Geochemical Fingerprinting - Application of Geochemical Tracers to Fluvial Sediment - page 22

in sediment source to a river through time (Fig. 2.4 ) (Foster et al. 1998 ; Owens
et al. 1999 ; Walling et al. 2003a , b ; Miller et al. 2005 , 2013 ; Pittam et al. 2009 ;
Collins et al. 2010c ). Essentially, it is assumed that the sampled deposits represent
an historical record of sediment transport within the basin, where the age of the
sediment varies as a function of depth below the ground surface. Thus, fingerprinting
can be carried out on samples collected at differing depths to reconstruct the changes
in sediment provenance to the depositional site through time. The method is useful
in that it allows an understanding of the contemporary sediment sources to be placed
into an historical framework. It also illustrates that fingerprinting can be used to
retrospectively determine the primary sources of sediment to the channel, something
that cannot be done using monitoring data.
2.3.3 Identifying Effective Geochemical Fingerprints
Studies of sediment provenance in the 1980s and 1990s often relied on a single para-
meter, many of which were based on the physical characteristics of the sediment,
such as it grain size distribution, mineralogy, or magnetic properties (Table 2.1 ). Later
investigations, beginning in the late 1990s, showed that erroneous sediment-source
area associations were common when only a single fingerprinting parameter was
utilized (Collins and Walling 2002 ). Thus, there was a move to use multiple parame-
ters to fingerprint source area sediments (Collins et al. 1997a , b ; Miller et al. 2005 ;
Mukundan et al. 2012 ; Collins et al. 2010a , 2013 ; Miller et al. 2013 ). This compos-
ite fingerprinting approach was aided by (1) advances in analytical chemistry that
greatly expanded the number and rate for which samples that could be analyzed for a
large number of constituents (Walling et al. 2013 ), and (2) the increased use of mul-
tivariate statistical methods to manipulate the composite fingerprinting data, thereby
allowing for the quantification of the results. Both factors also increased the use of
geochemical parameters as fingerprints, particularly the elemental concentrations of
trace metals (Lewin and Wolfenden 1978 ; Macklin 1985 ; Knox 1987 , 1989 ; Pass-
more andMacklin 1994 ; Miller et al. 2005 , 2013 ), rare earth elements (Morton 1991 ;
Miller et al. 2013 ), organic substances (Hasholt 1988 ), fallout radionuclides (Peart
and Walling 1986 ; Walling and Woodward 1992 ; Wallbrink and Murray 1993 ), and
various radiogenic or stable isotopes (Douglas et al. 1995 ). In many cases, the uti-
lized geochemical constituents are natural, but in others, investigators have made use
of anthropogenic pollutants, such as heavy metals, pesticides and fertilizers (Bravo-
Espinoza et al. 2009 ; Takeda et al. 2004 ). Takeda et al. ( 2004 ), for example, found
that while phosphate fertilizers contained 10-200 times more U than soils, they con-
tained lower Th concentrations than the soils. Thus, the U/Th ratio proved to be an
effective fingerprinting tool (Evrard et al. 2013 ).
In general, the type of tracer used for a given study depends on how the sediment
sources are defined. For example, if the intent is to determine the relative contributions
of sediment on the basis of source type (e.g., sheet, rill, gully, and bank erosion), then it
will be important to consider constituents that are elevated in surfacematerials eroded