Quaternary Science Reviews 30 (2011) 443e459
Contents lists available at ScienceDirect
Quaternary Science Reviews
journal homepage: www.elsevier.com/locate/quascirev
The significance of chemical, isotopic, and detrital components in three coeval
stalagmites from the superhumid southernmost Andes (53 S) as high-resolution
palaeo-climate proxies
Daniel Schimpf a, Rolf Kilian a, *, Andreas Kronz b, Klaus Simon b, Christoph Spötl c, Gerhard Wörner b,
Michael Deininger d, Augusto Mangini d
a
Lehrstuhl für Geologie, Fachbereich VI, Universität Trier, Behringstr. 16, D-54286 Trier, Germany
Geowissenschaftliches Zentrum Göttingen, Geochemie, Universität Göttingen, Goldschmidtstr. 1, D-37077 Göttingen, Germany
Institut für Geologie und Paläontologie, Universität Innsbruck, Innrain 52, A-6020 Innsbruck, Austria
d
Heidelberger Akademie der Wissenschaften, c/o Institut für Umweltphysik, Im Neuenheimer Feld 229, D-69120 Heidelberg, Germany
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 June 2010
Received in revised form
4 December 2010
Accepted 6 December 2010
Available online 5 January 2011
Stalagmites are important palaeo-climatic archives since their chemical and isotopic signatures have the
potential to record high-resolution changes in temperature and precipitation over thousands of years.
We present three U/Th-dated records of stalagmites (MA1eMA3) in the superhumid southern Andes,
Chile (53 S). They grew simultaneously during the last five thousand years (ka BP) in a cave that
developed in schist and granodiorite. Major and trace elements as well as the C and O isotope compositions of the stalagmites were analysed at high spatial and temporal resolution as proxies for palaeotemperature and palaeo-precipitation. Calibrations are based on data from five years of monitoring the
climate and hydrology inside and outside the cave and on data from 100 years of regional weather
station records.
Water-insoluble elements such as Y and HREE in the stalagmites indicate the amount of incorporated
siliciclastic detritus. Monitoring shows that the quantity of detritus is controlled by the drip water rate
once a threshold level has been exceeded. In general, drip rate variations of the stalagmites depend on
the amount of rainfall. However, different drip-water pathways above each drip location gave rise to
individual drip rate levels. Only one of the three stalagmites (MA1) had sufficiently high drip rates to
record detrital proxies over its complete length. Carbonate-compatible element contents (e.g. U, Sr, Mg),
which were measured up to sub-annual resolution, document changes in meteoric precipitation and
related drip-water dilution. In addition, these soluble elements are controlled by leaching during
weathering of the host rock and soils depending on the pH of acidic pore waters in the peaty soils of the
cave’s catchment area. In general, higher rainfall resulted in a lower concentration of these elements and
vice versa. The Mg/Ca record of stalagmite MA1 was calibrated against meteoric precipitation records for
the last 100 years from two regional weather stations. Carbonate-compatible soluble elements show
similar patterns in the three stalagmites with generally high values when drip rates and detrital tracers
were low and vice versa. d13C and d18O values are highly correlated in each stalagmite suggesting
a predominantly drip rate dependent kinetic control by evaporation and/or outgassing. Only C and O
isotopes from stalagmite MA1 that received the highest drip rates show a good correlation between
detrital proxy elements and carbonate-compatible elements. A temperature-related change in rainwater
isotope values modified the MA1 record during the Little Ice Age (w0.7e0.1 ka BP) that was w1.5 C
colder than today. The isotopic composition of the stalagmites MA2 and MA3 that formed at lower drip
rates shows a poor correlation with stalagmite MA1 and all other chemical proxies of MA1. ‘Hendy tests’
indicate that the degassing-controlled isotope fractionation of MA2 and MA3 had already started at the
cave roof, especially when drip rates were low. Changing pathways and residence times of the seepage
water caused a non-climatically controlled isotope fractionation, which may be generally important in
Keywords:
Stalagmite
Andes
Palaeo-climate
Holocene
Element composition
Stable isotopes
Medieval Warm Period
Little Ice Age
* Corresponding author. Tel.: þ49 651 201 4644; fax: þ49 651 201 3915.
E-mail addresses: [email protected] (D. Schimpf), [email protected] (R. Kilian), [email protected] (A. Kronz), [email protected] (K. Simon), [email protected]
(C. Spötl), [email protected] (G. Wörner), [email protected] (M. Deininger), [email protected] (A. Mangini).
0277-3791/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.quascirev.2010.12.006
444
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
ventilated caves during phases of low drip rates. Our proxies indicate that the Neoglacial cold phases
from w3.5 to 2.5 and from w0.7 to 0.1 ka BP were characterised by 30% lower precipitation compared
with the Medieval Warm Period from 1.2 to 0.8 ka BP, which was extremely humid in this region.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
Stalagmites represent important palaeo-climatic archives due to
a precise U/Th age control and the possibility of sub-annual resolution of their climate-related isotopic and geochemical signals
(Henderson, 2006). However, palaeo-climate interpretations are
often based on a single stalagmite record from one cave. Only a few
case studies have reported variations of proxies of multiple stalagmites from a given cave (e.g. Dorale et al.,1998; Williams et al., 2004;
Vollweiler et al., 2006). The variability of chemical and isotopic
proxies between different stalagmites and/or their dependence on
individual drip rate characteristics has rarely been investigated.
Changes of the drip-water pathways may produce chemical and
isotopic signals which are not directly related to climate, especially
in caves with enhanced evaporation due to ventilation.
Here we present data from three Late Holocene stalagmites
(MA1, MA2, and MA3) that grew close together in a small non-karst
cave in the fjord region of the southernmost Andes (Chile, 53 S,
Fig. 1). The location represents the core section of the Southern
Hemispheric Westerlies (SHW). To our knowledge, this is the first
study of speleothem records from this region. The aim of this study
is to evaluate the climatic sensitivity of chemical and isotopic
tracers in these stalagmites and calibrate them against instrumental data (temperature and precipitation) over the past 100
years. Furthermore, we monitored the cave for five years. A major
aim is to resolve extreme short-term climate events in our highresolution proxy record, such as three-day 1000 mm precipitation
periods that have been documented to be associated with extreme
storms in the area (Schneider et al., 2003). We analysed the stable
isotopes of C and O, and the concentrations of the carbonatecompatible elements Mg, U, and Sr as palaeo-environmental
proxies. Elements like Y, Al, Ti, Zr, and HREE are usually not dissolved in the cave drip water and are fixed in siliciclastic particles,
or sometimes adsorbed on colloidal particles (Borsato et al., 2007).
In non-karst caves, such detrital minerals can be transported with
the drip water onto the stalagmite. The variations of the detritusindicating trace elements in the stalagmites were investigated in
order to detect their sensitivity to drip rate and precipitation
changes. Our new speleothem-based high-resolution temperature
and precipitation records from southern Patagonia are compared
with other palaeo-climatic reconstructions from the southern
Andes, which is a key region for understanding past variability of
the SHW and their forcing mechanisms.
2. Site description
The “Marcelo Arévalo” cave (location MA in Fig. 1; 52 41.70 S/
73 23.30 W) was unofficially named after its discoverer in 2002. It is
located 15 km NW of the centre of the 200 km2 large Gran Campo
Nevado (GCN) ice field, which represents the highest elevation and
a significant climate divide in this section of the Andes (Schneider
et al., 2007). The cave is situated about 20 m above sea level on the
shore of a small bay in a fjord system along the Pacific coast (Fig. 1).
This extremely windy and superhumid area in the western range of
the Andes is unpopulated and can only be reached by ship.
South of 40 S the Andes represent one of the most pronounced
climate divides in the world since this mountain range is located
perpendicular to the SHW belt. Southernmost South America is the
only continental landmass within the SHW core. Hence, the southernmost Andes are the only barrier to these winds that would
otherwise blow nearly unimpeded around the globe. The very cold
Antarctic continent causes a steep temperature and pressure
gradient between the high latitudes and the subtropics giving rise to
stronger westerly winds than at similar latitudes in the Northern
Hemisphere (Schneider et al., 2003). The SHW belt stretches from
35 to 60 S and exhibits distinct seasonal as well as millennial-scale
changes in its wind and precipitation patterns (e.g. Lamy et al., 2007,
2010). At the Southern Andes climate divide, precipitation is mainly
controlled by the SHW intensities.
The MA cave is located in the core of the SHW (Fig. 1) where
throughout the year strong winds and very high precipitation occur
(up to 10,000 mm per year at the Gran Campo Nevado with maxima
of up to 500 mm per day; Schneider et al., 2003). NCEP-NCAR1 data
between 1960 and 2000 indicate that precipitation and wind speed
in the SHW core are higher in summer and lower in winter (Fig. 1a
inset), whereas the northern margin of the SHW in central Chile
(30e40 S) shows higher humidity in winter and significantly lower
humidity in summer (Lamy et al., 2010).
The cave was formed by coastal weathering and erosion in
a fracture zone during the Lateglacial or Early Holocene when
the coastline was 20 m higher than today. The host rock consists
of granite, granodiorite and gneiss; hence, karst dissolution
did not play a role. The Ca in the drip water is derived from
Ca-bearing silicate minerals (feldspar, hornblende, clinopyroxene)
leached from the soil-weathering horizon in the cave catchment
area and from sea spray. The restricted catchment area of the
cave (less than 1 km2) lies on a relatively flat crest that is
covered by peat vegetation with acidic (pH 3e5) soil water. The
well-drained steep slopes are overgrown by evergreen Magellanes rain forest (Fig. 1b and c). The MA cave is situated on
a steep southeastern slope (no direct sunlight) and is 15 m deep,
up to 6 m high and 2e3 m wide. Large trees (up to 20 m) cover
the relatively wide cave entrance. Due to this and the cave’s
easterly exposure only very weak winds are noticeable inside.
Thus, the cave is relatively well protected during the frequent
and strong storms from mainly western directions. However,
cave temperature and humidity are closely linked to the external
atmospheric conditions (see Section 4.1).
The three stalagmites (MA1eMA3) grew less than 1 m apart from
each other, approximately 7 m behind the cave entrance. Due to the
very close distance between the stalagmites, we assume that they
were probably fed by the same drip-water source. Given their location
nearby the wide entrance, the speleothems are influenced by large
humidity and temperature variations (comparison of inside and
outside temperatures in Fig. 2c) and sea spray (Biester et al., 2002,
2004) which can enter the cave as fine drifts of mist. Salty encrustations near the rain-protected entrance confirm the occasional deposition of sea spray. Furthermore, mosses and algae grew locally on the
surface of the stalagmites during periods of low drip rates.
Two Plinian eruptions of the Mt. Burney volcano, located about
40 km north of the cave (Fig. 1a, 1520 m asl, 52 200 S/73 230 W),
occurred during the growth interval of the stalagmites: a major
1
http://dss.ucar.edu/pub/reanalysis/.
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
445
Fig. 1. (a) The Pacific continental margin of southernmost South America in Chile with the locations of the MA cave (and the nearby automatic weather station Arévalo), the two
weather stations Evangelistas and Félix (both about 100 years old), and the newer automatic weather stations Passo and Bahamondes (10 years old). The glacier field of Gran Campo
Nevado (GCN) and the Mt. Burney volcano to the north of the MA cave are shown. Dashed curves represent isohyets for mean annual precipitation (in mm/year; details see Lamy
et al., 2010), which is highest at the Cordillera and near the GCN. The inset map shows the latitudinal changes of average westerly wind intensities in summer and winter between
1960 and 2000 AD from NCEP-NCAR data. (b) A “Google Earth” image shows the morphology around the Arévalo Bay with the location of the MA cave and the automatic weather
station AWS Arévalo. The catchment area of the MA cave with its major fracture zones as well as the distribution of forest and peat land is also shown. (c) Westeeast cross-section of
the MA cave. The cave entrance is oriented towards the east and is protected from wind and direct sunlight by dense forest. The three stalagmites MA1, MA2 and MA3 are situated in
the backmost part of the cave. There is only weak tree growth on the peaty soil above the cave host rock because of wet/waterlogged soils.
eruption at 4.2 cal ka BP (14C age of the Mt. Burney tephra after
Stern, 2008; calibrated for the Southern Hemisphere using the
SHcal04 database) and a minor eruption at 2.0 ka BP (Kilian et al.,
2006). The air-fall ash of the 4.2 ka BP eruption covered the area
around the cave with a 10 cm tephra layer, whereas the tephra layer
of the 2.0 ka BP eruption was only a few millimetres thick (Kilian
et al., 2003). Kilian et al. (2006) showed that the abundances of U
in stalagmite MA1 increased significantly during the two millennia
after 4.2 ka BP due to a long-term sulphur release from the
deposited tephra that caused a long-term acidification process in
446
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
Fig. 2. (a) Drip rate monitoring and precipitation at the MA cave between 2004 and
2009. The gap in the drip rate record was caused by a logger breakdown. Precipitation
data with a 30-day running mean from the nearby automatic weather station at Bahia
Bahamondes are shown for comparison (location in Fig. 1a). (b) Relative humidity and
temperature monitoring between 2004 and 2008. (c) Comparison of temperatures
inside and in front of the cave.
the peat overlying the cave. In contrast, the 2.0 ka BP tephra did not
induce such a perturbation.
3. Methods
The stalagmites were dated by the 230Th/234U disequilibrium
method. The isotopes were measured with a thermal ionisation
mass spectrometer (MAT 262 RPQ) using the double filament
technique at the Heidelberg Academy of Sciences (Frank et al.,
2000). All ages were calculated using half-lives of 75,381 years
for 230Th, 244,600 years for 234U, and 4.4683 $ 109 years for 238U
(Cheng et al., 2000). The quoted age uncertainties do not include
the half-life uncertainties. Ages were corrected for initial detrital
230
Th assuming an activity ratio of 230Th/232Th from an average
crustal value. The correction for detrital 232Th was performed using
a 232Th/238U mass ratio of 4.8, which is within the uncertainty limits
of the value 3.8 1.0 reported by Wedepohl (1995, see Section 4.3)
for average upper continental crust. In this article, all ages are
reported as ‘Before Present’ (BP, before the year 1950.
The ageedepth models of the stalagmites were constructed
using ‘Akima interpolation’ (Akima, 1970) based on detritus-
corrected Th/U ages. For MA1 and MA3 an additional age correction
(U fine-tuning using the U content) was applied (see Section 4.3).
Dating results are given in Table 1.
Two different sampling strategies were used for the isotope
measurements: first, d18O and d13C values were determined along
the growth axes of the three stalagmites at a resolution of
150 mm (micromilled samples, Fig. 3). Second, isotope measurements were performed within several individual growth layers
(‘Hendy test’; Hendy, 1971) for MA1 and MA2. Sampling started
at the drip centre and followed the individual lamina towards its
end at the margin of the stalagmite for about 90 mm at a resolution of 2.2e3.3 mm (seven growth layers for MA1 and 5 layers
for MA2). Samples were analysed using an online, automated
carbonate preparation system linked to a triple-collector gas
source mass spectrometer at Innsbruck University, Austria. Raw
data are calibrated against NBS19 and values are given relative to
Vienna Pee Dee Belemnite (VPDB) standard. The precision of the
d18O and d13C values is 0.08& and 0.07&, respectively (1-sigma
standard deviation of replicate analyses, Spötl and Vennemann,
2003).
Samples of 1.5 mm diameter were drilled at ca 2 mm intervals
along the stalagmite axes (Fig. 3) and were measured for major,
minor and trace elements by ICP-MS (FISONS VG PQ STE) and ICPOES at Göttingen University, Germany.
Sixteen overlapping thin sections (48 24 mm, 200 mm thick)
were prepared from the axial part of MA1 (Fig. 3), as well as from
the uppermost 50 mm of MA2 and MA3. Four thin sections were
also prepared from representative wall rock samples of the cave. An
electron microprobe (JEOL JXA-8900RL) at GZG, Göttingen University, Germany, was used to measure in 25 mm steps the Mg and Ca
content of the calcite in the stalagmites. Measured spots along track
lines were previously checked by backscatter electron images to
avoid analyses of non-carbonate detrital minerals (see BSE image
with spots in Fig. 3 inset). The chemistry of separated siliciclastic
detritus was analysed in the same way. Quantitative wavelengthdispersive analyses were conducted at an accelerating voltage of
15 kV (beam current 20 nA, spot size 20 mm and 60 s counting time)
using natural and synthetic standards. The detection limit for Mg
was 150 ppm. The organic and inorganic components of the separated detritus were also analysed by a LEO435VP backscatter
electron microscope at the Geology Department of the University of
Trier. Thin sections of cave wall rock samples were analysed by
optical microscopy and electron microprobe.
In March 2004, a drip-water counter as well as a temperature
and humidity sensor was installed in the MA cave about 8 m behind
the entrance next to the stalagmite location. Data were logged
every 3 h until September 2004 and every hour thereafter.
Three Campbell automatic weather stations (AWS) are operated
by the authors in close vicinity of the cave site (locations in Fig. 1a
and b): Puerto Bahamondes (26 m asl, 52 480 S/72 560 W, since
1999), 25 km southeast of the cave at the Canal Gajardo just eastern
of the Gran Campo Nevado ice field (1640 m asl, 52 480 S/73 060 W),
Passo (355 m asl, 52 400 S/73 070 W, since 2000), 19 km east of the
cave, and Arévalo (87 m asl, since September 2007), located 180 m
NW of the cave.
Long-term meteorological records are available from three
regional weather stations (Fig. 1a). The Evangelistas station, which
is off the Pacific coast, has recorded data for more than 100 years
and is situated 127 km WNW of the MA cave on the small off-shore
island Evangelistas (52 240 S/75 060 W, Schneider et al., 2003). The
Félix weather station, situated 77 km WSW of the cave site at Bahía
Félix (52 570 S/74 040 W), provided precipitation data from 1914 to
1985. The weather station at Punta Arenas (Atlantic coast), located
at 166 km to the ESE across the climate divide (53 090 S/70 540 W),
has been operating since 1890.
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
447
Table 1
Results of the U/Th dating for stalagmites MA1, MA2, and MA3. Errors are quoted as 2-s standard deviations.
Distance from top
d234U
(mm)
(&)
(&)
(mg/g)
(mg/g)
(ng/g)
(ng/g)
(pg/g)
(pg/g)
ka BP
(ka)
MA1
3654
3435
3922
3674
3849
3628
3923
3692
3653
3629
3693
3747
3737
3924
3652
4234
7
21
40
55
75
90
103
113
140
180
205
214
223
232
245
274
16.6
18.7
17.5
19.5
12.8
18.1
19.9
15.5
18.9
18.2
14.6
14.0
13.5
12.9
1.7
12.8
1.5
2.1
2.8
2.3
2.0
1.4
1.5
1.9
1.5
1.4
1.7
1.7
2.4
2.0
1.9
3.1
1.632
1.545
1.753
1.970
1.739
1.693
2.433
2.144
2.993
4.381
4.916
4.980
4.631
3.441
2.952
1.619
0.002
0.002
0.003
0.002
0.002
0.002
0.002
0.002
0.003
0.004
0.005
0.005
0.007
0.003
0.003
0.003
0.73
1.59
6.77
8.54
4.30
14.72
12.96
2.87
2.10
39.94
54.46
79.77
16.10
19.73
1.94
8.39
0.01
0.06
0.05
0.10
0.02
0.05
0.05
0.02
0.01
0.10
0.22
0.40
0.06
0.09
0.01
0.04
0.035
0.149
0.249
0.351
0.376
0.449
0.714
0.657
1.053
2.059
2.530
2.743
2.486
1.945
1.704
1.142
0.002
0.012
0.007
0.017
0.006
0.008
0.008
0.014
0.010
0.013
0.023
0.030
0.023
0.020
0.021
0.017
0.082
0.581
0.824
1.064
1.357
1.562
1.837
2.009
2.348
2.975
3.231
3.367
3.557
3.702
3.859
4.697
0.009
0.051
0.026
0.055
0.022
0.029
0.023
0.043
0.025
0.021
0.032
0.039
0.035
0.041
0.050
0.075
MA2
3879
3880
3437
3828
3798
3630
3918
3815
3796
3826
3834
3919
3835
3920
3631
3827
3921
3882
3438
7
12
17
32
52
82
93
102
119
132
150
159
171
181
192
204
209
243
252
19.3
18.2
18.9
20.5
18.6
18.3
18.9
19.8
18.6
22.1
15.9
14.8
6.6
4.6
2.1
3.4
3.5
0.9
0.5
1.7
1.8
2.1
1.7
1.4
1.6
1.6
1.8
1.6
1.8
1.4
1.5
1.8
1.8
1.7
1.9
1.4
2.5
1.9
1.610
1.579
1.862
1.788
1.887
2.438
2.110
3.353
3.059
3.913
4.566
4.538
2.685
1.544
1.659
1.657
1.688
2.028
1.804
0.002
0.002
0.002
0.002
0.002
0.002
0.002
0.003
0.003
0.004
0.005
0.005
0.003
0.002
0.002
0.002
0.002
0.003
0.002
1.091
2.409
12.145
2.259
2.234
1.057
4.888
1.426
2.002
6.105
3.534
3.571
1.707
2.536
1.910
4.156
0.355
3.094
1.113
0.021
0.043
0.070
0.008
0.010
0.008
0.023
0.009
0.011
0.094
0.014
0.017
0.016
0.015
0.010
0.038
0.004
0.032
0.011
0.040
0.116
0.192
0.205
0.293
0.658
0.617
1.094
1.070
1.518
2.074
2.353
1.564
0.985
1.123
1.170
1.198
1.561
1.516
0.004
0.006
0.010
0.003
0.006
0.017
0.010
0.022
0.017
0.050
0.020
0.029
0.032
0.018
0.020
0.022
0.024
0.056
0.029
0.096
0.411
0.498
0.699
0.978
1.785
1.896
2.179
2.334
2.589
3.050
3.496
3.916
4.268
4.522
4.720
4.800
5.176
5.691
0.015
0.023
0.029
0.013
0.020
0.047
0.033
0.045
0.039
0.088
0.032
0.045
0.082
0.083
0.082
0.092
0.100
0.192
0.112
MA3
3998
3955
3999
4000
3954
4001
3981
3982
3983
3984
3985
3986
3987
6
15
30
46
61
76
90
110
130
150
170
190
207
15.3
19.4
18.2
19.2
13.1
19.3
22.6
19.6
18.2
7.9
4.1
2.4
0.1
1.7
2.2
1.7
1.6
3.9
1.8
1.6
1.4
1.5
1.7
1.9
2.1
2.6
1.567
1.534
1.725
2.679
1.770
3.075
3.632
4.447
4.821
3.281
2.052
1.399
1.593
0.002
0.002
0.002
0.003
0.004
0.003
0.004
0.004
0.005
0.003
0.002
0.001
0.002
0.213
0.227
12.863
1.002
7.347
1.029
1.604
3.691
0.726
27.424
0.430
9.881
3.003
0.004
0.002
0.116
0.008
0.088
0.009
0.010
0.014
0.002
0.376
0.002
0.034
0.015
0.133
0.162
0.410
0.689
0.501
1.034
1.455
1.921
2.299
1.874
1.252
0.926
1.072
0.011
0.006
0.011
0.015
0.016
0.023
0.020
0.018
0.019
0.069
0.018
0.011
0.023
0.518
0.664
1.398
1.700
1.778
2.247
2.704
2.905
3.230
3.662
4.105
4.304
4.487
0.047
0.028
0.039
0.038
0.060
0.051
0.040
0.029
0.028
0.137
0.062
0.055
0.101
Lab. #
Conc.
238
U
Conc.
4. Results
4.1. Regional climate and cave monitoring
Precipitation in the southernmost Andes strongly depends on
local morphology and local wind distribution. Since 1999,
3500e5500 mm of annual precipitation was measured at AWS
Bahamondes (Fig. 1a). The short data set of the Arévalo station at
the MA cave site (since 2007) shows a good correlation with the
Bahamondes station (Fig. 2a). Amounts of 5000e8000 mm were
recorded at AWS Passo at 350 m elevation (Schneider et al., 2003,
and unpublished weather station data until September 2010). At
the Félix weather station, precipitation shows a long-term 30%
decrease from about 5400 to 3800 mm/year between 1914 and
232
Th
Conc.
230
Th
Age
1984 AD. During the same period, the weather station at Evangelistas recorded a similar long-term decrease in precipitation
from approximately 2900 to 2000 mm/year (31%). In general,
precipitation is high throughout the year with a clear maximum
during the summer months between January and March.
Between 2005 and 2008, the MA cave temperature ranged from
0.5 to 11.4 C and the annual mean was 5.3 C (Fig. 2b). Relative
humidity ranged from 71 to 100% with a mean of 94%. There is
a general correlation between temperature and humidity: during
winter (JuneeAugust) the average humidity value decreases to
about 85%, whereas during summer (DecembereFebruary) it
increases to almost 100%.
Compared with deeper caves, the daily and seasonal variations
in humidity and temperature are much more pronounced in the
448
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
Fig. 3. Cross-section of stalagmite MA1 with sampling strategy, showing the 150 mm sampling track for the C and O isotope analyses, the sampling position for the Th/U age
determination (ages in years BP), the drill holes (1.5 mm diameter) for the ICP-MS analyses and the position of thin sections with the tracks of the electron microprobe profiles. A
backscatter electron (BSE) image shows the spots measured by electron microprobe which are oriented along a general track-line avoiding spots on non-carbonate minerals. The
clear layering enables a good correlation between different cross-sections. Darker layers generally indicate a higher amount of organic material.
MA cave due to the wide entrance of this small cave. Its good
ventilation causes temperature and humidity variations to be only
slightly dampened compared with the atmospheric values measured outside at the cave entrance (Fig. 2c). Temperatures below
freezing point were measured only during very few days per year.
The cave temperature shows mildly buffered dayenight variations
as well as seasonal variations. They correlate well with the nearest
AWS Bahamondes, Passo, and Arévalo near to the Gran Campo
Nevado area (Fig. 1a). These stations in turn correlate well with
other stations at this latitude (Schneider et al., 2003), such as the
stations of Evangelistas and Punta Arenas (r 0.9; e.g. Rosenblüth
et al., 1997; Villalba et al., 2003). Punta Arenas, east of the climate
divide, represents a significantly drier climate (w500 mm of
precipitation per year lacking a seasonal pattern).
The drip-water counter collected water from three drip locations a few centimetres apart from each other and therefore
recorded an integrated signal (Fig. 2a). A malfunction of the device
caused a 12-month data gap from October 2004 to September
2005. A comparison of the timing of drip rate maxima and phases of
high precipitation suggests a delay of about 4e6 weeks, caused by
residence in the overlying peat. Cold phases are characterised by
lower humidity and coincide with lower drip rates. At such low drip
rates the evaporation of drip water is relatively high during cold
and less humid phases, which affects the C and O isotope values
(see Section 5.2).
4.2. Mineralogical composition, microfabrics, and detritus
The 28 cm long stalagmite MA1 exhibits portions with regular
laminae (30e100 mm thick), each separated by thin, dark, and
organic-rich layers (Fig. 3). The laminae have a thickness consistent
with average growth rates as determined by U/Th dating and they
are thus of annual origin. This lamination, however, is not continuous along the entire stalagmite. MA2 is 31 cm long and its colouration is weaker due to less organic detritus. MA3 is 26 cm long
and shows the lightest colour of all three stalagmites indicating
a low organic content.
The wall rocks of the cave represent a likely source for siliciclastic detritus. The southern inner walls of the cave (about 50% of
the wall rocks) consist of coarse crystalline granite to granodiorite
with a slight cumulate texture. Idiomorphic K-feldspar up to
0.5 mm (35e45 vol. %), hypidiomorphic plagioclase of intermediate
composition (15e20 vol. %), quartz (20e35 vol. %) and muscovite
(5e10 vol. %) are the major components. Plagioclase shows partial
sericitisation mostly restricted to the crystal cores. Biotite (5e10
vol. %) is locally chloritised. Zircon, apatite and opaques occur as
accessory minerals. Occasional fissures are filled by quartz and
calcite. No epidote was found as a low-grade metamorphic alteration product. The northern inner walls of the cave are dominated
by orthogneiss which consists primarily of quartz (45 vol. %),
plagioclase (25 vol. %, which is partly sericitised) and K-feldspar (10
vol. %). Biotite makes up to 30 vol. % and is rarely chloritised. Minor
apatite, opaques and zircon are present. Again, epidote was not
observed. The granitoids and gneisses only show a slight retrograde
overprint under greenschist metamorphic conditions as evidenced
by mild sericitisation of feldspar and limited chloritisation of biotites. However, this alteration is clearly unrelated to modern
meteoric weathering or soil-forming processes. Obviously, the
rocks that build up the cave’s walls are not representative of
weathering and soil formation processes in the catchment area
from which the drip water is derived. Lithologies in the catchment
area also include a variety of mafic to intermediate metavolcanic
and metasedimentary rock types. Details of soil formation and the
dynamics of surface meteoric weathering processes have not yet
been studied in this superhumid area of the southernmost Andes
and, therefore, could not be taken into consideration with respect
to element leaching and mineral formation in these soils. However,
in this region with very low and generally buffered temperatures
we expect that short-term changes (decades to centuries) are
unlikely to have a significant impact on the rockesoil system
(Migon, 2006).
Detritus was separated from 12 samples from the axial part of
MA1. The detrital fraction has relatively high values of up to 1.3 wt.%
(Hofmann, 2006). Electron microprobe analyses of the detritus of
MA1 show a large variety of siliciclastic components, which can
only be partly derived directly from the wall rocks of the cave.
Quartz, K-feldspar, plagioclase, muscovite, biotite, magnetite,
chlorite, zircon, apatite, monazite (potential source for Th, Y and
rare earth elements), and ilmenite may be derived from granites,
granodiorites and gneisses. Olivine, enstatite, clinopyroxenes, and
chromite may be derived from mafic metavolcanic rocks or dikes.
Ferrosilite and amphibole probably stem from metamorphic rocks.
Goethite and laumontite are related to the soil horizon. This clearly
suggests that other rock types, which do not crop out in the
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
The three stalagmites MA1, MA2, and MA3 were recovered from
the cave in 2004. MA1 was dated by 16 Th/U ages (Fig. 4, Table 1). It
grew during the past 4.9 ka with a rather constant mean growth
rate of about 67 mm/year. Only few sections show lower (down to
25 mm/year) and higher growth rates (up to 100 mm/year).
Stalagmite MA2 was dated by 19 Th/U ages. It grew during the
past 5.7 ka and shows larger fluctuations of the growth rates
compared with MA1 (mean value amounts to approximately 56 mm/
year with a range of 15e110 mm/year). 13 U/Th ages were obtained
from MA3, which grew continuously since 4.7 ka BP with a mean
growth rate of about 66 mm/year (range 10e120 mm/year). None of
the stalagmites shows any significant growth hiatus (Fig. 4).
The locally high detrital 232Th content of MA1 and MA3 results
in a large age correction of the isotope ratios for detrital-rich
samples. We used a 232Th/238U ratio of 4.8 for the detritus (and an
uncertainty of 50%), which resulted in a smooth deptheage relationship that is consistent with the ages from purer samples that
required only a small correction. The 232Th/238U value used here is
within the range of typical literature values for the upper continental crust (e.g. 3.8 1.0, Wedepohl, 1995) and falls into the range
of values measured for the cave’s host rocks. The age correction for
MA2 is much smaller due to the higher degree of purity.
The U content records of the three stalagmites show consistent
and well-correlated patterns through time (r z 0.9), including
a broad U maximum from 4 to 2 ka BP accompanied by several
narrow peaks and minima. They provide a possibility for finetuning the U/Th-based age models (Fig. 5a). For this purpose the
age model of MA2 was used as a reference because its U/Th dating
points required the smallest detrital corrections. The U patterns of
MA1 and MA3 were adjusted in the direction of younger ages by up
4.4. Elements controlled by siliciclastic detritus
Given the geologic setting of the MA cave, relatively insoluble
trace elements (e.g. Y and Heavy Rare Earth Elements (HREE)) are
bound to fine-grained detrital silicate minerals (e.g., pyroxene,
biotite, monazite, and clay minerals). These particles are preferentially transported during periods of intense precipitation, which
result in high drip rates. These minerals are deposited on the flat
tops of the stalagmites where they are incorporated into the
growing speleothem. Therefore, the concentration of these
elements should directly correlate with the amount of siliciclastic
detritus. However, the drip rates have only been sufficiently high at
MA1 to record systematic changes in the deposition of the detritus,
whereas in the other two stalagmites (MA2 and MA3) detritus was
only deposited during the most humid phases at maximum drip
rates. The mean REE concentrations are highest for stalagmite MA1,
whereas the concentrations in MA2 and in portions of MA3 are
close to the detection limit. Especially HREE (here Tb e Lu) show
good internal correlations: correlation coefficients between individual HREEs vary from 0.98 to 1.0 for MA1. The correlations of
these elements are weaker for MA2 (r ¼ 0.74e0.99), and variable
for MA3 (r ¼ 0.1e1.0). In addition, HREE of MA1, MA2 and MA3
correlate very well with Y (r ¼ 0.74e1.0) for each stalagmite. Y, Pb
and Cd are also highly correlated with the REE (r > 0.97).
As described above, Y is used as an indicator for the drip rate
dependent deposition of siliciclastic detritus, which shows characteristic patterns (Fig. 5b) with high concentrations of up to
a
4
3
2
14
5
3
MA3
2
Y content (ppm)
12
MA1
Less humid
period
1
MA2
4
MA1
MA2
MA3
5
b 16
6
age [kyrs B.P.]
6
4.2 kyrs eruption
4.3. Dating
to 200 years until they matched the age model of MA2. The largest
corrections were required in the uppermost detrital-rich section of
the last 2 ka.
U content (ppm)
immediate area of the cave, also contribute to the siliciclastic
detritus and its U/Th signature. Detrital volcanic glass in some
layers of MA1 can be correlated by their major element composition with the two Mt. Burney eruptions in the Late Holocene at
around 2.0 and 4.2 ka BP (Kilian et al., 2003, 2006). Organic
components (i.e. plant residuals, bat hairs, fragments of snail shells)
were also detected in the detritus. Dark brown layers in the axial
section of MA1 (Fig. 3) are rich in organic matter. Gypsum was also
observed in the MA1 thin sections. We suggest that it is derived
from sea spray that enters the cave and partly forms an evaporitic
coating on the cave wall rocks.
449
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
MA1
MA2
MA3
10
8
Less humid
period
6
4
2
1
0
0
0
50
100
150
200
distance from top [mm]
250
Fig. 4. Ageedepth models of the stalagmites MA1, MA2 and MA3, performed by
‘Akima interpolations’ (Akima, 1970). The age model of MA2 is based on the original U/
Th data, whereas the age models of MA1 and MA3 are tuned to the MA2 model (see
Section 4.3). Large negative error bars of MA1 and MA3 are caused by higher uncertainties due to the high detrital 232Th contents.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
age [kyrs B.P.]
Fig. 5. (a) U and (b) Y contents of the stalagmites MA1, MA2 and MA3. The Mt. Burney
eruption in 4.2 ka BP is indicated as well as two phases with relative low Y and
presumably low precipitation. Y records reflect exclusively drip rate changes above
a critical flow threshold, which was generally exceeded during the more humid phases
in the MA1 record. A millennium-scale U-enrichment from 4.1 to 1.5 ka BP is partly
related to a long-term acidification event after a sulphur-bearing tephra deposition.
450
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0
Y (MA1)
a
5
4
10
U (MA1)
15
b
2
1
1200
Mg [ppm]
1000
3
U [ppm]
Y [ppm]
5
c
800
Mg (MA1)
600
400
Sr (MA1)
d
Less humid
period
Sr (MA2)
e
350
200
Sr (MA3)
300
150
250
Sr [ppm]
Sr [ppm]
200
Very humid
period
300
250
250
Sr [ppm]
300
200
f
200
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
age [kyrs B.P.]
Fig. 6. (a) Y content of MA1 as indicator for siliciclastic detritus. Light grey bars from 3.5 to 2.5 and from 0.6 to 0.1 ka BP indicate phases when drip rates were below a critical
threshold value, (b) U content of MA1, (c) Mg content of MA1, (d) Sr content of MA1, (e) Sr content of MA2, and (f) Sr content of MA3 represent water-soluble and carbonatecompatible elements, which are influenced by rain dilution and the degree of soil acidification. Correlated peaks and patterns are marked by dark grey bars.
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
a Mg/Ca MA1
7000
6000
0.4
5000
0.5
calibrated section
100 x Mg/Ca
0.3
0.6
0.7
4000
3000
calibrated
precipitation [mm/yr]
0.2
451
2000
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
age [kyrs B.P.]
b
0.6
0.8
1.0
4000
3000
0.4
0.6
d
8000
2000
1000
0.2
e
7000
1.0
6000
5000
4000
U/Th
0.8
precipitation
[mm/yr]
precipitation [mm/yr]
c
3000
100 x Mg/Ca
0.4
100 x Mg/Ca
0.2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990
year (AD)
Fig. 7. (a) Mg/Ca record of MA1 measured by electron microprobe with a sub-annual resolution for the last 5 ka with a 40-point running mean. Calibration of the uppermost youngest
Mg/Ca records of MA1 (b) and MA2 (c) with respect to annual precipitation averages from the weather stations from Evangelistas (d) and Félix (e). The similar patterns allow a calibration
of the Mg/Ca ratios with respect to precipitation data from Félix. The U/Th dating point at 1922 AD for MA1 is given. For further details on the age constraints see Section 4.5.2.
16 ppm in MA1, 4.5 ppm in MA2 and 6 ppm in MA3. The Y record of
MA1 starts at 4.1 ka BP and is characterised by two periods with
high Y contents: between 3.8 and 3.5 ka BP (up to 7 ppm) and from
2.4 to 0.6 ka BP (3e12 ppm, showing seven cycles of about 200e250
years). From 3.5 to 2.5 ka BP and from 0.6 to 0.1 ka BP, the Y contents
are very low and near the detection limit. The Y record of MA2
(starting at 4.75 ka BP) shows a major peak from 1.0 to 0.7 ka BP
(synchronously to very high Y values in MA1) and smaller peaks at
4.6 ka BP and during the last century. Apart from these peaks, the
record shows low Y contents (<1 ppm), which are often near to or
below the detection limit.
The Y record of MA3 starts at 4.35 ka BP and exhibits only one
major peak between 4.2 and 3.8 ka BP. The rest of the record shows
only Y contents below 1.5 ppm.
4.5. Carbonate-compatible elements
4.5.1. Low resolution data
Several carbonate-compatible elements were measured in
stalagmites MA1eMA3, including U, Sr, and Mg with a temporal
resolution of about 40e50 years (Figs. 5a and 6). These elements
show a similar broad long-term maximum between about 4 and
2 ka BP. High U contents during this time interval can be related
to acidification due to a regional tephra deposition (Kilian et al.,
2006) and/or by a millennium-scale climate perturbation. The
highest element concentrations occur in the period between 3.5
and 2.5 ka BP, for which very low Y contents suggest low input of
detritus and low drip rates and, thus, a less humid climate (see
Section 5.2).
U contents of the analysed stalagmites range from 1.2 to 6.2 ppm
(Fig. 5a). Besides comparable long-term trends, all three stalagmites show correlated negative and positive short-term anomalies
at 0.6, 1.65, 2.05, 2.45, 2.75, 2.9, and 3.75 ka BP, which were used to
adjust the age models of MA1 and MA3 to the one of MA2, as
described above.
Similar correlated anomalies are observed in the Sr and Mg
records. Fig. 6 shows the Sr variations for all three stalagmites
together with the records for Mg, Y and U in MA1. In all records, the
increase of the millennium-scale acidification trend (Kilian et al.,
2006) started at about 4.0 ka BP, but the highest values of
different elements were reached at different times: U peaks at
3.2 ka BP, Mg at 2.75 ka BP (disregarding the peak at 3.6 ka BP), and
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
4.6. Stable isotopes
d13C and d18O values measured along the growth axes are
plotted inversely in Fig. 8 reflecting the interpretation that higher
precipitation levels caused less fractionated isotope values (see
Section 5.2). The average resolution is 2.5 years for MA1, 3.3 years
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
a
MA1
-6
-5
-12
-10
-3
-8
-6
-4
-7
13
-14
-4
δ C [ ‰ VPDB]
-16
b
MA2
-6
-5
-14
-4
-12
-10
-8
-6
13
18
-16
δ C [ ‰ VPDB]
18
δ O [ ‰ VPDB]
-7
δ O [ ‰ VPDB]
-4
c
MA3
-6
-5
-16
-14
-4
-12
-10
-3
-8
-6
13
-7
δ C [ ‰ VPDB]
4.5.2. High-resolution data
The Mg/Ca record of stalagmite MA1 is shown in Fig. 7. The
(100) Mg/Ca ratios vary between 0.02 and 2.00 (mean 0.37).
Short-term variations measured at a sub-annual resolution (2e3
data points per year) represent a few months and show larger
variability than the long-term variations. Mg/Ca ratios are possibly
overprinted by a long-term acidification effect (see Section 5.3;
Tipping et al., 2003). In general, however, low Mg/Ca values suggest
high precipitation rates caused predominantly by higher dilution of
Mg compared with Ca, and vice versa (McMillan et al., 2005). The
prominent maximum in concentrations of carbonate-compatible
elements can be also observed in the Mg/Ca ratios, which show
long-term elevated values (from 0.25 to 0.55) between 4 and 2 ka
BP. Sharp peaks with elevated Mg/Ca ratios are superimposed on
this long-term trend at 4.25, 3.4, and 3.3 ka BP, probably reflecting
phases of decreased precipitation. The correlation coefficient
between d18O and Mg/Ca ratios is 0.34, and 0.24 between d13C and
Mg/Ca.
The calibration of the Mg/Ca records from MA1 and MA2 with
respect to modern precipitation data measured at the weather
stations Evangelistas and Félix is illustrated in Fig. 7bee. The MA1
age model relies on the age of the youngest U/Th-dated subsample
(1922 8 AD) and on the assumption that the stalagmite was still
growing when it was sampled. However, the most recent end of the
Mg/Ca records in the thin sections is somewhat older since some
tenth of millimetres crumbled away from the topmost layers during
preparation of the thin sections. As a result, the Mg/Ca record of
MA1 was inferred to end at about 1984 AD and of MA2 at 1990 AD.
Based on the high-resolution U/Th age model of the two Mg/Ca
records for the last 100 years the observed laminae can be interpreted as reflecting annual variations. However, the ages were
slightly tuned (up to 5 years) to improve the correlation with the
two precipitation records. This tuning is considered reasonable
since we estimate a rapid transmissivity of the geochemical signal
from the overlying soilerock interface.
The comparison of Mg/Ca ratios with the precipitation records
from Félix and Evangelistas encompasses 71 and 85 years, respectively. For Evangelistas data are only available with sufficient
quality until 1984. Precipitation at Evangelistas and Félix decreased
between 1914 and 1984 by 40% and 25%, respectively, whereas the
temperatures did not change significantly. Since the Mg/Ca ratios of
MA1 and MA2 increased during the calibration periods by about
70% and 26%, respectively, higher Mg/Ca ratios indicate lower
precipitation, and vice versa.
Félix shows about 20% less precipitation than the AWS Bahia
Arévalo at the cave site, and precipitation at Evangelistas is about
70% lower than at the MA cave. This confirms the same tendency for
all three weather station sites. To calculate palaeo-precipitation at
the location of the MA cave for the last 5 ka, the amplitudes of the
weather station record of Félix were compared with the Mg/Ca
records of MA1 and MA2. Due to the orographic increase of
precipitation from westeeast towards the Andes, we assumed 20%
higher precipitation at the MA cave compared with the Felix
weather station (Fig. 1a). The calibration (Fig. 7bee) indicates
reasonable precipitation values between 4200 and 6500 mm/year
during the last five millennia. On a sub-annual scale, however, the
variability of the Mg/Ca ratios is much higher, indicating periods of
weeks to a few months with extraordinary high and low precipitation rates.
18
Sr between 2.5 and 3.0 ka BP. This could reflect distinct elementspecific response times to the pH-dependent leaching process. At
least since about 1 ka BP, the element concentrations returned to
values similar to those prior to the long-term leaching process
before the 4.2 ka BP tephra deposition. U, Mg, and Sr concentrations
were highest during the period between 3.5 and 2.5 ka BP, which
was low in detrital deposition and, thus, less humid. This phase is
highlighted in light grey in Fig. 6. The overall very humid period
from 2.0 to 0.8 ka BP is marked by high and strongly fluctuating Y
contents and lower contents in soluble elements (U, Sr). Prominent
excursions from the overall variation with strong peaks in the
concentration of these elements occur throughout the records and
are correlated with darker laminae. Examples are the short-term
minima at 0.6, 1.5, 2.05, and 2.9 ka BP. Correlations are highest for
U/Sr (r ¼ 0.86) and weaker for Sr/Mg (0.70) and U/Mg (0.51).
δ O [ ‰ VPDB]
452
-4
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
age [kyrs B.P.]
Fig. 8. d13C and d18O values of the stalagmites MA1 (a), MA2 (b) and MA3 (c). The d18O
records are in black, d13C in grey. The obvious correlation between the d13C and d18O
records of each stalagmite indicates kinetic fractionation (Rayleigh fractionation)
during the precipitation of calcite. However, there is no correlation of the d13C or d18O
records between the individual stalagmites.
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
for MA2, and 3.1 years for MA3. The comparisons between the d13C
and d18O records for each stalagmite show a high degree of
covariance (r2 ¼ 0.65 for MA1, 0.76 for MA2 and 0.69 for MA3).
Therefore, the two isotope proxies are well correlated and will be
discussed together below.
The C and O isotope records of MA1 start at 4.3 ka BP. d18O values
range between 4.6 and 7.2& and the values of d13C vary between
10 and 16& (excluding a few outliers). The MA2 record starts at
5.6 ka BP and shows slightly larger variations in stable isotopes than
MA1 (Fig. 10b): d18O values vary between 3.6 and 6.8& and d13C
values between 5.5 and 15&. The MA3 record starts at 4.6 ka BP
and shows the largest range: d18O values vary between 3.1 and
7&, d13C values between 4 and 15.5&. MA1 shows by far the
smallest amplitudes, whereas the records of MA2 and MA3 show
larger amplitudes. However, it is striking that there is no good
correlation between the isotopic patterns of the three stalagmites.
‘Hendy tests’ (Hendy, 1971) were performed on stalagmites MA1
and MA2. The data of each of these tests show very high correlation
coefficients between d13C and d18O (r2 ¼ 0.85e0.97 for MA1 and
0.87e0.92 for MA2). Fig. 9 shows ‘Hendy test’ data for both
stalagmites with clear trends towards higher values with increasing
distance from the stalagmite axis. Different slopes can be noticed at
each ‘Hendy test’ curve. Fig. 10 displays the relationship between
d13C and d18O values measured along the extension axes of the
stalagmites. The resulting positive correlation strongly suggests
isotope disequilibrium during the precipitation of calcite (see
Section 5.2).
453
a better comprehension of the climate significance of the proxies.
Many important palaeo-climatic reconstructions are based on
single stalagmite records and only few studies have presented
precisely dated proxies of more than one coeval stalagmite from
a given cave (Linge et al., 2001; Proctor et al., 2002; Frisia et al.,
2003; Williams et al., 2004; Cobb et al., 2007; Denniston et al.,
2007; Boch et al., 2009). Therefore, still little is known about the
cave-specific and climate-independent variability of stalagmite
proxies despite the results of a few years of cave monitoring that
indicates an influence of cave-specific processes on both stable
isotope values and growth rates (Mickler et al., 2006; Baldini et al.,
2008; Kluge et al., 2010).
In order to better characterise and interpret proxy records and
their climate relationship, we document several proxies for three
coeval stalagmites from the MA cave. These proxies include: (1)
trace elements related to siliciclastic detritus (e.g. Y, HREE) which
depend especially on drip rates. (2) Carbonate-compatible trace
elements (e.g. U, Mg, and Sr) incorporated into the calcite that
document changes in precipitation, basement alteration, soil
development, soil pH, and temperature. (3) C and O isotopes
reflecting changes in drip-water source (e.g. hydrological cycle),
temperature, drip rate dependent CO2 outgassing, and evaporation.
Given the fact that the three stalagmites have grown side by side in
this small cave we assume that there is a common drip-water
source.
5.1. Y and HREE as indicators for detrital influx
A thorough understanding of the processes controlling the
chemical and isotopic composition of stalagmites is crucial for
-2
b
-6
δ C range
13
-8
13
-10
-12
-12
30
40
50
60
70
80
90
0
d
18
VPDB]
δ O range
δ O[
-5
18
-6
-7
10
20
30
40
50
60
70
80
90
-3
-4
18
20
δ O range
10
-3
-4
VPDB]
-10
-16
0
18
-8
-14
-16
δ O[
13
-6
VPDB]
-4
-14
c
-2
-4
δ C[
13
δ C[
VPDB]
a
Stalagmites in karst caves commonly contain very small
amounts of siliciclastic detritus, unless they are contaminated by
sediment during flooding events (Niggemann et al., 2003).
However, significant detritus may occur in stalagmites, especially if
δ C range
5. Discussion
-5
-6
-7
0
10
20
30
40
50
60
70
80
Distance from central axis [mm]
90
0
10
20
30
40
50
60
70
80
Distance from central axis [mm]
90
Fig. 9. ‘Hendy tests’ of stalagmites MA1 (a, c) and MA2 (b, d). The distinct enrichment along different laminae leads to different slopes. Arrows with labels ‘d13C range’ and ‘d18O
range’ show minimum and maximum values of the isotope values measured along the growth axes (see Fig. 10).
454
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
Fig. 10. d13C vs. d18O values of stalagmites MA1 (a), MA2 (b) and MA3 (c) measured along the growth axes. The covarying values suggest disequilibrium conditions during the
precipitation of calcite. The fractionation between d13C and d18O is less pronounced for stalagmite MA1.
they were formed in non-carbonate host rocks such as the MA cave
(Hill and Forti, 1997; Fairchild et al., 2006). In such caves the
amount of detritus in the stalagmite can be constrained by chemical
tracers which integrate the amount of clay- and silt-sized mineral
particles in the deposit.
The abundance of detrital particles in the drip water depends on
(i) the characteristics of the meteoric weathering process, (ii) the
network of fissures in the aquifer, and (iii) the rate of water flow.
Our cave monitoring suggests that there is a positive relationship
between drip rate and the amount of transported/deposited siliciclastic detritus once a certain threshold of drip-water flow has been
exceeded. This proxy does not depend on temperature or humidity
in the cave and is therefore discussed first.
The HREE and Y contents of the MA stalagmites are highly
correlated due to their similar ionic radii and similar charge, which
enable these cations to be substituted in detrital silicate minerals
(e.g. chlorite, biotite, zircon, hornblende, and monazite). Since
these elements are also insoluble in seepage water, their abundance
can be used to trace the insoluble detritus (Fig. 11) if the drip rate
range is large enough to transport significantly large particles.
Above a critical level, drip rates correlate positively with the
amount of the transported detritus. Below this level detrital
particles are not transported (e.g. during the periods of 3.5e2.5 ka
BP and 0.6e0.1 ka BP), which results in Y values close to the
detection limit. MA2 and MA3 exhibit long growth intervals with
very low Y values (Fig. 5b), indicating that the drip rates were not
high enough at these sites to transport significant amounts of
detritus.
Electron microprobe analyses indicate that light-coloured
laminae (high precipitation) are enriched in detritus (high Y
contents), whereas darker and/or brownish laminae are poor in
detritus, formed at low precipitation and are accompanied by
higher contents of organic carbon (Fig. 3). The higher sensitivity of
the MA1 Y record is related to its higher and more variable drip
rates (see discussion about C and O isotopes in Section 5.2, Zhou
et al., 2008), whereas the lower drip rates for MA2 and MA3 yielded significantly lower detrital components, especially in MA2.
Only between 1.0 and 0.7 ka BP, when MA1 shows the highest Y
concentrations indicating the highest drip rates, the threshold level
for detritus transport was surpassed for MA2 and MA3.
Y peaks of MA1 correlate with negative d18O anomalies and low
Mg/Ca ratios (Fig. 11), in particular at 1.0, 1.3, 1.5, 1.8, 2.0, 2.25, 3.0,
3.5, and 3.75 ka BP. This confirms our interpretation that the Y
record of MA1 is controlled primarily by rainfall and associated drip
rate changes. All stalagmites show very low Y values between 2.5
and 3.5 ka BP, which are consistent with low precipitation indicated
by high d18O values and high Mg/Ca ratios of MA1 at these times.
5.2. Isotope records and ‘Hendy tests’
Strong isotope variations in meteoric waters are associated with
changes in global ice volume, and locally in the amount of rainfall
and air temperature (Lachniet, 2009). In the case of the MA cave,
altitude-related and orographic effects on the d18O isotope
composition are insignificant due to its coastal position at low
elevation and the prevailing westerly winds. However, centennial
to millennium-scale changes in the rainwater isotopic composition
in the Southern Andes are still unexplored. The short transit time of
the seepage water through the soil and bedrock (4e6 weeks,
Fig. 2a) enables climate events to be captured locally at high
temporal resolution.
Due to the small size of the cave and its large entrance, we
assume that the pCO2 values in the cave are similar to the atmospheric pCO2, which was 280 20 ppm during the pre-industrial
Holocene (Indermühle et al., 1999). The pCO2 gradient between the
drip water and the cave air causes rapid degassing of CO2, especially
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
455
Fig. 11. Comparison of precipitation-controlled proxies of stalagmite MA1. Dark grey shaded background areas mark common precipitation maxima, especially during the Medieval
Warm Period (MWP). The less humid phases from 3.5 to 2.5 ka BP and from 0.7 to 0.1 ka BP during the Little Ice Age (LIA) are illustrated by light grey bars. Local precipitation
maxima (minima) are marked by areas of dark (light) grey. (a) Sr/U ratios are plotted as an indicator for local sea spray deposition. The possible effect of this sea spray buffering of
soil water during the LIA and resulting lower Mg/Ca ratios is indicated in (c). (b) Mg/Ca ratios smoothed by a 40-point running mean with a precipitation axis, which results from the
calibration of the Mg/Ca values (see Fig. 7; dashed curve represents the correction of possible buffering effects in the soil). (c) d18O of biogenic carbonate in a lake of Torres del Paine
at 51 S for the MWP and LIA (Moy et al., 2008), reflecting the relationship between precipitation (lighter O isotopes) and wind-induced evaporation (heavier O isotopes). (d)
Temperatures based on modelling of C and O isotopes of MA1 are shown (Mühlinghaus et al., 2008) together with the possible effect of less rainwater during the cold LIA (dashed
curve). (e) d18O isotope values smoothed by a 40-point running mean (the dashed line indicates possible lighter O isotopes of meteoric water). (f) Yttrium content. Grey horizontal
bars indicate duration of Neoglacial glacier advances in the southernmost Andes (Mercer, 1982; Koch and Kilian, 2005).
at low drip rates during winter and/or colder climate phases,
producing a strong enrichment especially in 13C and a covariance
between d13C and d18O values. This kinetic fractionation known as
‘Hendy effect’ (Hendy, 1971; Wiedner et al., 2008) is clearly illustrated by all MA stalagmites, which show a high degree of covariance between C and O isotope values along the drip centre of the
stalagmite (r2 > 0.65; Figs. 8 and 10) and an increase along individual growth layers away from it (r2 z 0.9; Fig. 9). The steeper the
slope of the ‘Hendy test’ curves, the longer is the time interval
between two consecutive drips (Fig. 9). Additionally, the range of all
‘Hendy test’ isotope values measured at the drip centre is within
the range of the d13C and d18O values measured along the growth
axes (arrows in Fig. 9), hence both data sets are consistent. Therefore, we conclude that none of the stalagmites of the MA cave
formed under isotopic equilibrium. Given the good ventilation of
the cave, evaporation probably also contributes to this isotopic
covariance by increasing the d18O values during periods of low drip
rates and/or low humidity. Quantifying the role of evaporation,
however, was not possible, partly because prior calcite precipitation (see below) complicated the overall picture.
Intervals of large isotopic variations coincide with smaller
growth rates in MA2 and MA3. Mühlinghaus et al. (2008) modelled
isotope fractionation in the MA stalagmites and concluded that the
drip intervals for stalagmite MA2 were longer than those for MA1.
Therefore, we assume that the drip intervals for MA3 were also
longer than for MA1, which is also consistent with our cave
monitoring.
Mühlinghaus et al. (2008) calculated temperatures for the MA1
and MA2 record based on the difference in the temperaturedependent O and C isotopic fractionation and their response to
kinetic effects (Fig. 11). The temperatures of MA1 and MA2 show
a variability of 3e4 and 4e5 C, respectively. Although both
stalagmites show quite distinct patterns on C and O isotopes (Fig. 8)
the calculated temperatures are highly correlated (r2 ¼ 0.76).
456
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
However, the amplitude of Late Holocene temperature variations in
Patagonia was probably smaller than 2 C (Lamy et al., 2004;
Mohtadi et al., 2007; Sepulveda et al., 2009). Therefore, the calculated temperature data should be viewed cautiously, as they were
based on the assumption of a constant d18O value for the meteoric
water and drip water. For example, during the interval between 4.2
and 2.5 ka BP, stalagmite MA2 shows significantly higher d18O and
d13C values than the coeval stalagmite MA1 (Fig. 8). This intersample heterogeneity cannot be explained by evaporation, for
example, as humidity changes would equally affect all stalagmites
within this small cave. We suggest that these isotopic differences
reflect the sensitivity of individual drip sites to prior calcite
precipitation (PCP) at the cave roof and stalactite, i.e. prior to
reaching the stalagmite’s top. Stalagmite MA2 formed at a drip
location that apparently experienced extensive PCP during the
interval 4.2e2.5 ka BP resulting in elevated stable isotope values
compared to the neighbouring MA1 site. Assuming an average
slope of the ‘Hendy test’ curves of 0.25 the higher d13C values in
MA2 suggest that at least 1& of the d18O increase in MA2 is due to
the PCP effect. Subtracting 1& from the measured d18O value
(4&) over the interval of slow growth of MA2 yields a value closer
to that of MA1 (5&) for the same growth interval, as would be
expected for two stalagmites that were fed by the same drip water.
In this case, a common source of drip water for both stalagmites is
indicated and a significant part of the variability of the isotope
signal during periods of slower growth of MA2 is attributed to local
PCP effects. Consequently, the temperature reconstruction by
Mühlinghaus et al. (2008) for MA1 is regarded to be more reliable
than that for MA2.
The isotopic records of MA1 show a good correlation with the
variations of the Y concentrations, which indicate variations of the
drip rate (see above, Fig. 11). Therefore, we assume that the stable
isotope data of MA1 represent primarily temperature and drip rate
controlled variability, which makes this record highly suitable for
comparison with precipitation records from other archives in the
region.
Modelling shows that during the slower growth phases of MA2
(between 4.2 and 2.5 ka BP) the drip interval was about 1000 s,
significantly longer than in the fast-growing section (w300 s) of the
stalagmite. For a relative humidity of 70% (a typical value presently
measured in the MA cave), we estimate the buffering effect on the
isotopic composition of the precipitated calcite at the cave roof and
the evaporation effect to be 1& for 13C and 0.8& for 18O. These
values are based on a model that determines the stable isotope
composition of the drip water during the precipitation of calcite at
the cave roof (Deininger, 2010) in analogy to the model applied for
the formation of stalagmites by Mühlinghaus et al. (2009). For the
MA2 stalagmite, the model of Deininger (2010) shows that the
increase in the drip interval from 300 s to 1000 s raises the d18O and
d13C by 1.2& and 0.2&, respectively.
In general, low drip rates result in a longer residence time of the
solution on the stalactite and subsequently on the stalagmite,
enhancing the exchange with the cave atmosphere. The exchange
time constant for CO2 at a temperature of 6 C and a thickness of the
diffusive layer of the solution on the stalagmite of 0.1 mm amounts
to approximately 10 s. However, as already discussed by Salomons
and Mook (1986) and Mook (2000), the exchange only becomes
relevant when the solution has a low pCO2 compared to the cave
atmosphere. During MA2’s periods of slower growth the pCO2 of
the drip solution may come close to the atmospheric value and the
enrichment of 13C would amount to several per mill. This could
explain the difference in the isotope values between MA1 and MA2,
suggesting that periods of slower stalagmite growth tend to
enhance the amplitude of the isotope values. This example
emphasises the potential discrepancies observed between different
records from a single cave (e.g. Williams et al., 2004) with high
variability and only limited common signals on millennial time
scales.
In conclusion, if isotope variations in slowly growing stalagmite
sections are compared with sections with higher growth rates, the
response of the signals to the different effects needs to be
considered.
In order to verify the interpretation that the isotopic composition of MA1 is sensitive to precipitation and drip rate, we will now
compare these proxies with another precipitation record of MA1,
the Mg/Ca record (Section 5.3).
5.3. Carbonate-compatible elements: U, Sr, and Mg
Dilution by rainwater and pH-dependent leaching of elements
from overlying rocks and soil controls the abundance of watersoluble trace elements in the stalagmites in the MA cave. Maxima
and minima in the records are mainly caused by high- and lowprecipitation rates, respectively. All three stalagmites show very
similar, correlated trends for soluble elements (for U: r > 0.9;
Fig. 5a) with characteristic anomalies. Their concentrations were
primarily controlled by variations in the drip-water composition
because the detritus does not contain significant amounts of U or
other soluble elements. Y as well as Zr and Hf contents, however,
are anticorrelated with concentrations of U, Sr, and Mg, which is in
support of our interpretation (Fig. 6). However, the pronounced
effect of the sulphur-rich tephra layer from the Mt. Burney eruption
in 4.2 ka BP caused an additional long-term acidification in the pore
water (pH 3e5) of the soil above the cave (Kilian et al., 2006). This
acidification effect caused enhanced leaching of water-soluble
elements from the soils and is documented in all three stalagmites
by a broad peak with high U, Mg, and Sr contents over more than 2
millennia following this volcanic event (Fig. 6; Section 4.5). During
the period between 3.5 and 2.5 ka BP, for which the Y contents
suggest low drip rates and reduced precipitation, the amplitude of
this peak is most pronounced.
The Sr and Mg records of MA1eMA3 and the Mg record (MA1)
are moderately well correlated with each other and also with U.
Occasionally, there are large detrital mineral grains within the
1.5 mm small sample cores, which may cause deviations from these
trends. Sr patterns partly follow U, especially in MA1. Individual Sr
peaks and deviations from the correlated trends may have been
caused by feldspar and/or clay mineral particles and/or sea spray
from storm events. Similarly, individual Mg peaks can be explained
by Mg-rich and U-poor detrital particles (e.g. chlorite).
Pronounced negative anomalies of the U and Sr contents in MA3
and e less pronounced in MA2 e at around 1.0 ka BP are coeval to
very high Y concentrations in the MA1 record. This indicates
another short, very humid phase with increased detritus and U and
Sr dilution (Fig. 6). A particular peak of high U concentration, which
occurs only in MA3 at around 1.0 ka BP and thus during times of
high drip rates, may be related to accidental deposition of zircon
grains.
The high-resolution Mg/Ca ratios in Fig. 7 only reflect Mg and Ca
variations of the calcite because only clean, detritus-free areas were
analysed by electron microprobe. Mg/Ca ratios of speleothem
calcite thus probably reflect changes in temperature, abundance of
sea spray, and dilution effects (Katz et al., 1972; Mucci, 1987;
Oomori et al., 1987; Burton and Walter, 1991; Bar-Matthews et al.,
1999; Huang and Fairchild, 2001; Tooth and Fairchild, 2003;
McDonald et al., 2004; Musgrove and Banner, 2004; McMillan
et al., 2005). Y concentrations of MA1 partly correlate with the
Mg/Ca ratio, indicating that both proxies are mainly controlled by
drip rate where the dilution effect reduces Mg/Ca. However, during
phases of strong sea spray deposition and/or low precipitation the
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
pH values of the soil water decreases (local monitoring and e.g.
Tipping et al., 2003). Such conditions increase the Mg leaching from
basement rocks above the cave and would also increase the Mg/Ca
ratios. Therefore, Mg/Ca variations may reflect a combination of
both: precipitation rates and sea spray effects.
At AWS Bahamondes, the monthly amount of precipitation
varies between 80 and 150 mm during winter (JulyeAugust) and
between 380 and 450 mm during summer (DecembereFebruary;
Schneider et al., 2003). Furthermore, drip rates in the cave are 2- to
3-fold higher in summer than in winter. Intra-seasonal variations
within a year are therefore higher than inter-seasonal variations
between different years. A 40-point running mean of the Mg/Ca
ratios is shown in Fig. 7 and integrates these values over 10e15
years. The averaged values show a good correlation between Mg/Ca
and d18O in MA1 (r ¼ 0.71, and r ¼ 0.61 for d13C, not shown),
especially on a centennial and millennial time scale. Noticeable is
the period between 3.5 and 2.5 ka BP, where both proxies suggest
a precipitation minimum. Low Mg/Ca ratios are correlated with
high Y values and vice versa (where the Y record shows high
values). In particular, during the highest inferred drip rates at
around 1 ka BP the Mg/Ca ratios are lowest. This confirms that the
Mg/Ca ratio values are primarily controlled by precipitation. This
interpretation is also supported by the correlation between present
drip rates and modern precipitation data from the weather station
at Evangelistas (Schneider et al., 2003; Fig. 2a).
However, there is a clear discrepancy between the Mg/Ca ratios,
the d18O and Y contents when comparing the Medieval Warm
Period (MWP) with the Little Ice Age (LIA; Fig. 11): Y contents
indicate relatively low precipitation (below the threshold) during
the LIA, whereas the Mg/Ca ratios are nearly at the same low level
as during the MWP. d18O values are lowest during the very humid
MWP, but only slightly higher during the LIA. This can only be
explained if rainwater isotope compositions were significantly
different through time. Sea surface temperatures derived from
marine sediment cores from the Southern Andes continental
margin (Mohtadi et al., 2007; Sepulveda et al., 2009) and from the
fjord region at 50 S (F. Lamy, personal communication) suggest
a distinct cooling event of about 1e2 C during the LIA (0.7e0.1 ka
BP). This cooling of the South Pacific should indeed have produced
lighter meteoric O isotope values suppressing the increase in d18O
values even for the significantly less humid LIA compared with the
MWP (Fig. 11). d18O values of biogenic carbonate from lacustrine
sediments at Torres del Paine near 51 S (Moy et al., 2008) reflect the
relationship between precipitation (lighter O isotopes) and windinduced evaporation (heavier O isotopes). These data indicate less
precipitation and/or more evaporation during the LIA after 0.6 ka
BP (which means less westerly influence on the lee-side of the
Andes) and more precipitation and/or less evaporation during the
MWP consistent with other proxies. Even during similar westerly
intensities, the LIA temperature decrease of approximately 2 C
would have caused about 15e20% less precipitation, which,
however, is not documented in the Mg/Ca record. One explanation
is an increased buffering of the soil pH (causing less Mg leaching) by
increased sea spray around the cave. This is suggested by increasing
Sr/U ratios during the last 5 ka, culminating during the LIA (Fig. 11).
The Ca derived from sea spray buffered the acidic soil water, so that
Mg leaching from soil and basement rocks decreased, leading to
lower Mg/Ca values. The increased sea spray influence can be
explained by higher salinities in the fjord system around the cave
due to (1) less precipitation (Kilian et al., 2007), especially during
the LIA, and (2) glacio-isostatic depression of the coastline during
the Neoglacial of the last 5 ka as documented in a fjord sediment
core drilled near the cave. This would have led to saltier fjord
waters and increased water surface area as a source for increased
sea spray.
457
On account of the possible influences (especially by sea spray)
on the Mg/Ca record, the proposed correlation between Mg/Ca and
precipitation presented above (Fig. 11) must be considered with
care, in particular, since the calibration is based only on a period of
80 years. Nevertheless, it is the first time that a calibrated palaeoprecipitation proxy from this region has been presented.
5.4. Comparison with regional climate records
The previous evaluation and calibration of the investigated
proxies indicate that they provide useful information on temperature and precipitation for the last 5 ka. However, we discuss this
only briefly below since this is not the main focus of our research.
At the MA cave site, precipitation is strongly related to average
wind velocities and SHW intensities (Schneider et al., 2003).
Furthermore, NCEP-NCAR data of the last 40 years indicate that the
MA cave is situated in the core section of the SHW (53 S).
Examination of the effect of drip rate and temperature on the
d18O and d13C values by Mühlinghaus et al. (2008, 2009) suggests
that the period between 3.5 and 2.5 ka BP was the coldest phase of
the Neoglacial, even colder than the relatively cold period between
0.7 and 0.1 ka BP (Fig. 11). Peat and pollen records were interpreted
to document a cold and dry Neoglacial period from 2.8 to 2.7 ka BP
(Van Geel et al., 2000; Chambers et al., 2007). For this period, the
MA1 record shows the highest d18O and d13C values as well as the
highest Mg/Ca ratios. Such Neoglacial cold phases (3.5e2.5 ka BP)
have also been detected in several other records worldwide (Bond
et al., 2001; Wanner et al., 2008) and were probably triggered by
relatively low solar activity which led to a decrease in temperature
in the northern Hemisphere, particularly at mid-high latitudes. This
could also have caused a weakening and/or shifting of the SHW.
However, Lamy et al. (2004), Mohtadi et al. (2007) and Sepulveda
et al. (2009) argued that the strongest Neoglacial temperature
decrease of about 2 C occurred during the LIA (0.7e0.1 ka BP),
which led to significant glacier advances (Mercer, 1982; Koch and
Kilian, 2005). Although the very low Y values suggest relatively
low precipitation for the LIA, the pronounced Holocene cooling can
explain a positive glacier mass balance for the LIA advances even
during times of relatively low precipitation (Neukom et al., 2010).
We argue that the Y record of MA1 was predominantly precipitation-controlled, whereas the d18O record and the Mg/Ca ratios
also reflect temperature changes and other factors. The reconstruction of palaeo-precipitation based on the calibration of Mg/Ca
ratios (Section 5.3 and Fig. 7) shows that precipitation was also
lowest during the first Neoglacial cold phase from 3.5 to 2.5 ka BP.
However, reduced precipitation during the LIA is only evident in the
Y record, but not in the Mg/Ca record. Lower temperatures in the
soil-weathering horizon could have hampered Mg leaching which
led to the lower Mg/Ca ratios despite low precipitation. The highest
precipitation (up to 6500 mm/year) documented by high Y and low
Mg/Ca values occurred from 4.2 to 3.8 ka BP and from 2.0 to 0.8 ka
BP with very high values during the Northern Hemispheric Medieval Warm Period between 1.2 and 0.8 ka BP (e.g. Crowley and
North, 1991).
6. Conclusions
We provide the first ultra-high resolution (up to sub-annual)
palaeo-environmental records based on stalagmites from southern
South America. The three stalagmites (MA1, MA2, and MA3) have
been recovered from the same non-karst cave located on the Pacific
coast of southernmost Chile in the centre of the Southern Hemispheric Westerlies (SHW). The stalagmites grew simultaneously
and close to each other since about 5.7 ka BP, each one dated by
more than a dozen U/Th ages.
458
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
Comparisons of different palaeo-climatic proxies in three
stalagmites from a single cave provide important insights into the
site-specific (within the cave) variability of climate proxies.
Due to distinct pathways of the seepage water each speleothem
represents a different drip rate regime. Our reconstructions show
that the overall drip rates for stalagmite MA1 were significantly
higher than for MA2 and MA3.
Chemical tracers of siliciclastic detritus (e.g. Y and HREE) in nonkarst caves represent new and hitherto unexplored proxies of
palaeo-precipitation if the drip rate exceeds a certain threshold
level. In the case of stalagmite MA1 the detritus-controlled element
concentrations show a very good correlation with the (mostly) drip
rate dependent d18O and Mg/Ca records.
The fractionation of C and O isotopes was partly controlled by
temperature, which was lowest during the cold period between 3.5
and 2.5 ka BP and the Little Ice Age (0.7e0.1 ka BP). However, the C
and O isotope variations are mostly controlled by kinetic fractionation processes. The isotopic fractionation is only related directly to
drip rate and precipitation if the drip rate range is sufficiently high
(e.g. MA1). At low drip rates, individual drip sites are affected by
kinetic fractionation of the seepage water during prior calcite
precipitation (PCP) at the cave roof and at the stalactites. This
resulted in an isotope signature that is partly unrelated to climate.
The contents of carbonate-compatible elements are similar in
each of the three stalagmites and depend on precipitation. Mg/Ca
ratios were calibrated using precipitation data over the last 100
years from two different weather stations in the southern Andes.
This provides reliable estimates for annual precipitation, which
ranges between 4000 and 6500 mm/year during the last 5 ka.
The reconstructed temperature and precipitation changes indicate that cold phases were characterised by lower precipitation and
weaker westerlies in this region, especially during the Little Ice Age
and from 3.5 to 2.5 ka BP. The opposite, i.e. high precipitation and
strong westerlies in the SHW core section, prevailed during warm
phases especially during the Medieval Warm Period.
Acknowledgements
This work was funded by the German Research Foundation (DFG;
Grants AR 367/4, Ki 456/10-11, KR 2061/1, MA 821/32, Wo 362/18).
We thank Marcelo Arévalo for cave exploring and logistic support.
References
Akima, H.,1970. A new method of interpolation and smooth curve fitting based on local
procedures. Journal of the Association for Computing Machinery 17 (4), 589e602.
Baldini, J.U.L., McDermott, F., Hoffmann, D.L., Richards, D.A., Clipson, N., 2008. Very
high-frequency and seasonal cave atmosphere pCO2 variability: implications for
stalagmite growth and oxygen isotope-based paleoclimate records. Earth and
Planetary Science Letters 272 (1e2), 118e129.
Bar-Matthews, M., Ayalon, A., Kaufman, A., Wasserburg, G.J., 1999. The Eastern
Mediterranean paleoclimate as a reflection of regional events: Soreq cave,
Israel. Earth and Planetary Science Letters 166, 85e95.
Biester, H., Kilian, R., Franzen, C., Woda, C., Mangini, A., Schöler, H.F., 2002. Elevated
mercury accumulation in a peat bog of the Magellanic Moorlands, Chile (53 S)
e an anthropogenic signal from the Southern Hemisphere. Earth and Planetary
Science Letters 201 (3e4), 609e620.
Biester, H., Keppler, F., Putschew, A., Martinez-Cortizas, A., Petri, M., 2004. Halogen
retention, organohalogens, and the role of organic matter decomposition on
halogen enrichment in two Chilean Peat Bogs. Environmental Science and
Technology 38 (7), 1984e1991.
Boch, R., Spötl, C., Kramers, J., 2009. High-resolution isotope records of early
Holocene rapid climate change from two coeval stalagmites of Katerloch Cave,
Austria. Quaternary Science Reviews 28, 2527e2538.
Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffmann, S.,
Lotti-Bond, R., Hajdas, I., Bonani, G., 2001. Persistent solar influence on North
Atlantic climate during the Holocene. Science 294, 2130e2136.
Borsato, A., Frisia, S., Fairchild, I.J., Somogyi, A., Susini, J., 2007. Trace element
distribution in annual stalagmite laminae mapped by micrometer-resolution Xray fluorescence: implications for incorporation of environmentally significant
species. Geochimica et Cosmochimica Acta 71, 1494e1512.
Burton, E.A., Walter, L.M., 1991. The effects of pCO2 and temperature on magnesium
incorporation in calcite on seawater and MgCl2-CaCl2 solutions. Geochimica et
Cosmochimica Acta 55, 777e785.
Chambers, F.M., Mauquoy, D., Brain, S.A., Blaauw, M., Daniell, J.R.G., 2007. Globally
synchronous climate change 2800 years ago: proxy data from peat in South
America. Earth and Planetary Science Letters 253, 439e444.
Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmerom, Y., 2000. The
half-lives of uranium-234 and thorium-230. Chemical Geology 169, 17e33.
Cobb, K.M., Adkins, J.F., Partin, J.W., Clark, B., 2007. Regional-scale climate influences
on temporal variations of rainwater and cave dripwater oxygen isotopes in
northern Borneo. Earth and Planetary Science Letters 263 (3e4), 207e220.
Crowley, T.J., North, G.R., 1991. Paleoclimatology. Oxford University Press, New York.
Deininger, M., 2010. Klimaarchiv Stalagmit. Modellierung der Bildungsprozesse und
Rekonstruktion von Paläoklimaparametern. Diploma Thesis, Faculty of Physics
and Astronomy, Heidelberg University, Germany.
Denniston, R.F., DuPree, M., Dorale, J.A., Asmerom, Y., Polyak, V.J., Carpenter, S.J.,
2007. Episodes of Late Holocene aridity recorded by stalagmites from Devil’s
Icebox Cave, central Missouri, USA. Quaternary Research 68 (1), 45e52.
Dorale, J.A., Edwards, R.L., Ito, E., González, L.A., 1998. Climate and vegetation
history of the midcontinent from 75 to 25 ka: a speleothem record from Crevice
cave, Missouri, USA. Science 282, 1871e1874.
Fairchild, I.J., Tuckwell, G.W., Baker, A., Tooth, A.F., 2006. Modelling of dripwater
hydrology and hydrogeochemistry in a weakly karstified aquifer (Bath, UK):
implications for climate change studies. Journal of Hydrology 321, 213e231.
Frank, N., Braun, M., Hambach, U., Mangini, A., Wagner, G., 2000. Warm period
growth of travertine during the Last Interglaciation in southern Germany.
Quaternary Research 54, 38e48.
Frisia, F., Borsato, A., Preto, N., McDermott, F., 2003. Late Holocene annual growth in
three Alpine stalagmites records the influence of solar activity and the North
Atlantic oscillation on winter climate. Earth and Planetary Science Letters 216
(3), 411e424.
Henderson, G.M., 2006. Caving into new chronologies. Science 313, 620e622.
Hendy, C.H., 1971. The isotopic geochemistry of speleothems: I. The calculation of
the effects of different modes of formation on the isotopic composition of
speleothems and their applicability as palaeoclimatic indicators. Geochimica et
Cosmochimica Acta 35, 801e824.
Hill, C., Forti, P., 1997. Cave Minerals of the World. National Speleological Society,
Huntsville, Alabama.
Hofmann, M., 2006. Der Stalagmit MA-1 aus Südpatagonien als hochauflösendes
Klimaarchiv: Chemische Variabilität und Identifikation detritischer Komponenten im Zeitabschnitt der Mount Burney-Eruptionen (4300 und 2000 Jahre
B.P.). Bachelor thesis, Centre for Earth Science, Department of Geochemistry,
Göttingen University, Germany.
Huang, Y., Fairchild, I.J., 2001. Partitioning of Sr2þ and Mg2þ into calcite under karstanalogue experimental conditions. Geochimica et Cosmochimica Acta 65,
47e62.
Indermühle, A., Stocker, T.F., Joos, F., Fischer, H., Smith, H.J., Wahlen, M., Deck, B.,
Mastroianni, D., Tschumi, J., Blunier, T., Meyer, R., Stauffer, B., 1999. Holocene
carbon-cycle dynamics based on CO2 trapped in ice at Taylor Dome, Antarctica.
Nature 398, 121e126.
Katz, A., Sass, E., Starinsky, A., Holland, H.D., 1972. Strontium behaviour in the
aragonite-calcite transformation: an experimental study at 40e98 C. Geochimica et Cosmochimica Acta 36 (4), 481e496.
Kilian, R., Hohner, M., Biester, H., Wallrabe-Adams, H., Stern, C., 2003. Holocene peat
and lake sediment tephra record from the southernmost Andes (53e55 S).
Revista Geológica de Chile 30, 47e64.
Kilian, R., Biester, H., Behrmann, J., Baeza, O., Fesq-Martin, M., Hohner, M.,
Schimpf, D., Friedmann, A., Mangini, A., 2006. Millennial-scale volcanic impact
on pristine and superhumid ecosystem. Geology 34 (8), 609e612.
Kilian, R., Baeza, O., Steinke, T., Arévalo, M., Rios, C., Schneider, C., 2007. Late
Pleistocene to Holocene marine transgression and thermohaline control on
sediment transport in the western Magellanes fjord system of Chile (53 S).
Quaternary International 161, 90e107.
Kluge, T., Riechelmann, D.F.C., Wieser, M., Spötl, C., Sültenfuß, J., SchröderRitzrau, A., Niggemann, S., Aeschbach-Hertig, W., 2010. Dating cave drip water
by tritium. Journal of Hydrology 394 (3e4), 396e406.
Koch, J., Kilian, R., 2005. Little ice age glacier fluctuations, Gran Campo Nevado,
southernmost Chile. The Holocene 15 (1), 21e29.
Lachniet, M., 2009. Climatic and environmental controls on speleothem oxygenisotope values. Quaternary Science Reviews 28 (5e6), 412e432.
Lamy, F., Kaiser, J., Ninnemann, U., Hebbeln, D., Arz, H., Stoner, J., 2004. Antarctic
timing of surface water changes off Chile and Patagonian ice sheet response.
Science 304, 1959e1962.
Lamy, F., Kaiser, J., Arz, H.W., Hebbeln, D., Ninnemann, U., Timm, O.,
Timmermann, A., Toggweiler, J.R., 2007. Modulation of the bipolar seesaw in the
Southeast Pacific during Termination I. Earth and Planetary Science Letters 259
(3e4), 400e413.
Lamy, F., Kilian, R., Arz, H., Francois, J.-P., Prange, M., Steinke, T., 2010. Holocene
changes in the position and intensity of the southern westerly wind belt. Nature
Geoscience 3, 695e699.
Linge, H., Lauritzen, S.E., Lundberg, J., Berstad, I.M., 2001. Stable isotope stratigraphy
of Holocene speleothems: examples from a cave system in Rana, northern
Norway. Palaeogeography, Palaeoclimatology. Palaeoecology 167, 209e224.
McDonald, J., Drysdale, R., Hill, D., 2004. The 2002e2003 El Niño recorded in
Australian cave drip waters: implications for reconstructing rainfall histories
D. Schimpf et al. / Quaternary Science Reviews 30 (2011) 443e459
using stalagmites. Geophysical Research Letters 31, L22202. doi:10.1029/
2004GL020859.
McMillan, E., Fairchild, I.J., Frisia, S., Borsato, A., 2005. Calcite-aragonite trace element
behaviour in annually layered speleothems: evidence of drought in the Western
Mediterranean 1200 years ago. Journal of Quaternary Science 20, 423e433.
Mercer, J.H., 1982. Holocene glacial variations in southern Patagonia. Striae 18,
35e40.
Mickler, P.J., Stern, L.A., Banner, J.L., 2006. Large kinetic isotope effects in modern
speleothems. The Geological Society of America Bulletin 118 (1e2), 65e81.
Migon, P., 2006. Granite Landscapes of the World, Geomorphological Landscapes of
the World. Oxford University Press, New York.
Mohtadi, M., Romero, O.E., Kaiser, J., Hebbeln, D., 2007. Cooling of the southern high
latitudes during the Medieval Period and its effect on ENSO. Quaternary Science
Reviews 26, 1055e1066.
Mook, W., 2000. Environmental Isotopes in the Hydrological Cycle. Principles and
Applications. In: Introduction e Theory, Methods, Review, vol. 1. UNESCO/IAEA,
Paris.
Moy, C.M., Dunbar, R.B., Moreno, P.I., Francois, J.-P., Villa-Martinez, R.,
Mucciarone, D.M., Guilderson, T.P., Garreaud, R.D., 2008. Isotopic evidence for
hydrological change related to the westerlies in SW Patagonia, Chile, during the
last millennium. Quaternary Science Reviews 27, 1335e1349.
Mucci, A., 1987. Influence of temperature on the composition of magnesian calcite
overgrowths precipitated from seawater. Geochimica et Cosmochimica Acta 51,
1977e1984.
Mühlinghaus, C., Scholz, D., Mangini, A., 2008. Temperature and precipitation
records from stalagmites grown under disequilibrium conditions: a first
approach. Pages News 16 (3), 19e20.
Mühlinghaus, C., Scholz, D., Mangini, A., 2009. Modelling fractionation of stable
isotopes in stalagmites. Geochimica et Cosmochimica Acta 73, 7275e7289.
Musgrove, M., Banner, J., 2004. Controls on the spatial and temporal variability of
vadose drip water geochemistry: Edwards Aquifer, Central Texas. Geochimica et
Cosmochimica Acta 68, 1007e1020.
Neukom, R., Luterbacher, J., Villalba, R., Küttel, M., Frank, D., Jones, P.D., Grosjean, M.,
Wanner, H., Aravena, J.-C., Black, D.E., Christie, D.A., D’Arrigo, R., Lara, A.,
Morales, M., Soliz-Gamboa, C., Srur, A., Urrutia, R., von Gunten, L., 2010. Multiproxy summer and winter surface air temperature field reconstructions for
southern South America covering the past centuries. Climate Dynamics,
Springer. doi:10.1007/s00382-010-0793-3.
Niggemann, S., Mangini, A., Richter, D.K., Wurth, G., 2003. A paleoclimate record of
the last 17,600 years in stalagmites from the B7 cave, Sauerland, Germany.
Quaternary Science Reviews 22 (5e7), 555e567.
Oomori, T., Kaneshima, H., Maezato, Y., Kitano, Y., 1987. Distribution coefficient of
Mg2þ ions between calcite and solution at 10e50 C. Marine Chemistry 20,
327e336.
Proctor, C.J., Baker, A., Barnes, W.L., 2002. A three thousand year record of North
Atlantic climate. Climate Dynamics 19, 449e454.
Rosenblüth, B., Fuenzalida, H., Aceituno, P., 1997. Recent temperature variations
in southern South America. International Journal of Climatology 17, 67e85.
Salomons, W., Mook, W.G., 1986. Isotope geochemistry of carbonates in the
weathering zone. In: Fritz, P., Fontes, J.C. (Eds.), The Terrestrial Environmental B.
459
Handbook of Environmental Isotope Geochemistry, vol. 2. Elsevier, Amsterdam,
pp. 239e269.
Schneider, C., Glaser, M., Kilian, R., Santana, A., Butorovic, N., Casassa, G., 2003.
Weather observations across the southern Andes at 53 S. Physical Geography
24 (2), 97e119.
Schneider, C., Schnirch, M., Acuña, C., Casassa, G., Kilian, R., 2007. Glacier inventory
of the Gran Campo Nevado Ice Cap in the Southern Andes and glacier changes
observed during recent decades. Global and Planetary Change 59, 87e100.
Sepulveda, J., Pantoja, S., Hughen, K.A., Bertrand, S., Figueroa, D., Leon, T.,
Drenzek, N.J., Lange, C., 2009. Late Holocene sea-surface temperature and
precipitation variability in northern Patagonia, Chile (Jacaf Fjord, 44 S).
Quaternary Research 72, 400e409.
Spötl, C., Vennemann, T., 2003. Continuous-flow IRMS analysis of carbonate
minerals. Rapid Communications in Mass Spectrometry 17, 1004e1006.
Stern, C.R., 2008. Holocene tephrochronology record of large explosive eruptions in
the southernmost Patagonian Andes. Bulletin of Volcanology 70 (4), 435e454.
Tipping, E., Smith, E.J., Lawlor, A.J., Hughes, S., Stevens, P.A., 2003. Predicting the
release of metals from ombrotrophic peat due to drought-induced acidification.
Environmental Pollution 123 (2), 239e253.
Tooth, A.F., Fairchild, I.J., 2003. Soil and karst aquifer hydrological controls on the
geochemical evolution of speleothem-forming drip waters, Crag Cave, southwest Ireland. Journal of Hydrology 273, 51e68.
Van Geel, B., Heusser, C.J., Renssen, H., Schurmanns, C.J.E., 2000. Climate change in
Chile at around 2700 BP and global evidence for solar forcing: a hypothesis. The
Holocene 10, 659e664.
Villalba, R., Lara, A., Boninsegna, J.A., Masiokas, M., Delgado, S., Aravena, J.C.,
Roig, F.A., Schmelter, A., Wolodarsky, A., Ripalta, A., 2003. Large-scale temperature changes across the southern Andes: 20th century variations in the
context of the past 400 years. Climatic Change 59, 177e232.
Vollweiler, N., Scholz, D., Mühlinghaus, C., Mangini, A., Spötl, C., 2006. A precisely
dated climate record for the last 9 kyrs from high alpine stalagmites.
Geophysical Research Letters 33, L20703. doi:10.1029/2006GL027662.
Wanner, H., Beer, J., Bütikofer, J., Crowley, T.J., Cubasch, U., Flückiger, J., Goosse, H.,
Grosjean, M., Joos, F., Kaplan, J.O., Küttel, M., Müller, S.A., Prentice, I.C.,
Solomina, O., Stocker, T.F., Tarasov, P., Wagner, M., Widmann, M., 2008. Mid- to
Late Holocene climate change: an overview. Quaternary Science Reviews 27
(19e20), 1791e1828.
Wedepohl, K.H., 1995. The composition of the continental crust. Geochimica et
Cosmochimica Acta 59 (7), 1217e1232.
Wiedner, E.A., Scholz, D., Mangini, A., Polag, D., Mühlinghaus, C., Segl, M., 2008.
Investigation of the stable isotope fractionation in speleothems with laboratory
experiments. Quaternary International 187 (1), 15e24.
Williams, P.W., King, D.N.T., Zhao, J.-X., Collerson, K.D., 2004. Speleothem master
chronologies e combined Holocene 180 and 13C records from the North Island of
New Zealand and their palaeoenvironmental interpretation. The Holocene 14 (2),
194e208.
Zhou, H., Wang, Q., Zhao, J.-X., Zheng, L., Guan, H., Feng, Y., Greig, A., 2008. Rare
earth elements and yttrium in a stalagmite from Central China and potential
paleoclimatic implications. Palaeogeography, Palaeoclimatology, Palaeoecology
270, 128e138.