Geoscience Reference
In-Depth Information
To date, there have been very few studies that have attempted to model both flow and water qual-
ity in a totally consistent way. To do so requires that the different dynamics of celerities and the full
distribution of pore water velocities would need to be properly represented (as well as all the other
biogeochemical processes that affect water quality in the soil, groundwater and stream channels with
their own uncertainties in process representations and effective parameter values). Most water quality
models at the catchment scale rely on mixing in conceptual storage elements in some way, with no
attempt to represent flow pathways (understandably so, given the complexity and time variability of
flow pathways).
This is, however, a promising direction for future work and pointers in that direction are provided by
the work of Botter
et al.
(2009, 2010) and Duffy (2010). Other techniques in this chapter, such as the
end member mixing analysis (Section 11.5) can also be used in water quality contexts (see, for example,
Jarvie
et al.
(2011) and references therein).
11.11
Key Points from Chapter 11
•
The use of artificial and environmental tracers to understand and improve the respresenta-
tion of hydrological processes is a valuable tool and has led to new insights into catchment
behaviour.
•
The “paradox” of how catchments can store water for such long periods of time but so easily mobilise
so much “old” water during storms is, in part at least, explained by the different storage mechanisms
involved in the wave celerities that control the hydrograph response and the pore water velocities that
control tracer transport.
•
The assumptions required in analysing sources of water and residence time distributions represent
significant simplifications, so that inferences may be subject to multiple sources of uncertainty. Un-
certainty may also be introduced by the limitations of the sampling of concentrations in inputs and
discharges. Consideration of the uncertainties associated with the analysis should be a matter of good
practice.
•
Collection of tracer concentrations at shorter time scales, including oxygen and hydrogen isotopes,
is becoming cheaper and easier. This should mean that assumptions of linearity and stationarity in
residence time distributions can be relaxed. Theoretical constructs that allow time varying residence
time distributions still need to be properly tested.
•
The use of both flow and tracer data to calibrate hydrological models needs to take account of the fact
that the effective storages involved in the hydrograph and tracer responses are different. Thus, using
tracer observations requires additional parameters to be identified, which means that the uncertainties
in the discharge predictions may not necessarily be significantly reduced.
