Geoscience Reference
In-Depth Information
Freshwater
Ecosystem
Surface
water
Soil water
Out-flow areas
In-flow areas
Lakes
Rivers
Target ecosystem
for metal
contamination
Oligotrophic
Mesotrophic
Eutrophic
Fig. 4.17
The fresh water ecosystem may be
divided into several subecosystems (such as
oligotrophic, mesotrophic and eutrophic lakes)
where different key organisms prevail. As part of
this differentiation process, one must decide which
subsystem should be used as the target ecosystem
for a given chemical threat, which target organisms,
functional groups or effect variables should be used
for a given threat, and what load and sensitivity
variables should be used relative to a given
effect variable. The target ecosystem for metals
(such as Hg) and radionuclides (such as
137
Cs)
is low-productive (oligotrophic) lakes (shaded
in this figure): ELS, effect-load-sensitivity.
(From Håkanson & Peters 1995.)
1 2 3 4 5 6 7 8 9 10 11 12
ELS models available
for
137
Cs and Hg
12 environmental
threats
tool for quantitative predictions which relate
operationally defined ecological effects to com-
patible load and sensitivity variables (Håkanson
1999).
Differential equations and mass-balance
models are often used to handle fluxes (e.g.
g
X
yr
−1
), amounts (g
X
) and concentrations
(g
X
m
−3
) of all types of materials (such as gases,
carbon, toxins and nutrients), but not ecosystem
effect variables (
E
). Statistical methods, such as
regressions based on empirical data, are generally
necessary to relate concentrations of chemicals
to effect variables (
E
). In theory, both these
model approaches (see Fig. 4.18) may be used
for the ELS models, provided that at least one
operationally defined ecological effect variable
relevant for the load variable(s) in question is
included in the model. Ideally, the
E
variable
should express the reproduction, abundance,
mass or status of defined functional organisms
(preferably fish at the top trophic level) that
characterize the given ecosystem. Such ideal effect
variables cannot normally be predicted with dif-
ferential equations and mass-balance models. If
the more ideal effect variables cannot be deter-
mined, then in practice one has to do the second
best and define operational effect variables, such
as toxic concentrations in fish eaten by humans,
the oxygen concentration in the deep-water
zone and the Secchi depth (see Table 4.10). The
sensitivity variables express how different lake
characteristics (such as pH, colour, total-P con-
centrations and lake mean depth) regulate the
'road between load and effect'. One and the same
load will cause different ecosystem effects in lakes
of different sensitivities. For example, in lakes
