Integrated Environmental Quality Assessments of Surface Water around Obajana Cement Production Area

Due to industrialization, there is enormous amount of heavy metals been released from anthropogenic sources into the environment. Heavy metals are considered as one of the main sources of environmental pollution since they have significant effect on the ecological quality and water in particular. These pollutants are hazardous to consumers of water that have significant quantity of these heavy metals. The population most exposed to cement polluted water includes workers in cement factories, families of workers living in Staff houses of factories like in Obajana and other neighborhood habitations. The Obajana cement factory consists of cement kilns/coolers with clinkers. The kilns are equipped with pre-heaters and Electro-Static Precipitators (ESP). The facility has raw mills, crushing operations, cement mills that are potential source of pollutants into the water bodies. Storage silos, conveyors, vehicular travel, and other unquantified fugitive source of water contamination exist in the factory. Monitoring the contamination of water with respect to heavy metals is of interest due to their influence on humans, animals and to some extent plants. A good approach to estimate how much of the water is impacted is by using the heavy metal pollution index and metal index for metal concentrations above the control points in water bodies around Obajana cement.


INTRODUCTION
Obajana plant has two production lines with a combined capacity of 15,000 metric tons of cement per day.All the raw materials-marble, schist, clay and laterite are available within the mines.The raw materials are spread over 900 ha with an average depth of 130 m.Transport of the raw materials from the quarry to the plant site is through a conveyor belt line of 8.0 km long.The raw materials are crushed at the mines into smaller pebbles and transported by the conveyor belt to the raw material storage shed at the factory.The raw meal is passed through a pre-heater where it is heated up to 900°C before it is fed into kilns where the raw meal is heated up to a temperature of 1400°C.The resulting product-clinker-is cooled in coolers and stored in clinker silos.The clinker is mixed with gypsum and ground in the cement mills to produce cement.The cement is stored in cement silos.From the cement silos, the cement is distributed through automatic truck loaders.
The potential impacts of production could result from construction and operation of the ropeway; on-site storage of the crushed limestone, clay, laterites and gypsum; dust from grinding and mixing of raw material; dust and combustion by products from the calciner; dust from clinker grinding and bagging; wastewater used for bearing cooling and sewage discharge from staff housing and offices and disposal of solid wastes.
The major constituents in dust from cement manufacturing plants are alumina, silica, metallic oxides and clay, trace amounts of organic chemicals (dioxins and furans), heavy metals (cadmium, lead and selenium) and radio nuclides (Ziadat et al., 2006;Nitish, 2008).Typical gaseous emissions to air from cement manufacturing plants include Nitrogen Oxides (NOx), Sulphur dioxide (SO x ), Carbon Oxides (CO and CO x ) and dust.Effluent discharge is generated from cooling process equipment, wet scrubbing kiln stack emission for recovering cement kiln dust and runoff water from the outdoor areas.Effluents contain mainly dissolved solids (potassium and sodium hydroxides, chlorides and sulphates), suspended solids (calcium carbonate), waste heat and heavy metals (Legator et al., 1998).
The objective of the study is to analyzed the water samples, use heavy metal pollution and metal indexing in conjunction with factor and cluster analysis to evaluate cement impacts on surface water.

Geology of study area:
The study area lies within the Benin-Nigeria shield, situated in the Pan-African mobile zone extending between the ancient Basements of West African and Congo Cratons in the region of Late Precambrian to Early Palaeozoic orogenies (Rahaman, 1976;Odigi, 2002;Ekwueme, 2003).The Basement Complex rocks of Nigeria are composed predominantly of migmatite gneiss complex; slightly migmatised to unmigmatised paraschists and metaigneous rocks; charnockitic, older granite suites and unmetamorphosed dolerite dykes (Rahaman, 1976).

MATERIALS AND METHODS
Sampling was carried out in February (dry season).A total of (8) water samples were collected (Fig. 2).Sampling was random but evenly distributed.Samples were collected from mid-point and a foot below surface water in duplicates-one for heavy metal and the other for anion analyses.Samples were filtered as soon as they were collected using cellulose nitrate filter with pores of 0.45 micron diameter.Polyethylene plastic bottles were used as sample containers.New bottles were cleaned with strong-metal free acid.The containers were rinsed with sample water prior to collection.Sufficient air space was allowed and sample stored upright.Teflon lined caps were screwed on tightly to prevent leakage.Water samples for cations and heavy metal analyses were acidified with metal free HNO 3 to a pH of 1-2.The samples were stored between 1 and 4°C on cool ice packs from the field to the Lab.for analyses.
Analytical methods: Insitu measurements of temperature, pH, Tds and EC were carried out intrusively with appropriate probes.Spectrophotometer (Model Genesys 20) was used to determine the concentrations of K, Na, Ca, NO 3 and SO 4 while AAS (Model 210 VGP) was used to determine the concentrations of Mg, Pb, Zn, Ni, Cu, Cd and Fe.Titration method was used for the determination of the concentrations of Cl and alkalinity.All analyses were performed according to APHA (2000) in the Dept. of Soil Science Lab, Faculty of Agriculture, Kogi State University, Anyigba.
Data evaluations: Heavy metal pollution index and metal index approaches were used for this study.HPI is a method that rates the aggregate influence of individual heavy metal on the overall quality of water.It is defined as Wi taken as inversely proportional to the recommended standard (Si) for each parameter.HPI model is given as: where, Qi = Subindex of the i th parameter Wi = The unit weightage of i th parameter n = The number of parameters considered The subindex (Qi) of the parameter is calculated by: where, Mi = The monitored value of heavy metal of the i th parameter Ii = The ideal/baseline value of i th parameter Si = The standard value of i th parameter The sign (-) indicates the numerical difference of the two values, ignoring the algebraic sign (Prasad and Bose, 2001).The critical pollution index value is 100 (Reza and Singh, 2010;Bakan et al., 2010).
Another index used is the general Metal Index (MI) for drinking water (Caeiro et al., 2005) which takes into account possible additive effect of heavy metals on the human health that help to quickly evaluate the overall quality of drinking waters.MI = ∑ [Ci/ (MAC) i] proposed by Caeiro et al. (2005).Where MAC is maximum allowable concentration and Ci is concentration of each metal.The higher the concentration of a metal compared to its respective MAC value, the worse the quality of water.MI value >1 is a threshold of warning (Caeiro et al., 2005).
Univariate and multivariate statistical methods of analysis were also used in the study.The software SPSS 15.0 was used for statistical analysis.The correlation matrix which is based on the Pearson's correlation coefficient was utilized for displaying relationships between variables.The obtained matrix was subjected to multivariate analytical technique.Factor analysis which aims to explain an observed relationship between numerous variables in terms of simple relations was applied.Cluster analysis was also used for investigating the similarities between variables found in River Mimi and its tributaries.Evaluations of similarity were based on the average linkage between groups (Praveena et al., 2007;Harikumar and Jisha, 2010).

RESULT ANALYSIS
Analytical results expressed in milligram per liter (mg/L) were log-transformed and statistically treated using SPSS version 15.0 software.Factor and cluster analyses were computed for the variables using the SPSS version 15.0 software.The total variance in each factor is re-expressed by the eigenvalue provided by the principal component solution as an initial set of uncorrelated linear transformations of original variables.The factor loadings, which can be regarded as combination between the elements, were then computed after rotating the original principal component solution according to Kaiser's varimax criterion (Praveena et al., 2007).In this study, factors with eigenvalues greater than 1 were retained and only variables with loadings greater than 0.40 were considered significant groups of a particular factor.Cluster analysis was also used for investigating the similarities between major variables and heavy metals from in the samples.Evaluations of similarity were based on the average linkage between groups (Praveena et al., 2007;Aprile and Bouvy, 2008).
Table 1 shows the summary of parameters measured in Obajana dry season water samples.The order of mean concentration among the physiochemical is Tds>alkalinity>temperature>pH>Ec.pH of the area is generally alkaline because of the presence of Carbonates (CO 3 ).The major cations have this order of concentration: Mg>K>Na>Ca.Because of cement production and the geology of the area, this order among the major cations is expected.The order observed also among the anions is: CO 3 >NO 3 >SO 4 >Cl.CO 3 is expected to dominate the anions because of its presence in marble while NO 3 maybe from fertilizer application to improve yields.The heavy metals show this concentration trend: Fe>Ni>Pb>Cu>Cd/Zn.Fe is generally high in Nigeria because of our tropical climatic conditions (Fagbote and Olanipekun, 2010) but the relatively higher concentrations of Ni and Pb is worthy of note.
Temperature shows (Table 2) negative, weak correlations with K, Ni; moderate correlation with Ca and strong correlation with Mg.Moderate correlations were recorded with pH-alkalinity and pH-K while strong correlations exists between pH-Zn, pH-NO 3 .Alkalinity shows negative and weak correlations with Tds, weak correlations with Zn and Cl while the correlation between Fe and NO 3 is moderate.The same alkalinity correlates moderately with Ni but negative.Tds recorded very strong correlations with Ec and Na, moderate relationship with K and Pb, negative and moderate relation with Fe.Tds exhibits negative and weak relationships with Cd and Cl and weak correlation with Cu.Ec-Na shows very strong correlation.Moderate correlation exists between Ec-Pb and negative, moderate relations exist between Ec-Fe.The correlations between Ec-K, Cu-Cd are weak.Between Ec-Cd is negative.Na shows very strong correlation with Pb (r>0.8),strong but negative correlation with Cd and a moderate correlation with Cu.Weak, negative correlation also exist between Na-Fe.The relationship between Na-Pb is very strong, with Cd it is negative but strong, moderate correlation exists with Cu.Weak and negative correlation was evident between Na-Fe.Weak correlations were observed with K-NO 3 , K-CO 3 while same weak but negative correlations exists with K-Fe and K-Cl.K-Mg shows strong but negative correlation.The correlation between Ca-Mg, Ca-NO 3 is weak while that between Ca-Zn is moderate.Mg shows negative, weak correlations with Cd and Ni, weak correlation with SO 4 .Zn shows weak correlation with Cl, weak and negative correlation with SO 4 and moderate correlation with NO 3 .Interestingly, Zn shows no correlation with any of the heavy metals.Cu correlates negatively but weakly with Cd, SO 4 and CO 3 and a positive, weak correlation with Cl.The correlation between Cu-Pb is moderate.Cd correlates negatively but moderate with only Pb among the heavy metals.Lead on the other hand has weak, negative correlations with only Fe.Fe shows negative and weak correlations with Ni, SO 4 and a weak correlation with Cl.Ni correlates negatively and weakly with NO 3 and CO 3 .From the correlation (Table 2), the heavy metals shows no correlations among themselves, where it exist, they are weak and negative.This suggests anthropogenic source for the heavy metals.The correlation between Cl-CO 3 is strong (r>-0.7)but negative.Between Cl and SO 4 is weak and also negative.SO 4 correlates negatively and weakly with NO 3 .As observed with the heavy metals, the anions show no relations among themselves.Where relationship exists, they are negative and weak.This is also an indication of anthropogenic inputs for these anions.Overall, strong and positive correlations exist among the physiochemical and major cations while this is not true for the heavy metals and anions.R-mode varimax factor analysis revealed three factors.Factor one consist of high factor loadings of Cu, Cd and Pb with eigenvalue of 2.275 and total varimax of 37.914%.Factor two also consists of high loadings of Fe and Ni with eigenvalue of 1.710 and total variance of 28.497%.The association in factor three is between Zn and Cu with Zn having high, positive loading and weak factor loading for Cu.Factor three has eigenvalue of 1.296 and total variance of 21.599% (Table 3).R-mode cluster extracted two clusters.Cluster one consist of Cu, Pb, Zn and Fe while cluster two is an association between Cd and Ni (Fig. 3).
Q-mode varimax rotated factor analysis of heavy metals yield three factors.Factor one consists of high factor loadings of locations Obj07, Obj16 and Obj19 with total variance of 33.762% and eigenvalue of 2.363.Factor two has eigenvalue of 1.802 and total variance of 25.738%.Factor two is an association of locations Obj16, Obj20, Obj21 and Obj23.Obj21 has high factor loading and the rest showed weak factor loading.Factor three consists of locations Obj05, Obj19 and high loading for Obj23.Factor three has eigenvalue of 1.650 and total variance of 23.566 (Table 4).
Q-mode cluster extracted two clusters Figure 4. Cluster one consists of locations Obj07, Obj19, Obj23, Obj05 and Obj20.Cluster two on the other hand consists of locations Obj16 and Obj21.

DISCUSSION
Moderate to strong correlation exist (in few cases) between the cations and the physiochemical, a clear indication of the contribution of these ionic components to the overall Tds and Ec (Abimbola et al., 2005).Among the major cations, strong correlation also exist between Mg-K while the rest variables show weak to no correlation-suggesting diverse origin for the major cations (Abimbola et al., 2005;Olayinka and Olayiwola, 2001).The heavy metals exhibit weak to moderate correlations and in most cases negative with one another as well as with Ec and Tds.This may be attributed to the fact that dissolution of metals in the aqueous phase is controlled by the chemical character and solubility (mobility) of the respective metals (Tijani et al., 2004).Correlation between heavy metals and major cations were also very weak, moderate and negative in most cases except that between Na-Cd; Na-Pb which is strong and is a reflection of a source area deposited together under similar conditions (Tijani et al., 2004).The relationship between cations and physiochemical were weak except between pH-SO 4 ; pH-NO 3 ; alkalinity-SO 4 which were strong, an indication of same source and affinity of the physiochemical for NO 3 and SO 4 .Between major cations and anions, it is generally weak, same for between anions and heavy metals.Among the anions, strong but negative relationship was observed between CO 3 and Cl, an indication also of same source/origin of deposition (Olayinka and Olayiwola, 2001;Tijani et al., 2004).
The R-mode factor analysis of heavy metals extracted three factors.Factor one which consists of high loadings of Cu, Cd and Pb and showed fairly significant correlation is an indication of similar source related to anthropogenic inputs considering the presence of Cd and dual sources for Cu (Shakeri et al., 2009).Ni and Fe in factor two shows high factor loadings, weak and negative relationship (r = -0.447)indicating that the dynamics of the two heavy metals though linked are not closely related (Idowu et al., 2007).Zn and Cu in factor three shows high and moderate factor loadings.It is a normal association but the weak relationship that exists (r = 0.371) may implies psudo anthropogenic factor (Alkarkhi et al., 2008).The R-mode cluster (Fig. 3) is characterized by Cu, Pb, Zn and Fe in the first cluster.In this cluster, Cu and Pb shows first degree of similarity reflecting their strong common sources.Zn and Fe are not related to Cu and Pb but bounded loosely to them at a farther Euclidean distance.Cluster two also show similar pattern.Though Cd and Ni are in a cluster, there seems to be little relationship which could imply diverse origin (Pathak et al., 2008).
Q-mode rotated varimax analysis extracted three factors (Table 4).High factor loadings of locations Obj07, Obj10 were recorded.Obj07 in this factor is close and under the influence of Obajana cement production activities while Obj16 and 19 are far from the factory's direct influence.Factor two has high, positive loadings of Obj21 while the rest factor loadings are low and negative for Obj23 and Obj20.All the locations in this factor are far from the influence of the factory suggesting similar concentrations (Tauhid- Ur-Rahman et al., 2008).High but negative factor loadings was recorded for location Obj23 while weak to moderate factor loadings were observed for Obj20 and Obj25 respectively.These locations also are far from the factory and hence little influence from it (Tauhid- Ur-Rahman et al., 2008).
Q-mode cluster analysis revealed two clusters.Cluster one include locations Obj07, 19, 23, 05 and 20.The closest association is between location Obj07 and Obj19 while the rest locations may not have been influenced by same process as in Obj07 and Obj19.These two locations are directly under the influence of the factory while the rest are not.Cluster two though consist of locations 16 and 21 are not strongly linked indicating a mixture of factors/influences (Ziadat et al., 2006).
The HPI value for dry season water samples is 0.43 while the MI = 114.90.This HPI imply that the water is not contaminated with respect to heavy metals.The MI value of 114.90 suggests that the water is contaminated because MI value>1 is a threshold of warning (Caeiro et al., 2005;Reza and Singh, 2010;Bakan et al., 2010).

CONCLUSION
Mean concentration of Fe, Ni and Pb were relatively higher compared to other heavy metals.Factor and cluster analysis showed various degree of anthropogenic influence on the water.Heavy metal pollution index reveals that pollution has not reached the critical level while metal index suggests low water quality.Pollution prevention at source through substitution of less hazardous materials, improved maintenance and effective production processes is recommended for a sustainable environment and healthy living.

Table 1 :
Descriptive statistics of obajana dry season water

Table 2 :
Correlation matrix of obajana dry season water samples

Table 3 :
R-mode varimax factor analysis

Table 5 :
HPI and MI indices of obajana dry season water samples