Grain Size and Source Apportionment of Heavy Metals in Urban Stream Sediments

K. Sekabira1, H. Oryem-Origa2, T. A. Basamba3, G. Mutumba4 and E. Kakudidi5 1Department of Environment, School of Engineering and Applied Sciences Kampala International University, Kampala 2Department of Biology, School of Biological Sciences, Makerere University, Kampala 3Soil Science Department of Agricultural Production College of Agricultural and Environmental Sciences, Makerere University, Kampala Uganda


Introduction
Heavy metals are ubiquitous environmental pollutants and concern over possible health risks and ecosystem effects in sediments have increased in recent years. In aquatic sediments, heavy metals are mostly enriched in the fine grained fractions (Salomon and Förstner 1984;Muwanga, 1997;Prego et al., 1999). Studies have therefore used clay (< 2 µm) (Förstner 1987) and < 63 µm (Muwanga 1997) fractions in assessment of heavy metal concentrations in sediments. The specific surface area of sediments is dependant on granulometric parameter and mineral composition (Juracic et al., 1982). Association of metals with smaller grain-size particles is attributed to co-precipitation and complexation of metals on particle surfaces and this determines the distribution pattern of heavy metals in sediments (Ho et al., 2010). Most of the metal content occurs in complex form with insoluble inorganic and organic ligands. Heavy metal emissions from anthropogenic activities occur in stream or river sediments, where they are absorbed onto clays and other fine grained materials (Ho et al., 2010) Heavy metals can be absorbed on negatively charged surfaces of clay minerals, organic matter or iron and manganese oxides and hydroxides. Sediments are considered to be important carriers as well as sinks for heavy metals in the hydrological cycle (Muwanga, 1997). The objectives of this study were: (1) to determine heavy metal content in urban stream sediment fractions and, (2) to assess source apportionment of heavy metals in stream sediments. This study was conducted between the months of August 2008 to November 2009, along the Nakivubo Channelized stream, Kampala Uganda. Victoria , with a total area of 190 km 2 (Fig. 1). The study area has a tropical climate that is attributed to high altitude, relief, proximity to Lake Victoria and long distance from the sea (Matagi et al., 1998). The Lake Victoria basin has warm temperatures ranging between 23°C to 32°C and a bi-modal rainfall pattern averaging approximately 1260 mm annually. The area is underlain by granites and granitoid gneisses of the Precambrian. These are overlain by phyllites and schist of the Buganda-Toro system with a mixture of alluvial and lacastrine sand, silt and clay that characterise Nakivubo swamp soils. Soils have been derived from weathering of the rocks. The alluvial soils from the upper (profile A) layers are composed of semi-liquid organic material and those underneath consists of reddish ferruginous loams and clays (profile B) attributed to organic decomposition and runoff (Kansiime and Nalubega, 1999). In this study, the stream was subdivided into three sections ( Fig. 1 .1), using a hand trowel. Watindo stream was chosen to be outside the study area, 30 km along Gulu road for comparison purposes. The sam p l e s w e r e p l a c e d i n Z i p l o c b a g s a n d transported to the laboratory. The hand trowel [plastic] was washed with a detergent, rinsed and dried before each use so as to minimize contamination. Sediment samples were air dried and 300.0 g were transferred into a set of standard sieves (0.063 mm, 0.125 mm, 0.25 mm, 0.5 mm, 1.0 mm and 2.0 mm) with the largest mesh size on top and the smallest at the bottom. The sieving was carried out using a vibrating machine. The weight of soil retained www.intechopen.com on each sieve was determined and a cumulative percentage calculated. Coefficient of curvature and uniformity were assessed to determine particle-size distribution fractions of each sample. The total content of Pb, Cd, Cu, Zn, Mn and Mn in sediment fractions was determined using the method as described by Sekabira et al. (2010). Ideally, 1.25 g of each sample was digested with 20 mL aqua regia (HCl/HNO 3 3:1) in a beaker (open-beaker digestion) on a thermostatically controlled hot plate. The digest was heated to near dryness and cooled to ambient temperature. Then 5.0 mL of hydrogen peroxide was added in parts to complete the digestion and the resulting mixture heated again to near dryness in a fume cupboard. The beaker wall was washed with 10 ml of de-ionised water and 5.0 ml HCl were added, mixed and heated again. The resulting digest was allowed to cool and transferred into a 50 mL standard flask and made up to the mark with de-ionised water. Pb, Cd, Cu, Zn, Mn and Fe were then analyzed by direct aspiration of the sample solution into a Perkin-Elmer model 2380 Flame Atomic Absorption Spectrophotometer (AAS). All metals were analysed using lean-blue acetylene flame at wavelength 324.8 nm, slit width 0.2 mm and sensitivity check of 5.0 mg/L Cu; wavelength 228.8 nm, slit width 0.7 and sensitivity check of 2.0 mg/L Cd; wavelength 213.9 nm, slit size 0.7 nm and sensitivity check 9.0 mg/L Pb and wavelength 279.5 nm, slit size 0.2 nm and sensitivity check of 2.5 mg/L Mn. Sediment pH was measured in a suspension of 1:2.5, sediment to water ratio using a calibrated pH meter (WE-30200). Accuracy of the analytical method was evaluated by comparing the expected metal concentrations in certified reference materials with the measured values. Simultaneous performance of analytical blanks, standard reference (JG-3) (Imai et al., 1995) and calculation of the average recoveries of heavy metals confirmed that the accuracy of the method was within acceptable limits (

Assessment of heavy metal distribution in sediment
Enrichment Factor (EF): As proposed by Simex and Helz (1981), EF was employed to assess the degree of contamination and to understand the distribution of heavy metal elements of anthropogenic origin from sites by individual elements in sediments. Iron (Fe) was chosen as the normalizing element while determining EF-values, since in wetlands it is mainly supplied from sediments and is one of the widely used reference element ( Loska et al., 2002;Kothai et al., 2009: Chakravarty andPatgiri, 2009;Seshan et al., 2010). Other widely used reference metal elements include Al and Mn (Nyangababo et al., 2005a;Kamaruzzaman et al., 2008;Ong and Kamaruzzaman, 2009).

Enrichment Factor (EF) = (C n /Fe) sample/ (C n /Fe) background
www.intechopen.com where C n is the concentration of element "n". The background value is that of average shale (Turekian and Wedepohl, 1961). Elements which are naturally derived have an EF value of nearly unity, while elements of anthropogenic origin have EF values of several orders of magnitude. Six categories are recognised: ≤ 1 background concentration, 1 -2 depletion to minimal enrichment, 2 -5 moderate enrichment, 5 -20 significant enrichment, 20 -40 very high enrichment and > 40 extremely high enrichment (Sutherland, 2000). Analysis of variance (ANOVA): ANOVA was employed to determine whether groups of variables have the same means on data that are continuous or normally distributed and with homogeneous variance. Additionally, it was employed to assess the relationship between heavy metal concentrations and their interaction between sections of the stream. Correlation analysis: Pearson's correlation analysis was adopted to analyse and establish inter-metal relationship and physico-chemical characteristics of the stream water. Cluster Analysis (CA) and Factor analysis (FA): CA was performed to classify elements of different sources on the basis of their similarities using dendrograms and to identify relatively homogeneous groups of variables with similar properties. FA was employed on the variables that are correlated to isolate or determine specific factors that are associated with such groupings of metal concentrations so as to establish their origin and distribution. The data was standardised to give a normal distribution with a mean of 0 and a variance of 1. Sample means were standardised by subtracting the mean of their distribution and dividing by standard error (SE) or square root of the variance.

Heavy metal concentrations in various fractions
The mean pH ranged between 5.71±1 (slightly acidic) and 7.25±1 (neutral), but at Watindo stream, the mean pH of the sediments ranged from highly acidic (4.53±1) to acidic (5.40±1) ( Table 3). Total heavy metal concentrations in < 63 µm and 63-125 µm sediments as well as distribution pattern along the Nakivubo Channelized stream are indicated in Table 3, Fig. 2 and Fig. 4

Sediment enrichment
The results show that enrichment factor values can be assessed with respect to the average shale in reference to the degree of contamination (Harikumar et al., 2009;Ong and Kamaruzzaman, 2009 Fig. 6b. A linear correlation between Fe-Cd and Fe-Cu elemental pairs suggest that Cd and Cu were naturally occurring heavy metal concentrations (terrigenous) and the outliers would suggest anthropogenic sources. Cadmium and Copper were strongly adsorbed to manganese oxides and hydroxides in the clay-silt fractions. However, Pb and Zn were poorly correlated with Mn suggesting anthropogenic influence. Results for Pb-Fe, Cd-Fe, Cu-Fe and Zn-Fe scatter plots are shown in Fig.7a. A linear correlation between Fe-Cu and Fe-Mn elemental pairs suggest that Cu and Mn were naturally occurring heavy metal concentrations (terrigenous) and the outliers above the threshold would suggest an anthropogenic source. However, Pb, Cd and Zn were poorly correlated with Mn suggesting anthropogenic influence in fine-sand fractions. Results for Pb-Mn, Cd-Mn, Cu-Mn and Zn-Mn scatter plots are shown in Fig.7b. A linear correlation between Fe and Cu elemental pairs suggest that Cu was naturally occurring heavy metal concentrations (terrigenous) and the outliers above the threshold would be regarded as anthropogenic. Copper showed strong adsorption to manganese oxides and hydroxides in the clay-silt fractions. However, Pb, Cd and Zn were poorly correlated with Mn suggesting anthropogenic influence in fine-sand fractions. Heavy metal concentration data in the sediment fractions were subjected to ANOVA and showed no significant variation in the means of clay-silt and fine sand for Pb, Cd and Cu, except for Zn along the Nakivubo Channelized stream (Kruopiene, 2007), Cd and Cu along the tributaries. However, the mean values were higher in clay-silt fractions in all the samples. Elemental concentrations were within the same order of magnitude as observed by Sekabira et al., (2010) along the Nakivubo drainage system except for Cu and Zn which showed extreme high elemental concentrations in sediment fractions of clay-silt and fine sand. Enrichment Factor values of heavy metals in clay-silt fractions (< 63 µm) for Pb, Cd, Cu, Zn and Mn increased gradually downstream, whereas the EF values for fine sand fractions showed an irregular decrease. This phenomenon of increasing heavy metal concentration downstream may be attributed to the increased pollution downstream and adsorption of heavy metals from the water by fine grained sediments with large surface area and clay with negative surface charge. Irregular distribution of heavy metals in fine sand fractions may indicate a localized source of the pollutants, sink and/or retention phenomena (Zanganeh et al., 2008).

Source apportionment of pollutants
Inter-metal and sediment property association was also evaluated by using Pearson correlation coefficient (r) and the results are presented in Table 5.  -c), Zn-c, Pb-s, Zn-s and BOD were significantly associated with sediment pH, suggesting its influence as a controlling factor. Elemental associations were assessed using Pearson correlation coefficient (r) and indicated that each paired elements had an identical source, geochemistry, and/or common sink (Nyangababo et al., 2005b;Sekabira et al., 2010). Cadmium elemental association with grain size fraction contents may signify its influence as a controlling factor. In aquatic sediments, heavy metals are mostly enriched in and are associated with the fine grained fractions (Muwanga, 1997;Prego et al., 1999;El-Moselhy and Abd El-Azim, 2005). Association of copper in sediments with Mn and Fe-oxides/hydroxides may suggest specific adsorption and co-precipitation by isomorphic substitution. **Correlation is significant at the 0.01 level (2-tailed); *. Correlation is significant at the 0.05 level (2tailed); BOD-biological oxygen demand s-silt-clay fractions; s-Fine-sand fractions  Cluster Analysis (CA) was performed on the data using Ward or Average linkage and correlation coefficient distance. Results for CA are shown in Fig. 8. Four clusters of elemental associations were identified based on the fusion of the clusters that are similar. The dendrogram explains the influence and association of the heavy metal clusters or groups by their relative elemental concentrations at each site. CA showed the association of pH and BOD with Pb and Zn in clay-silt and fine-sand fractions as well as Cu in clay-silt fraction in the first group (I). This may suggest the association of Pb with organic matter and pH as a controlling factor. The second group (II) showed the association of Cd and Cu in fine-sand fractions as well as Cd in clay-silt fractions in the Nakivubo stream sediments. Elements in group III (Mn and Fe) originate from terrigenous sources (Sekabira et al., 2010) in both clay-silt and fine-sand fractions. Group IV contains percentage fractions of clay-silt and fine-sand. A biplot of sites and elemental concentrations associated Agakhan High School Bridge and Lugogo stream with Cu and Cd (Fig. 9). This may be attributed to car washing bay, petrol stations and vehicular emissions.  Table 6. Varimax rotated factor loadings and communalities of the Nakivubo stream, tributaries, industrial outfall sediment fractions and Watindo stream (n=15) Factor Analysis was carried out to establish the influence of sediment grain size, pH and BOD on heavy metal concentrations in clay-silt and fine-sand fractions (Table 6). Four factors with eigenvalues > 1 were extracted in the analysis to help explain the data. The first four factors account for 80.1 % of the total variance/inertia in the data set. The rotated factor matrix is explained by four factors with high communalities of elements except pH. The first factor accounts for 24.9 % of the total variance and contains Pb, Cu and Zn, as well as BOD in water and pH with high variable loading on this factor and corresponds to group I of the cluster analysis. This suggests pH as the controlling factor (Muwanga, 1997;Prego et al., 1999;Abílio et al., 2006 andHo et al., 2010) and the influence of organic matter on Pb. At neutral pH, clays possess negative surface charges that attract Pb, Cu nd Zn cations into bottom sediments. The second factor accounts for 22.0 % of the total inertia and contains Mn and Fe as well as Cu in fine sand fractions with high variable loadings and corresponds to group III of cluster analysis. This association may be due to their common occurrence in the basic rocks [terrigenous], since their concentrations were within background values (EF ≤ 1) (Sekabira et al., 2010). The Third factor accounts for 19.0 % of the total inertia and contains Cd, Cu and Zn in fine-sand fractions as well as clay-silt fractions and BOD. The association of Cd and Zn may be attributed to their similar geochemistry and may indicate a source of mixed origin and/or sink of vehicular emissions, Katwe metal works and cement stores at Good Shade. The association of heavy metals and BOD may suggest the role of organic matter (OM) in heavy metal sequestration and the dual origin of Cu and Zn. This causes the transfer of heavy metals into bottom sediments. The Fourth factor accounts for 14.3 % of the total inertia and contains percentage clay-silt and fine-sand fractions with high variable loadings and corresponds to group IV of the cluster analysis.

Conclusions
1. Clay-silt fraction sorting increased for zinc concentrations in the Nakivubo stream sediments and generally for lead. Heavy metal concentration increased downstream with percentage increase in clay-silt fractions and enrichment in the very fine grained fractions, probably due to their negative surface charge and higher particulate surface area. 2. The distribution patterns of the heavy metals are controlled by the sorting of finegrained fractions, pH and organic matter as indicated by BOD.

This study showed that stream sediments have background concentrations for Fe and
Mn at most of the sites. 4. Factor analysis also indicated three sources of pollutants; (1) mixed origin or retention phenomena of Pb, Cu and Zn as well as BOD of industrial and municipal waste effluents [NWSC]; (2) industrial and vehicular emissions of Cd, Cu and Zn and terrigenous fraction sources characterised by Cu, Mn and Fe. Water pollution is a major global problem that requires ongoing evaluation and revision of water resource policy at all levels (from international down to individual aquifers and wells). It has been suggested that it is the leading worldwide cause of deaths and diseases, and that it accounts for the deaths of more than 14,000 people daily. In addition to the acute problems of water pollution in developing countries, industrialized countries continue to struggle with pollution problems as well. Water is typically referred to as polluted when it is impaired by anthropogenic contaminants and either does not support a human use, such as drinking water, and/or undergoes a marked shift in its ability to support its constituent biotic communities, such as fish. Natural phenomena such as volcanoes, algae blooms, storms, and earthquakes also cause major changes in water quality and the ecological status of water. Most water pollutants are eventually carried by rivers into the oceans.