514
11.4.3 Evolution of Closed Lake Basins
(Mixed Rock Types)
A more realistic, modem approach is the use of geochemical computer pro grams (e.g. The Geochemist's
Workbench; Bethke 1996) to study water-rock interactions and mineral successions in lakes. These programs are based on thermodynamic equilibria between solutions (inflowing water and lake water) and
solid phases (precipitating minerals including
backreactions with pre-existing minerals). If the
chemistry of inflowing waters and the hydrological
budget of a closed lake basin are known, one can
simulate the past or future evolution of the basin. The
validity of backward modeling can be checked by
investigating the sediments present in the lake. The
mass of authigenic lake sediments, including their
porewater composition, allows independent establishment of a mass balance of the dissolved matter
influx for a certain time period. Provided these tests
are possible and turn out satisfactorily, the modeling
can be extended to predict the lake development in
the future.
Figure 11.22 shows examples of this type of
foreward modeling for two modem closed lakes,
Lake Qinghai in Tibet and Lake Turkana in East Africa. Some data about the characteristics of these
lakes and their drainage areas are listed in Table
11.1.
Lake Qinghai, Tibet, lies in a rift depression (mainly sedimentary rocks) and has existed since the early Pleistocene
(Yan 1999; Yan et al. 2000). The simulation of this lake is
based on its development during the Holocene (10 ka)
when the lake became increasingly salty.
Lake Turkana, East Africa, lies in the eastem branch
ofthe East African rift zone (cf. Fig. 12.3) and has evolved
since about 4.5 Ma. It has a complex history (e.g. Cerling
1986). The data used below refer to the development during the past 4 ka. The northem drainage area of this lake,
which delivers 80-90% of the inflow, is characterized by
young basalts whereas in other parts granitic and metamorphic rocks also occur.
By extrapolating the mean sedimentation rates of the
lakes into the future, the lifetime of the lakes up to overflow was determined. Furthermore, the storage capacities
of the basins and their subsidence was taken into account
(Lake Qinghai 200 mmlka, Lake Turkana 230 mmlka).
Lake Qinghai was found to persist for about 300 ka, Lake
Turkana for 100-150 ka. These time periods were taken as
limits for the forward modeling.
Fig. 11.22. Simulation of the future evolution of two
naturallakes, L. Qinghai, Tibet, and L. Turkana, East
African rift. (After Yan 1999, modified). The two
hydrological models meet both reduction in water
volume and leakage ofthe lake basins. The solute input
(in 10 16 g) is based on modem measurements and gives
the masses for 300 ka and 100 ka, respectively). The
Chapter 11 Sediment Supply, Subsidence, and Basin Fill
The influx of dissolved matter was taken from chemical
analyses of river water and mean chemical denudation rates
in the drainage areas (Lake Qinghai 20 mmlka, Lake
Turkana 7.5 mmlka).
F or both lakes two different scenarios are assumed:
(1) the salt concentration increases with time due to
influx and, in addition, the initial (present-day) water
volume, VI' of the lakes is reduced to the ultimate
volume V E=O.l VI (additional evaporative concentration), or (2) VI remains constant with time (V E=VI)
and the salt concentration solely increases due to influx.
The input of dissolved species has to be subdivided into solutes derived from rocks and others
coming from the atmosphere (mainly HC0 3 and Cl).
Only part of the atrnospheric HC0 3 is used up for
mineral precipitation, the remainder is again released
into the atmosphere or stored in the lake. Due to the
various losses (part of HC0 3 , leakage, regradation of
clay minerals, sulfate reduction) the mass of minerals
and solutes remaining in the lake and its sediments is
less than the solute input. This is particularly so in
Lake Turkana. The succession of salt minerals and
their mean sedimentation rates in the two lakescan
be seen from Fig. 11.22.
Lake Qinghai is slightly leaky (loss of 25% of the
inflowing Na and Cl to groundwater) and Lake Turkana is
very leaky (75% loss). As a result, the salt concentration
increases faster in Lake Qinghai and in model (I) of Lake
Turkana than in model (2) ofLake Turkana with the result
that, even after a time period of 100 ka, the salinity is not
high enough for precipitation of highly soluble minerals.
With less leakage, model (2) of Lake Turkana would produce a mineral succession similar to that ofmodel (I).
Ions consumed by the regradation and neo-formation of
clay minerals are not considered in Fig. 11.22. In Lake
Qinghai they make up around 6% of the influx (80% of
K+), in Lake Turkana 25% of the influx. In Lake Qinghai,
Ca- and Mg-carbonates will be precipitated during the first
20 ka ofthe simulation. Primary magnesite will be replaced
by dolomite during early diagenesis. After this phase,
gaylussite (Ca-Na~carbonate) will form but, due to further
solute influx, the precipitation of calcite and magnesite
continues at a somewhat reduced rate. After 60 to 80 ka,
mirabilite (Na-sulfate) starts to precipitate along with the
carbonate minerals. Saturation with respect to halite (NaCl)
is reached after about 130 ka (model I) or 220 ka (model
2). Sylvite (KCI) only forms in model I at the very end of
the lake evolution.
It should be noted that in reality the various salts do not
precipitate simultaneously as indicated in the diagrams.
masses precipitated as authigenic minerals and left in
lake water are reduced by leakage, ion uptake by
weathered clay minerals and release of CO 2 into the
atmosphere. Note that second and following minerals
react with formerly precipitated minerals and thus
reduce their sedimentation rate. See text for further
explanation
11.4.3 Evolution of Closed Lake Basins
(Mixed Rock Types)
A more realistic, modem approach is the use of geochemical computer pro grams (e.g. The Geochemist's
Workbench; Bethke 1996) to study water-rock interactions and mineral successions in lakes. These programs are based on thermodynamic equilibria between solutions (inflowing water and lake water) and
solid phases (precipitating minerals including
backreactions with pre-existing minerals). If the
chemistry of inflowing waters and the hydrological
budget of a closed lake basin are known, one can
simulate the past or future evolution of the basin. The
validity of backward modeling can be checked by
investigating the sediments present in the lake. The
mass of authigenic lake sediments, including their
porewater composition, allows independent establishment of a mass balance of the dissolved matter
influx for a certain time period. Provided these tests
are possible and turn out satisfactorily, the modeling
can be extended to predict the lake development in
the future.
Figure 11.22 shows examples of this type of
foreward modeling for two modem closed lakes,
Lake Qinghai in Tibet and Lake Turkana in East Africa. Some data about the characteristics of these
lakes and their drainage areas are listed in Table
11.1.
Lake Qinghai, Tibet, lies in a rift depression (mainly sedimentary rocks) and has existed since the early Pleistocene
(Yan 1999; Yan et al. 2000). The simulation of this lake is
based on its development during the Holocene (10 ka)
when the lake became increasingly salty.
Lake Turkana, East Africa, lies in the eastem branch
ofthe East African rift zone (cf. Fig. 12.3) and has evolved
since about 4.5 Ma. It has a complex history (e.g. Cerling
1986). The data used below refer to the development during the past 4 ka. The northem drainage area of this lake,
which delivers 80-90% of the inflow, is characterized by
young basalts whereas in other parts granitic and metamorphic rocks also occur.
By extrapolating the mean sedimentation rates of the
lakes into the future, the lifetime of the lakes up to overflow was determined. Furthermore, the storage capacities
of the basins and their subsidence was taken into account
(Lake Qinghai 200 mmlka, Lake Turkana 230 mmlka).
Lake Qinghai was found to persist for about 300 ka, Lake
Turkana for 100-150 ka. These time periods were taken as
limits for the forward modeling.
Fig. 11.22. Simulation of the future evolution of two
naturallakes, L. Qinghai, Tibet, and L. Turkana, East
African rift. (After Yan 1999, modified). The two
hydrological models meet both reduction in water
volume and leakage ofthe lake basins. The solute input
(in 10 16 g) is based on modem measurements and gives
the masses for 300 ka and 100 ka, respectively). The
Chapter 11 Sediment Supply, Subsidence, and Basin Fill
The influx of dissolved matter was taken from chemical
analyses of river water and mean chemical denudation rates
in the drainage areas (Lake Qinghai 20 mmlka, Lake
Turkana 7.5 mmlka).
F or both lakes two different scenarios are assumed:
(1) the salt concentration increases with time due to
influx and, in addition, the initial (present-day) water
volume, VI' of the lakes is reduced to the ultimate
volume V E=O.l VI (additional evaporative concentration), or (2) VI remains constant with time (V E=VI)
and the salt concentration solely increases due to influx.
The input of dissolved species has to be subdivided into solutes derived from rocks and others
coming from the atmosphere (mainly HC0 3 and Cl).
Only part of the atrnospheric HC0 3 is used up for
mineral precipitation, the remainder is again released
into the atmosphere or stored in the lake. Due to the
various losses (part of HC0 3 , leakage, regradation of
clay minerals, sulfate reduction) the mass of minerals
and solutes remaining in the lake and its sediments is
less than the solute input. This is particularly so in
Lake Turkana. The succession of salt minerals and
their mean sedimentation rates in the two lakescan
be seen from Fig. 11.22.
Lake Qinghai is slightly leaky (loss of 25% of the
inflowing Na and Cl to groundwater) and Lake Turkana is
very leaky (75% loss). As a result, the salt concentration
increases faster in Lake Qinghai and in model (I) of Lake
Turkana than in model (2) ofLake Turkana with the result
that, even after a time period of 100 ka, the salinity is not
high enough for precipitation of highly soluble minerals.
With less leakage, model (2) of Lake Turkana would produce a mineral succession similar to that ofmodel (I).
Ions consumed by the regradation and neo-formation of
clay minerals are not considered in Fig. 11.22. In Lake
Qinghai they make up around 6% of the influx (80% of
K+), in Lake Turkana 25% of the influx. In Lake Qinghai,
Ca- and Mg-carbonates will be precipitated during the first
20 ka ofthe simulation. Primary magnesite will be replaced
by dolomite during early diagenesis. After this phase,
gaylussite (Ca-Na~carbonate) will form but, due to further
solute influx, the precipitation of calcite and magnesite
continues at a somewhat reduced rate. After 60 to 80 ka,
mirabilite (Na-sulfate) starts to precipitate along with the
carbonate minerals. Saturation with respect to halite (NaCl)
is reached after about 130 ka (model I) or 220 ka (model
2). Sylvite (KCI) only forms in model I at the very end of
the lake evolution.
It should be noted that in reality the various salts do not
precipitate simultaneously as indicated in the diagrams.
masses precipitated as authigenic minerals and left in
lake water are reduced by leakage, ion uptake by
weathered clay minerals and release of CO 2 into the
atmosphere. Note that second and following minerals
react with formerly precipitated minerals and thus
reduce their sedimentation rate. See text for further
explanation