Regional climate variability and ecosystem responses to the last deglaciation in the Northern Hemisphere from stable isotope data and calcite fabrics in two northern Adriatic stalagmites

Full text

Regional climate variability and ecosystem responses to the last
deglaciation in the northern hemisphere from stable isotope data and
calcite fabrics in two northern Adriatic stalagmites
R. Belli
a
,
*
, S. Frisia
a
, A. Borsato
a
, R. Drysdale
b
, J. Hellstrom
c
, J.-x. Zhao
d
, C. Spötl
e
a
School of Environmental and Life Science, The University of Newcastle, Callaghan, NSW 2308, Australia
b
Department of Resource Management and Geography, The University of Melbourne, Parkville, VIC 3010, Australia
c
School of Earth Sciences, The University of Melbourne, Parkville, VIC 3010, Australia
d
The University of Queensland, Brisbane, QLD 4072, Australia
e
Institute of Geology and Paleontology, The University of Innsbruck, 6020 Innsbruck, Austria
article info
Article history:
Received 15 June 2012
Received in revised form
18 April 2013
Accepted 19 April 2013
Available online
Keywords:
Stable isotopes
Calcite fabric
Lamina thickness
Speleothems
Younger Dryas
Early Holocene
abstract
Geochemical and physical changes along the growth axis of speleothems are controlled by climate as
well as the dynamics of the host karst system. Discriminating between the two is one of the major
challenges in assembling robust cave palaeoclimate records. Overcoming this dilemma may be achieved
by investigating multiple properties from two or more speleothems from the same cave, spanning the
same time interval. In this study, we use high-resolution stable oxygen and carbon isotope (
d
13
C and
d
18
O), morphological and petrographic data from two calcite stalagmites (SV1 and SV7) from Grotta Savi,
Italy, to help constraining the regional climate conditions during the Younger Dryas. The new SV1 record,
spanning 15.3e9.4 ka, builds on a previously published low-resolution stable isotope time series. It is
anchored by 22 new UeTh ages and lamina-counting, yielding a refined age model that supersedes the
previous one. The SV7 record is anchored by 8 UeTh ages and spans 13.1e10.5 ka, partially overlapping
the SV1 record. Together, the morphological and geochemical characteristics of SV1 and SV7 allow us to
identify site-specific processes driving the isotopic variations. This enables us to explore the degree of
complacency in speleothems whereby, periodically, a speleothem is insensitive to recording a climate
signal, due to aquifer hydrology and network evolution. The model suggests that
d
18
O is a proxy for the
infiltration amount that reaches the site of deposition, which, in the case of SV7, was buffered by the
aquifer hydrology. The interpretation of
d
13
C in both stalagmites is more complex, being influenced by
both temperature and hydrology. Only a comparison of the two stalagmite records allows us to recognise
a decoupling between the effects of hydrology and temperature. According to our interpretation, the
most distinctive climate information of the Younger Dryas was the high hydrological variability, with
phases dominated by seasonal infiltration within an overall cooling trend. Our record also reveals a
significant Early Holocene climate anomaly ecentred at 10.4 ka ewhose magnitude was possibly
amplified by local synoptic conditions. Our research demonstrates the importance of investigating
several complementary proxies, particularly from morphologically diverse stalagmites. When moni-
toring data are not available, fabric and lamina thickness can be fundamental to unravelling the domi-
nant control of hydrology or temperature on stable isotopes.
Ó2013 Elsevier Ltd. All rights reserved.
1. Introduction
The last deglaciation was characterised by significant global-
scale climate, atmospheric and ocean reorganisations (Clark et al.,
2012 and reference therein) that probably disrupted continental
and ocean ecosystems. For example, in the northern hemisphere,
large-scale meltwater release from the retreating Laurentide ice
sheet during the BøllingeAllerød warming resulted in a large influx
of freshwater to the Arctic Ocean and a shutdown of the North
Atlantic Meridional Overturning Circulation, which is thought to
have caused the Younger Dryas (YD) cooling event (e.g. Broecker
et al., 1989;Thornalley et al., 2010;Not and Hillaire-Marcel,
2012). Given that a cooling similar to the YD may present itself as
*Corresponding author. Current address: Centre for Integrative Biology (CIBIO),
University of Trento, Via Sommarive, I e38123 Povo, Trento.
E-mail addresses: Romina.Belli@uon.edu.au,Romina.Belli@unitn.it (R. Belli).
Contents lists available at SciVerse ScienceDirect
Quaternary Science Reviews
journal homepage: www.elsevier.com/locate/quascirev
0277-3791/$ esee front matter Ó2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.quascirev.2013.04.014
Quaternary Science Reviews 72 (2013) 146e158

a“surprise”of global warming (Cohen et al., 2012), unravelling the
detailed spatial patterns of temperature and hydrological changes
through the YD is of considerable importance, not only for under-
standing the Lateglacial in general but also for future climate
predictions.
There is, currently, a shortage of coupled palaeotemperature/
palaeohydrological information from terrestrial archives for the
Lateglacial and the transition to the Holocene, in particular for the
Mediterranean region, which is being gradually redressed by spe-
leothem research (e.g. Genty et al., 2006). Speleothems are, in fact,
an environmental archive capable of yielding reliable records of
palaeohydrological changes (Zanchetta et al., 2007).
Studies over the last decade, however, show that complex pro-
cesses operating at global, regional and site-specific scales combine
to drive variations in
d
18
O and
d
13
C of calcite speleothems (Fairchild
and Baker, 2012 and references therein). It has been also pointed
out that the interpretation of the
d
18
O and
d
13
C is particularly
complex for caves located in the mid latitudes (Baker et al., 2011),
and decoupling past moisture and temperature influences on these
geochemical proxies remains a great challenge in speleothem
research (Fohlmeister et al., 2012). One of the fundamental issues
concerns whether karst-specific processes, which may be unrelated
to palaeoclimate but rather associated with the inherent features of
the soil-karst-cave system (Baker et al., 1997;Mühlinghaus et al.,
2007,2009), impact upon, or perhaps even dominate, climate-
related signals recorded by speleothems. To this end, one can use
different approaches, such as modelling of soil and cave-
monitoring data (Frisia et al., 2011), calibrating modern proxies to
instrumental data and extrapolating back through time under the
assumption that instrumental data represent past conditions
(Schimpf et al., 2011), or utilising a suite of proxies from a single
speleothem to reconstruct palaeoclimate (Oster et al., 2010).
Bearing in mind the potential influence of karst-specific pro-
cesses on speleothem climate proxies, in this paper we re-examine
the previously published SV1 record of
d
13
C and
d
18
O with a new
high-resolution dataset spanning the Lateglacial to the Early Ho-
locene. Frisia et al. (2005) showed that several large and rapid
changes of the Lateglacial were recorded in the stable carbon and
oxygen isotope records of SV1 from Grotta Savi, northern Adriatic
(Italy), whereas relatively more stable climatic conditions were
reconstructed for the Holocene portion of this record. To date, this
stalagmite record remains the only one of this type for the Late-
glacial in the south-eastern Alps of Europe, a region strategically
positioned at the boundary between the Mediterranean and central
Europe (Davis et al., 2003). In refining the age model of SV1, we
determined that the original chronology of Frisia et al. (2005) was
incorrect. The portion of record herein reinvestigated is anchored in
time by an entirely revised age model based on 22 new UeTh ages
and supplemented by lamina counting. To complement the SV1
data, we studied a second Grotta Savi stalagmite (SV7), whose
Lateglacial record partially overlaps that of SV1. The two stalag-
mites show similar fabric but, crucially, different morphologies,
which indicate different karst-hydrological settings (Miorandi
et al., 2010). Therefore, they provide an opportunity to explore
whether the karst-specific hydrology exerted any significant in-
fluence on the calcite
d
13
C and
d
18
O(Mickler et al., 2006). In our
approach, we assumed that present-day monitoring data may not
reflect the cave conditions of many thousands of years ago,
particularly as far back as the last deglaciation. Without the pos-
sibility of making direct climate measurements, we utilise the
petrography and the inferred calcite precipitation rate to under-
stand the climatic significance of Savi
d
13
C and
d
18
O values in terms
of the relative contributions of temperature and moisture. Petrog-
raphy is fundamental to the interpretation of geochemical proxies
(Frisia et al., 2000;Mattey et al., 2010;Boch et al., 2011), yet as we
illustrate below, it can also be used as a proxy in its own right, an
approach whose potential has yet to be fully exploited.
2. Study area and cave setting
Grotta Savi is located in NE Italy (45

37
0
N, 13

53
0
E) on the
northernmost seaboard of the Adriatic Sea (Fig. 1). The main
entrance to the cave opens at 350 m a.s.l. on the NE slope of Monte
Stena, a large limestone overthrust, which forms a plateau at a
general elevation of circa 450 m a.s.l. The catchment of the cave drip
waters has developed entirely in 80e100 m-thick Palaeocene
Opicina Limestone (Lenaz et al., 2004), whose stable carbon iso-
topic composition is 1.2 0.1&. The bedrock is covered by less than
50 cm of soil that supports grass, shrubs and scattered conifers
(Pinus nigra). Grotta Savi is in a region with a strong seasonal
climate pattern. June and September to November account for 50%
of the annual rainfall (1320 mm/year). The mean annual air tem-
perature is 12.5

C (range: 3.6

C in January to 22.7

C in August)
(GNIP, 2006), which is reflected in the cave temperature where the
stalagmites were collected (12.3 0.2

C, measured hourly from
2004 to 2007). A further regional feature is the wind pattern. The
short distance between the Adriatic and Friulian mountain ranges
to the Adriatic Sea, which bracket the cave site, produces a large
Fig. 1. Location map of Grotta Savi and other palaeoclimate archives cited in the text.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 147

atmospheric pressure difference resulting in strong Bora-wind
events (Grisogono and Belu
si
c, 2009).
3. Materials and methods
3.1. Stalagmites
The candle-shaped, actively forming stalagmite SV1 (Fig. 2)was
extracted from the large ‘Sala Morpurgo’chamber, located
approximately 500 m from the original cave entrance. SV1 was fed
by a stalactite located more than 15 m overhead, with a mean
discharge of 6 drops/min at the time of sampling. The mean su-
persaturation index of the water with respect to calcite (SI
cc
)is
0.29 0.1 (n¼6, year 2003) (PHREEQC, Parkhurst and Appelo,
1999) and the mean Ca concentration in the drips is
84.2 6.1 mg/l (n¼6, year 2003).
The inactive stalagmite SV7 (Fig. 2) was collected from a narrow
passage w200 m from the SV1 chamber. The distance between SV7
and its fossil-feeding stalactite was circa 1.5 m, significantly less
than that of SV1. Since the cave floor on which the two stalagmites
grew is approximately at the same level and the terrain above the
cave is flat, such a difference indicates a greater thickness of
limestone above SV7.
The tabular core of SV1 and SV7 average 67 and 40 mm inwidth,
respectively. Despite the persistence of the tabular core
morphology, SV7 shows a progressive diametric narrowing towards
the top (Fig. 2). Fabric and micromorphology investigations were
conducted on thin sections using optical microscopy (ZEISS Axio-
plan, in bright-field mode, transmitted light). Thin sections and
geochemical analyses were performed on adjacent slabs.
3.2. Chronology
Seventeen new UeTh dating samples from SV1 and eight from
SV7 were collected, spanning the Lateglacial to the Early Holocene,
and analysed at the University of Melbourne using multi-collector
inductively coupled plasma mass spectrometry (MC-ICP-MS) and
employing the analytical method of Hellstrom (2003). A further
five SV1 samples were analysed by MC-ICP-MS at the University of
Queensland following the procedure of Zhou et al. (2011). In both
laboratories, an initial
230
Th/
232
Th activity ratio of 1.3 0.3 (2
s
)
was estimated for Savi stalagmites, based on the principal of
stratigraphical constraint (Hellstrom, 2006), and this was used to
calculate corrected ages in conjunction with the decay constants of
Cheng et al. (200 0). Through the portion of greatest interest in SV1
(248e207 mm), lamina-counting was performed using a Zeiss
optical microscope and digital-image processing software. The
mean error of the counting (3.9 0.4%) was estimated assuming a
minimum error of 1.0% for sections containing relatively thick,
clear laminae and a maximum error of 4.2% determined from
counts on four parallel transects through the thinner, less well-
resolved laminations. Time series and UeTh ages are reported in
thousands of years (ka) before present, where present is defined as
2000 AD.
3.3. Stable isotopes
Powder samples for stable isotope analysis were collected along,
and perpendicular to, the central growth axis at a fixed 150-
m
m
increment for both SV1 and SV7. Subsampling and isotopic analysis
of stalagmites SV7 and new SV1 were carried out at the University
of Innsbruck (Austria) and the University of Newcastle (Australia),
respectively. In both laboratories calcite powders were analysed by
conventional acid-digestion using the analytical protocols as out-
lined in Spötl and Vennemann (2003) and Drysdale et al. (2009).
The long-term reproducibility (1
s
) of analyses from the University
of Innsbruck laboratory is 0.06&for
d
13
C and 0.08&for
d
18
O(Spötl,
2011). The mean 1
s
analytical reproducibility of a Carrara Marble
internal working standard (NEW1) at the University of Newcastle
laboratory was 0.05&for
d
13
C and 0.09&for
d
18
O. Stable carbon
and oxygen isotopic ratios are reported in ‘per mille’(&) relative to
the Vienna Pee Dee Belemnite (VPBD) scale using the standard
d
notation.
4. Results
4.1. Revision of SV1 chronology
A major limitation of using SV1 for the purpose of this study was
the limited number of UeTh age determinations through the
Fig. 2. A) New SV1 UeTh ages (black circles) with 2
s
uncertainties, age model (solid black line) and 2
s
age-model uncertainties (dashed lines). The red line shows the best-fit of the
SV1 lamina-counting chronology to the UeTh chronology. B) SV1 extension rate (black line) and 2
s
uncertainty (grey lines). C) Stalagmite SV7 UeTh ages (black circles) with 2
s
uncertainties, age model (solid black line) and 2
s
age-model uncertainties (dashed lines). D) SV7 extension rate (black line) and 2
s
uncertainty (grey lines). The same scale is used
(subscale section ¼5 cm) for the images of stalagmites SV1 (left) and SV7 (right). The dotted lines on the stalagmite sections indicate the growth axes sampled for stable isotope
analyses. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158148

Lateglacial and the Early Holocene period (Frisia et al., 2005). Thus,
to constrain the timing of major features revealed by SV1, a new set
of UeTh age determinations were carried out on the interval of
interest (see Section 4.2). In refining the age model, it was revealed
that the original chronology of the Lateglacial portion (Frisia et al.,
2005), based on UeTh ages carried out in 2004 at the Laboratory of
Isotope Geochemistry, University of Bern (Switzerland), was
incorrect. Possible reasons for the discrepancies are related to the
high
230
Th/
232
Th in several samples, which led to large age un-
certainties, and to a memory effect that systematically affected the
mass spectrometer measurements in the Bern chronology
laboratory.
For this reason, we use the new UeTh ages for SV1 and retract
the ages and the age model of Frisia et al. (2005) for the older
portion of the record (ex 16.6 kae10.9 ka). The new age model
establishes the onset of SV1 record at w15.3 ka and shifts by w1000
years the major structures identified in record of Frisia et al. (2005),
with the Older Dryas of the original work now identified as the
Younger Dryas. Our new SV1 record ends at 9.4 ka and includes the
last major isotope peak of the entire stalagmite record. Because of
the systematic error in the original dataset, it is very likely that the
youngest portion of SV1 also experienced a similar error. Given that
the focus of the present study was the Lateglacial, and not the
Holocene, for which SV1 did not record significant sub-millennial
isotopic events, there are no new radiometric dates for the last
w9300 years. At the present state of knowledge, we suggest that
the Holocene record of SV1 be used with caution until it can be re-
dated.
4.2. SV1 and SV7 chronologies
The new set of SV1 UeTh ages ranges from 15.3 1.6 ka t o
9.4 0.2 ka (Fig. 2A), and entirely overlaps the 13.1 0.3 kae
10.5 0.5 ka timespan of the SV7 growth phase. Stalagmite SV1
began to grow over a dark 10-mm thick condensed calcite layer that
contains too much detritus for reliable dating. Age models were
determined using a Bayesian-Monte Carlo approach (Drysdale
et al., 2005,2007) with a detrital-thorium-based correction as
outlined in Hellstrom (2006). Details of the corrected UeTh ages of
SV1 and SV7 are reported in the Supplementary Data (S1). After age
correction, the new SV1 UeTh ages, however, still show large un-
certainties of up to 1.6 ka at 15.3 ka and a number of UeTh age
reversals between 227.5 and 215.7 mm (Table S1.1). The series of Ue
Th ages shows also a step of 840 years between 235 and 233 mm
(Fig. 2A), which includes the dark layers observed in this section
(Section 5.1). Under such conditions, two independent UeTh-based
age models better describe the pre- and post-step chronology
(Scholz and Hoffmann, 2011).
To refine the SV1 age model, lamina-counting was performed
between 248 and 219 mm. The floating lamina chronology (con-
tained within the 14.3e10.4 ka interval) was anchored to the UeTh
age model using two clear markers in the stalagmite and the small
uncertainties of two UeTh ages (at 236.5 and 232.0 mm, 2
s
0.09
and 0.07 ka, respectively). This successfully constrains the floating
lamina chronology by minimising age differences between the
lamina-counting and the UeTh age models. The best-fit of the SV1
lamina-counting chronology to the UeTh age model (within 2
s
Ue
Th age uncertainties) is shown in Fig. 2A. The lamina chronology
was used to derive an age-depth model for the 248.0e235.3 mm and
the 233.0e219.3 mm depth intervals, providing an annually
resolved chronology between 14.3 and 10.4 ka. Through sections
with no discontinuities (e.g. 233.0e220.0 mm), the correspondence
of the lamina chronology to the UeTh age-model slopes supports
the annual periodicity represented by each lamina. A condensed but
continuous growth was assumed from 235 to 233 mm throughout
the dark layers section (Section 5.1). Across this portion, the timee
depth relationship was anchored to the nearest lamina ages. The
SV1 time series are plotted on an age model that combines both the
UeTh age model and the lamina-counting chronology (Fig. 2A). This
age model reveals moderately fast axial growth (13e14
m
m/year)
between 15.3 and 13.3 ka and extremely slow growth rates (<5
m
m/
year) from 12.6 to 11.9 ka and 10.4 to 10.1 ka (Fig. 2B).
The distribution of SV7 UeTh ages versus stalagmite depth
shows linear growth (Fig. 2C). Given that the
230
Th/
232
Th ratios are
>500, the final ages were only minimally influenced by
230
Th
corrections (Table S1.1). The UeTh ages of SV7 show at least four
stages of calcite precipitation (w10e13 ka, 40e83 ka, 120e130 ka,
and 209e248 ka). In this study, we focus on the younger section,
which spans the 13.1 0.3 kae10.5 0.5 ka interval. Petrographic
observations and lack of significant steps in the age-depth rela-
tionship of SV7 indicate the absence of major hiatuses throughout
this section. SV7 growth rates vary from 5
m
m/year (11.5e10.8 ka) to
up to 29
m
m/year at 11.7 ka (Fig. 2D). At 4.5 mm from the top
(10.8 ka), SV7 preserves a hiatus. Calcite precipitation resumed to
form the last few millimetres of growth, which are dated at
8.1 0.6 ka.
4.3. Fabric and microstratigraphy
Both SV1 and SV7 stalagmites comprise translucent, light-
honey-coloured laminated calcite. They consist exclusively of
low-Mg calcite arranged as a columnar fabric with large, composite
crystals with uniform extinction (Frisia et al., 2000;Frisia and
Borsato, 2010). Those millimetre-size crystals are elongated along
the c-axis. In both stalagmites, the compact columnar fabric makes
several transitions to open columnar laminated calcite which dis-
plays elongated voids between crystals in the growth direction.
Sub-millimetre-thick laminae are visible under transmitted light
and occur within the large calcite crystals. These laminae are
composed of translucent, columnar calcite and are identified by a
regular (up to 1.5-
m
m thick) dark layer, similar to the annual
lamination described in Grotta di Ernesto stalagmites (Frisia et al.,
2003). The SV1 lamina thickness between 14.4 and 9.8 ka ranges
from 42.4 to 0.7
m
m (average 6.7 3.0
m
m, n¼3515). The regular
stacking of the flat, parallel laminae, which characterises the lower
section of the stalagmite, is abruptly terminated at 12.7 ka for 840
years and comprises either deposition of several darker layers up to
350
m
m thick, or the stacking of thin (<5
m
m) laminae. For the
period from 11.5 to 11.0 ka, the calcite fabric becomes predomi-
nantly open with visible lamination. Laminae again become
extremely thin (<2
m
m) from 10.6 to 10.1 ka. Similar features are
not present in SV7. Neither the open calcite fabric nor the dark,
impurity-rich layers disturb the growth of the columnar-composite
crystals, which preserve the same orientation as the substrate, with
no evidence of re-nucleation episodes. SV1 fabric and micro-
stratigraphy are described using a coding system, where 1 is
compact columnar calcite with flat, parallel laminae (Fig. 3A), 2 and
3 describe open columnar calcite, with and without clear laminae
respectively (Fig. 3B), and 4 pertains to sections containing darker
layers (Fig 3C). The SV1 lamina thickness and stratigraphic log are
illustrated in Fig. 4AeB.
4.4.
d
18
O and
d
13
C analyses
The
d
18
O(
d
18
O
c
) and
d
13
C(
d
13
C
c
) time series of stalagmites SV1
and SV7 are shown in Fig. 4CeD. The high-resolution SV1 record
replicates well the existing lower resolution SV1 record of Frisia
et al. (2005) (Supplementary Data Section. S2.1). As SV1 and SV7
exhibit similar
d
18
O
c
variations (Fig. 4C), the
d
18
O
c
maxima of SV7
were aligned with those of SV1 (Supplementary Data Section S2.2).
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 14 9

The SV1 age model is based on UeTh ages from the two different
laboratories, and have average 95% uncertainties of 4%, which is 2%
higher than that obtained for SV7 (Fig. 2). SV1, however, shows
visible lamination that not only provides a means for developing an
internal chronology, but also allows insights into internal cave
processes, which may have modified the stable isotope signatures.
This is particularly critical for the
d
13
C
c
values which are known to
be highly responsive to cave processes, in contrast to
d
18
O
c
(Mühlinghaus et al., 2007). Thus,
d
18
O
c
excursion maxima were
used to tune the SV7 to the SV1 record, whose chronology were
used as the target because based both on lamina-counting and
radiometric ages. We stress that this tuning was performed within
the age uncertainties of the SV7 age models (Supplementary Data
Section S2.2). The SV7
d
13
C
c
and
d
18
O
c
time series are plotted as a
function of the tuned chronology in Fig. 4EeF. As result of the
tuning, the abrupt shift in
d
13
C
c
of SV7 at w12 ka precedes that in
SV1 (Fig. 4F). The delay in the shift of SV1
d
13
C
c
corresponds to
changes in the calcite fabric (Fig. 4A) not observed in SV7. Therefore,
such a delay is unlikely to be an artefact of the tuning but rather a
consequence of a different response of the two stalagmites to a
change in the catchment or in the cave. This anomaly is discussed
further in Section 5.2.2.
During glacial-to-interglacial transitions the
d
18
O
c
value of at-
mospheric moisture changes due to the effects of ice-sheet melting
on ocean water
d
18
O values. The
d
18
O
c
series has been adjusted by
0.11&for every 10 m (Fairbanks and Matthews, 1978) of sea-level
rise (Siddall et al., 2003). The ice-volume-adjusted
d
18
O
c
values vary
from 6.0 to 8.4&in SV1 (mean: 7.1 0.3&)(Fig. 4E). Although
the adjustment reduced the total
d
18
O
c
variation, a Lateglacial to
Early Holocene trend of 1.1&towards more negative values is
preserved. Such a trend is more evident in the SV1
d
13
C
c
time series
(Fig. 4F), which ranges between 9.0 and 11. 5 &(mean:
10.4 0.4&). The overall correlation between the SV1
d
18
O
c
and
d
13
C
c
is moderate but statistically significant (r¼0.4 p<0.0001,
n¼241).
The SV7
d
13
C
c
values and the ice-volume-adjusted
d
18
O average
10.0 0.4&(from 8.6 to 10.7&) and 6.6 0.3&(from 5.5
to 7.7 &), respectively. During the overlapping period based on
the tuned chronology (13.0e10.8 ka), the SV7
d
18
O
c
versus
d
13
C
c
ratios show a stronger correlation (r¼0.8 p<0.0001, n¼276)
relative to SV1 (r¼0.2 p<0.1, n¼85). It is worth noting that across
this period, both the
d
18
O
c
and
d
13
C
c
values of SV7 are w0.7 and
0.8&, respectively, more positive than SV1 (Section 5.1), with some
exception as discussed in Section 5.2.
5. Discussion
5.1. Stalagmite growth rates and calcite fabrics
The mean vertical axial growth rate is slow in both SV1 and SV7
(7 versus 11
m
m/year, respectively), although the two stalagmites
show opposite trends. From 12.8 to 11.9 ka, SV7 reached its highest
extension rate of 21
m
m/year, while SV1 reached its lowest rate
(5
m
m/year). Lamina thickness in calcite is commonly controlled
by Ca content and saturation state, which are related to climate
parameters as well as cave ventilation (Dreybrodt, 2008;Genty
et al., 2001b). For example, annual lamina thickness in several
stalagmites from Grotta di Ernesto, at 1165 m a.s.l. and 200 km west
of Grotta Savi, has been positively correlated with surface tem-
perature, which in that cave drives seasonal patterns of degassing,
pH of the film of fluid wetting the speleothems and morphology, as
well as architecture of crystal surfaces (Frisia et al., 2003,2011). SV7
and SV1 show antipathetic growth trends. Given that the pCO
2
of
the cave air today is similar at the two sites where stalagmites have
grown (w8500 ppm for SV7 and w8000 ppm for SV1, June 2010), it
is unlikely that their axial growth rate was controlled by a different
pattern of cave ventilation. If this assumption is true, the growth
rate differences between the two stalagmites were most likely
related to the effect of changes in drip-rate interval on degassing at
the stalagmite surface (Kaufmann, 2003). Evidence for this can be
seen in the morphology of the stalagmites. The narrower inner core
and width of SV7 relative to SV1 most likely originated from a lower
potential for calcite deposition from its feeding drip, which resulted
in a mass of calcite precipitated circa 2.5 times lower than SV1, over
the same period of growth (11.8e10.8 ka). Assuming similar calcite
supersaturation levels for the two drips, such a mass difference is
likely to be caused by a longer drip interval (Kaufmann, 2003).
Lower, but constant discharge reconstructed for the SV7 drip must
have resulted in a stagnant film of fluid covering the stalagmite
surface. As modelled by Mühlinghaus et al. (2007,2009) for low
drip rates, Rayleigh distillation would be operative even at the apex
of a stalagmite, and calcite precipitation would occur in disequi-
librium. Hence, a higher drip interval for the SV7 feeding system,
relative to SV1, seems the possible cause for the measured isotopic
w0.7e0.8&offset between coeval layers of the two stalagmites.
Compact columnar calcite is the most common fabric type
observed in SV1 and SV7, and has been related to low inter-annual
drip rate variability (Frisia and Borsato, 2010). However, SV1 shows
a sequence of layers characterised by open columnar fabric, which
has been related to higher drip rate variability with respect to
compact fabrics (Boch et al., 2011). In SV1, the open columnar
calcite at 12.9 ka includes a few thick, dark layers (Figs. 3Ce4B). This
portion of stalagmite SV1 coincides with brighter luminescence,
which has been related to higher concentrations of organic com-
pounds (Stoykova et al., 2005). Similar dark layers in stalagmites
from Grotta di Ernesto were interpreted as reflecting the accumu-
lation of colloidal organic matter mobilised from the soil during
autumnal infiltration events, when annual calcite precipitation was
greatly reduced as a result of prolonged winters (Frisia et al., 2003;
Borsato et al., 2007;Baker et al., 2011;Hartland et al., 2012).
The SV1 dark layers show lateral correlation with laminated
calcite (Fig. 3C). No clear evidence for dissolution or dissolution-
Fig. 3. Microstructure and fabric of SV1 as observed in thin section (transmitted light). A) Compact columnar calcite with horizontal, parallel, clearly resolved laminae. B) Open
columnar calcite with poorly visible lamination. C) Dark calcite layers characterised by high density of impurities. The arrows indicate regions of condensed deposition.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158150

reprecipitation fabric is found within these layers. On the basis of
these observations and comparison with other alpine speleothems
(Frisia et al., 2003) and sedimentary carbonate systems (Miall,
2000), we interpret the SV1 dark layers as indicative of
condensed deposition. While in a few areas calcite could still form,
on most of the stalagmite surface a combination of increased flow
of water carrying impurities and lower supersaturation inhibited
calcite deposition, resulting in a proportionally higher content of
organic compounds. Thus, the dark layers are not hiatuses but,
rather, identify periods of extremely reduced calcite deposition.
5.2. The interpretation of
d
18
O and
d
13
C values in Savi stalagmites
In this section we investigate the contribution of climate and
karst-specific signals to stable isotope compositions. To this end, we
compare the time series from the two stalagmites by plotting the
d
18
O
c
and
d
13
C
c
deviations from the mean value (anomalies)
(Fig. 5AeB). By removing the positive shift in SV7 isotopic records,
this graph allows the visual comparison of the two stalagmite series
and the identification of two episodes, labelled Episodes 1 and 2.
These Episodes are defined by distinct changes in the correlation
Fig. 4. Stalagmite SV1 and SV7 time series. A) The SV1 lamina-thickness record (grey) and the 9-point running mean (black), plotted on a logarithmic scale. B) The SV1 calcite
fabrics. Codes are: 1 ¼compact columnar calcite, with flat parallel laminae; 2 ¼open columnar calcite with resolved lamination; 3 ¼open columnar calcite with less-resolved
lamination, 4 ¼dark layers. C) The high resolution
d
18
O
c
and D)
d
13
C
c
time series of stalagmites SV1 (grey line) and SV7 (black line). The mean stable oxygen (square) and car-
bon (triangle) isotope compositions for the SV1 Holocene record are shown (the symbol size represents standard deviations from the mean values) (from Frisia et al., 2005). E) The
SV1 (grey) and SV7 (black) global ice-volume adjusted
d
18
O
c
and F) the
d
13
C
c
time series, plotted on the tuned chronology (see main text for explanation). At the top, black (SV1) and
white (SV7) circles indicate the positions of the UeTh ages.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 151

among
d
18
O
c
anomalies in the two stalagmites. Based on the posi-
tive or negative
d
18
O
c
anomalies, Episode 2 is further divided into
2a and 2b.
Our interpretation of SV1 and SV7
d
18
O
c
and
d
13
C
c
from the
Lateglacial to the Early Holocene is based on the assumption that
SV1 calcite precipitated close to isotopic equilibrium (Frisia et al.,
2005), thus its
d
13
C
c
and
d
18
O
c
values likely reflect those of the
original drip water. This assumption is corroborated by the 11-cm
diameter of the stalagmite, which is considered by Dreybrodt and
Scholz (2011) as typical for stalagmites that reflect the original
composition of drip water. Consequently, SV1 stable isotope values
should reflect processes external to the cave, such as rainfall iso-
topic composition and soil dissolved inorganic carbon (DIC)
(Dreybrodt and Scholz, 2011). Under these premises, the
d
13
C
c
values depend on the combination of soileand limestoneederived
carbon, from which the DIC in groundwater derives (Genty et al.,
2001a,2003;Rudzka et al., 2011). The soil-carbon pool is mostly
derived from the extent of soil microbial respiration which is pri-
marily driven by temperature, when soil moisture is not a limiting
factor (Fairchild and Baker, 2012). The similarities of the SV1
d
13
C
c
anomalies to lamina thickness variations (Fig. 5BeC), which are
interpreted to be mostly modulated by temperature, supports the
hypothesis that the
d
13
C
c
variability is mostly influenced by
temperature-related ecosystem processes at the surface. In
addition, the lack of evidence for C4 pollen in the region (Monegato
et al., 2011), excludes any influence from changes in the ratio of C4
to C3 plants to the
d
13
C values of soil-derived carbon.
The
d
18
O
c
fluctuations, which have been corrected for changes in
the seawater
d
18
O, can be interpreted as changes in the rainfall
d
18
O
values. The
d
18
O of precipitation is mostly related to changes in air-
mass trajectories, rainfall amount, and temperature (Dansgaard,
196 4;Pausata et al., 2011). Our data, however, are insufficient to
speculate changes in the rainfall provenance. By contrast, we can
examine the possible control of temperature and rainfall amount
on the
d
18
O
c
. An increase of surface temperature or a decrease of
rainfall amount could both result in more positive
d
18
O
c
values
(Dansgaard, 1964;Rozanski et al., 1993). If surface temperature
controlled the SV1
d
18
O
c
values, the surface temperature recon-
structed from the
d
18
O
c
anomalies would contradict the observed
positive relationship between lamina thickness and temperature
(Frisia et al., 2003;Fairchild and Baker, 2012), such as from 12.4 to
11.9 ka. In this period, higher surface temperature would corre-
spond to the occurrence of thinner laminae in SV1 (Fig. 5AeC).
Although we discount changes in surface air temperature as the
major driver of the variation in
d
18
O
c
, the observed general agree-
ment between high lamina thickness and negative
d
18
O
c
anomaly
in SV1 could suggest the influence of cave temperature, whose ef-
fect contrasts that of surface temperature (Tremaine et al., 2011).
Fig. 5. The anomalies of the global-ice volume-adjusted
d
18
O
c
and
d
13
C
c
time series of SV1 (grey area) and SV7 (black line). A) The
d
18
O
c
and B)
d
13
C
c
anomalies (deviations from the
mean value). The mean values are as follows: the
d
18
O
c
of SV1 ¼7.1 0.4 and SV7 ¼6.6 0.5; the
d
13
C
c
of SV1 ¼10.4 0.6 and SV7 ¼9.8 0.5. C) The SV1 lamina-thickness
record (grey) and the 9-point running mean (black), plotted on a logarithmic scale. D) SV1 calcite fabrics. Codes are as for Fig. 4. At the top, black and white boxes are indicating the
subdivision into Episodes related to the model for Savi proxy interpretation esee main text for explanation. At the bottom, the chronology (and 2
s
uncertainty) of major isotopic
events as from SV1 and SV7 records is shown.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158152

This hypothesis, however, is not consistent with the observed
opposite pattern between the
d
13
C
c
(positive) and
d
18
O
c
(negative)
anomalies in the period from 12.8 to 12.4 ka. Consequently, the
d
18
O
c
values were more likely controlled by the isotopic signal of
the infiltrations, which includes the effect of rainfall amount,
provenance and seasonality, and possible kinetic fractionation
during pronounced dry periods. By considering the fabric type
during the episodes of positive
d
18
O
c
anomaly, we suggest that the
amount of infiltrating waters was the dominant effect. The con-
flicting pattern of
d
18
O
c
anomaly of SV7 and SV1 during the 12.8e
12.5 ka period is thought to reflect the different structure of the
aquifer, which buffered the climatic signal in one stalagmite with
respect to the other. This singularity will be used in interpreting
Episode 1 (Section 5.2.1).
In the following two sections, we discuss in detail the SV1 and
SV7 proxies throughout Episode 1 and 2. Being temperature and
amount of infiltrating waters the more likely controlling factors of
the
d
13
C
c
and
d
18
O
c
trends, respectively, a change towards more
negative
d
13
C
c
would indicate an increase of temperature while a
change towards more negative
d
18
O
c
values suggests an increase of
infiltrating waters. Although this interpretation should apply also
for SV7, the systematic shift of the
d
13
C
c
and
d
18
O
c
time series to-
wards more positive values than SV1 suggests that SV7 was
consistently affected by disequilibrium fractionation. This system-
atic shift, however, did not compromise the preservation of the
climatic signal in SV7, as confirmed by the comparison to SV1
(Fig. 5AeB). A summary of the proposed interpretation for the SV1
and SV7 proxies is shown in Table S3 (Supplementary Data).
5.2.1. Episode 1
This episode is characterised by a negative
d
18
O
c
anomaly for
SV1, contrasting
d
18
O
c
anomalies for SV7 (Fig. 5A), and positive
d
13
C
c
anomalies in both SV1 and SV7 (Fig. 5B). In addition, the
calcite deposition is extremely reduced on SV1 (Fig. 5C) but is
persistent on SV7 (Fig. 2D). SV1 fabric is dominant open columnar
with dark layers (Fig. 5D). The combination of these features sug-
gests a period of great infiltration variability for SV1, which was not
recorded by SV7. The different flow system feeding the two sta-
lagmites likely influenced their
d
18
O
c
values: under conditions of
high infiltration variability, the slower SV7 drip was probably un-
able to drain the water excess associated with the periods of high
recharge indicated by the SV1 fabric (codes 4e2, Fig. 5D). The SV7
aquifer water excess was re-routed to high capacity paths via
overflow mechanisms (Baker et al., 1997;Miorandi et al., 2010),
causing the loss of climate signal (Fig. 6A). By contrast, the faster
SV1 drip was able to drain the whole infiltrating water. Based on
this premise, therefore, the pronounced negative SV1
d
18
O
c
anomaly of Episode 1, which is the largest throughout the over-
lapping record, reflects a period of higher infiltration amount. It
follows that, the trend of SV1
d
18
O
c
towards less positive anomaly
indicates a decrease in discharge.
The SV1 fabrics (code 2e4) point to high, variable recharge and
inflow of colloidal material and impurities, possibly related to
autumnal flushing (Frisia et al., 2003;Borsato et al., 2007;Baker
et al., 2011). In analogy to Grotta di Ernesto and other mid-
latitude caves, extremely thin laminae in SV1 suggest prolonged
winters (Frisia et al., 2003;Baker et al., 2011), which along with the
input of colloids and impurities, suggest that the soil may not have
been completely protected by litter from erosion during infiltration
events. The coeval positive
d
13
C
c
anomaly of Episode 1 in both SV1
and SV7 supports the inference of a cooling. In summary, from
fabric, lamina thickness and isotope data, we can conclude that
Episode 1 was characterised by long winters (cooling) and infil-
tration episodes in autumn, within a general trend towards
dryness.
5.2.2. Episode 2
Episode 2 is characterised by covariant
d
18
O
c
and
d
13
C
c
isotope
values in both SV1 and SV7 records (Fig. 5AeB). The positive or
negative
d
18
O
c
anomaly allowed us to subdivide the period into 2a
and 2b. In Episode 2a, the persistence of thin laminae is indicative
of long winters (Fig. 5C). The covariant, strongly positive
d
18
O
c
anomalies, suggests low infiltration amount, whereas the SV1 open
columnar fabric with lower accumulation of dark layers relative to
Episode 1, strongly indicates a change in the infiltration pattern,
towards less pronounced seasonality. In conditions of pronounced
dryness, the positive
d
13
C
c
and
d
18
O
c
anomalies may have been
enhanced by greater kinetic fractionation (Mühlinghaus et al.,
2007;Lachniet, 2009).
The infiltration conditions are reversed towards the end of
Episode 2a. The occurrence of dark layers suggests a return to
Fig. 6. A conceptual model of the processes that operate during contrasting hydro-
logical states at Grotta Savi during periods of cooling. A) High effective infiltration. The
fast, fissure-dominated flow of SV1 made the stalagmite climatically sensitive to hy-
drological changes. During periods of high recharge and reduced vegetation cover, very
fine particulates from the soil were carried by SV1 flow and accumulate on the sta-
lagmite top (dark layer). By contrast, the slow, matrix-dominated flow of SV7 was
unable to drain the water excess, which was re-routed (overflow mechanism). This
made SV7 complacent to periods of high aquifer recharge. B) Low effective infiltration.
Particles accumulated in the soil zone. The more responsive SV1 drip declined to its
lowest rate, while the steady SV7 flow was maintained, fed by the matrix reservoir,
making the latter sensitive at low aquifer recharge.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 153

greater seasonality of the infiltrating waters from 12.1 to 11.9 ka.
This change marks a switch in SV1 and SV7
d
18
O
c
towards a less
positive anomaly and corresponds to the abrupt shift of the SV7
d
13
C
c
towards a negative anomaly. By contrast, SV1 shows a delay
in the shift towards a negative
d
13
C
c
anomaly with respect to SV7,
which may indicate more positive
d
13
C values of the DIC. Under
high recharge conditions the weathering of the host rock may be
enhanced, leading to more positive
d
13
C values of drip waters
(Bar-Matthews et al., 2000). Under such conditions, the delay in
the shift of SV1
d
13
C
c
records a possible increase of host-rock
dissolution. Conversely, the rapid shift of
d
13
C
c
values in SV7
may have been caused by a change in the relative contribution of
matrix flow.
In Episode 2b, the
d
18
O
c
and
d
13
C
c
anomalies are negative for
both records. This period in SV1 is marked by predominant
compact fabric, absence of dark layers, and well-developed annual
laminae. The combination of these features indicates a higher
amount of infiltration (particularly soon after the transition from 2a
to 2b) but lower infiltration differences between seasons, as well as
shorter winters relative to Episode 2a. At the same time, the SV7
growth decreased dramatically and definitively ceased at w10.8 ka.
In a period of relatively warmer and wetter conditions, as well as
reduced hydrological variability, the stop of SV7 growth was likely
caused by processes linked to the aquifer structure, such as the
clogging of the feeding system.
To summarise, the comparison of stable isotope, fabric and
growth rate records of the two stalagmites highlights the
complexity of interpreting speleothem proxies. By visually
comparing the two stalagmite records, the
d
18
O
c
variability can be
interpreted as reflecting changes in the amount of infiltrating wa-
ters, while the
d
13
C
c
variability reflects the surface temperature,
with an additional hydrological component during high-infiltration
episodes. In such a context, where decoupling the influences of past
moisture and temperature remains challenging, the use of fabric
types and lamina thickness, as well as the comparison of SV1eSV7
became the key to identify the dominant controlling factor on
d
13
C
c
,
with the temperature component being dominant when calcite is
compact columnar. By using the approach of comparing coeval
stalagmite properties described in this section, we discuss the cli-
matic and hydrologic interpretation of the SV1 record in the in-
terval 15.3e9.3 ka in the following sections.
5.3. Climatic and hydrologic interpretation of SV1 record from
15.3 to 9.3 ka
5.3.1. The Lateglacial (15.3-12.9 ka)
The Lateglacial is recorded only by SV1. Within age un-
certainties, the commencement of SV1 growth (w15.3 ka) co-
incides with the expansion of broadleaf forest into areas formerly
barren, or colonised by conifers, and the first evidence of soil humic
substances in piedmont lake deposits (Monegato et al., 2011). This
suggests that the type of vegetation above the cave influences the
development of stalagmites in Grotta Savi.
The SV1
d
18
O
c
trend towards more negative values (Fig. 7A)
between 15.3 and 11.8 ka suggests a progressive increase of infil-
trating waters. As the dominant fabric type is compact columnar,
the drip-rate variability was not the major controlling factor of the
SV1
d
13
C
c
trend towards 2&more negative values (Fig. 7B), which
at w11.9 ka became similar to the average value for the Holocene in
SV1 (11. 3 0.2&;n¼532, Fig. 4D). Given that the
d
13
C
c
of SV1 is a
proxy of temperature, the Lateglacial
d
13
C
c
millennial-scale trend
encodes the response of the ecosystem above the cave to the post-
glacial warming, which accompanied the increase in moisture. The
most negative peak of SV1
d
13
C
c
values in the pre-Younger Dryas
(YD) record, marks the Allerød within age uncertainty (Fig. 7B). This
reconstruction agrees with the rise of chironomid-based summer
temperature by 2e3

C, reconstructed for the BøllingeAllerød (Be
A) in the European Southern Alp region (Heiri et al., 2007;Larocque
and Finsinger, 2008;Samartin et al., 2012).
A deglacial
d
13
C
c
evolution similar to SV1 characterises other
stalagmites from mid-latitude European caves, such as Chau-stm6
from Chauvet Cave (Fig. 7C) and Vil-11 from Villars Cave (France)
(Genty et al., 2006), and stalagmite CAN from Pindal Caves (Spain)
(Moreno et al., 2010). The 2e3&more negative
d
13
C
c
values of SV1
calcite relative to these other records are likely to be related to its
formation closer to equilibrium conditions (Fairchild and Baker,
2012). The Lateglacial trend of SV1
d
13
C
c
is similar to those of the
candle-shaped stalagmite So-1 from Sofular cave, in Northern
Turkey (Fig. 7D) (Fleitmann et al., 2009). This similar pattern implies
that the
d
13
C
c
values of SV1 and So-1 were governed by a similar
ecosystem process superimposed on local temperature, rainfall
variability, and the nature of the C pool (Brüggemann et al., 2011).
The increase of infiltrating waters recorded at w13.4 ka (more
negative
d
18
O
c
values) was followed by a shift towards colder
conditions (more positive
d
13
C
c
values) (Fig. 7AeB). A sea surface
temperature (SST) cooling starting before the YD is also docu-
mented in the Central and Western Mediterranean Sea (Fig. 7E)
(Cacho et al., 2001), although uncertainty over the reservoir effect
makes a robust comparison difficult.
It has been suggested that Arctic sea-ice melting, caused by the
warming during the BeA, resulted in an increase of moisture
availability, thus increase of precipitation, which enhanced the
freshwater input into oceans. This led to a weakening of the North
Atlantic overturning circulation which, ultimately, resulted in a
cooling (Eisenman et al., 2009). The greater amount of infiltrating
waters and increase of precipitation associated with cooler tem-
peratures towards the end of the B-A, reconstructed from Savi re-
cords, would agree with this hypothesis.
5.3.2. The Younger Dryas (12.8e11.9 ka)
At 12.8 0.3 ka, the lamina thickness of SV1 drastically decreased
(from 17 to less than 3
m
m in the smoothed record, Fig. 5C). It is
worth noting that this sudden decrease in lamina thickness
occurred within a century, based on lamina-counting chronology. In
spite of the relatively large uncertainties of the SV1 age model, we
believe that this lamina reduction represents the beginning of the
YD response in the southern alpine border, which coincides with the
NGRIP chronology of the event, within age uncertainty (Fig. 7F)
(Rasmussen et al., 2006;Svensson et al., 2008).
The combination of SV1 and SV7 proxies indicates a pattern of
high seasonal variability in the infiltrating water, with high
autumnal recharge, and long winters for an earlier YD (cf. Episode 1,
Section 5.2.1). As a consequence of the cooling, a reduced vegeta-
tion cover and more exposed soil resulted in fast infiltrating waters
carrying colloidal material and impurities. Under such conditions,
SV1 drip water was probably barely at saturation, and calcite for-
mation was at its minimum. Independent chironomid data indicate
a drop of circa 1.5

C at the Allerød-YD transition for northern Italy
at the same altitude as Savi (Lake Piccolo di Avigliana, 365 m a.s.l.)
(Larocque and Finsinger, 2008). A summer temperature decrease of
w2

C was also reconstructed from the chironomid record of Lago
di Lavarone (at 1200 m a.s.l., NE Italy) (Heiri et al., 2007). The
extreme reduction in lamina thickness of SV1 suggests a decrease of
winter temperatures.
The YD was characterised by a further deterioration of the
climate (cf. Episode 2a, Section 5.2.2). Conditions of lower recharge,
relatively reduced hydrological variability, and persistence of long
winters are inferred from SV1 and SV7 proxies. This climate
reconstruction suggests that this period coincides with the summer
temperature minima (9

C) reconstructed for the southern Swiss
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158154

Alps at w12.2 cal ka BP (Samartin et al., 2012), when Alpine glacier
extent reached its maximum (Ivy-Ochs et al., 2009).
The most distinctive climate information on the YD emerging
from the Savi records is the hydrological variability. This is evident
by the predominance of autumnal flushing layers over clear calcite
deposition in SV1. Strong seasonality in infiltration (rainfall) could
be the local response to a large-scale reorganisation of atmospheric
circulation. The Meerfelder Maar lake records suggest that the
beginning of the YD in central Europe was characterised by rapid
atmospheric response to the suppression of the Atlantic Meridional
Overturning Circulation (AMOC) and a southward shift of the
westerlies, which could be linked to the structure of the Arctic
Oscillation (AO) (Brauer et al., 2008). Similarities between climate
trends and the AO zonal pattern have been recognised by
Thompson and Wallace (2000). By analogy with recent events, a
negative Arctic Oscillation Index (AOI) following sea-ice melting
and higher moisture in the atmosphere would result in more
moisture available for transport into the Eurasian region (Cohen
et al., 2012). A negative AOI has been linked to a strengthening of
the Siberian High, which in turn has been correlated with low
temperatures in eastern China (Gong et al., 2001). Similarly, an
increase in Siberian High would result in the strengthening of Bora
winds producing cooler, and possibly drier conditions in the Savi
region (Grisogono and Belu
si
c, 2009). The variable infiltrating wa-
ter amount reconstructed for the YD at Grotta Savi, therefore, may
reflected changes in wind strength (McVicar et al., 2012).
The change from condensed layers and thin laminae to more
calcite precipitation and lower seasonality of infiltration (from
Episode 2a to 2b Fig. 5D) marks the end of the YD (11.9 0.2 ka).
The YD to Preboreal transition is characterised by increase of
infiltrating waters and overall warmer winters. In central-northern
Europe, a change towards wetter conditions at circa 12.2 ka has
been inferred from the sedimentation rate at Meerfelder Maar lake
(Lücke and Brauer, 2004). The switch towards wetter and milder
conditions at Savi appears to follow the increase of SSTs recon-
structed from the alkenone records of the central-western
Fig. 7. Comparison between the SV1 and SV7 isotopic records and other palaeoclimate records. B) The
d
13
C
c
time series of SV1 (solid line) and SV7 (dotted line). A) SV1 (solid line)
and SV7 (dotted line)
d
18
O
c
time series adjusted for the global ice-volume effect esee text for explanation. C) Chau-stm6
d
13
C
c
record from Chauvet Cave, France (Genty et al., 2006).
D) So-1
d
13
C
c
time series from Sofular Cave, Turkey (Fleitmann et al., 2009). E) Uk
37
alkenone-based sea-surface temperatures (SST) from the Mediterranean Sea core MD952043
(solid line) and Tyrrhenian Sea core BS7938 (dotted line) (Cacho et al., 2001). F) North Greenland Ice Core Project
d
18
O
ice
(Rasmussen et al., 2006).
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 155

Mediterranean Sea (Cacho et al., 2001)(Fig. 7E) and a weakened
AMOC (Thornalley et al., 2010). In the Savi region, a weakening of
the Bora wind (possibly due to a weakening of the Siberian high-
pressure system) might have played a role in promoting the
climate amelioration.
5.3.3. The Early Holocene
From our Savi records, relatively warmer and humid conditions
characterised the Early Holocene. The SV1 records give evidence of
a further reversal between w10.8 0.2 and 10.1 0.2 ka, which in
SV7 coincides with the cessation of stalagmite growth. The SV1
d
18
O
c
and
d
13
C
c
anomalies became more positive (Fig. 5AeB) while
calcite formation was markedly reduced (Fig. 5C). The fabric sug-
gests an increase of hydrological variability at w10.4 ka (Fig. 5D).
The extremely thin laminae coincide with darker luminescence,
thus evidencing lower concentrations of organic compounds
(Stoykova et al., 2005). Therefore, the Savi records give evidence of
a significant reversal towards cooler conditions, which lasted w700
years.
Climate oscillations in the Early Holocene have been observed in
other archives from the greater Mediterranean region. In the Alps,
chironomid-inferred July temperatures fell w1

C between 10.7
and 10.5 cal ka BP (Heiri et al., 2004).
10
Be exposure ages indicate a
re-advance of Alpine glaciers, with final moraine stabilisation by
10.5 ka (Ivy-Ochs et al., 2009). Arboreal taxa diminished between
11.0 and 10.5 cal ka BP (Filippi et al., 2007;Vescovi et al., 2007).
Similarly, pollen records from marine cores from the Aegean (11.4
and 10.9 ka BP) and Alboran Seas (10.8, 10.3 and 10.0 ka BP) indicate
that several short-term cold and dry fluctuations occurred (Dormoy
et al., 2009). Changes were also observed in pollen and foraminifera
records of the southern Adriatic Sea between 10.8 and 10.4 and
10.0 cal ka BP (Favaretto et al., 2008), and in
d
13
C
c
records of
northern Turkey (10.6 and 10.4 ka Fig. 7D) (Fleitmann et al., 2009)
and eastern rim of Austria (10.0 ka) (Boch et al., 2009).
Although the Lateglacial to Holocene transition was charac-
terised by unstable climate conditions in the greater Mediterranean
region, there is no evidence in Early Holocene records of an event as
marked as that in the SV1. It is likely, therefore, that the magnitude
of the climate anomaly centred at 10.4 ka at Savi was intensified by
local factor(s). A potential cause for this amplification may be
related to a regional atmospheric circulation. Today, occasional
severe events of Bora wind govern temperature and humidity in the
region. In the northern Adriatic, for example, a higher-than-average
pressure difference between the cell of high pressure over Siberia
and low pressure in the Mediterranean in February 2012 (NCEP/
NCAR Reanalysis Data) resulted in Bora wind speeds of 150 km/h,
with the temperature plummeting down to 7

C at sea level
(Servizio Meteorologico Aeronautica Militare, www.meteoam.it). It
can be tentatively speculated that a similar synoptic pattern may
have been dominant for several hundred years in the region during
the Early Holocene, causing the colder and possibly drier conditions
between 10.8 and 10.1 ka.
6. Conclusions
We have presented an approach aimed at recognising the in-
fluence of cave-specific processes on speleothem proxy values and
discriminating the effect of moisture and temperature on the
d
18
O
c
and
d
13
C
c
values. Our approach is based on the assumption that
modern-day monitoring data may not represent conditions of
many thousands of years ago. Therefore, we utilised fabric types
and growth rates to understand the climatic significance of stable
isotope ratio data from two calcite stalagmites (SV1 and SV7). The
physical morphologies and isotope values of the stalagmites have
been related to different drip-rate regimes, which were interpreted
to be governed by different flow structures. By visually comparing
the records of SV1 and SV7, whose chronology was tuned, we
established that the SV1
d
18
O
c
is a proxy for infiltrating water
amount. Our data were, however, not enough to speculate about
changes in rainfall provenance.
Assuming a relationship between temperature, lamina thick-
ness and
d
13
C
c
values similar to that observed at Grotta di Ernesto,
the
d
13
C
c
variability of SV1 should reflect changes in surface
temperature, with a hydrological component during high infiltra-
tion episodes. Fabric and lamina thickness were the keys to
recognise the possible predominance of one factor (hydrology)
over another (temperature) to the
d
13
C
c
. Our reconstruction
highlights, however, that the decoupling of the two variables is
possible only if dealing with two records from speleothems with
different hydrological behaviour. In particular, when the fissure-
fed SV1 displays laminated calcite and compact fabric, we sug-
gest that temperature-related ecosystem processes at the surface
mostly influence the
d
13
C
c
values. By contrast, a hydrological
component must be considered in interpreting the
d
13
C
c
values,
when condensed, less-resolved lamination and open calcite fabric
are observed.
The Savi stalagmite records show that the post-glacial warming
was punctuated by several climate reversals, the most complex and
profound of which was the Younger Dryas (YD). The effect of the
climate change at the BøllingeAllerød to YD transition was
apparent within a century, according to the SV1 lamina-counting
chronology. The unique information extracted from the compari-
son of all proxies in the two stalagmites is that the YD appears to be
characterised by an initial period of higher infiltration variability,
with predominant autumnal flushing, within a trend towards
longer winters. The peak of climate deterioration is marked by dry
conditions, low hydrological seasonality, although still long win-
ters. A change towards wetter and warmer conditions at w11.9 ka
heralded the YD to Early Holocene transition in the region.
The Savi record also provides evidence of a significant w700-
year-long climate anomaly in the Early Holocene (10.8e10.1 ka),
which was characterised by cooling conditions and possibly soil-
moisture deficit, although not as severe as in the YD. Early Holo-
cene climate oscillation(s) have been recognised in other Alpine
and Mediterranean archives, even though less marked than at Savi.
We relate the inferred cooler and drier conditions in the Savi region
between 10.8 and 10.1 ka to a possible strengthening of the Bora
wind system, which may have locally amplified the magnitude of
climate conditions relative to other palaeoclimate records.
The present study is a further demonstration that stable C and O
isotope ratios in speleothems from mid-latitude regions are not
always interpretable as “simple palaeoclimate proxies”(Baker et al.,
2011) and that it is recommended to use several complementary
proxies, particularly from stalagmites with different morphologies.
When monitoring data are not available, we suggest that petro-
graphic observations may offer a low-cost means by which to un-
ravel proxy interpretation in speleothems.
Acknowledgements
The authors are grateful to F. Cucchi, L. Visintin and B. Grillo for
field work (Trieste University, Italy). D. Genty (Laboratoire des
Sciences du Climat et de l’Environnement, Gif-Sur-Yvette Cedex,
France) is acknowledged for providing the Chau-stm6 data. A.
Wainwright (Melbourne University, Australia) is thanked for her
support in sample preparations for UeTh dating. M. Wimmer
(Innsbruck University, Austria) assisted in isotope analytical work
for SV7. Water analyses were performed by F. Corradini (Istituto
Agrario S. Michele all’Adige, Trento). S. Moggio, Y.Y. Sun, O. Rey-
Lescure and R. Offler (Newcastle University, Australia) are
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158156

thanked for contributions to the discussion and technical assis-
tance, thin section preparations, map and paper revision, respec-
tively. The authors are grateful to the valuable comments of G.
Zanchetta and an anonymous reviewer who greatly improved the
manuscript. RB is the recipient of an Endeavour International
Postgraduate Research Scholarship (EIPRS) funded by the Austra-
lian Government and a Ph.D. scholarship from the University of
Newcastle.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.quascirev.2013.04.014.
References
Baker, A., Barnes, W.L., Smart, P.L., 1997. Variations in the discharge and organic
matter content of stalagmite drip waters in lower cave, Bristol. Hydrological
Processes 11, 1541e1555.
Baker, A., Wilson, R., Fairchild, I.J., Franke, J., Spötl, C., Mattey, D., Trouet, V., Fuller, L.,
2011. High resolution
d
18
O and
d
13
C records from an annually laminated Scot-
tish stalagmite and relationship with last millennium climate. Global and
Planetary Change 79, 303e311.
Bar-Matthews, M., Ayalon, A., Kaufman, A., 2000. Timing and hydrological condi-
tions of Sapropel events in the Eastern Mediterranean, as evident from spe-
leothems, Soreq cave, Israel. Chemical Geology 169, 145e156.
Boch, R., Spötl, C., Frisia, S., 2011. Origin and palaeoenvironmental significance of
lamination in stalagmites from Katerloch Cave, Austria. Sedimentology 58,
508e531.
Boch, R., Spötl, C., Kramers, J., 2009. High-resolution isotope records of early Ho-
locene rapid climate change from two coeval stalagmites of Katerloch Cave,
Austria. Quaternary Science Reviews 28, 2527e2538.
Borsato, A., Frisia, S., Fairchild, I.J., Somogyi, A., Susini, J., 2007. Trace element dis-
tribution in annual stalagmite laminae mapped by micrometer-resolution x-ray
fluorescence: implication for incorporation of environmentally significant
species. Geochimica et cosmochimica Acta 71, 1494e1512.
Brauer, A., Haug, G.H., Dulski, P., Sigman, D.M., Negendank, J.F.W., 2008. An abrupt
wind shift in western Europe at the onset of the Younger Dryas cold period.
Nature Geoscience 1, 520e523.
Broecker, W.S., Kennett, J.P., Flower, B.P., Teller, J.T., Trumbore, S., Bonani, G.,
Wolfli, W., 1989. Routing of meltwater from the Laurentide Ice Sheet during the
Younger Dryas cold episode. Nature 341, 318e321.
Brüggemann, N., Gessler, A., Kayler, Z., Keel, S.G., Badeck, F., Barthel, M., Boeckx, P.,
Buchmann, N., Brugnoli, E., Esperschütz, J., Gavrichkova, O., Ghashghaie, J.,
Gomez-Casanovas, N., Keitel, C., Knohl, A., Kuptz, D., Palacio, S., Salmon, Y.,
Uchida, Y., Bahn, M., 2011. Carbon allocation and carbon isotope fluxes in the
plant-soil-atmosphere continuum: a review. Biogeosciences 8, 3457e3489.
Cacho, I., Grimalt, J.O., Canals, M., Sbaffi, L., Shackleton, N.J., Schönfeld, J., Zahn, R.,
2001. Variability of the western Mediterranean Sea surface temperature during
the last 25,000 years and its connection with the northern Hemisphere climatic
changes. Paleoceanography 16, 40e52.
Cheng, H., Edwards, R.L., Hoff, J., Gallup, C.D., Richards, D.A., Asmeron, Y., 2000. The
half-lives of uranium-234 and thorium-230. Chemical Geology 169, 17e33.
Clark, P.U., Shakun, J.D., Baker, P.A., Bartlein, P.J., Brewer, S., Brook, E., Carlson, A.E.,
Cheng, H., Kaufman, D.S., Liu, Z., Marchitto, T.M., Mix, A.C., Morrill, C., Otto-
Bliesner, B.L., Pahnke, K., Russell, J.M., Whitlock, C., Adkins, J.F., Blois, J.L.,
Clark, J., Colman, S.M., Curry, W.B., Flower, B.P., He, F., Johnson, T.C., Lynch-
Stieglitz, J., Markgraf, V., McManus, J., Mitrovica, J.X., Moreno, P.I., Williams, J.W.,
2012. Global climate evolution during the last deglaciation. Proceedings of the
National Academy of Sciences of the United States of America 109, E1134e
E1142.
Cohen, J.L., Furtado, J.C., Barlow, M.A., Alexeev, V.A., Cherry, J.E., 2012. Arctic
warming, increasing snow cover and widespread boreal winter cooling. Envi-
ronmental Research Letters 7, 014007.
Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436e468.
Davis, B.A.S., Brewer, S., Stevenson, A.C., Guiot, J., Contributors, D., 20 03. The tem-
perature of Europe during the Holocene reconstructed from pollen data. Qua-
ternary Science Reviews 22, 1701e171 6.
Dormoy, I., Peyron, O., Combourieu-Nebout, N., Goring, S., Kotthoff, U., Magny, M.,
Pross, J., 2009. Terrestrial climate variability and seasonality changes in the
Mediterranean region between 15 000 and 4000 years BP deduced from marine
pollen records. Climate of the Past 5, 615e632.
Dreybrodt, W., 2008. Evolution of the isotopic composition of carbon and oxygen in
a calcite precipitating H
2
OeCO
2
eCaCO
3
solution and the related isotopic
composition of calcite in stalagmites. Geochimica et Cosmochimica Acta 72,
4712e4724 .
Dreybrodt, W., Scholz, D., 2011. Climatic dependence of stable carbon and oxygen
isotope signals recorded in speleothems: from soil water to speleothem calcite.
Geochimica et Cosmochimica Acta 75, 734e752.
Drysdale, R.N., Hellstrom, J.C., Zanchetta, G., Fallick, A.E., Sánchez Goñi, M.F.,
Couchoud, I., McDonald, J., Maas, R., Lohmann, G., Isola, I., 2009. Evidence for
obliquity forcing of glacial termination II. Science 325, 1527e1531.
Drysdale, R.N., Zanchetta, G., Hellstrom, J.C., Fallick, A.E., McDonald, J., Cartwright, I.,
2007. Stalagmite evidence for the precise timing of North Atlantic cold events
during the early last glacial. Geology 35, 77e80.
Drysdale, R.N., Zanchetta, G., Hellstrom, J.C., Fallick, A.E., Zhao, J.X., 2005. Stalagmite
evidence for the onset of the Last Interglacial in southern Europe at 129 1 ka.
Geophysical Research Letters 32, 1e4.
Eisenman, I., Bitz, C.M., Tziperman, E., 2009. Rain driven by receding ice sheets as a
cause of past climate change. Paleoceanography 24, PA4209.
Fairbanks, R.G., Matthews, R.K., 1978. The marine oxygen isotope record in Pleis-
tocene coral, Barbados, West Indies. Quaternary Research 10, 181e196.
Fairchild, I.J., Baker, A., 2012. Speleothem Science: From Process to Past Environ-
ments. John Wiley & Sons, Oxford, UK.
Favaretto, S., Asioli, A., Miola, A., Piva, A., 2008. Preboreal climatic oscillations
recorded by pollen and foraminifera in the southern Adriatic Sea. Quaternary
International 190, 89e102.
Filippi, M.L., Heiri, O., Arpenti, E., Angeli, N., Bortolotti, M., Lotter, A.F., Van der
Borg, K., 2007. Evoluzione paleoambientale dal Tardoglaciale a oggi ricostruita
attraverso lo studio dei sedimenti del Lago di Lavarone (Altopiano di Folgaria e
Lavarone, Trentino). Studi Trentini Scienze Natutali Acta Geologica 82, 279e
298.
Fleitmann, D., Cheng, H., Badertscher, S., Edwards, R.L., Mudelsee, M., Göktürk, O.M.,
Fankhauser, A., Pickering, R., Raible, C.C., Matter, A., Kramers, J., Tüysüz, O.,
2009. Timing and climatic impact of Greenland interstadials recorded in sta-
lagmites from northern Turkey. Geophysical Research Letters 36, L19707.
Fohlmeister, J., Schroder-Ritzrau, A., Scholz, D., Spotl, C., Riechelmann, D.F.C.,
Mudelsee, M., Wackerbarth, A., Gerdes, A., Riechelmann, S., Immenhauser, A.,
Richter, D.K., Mangini, A., 2012. Bunker Cave stalagmites: an archive for central
European Holocene climate variability. Climate of the Past 8, 1751e176 4 .
Frisia, S., Borsato, A., 2010. Karst, Developments in Sedimentology.
Frisia, S., Borsato, A., Fairchild, I.J., McDermott, F., 2000. Calcite fabrics, growth
mechanisms, and environments of formation in speleothems from the Italian
Alps and Southwestern Ireland. Journal of Sedimentary Research 70,1183e1196 .
Frisia, S., Borsato, A., Pretto, 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,
411e424.
Frisia, S., Borsato, A., Spötl, C., Villa, M.I., Cucchi, F., 2005. Climate variability in the
SE Alps of Italy over the past 17000 years reconstructed from a stalagmite re-
cord. Boreas 34, 445e455.
Frisia, S., Fairchild, I.J., Fohlmeister, J., Miorandi, R., Spötl, C., Borsato, A., 2011. Carbon
mass-balance modelling and carbon isotope exchange processes in dynamic
caves. Geochimica et Cosmochimica Acta 75, 380e400.
Genty, D., Baker, A., Massault, M., Proctor, C., Gilmour, M., Pons-Branchu, E.,
Hamein, B., 2001a. Dead carbon in stalagmite: carbonate bedrocks paleo-
dissolution vs. ageing of soil organic matter. Implication for delta
13
C variations
in speleothems. Geochimica et Cosmochimica Acta 65, 3443e3457.
Genty, D., Baker, A., Vokal, B., 2001b. Intra- and inter-annual growth rate of modern
stalagmites. Chemical Geology 176, 191e212.
Genty, D., Blamart, D., Ghaleb, B., Plagnes, V., Causse, C., Bakalowicz, M., Zouari, K.,
Chkir, N., Hellstrom, J., Wainer, K., Bourges, F., 2006. Timing and dynamics of the
last deglaciation from European and North African
d
13
C stalagmite profiles-
comparison with Chinese and South Hemisphere stalagmites. Quaternary Sci-
ence Reviews 25, 2118e2142.
Genty, D., Blamart, D., Ouahdi, R., Gilmour, M., Baker, A., Jouzel, J., Van-Exter, S.,
2003. Precise dating of DansgaardeOeschger climate oscillations in western
Europe from stalagmite data. Nature 421, 833e837.
GNIP, I.A.E.A./W.M.O., 2006. Global Network of Isotopes in Precipitation. The GNIP
Database. Accessible at: http://www.iaea.org/water.
Gong, D.Y., Wang, S.W., Zhu, J.H., 2001. East Asian winter monsoon and Arctic
Oscillation. Geophysical Research Letters 28, 2073e2076.
Grisogono, B., Belu
si
c, D., 2009. A review of recent advances in understanding the
meso- and micro-scale properties of the severe Bora wind. Tellus A 61, 1e16 .
Hartland, A., Fairchild, I.J., Lead, J.R., Borsato, A., Baker, A., Frisia, S., Baalousha, M.,
2012. From soil to cave: transport of trace metals by natural organic matter in
karst dripwaters. Chemical Geology 304e305, 68e82.
Heiri, O., Filippi, M.L., Lotter, A.F., 2007. Lateglacial summer temperature in the
Trentino area (Northern Italy) as reconstructed by fossil chironomid assem-
blages in Lago di Lavarone (1100 m a.s.l.). Studi Trentini Scienze Natutali Acta
Geologica 82, 299e308.
Heiri, O., Tinner, W., Lotter, A.F., 2004. Evidence for cooler European summers
during periods of changing meltwater flux to the North Atlantic. PNAS 101,
15285e15288.
Hellstrom, J., 2003. Rapid and accurate U/Th dating using parallel ion-counting
multicollector ICP-MS. Journal of Analytical Atomic Spectrometry 18, 1346e
1351.
Hellstrom, J., 2006. UeTh dating of speleothems with high initial
230
Th using
stratigraphical constraint. Quaternary Geochronology 1, 289e295.
Ivy-Ochs, S., Kerschner, H., Maisch, M., Christl, M., Kubik, P.W., Schlüchter, C., 2009.
Latest Pleistocene and Holocene glacier variations in the European Alps. Qua-
ternary Science Reviews 28, 2137e2149.
Kaufmann, G., 2003. Stalagmite growth and palaeo-climate: the numerical
perspective. Earth and Planetary Science Letters 214, 251e266.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158 157

Lachniet, M.S., 2009. Climatic and environmental controls on speleothem oxygen-
isotope values. Quaternary Science Reviews 28, 412e432.
Larocque, I., Finsinger, W., 2008. Late-glacial chironomid-based temperature re-
constructions for Lago Piccolo di Avigliana in the southwestern Alps (Italy).
Palaeogeography, Palaeoclimatology, Palaeoecology 257, 207e223.
Lenaz, D., Potleca, M., Zini, L., 2004. Caratterizzazione mineralogica delle frazioni
argillose nei depositi di riempimento della Grotta G. Savi (Carso Triestino, Ita-
lia). Atti e Memorie della Commissione Grotte “E. Boegan”, Trieste 39, 101e106 .
Lücke, A., Brauer, A., 2004. Biogeochemical and micro-facial fingerprints of
ecosystem response to rapid Late Glacial climatic changes in varved sediments
of Meerfelder Maar (Germany). Palaeogeography, Palaeoclimatology, Palae-
oecology 211, 139e155 .
Mattey, D.P., Fairchild, I.J., Atkinson, T.C., Latin, J.P., Ainsworth, M., Durell, R., 2010.
Seasonal microclimate control of calcite fabrics, stable isotopes and trace ele-
ments in modern speleothem from St. Michaels Cave. In: Gibraltar Geological
Society Special Publication.
McVicar, T.R., Roderick, M.L., Donohue, R.J., Li, L.T., Van Niel, T.G., Thomas, A.,
Grieser, J., Jhajharia, D., Himri, Y., Mahowald, N.M., Mescherskaya, A.V.,
Kruger, A.C., Rehman, S., Dinpashoh, Y., 2012. Global review and synthesis of
trends in observed terrestrial near-surface wind speeds: implications for
evaporation. Journal of Hydrology 416-417, 182e205.
Miall, A.D., 2000. Principles of Sedimentary Basin Analysis. Springer, Berlin; New
York.
Mickler, P.J., Stern, L.A., Banner, J.L., 2006. Large kinetic isotope effects in modern
speleothems. GSA Bulletin 118, 65e81.
Miorandi, R., Borsato, A., Frisia, S., Fairchild, I.J., Richter, D.K., 2010. Epikarst hy-
drology and implications for stalagmite capture of climate changes at Grotta di
Ernesto (NE Italy): results from long-term monitoring. Hydrological Processes
24, 3101e3114.
Monegato, G., Pini, R., Ravazzi, C., Reimer, P.J., Wick, L., 2011. Correlating Alpine
glaciation with Adriatic sea-level changes through lake and alluvial stratig-
raphy. Journal of Quaternary Science 26, 791e804.
Moreno, A., Stoll, H., Jiménez-Sánchez, M., Cacho, I., Valero-Garcés, B., Ito, E.,
Edwards, R.L., 2010. A speleothem record of glacial (25e11.6 kyr BP) rapid cli-
matic changes from northern Iberian Peninsula. Global and Planetary Change
71, 218e231.
Mühlinghaus, C., Scholz, D., Mangini, A., 2007. Modelling stalagmite growth and
d
13
C as a function of drip interval and temperature. Geochimica et Cosmochima
Acta 71, 2780e2790.
Mühlinghaus, C., Scholz, D., Mangini, A., 2009. Modelling fractionation of stable
isotopes in stalagmites. Geochimica et Cosmochima acta 73, 7275e7289.
Not, C., Hillaire-Marcel, C., 2012.Enhanced sea-ice export from the Arctic during the
Younger Dryas. Nature Communications 3, 647.
Oster, J.L., Montañez, I.P., Guilderson, T.P., Sharp, W.D., Banner, J.L., 2010. Modeling
speleothem
d
13
C variability in a central Sierra Nevada cave using
14
C and
87
Sr/
86
Sr. Geochimica et Cosmochima Acta 74, 5228e5242.
Parkhurst, D.L., Appelo, C.A.J., 1999. User’s guide to PHREEQC (Version 2): a com-
puter program for speciation, batch-reaction, one-dimensional-transport, and
inverse geochemical calculations. U.S. Geological Survey Water-Res. Invest. Rep.
Pausata, F.S.R., Battisti, D.S., Nisancioglu, K.H., Bitz, C.M., 2011. Chinese stalagmite
d
18
O controlled by changes in the Indian monsoon during a simulated Heinrich
event. Nature Geoscience 4, 474e480.
Rasmussen, S.O., Andersen, K.K., Svensson, A.M., Steffensen, J.P., Vinther, B.M.,
Clausen, H.B., Siggaard-Andersen, M.L., Johnsen, S.J., Larsen, L.B., Dahl-
Jensen, D., Bigler, M., Röthlisberger, R., Fischer, H., Goto-Azuma, K.,
Hansson, M.E., Ruth, U., 2006. A new greenland ice core chronology for the last
glacial termination. Journal of Geophysical Research D 111, D06102.
Rozanski, K., Araguas-Araguas, L., Gonfiantini, R., 1993. Isotopic patterns in modern
global precipitation. In: Swart, P.K., Lohmann, K.C., McKenzie, J., Savin, S. (Eds.),
Climate Change in Continental Isotopic Records. AGU, Washington, D.C.
Rudzka, D., McDermott, F., Baldini, L.M., Fleitmann, D., Moreno, A., Stoll, H., 2011.
The coupled
d
13
C-radiocarbon systematics of three Late Glacial/Early Holocene
speleothems; insights into soil and cave processes at climatic transitions.
Geochimica et Cosmochimica Acta 75, 4321e4339.
Samartin, S., Heiri, O., Vescovi, E., Brooks, S.J., Tinner, W., 2012. Lateglacial and early
Holocene summer temperatures in the southern Swiss Alps reconstructed using
fossil chironomids. Journal of Quaternary Science 27, 279e289.
Schimpf, D., Kilian, R., Kronz, A., Simon, K., Spötl, C., Wörner, G., Deininger, M.,
Mangini, A., 2011. The significance of chemical, isotopic, and detrital compo-
nents in three coeval stalagmites from the superhumid southernmost Andes
(53

S) as high-resolution palaeo-climate proxies. Quaternary Science Reviews
30, 443e459.
Scholz, D., Hoffmann, D.L., 2011. StalAge ean algorithm designed for construction of
speleothem age models. Quaternary Geochronology 6, 369e382.
Siddall, M., Rohling, E.J., Almogi-Labin, A., Hemleben, C., Meischner, D., Schmelzer, I.,
Smeed, D.A., 2003. Sea-level fluctuations during the last glacial cycle. Nature
423, 853e858.
Spötl, C., 2011. Long-term performance of the Gasbench isotope ratio mass spec-
trometry system for the stable isotope analysis of carbonate microsamples.
Rapid Communications in Mass Spectrometry 25, 1683e1685.
Spötl, C., Vennemann, T.W., 2003. Continuous-flow IRMS analysis of carbonate
minerals. Rapid Communications in Mass Spectrometry 17, 1004e1006.
Stoykova, D., Shopov, Y., Sauro, U., Borsato, A., Cucchi, F., Forti, P., 2005. High-res-
olution luminescence speleothem records from Savi Cave, Trieste, NE Italy. Acta
Geologica 80, 169e173 .
Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-Jensen, D., Davies, S.M.,
Johnsen, S.J., Muscheler, R., Parrenin, F., Rasmussen, S.O., Röthlisberger, R.,
Seierstad, I., Steffensen, J.P., Vinther, B.M., 20 08. A 60 000 year greenland
stratigraphic ice core chronology. Climate of the Past 4, 47e57.
Thompson, D.W.J., Wallace, J.M., 2000. Annular modes in the extratropical circu-
lation. Part I: month-to-month variability. Journal of Climate 13, 1000e1016.
Thornalley, D.J.R., McCave, I.N., Eld erfield, H., 2010. Freshwater input and abrupt
deglacial climate change in the North Atlantic. Paleoceanography 25,
PA1201.
Tremaine, D.M., Froelich, P.N., Wang, Y., 2011. Speleothem calcite formed in situ:
modern calibration of
d
18
O and
d
13
C paleoclimate proxies in a continuously-
monitored natural cave system. Geochimica et Cosmochimica Acta 75, 4929e
4950.
Vescovi, E., Ravazzi, C., Arpenti, E., Finsinger, W., Pini, R., Valsecchi, V., Wick, L.,
Ammann, B., Tinner, W., 2007. Interactions between climate and vegetation
during the Lateglacial period as recorded by lake and mire sediment archives in
Northern Italy and Southern Switzerland. Quaternary Science Reviews 26,
1650e1669.
Zanchetta, G., Drysdale, R.N., Hellstrom, J.C., Fallick, A.E., Isola, I., Gagan, M.K.,
Pareschi, M.T., 2007. Enhanced rainfall in the Western Mediterranean during
deposition of sapropel S1: stalagmite evidence from Corchia cave (Central Italy).
Quaternary Science Reviews 26, 279e286.
Zhou, H., Zhao, J., Qing, W., Feng, Y.-x., Tang, J., 2011. Speleothem-derived Asian
summer monsoon variations in Central China, 54e46 ka. Journal of Quaternary
Science 26, 781e790.
R. Belli et al. / Quaternary Science Reviews 72 (2013) 146e158158