analyses were combined in validating
the data for further analysis.
2.2 Data preparation
The data obtained from laboratory analysis were standardized to their standard scores (z-scores) by setting the
mean and standard deviation to zero and one respectively so that each of the variables will have equal weight in
the multivariate statistical analyses. This helps in minimizing the effects of different units and variance of
variables by rendering the data dimensionless (Liu, et al., 2003; Singh et al., 2004). Data were evaluated to screen
for outliers and assess normality and linearity using Mahalanobis distance. A scatterplot matrix reveals fairly
normal distributions and linear relationship among variables. The results were subjected to statistical analysis at
5% level of significance (i.e. α=0.05).
2.3 Multivariate data analysis
Multivariate sequential analysis which employs standard methods in a way that each step builds on the prior
analysis was adopted. The effort is expected to provide increasing confidence and greater insight into the
hydrochemical evolution of the study area, culminating in a robust interpretation. Q-mode Hierarchical cluster
analysis (HCA) was performed with assumptions of homoscedasticity of variables (Mertler and Vannatta, 2005).
Samples were grouped into distinct clusters of hydrogeologic and statistical significance. The Group Average
(Unweighted Pair-Group) linkage and Euclidean distance methods were used. R-mode factor analysis (FA) by
principal component extraction method with varimax rotation and Kaiser normalization was employed to reduce
bulk hydrochemical data into components that explain possible underlying structures existing among the clustered
variables.
Three steps were followed in carrying out the FA, viz: computation of correlation matrix by determining the
correlation coefficient, estimation of the factor loadings, and factor rotation and interpretation. For the initial
factor extraction, Kaiser’s (1958) eigenvalue criterion of retaining only those components/factors whose
eigenvalues are greater than 1 was adopted. Other criteria used to determine the appropriate number of
components to retain include, scree plot, variance and residuals. Kaiser’s varimax rotation, an orthogonal rotation
procedure which produces a set of component loadings having the maximum variance of the squares of the
loadings is used in conducting the principal component analysis to make the factor solutions more interpretable
without altering the underlying mathematical structure (Mertler and Vannatta, 2005). The scheme operates by
adjusting the factor loadings to either near ±1 or near zero which is a way of maximizing the variance of the
loadings on the factors. Tests indicating factorability such as Kaiser-Meyer-Olkin (KMO) and Bartlett’s test were
conducted and the various indicators were good, and the residuals indicate that the factor solution was a good one
too. SPSS® Statistics 17.0 (SPSS Inc., 2008) was used for the multivariate statistical analysis.
3.0 Results and Discussion
Table 1 gives the summary of the result of laboratory analysis which shows that the water in the area is of
moderate mineralization on the basis of the moderate TDS values ranging from 50mg/l to 297mg/l.
3.1 Statistical Analysis
The univariate descriptive statistical overview of the hyrochemical data of groundwater sampled in August, 2010
including major cations and anions is presented in Table 2. The distributions of the water quality parameters were
assessed by determining minimum, maximum, mean and standard deviation for each of the 11 variables. Na+ and
K+ are relatively high in abundance compared with Ca 2+ and Mg2+. This may be responsible for the moderate
TDS, very low hardness and alkalinity of the water samples. Ca 2+ and Mg2+ values are low, and range from 0.9
mg/l to 5.7 mg/l and 0.2mg/l to 0.92mg/l, with means.d. of 2.431.15mg/l and 0.50.2mg/l respectively. Na+ and
K+ values range from 12.0mg/l to 76.0mg/l and 15.0mg/l to 88.0mg/l, with means.d. of 33.8315.67mg/l and
31.0916.07mg/l respectively.
40
© Centre for Promoting Ideas, USA www.aijcrnet.com

The total dissolved solids (TDS) in the water samples varied over a wide range of 50 to 297 mg/l with an overall
mean s.d. of 126.26 62.16mg/l. The pH values varied between 5.42 and 6.52 with means.d. of 5.990.31.
3.2 Q-and R-mode Analyses
Q-mode analysis which grroups samples on the basis of similarities in multidimensional space was employed to
classify the dataset into clusters and shown as a dendogram (Fig. 3). From the dendogram, 3 preliminary clusters
were selected by visual inspection to depict different hydrochemical facies. Generally, the order of enrichment in
all the chemical constituents is such that cluster 3 > cluster 2 > cluster 1 (Table 3). The correlation matrix of these
groundwater quality parameters was composed for R-mode factor analysis (Table 4) using principal component
extraction method with varimax rotation and Kaiser normalization. Three components were retained and
interpreted on the basis of Kaiser’s criterion of eigenvalues >1.0 (Tables 5 & 6) and indications from the scree
plot (Fig. 4). The three components together explain 83.48 % of the total variance in the dataset (Table 7). Factor
1 shows high positive loadings of Total hardness (TH), Ca2+, Mg2+, Cl- and TDS, and explains 35.82 % of the total
variance. It indicates atmospheric controls and silicate mineral weathering processes. Factor 2 with high loadings
of Na+, K+, HCO3- and Alkalinity, and average loadings of Total hardness, Ca 2+ and Mg2+ concentrations explains
35.41 % of the total variance. It reflects atmospheric controls and speciation reactions producing inorganic carbon
ions in solution. Factor 3 associated with high pH explains 12.24 % of the total variance. It reflects silicate
mineral weathering processes associated with increased pH. There is a very high degree of consistency between
the results of factor analysis by principal component extraction and those of HCA. The plot of the first two factor
scores produced three distict clusters which correspond well with the clusters produced by HCA (Fig. 5).
3.3 Hydrochemistry
Plots of hydrochemical data on Piper (1944) trilinear diagram (Fig. 6) show the distribution of chemical
constitiuents (cations and anions) in groundwater within the the area. This was accomplished by using the
GW_Chart package (USGS, 2000). From the plot, three different water types were identified in the study area,
viz: Na + - K + – Cl- - SO42- - HCO3-, Na+- K+ - HCO3- - Cl- - SO42-, and Na+ - Ca2+ - Cl- - SO42- - HCO3- water
types. The water types tend towards sodium chloride bicarbonate which is consistent with expected water-rock
interactions in such basement complex terrain (Bala and Onugba, 2001; Ehinola, 2002). These water types would
probably have resulted from precipitation of meteoric water through the different lithologic units (granite gneiss,
porphyritic gneiss, pelitic schist, charnockite, and migmatite gneiss) in the area. The Na+ and K+ may have
resulted from silicate mineral weathering while Cl - may have originated from evapotranspiration process, and
HCO3- may be inorganic carbon ions produced when infiltrating water reacted with soil CO2. The order of
dominance of cations and anions is Na+ > K+ > Ca2+ > Mg2+ and HCO3- > Cl- > SO42- respectively. The spatial
distribution of statistical water groups in the study area is as presented in Figure 7. It is observed that water
samples in the same statistical group are in close proximity spatially.
Generally, there is a high degree of spatial, statistical and hydrochemical coherence (Figs. 3, 5 & 7). The low TDS
group 1 water samples occur at relatively highest elevations, while Group 2 water samples which are
characterized by intermediate TDS values occur at lower elevations than group 1. Group 3 water samples which
are associated with the highest TDS values occur along topographic lows and flowpaths where more water-rock
and soil-rock interactions that would increase TDS is expected. This agrees with the findings of Thyne et al.
(2004).
4.0 Conclusions
This study has employed an integrated approach to groundwater hydrochemical system characterisation using
standard multivariate statistical techniques, spatial analysis, geographic information systems (GIS) and traditional
graphical methods in a sequential manner in Shagari area, southwestern Nigeria. The study showed that
atmospheric, geologic and hydrogeologic controls are possibly behind the geochemical processes responsible for
variation in water quality in area. Three different water types were identified in the study area, viz: Na + - K + –
Cl- - SO42- - HCO3- water type, Na+- K+ - HCO3- - Cl- - SO42- water type, and Na+ - Ca2+ - Cl- - SO42- - HCO3- water
type. The water types tend towards sodium chloride bicarbonate which is consistent with expected water-rock
interactions in such basement complex terrain.

41
American International Journal of Contemporary Research Vol. 3 No. 3; March 2013

Generally, there is a high degree of spatial, statistical and hydrochemical coherence. The study had demonstrated
the integration of the strengths of multivariate statistical, spatial, and conventional graphical methods in a
sequential manner for a robust interpretation of hydrochemical dataset.

Acknowledgments
The Department of Applied Geology, The Federal University of Technology, Akure (FUTA), Nigeria supported
this study by providing analytical laboratory facilities. The authors acknowledge the assistance of Mrs. Tolulope
Ogunsuyi (Marine Science and Technology, FUTA) for assisting with chemical analyses.

References
Adekoya, J. A., Kehinde-Philips, O. O. and Odukoya, A. M. (2003). Geological distribution of mineral resources in
southwestern Nigeria. In: Prospects for investment in mineral resources of southwestern Nigeria (Elueze AA,
ed.), pp. 1 – 13.
Bala, A. E. and Onugba, A. (2001). Preliminary chemical assessment of groundwater in the Basement Complex Area
within Bunsuru and Gagere Sub-basins, Northwestern Nigeria. Journal of Mining and Geology. 37(1): 45-52.
Belkhiri, L., Boudoukha, A., Mouni, L. and Baouz, T. (2010). Multivariate statistical characterization of groundwater
quality in Ain Azel plain, Algeria. African journal of Environmental Science and Technology. 4(8): 526-534.
Davis, J. C. (1973). Statistics and data analysis in geology. John Wiley and Sons, Inc., New York 550pp.
Ehinola, A. E. (2002). Hydrochemical characteristics of groundwater in parts of the Basement Complex of
Southwestern, Nigeria. Journal of Mining and Geology. 38(2): 125-133.
Elueze, A. A., Ephraim, B. E. and Nton, M. E. (2001). Hydrochemical assessment of surface water in parts of
southwestern Nigeria. Mineral Wealth. 119: 45-58.
Kaiser, H. F. (1958). The varimax criterion for analytical rotation in factor analysis.Psychometrika 23: 187-200.
Liu, C. W., Lin, K. H., & Kuo, Y. M. (2003). Application of factor analysis in the assessment of groundwater quality in
a blackfoot disease area in Taiwan. Science of the Total Environment. 313: 77–89.
Locsey, K. L. and Cox, M. E. (2003). Statistical and hydrochemical methods to compare basalt- and basement rock-
hosted groundwaters: Atherton Tablelands, north-eastern Australia. Environmental Geology. 43: 698-713.
Mertler, C. A. and Vannatta, R. A. (2005). Advanced and Multivariate statistical methods: Practical application and
interpretation (3rd ed) Pyrczak, Glendale, CA. ...pp.
Olobaniyi, S. B. and Owoyemi, F. B. (2006). Characterization by factor analysis of the chemical facies of groundwater
in the Deltaic plain sands aquifer of Warri, western Niger Delta, Nigeria. African Journal of Science and
Technology. 7(1): 73-81.
Olorunfemi, M. O., Ojo, J. S. and Akintunde, O. M. (1999). Hydrogeophysical evaluation of ground water potentials of
Akure metropolis, southwestern Nigerian Journal of Mining and Geology 35(2): 207-228.
Papatheodorou, G., Lambrakis, N. and Panagopoulos, G. (2007). Application of multivariate statistical procedures to
the hydrochemical study of a coastal aquifer: an example from Crete, Greece. Hydrol. Process. 21: 1482-1495.
Petalas, C., Lambrakis, N. and Zaggana, E. (2006). Hydrochemistry of waters of volcanic rocks: The case of the
volcanosedimentary rocks of Thrace, Greece. Water, air and soil pollution. 169: 375-394.
Piper, A. M. (1944). A graphical procedure in geochemical interpretation of water analysis. Trans American
Geophysical Union. 25: 914-923.
Shihab, A. S. and Hashim, A. (2006). Cluster analysis classification of groundwater quality in wells within and around
Mosul city, Iraq. Journal of Environmental Hydrology. 14(24): 1-11.
Singh, K. P., Malik, A., Mohan, D., & Sinha, S. (2004). Multivariate statistical techniques for the evaluation of spatial
and temporal variations in water quality of Gomti River (India) – A case study. Water Research. 38: 3980–
3992.
SPSS Inc. (2008). SPSS Statistics 17.0, Rel. 17.0.0. Polar Engineering and Consulting. Chicago: SPSS Inc.
Thyne, G., Guler, C. and Poeter, E. (2004). Sequential analysis of hydrochemical data for watershed characterization.
Ground Water 42(5): 711–723.
United States Geological Survey. (2000). GW_Chart Version 4. US Geological Survey.
Yidana, S. M. (2011). Hydrochemical Characterization of Aquifers Using Sequential Multivariate Analyses and
Geographic Information Systems in a Tropical Setting Journal of Environmental Engineering. 137(4): 258-
272.

42
© Centre for Promoting Ideas, USA www.aijcrnet.com

Table 1: Water Analysis Result

SO4 (mg/L)
Longitude

Alkalinity
Location

(E 005O)

Latitude
(N 07O)

(mg/L)

(mg/L)

(mg/L)

(mg/L)

(mg/L)

(mg/L)

(mg/L)

(mg/L)
HCO3
TDS

Mg
TH
pH

Na

Ca

Cl
K
1 11.871 16.898 129.00 5.93 4.40 38.00 52.00 2.30 0.62 1.30 2.70 0.05 2.50
2 11.483 17.150 157.00 5.99 3.70 30.00 42.00 2.50 0.50 2.40 2.20 0.10 1.80
3 11.413 17.252 148.00 5.87 3.10 22.00 34.00 2.10 0.40 2.60 1.60 0.06 1.50
4 11.437 16.978 146.00 6.10 4.70 20.00 31.00 2.70 0.64 2.00 2.10 0.05 1.50
5 11.288 17.034 73.00 5.98 2.90 32.00 18.00 1.00 0.50 1.00 1.40 0.02 1.00
6 11.257 16.953 120.00 6.01 3.90 34.00 40.00 1.90 0.60 1.60 2.00 0.03 1.70
7 11.307 16.768 86.00 5.93 3.60 15.00 24.00 2.10 0.50 1.40 4.10 0.04 1.20
8 11.357 16.470 64.00 5.99 2.20 18.00 26.00 1.80 0.24 1.00 1.30 0.08 1.22
9 10.820 16.918 60.00 5.88 1.40 22.00 19.00 1.00 0.20 1.20 0.90 0.01 1.00
10 10.971 16.730 57.00 5.94 2.10 24.00 32.00 1.60 0.26 0.70 1.30 0.04 1.30
11 10.966 16.595 64.00 6.09 2.12 20.00 28.00 1.62 0.23 0.80 1.10 0.20 1.32
12 10.949 16.486 91.00 6.15 3.10 38.00 51.00 1.81 0.42 1.00 2.20 0.02 2.20
13 11.051 16.346 135.00 5.83 3.60 15.00 21.00 1.83 0.52 2.00 1.10 0.03 1.20
14 11.727 16.215 234.00 6.49 5.80 24.00 32.00 4.20 0.70 3.60 1.70 0.10 1.40
15 12.034 16.611 261.00 6.61 7.70 88.00 76.00 5.71 0.92 3.30 5.20 0.12 4.30
16 11.792 16.631 137.00 6.09 5.10 32.00 42.00 3.20 0.70 1.70 1.40 0.03 1.70
17 10.820 17.050 105.00 6.21 4.90 40.00 53.00 2.60 0.71 1.20 2.50 0.04 2.20
18 10.728 17.292 102.00 6.58 4.20 31.00 50.00 3.60 0.50 0.80 2.20 0.01 2.40
19 11.077 17.143 147.00 5.98 4.80 48.00 12.00 3.90 0.54 2.20 1.90 0.15 0.50
20 11.199 17.303 103.00 5.70 3.10 16.00 13.00 2.00 0.22 1.50 1.20 0.05 0.70
21 10.884 16.388 170.00 6.30 3.70 26.00 32.00 2.20 0.52 2.80 1.60 0.09 1.40
22 11.824 16.198 262.00 6.55 6.10 30.00 37.00 3.50 0.83 3.90 1.70 0.11 1.50
23 11.976 16.392 83.00 5.64 1.60 54.00 14.00 0.90 0.23 1.50 0.60 0.06 0.70

Table 2: Descriptive statistics of water quality data from Shagari

Variable Min Max Mean Std. Deviation

TH 1.40 7.70 3.82 1.52
Ca2+ 0.90 5.70 2.43 1.15
Mg2+ 0.20 0.92 0.50 0.20
Cl- 0.70 3.90 1.80 0.92
Na+ 12.00 76.00 33.83 15.67
K+ 15.00 88.00 31.09 16.07
HCO3- 0.60 5.20 1.91 1.02
Alkalinity 0.50 4.30 1.57 0.79
pH 5.42 6.52 5.99 0.31
TDS 50.00 297.00 126.26 62.16
SO42- 0.01 0.20 0.06 0.05

Table 3: Mean water chemistry of the HCA water clusters

Cluster TH Ca2+ Mg2+ Cl- Na+ K+ HCO2- Alk. pH TDS SO42-
Cluster 1 2.86 1.87 0.36 1.64 24.17 27.25 1.35 1.13 5.94 105.08 0.07
Cluster 2 4.24 2.53 0.59 1.38 42.88 30.75 2.40 1.93 6.16 114.50 0.03
Cluster 3 6.53 4.47 0.82 3.60 48.33 47.33 2.87 2.40 6.55 252.33 0.11

43
American International Journal of Contemporary Research Vol. 3 No. 3; March 2013

Table 4: Correlation matrix of water quality parameters from Shagari

TH Ca2+ Mg2+ Cl- Na+ K+ HCO3- Alk. pH TDS SO42-

TH 1.000
Ca2+ 0.913 1.000
Mg2+ 0.947 0.765 1.000
Cl- 0.684 0.633 0.628 1.000
Na+ 0.608 0.567 0.623 0.181 1.000
K+ 0.499 0.534 0.455 0.253 0.548 1.000
HCO3- 0.644 0.618 0.611 0.245 0.672 0.530 1.000
Alk. 0.619 0.600 0.606 0.196 0.962 0.630 0.750 1.000
pH 0.517 0.455 0.537 0.258 0.527 0.012 0.338 0.468 1.000
TDS 0.783 0.680 0.769 0.925 0.329 0.270 0.311 0.306 0.434 1.000
SO42- 0.235 0.373 0.102 0.409 -0.018 0.216 0.083 0.005 -0.102 0.346 1.000

Table 5: Eigenvalues of Factors extracted through PCA, Differences between Factors and proportion of
variance explained by factors

Factors 1 2 3 4 5 6 7 8 9 10 11
Eigenvalues 6.080 1.889 1.213 0.613 0.417 0.313 0.221 0.178 0.043 0.022 0.010
Difference 4.191 0.676 0.600 0.196 0.104 0.092 0.043 0.135 0.021 0.012
Proportion 0.553 0.172 0.110 0.056 0.038 0.028 0.020 0.016 0.004 0.002 0.001
Cummulative 0.553 0.725 0.835 0.891 0.929 0.957 0.977 0.993 0.997 0.999 100.000
Eigenvalues of the Correlation Matrix; Total = 11; Average = 1

Table 6: Variance explained for factors retained before and after varimax rotation

Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings

Factor Total % of Cumulative Total % of Cumulative
Variance % Variance %
1 6.080 55.273 55.273 3.941 35.824 35.824
2 1.889 17.173 72.446 3.895 35.407 71.231
3 1.213 11.029 83.475 1.347 12.244 83.475

Table 7: Varimax rotated factor loadings and communalities

Variable Factor 1 Factor 2 Factor 3 Communalities
TH 0.785 0.554 0.067 0.928
Ca2+ 0.710 0.567 -0.091 0.833
Mg2+ 0.739 0.529 0.193 0.864
Cl- 0.918 0.038 -0.183 0.877
Na+ 0.194 0.874 0.269 0.874
K+ 0.125 0.778 -0.402 0.783
HCO3- 0.235 0.812 0.045 0.716
Alk. 0.170 0.928 0.181 0.924
pH 0.473 0.275 0.687 0.771
TDS 0.951 0.134 -0.014 0.922
SO42- 0.409 0.018 -0.723 0.690
% Variance 35.824 35.407 12.244 83.475

44
© Centre for Promoting Ideas, USA www.aijcrnet.com

Figure 1: Topographical map of study area (Fed. Survey Nig., 1964)

Figure 2: Geological map of study area

45
American International Journal of Contemporary Research Vol. 3 No. 3; March 2013

Figure 3: Dendogram of hydrochemical data depicting associations between groundwater samples from
Shagari.

Figure 4: Scree plot for selecting number of factors to retain and interpret.

46
© Centre for Promoting Ideas, USA www.aijcrnet.com

Figure 5: Distribution of HCA-derived classification of samples using factor analysis results

Figure 6: Piper trilinear diagram depicting variation in groundwater chemistry of Shagari area

47
 American International Journal of Contemporary Research Vol. 3 No. 3; March 2013

Figure 7: Spatial distribution of HCA-derived water groups for Shagari area.

48

View publication stats