Geochemical Fingerprinting - Application of Geochemical Tracers to Fluvial Sediment - page 17

portion of the suspended load, is largely responsible for decreasing water quality, and
is chemically reactive, thereby serving as an important conveyor of hydrophobic con-
taminants. Sand-sized sediment, however, may also be of importance. Within many
of the gravel-bed rivers of the Southern Appalachian Mountains of the southeastern
U.S., for instance, aquatic habitats are predominantly affected by the deposition of
sand-sized sediments on the channel bed, and their infiltration into the interstitial
spaces between gravel sized clasts.
Another fundamental problem inherent in inverse mixing models is the potential
for sediments to be eroded from a defined sediment source, transported downvalley
and temporarily depositedwithin the channel (or some other unit) before being 'remo-
bilized'. These reworked sediments are often difficult to recognize (Weltje 2012 );
thus, it is generally assumed that the source area sediment travels directly from its
point of detachment to its point of sampling. The degree to which this assump-
tion is violated depends largely on the size of the basin and the degree of physical
connectivity that exists along the drainage network; the chances of determining the
ultimate source of sediment, and not its proximal one, decreases with increasing
catchment size and decreasing connectivity (Miller et al. 2013 ).
Inverse modeling, as defined above, is aimed at determining the relative contri-
bution of sediment from defined source areas to a specific type of river sediment.
Emphasis is placed on the composition of the river sediment and the origin of the
particles contained within it. Some geochemical fingerprinting studies, however, pro-
pose a slightly different objective: to assess the relative amount of sediment eroded
from the defined source areas or source types. The difference between these two
objectives is subtle, but important. When the goal is to determine the amount of sed-
iment eroded from each of the sediment sources, an additional assumption is applied
to geochemical fingerprinting. It must be assumed that the sediment leaves all sources
at the same time and is transported downstream at an equal rate so that it arrives at
the sampling point simultaneously. This assumption is often violated by differences
in the proximity of a source to the sampled depositional area, or by differences in
the rate at which particles of differing size or shape are transported downstream
(the transport variance problem). Take, for example, a 2 cm thick sample collected
from the surface of a floodplain that received sediment from two upstream sources.
One source is located immediately adjacent to the sampling site, whereas the other is
located a considerable distance upstream. Also assume that equal amounts of material
are eroded from both sediment sources, and the rate of sediment deposition from both
sources is the same. At the onset of the runoff event sediment from the closest source
will reach the site first; thus, the lower portions of the 2 cm thick sampling interval
will be composed of material from only this source. As the event continues, material
from the other source reaches the site, and equal proportions of sediment from both
sources will be deposited at the site. If the entire 2 cm of sediment is not composed
of a single event, the other events will follow the same pattern until 2 cm of sediment
has been accumulated. When the inverse/mixing model is applied to the sample, it
will correctly indicate that a larger relative percent of sediment was derived from the
closest site over the timeframe represented by the 2 cm increment (this is the objective
of the inverse modeling as defined earlier). Thus, sediment provenance with respect