freshwater at 4 °C. The exact temperature for the overturn
might, however, deviate slightly from 4 °C, if the wind
forcing during the holomixis is strong enough to overcome
the density gradients. Nevertheless, the possible temperature
range of 3–5 °C is relatively small. In regions where air
temperatures of the coldest month are equal to, or below, 4 °
C, positive deviations from 4 °C are restricted to a short
period during the onset of the holomixis, and negative
deviations are limited to the winter and to cases of a continuing wind-forced overturn. Most probably, the deep water
is at 4 °C at the moment of stabilisation due to spring
warming. During summer, the hypolimnion is efficiently
insulated from irradiative warming. Small temperature
increases up to 6 °C are, however, observed in lakes in
regions with high geothermal gradients. Warmer
bottom-water temperatures occur in meromictic lakes, in
which the winter overturn only reaches down to a limited
depth and in which a deeper and denser water body (monimolimnion) exists. However, as mentioned above, no benthic in situ ostracod and mollusc fauna will be found there
due to the absence of oxygen in the monimolimnion. In
order to further minimize any error in the reconstruction of
d
18 O L from benthic carbonate fossils, bottom-water temperatures of the respective lakes should be followed over
several seasons and compared with those simulated by
energy- and water-balance models driven by observed climate conditions for the same period. The models, validated
in this way, can then be used to estimate bottom water
temperatures and thus provide error estimates for d
18 O L for a
large range of climate conditions. For Ammersee and Lac
d’Annecy, two of the few lakes which should be excellently
suited for quantitative reconstruction of d
18 O P , this modelling approach was used, confirming deep water temperatures astonishingly constant at 4 °C for a very wide range of
climate conditions, colder than today for Ammersee, but
with the risk of meromixis in warmer climates, especially
when winter mean daily temperatures remain significantly
above 4 °C. In contrast, Lac d’Annecy continues to be
episodically meromictic during modelled warmer periods,
because of its higher transparency during the warmest
summer month, allowing deep water to reach the warmer
winter minimal air temperatures within a couple of years
(Danis et al. 2003, 2004)
In our example from the Ammersee, the average temperature at 80 m over the last 20 years (1980–2000, data
provided by Dr. B. Lenhardt, WWA Weilheim) was 4.15 °
C, with a standard deviation of 0.45 °C, and extremes of
3.2 °C and 5.0 °C. With such a narrow range of water
temperatures and the relatively constant d
18 O L of the
hypolimnion (see Fig. 15.1c), the preferred moulting and
calcification period of the ostracods and their instars is
almost irrelevant. This is in strong contrast to sites within the
epilimnion, where differences of up to 3‰ between the d
18
O
of summer and winter produced valves (after correction for
vital offsets) are common (von Grafenstein et al. 1994;
1999a, b, 2013; Dettman et al. 1995). However, a problem
with such deep-lake studies is the low abundance of ostracods compared to littoral sites; this is most probably related
to the combination of low population densities of the different species with relatively elevated sediment accumulation rates (in average *1 mm per year in core AS96-1 from
Ammersee). High-resolution d
18 O records approaching the
hydrologic resolution (ca. 8 years/sample) are therefore only
possible if 10 to 20 valves of juveniles (instars A-5 to A-2)
of the most abundant species Fabaeformirscandona levanderi, F. tricicatricosa or Candona candida are grouped
together to produce samples of >10 lg calcite. This mixing
of species and instars is not problematic, however, as the
studied European Candonidae share identical vital offsets for
d
18 O.
The error of d
18 O L reconstruction from Ammersee deep
lake ostracod d
18 O, is about ±0.2‰. The error of the d
18 O P
calculation, introduced by assuming that evaporative
enrichment and catchment effects were as today, could, in
times of extreme hydrological conditions, range from
−0.3‰ to +0.6‰.
The Fossil Ostracod Record
Calibration Against the Instrumental Air
Temperature Record
Despite the efforts to understand and quantify the transfer of
an atmospheric d
18 O P signal into a sedimentary archive, two
points of significance with respect to paleoclimatic studies
could not be addressed due to the short period of the relevant
field observation. The first concerns the relationship between
mean annual d
18 O P and mean annual air temperature based
on regional and European-wide inter-annual variability of
the of the last three decades (Rozanski et al. 1992). This
comparison had to be extended to at least the duration of
existing instrumental records, to maximize the range of
observed temperature changes and to meet the standards for
the calibration of other paleotemperature proxies. The second open point was the overall reactivity of the coupled
drainage basin-lake system to changes in d
18
O P , where,
especially, the average retention time of the slowest runoff
(groundwater) could only be roughly estimated to *3 years
from a tentative match between a four-year shift of d
18 O L
and long-term averages of d
18 O P .
Both concerns could be addressed by establishing a
300-year long ostracod-derived record of d
18 O L from the
uppermost 120 cm of sediment in 80 m water depth in
Ammersee (von Grafenstein et al. 1996). The age-depth
model of the core was based on clearly identifiable annual
15 Air-Interface: d
18 O Records of Past Meteoric Water Using …
187
might, however, deviate slightly from 4 °C, if the wind
forcing during the holomixis is strong enough to overcome
the density gradients. Nevertheless, the possible temperature
range of 3–5 °C is relatively small. In regions where air
temperatures of the coldest month are equal to, or below, 4 °
C, positive deviations from 4 °C are restricted to a short
period during the onset of the holomixis, and negative
deviations are limited to the winter and to cases of a continuing wind-forced overturn. Most probably, the deep water
is at 4 °C at the moment of stabilisation due to spring
warming. During summer, the hypolimnion is efficiently
insulated from irradiative warming. Small temperature
increases up to 6 °C are, however, observed in lakes in
regions with high geothermal gradients. Warmer
bottom-water temperatures occur in meromictic lakes, in
which the winter overturn only reaches down to a limited
depth and in which a deeper and denser water body (monimolimnion) exists. However, as mentioned above, no benthic in situ ostracod and mollusc fauna will be found there
due to the absence of oxygen in the monimolimnion. In
order to further minimize any error in the reconstruction of
d
18 O L from benthic carbonate fossils, bottom-water temperatures of the respective lakes should be followed over
several seasons and compared with those simulated by
energy- and water-balance models driven by observed climate conditions for the same period. The models, validated
in this way, can then be used to estimate bottom water
temperatures and thus provide error estimates for d
18 O L for a
large range of climate conditions. For Ammersee and Lac
d’Annecy, two of the few lakes which should be excellently
suited for quantitative reconstruction of d
18 O P , this modelling approach was used, confirming deep water temperatures astonishingly constant at 4 °C for a very wide range of
climate conditions, colder than today for Ammersee, but
with the risk of meromixis in warmer climates, especially
when winter mean daily temperatures remain significantly
above 4 °C. In contrast, Lac d’Annecy continues to be
episodically meromictic during modelled warmer periods,
because of its higher transparency during the warmest
summer month, allowing deep water to reach the warmer
winter minimal air temperatures within a couple of years
(Danis et al. 2003, 2004)
In our example from the Ammersee, the average temperature at 80 m over the last 20 years (1980–2000, data
provided by Dr. B. Lenhardt, WWA Weilheim) was 4.15 °
C, with a standard deviation of 0.45 °C, and extremes of
3.2 °C and 5.0 °C. With such a narrow range of water
temperatures and the relatively constant d
18 O L of the
hypolimnion (see Fig. 15.1c), the preferred moulting and
calcification period of the ostracods and their instars is
almost irrelevant. This is in strong contrast to sites within the
epilimnion, where differences of up to 3‰ between the d
18
O
of summer and winter produced valves (after correction for
vital offsets) are common (von Grafenstein et al. 1994;
1999a, b, 2013; Dettman et al. 1995). However, a problem
with such deep-lake studies is the low abundance of ostracods compared to littoral sites; this is most probably related
to the combination of low population densities of the different species with relatively elevated sediment accumulation rates (in average *1 mm per year in core AS96-1 from
Ammersee). High-resolution d
18 O records approaching the
hydrologic resolution (ca. 8 years/sample) are therefore only
possible if 10 to 20 valves of juveniles (instars A-5 to A-2)
of the most abundant species Fabaeformirscandona levanderi, F. tricicatricosa or Candona candida are grouped
together to produce samples of >10 lg calcite. This mixing
of species and instars is not problematic, however, as the
studied European Candonidae share identical vital offsets for
d
18 O.
The error of d
18 O L reconstruction from Ammersee deep
lake ostracod d
18 O, is about ±0.2‰. The error of the d
18 O P
calculation, introduced by assuming that evaporative
enrichment and catchment effects were as today, could, in
times of extreme hydrological conditions, range from
−0.3‰ to +0.6‰.
The Fossil Ostracod Record
Calibration Against the Instrumental Air
Temperature Record
Despite the efforts to understand and quantify the transfer of
an atmospheric d
18 O P signal into a sedimentary archive, two
points of significance with respect to paleoclimatic studies
could not be addressed due to the short period of the relevant
field observation. The first concerns the relationship between
mean annual d
18 O P and mean annual air temperature based
on regional and European-wide inter-annual variability of
the of the last three decades (Rozanski et al. 1992). This
comparison had to be extended to at least the duration of
existing instrumental records, to maximize the range of
observed temperature changes and to meet the standards for
the calibration of other paleotemperature proxies. The second open point was the overall reactivity of the coupled
drainage basin-lake system to changes in d
18
O P , where,
especially, the average retention time of the slowest runoff
(groundwater) could only be roughly estimated to *3 years
from a tentative match between a four-year shift of d
18 O L
and long-term averages of d
18 O P .
Both concerns could be addressed by establishing a
300-year long ostracod-derived record of d
18 O L from the
uppermost 120 cm of sediment in 80 m water depth in
Ammersee (von Grafenstein et al. 1996). The age-depth
model of the core was based on clearly identifiable annual
15 Air-Interface: d
18 O Records of Past Meteoric Water Using …
187