Nitrate contamination of groundwater in two areas of the Cameroon Volcanic Line (Banana Plain and Mount Cameroon area)

Abstract

Water containing high concentrations of nitrate is unfit for human consumption and, if discharging to freshwater or marine habitats, can contribute to algal blooms and eutrophication. The level of nitrate contamination in groundwater of two densely populated, agro-industrial areas of the Cameroon Volcanic Line (CVL) (Banana Plain and Mount Cameroon area) was evaluated. A total of 100 samples from boreholes, open wells and springs (67 from the Banana Plain; 33 from springs only, in the Mount Cameroon area) were collected in April 2009 and January 2010 and analyzed for chemical constituents, including nitrates. The average groundwater nitrate concentrations for the studied areas are: 17.28 mg/l for the Banana Plain and 2.90 mg/l for the Mount Cameroon area. Overall, groundwaters are relatively free from excessive nitrate contamination, with nitrate concentrations in only 6 % of groundwater resources in the Banana Plain exceeding the maximum admissible concentration for drinking water (50 mg/l). Sources of NO3 ^sup -^ in groundwater of this region may be mainly anthropogenic (N-fertilizers, sewerage, animal waste, organic manure, pit latrines, etc.). Multivariate statistical analyses of the hydrochemical data revealed that three factors were responsible for the groundwater chemistry (especially, degree of nitrate contamination): (1) a geogenic factor; (2) nitrate contamination factor; (3) ionic enrichment factor. The impact of anthropogenic activities, especially groundwater nitrate contamination, is more accentuated in the Banana Plain than in the Mount Cameroon area. This study also demonstrates the usefulness of multivariate statistical analysis in groundwater study as a supplementary tool for interpretation of complex hydrochemical data sets.

Appl Water Sci (2014) 4:99113 DOI 10.1007/s13201-013-0134-x

ORIGINAL ARTICLE

Nitrate contamination of groundwater in two areasof the Cameroon Volcanic Line (Banana Plain and Mount Cameroon area)

Andrew Ako Ako Gloria Eneke Takem Eyong Jun Shimada Katsuaki Koike

Takahiro Hosono Kimpei Ichiyanagi Akoachere Richard Beatrice Ketchemen Tandia

George Elambo Nkeng Ntankouo Njila Roger

Received: 27 June 2012 / Accepted: 11 October 2013 / Published online: 1 November 2013 The Author(s) 2013. This article is published with open access at Springerlink.com

Abstract Water containing high concentrations of nitrate is unt for human consumption and, if discharging to freshwater or marine habitats, can contribute to algal blooms and eutrophication. The level of nitrate contamination in groundwater of two densely populated, agro-industrial areas of the Cameroon Volcanic Line (CVL) (Banana Plain and Mount Cameroon area) was evaluated. A total of 100 samples from boreholes, open wells and springs (67 from the Banana Plain; 33 from springs only, in the Mount Cameroon area) were collected in April 2009 and January 2010 and analyzed for chemical constituents, including nitrates. The average groundwater nitrate concentrations for the studied areas are: 17.28 mg/l for the Banana Plain and 2.90 mg/l for the Mount Cameroon area. Overall, groundwaters are relatively free from excessive nitrate contamination, with nitrate concentrations in only 6 % of groundwater resources in the Banana Plain exceeding the maximum admissible concentration for

drinking water (50 mg/l). Sources of NO3- in groundwater of this region may be mainly anthropogenic (N-fertilizers, sewerage, animal waste, organic manure, pit latrines, etc.). Multivariate statistical analyses of the hydrochemical data revealed that three factors were responsible for the groundwater chemistry (especially, degree of nitrate contamination): (1) a geogenic factor; (2) nitrate contamination factor; (3) ionic enrichment factor. The impact of anthropogenic activities, especially groundwater nitrate contamination, is more accentuated in the Banana Plain than in the Mount Cameroon area. This study also demonstrates the usefulness of multivariate statistical analysis in groundwater study as a supplementary tool for interpretation of complex hydrochemical data sets.

Keywords Nitrate contamination Groundwater

Banana Plain Mount Cameroon area Cameroon Volcanic

Line

A. A. Ako (&) G. E. T. Eyong

Hydrological Research Centre Yaound, P.O. Box 4110, Yaound, Cameroone-mail: [email protected]; [email protected]

J. Shimada K. Ichiyanagi

Graduate School of Science and Technology, Kumamoto University, Kurokami 2-29-1, Kumamoto 860-8555, Japan

K. KoikeLaboratory of Environmental Geosphere Engineering, Department of Urban Management, Graduate School of Engineering, Kyoto University, Katsura C1- 2-215, Kyoto 615-8560, Japan

T. HosonoPriority Organization for Innovation and Excellence, Kumamoto University, Kurokami 2-29-1, Kumamoto 860-8555, Japan

A. RichardDepartment of Geology, University of Buea, P.O. Box 63, Buea, Cameroon

B. K. TandiaDepartment of Earth Sciences, University of Douala, P.O. Box 12407, Douala, Cameroon

G. E. NkengNational Advanced School of Public Works Yaound, P.O. Box 510, Yaound, Cameroon

N. N. RogerWater Resources Management Laboratory, Department of Agric Engineering, University of Dschang, P.O. Box 222, Dschang, Cameroon

123

100 Appl Water Sci (2014) 4:99113

Introduction

Nitrate contamination in groundwater is a common problem in many part of the world arising from diffuse reasons, e.g., intensive agriculture, unsewered sanitation in densely populated areas or from non-point sources such as irrigation of land by sewage efuents. Nevertheless, the heavy use of nitrogenous fertilizers in cropping system is the largest contributor to anthropogenic nitrogen in ground-water worldwide (Suthar et al. 2009). In particular, shallow aquifers in agricultural elds are highly vulnerable to nitrate contamination, due to the widespread application of fertilizers and manure (Bhlke 2002; Kyoung-Ho et al. 2009).

In many groundwater basins, urbanization has created a growing demand for drinking water, while long histories of agricultural activity have left aquifers potentially at risk from NO3- (Moore et al. 2006). Some of the major concerns of nitrate contamination/loading to both groundwater and surface water systems include health risks to humans through drinking water (i.e., methemoglobinemia and potential carcinogenic effects) and degradation of the local ecosystems (i.e., excessive plant and algal growth) (Murgulet and Tick 2008).

Extensive research has indicated that agricultural practices may cause nitrate contamination to be high so as to exceed the maximum acceptable level for drinking water (Bhlke 2002). The maximum admissible concentration (MAC) in most countries is within the range of 4550 mg/l, which is equivalent to 1011 mg/l nitrate as nitrogen. Also, the World Health Organization (WHO) indicates a guideline value (GV) of 50 mg/l (11 mg/l nitrate as nitrogen) for drinking water (WHO 2004), which has been integrated into Cameroon Water Quality Norms for mineral and drinking waters (ANOR 2001, 2003). Various authors have studied the issue of nitrate contamination of groundwater resources in Cameroon, especially in the north of Cameroon (Njitchoua and Ngounou Ngatcha 1997; Djoret 2000; Ngounou Ngatcha and Djoret 2010), which is the main agro-pastoral zone of Cameroon. Ako et al. (2011) also reported about nitrate contamination of groundwater and surface water resources in the Banana Plain. Endeley et al. (2001) in a study of water sources around Mount Cameroon made up of weathered volcanic and sedimentary rocks like the Banana Plain showed that these waters did contain trace elements and a dominance of Na? and K? as major cations closely associated with nitrates.

Currently, there is a lack of information and understanding of NO3- input and its effect on the quality of groundwater resources and ecosystem health within these densely populated, agro-industrial areas of the Cameroon Volcanic Line (CVL). Therefore, it is essential to conduct

detailed assessments of the degree of nitrate contamination of their groundwater resources.

The objective of this research was to apply hydro-chemistry and principal component/factor analysis for the identication of factors responsible for nitrate contamination of groundwater in the Banana Plain and Mount Cameroon area, two important areas of the Cameroon Volcanic Line. Research on these areas can contribute to the understanding of nitrate contamination of groundwater in active and dormant volcanic areas under both intense urban development and agro-industrial activities.

Materials and methods

Study areas

The study areas are located along the Cameroon Volcanic Line, with the Banana Plain located in Mbanga, Njombe-Penja, (43004530N; 93709500E) in the Mungo Division, 70 km north of Douala, economic capital of Cameroon. The Mount Cameroon area is located between 4090

4130N and 91609210E. These areas will be referred to as Banana Plain and Mount Cameroon, respectively (Fig. 1).

The development of plantation agriculture and the fertility of the soils in the Banana Plain and the Mount Cameron area have created dense human settlements. In the Banana Plain, the towns of Mbanga, Njombe and Penja are all ourishing because of plantation agriculture. Mbanga town has a population of about 140,000 and covers a surface area of about 544 km2 (population density of 257 inhabitants/km2), while Njombe and Penja have about 50,800 inhabitants with a surface area of about 260 km2 (195 inhabitants/km2) (Mbanga Rural Council 2008). The populations of these towns have witnessed rapid increases between 1987 and 2006 (the population of Njombe and Penja increased from 33,000 to 50,000 inhabitants). The population in the Mount Cameroon area is estimated at about 450,000 people of whom two-thirds live in urban and semi-urban areas, while the rest live in villages. The population density in the Mount Cameroon area is about 150 inhabitants/km2. The settlement pattern forms a closed ring around the foot of the mountain with no permanent settlements on altitudes above 1,500 m. At lower altitudes, extensive plantations of bananas, palms, rubber and the creation of agglomeration have totally replaced the primary forest. Mount Cameroon is unique for having a relatively unbroken sequence of natural vegetation from lowland evergreen forest almost at sea level right up to about 2,000 m then to subalpine prairies near its summit.

Geologically, the CVL is a 1,600 km-long Tertiary Recent intraplate alkaline volcanic province that trends NESW, from the Atlantic Ocean through the Gulf of

123

Appl Water Sci (2014) 4:99113 101

Fig. 1 Location and geology of Banana Plain (Mbanga, Njombe and Penja) and the Mount Cameroon area. Inset shows a map of the CVL modied after Fitton and Dunlop (1985); Druelle et al. (1987); Marzoli et al. (1999); Oyebog et al. (2012)

123

102 Appl Water Sci (2014) 4:99113

Guinea onto African continental lithosphere (Fitton and Dunlop 1985). Recent isotope (K/Ar and Ar/Ar) ages have shown that the Cameroon Line volcanism is essentially Cenozoic (Marzoli et al. 1999; Aka et al. 2004). Geologic settings in the two areas are quite contrasting, with Mount Cameroon being an active volcano made up of younger basaltic rocks (\3 Ma). The Banana Plain is composed of older volcanic rocks (*3 Ma) mixed with sedimentary rocks.

Rapid increase in population in these two areas was not accompanied by the development of basic infrastructure (water supply and sanitation). In the Banana Plain, \10 % of the population is connected to the potable water network of CAMWATER (the national potable water utility) and \5 % have adequate sanitation facilities (GTZ 2006). Only about 35 % of the population in the Mount Cameron area have access to the potable water network (Folifac et al. 2009). Thus, many people rely on bottled water for safety purposes for infants (for formula preparation, drinking water and reconstitution of food), health reasons and for drinking water. Most of the bottled water plants are located on the volcanics along the CVL (Fig. 1) and they obtain bottled mineral water from springs and boreholes (Oyebog et al. 2012). There are also many springs, boreholes and wells from which the local people fetch drinking water and for other domestic purposes. Groundwater (springs) is also the main source of irrigation water supply in these important agro-industrial zones.

In the Banana Plain, the climate is equatorial type inuenced by monsoon winds (July to September) during the rainy season with maximum rainfall in September with about 175200 rainfall days per year. There are two distinct seasons, a dry season from November to June and rainy season from July to October. Ten-year (19851994) metrological data indicate that in Mbanga, the annual average rainfall is 1,626 mm, 3,197 mm in Njombe and 2,769 mm in Penja. Temperatures vary between 19 and 32 C with an average of 25 C (National Metrological Service 2008). The climate is humid tropical in the Mount Cameroon area characterized by extreme rainfalls and elevated temperatures all year long. The combination of high relief (4,095 m) and proximity of the sea leads to strong local climatic contrast. Mean annual temperature decreases from 2629 C at sea level to 0 C at the top of the edice and this large temperature drop is associated with a decrease in rainfall. The highest rainfall is recorded on the southwest ank of Mount Cameroon where rainfall can reach 12 m/ year. Lower rainfalls occur on the opposite ank because it is partially sheltered from the oceanic inuence (for example, rainfall at Ekona is 1,800 mm/year) (Fraser et al. 1998).

Methods

Sampling and chemical analysis

A total of 100 samples from boreholes, open wells and springs (67 from the Banana Plain; 33 from springs only, in the Mount Cameroon area) were collected in April 2009 and January 2010. Available water resources in the study areas were randomly sampled with emphasis laid on uniform spatial distribution. Locations and altitudes of selected sample sites were determined on the eld using a Garmin GPS. All the samples, collected in tightly capped high-quality polyethylene bottles, were immediately transported under low temperature conditions in ice box and stored at 4 C until analysis. Water from most of the springs sampled was pumped to the surface using a peristaltic pump equipped with a tubing. The end of the tubing was placed as near as possible to the spring vent. Some spring samples were collected without the use of the pump when sampling containers could easily be placed directly in the spring outow. Water was collected from open wells using drawing buckets tied with ropes, while boreholes were pumped for 515 min before sampling. These waters were poured into 500-ml-capacity plastic bottles after rinsing with the sample and preserved airtight to avoid evaporation. Physico-chemical parameters like pH, electrical conductivity (EC), water temperature, oxygen reduction potential (ORP) and dissolved oxygen (DO) of sampled springs were measured in situ using pH/EC/temperature TOA-DKK meter, TOA-DKK ORP meter and a TOA-DKK DO meter, respectively.

Major ion analyses were done at the Isotope Hydrology Laboratory of Kumamoto University, Japan, for major ions, while dissolved silica (SiO2) was analyzed at the

National Agriculture Center for Kyushu-Okinawa Region, Kumamoto. The major ions that were analyzed included Na?, K?, Ca2?, Mg2?, Cl-, HCO3-, SO42-and NO3-.

Na?, K?, Ca2?, Mg2?, Cl-, SO42- and NO3- concentrations were determined by ion chromatography and dissolved silica using an ICP Dionex 3000 Spectrometer. Bicarbonate (HCO3-) ion concentrations were determined by acid titration of 0.02 N H2SO4. Samples with EC [200 lS/cm were subjected to 20 times dilution using distilled water before analyzing for major ions and dissolved SiO2. No dilution was done for samples with EC \200 lS/cm. Both the diluted and undiluted samples were ltered through 0.2 lm lters, prior to major ions and dissolved silica determination. The analytical precision of cations (Na?, K?, Ca2?, Mg2?) and anions (Cl-, HCO3-,

SO42-and NO3-) was veried using ionic balance error (IBE) on the basis of ions expressed in meq/l (Appelo and

Postma 1999). IBE was observed to be within a limit of 5 % for all the investigated samples.

123

Appl Water Sci (2014) 4:99113 103

For quality control purposes, duplicate samples were taken in the eld for 10 % of the total samples. These were treated as completely separate samples, receiving their own sample number. The duplicates were thus not recognizable in the laboratory. Also, for quality control of the chemical measurements, standards (used for not more than 3 days after preparation) and blanks were used in between runs to provide a measure of background noise, accuracy and precision. The equipment and instruments were tested and calibrated with calibration blanks and a series of calibration standards as per specications outlined in standard methods of water and wastewater (APHA 1998). Total dissolved salts (TDS) were calculated by summing the main ionic species (Na?, K?, Ca2?, Mg2?, Cl-, HCO3-, SO42- and

NO3-) and dissolved silica. PHREEQC program (Parkhurst and Appelo 1999) was used to calculate saturation indices, partial pressure of CO2 (PCO2) and the activities of aqueous species in the groundwaters.

After all sample analyses were completed, descriptive statistics, correlation matrix of variables and multivariate statistical analysis [principal component analysis (PCA) and factor analysis (FA) were done using the computer package, Statistical Package for Social Sciences (SPSS) (SPSS 19.0) software on 16 hydrochemical variables (pH, EC, temperature, ORP, DO, SiO2, Na?, K?, Ca2?, Mg2?,

Cl-, HCO3-, SO42-, NO3-, TDS, PCO2].

Principal component analysis (PCA) and factor analysis (FA)

Among multivariate techniques, R-mode FA has been widely employed for understanding hydrogeochemical associations and processes controlling them (Drever 1988; Razack and Dazy 1990; Melloul and Collin 1992; Brizkishore and Murali 1992; Ballukraya and Ravi 1999). As a rst step, correlation analysis which reveals the relationship between two variables is calculated for the major ion chemical data from the study area. Since correlation analysis reveals similarities or differences in the behavior of pairs of ions, and does not conveniently identify groups of ions that behave similarly, FA is carried out for the chemical data from the study area to help in hydrogeo-chemical interpretation. For factor analysis, rst principal components are calculated, which give the eigenvalue and the percentage of the variance explained by each factor.

Principal component analysis gives communality of unity for each component. Only those factors having eigenvalues greater than unity (Kaiser 1958) are considered for nal analysis. Though FA reduces the dimensionality of the problem, the meaning of these factors may sometimes be difcult to deduce (Davis 1986). The interpretation can be simplied using certain rotational procedures. For the present study, Kaisers varimax rotation was applied to

obtain a simple structure. A set has been taken to rotate the factors (varimax rotated) in such a way that all their components are closer to ?1, 0 and -1, representing the contributions of corresponding variables to the total variance as a positive contribution, a no contribution and a negative contribution, respectively. The factor loadings depict the inuence of a factor on a variable and vice versa. Thus, Factor 1 gives the largest eigenvalue and explains the greatest amount of variance in the data set. Factor 2 represents low eigenvalue and explains the greatest of the remaining variance and so forth.

The nal step of the FA is to project the data on the rotated signicant factors. The scores obtained by this projection are called factor scores, which are used to understand the nature of variables. Dalton and Upchurch (1978) have stated that the factor scores are related to the intensity of the chemical process described by each factor. Negative numbers reect areas unaffected by the process, positive numbers indicate areas most affected and near-zero numbers affect to an average degree (Lawrence and Upchurch 1983).

Results and discussion

Descriptive statistical analysis and major water types

Table 1 shows the descriptive statistics (minimum, maximum and average values) of chemical constituents (pH, EC, temperature, Eh, TDS, DO, SiO2, PCO2, nitrate and major ion concentrations) of groundwater samples (67 in the Banana Plain and 33 in the Mount Cameroon area). Groundwaters in the Banana Plain are moderately acidic to mildly alkaline with pH values varying between 4.30 and7.85, while in the Mount Cameroon area groundwaters are neutral to mildly alkaline (pH 6.977.98). TDS, DO, EC, Eh and SiO2 vary in a similar manner in both study sites, but they are generally enhanced in the Mount Cameroon area.

Average cation concentrations occur in the order Ca [ Na [ Mg [ K in the Banana Plain, while in the

Mount Cameroon area the order is Ca [ Na [ K [ Mg. For the anions, the order is HCO3 [ NO3 [ SO4 [ Cl in the Banana Plain and HCO3 [ SO4 [ Cl [ NO3 in the

Mount Cameroon area.

Sulfate and chloride are major constituents of natural water and are generally good tracers of the sources of salinity.

Natural sources of sulfate and chloride are seawater and the dissolution of evaporitic rocks, and the main anthropogenic sources are urban, industrial and agricultural wastewater. Different origins of these contaminants have often been distinguished on the basis of their bromide-to-

123

104 Appl Water Sci (2014) 4:99113

Table 1 Descriptive statistics of groundwater physico-chemical and chemical parameters in the study areas

Constituents Area No. samples Min. Max. Average Standard deviation

pH Banana Plain 67 4.30 7.85 5.79 0.87

Mount Cameroon 33 6.97 7.98 7.57 0.29

EC (lS/cm) Banana Plain 67 33.10 530.00 170.33 123.74 Mount Cameroon 33 70 500 209.36 109.49

Temp (C) Banana Plain 67 25.00 30.40 27.60 1.31 Mount Cameroon 33 16.20 27.90 22.60 2.33

DO (mg/l) Banana Plain 67 2.70 48.99 14.06 11.90

Mount Cameroon 33 5.50 27.50 13.62 5.59

Eh (mV) Banana Plain 67 -58.00 280.00 190.08 65.38

Mount Cameroon 33 82 246 189.59 32.81

SiO2 (mg/l) Banana Plain 67 5.11 27.73 12.26 7.00 Mount Cameroon 33 7.60 18.12 13.29 2.96

Ca2? (mg/l) Banana Plain 67 1.53 46.31 11.40 8.60

Mount Cameroon 33 3.87 20.64 12.18 4.42

Mg2? (mg/l) Banana Plain 67 0.57 18.79 5.53 3.45 Mount Cameroon 33 2.55 15.25 8.33 3.43

Na? (mg/l) Banana Plain 67 1.03 43.01 6.81 6.33

Mount Cameroon 33 5.69 64.90 14.07 11.30

K? (mg/l) Banana Plain 67 0.88 26.46 5.47 4.32

Mount Cameroon 33 2.60 16.75 6.50 3.05

HCO3- (mg/l) Banana Plain 67 7.76 377.76 75.56 80.47 Mount Cameroon 33 44.15 203.60 122.51 41.64

SO42- (mg/l) Banana Plain 67 0.23 171.59 8.00 26.74 Mount Cameroon 33 0.27 325.39 15.31 57.12

Cl- (mg/l) Banana Plain 67 0.53 65.08 5.65 8.09

Mount Cameroon 33 0.52 104.42 6.07 18.16

NO3- (mg/l) Banana Plain 67 0.05 146.61 17.28 21.54 Mount Cameroon 33 0.29 12.37 2.90 2.69

TDS (mg/l) Banana Plain 67 43.18 493.48 147.63 108.05

Mount Cameroon 33 75.99 503.06 201.16 80.60

chloride and bromide-to-sulfate ratios. No health-based GV has been proposed by the WHO for sulfate and chloride in drinking water (WHO 2004). However, concentrations in excess of about 200 mg/l for both anions can give rise to detectable taste in water. At higher concentration, SO42-may cause gastrointestinal irritation, particularly when

Mg2? and Na? are also present in drinking water resources (Suthar et al. 2009). In this study, sulfate (as SO42-) concentration ranged from 0.23 to 171. 56 mg/l (average8.00 mg/l) in the Banana Plain and 0.27325.39 mg/l (average 15.31 mg/l) in the Mount Cameroon area (Table 1). Thirty percent of Mount Cameroon groundwater resources are not suitable for drinking according to the WHO drinking water sulfate concentration limit of 200 mg/l (WHO 2004). Chloride concentrations ranged from 0.53 to 65.08 mg/l (average 5.65 mg/l) in the Banana Plain and 0.52104.42 mg/l (average 6.07 mg/l) in the Mount Cameroon area (Table 1).

The Chadha diagram (Chadha 1999) was used to dene the major water types in the areas. It is a modied version of the Piper (1944) diagram and plots the difference in milliequivalent percentage between alkaline earths (Ca ? Mg) and alkali metals (Na ? K) expressed as percentage and plotted on the x-axis. The difference in milli-equivalent percentage between weak acidic anions (CO3 ? HCO3) and strong acidic anions (Cl ? SO4) is plotted on the y-axis. The rectangular eld resulting from this is similar to the diamond-shaped eld in the Piper diagram and describes the overall character of the water. Thus, there are eight elds in the rectangular plot which represent eight different water types as in the case of Piper diagram. Those eight water types are: (1) alkaline earths exceed alkali metals; (2) alkali metals exceed alkaline earths; (3) weak acidic anions exceed strong acidic anions;(4) strong acidic anions exceed weak acidic anions; (5) alkaline earths and weak acidic anions exceed both alkali

123

Appl Water Sci (2014) 4:99113 105

metals and strong acidic anions, respectively (such water has temporary hardness); (6) alkaline earths exceed alkali metals and strong acidic anions exceed weak acidic anions (such water has permanent hardness); (7) alkali metals exceed alkaline earths and strong acidic anions exceed weak acidic anions; (8) alkali metals exceed alkaline earths and weak acidic anions exceed strong acidic anions (Chadha 1999).

Figure 2a, b shows the Chadha diagrams for ground-water samples from the study areas. This diagram shows that 89.6 % of the groundwater samples from the Banana Plain (Fig. 2a) fall in the subeld 5-(CaMg-HCO3), 9 %

samples in subeld 6-(CaMg-Cl) and 1.4 % in the subelds 7-(NaK-Cl) and 8-(NaK-HCO3). For Mount

Cameroon groundwaters (Fig. 2b), 87.5 % fall in the subeld 5-(CaMg-HCO3), 6.3 % in subeld 8-(NaK-HCO3)

and 3.1 % in the subelds 6-(CaMg-Cl) and 7-(NaK-Cl). The dominance of CaMg-HCO3 water type in both study areas may represent groundwaters that either contain

waters at the early stages of geochemical evolution (recent recharge) or rapidly circulating groundwaters which have not undergone signicant waterrock interactions (Kebede et al. 2005; Marghade et al. 2011). Water groups represented by CaMg-HCO3 and NaK-HCO3 are weakly mineralized waters circulating within the basaltic and scoriaceous aquifers of the Banana Plain and the Mount Cameroon area. CaMg-Cl and NaK-Cl water types are symptomatic of anthropogenic pollution. In groundwaters, the source of Cl, SO4, NO3 and Na ions are mostly agricultural fertilizers, animal wastes and industrial and municipal sewage. This indicates the inuence of human activities on water chemistry (Jalali 2009).

Nitrate contamination of groundwaters

The average groundwater nitrate concentration for the study areas is 17.28 mg/l in the Banana Plain and 2.90 mg/ l in the Mount Cameroon area. All 100 % of the wells, springs and boreholes had detectable nitrate (Table 2) and about 6 % (in Banana Plain) and 0 % (in Mount Cameroon) of the groundwater resources were not suitable for drinking without treatment, according to the WHO and Cameroon Water Quality Norms drinking water nitrate concentration limit of 50 mg/l (WHO 2004; ANOR 2003). A high nitrate concentration is always related to anthropogenic contamination, the main sources being fertilizers and human or animal bodily waste (Spalding and Exner 1993). Nitrate can be the main anionic species in both agricultural and urban areas. In the Banana Plain, it is the second most abundant anionic species, while in the Mount Cameroon area, it occupies the last position. Thus, it is suggested that the impact of anthropogenic activities, especially groundwater nitrate contamination, is more accentuated in the Banana Plain than in the Mount Cameroon area.

A large majority of nitrate sources are of anthropogenic origin and are commonly conned to the same watershed and groundwater basin. In aquifer systems more prone to nitrate contamination, surcial aquifers are more likely to be contaminated and at a greater extent compared to deeper, conned or unconned aquifers. Inverse relationships between nitrate concentration and aquifer depth have been noticed in previous studies (Spalding and Exner 1993; Ray and Schock 1996; Ako et al. 2011). The contaminants coming from these various sources can be transported to surface waters by surface runoff and groundwater discharge. Several studies indicate that rural land uses, especially agricultural practices, can cause nitrate contamination of underlying groundwater. In addition, regions experiencing signicant population growth may be substantially contributing to the groundwater nitrate contamination (Ako et al. 2011). In some instances, it has been

a

(HCO3)-(SO4+Cl)meq%

120.0

100.0

80.0

8

8

8

8 60.0

5

6

40.0

20.0

0.0

-20.0 0.0 20.0 40.0 60.0 80.0 100.0

-20.0

7 6

7 7 6

7 -40.0

6

-60.0

(Ca+Mg)-(Na+K)meq%

Borehole Spring Well

b

(HCO3)-(SO4+Cl)meq%

120.0

100.0

8

80.0

60.0

40.0

5

20.0

0.0

-60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0

-20.0

7

6

-40.0

-60.0

(Ca+Mg)-(Na+K)meq%

Borehole Spring

Fig. 2 a Chadha diagram for Banana Plain groundwaters. b Chadha diagram for Mount Cameroon area groundwaters

123

106 Appl Water Sci (2014) 4:99113

Table 2 Nitrate concentration and percentage distribution in groundwater of the study areas

Area No. samples Number of samples with NO3 concentration of: Percentage of samples with NO3 concentration of:

010 mg/l 1045 mg/l [45 mg/l 010 mg/l 1045 mg/l [45 mg/l

Banana Plain 67 24 39 4 36.0 58.00 6

Mount Cameroon 33 32 1 0 96.97 3.03 0

observed that elevated nitrate concentrations are accompanied by increased chloride concentrations. In places where fertilizers are applied to the land surface, there is an increased potential for groundwater contamination due to the direct downward inltration of nitrate to the aquifer system. These sources, however, do not necessarily coincide with increased concentrations of chloride in ground-water. Alternative sources such as sewer breakthrough and animal waste can cause a signicant increase of both nitrate and chloride concentrations (Hudak and Blanchard 1997). Ako et al. 2011 observed a very strong positive correlation

between NO3- and Cl- (R2 = 0.83) in the Banana Plain and postulated that this was a diagnostic indicator of anthropogenic activity on groundwater quality in the area.

Principal component analysis (PCA) and factor analysis (FA)

The rst step in multivariate statistical analysis is to establish relationships between variables for the physico-chemical and chemical data of the study area. Table 3a, b presents the matrix of Spearmans correlation coefcient

Table 3 Matrix of correlation between 12 variables in (a) Banana Plain and (b) Mount Cameroon area

pH EC Eh Ca2? Mg2? Na? K? Cl- NO3- SO42- HCO3- TDS

(a) Banana Plain

pHEC 0.4

Eh -0.02 -0.41

Ca2? 0.2 0.49 -0.64

Mg2? 0.16 0.13 -0.25 0.68

Na? -0.03 0.31 -0.42 0.68 0.6

K? 0.04 0.00 -0.09 0.5 0.5 0.29

Cl- -0.19 -0.18 -0.03 0.34 0.51 0.48 0.22

NO3- -0.27 -0.28 0.03 0.2 0.45 0.38 0.14 0.91SO42- 0.1 0.23 -0.22 0.11 0.03 -0.01 -0.04 -0.06 -0.09HCO3- 0.4 0.76 -0.58 0.76 0.35 0.56 0.25 -0.07 -0.21 0.23TDS 0.31 0.64 -0.6 0.86 0.56 0.7 0.33 0.27 0.14 0.4 0.9

log (PCO2) -0.43 0.08 -0.25 0.19 0.26 0.18 0.16 0.00 0.01 0.05 0.13 0.15(b) Mount Cameroon area

pH

EC -0.02

Eh 0.17 0.04Ca2? 0.30 -0.01 0.04

Mg2? 0.30 0.42 -0.02 0.79

Na? -0.05 0.70 0.03 0.09 0.42

K? -0.04 0.36 0.18 0.35 0.53 0.68

Cl- -0.13 0.53 -0.07 -0.05 0.14 0.86 0.39

NO3- -0.03 -0.12 0.05 0.28 0.13 -0.17 -0.16 -0.06SO42- -0.15 0.09 0.23 -0.22 -0.17 -0.06 -0.17 0.02 0.32HCO3- 0.10 0.63 0.04 0.33 0.57 0.11 0.08 -0.12 -0.08 0.08TDS -0.05 0.64 0.19 0.14 0.39 0.39 0.18 0.30 0.19 0.72 0.62

log (PCO2) 0.31 -0.30 0.16 0.04 0.00 -0.20 -0.02 -0.19 -0.20 -0.19 -0.17 -0.29

Correlation coefcients [0.5 are marked in bold font

123

Appl Water Sci (2014) 4:99113 107

for physico-chemical and chemical data of the study areas; most of the physico-chemical and chemical parameters were not normally distributed. EC and TDS showed a strong positive correlation with the major cations (calcium, magnesium and sodium and the bicarbonate anion) in both study areas. Very strong correlation exists between chlo-ride and nitrate in the Banana Plain (R2 = 0.91). In the Mount Cameroon area, very strong positive correlation exists between calcium and magnesium and also between sodium and chloride. In both areas, the main contributor to ionic content is bicarbonate deriving from the titration of magmatic carbon dioxide within the aquifers and to metals (mainly alkaline and alkaline earth elements) released by waterrock interaction processes (Aiuppa et al. 2003). Demlie et al. (2007) illustrated that a positive correlation of NO3- and Cl- was a diagnostic indicator of anthropogenic activity. The very strong correlation between chloride and nitrate in the Banana Plain is a diagnostic indicator of anthropogenic activity on groundwater quality. The high salinity values in Mount Cameroon area are probably due to the presence of seawater or from dissolution of evaporate rocks in the active volcanic area of Mt. Cameroon. According to Vengosh and Rosenthal (1994), the Na/Cl ratio during mixing between groundwater and seawater is0.86 and for fresh groundwater Na/Cl [ 1, while Na/ Cl = 1 indicates dissolution of halite minerals. Na/Cl ratios of Mount Cameroon groundwaters vary between 0.1 and 64.1 (median value 0.4) with only 4 % of samples having Na/Cl ratios less than that of seawater (0.86). Ninety-six percent of Mount Cameroon groundwaters have high Na/Cl ratio ([1) which is typical of anthropogenic sources like domestic wastewaters and distinguishable from the low ratio of seawater intrusion (El Moujabber et al. 2006). Thus, the coastal aquifers of the Mount Cameroon area are not much subject to saltwater intrusion. In the Banana Plain, Na/Cl ratios vary between 0.7 and24.4 (median value 2.1) with 85 % of samples having Na/ Cl ratios greater than that of seawater (0.86) and 81 % having Na/Cl ratios [1. Equally, the Banana Plain aquifer is not much subjected to seawater intrusion, but more of the inuence of anthropogenic activities. Saturation indices (SI) for groundwater in the Mount Cameroon area have been calculated with respect to carbonate, halite and sulfate minerals (Ako et al. 2012). SI values suggest that all the groundwaters are undersaturated with respect to carbonate, halite and sulfate minerals. SI also suggests that the dissolution of carbonate, sulfate and halite minerals is insignicant in the aquifer of the Banana Plain (Ako et al. 2011). Thus, evaporate mineral phases are minor or absent in the host rocks despite the presence of SO42- and Clions in the groundwaters. Local contamination by septic tanks, sewage systems and agricultural fertilizers may contribute sulfates and chlorides to the groundwaters of the

Banana Plain and Mount Cameroon area (Ako et al. 2011, 2012).

Ako et al. (2012) also established an inverse relationship between Cl- concentrations in spring water and altitude in the Mount Cameroon area. Benedetti et al. (2003) also reported the same ndings, and this reects the input of the oceanic monsoon chloride-rich rainwater at low altitudes. At higher altitudes, the decrease in Cl- concentrations is due to smaller rainfall inputs caused by the rapid decrease in the absolute humidity and temperature with elevation resulting in a smaller condensation. However, chloride ions can also result from a variety of human activities (domestic sewage, fertilizers, etc.). Enrichment of TDS, Na and Cl is also possible, because of the effect of urban wastewaters (Subba Rao and Krishna Rao 1990; Somasundaram et al. 1993). This strong positive correlation between Na? and Cl- is typically observed in wet environments like the Banana Plain and Mount Cameroon area where precipitation inputs dominate evapotranspiration (Appelo and Postma 1999).

Since correlation analysis reveals similarities or differences in the behavior of pairs of ions and does not conveniently identify groups of ions that behave similarly, PCA/FA is carried out on the data set for better hydrogeochemical interpretation. For PCA/FA, rst principal components are calculated which give the eigenvalues and the percentage of the variance explained by each factor. The principal components were extracted in the decreasing order of importance, so that the rst PC accounts for as much of the variation as possible and each successive component, a little less. As the rst PC accounts for the covariation shared by all the attributes, this may be a better estimate than the simple or weighted averages of the original variables (Mahloch 1974). Eigenvalues and eigenvectors were thus calculated for the covariance matrix. Then, the data were transformed into factors. Table 4 presents the eigenvalues, percent of variance, cumulative eigenvalue and cumulative percent of variance associated with each factor or component. These values are summed to express as a cumulative eigenvalue and percentage of variance, respectively. The number of factors extracted is to be determined. This study retains only factors with eigenvalues that exceed one. This criterion was proposed by Kaiser (1958) and is probably the one most widely used (Matalas and Reiher 1967; Davis 1973; Reyment and Joreskog 1993; Miller and Miller 2000). In the Banana Plain from the 17 variables, 5 explain 77.48 % of the total variance. For the Mount Cameroon data set, from the 15 variables 6 explain 80.78 % of the total variance.

Though FA reduces the dimensionality of the problem, the meaning of these factors may sometimes be difcult to deduce. The interpretation can be simplied using certain rotational procedures. The ve components, which

123

108 Appl Water Sci (2014) 4:99113

Table 4 Eigenvalues, percent of variance, cumulative eigenvalue, cumulative percent of variance for the factor analysis of hydrochemical data in the Banana Plain and the Mount Cameroon area

Component Initial eigenvalues Total variance explained

Extraction sums of squared loadings

Rotation sums of squared loadings

Total % of variance

Cumulative %

Total % of variance

Cumulative %

Total % of variance

Cumulative %

(a) Banana Plain

1 6.121 36.006 36.006 6.121 36.006 36.006 5.557 32.686 32.686

2 2.978 17.519 53.525 2.978 17.519 53.525 2.787 16.394 49.08

3 1.698 9.987 63.512 1.698 9.987 63.512 1.971 11.595 60.675

4 1.332 7.833 71.345 1.332 7.833 71.345 1.483 8.722 69.397

5 1.043 6.134 77.478 1.043 6.134 77.478 1.374 8.082 77.478

6 0.971 5.714 83.193

7 0.775 4.559 87.751

8 0.642 3.778 91.530

9 0.387 2.278 93.807

10 0.358 2.104 95.911

11 0.263 1.549 97.461

12 0.147 0.835 98.325

13 0.142 0.835 99.160 14 0.081 0.479 99.639

15 0.058 0.344 99.982

16 0.002 0.012 99.995

17 0.001 0.005 100.000

(b) Mount Cameroon area

1 3.949 24.679 24.679 3.949 24.679 24.679 2.936 18.348 18.348

2 2.743 17.142 41.821 2.743 17.142 41.821 2.896 18.803 36.451

3 2.149 13.431 55.252 2.149 13.431 55.252 2.608 16.297 52.749

4 1.587 9.92 65.172 1.587 9.92 65.172 1.64 10.297 63.002

5 1.358 8.485 73.657 1.358 8.485 73.657 1.577 9.854 72.856

6 1.14 7.124 80.782 1.14 7.124 80.782 1.268 7.926 80.782

7 0.75 4.688 85.47

8 0.637 3.98 89.45

9 0.632 3.953 93.403

10 0.44 2.75 96.153 11 0.291 1.819 97.972

12 0.201 1.255 99.227

13 0.058 0.362 99.589

14 0.048 0.299 99.889

15 0.018 0.111 100.000

explained 77.47 % of the total variance of the Banana Plain data set and the six components which explained 80.78 % of the total variance in the Mount Cameroon area, were rotated by the varimax method to make interpretation in terms of the original variables easier by adjusting the loadings so that they were either near 1 or near 0 (Davis 1973). Table 5 lists the PCA/FA results after varimax rotation.

The terms strong, moderate and weak as applied to factor loadings refer to absolute loading values of [0.75,

0.750.5 and 0.50.3, respectively (Liu et al. 2003). For the Banana Plain, Factor 1 which explains 36.00 % of the total variance has strong positive loadings on TDS, HCO3, Ca and TH, moderate loading on EC, Eh (negative), Na, Mg and SiO2 and weak loading on SO4. This PC can be ascribed to natural hydrogeochemical evolution of groundwater by groundwatergeological medium interactions and hence a geogenic factor. Factor 2 accounts for17.52 % of the variance of the data set. It has strong positive loadings on NO3 and Cl and weak loadings on Mg

123

Appl Water Sci (2014) 4:99113 109

Table 5 Results of the factor analysis after varimax rotation

Parameter Rotated component matrix

Component

1 2 3 4 5

(a) Banana Plain

TDS 0.960 0.18

HCO3 0.925 -0.186Ca 0.836 0.257 0.111 0.308

TH 0.792 0.384 -0.192 0.198 0.308

EC 0.742 -0.305 0.199 -0.104

Eh -0.696 0.112 Na 0.645 0.474 0.102 0.145

Mg 0.516 0.493 -0.424 0.262 0.323

NO3 0.955 Cl 0.949

Temp 0.774 0.21

Ph 0.3 -0.294 0.682 0.228

PCO2 0.132 -0.131 -0.617 0.15 0.125 DO 0.21 0.898

SiO2 0.548 -0.439 0.555 -0.134 K 0.306 0.125 -0.17 0.807

SO4 0.408 -0.292 -0.574

Parameter Rotated Component Matrix

Component

1 2 3 4 5 6

(b) Mount Cameroon area

Na 0.96 0.089 0.158 -0.075 -0.009 0.059

Cl 0.882 -0.147 0.048 0.052 -0.014 0.074 K 0.705 0.414 -0.021 -0.059 0.15 -0.208

Ca 0.052 0.889 -0.056 0.283 -0.068 0.09

Mg 0.325 0.868 0.207 0.062 -0.062 0.036

SiO2 -0.216 0.684 0.185 -0.452 0.098 -0.131

TDS 0.248 0.121 0.836 0.189 0.129 0.107

HCO3 -0.047 0.457 0.75 -0.248 -0.227 0.084 SO4 -0.129 -0.321 0.651 0.407 0.346 0.062 EC 0.603 0.108 0.638 -0.224 -0.087 0.09

NO3 -0.166 0.181 0.069 0.854 0.074 0.038

ORP -0.024 0.083 0.198 -0.047 0.759 0.027

DO 0.175 -0.236 -0.119 0.23 0.666 -0.046

PCO2 -0.201 0.167 0.187 -0.387 0.483 0.041 Temp 0.125 -0.122 0.271 0.184 -0.109 0.848

pH -0.158 0.457 -0.146 -0.261 0.261 0.659

Factor loadings [0.4 are marked in bold font and blanks represent very insignicant loadings

and Na. This PC indicates anthropogenic contamination from domestic and agricultural sources and thus can be termed the anthropogenic factor (Esteller and Andreu

2005). Factor 3 which accounts for 9.99 % of the total variance has strong positive loading on temperature only, moderate loadings on pH and PCO2 (negative) and weak loading on SiO2 (negative). PC3 seems to represent recharge water because of positive loadings on temperature and pH which could be acquired during inltration of rainwater through soil zones. The loadings on PCO2 and

SiO2 represent the interaction between CO2-charged recharge water and basaltic rocks. Factor 4 has strong positive loading on DO and weak loading on SiO2 and accounts for 7.83 % of the total variance of the data set.

Factor 5 accounts for 6.13 % of the total variance and has a strong positive loading on K with weak (negative) loading on SiO2. Major cations including Ca, Na and Mg have positive correlations with alkalinity (HCO3) and EC (Table 3). This can be explained by acidic hydrolysis of mac minerals in basaltic rocks. This hydrolysis reaction consumes water and acid that might have originated from CO2 and increased the pH, alkalinity and cation concentration of groundwater.

In the Mount Cameroon area, Factor 1, which explains24.68 % of the total variance, has strong positive loadings on Na and Cl and moderate loadings on K and EC. This association strongly suggests that PC1 may represent an ionic enrichment factor: gentle slope and sluggish drainage conditions supporting longer residence time of groundwater, more waterrock interaction and higher solubility of minerals together with saline water leading to enrichment of Na and Cl (UNESCO 1984; Subba Rao et al. 1997). Enrichment of EC, Na, Cl and K is also possible, because of the effect of urban wastewaters (Subba Rao and Rao 1990; Somasundaram et al. 1993) and high rate of evapotranspiration (Drever 1988; Karanth 1991). Factor 2 has strong positive loadings on Ca and Mg, moderate loading on SiO2 and weak loadings on HCO3 and pH. It accounts for 17.14 % of the total variance of the data set. PC2 represents the natural process by which water acquires its chemical characteristics through rockwater interactions, and hence a geogenic factor. Factor 3 which accounts for 13.43 % of the total variance has strong positive loadings on TDS and HCO3 and moderate loadings on SO4 and EC. PC3 also seems to represent a geogenic factor. Factor 4 accounts for 9.92 % with strong positive loading on NO3 only and weak loadings on SO4 and SiO2 (negative). PC4 can be considered an anthropogenic factor.

Factor 5 has strong loading on ORP, moderate loadings on DO, weak loading on PCO2 and accounts for 8.49 % of the total variance. Factor 6 accounts for 7.12 % of variance of the data set having strong loading on temperature and moderate loading on pH. PC5 and PC6 can be attributed to the physico-chemical properties of recharge water.

Figure 3a, b illustrates a plot of component scores for PC1 against PC2 in the two study areas. As stated above, in the Banana Plain, PC1 is a geogenic factor, while PC2 is an

123

110 Appl Water Sci (2014) 4:99113

anthropogenic factor. It is obvious from Fig. 3a that in the Banana Plain, samples lean more closely toward PC1 than PC2, suggesting that groundwaters from its aquifers are predominantly under the effects of anthropogenic activities than weathering. Groundwater samples from the Mount Cameroon area (Fig. 3b) are high on the ionic enrichment side and have very signicant contributions from the two components, PC1 and PC2.

The result of the principal component analysis/FA suggests that although two areas are part of the CVL, there are differences in their water chemistry that are very apparent. The major differences in the composition of groundwater from the two sampled areas are in the concentrations of nitrate, bicarbonate and sulfate. It can be concluded that three factors are responsible for the groundwater chemistry of the selected areas. The rst factor represents the natural hydrogeochemical processes by which the groundwaters acquire their chemical characteristics by interaction with aquifer material and hence a geogenic factor. Anthropogenic factor due to anthropogenic contamination from domestic and agricultural sources represents the second factor. Ionic enrichment due to the input of oceanic

monsoon chloride-rich rainwater and the enrichment of Na and Cl from urban wastewaters is the third factor. In the Banana Plain, the geogenic and anthropogenic factors are the most preponderant, while in the Mount Cameroon area, the ionic enrichment and geogenic factors are the most important.

Kross et al. (1993) indicated that the depth of a well was an important determinant of nitrate contamination. In the Banana Plain, a correlation (R = 0.44) was observed between well and borehole depths and nitrate concentrations (Fig. 4). It can be seen that the amount of nitrates are signicantly higher in the rst 20 m of the aquifers and decreases with depth. Such a distribution clearly indicates that nitrate species in most of the considered groundwaters are of supercial origin (anthropogenic). The signicant concentrations of NO3- in groundwater are indicative of the unconned nature of the system. Shallow groundwater resources (open wells and springs) are more vulnerable to nitrate contamination due to natural N-xing under aerobic conditions, while denitrication in reducing conditions is the probable cause of low nitrate concentrations in waters from deep boreholes of the Banana Plain.

The following point and non-point sources could have contributed nitrates to the groundwater of the Banana Plain and the Mount Cameroon area(1) Fertilizer use: the fertilizer consumption rate is relatively higher in these areas due to well-established agro-industrial farming system of bananas. Commonly used fertilizers in the banana plantations are NPK, urea and potash which are applied on a monthly basis, whereas lime amendments are applied on a 2-year basis and rotation with other crops (pineapple) is sometimes used (Opfergelt 2008). (2) Freely drained, well-aerated, deep, fertile loam soils: the soil of these areas is typically ash and pumice deposits from volcanic cones and sandy or sandy clay with high coarse texture. Such soils have high water ltration rate and possibly contribute in

a

PC2(anthropogenic factor)

2.5

2

1.5

1

0.5

0

-1 -0.5 0 0.5 1 1.5

-0.5

PC1(geogenic factor)

b

2.5

NO3 (mg/l)

PC2(geogenic factor)

2

0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0

0

1.5

-20

-40

1

Depth (m)

-60

0.5

Borehole Well

-80

0

-100

-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2

-120

-0.5

PC1(salinity factor)

-140

Fig. 3 a Plot of the loadings of PC1 against PC2 in the Banana Plain. b Plot of the loadings of PC1 against PC2 in the Mount Cameroon area

-160

Fig. 4 Plot of depth of wells and boreholes versus nitrate concentrations of groundwaters in the Banana Plain

123

Appl Water Sci (2014) 4:99113 111

nitrate leaching to underground waters. (3) Irrigation mechanisms: the two areas are rich in agro-industrial production of banana and other export crops (rubber and palm oil) based on a well-developed system of sprinkler irrigation. High irrigation rate may contribute to nitrate leaching in soils of these areas. Due to intensive agro-industrial activities, nitrogenous fertilizer consumption rate is comparatively higher in these areas than in other agro-zones of the country (Jayne et al. 2003). These nitrogenous fertilizers are easily converted to NO3- when applied to the soil.

Nitrates are highly soluble and easily leachable on irrigation (Kundu et al. 2008). It is therefore concluded that heavy fertilizer consumption, highly coarse soils, high irrigation rates and well-developed sprinkler irrigation network are some factors responsible for NO3- leaching to groundwater of the studied areas alongside the inltration of wastewater from sanitation facilities into groundwater.

Apart from non-point sources, some other point sources also contribute to groundwater NO3-pollution. Wakida and

Lerner (2005) reviewed the non-agriculture sources of groundwater nitrate and suggested that waste disposal network, animal wastes including livestock and human excreta, industry, riveraquifer interactions, house building, etc. are some important factors indirectly enriching nitrate in groundwater. It has been observed that open dumping (in heaps) is the common practice to dispose garbage in the two areas. Due to weathering and open dumping, a considerable amount of soluble forms of nitrogen could be leached into deep soil layers especially during the wet months of the year. Other sources such as human excreta may contribute to nitrate leaching in the areas. Pit latrines may enhance NO3- in groundwater, but further detailed study on groundwater microbiology is needed to support the hypothesis (Suthar et al. 2009).

Continuous urbanization in the selected areas is expected to increase concentrations of contaminants including nitrates in groundwater. The consequence of urban expansion and population growth is that water tables are lowered and contamination of shallow groundwater occurs through indiscriminate disposal of domestic and industrial wastes. In the Banana Plain, a series of cholera epidemics have been recorded in the last few years (222 cases in 2004; 1,112 cases in 2005; 255 cases in 2006) (GTZ 2006). Folifac et al. (2009) revealed that in the Buea municipality of the Mount Cameroon area, anthropogenic activities around the six major drinking water sources are potential threats and pathways for contamination and that source water protection has not been given adequate attention in the planning and development of Buea. Most water sources around the Buea area are not potable with respect to total coliform counts with values ranging from 3 CFU/100 ml to [4,800 CFU/100 ml during the dry and rainy seasons, respectively (Azah 2009). Poor sanitation and hygiene

conditions in the studied areas lead not only to nitrate contamination of groundwaters, but also to outbreak of diseases. High prevalence of waterborne and water-related diseases is a direct manifestation of this situation.

Conclusions

Principal component analysis (PCA)/factor analysis (FA) is an effective means of manipulating, interpreting and representing data concerning groundwater pollutants. PCA/FA was applied to groundwater samples for the assessment of nitrate contamination of groundwater in two agro-industrial areas of the CVL. The results showed that three factors describe the main processes affecting the groundwater chemistry of the studied areas: (1) Factor 1 (geogenic) marked by high loadings on TDS, HCO3, Ca, Mg and TH.

This factor reects the natural (geogenic) hydrogeo-chemical processes, connected mainly with interaction between CO2-charged recharge water and basaltic rocks of the aquifers; (2) Factor 2 (anthropogenic) marked by high loadings on NO3 and Cl. The concentrations of these elements have mostly an anthropogenic origin and are associated with the inuence of pollution from land use activities in the study area; (3) Factor 3 (ionic enrichment) marked by high loadings on Na and Cl. The input of oceanic monsoon chloride-rich rainwater and/or the effects of urban wastewaters are responsible for this factor. The discovery of the human activity inuence on groundwater chemistry in the densely populated and agro-industrialized areas of the CVL is the main contribution of this research. Based on the results obtained it can be concluded that the impact of anthropogenic activities, especially groundwater nitrate contamination, is more accentuated in the Banana Plain than in the Mount Cameroon area This study clearly demonstrates that PCA/FA is an effective and reliable method for differentiating natural and anthropogenic processes affecting groundwater quality.

Acknowledgments This write up constitutes part of data generated during the PhD study of the corresponding author at Kumamoto University, Japan, who was supported by Monbukagakusho Scholarship from MEXT (The Japanese Ministry of Education, Science, Sports and Culture). Material support was also provided by Grant-in-aid for scientic research No. 21606 from SASAGAWA Foundation, Japan (Japan Science Society). We thank the Ministry of Scientic Research and Innovation (MINRESI) of Cameroon and the Institute for Geological and Mining Research (IRGM) Yaound-Cameroon for putting the corresponding author on study leave during his PhD studies. Thanks are due to the anonymous referees of the journal, who helped immensely in rewriting the paper in terms of content and language.

Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.

123

112 Appl Water Sci (2014) 4:99113

References

Aiuppa A, Bellomo S, Brusca L, DAlessandro W, Federico WC

(2003) Natural and anthropogenic factors affecting groundwater quality of an active volcano (Mt. Etna, Italy). Appl Geochem 18:863882Aka FT, Nagao K, Kusakabe M, Sumino H, Tanyileke G, Ateba B,

Hell J (2004) Symmetrical helium isotope distribution on the Cameroon Volcanic Line, West Africa. Chem Geo 203(34): 205223Ako AA, Jun S, Takahiro H, Kimpei I, Nkeng GE, Fantong WY,

Eyong GET, Ntankouo NR (2011) Evaluation of groundwater quality and its suitability for drinking, domestic, and agricultural uses in the Banana Plain (Mbanga, Njombe, Penja) of the Cameroon Volcanic. Environ Geochem Health 33:559575 Ako AA, Jun S, Takahiro H, Makoto K, Akoachere RA, George EN,

Gloria ETE, Alain LFT (2012) Spring water quality and usability in the Mount Cameroon area revealed by hydrogeochemistry. Environ Geochem Health 34:615639ANOR (2001) Cameroon mineral water quality norms (NC 05:

2001-02) Agence Nationale des Normes et Qualit, Yaound ANOR (2003) Cameroon drinking water quality norms (NC 207:

2003-02) Agence Nationale des Normes et Qualit, Yaound APHA (1998) Standard methods for the examination of water and wastewater, 20th edn. American Public Health Association, Washington DCAppelo CAJ, Postma D (1999) Chemical analysis of groundwater, geochemistry, groundwater and pollution. Balkema, Rotterdam Azah LA (2009) A physico-chemical and microbiological analysis of some drinking water sources in Buea. MSc Thesis, Department of Geology and Environmental Science. University of Buea, CameroonBallukraya PN, Ravi R (1999) Characterization of groundwater in the unconned aquifers of Chennai City, India. J Geol Soc India 54:1322Benedetti MF, Dia A, Riotte J, Chabaux F, Gerald M, Boulegue J et al (2003) Chemical weathering of basaltic lava ows undergoing extreme climatic conditions: the water geochemistry record. Chem Geol 201:117Bhlke JK (2002) Groundwater recharge and agricultural contamination. Hydrogeol J 10:153179Briz-kishore BH, Murali G (1992) Factor analysis for revealing hydrogeochemical characteristics of a watershed. Environ Geol 19:39Chadha DK (1999) A proposed new diagram for geochemical classication of natural waters and interpretation of chemical data. Hydrogeo J 7:431439Dalton MG, Upchurch SB (1978) Interpretations of hydrogeochemical facies by factor analysis. Groundwater 10:228233Davis JC (1973) Statistics and data analysis in geology. John Wiley and Sons, New YorkDavis JC (1986) Statistics and data analysis in geology. John Wiley and Sons, New YorkDemlie M, Wohnlich S, Wisotzky F, Gizaw B (2007) Groundwater recharge, ow and hydrogeochemical evolution in a complex volcanic aquifer system, central Ethiopia. J Hydrogeo 15:11691181Druelle B, Nni J, Kambou R (1987) Mount Cameroon: an active volcano of the Cameroon line. J Afr Earth Sci 6:197214 Djoret D (2000) Etude de la recharge de la nappe du Chari Barguimi

(Tchad) par les methods chimiques et isotopiques. Thse Doctorat, Universite dAvignon et les pays de Vaucluse Drever JI (1988) The geochemistry of natural waters. Prentice-Hill

Inc., New York

El Moujabber M, Bou Samra B, Darwish T, Atallah T (2006)

Comparison of different indicators for groundwater contamination by seawater intrusion on the Lebanese coast. Water Resour Manage 20:161180Endeley RE, Ayonghe SN, Tchuenteu F (2001) A preliminary hydrogeochemical baseline study of water sources around the Mount Cameroon. J Cameroon Acad Sc 1(3)

Esteller MV, Andreu JM (2005) Anthropic effects on hydrochemical characteristics of the Valle de Toluca aquifer (Central Mexico). Hydrogeo J 13(2):378390Fitton JG, Dunlop HM (1985) The Cameroon Line, West-Africa, and its bearing on the origin of oceanic and continental alkali basalt. Earth Planetary Sc Lett 72(1):2338Folifac F, Lifongo L, Nkeng G, Gaskin S (2009) Municipal drinking water source protection in low income countries: case of Buea municipalityCameroon. J Ecolo Nat Env 1(4):073084 Fraser PJ, Hall JB, Healey JR (1998) Climate of the Mount Cameroon region: long and medium term rainfall, temperature and sunshine data. University Wales, Bangor; Mount Cameroon Project and Cameroon Development Corporation. School of Agricultural and Forest Sciences Publication Number 16GTZ (German Technical Coorperation) (2006) Projet pour l Amelioration de lacces a leau potable et de lassainissement de base dans les trios communes rurales de Manjo, Loum et Penja/ Njombe, Province de Littoral-Cameroun. German Technical Coorperation, Douala-CameroonHudak PF, Blanchard S (1997) Land use and groundwater quality in the Trinity Group outcrop of north-central Texas, USA. Environ Int 23(4):507517Jalali M (2009) Geochemistry characterisation of groundwater in an agricultural area of Razan, Hamadan, Iran. Environ Geol 56:14791488Jayne TS, Valerie K, Eric C (2003) Fertilizer consumption trends in

Sub-Saharan Africa. http://www.msu.edu/agecon/fs2/psynindx.htm

Web End =http://www.msu.edu/agecon/fs2/psynindx. http://www.msu.edu/agecon/fs2/psynindx.htm

Web End =htm . Accessed 25 August 2010Kaiser HF (1958) The varimax criteria for analytical rotation in factor analysis. Psychometrika 23(3):187200Karanth KR (1991) Impact of human activities on hydrogeological environment. J Geol Soc India 38:195206Kebede S, Travi Y, Alemayehu T, Ayenew T (2005) Groundwater recharge, circulation and geochemical evolution in the source region of the Blue Nile River, Ethiopia. Appl Geochem 20(9):16581676Kross BC, Hallberg GR, Bruner DR, Sherry Holmes K, Johnson JK

(1993) The nitrate contamination of private well water in Iowa. American J Public Health 83(2):270272Kundu MC, Mandal B, Sarkar D (2008) Assessment of the potential hazardous of nitrate contamination in surface and groundwater in a heavily fertilized and intensively cultivated district of India. Environ Monit Assess 146:183189Kyoung-Ho K, Seong-Taek Y, Byoung-Young C, Gi-Tak C, Yong-sung J, Kangjoo K, Hyoung-Soo K (2009) Hydrochemical and multivariate statistical interpretations of spatial controls of nitrate concentrations in a shallow alluvial aquifer around oxbow lakes (Osong area, central Korea). J Contam Hydrol 107:114127Lawrence FW, Upchurch SB (1983) Identication of recharge areas using geochemical factor analysis. Groundwater 20:680687 Liu CW, Lin KH, Kuo YM (2003) Application of factor analysis in the assessment of groundwater quality in a blackfoot disease area in Taiwan. Sci Tot Environ. doi:http://dx.doi.org/10.1016/S0048-9697(02)00683-6

Web End =10.1016/S0048-9697(02) http://dx.doi.org/10.1016/S0048-9697(02)00683-6

Web End =00683-6 Mahloch JL (1974) Graphical interpretation of water quality data.

Water Air Soil Pollut 3:217236

123

Appl Water Sci (2014) 4:99113 113

Marghade D, Malpe DB, Zade AB (2011) Major ion chemistry of shallow groundwater of a fast growing city of Central India. Environ Monit Assess. doi:http://dx.doi.org/10.1007/s10661-011-2126-3

Web End =10.1007/s10661-011-2126-3

Marzoli A, Renne PR, Piccirillo EM, Francesca C, Bellieni G, Mel

AJ, Nyobe JB, Nni JN (1999) Silicic magmas from the continental Cameroon Volcanic Line (Oku, Bambouto and Ngaoundere): Ar-40-Ar-39 dates, petrology, SrNd O isotopes and their petrogenetic signicance. Contribs Mineral Petrol 135(23):133150Matalas CN, Reiher JB (1967) Some comments on the use of factor analysis. Water Resour Res 3(1):213223Mbanga Rural Council (2008) Monographie de la Commune de

Mbanga. Mbanga Rural CouncilMelloul A, Collin M (1992) The principal component statistical method as a complementary approach to geochemical methods in water quality factor identication: application to coastal plain aquifer of Israel. J Hydrol 140:4973Miller NJ, Miller JC (2000) Statistics and chemometrics for analytical chemistry. Pearson Education, Englewood CliffMoore KB, Ekwurzel B, Esser BK (2006) Sources of groundwater nitrate revealed using residence time and isotope methods. Appl Geochem 21:10161029Murgulet D, Tick GR (2008) Assessing the extent and sources of nitrate contamination in the aquifer system of southern Baldwin County. Alabama. Environ Geol. doi:http://dx.doi.org/10.1007/s00254-008-1585-5

Web End =10.1007/s00254-008-1585-5 National Metrological Service (2008) Metrological data for Mbanga.

Njombe and Penja, National Metrological ServiceNgounou Ngatcha B, Djoret D (2010) Nitrate pollution in ground-water in two selected areas from Cameroon and Chad in the Lake Chad basin. Water Policy 12(5):722733Njitchoua R, Ngounou Ngatcha B (1997) Hydrogeochemistry and environmental isotope investigations of the North Diamare Plain, northern Cameroon. J Afr Earth Sci 25(2):307316Opfergelt S (2008) Silicon cycle in the soilplant system: biogeo-chemical tracing using Si isotopes. Dissertation. Catholic University of Louvain, BelgiumOyebog SA, Ako AA, Nkeng GE, Suh EC (2012) Hydrogeochemical characteristics of some Cameroon bottled waters, investigated by multivariate statistical analyses. J Geochem Explor 112:118130

Parkhurst DL, Appelo CAJ (1999) PHREEQC for windows version1.4.07. A hydrogeochemical transport model. US Geological Survey softwarePiper AM (1944) A graphic procedure in geochemical interpretation of water analysis. Trans-Am Geophys Union 25(6):914928 Ray C, Schock SC (1996) Comparability of large-scale studies of agricultural chemical contamination of rural private wells. Ground Water Monit Remed 16(2):92102Razack M, Dazy J (1990) Hydrogeochemical characterization of groundwater mixing in sedimentary and metamorphic reservoirs with combined use of Pipers principle and factor analysis. J Hydrol 114:371393Reyment RA, Joreskog KH (1993) Applied factor analysis in the natural sciences. Cambridge University Press, New York Somasundaram MV, Ravindran G, Tellam JH (1993) Groundwater pollution of the Madras urban aquifer, India. Groundwater 31:411Spalding RF, Exner ME (1993) Occurrence of nitrate in groundwatera review. J Environ Qual 22(3):392402Subba Rao NG, Rao Krishna (1990) Intensity of pollution of groundwater in Visakhapatnam area, Andhra Pradesh. J Geol Soc India 36:670673Subba Rao N, Prakasa Rao J, Chandra Rao P, Niranjan Babu PG, Rao

Krishna (1997) Hydrogeochemical zoning in crystalline terrain and its signicance to water quality. J Geol Soc India 49:715719Suthar S, Preeti B, Sushma S, Pravin KM, Arvind KN, Nagraj SP

(2009) Nitrate contamination in groundwater of some rural areas of Rajasthan, India. J Hazard Mater. doi:http://dx.doi.org/10.1016/j.jhazmat.2009.05.111

Web End =10.1016/j.jhazmat.2009. http://dx.doi.org/10.1016/j.jhazmat.2009.05.111

Web End =05.111 UNESCO 1984. Groundwater in hard rock, studies and reports in hydrology, 32, Paris, pp 228Vengosh A, Rosenthal E (1994) Saline groundwater in Israel: its bearing on the water crisis in the country. J Hydrol 156:389430 Wakida FD, Lerner DN (2005) Non-agriculture sources of ground-water nitrogen: a review and case study. Water Res 39:316 WHO (2004) Guidelines for drinking water quality (vol 2). Health criteria and other supporting information, WHO

123