Geoscience Reference
In-Depth Information
Figure 11.2 Mixing diagram for direct precipitation, soil water and groundwater end members defined for
silica and calcium in the Haute-Mentue catchment, Switzerland (after Iorgulescu et al., 2005, with kind per-
mission of John Wiley and Sons).
The first assumption can be checked by plotting the tracer concentrations for each end member on a
mixing diagram similar to Figure 11.2, which shows soil water, rainfall and baseflow silica and calcium
concentrations for the Haute-Mentue catchment in Switzerland. In this case, these two tracers separate the
components quite nicely, since an acid soil (high in silica, low in calcium) has developed over a regolith of
molasse bedrock (high in calcium, low in silica), while the rainfall should be low in both calcium and silica.
The analysis might still only be approximate, however. In particular, it might be expected that soil
water concentrations would vary over time as the soil wets and dries, so that a set of samples might
be needed prior to each event to improve the accuracy of the separation. There is also an issue about
whether the soil concentrations (measured on samples that are usually obtained by imposing a suction)
are representative of the mobile water stored in the soil that might contribute to the hydrograph. There is
also the possibility that, when the event water interacts with the soil, it takes up silica into solution and
rather quickly starts to appear as if it were soil water, thereby resulting in an overestimation of the soil
water contribution and an underestimation of the event water contribution.
Similar issues arise in most EMMA applications and the validity of the assumptions of such an analysis
should always be examined with some care. One way of doing so, as matter of good practice, is to evaluate
the uncertainties in the resulting separations. Hooper et al. (1990) and Genereux (1998) used analytical
forward uncertainty propagation to do this while Bazemore et al. (1994), Joerin et al. (2002) and Soulsby
et al. (2003) have used Monte Carlo simulation.
EMMA can also be used with multiple chemical species contributing to the definition of the end
members (Burns et al. , 2001; Hooper, 2003) and to test for the conservativity of pollutants in the system.
Christophersen et al. (1990), using two source components, tested for deviations from the 1:1 mixing
line between end members, while recently Jarvie et al. (2011) showed how EMMA can be extended for
end members that are expected to vary with discharge.
11.6 On the Implications of Tracer Information for
Hydrological Processes
The results of such mixing model calculations often suggest that a large part of the hydrograph is made
up of water that has been stored in the catchment for a long time (see also Section 1.4). Kirchner
Search WWH ::




Custom Search