LITERATURE REVIEW
2. Introduction
2.4. Chemical composition
2.4.1. Elemental analyses
differences in the method of sampling, separation of grain size fractions, frequency and timing (season) of sampling and identification and estimation of the respective minerals (Datta and Subramanian, 1997).
than K2O, tending to produce initial enrichment in potassium in residues of weathering (the source material for all sediments) (Cited by Hutcheon et al., 2000).
Subramanian et al. (1985) observed lower values of Al, Si and Ca, and higher values of Fe, Mn and Ni in the Brahmaputra sediments than those in the Ganges sediments. They ascribed this to the Brahmpautra sediments having mean grain size coarser than that of the Ganges sediments. They cited Subramanian (1979) and Milliman and Meade (1983) as reporting that world surface rocks (WSR) differed from the composition of sediments in Indian rivers in case of mobile and reactive elements due to dominant chemical weathering suffered by these sediments. Concentrations of Ca, Mg, Ba and Sr were lower for the Indian river sediments than those of WSR due to high carbonate equilibria in the waters of Indian rivers (Subramanian, 1983—cited by Subramanian et al., 1985).
Biksham and Subramanian (1988) observed a general increase of Mg, Al, Si, Cu, Zn, decrease of K and Ca, but no variations of Fe and Ti in the elemental compositions of Godavari River sediments. They opined that the large variations in the values of the elements emphasized the complexity of data generation on river sediments.
Ramesh et al. (1989) observed enrichment of minor elements (V, Cr, Co, Ni, Cu and Zn) in suspended sediments compared to bed sediments. They stated that similar observations were reported for several other major rivers such as the Ganges and the Brahmaputra (Subramanian et al., 1987) in India, and the Mississippi in the USA (Tefry and Presley, 1976). They observed good correlations between major elements (Al-Si = - 0.67; Al-Mg = 0.87; Na-K = 0.99) and between transition elements (Fe-V = 0.96; Fe-Ni =
0.93; Fe-Zn = 0.99; V-Zn = 0.95; Co-Ni = 0.94). They pointed to common sinks for elements showing good correlations.
Martin and Meybeck (1979) observed lower concentrations of Ca and Na, and higher concentrations of Al in suspended sediments of 15 major tropical rivers. Ramesh (1985) observed increase in Al corresponding to increase in clay minerals (Cited by Ramesh et al., 1989).
Martinelli et al. (1993) found 68.5% SiO2, 15.2% Al2O3, 5.6% Fe2O3, 3.33%
Na2O, 2.54% CaO, 2.40% K2O, 1.46% MgO and 1.00% TiO2 in the Varzea sediment samples collected along the Amazon River floodplain. They observed losses of major elements such as Al, Fe, Ca, Mg and Na and enrichment of SiO2 in the sediment as it traveled downriver and was deposited and eroded from the floodplain of the Amazon River. They attributed this observation to chemical and physical weathering of the sediment in its downward transport.
According to Manjunatha et al. (1996), chemical weathering predominates over physical weathering resulting in the enrichment of Al and Fe and depletion of Ca, Na, Mg, K, Rb, Sr and Ba in weathered products with respect to the average composition of surficial rocks.
Dupre et al. (1996) observed strong depletion of U, Rb, K, Ba, Sr, Na and Ca in the suspended sediments (Congo Basin) relative to the mean upper crust values (Taylor and McLennan, 1985) and an enrichment of the same elements in the dissolved phase.
Viers et al. (2009) stated that Mg, Na and K were more strongly depleted in tropical environments than in temperate or boreal zones.
Canfield (1997) reported a significant positive correlation of Al, Fe and Mn
concentrations (suspended sediments) versus runoff. According to Viers et al. (2009), higher proportion of soluble rock components leaches from the rock at higher runoff, concentrating Al, Fe and Mn in the particulate phase. Pokrovsky et al. (2005) reported that presence of ferromagnesian silicates or Fe-rich secondary phases in the drainage basin resulted in high Fe concentrations in suspended sediments. Viers et al. (2009) stated that high Mn concentrations might be encountered in both polluted and pristine rivers as Mn-enrichment might result from natural processes also [e.g. formation of authigenic particles in the fluvial environments (Ponter et al., 1992; Andersson et al., 1998)].
Singh and France-Lanord (2002) studied the chemical and isotopic compositions of sediments from the Brahmaputra River and its tributaries to trace sediment provenance and to understand erosion pattern in the basin. They concluded that erosion of the Himalayan rocks represented ca 70% of the detrital flux and the Siang-Tsangpo River were the major source of sediment to the whole Brahmaputra. According to them, erosion of the Namche Barwa represented about 45% of the total flux at its outflow before confluence with the Ganga. They also observed that sediments of the Brahmaputra, Ganga or other western Himalayan river contained lower Na and K values compared to average continental crust, which, according to them, showed that the sources were recycled crustal formations.
Mahanta and Subramanian (2004) reported the abundance and relative mobility of analysed elements in Brahmaputra bed sediments as Si> Al> Fe> Ca> Mg> Na> K> P>
Mn> Zn> Cu> Cr> Pb> Cd and suspended sediments as Si> Al> Fe> Na> K> Ca> Mg>
P> Mn> Zn> Cu> Cr> Pb> Cd. They observed enrichment of Fe (1.3×), K, P, Mn (1.4×), Zn (1.2×), Cu (1.6×), Cr (1.6×), Pb (1.8×), and Cd (1.3×) in suspended sediments over
bed sediments in bulk phases. They found more Si in bed sediments than the world average and attributed its cause to the extensive physical weathering experienced by the basin comprising granite, gneiss and sandstone terrains under the influence of wet humid climate in the downstream region and transport of generated silicate minerals by the tributaries. Amongst the heavy metals, they found some correlation among Cu, Pb, Zn and Mn in the suspended sediments, but not in the bed sediments. According to them, metals in the sediments owe their origin more to natural than to anthropogenic sources.
They stated that enormous amount of sediment load might have diluted the inputs from domestic, agricultural and industrial sources, which was also reflected by samples collected downstream of Guwahati, the biggest urban settlement along the river, showing no significant rise in elemental concentrations.
Viers et al. (2009) presented a new database on the chemical composition of suspended matter in World Rivers, together with the associated elemental fluxes. A part of the database is presented in Table 2.8.
Table 2.8: Average concentration of major and trace elements in the suspended sediments of world rivers (Viers et al., 2009) [σ = standard deviation; n = no. of samples]
Elements Unit Concentration σ n upper crust Martin and Meybeck Savenko
Si % 25.4 11.7 68 30.8 28.5 25.6
Al % 8.72 6.01 140 8.04 9.4 8.63
Fe % 5.81 4.81 144 3.5 4.8 5.03
Ca % 2.59 2.8 118 3 2.15 2.6
K % 1.69 1.04 119 2.8 2 2.15
Mg % 1.26 1.4 93 1.33 1.18 1.44
Na % 0.71 0.93 108 2.89 0.71 0.82
Mn µg/g 1679 5011 148 600 1050 1150
Zn µg/g 208 237 134 71 250 130
Cr µg/g 130 155 149 35 100 85
Cu µg/g 75.9 135 143 25 100 45
Ni µg/g 74.5 100 164 20 90 50
Pb µg/g 61.1 91.2 132 20 100 25
Cd µg/g 1.55 4.15 70 0.098 - 0.5
The table also contains composition of the continental crust proposed by Taylor and McLennan (1985) as well as the chemical composition of riverine suspended sediments given by Martin and Meybeck (1979) and Savenko (2006), as reported by Viers et al. (2009).