Multiproxy record for the last 4500 years from Lake
Shkodra (Albania/Montenegro)
GIOVANNI ZANCHETTA,1,2,3* AURELIEN VAN WELDEN,4
ILARIA BANESCHI,3
RUSSELL DRYSDALE,5
LAURA SADORI,6
NEIL ROBERTS,7
MARCO GIARDINI,6
CHRISTIAN BECK,8
VINCENZO PASCUCCI9
and ROBERTO SULPIZIO10
1
Dipartimento di Scienze della Terra, University of Pisa, Via S. Maria, 53, 56126 Pisa, Italy
2
INGV sez. Pisa, Pisa, Italy
3
IGG-CNR sez. Pisa, Pisa, Italy
4
Geological Survey of Norway (NGU), Trondheim, Norway
5
Department of Resource Management and Geography, University of Melbourne, Victoria, Australia
6
Dipartimento di Biologia Ambientale, Universita` ‘La Sapienza’, Rome, Italy
7
School of Geography, Earth and Environmental Sciences, University of Plymouth, UK
8
Laboratoire de Ge´odynamique des Chaıˆnes Alpines, Universite´ de Savoie, Le Bourget du Lac, France
9
Dipartimento Scienze Botaniche, Ecologiche e Geologiche University of Sassari, Sassari, Italy
10
Dipartimento Scienze della Terra e Geoambientali, University of Bari, Bari, Italy
Received 25 January 2012; Revised 5 June 2012; Accepted 7 June 2012
ABSTRACT: A multi-proxy record is presented for approximately the last 4500 cal a BP from Lake Shkodra, Albania/
Montenegro. Lithological analyses, C/N ratio and d13
C of the organic and inorganic carbon component suggest that
organic matter and bulk carbonate are predominantly authigenic. The d18
O record of bulk carbonate indicates the
presence of two prominent wet periods: one at ca. 4300 cal a BP and one at ca. 2500–2000 cal a BP. The latter phase is
also found in southern Spain and Central Italy, and represents a prominent event in the western and central
Mediterranean. In the last 2000 years, four relatively wet intervals occurred between ca. 1800 and 1500 cal a BP
(150–450 AD), 1350–1250 (600–700 AD), 1100–800 (850–1150 AD), and at ca. 90 cal a BP (1860 AD). Between
ca. 4100 and 2500 cal a BP d18
O values are relatively high, with three prominent peaks indicating drier conditions at
ca. 4100–4000 cal a BP, ca. 3500 and at ca. 3300 cal a BP. Four additional drier events are identified at 1850 (ca. 100
AD), 1400 (ca. 550 AD), 1150 (800 AD) and ca.750 cal a BP (1200 AD). The pollen record does not show changes in
accordance with these episodes owing to the poor sensitivity of vegetation in this area, which is dominated by an
orographic rainfall effect and where changes in altitudinal vegetation belts do not affect the pollen rain in the lake
catchment. However, since ca. 900 cal a BP a significant decrease in the percentage arboreal pollen and in pollen
concentrations suggest major deforestation produced by human activities. Copyright # 2012 John Wiley & Sons, Ltd.
KEYWORDS: Lake Shkodra; late Holocene; Mediterranean; palaeoclimate; stable isotopes.
Introduction
Future climate predictions suggest that changes in rainfall and
water resources will have important socio-economic and
political impacts over the Mediterranean region (Lionello et al.,
2006). Therefore, understanding the past hydrological variability
in this region is an essential prerequisite for establishing
future climate scenarios. The last ca. 5000–6000 years is
regarded as being particularly relevant because the boundary
conditions of the climate system have not changed substantially
(in comparison with larger glacial–interglacial changes or at the
beginning of the Holocene), and represents the period when an
environment and climate comparable with that of today was
established (e.g. Wanner et al., 2008).
However, in the Mediterranean basin the long history of
human occupation and activities makes it problematic to
discriminate unequivocally between climate and non-climatic
influences on the environment, especially during the mid to late
Holocene (e.g. Roberts et al., 2004, 2011a; Magny et al., 2012).
One means by which to sidestep this problem is through the
application of stable isotope analyses to environmental
archives (e.g. speleothems, lake sediments, pedogenic carbonates),
whose isotopic composition is mostly independent of
direct anthropogenic activity (Roberts et al., 2010). In
particular, stable isotopes on lake deposits can be used to
assess the spatial coherency of the climate changes across the
Mediterranean region, especially given the wide geographical
distribution of lakes. Recently, the number of available records
has increased, allowing the compilation of a regional synthesis
(Roberts et al., 2008). However, many records possess poor
chronologies (Zanchetta et al., 2007) and low resolution (for the
late Holocene in particular there is poor geographic coverage),
implying that the palaeohydrological and palaeoclimatic
information for most of the Mediterranean region is incomplete
and thus poorly understood. Here we seek to redress this
situation by presenting a high-resolution (multi-decadal) multiproxy
record from the freshwater coastal Lake Shkodra (Fig. 1)
for the last ca 4500 cal yr BP.
Site description
Lake Shkodra is a 45-km-long, 15-km-wide, shallow lake (5 m
mean depth) situated at ca. 5 m a.s.l. (Fig. 1). It covers part of a
flat depression surrounded by NW–SE-elongated relief (up to
2750 m a.s.l.). The catchment consists of carbonate rocks
(limestones and dolomites), with minor exposures of siliciclastics,
which supply sediments to the Moraca River, the main
surface inflow to the lake. The only outlet of the lake is the
Bojana River, which flows to the Adriatic Sea. The area is
dominated by a Mediterranean-type climate. Rainfall received
in the lake catchment (Podgorica and Shkodra meteorological
stations) generally ranges between 2000 and 2800 mm aÀ1
but
some areas receive over 3000 mm annually on average.
Summer aridity is pronounced with a mean monthly rainfall
of <50 mm for July. The average annual air temperature is ca.
JOURNAL OF QUATERNARY SCIENCE (2012) 27(8) 780–789 ISSN 0267-8179. DOI: 10.1002/jqs.2563
Copyright ß 2012 John Wiley & Sons, Ltd.
*Correspondence: G. Zanchetta, as above.
E-mail: zanchetta@dst.unipi.it
158C. Occasional flash floods combined with a relatively high
catchment/lake ratio (11:1) and the bathymetry can produce
large (up to 5 m) lake-level fluctuations, such as those observed
between 1956 and 1970 (Lasca et al., 1981; Keukelaar et al.,
2006).
The water residence time in the lake is estimated to be
ca. 120 days (Keukelaar et al., 2006). A significant karst system
with submerged springs contributes to inflow to the lake
(Karaman and Beeton, 1981), which compensates for summer
evaporation. This plus the rapid residence time suggests that
water level is not dominated by evaporative effects. van
Welden et al. (2008) studied the last 500 years of sedimentation
of the lake from two short cores (SK-17 and SK-06) and showed
that this period was characterized by undisturbed sedimentation
of fine-grained carbonate-dominated material, despite
the incidence of destructive earthquakes that affected the area
in 1905 and 1979.
Materials and methods
Several cores were retrieved in September 2003 using a
UWITECTM
platform and coring device. Core SK13 (7.8 m long)
was selected due to its more central position within the lake
and for the evident continuity in sedimentation (Fig. 1).
The chronological framework was established by four
14
C accelerator mass spectrometry measurements (Table 1)
obtained from terrestrial plant macrofossils and on the basis of
the presence of four tephra layers of known age (Sulpizio et al.,
2010; Fig. 2A). The 14
C ages were calibrated using the IntCal
04 curve (Reimer et al., 2004; Table 1). The youngest tephra is
historically documented (the AD 472 ‘Pollena’ eruption, Rosi
and Santacroce, 1983) whilst for the others the best estimated
ages available in the literature have been used (Table 1). The
age model was discussed in detail by Sulpizio et al. (2010) and
for completeness of information is reported in Fig. 2(B). The age
model indicates that the sedimentation rate is, on average,
ca. 0.2 cm aÀ1
. One sample (thickness ca. 1 cm) corresponds to
ca. 5 years.
Samples for stable isotopes on bulk carbonate were collected
every 5 cm (indicating a sample resolution of ca. 30–50 years)
and dried in an oven at 508C for 48 h. Samples were gently
disaggregated and sieved at 100 mm to separate biogenic
remains (e.g. ostracods and shells) from the sediments and then
the fraction below 100 mm was powdered and homogenized. A
preliminary test on ten random samples was performed to verify
the effect on the presence on organic matter. Aliquots of the
homogenized sample were digested for 24 h with H2O2 to
oxidize organic matter. One aliquot was analysed for
13
C/12
C and 18
O/16
O ratios without pre-treatment. Because
we do not find significant differences between treated and
untreated samples, the entire succession was analysed without
further pre-treatment following the recommendation of Wierzbowski
(2007). Herein we refer to the isotopic composition of
these samples as ‘bulk’, and represented by the notation d18
Oc
and d13
Cc. Measurements were made using a GV Instruments
GV2003 continuous-flow isotope-ratio mass spectrometer at
the Advanced Mass Spectrometry Unit, University of Newcastle,
Australia. Samples were digested in 105% phosphoric
acid at 708C. Mass spectrometric measurements were made on
the evolved CO2 gas. Results were normalized to the Vienna
Pee Dee Belemnite scale using an internal working standard of
Carrara Marble (NEW1–Newcastle), previously calibrated
against the international standards NBS18 and NBS19. Mean
analytical precision for both d18
O and d13
C is better than 0.1%.
One aliquot of the sample was treated with 10% HCl to
remove carbonate, then washed several times with deionized
water to neutral pH, and dried again at 408C. The isotopic
composition of bulk organic matter was obtained at the IGGCNR
of Pisa by producing CO2 by combustion using a Carlo
Erba 1108 elemental analyser, interfaced to a Finnigan
DeltaPlusXL via the Finnigan MAT Conflo II interface,
calibrated using a within-run laboratory standard (graphite
and ANU-sucrose). Average analytical reproducibility for these
ADRIATIC
SEA
ALBANIA
MONTENEGRO
SHKODRA LAKE
Bojana
Drin
Moraca
Moraca
SK-13
N
10 km
2m
3m
4m
5
m
6m
7m
8
m
Country boundary
Catchment boundary
48°N
30°N
3
2
4
5
6
1
7
6°O 42°O
Figure 1. Location map of Lake Shkodra and core location (lower
panel) and location of the sites quoted in the text (upper panel): 1, Zon˜ar
Lake; 2, Renella Cave; 3, Accesa Lake; 4, Shkodra Lake; 5, Prespa and
Orhid lakes; 6, Ioannina (Pomvotis); 7, Go¨lhisar Lake.
Table 1. Radiocarbon data, tephra layers and depths from core SK13.
Note that the age of the Agnano Mt Spina tephra (AMS) is reported as the
best estimate performed by Blockley et al. (2008).
Lab. label/Tephra
Core
depth (cm)
Age,
14
C a BP
Age,
cal a BP (2s)
SacA 17 200 Æ 30 305–0
Poz-15211 115 745 Æ 30 728–661
Poz-15212Ã
288 1695 Æ 30 1694–1535
Pollena tephra 300–295 — 14781
FL tephra 514 3150 Æ 602
3215–3480
Avellino tephra 571 3530 Æ 403
3920–3690
Poz-15214 614 3760 Æ 35 4238–3990
AMS tephra 726–716 — 4690–43004
1
Rosi and Santacroce (1983); 2
Coltelli et al. (2000), 3
Zanchetta et al.
(2011), 4
Blockley et al. (2008). Ã
This age has been excluded from the
age model for the obvious incongruence with the historical age of the
tephra.
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
MULTIPROXY RECORD FROM LAKE SHKODRA 781
samples was <0.1%. Concentrations of total carbon (TC), total
organic carbon (TOC) and total nitrogen (TN) were measured
with a Carlo Erba 1108 elemental analyser, with measurements
calibrated against an Acetanilide standard (precision generally
<0.1%). The high-resolution carbonate content of Lake
Shkodra was determined by combining hydrochloric acid to
the laser grain-sizer according to the method of Arnaud (2005).
Randomly selected samples (n ¼ 20) were ground to a fine
powder and analysed by X-ray diffraction for defining the main
mineral component. A few selected samples at different depths
were mounted on glass and observed under a scanning electron
microscope (SEM) coupled to an energy-dispersive X-ray
analyser (Fig. 2C).
The isotopic composition of regional meteoric precipitation
and lake isotope data are necessary for proposing a model for
the interpretation of the oxygen isotope composition of lake
carbonate (e.g. Baroni et al., 2006; Leng et al., 2010b). As there
are no local monitoring programmes for the stable isotopes of
local rainfall and lake waters, we have used the IAEA data
(http://ishohis.iaea.org for Dubrovnik, Komiza-Vis island,
Zadar Malinka-Krk island) and those reported by Longinelli
and Selmo (2003) and Anovski (2001) to obtain a regional
estimate. In addition, lake-water samples were collected
sporadically between 2008 and 2011 (Table 2). The oxygen
isotopic composition was determined by the water–CO2
Figure 2. (A) Lithology, magnetic susceptibility, tephra layers and position of radiocarbon dates for core SK13; (B) age model for core SK13; (C) SEM
image of sediments – scale bar ¼ 10 mm. Note the calcite crystal.
Table 2. Water isotopic composition of Lake Shkodra and the inlet,
the Moraka River.
Date d18
O (%) d2
H (%)
August 2008 (Moraca River) À7.43 À44.1
August 2008 (lake water) À5.84 À33.6
November 2009 (lake water) À6.01 À36.2
June 2010 (lake water) À7.69 À49.9
June 2010 (Moraka River) À9.33 61.7
June 2010 (Moraka River: mouth) À8.70 À56.5
May 2011 (lake water) À7.12 À42.4
May 2011 (lake water) À7.21 À44.3
May 2011 (lake water) À7.25 À44.2
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
782 JOURNAL OF QUATERNARY SCIENCE
equilibration mass spectrometry method at 25 8C using a
Finnigan MAT252 at IGG-CNR. For hydrogen isotope analysis,
waters were reduced to H2 at 450 8C using Mg and analysed
using a Europa Scientific GEO 20-20 mass spectrometry. Water
isotope data are reported with respect to V-SMOW.
Pollen samples were prepared at ca. 10- to 20-cm intervals.
About 0.5 g of dry sediment was chemically treated with HCl
(30%), HF (40%) and hot NaOH (10%). Lycopodium tablets
were added to each sample to calculate the pollen concentration.
More than 50 samples were analysed under the
microscope, with counts of ca. 300 terrestrial pollen grains on
average per sample. The pollen content was very rich and
generally well preserved. All pollen count data were reduced to
the arboreal pollen/non-arboreal pollen (AP/NAP) curve.
Results
Lithology
Core SK13 core is dominated by relatively massive light-brown
calcareous mud (Fig. 2A). X-ray diffraction (XRD) and SEM
analyses on the non-clay fraction indicate that the prominent
mineral phase throughout the core is calcite followed by minor
quartz. The carbonate fraction is mainly composed of calcite
crystals of ca. 10 mm (Fig. 2C), typical of bioinduced
precipitation (e.g. Kelts and Hsu¨, 1978). XRD for some samples
revealed a peak just above background value at 30.80–31.00
2u, which may correspond to the principal peak of dolomite.
An estimate performed with XRD using mixtures of known
amounts of calcite and dolomite suggest that the dolomite
content, if present, should be significantly less than 1%.
Visual inspection and magnetic susceptibility investigations
from the core have previously revealed the presence of four
tephra layers (Table 1), which have been extensively discussed
by Sulpizio et al. (2010). The CaCO3 content shows large
variability (Fig. 3) with the highest values at ca. 1000 cal a BP
(ca. 60%), but mirroring generally the trend of d13
Cc (i.e. when
calcite increases d13
Cc decreases and vice versa, Fig. 3). This is
particularly evident between 3500 and 3000 cal a BP or at ca.
1000 cal a BP where lowest d13
Cc values correspond to highest
CaCO3 content or since 1000 cal a BP when d13
Cc values
increase and percentage carbonate progressively decreases.
Statistically, this negative correlation is very poor considering
the whole record (r2
< 0.2), but it is higher in the first 1000–
1500 years (r2
¼ 0.7 and 0.5, respectively) and between 3200
and 3700 years (r2
¼ 0.5). Three prominent spikes with very low
CaCO3 values are centred at ca. 1430, 2550 and 4350 cal a BP.
The first and the last correspond relatively well to episodes of
tephra deposition (Pollena and Agnano Mt Spina). Although the
macroscopic tephra thicknesses were excluded from the age
model, volcaniclastic sedimentation, settling of the finest tail of
the tephra and bioturbation appear responsible for these spikes
having diluted the carbonate fraction. Excluding these intervals,
the carbonate content record shows five main intervals. After a
first phase with a decreasing trend from ca. 4500 to ca. 4200 cal
a BP, values remain relatively high (ca. 50%) and stable up to
ca. 3700 cal a BP, after which an interval with lower percentage
carbonate occurs, lasting until ca. 1250 cal a BP. In this interval
there are four main phases of significant decrease, two, as
already noted, at ca. 2250 and 1430 cal a BP and the other
centred at ca 1700cal a BP and lasting ca. 200 years. The longer
phase of lower values corresponds to the first part of this interval,
lasting ca. 1000 years (3700–2700 cal a BP). At ca. 1250 there is
a sharp increase in carbonate content reaching the highest values
at ca. 1050 followed by progressively lower values.
TOC (Fig. 3) shows a first phase, ending at ca. 2500 cal a BP,
that is characterized by values between ca. 3 and 6% followed
by an interval displaying a sustained increase and reaching the
highest values of the record at ca. 2100 cal a BP (8.7%). A
decrease follows this peak, reaching a low at ca. 1450 cal a BP,
before rising again. The lowest value occurs at ca. 950 cal a BP.
The last part of the record is fairly constant at around 2%. The
C/N ratio is, on average 8.3 Æ 1.7. From ca. 4500 to 900 cal a
BP they are slightly higher (8.9 Æ 1.0) than the interval after
ca. 900 cal a BP (6.8 Æ 0.6).
Stable Isotope Record
The d18
Oc record shows a variation of more than 2% (from ca.
À8.7% to À6.4%, Fig. 3). From ca. 4500 to 1100 cal a BP the
record shows frequent and large fluctuations. After ca. 1100–
1200 cal a BP the variability progressively decreases and after
ca. 700 cal a BP the values remains close to ca. À7.2%. Two
intervals with, on average, 18
O-depleted values occur between
ca. 4500–4100 and ca. 2600–2000 cal a BP. More 18
Oenriched
intervals occur between ca. 4100 and 2600 cal a BP
and from ca. 2000 cal a BP onward.
The d13
Cc time series shows less short-term variability but a
more complex structure over the long-term, compared with the
d18
Oc record. The d13
Cc values have a range of ca. 2% (from
À5.1% to À3.1%). Between ca. 4500 and 2900 cal a BP the
values are generally higher when compared with the interval
between ca. 2900 and 1700 cal a BP. A decreasing trend
starting at ca. 1400 cal a BP reaches the lowest d13
Cc values at
ca. 1200 cal a BP. Following this, there is a long trend of
increasing values up to ca. 600 cal a BP followed by relatively
invariant values up to the top of the core.
Figure 3. Proxy data from core SK13 plotted versus age (cal a BP).
d18
Oc and d13
Cc correspond to bulk carbonate; d13
Com corresponds to
organic matter; AP, arboreal pollen percentage.
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
MULTIPROXY RECORD FROM LAKE SHKODRA 783
The carbon isotope composition of the organic matter
(d13
Com) oscillates between ca. À30.5 and À26.5% (Fig. 3).
Between ca. 3600 and 2600 cal a BP there is a clear phase
characterized by higher (on average) values, followed by an
interval of lower values at ca. 2500–2100 cal a BP. A phase of
relatively higher values again occurs at 850–300 cal a BP. More
prominent lower values instead occur at the base of the record,
at ca. 4000, 1300 and 900 cal a BP and at the top of the record.
Pollen Data
Here we discuss only the AP and total pollen concentrations.
AP and total concentrations show a relatively high correlation
(r2
¼ 0.52). AP% is mostly greater than 85% up to ca. 900 cal a
BP, after which it decreases down to ca. 70% (Fig. 3). The
record starts with high AP percentages then progressively
decreasing with the first notable reduction at ca. 3100 cal a BP,
after which they increase again to the highest values between
ca. 1950 and 1800 cal a BP. Following this, a sharp decline
occurs, with another prominent reduction at ca. 1450 cal a BP,
followed by a recovery before the final decrease. Similar to
AP%, total pollen concentrations show a strong decline from
ca. 900–1000 cal years BP up to the top of the record. Before
this decline, the concentration values show a first phase of
increase (ca. 4500–4000 cal a BP), followed by a rapid decline
centred at ca. 3900 cal a BP. After a short period of recovery
(centred at ca. 3600 cal a BP), there is an interval with lower
values between ca. 3500 and 2800 cal a BP, with the lowest
values of this interval at ca. 2900 cal a BP. Between ca. 2700
and 1900 cal a BP there is an interval characterized by the
highest concentration values, although large fluctuations are
observable (e.g. the sharp declines at ca. 2550 and 2100 cal a
BP). From ca. 1900 cal a BP there is a long decline of pollen
concentration, reaching one of the lowest values at ca. 1450 cal
a BP, and after a rapid recovery lasting about 100 years there is
a final dramatic decline in total pollen concentration.
Discussion
Oxygen Stable Isotope Interpretation
The possible presence of trace amounts of dolomite suggests
clastic contamination from the lake catchment. Significant
input of clastic carbonates limit the use of stable isotope data
from the bulk fraction as a palaeolimnological proxy (e.g.
Leng et al., 2010a). However, we note that: (i) if any, the
contamination is relatively minor, as suggested by XRD ((1%
or below the detection limit); (ii) there is no covariance
between d18
Oc and d13
Cc, which could be expected if
contamination is significant and its input varies through time
(Hammarlund and Buchardt, 1996), even if a positive
correlation between clastic contaminant and isotope composition
it is not always obvious (see, for instance, the data for
Lake Pamvotis, discussed by Leng et al., 2010a); (iii) there is a
crude negative association (statistically low for the whole
record) between d13
Cc and carbonate content (Fig. 3) – this
would be difficult to attribute to high bedrock carbonate
contamination in the absence of a positive correlation of d18
Oc
and d13
Cc; (iv) d13
Cc and d13
Com show, for most of the record, a
positive association (again not statistically significant),
suggesting that both organic matter and carbonate originate
from within the lake; and (v) the C/N ratio (on average <10)
suggests that organic matter has originated mainly from within
the lake (e.g. Meyers and Lallier-Verge`s, 1999): if a large
amount of clastic carbonate was present it should have been
washed into the lake accompanied by terrestrial organic matter,
the ratio of which would be higher.
The few water samples collected in the lake plot along the
regional meteoric water line (Fig. 4), indicating that lake waters
are, if anything, minimally affected by evaporation. This is not
completely surprising owing to the very short residence time of
the lake water and the presence of submerged springs that
undoubtedly dampen the effects of evaporation during the
summer months. Preliminary data on diatoms (N. Roberts,
unpublished data) indicate that freshwater conditions occurred
throughout the investigated period.
The large lake-level variability monitored in the last few
decades (Keukelaar et al., 2006) and the relatively large
oscillations observed in the d18
Oc record suggests that the
system is sensitive to hydrological changes, but the isotopic
balance is not affected by local evaporation. Thus, the isotopic
composition of the lake water is likely to respond mainly to
changes in the isotopic composition of the inflow waters, which
in turn are a proxy for the isotopic composition of meteoric
precipitation in the catchment.
The analyses of rainfall isotopic data from the region show a
complex behaviour (Fig. 5), and the data collected by Anovski,
(2001) for Lake Ohrid (Fig. 1) can be considered, as a first
approximation, representative of the recharge area of the
Moraca River (Fig. 1), as shown in Fig. 4. Although the
temperature effect and the relationship with the amount of
precipitation are not particularly strong, there is a tendency for
rainfall with higher isotopic composition during the summer at
the time of higher temperatures and lower precipitation; the
opposite prevails during winter (Fig. 5). Spring and autumn
have intermediate values. Therefore, these figures indicate that
lake water could be forced towards lower isotopic values
during cool periods, especially during autumn/winter/early
spring, when accompanied by higher precipitation: in other
words, a low summer-to-non-summer precipitation ratio. By
contrast, a shift towards a higher ratio of summer-to-nonsummer
precipitation, and/or higher summer temperatures,
should favour heavy-isotope enrichment in the lake waters. As
most of the precipitation in the area occurs during the winter
months, increased winter recharge may produce an average
decrease of lake water d18
O, and vice versa. This should be
partially affected by the change in the fractionation factor of
Figure 4. d2
H versus d18
O for the area under study. Monthly rainfall
data are from http://ishohis.iaea.org, Longinelli and Selmo (2003) and
Anovsky (2001).
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
784 JOURNAL OF QUATERNARY SCIENCE
calcite–water with temperature at the time of calcite precipitation
(usually during the warmer part of the year, e.g. Leng and
Marshall, 2004). Interestingly, many Mediterranean stations
show a d18
Op/T of ca 0.2% 8CÀ1
(e.g. Bard et al., 2002; Zoto,
2006), including the data set shown in Figs. 4 and 5 (0.22%
8CÀ1
), even if the correlation for these stations is not particularly
high (r2
¼ 0.28), partly due to the relatively short period of
measurements. Because the fractionation factor for carbonate
precipitating close to isotopic equilibrium dictates a d18
Oc/T
gradient of ca. À0.2% 8CÀ1
in the range ca. 10–30 8C (e.g. Kim
and O’Neil, 1997), this indicates that the temperature effect on
rainfall and on calcite precipitation could cancel out each other
(e.g. Bard et al., 2002). Therefore, the d18
Oc record in Lake
Shkodra is potentially a near-pure signal of the change in the
isotopic composition of the recharge water, given the effect of
the short residence time and the presence of recharge by
submerged springs. The changes in the isotopic composition of
d18
Oc should record changes in the lake’s seasonal recharge
patterns, although this can be complicated by altered stormtrack
trajectories, which themselves could alter rainfall isotope
composition (e.g. Celle-Jeanton et al., 2001).
The most striking feature of the d18
Oc record is the long-term
trend showing two prominent periods of wetter conditions
centred at ca. 4300 and 2100 cal a BP, which are followed
rapidly by d18
Oc increases indicative of drier conditions. The
drier phase between ca. 4100 and 2500 cal a BP is punctuated
by several oscillations suggesting unstable hydrological
conditions. Particularly prominent are drier conditions occurring
at ca. 4100–4000, 3500 and 3300 cal a BP. The following
wetter period lasted ca. 500 years and ended abruptly at
ca. 2000 cal a BP, and was rapidly followed by a short period at
ca. 1850 cal a BP (ca. 100 AD) of drier conditions.
Three evident short phases of wetter conditions occurred
between ca. 1800 and 1500 cal a BP (150–450 AD, spanning
the Roman Empire), 1350–1250 cal a BP (600–700 AD, at the
beginning of the Dark Ages), 1100–800 cal a BP [850–1150,
during the Middle Ages – possibly representing the so-called
Medieval Climate Anomaly (MCA)] and at ca. 90 cal a BP. Four
prominent drier events occur at ca. 1850 cal a BP (ca. 100 AD),
1400 cal a BP (ca. 550 AD), 1150 cal a BP (ca. 800 AD), and a
longer period of relatively high d18
O values lasting from ca. 780
to 200 years a BP, at ca. 750 cal a BP (ca. 1200 AD). This last
period covers most of the Little Ice Age.
Interpretation of carbon isotope composition of
organic matter and carbonate, C/N ratio and
carbonate content
The d13
Cc and carbon isotope composition of organic matter
(d13
Com) are broadly covariant (Fig. 3), supporting the notion
that organic and inorganic carbon originated from the same
dissolved inorganic carbon pool (DIC; Hollander and McKenzie,
1991). The average difference between d13
Cc and d13
Com is
24.9 Æ 0.8, in agreement with fractionation from the same DIC
source (e.g. Hollander and McKenzie, 1991). The fact that this
difference does not vary significantly implies that neither the
eutrophic state of the lake nor the level of CO2 consumption
varied significantly (Hollander and McKenzie, 1991; Mayer
and Schwark, 1999). The C/N ratio largely matches this
interpretation, indicating that lacustrine organic matter prevails.
After ca. 900 cal a BP the C/N ratio decreases, suggesting
that the lake sediment shows increased proportion of algal
matter with a lower C/N ratio compared with that of vascular
lacustrine plants, or a concurrent decrease in the terrestrial
input, if present (e.g. Meyers and Lallier-Verge`s, 1999). The
d13
Cc and carbonate content, with an apparent negative
correlation for some intervals (Fig. 3), may suggest, however,
that the amount of primary productivity changed in the lake.
This is supported by the TOC data.
Factors influencing the isotopic composition of DIC in karstic
lakes are many and varied (Hollander and McKenzie, 1991;
Mayer and Schwark, 1999; Leng and Marshall, 2004). Usually,
13
C-depleted intervals reflect an increased input of CO2 derived
by oxidation of organic matter whereas higher values suggest
increasing equilibration with atmospheric CO2 and/or hard
waters from the catchment (i.e. those strongly affected by
dissolution of old carbonates). Rivers normally carry lower
d13
C DIC values coming primarily from the leaching of soil CO2
(e.g. Mook and Tan, 1991), which can change in lakes due to
a progressive equilibration with atmospheric CO2 as well as
plant photosynthesis, respiration and microbially mediated
carbon cycling (e.g. Hollander and Smith, 2001). However,
karst springs can carry CO2 derived from dissolution of old
marine carbonate, which is usually more 13
C-enriched than
‘respired’ CO2. Intervals characterized by lower d13
Cc and
d13
Com at Shkodra could indicate a persistent input of CO2
originating from soil CO2 via river flow, whereas intervals
Figure 5. Isotopic composition from different stations for the region showing the dependence of isotopic composition on temperature and
precipitation. Data are from http://ishohis.iaea.org, Longinelli and Selmo (2003) and Anovsky (2001).
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
MULTIPROXY RECORD FROM LAKE SHKODRA 785
characterized by higher values may indicate increased
equilibration with the atmosphere and more 12
C consumption
by biological activity and/or relatively more recharge by
submerged springs.
Pollen record
Pollen results indicate that the physiognomy of the vegetation
was not overly affected by the hydrological changes inferred by
the oxygen isotopes, with percentage changes in AP showing
rather moderate variability, ranging mainly from 84% to 94%
from ca. 4500 to 900 cal a BP. This is not surprising taking into
account the orography of the area and the disposition of the
altitudinal vegetation belts. Pollen concentration values, by
contrast, suggest that important decreases in forest biomass
have occurred. However, a general tendency for pollen
concentration and AP% to decrease follows the general trend
of increasing d18
O, possibly suggesting progressive drying
(Fig. 3).
The first significant pollen concentration minimum occurs at
ca. 4000 cal a BP, close to but not exactly matching the oxygen
isotope peak, followed by a minimum at ca. 3300 cal a BP in
good agreement with the d18
Oc record. A further important
minimum occurs at ca. 2900 cal a BP, which does not correlate
with the d18
Oc record. At ca. 1400 cal a BP a significant
decrease is well matched by a drier phase identified in the
d18
Oc record. The start of the prominent decline of both AP%
and pollen concentration at ca. 900 cal a BP seems likely to
have been generated by phases of deforestation due to human
activity starting in the Middle Ages. A first relatively large
increase in pollen concentrations matches the arid phase in the
d18
Oc record, abruptly followed by the minimum cited above.
This suggests that the pollen response lags the isotope response,
notwithstanding the different resolution of these two proxies.
The maximum at ca. 3600 cal a BP is in agreement with a
relatively wetter phase indicated by isotope data.
Perhaps one of the most interesting features is the interval
containing the highest concentration values that corresponds to
the wetter period identified in the d18
Oc record. The highest
values are reached just prior to the abrupt end of this interval.
The relative maximum centred at ca. 1300 cal a BP is again
close to a wet phase as suggested by the d18
O.
Regional Significance of the Record
The oxygen isotope record in SK13 is one of the most robust and
informative proxies from a palaeohydrological point a view, as
the other proxies seem to indicate long-term trends that
are difficult to interpret owing to the relatively sparse
information on modern lake behaviour. Pollen percentages
seem strongly influenced by the orography of the area: any
climatic change is probably accommodated by changes in
the altitudinal position of the vegetation belts but without
wholesale changes in the pollen rain in the lake, even if a
progressive decrease in the pollen concentration and AP%
since 2000 years a BP (mirroring the progressive increase of the
d18
Oc values) may indicate a progressive drying. However,
total pollen concentration appears to be more informative than
AP%, correlating better with d18
O. Yet this correlation is not
entirely consistent, suggesting that the vegetation may have a
threshold of hydrological sensitivity somewhat different to the
lake d18
Oc.
Overall, the interval between ca. 4200 and 2500 cal a BP is
one of higher d18
Oc bracketed by two intervals characterized
by lower d18
Oc values. At Lake Malik (Albania), Fouache et al.
(2010) described a phase between ca. 4000 and 2600 cal a BP
of lower lake levels with deposition of peat. A similar pattern
was discussed by Roberts et al. (2011b) for stable isotope
records in lakes of the eastern Mediterranean, notwithstanding
differences in the quality of the various chronologies. In
particular, the isotopic record of Lake Go¨lhisar (Eastwood et al.,
2007) shows two humid phases separated by a prominent dry
event centred at ca. 3500 cal a BP.
By contrast, the interval between ca. 2500 and 2000 cal a BP
with lower d18
Oc values indicates wetter conditions. This
interpretation is also supported by the pollen data. We note that
during this interval there is an increase in TOC accumulation,
which could indicate increased preservation of organic matter.
This interval is also coincident with a decrease in d13
Cc. Lake
level at Lake Malik was higher between 2600 and 2000 cal a BP
(Fouache et al., 2010), whilst at Lake Accesa (Tuscany; Magny
et al., 2007) level was higher between ca. 2800 and 2000 cal a
BP (Fig. 6). Most notably, this period coincides with the highest
phase of lake level and amount of precipitation in southern
Spain reconstructed in Zon˜ar Lake (Martı´n-Puertas et al., 2010).
In Fig. 6 the Zon˜ar Lake Rb/Al ratio is interpreted as a proxy for
runoff input to the lake (Martı´n-Puertas et al., 2010). We also
note that this interval corresponds to a phase of lower
haematite-stained grain (HSG) deposition in subpolar cores
of the North Atlantic (Fig. 6).
At the base of the SK13 record we see evidence for a humid
phase centred at ca. 4300 cal a BP and lasting about 200 years.
Although not completely captured in its entirety, analogies can
be found in other records in the central Mediterranean. A multiproxy
speleothem record from Buca della Renella (Italy)
(Drysdale et al., 2006) indicates wetter conditions around this
time. Magny et al. (2009, and references therein) discuss a wet
phase at ca. 4500–4300 cal a BP, including a period of
increased floods in Spain, and soil development in Tunisia. At
Lake Malik (Albania), sedimentological evidence indicates high
lake levels between ca. 4200 and 4100 cal a BP (Fouache et al.,
2010).
At ca. 4000–4100 cal a BP there is a prominent arid event.
Our chronology is substantially in agreement with that of the
Buca della Renella multi-proxy record (Drysdale et al., 2006)
and this arid phase has been identified in the aforementioned
record of Spanish flood history (Thorndycraft and Benito, 2006)
and soil formation in Tunisia (Zielhofer and Faust, 2008). Many
pollen records of central Italy probably also capture this event
(e.g. Sadori et al., 2004, 2011). According to Di Rita and Magri
(2009), this event is consistent with an increase in summer
drought with progressive expansion of the North Africa
anticyclone zone over the central Mediterranean Basin.
Additionally, the Shkodra record may further suggest a decrease
of rainfall during the colder months. It is important to note that
this event is well recorded outside the Mediterranean at lower
and higher latitudes (Drysdale et al., 2006). It is also identified
in the HSG record from subpolar cores of the North Atlantic
(Bond et al., 2001), even if it does not appear among the most
prominent of such events (Fig. 6).
Two further significant phases of drier conditions occur at ca.
3500 and 3300 cal a BP. These events correspond to a peak in
the HSG record and the phase of cooling recorded in core LC21
in the Aegean Sea following the deposition of the Thera tephra
(Rohling et al., 2002). In the alluvial deposits around the
ancient city of Gibala (coastal Syria) Bretscheider and Van
Lerbeghe (2008) found pollen evidence for a severe drought
between ca. 3.3 and 2.7 cal ka BP, and Kaniewski et al. (2010)
have suggested that this arid phase may have produced regionwide
crop failure corresponding to the Late Bronze Age
collapse.
Four shorter phases of wetter conditions occurred at
ca. 1800–1500 cal a BP (150–450 AD, during the Roman
Empire), 1350–1250 (600–700 AD, at the beginning of the Dark
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
786 JOURNAL OF QUATERNARY SCIENCE
Ages), 1100–800 (850–1150 AD, during the Middle Ages –
possibly representing the MCA – Bradley et al., 2003) and
centred at ca. 90 cal a BP, i.e. at ca. 1860 AD. According to the
interpretation proposed under our age model, the MCA is
represented by a slightly wetter interval. It is interesting to note
that this period does not show any prominent feature compared
with the last 2000 years of our record. However, the accuracy
of our age model is questionable for the last ca. 1000 years and
this interval needs to be better resolved in future studies using,
for example, other cores which may contain this interval at a
higher resolution (van Welden et al., 2008).
Four prominent drier events occur at ca. 1850 (100 AD),
1400 (550 AD), 1150 (800 AD) and ca. 750 cal a BP (1200 AD).
The event at ca. 1150 cal a BP is in good agreement (within age
errors) with an arid period between ca. 1200 and 1100 cal years
BP reported by McMillan et al. (2005) from speleothems in two
caves of southern France. Still within age error is the prominent
arid event that can be inferred from the d18
O of bulk carbonate
from Lake Ohrid (Leng et al., 2010b). The drier conditions
which appear to extend from ca. 780 to 200 cal a BP with the
higher oxygen isotope values centred at ca. 750 cal a BP cover a
significant part of the Little Ice Age.
An interesting feature of the interval between 2000 and
300 cal a BP is the progressive reduction in magnitude of the
d18
Oc excursions. This may correspond to a progressive, longterm
aridification, supported somewhat by pollen data, even if
we cannot rule out other contributing factors: for instance, the
residence time of the lake water due to damming of the outlet,
or the lake ‘catchment’ effect triggered by a change in land use
due to human activity, as suggested by the pollen data for the
last 900 years. Additionally, human impacts could have
promoted catchment erosion, increasing clastic carbonate
input to the lake. van Welden et al. (2008), on the basis of grainsize
and magnetic data from cores SK-17 and SK-06, reported a
substantial increase in clastic input during the Little Ice Age.
This may have increased input of clastic carbonate to the lake,
dampening the isotopic signature of in situ-precipitated
CaCO3. However, the possible increase in clastic carbonate
input during the last 2000 years is not supported by the trend in
d13
Com, which shows a decrease in the last 2000 years, and the
C/N ratios, which suggest an increase of algal organic matter in
the last 1500 years, and finally %CaCO3, which shows a
decrease in the last 1000 years BP. In addition, the d13
Com
record reveals that the interval with the lowest values was
40003000200010000
Age (years BP)
-8.4
-7.8
-7.2
δ18O(‰PDB)
-12
-6
-4.5
-3
δ18O(‰PDB)
-3
-2
-1
0
1
2
δ18O(‰PDB)
-4.2
-3.6
-3
δ18O(‰PDB)
0
2
4
6
LakeLevel(m)
4
8
12
HRD(%)
0.2
0.24
Mg/Al
1.2
1.8
2.4
3
Rb/Al
SHKRODA LAKE
PRESPA LAKE
RENELLA CAVE
ACCESA LAKE
IOANNINA LAKE
VM29-191
MC52
ODP 976
ZOÑAR LAKE
PALEOSOIL
DRIER
WETTER
DRIER
WETTER
DRIER
WETTER
DRIER
WETTER
LOW
HIGH
WETTER
DRIER
Figure 6. Comparison among different
proxies discussed in the text and the d18
O
record of Lake Shkodra. Zon˜ar Lake Rb/Al
record: Martı´n-Puertas et al. (2010); marine
core ODP 976 Mg/Al record: Jimenez-Espejo
et al. (2008); marine cores VM29-191 and
MC52 HSG (%) record: Bond et al. (2001);
Accesa Lake level record: Magny et al.
(2007); Renella cave d18
O record: Drysdale
et al. (2006); Ioannina Lake d18
O record:
Frogley et al. (2001); Prespa Lake d18
O
record on bulk carbonate: Leng et al.
(2010b); Skhodra d18
O record on bulk
carbonate fraction (this study). The Shkodra
record is three-point averaged compared
with Fig. 3.
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
MULTIPROXY RECORD FROM LAKE SHKODRA 787
between ca. 1500 and 500 cal a BP. Taken together, these
patterns are difficult to interpret in terms of increases in the
terrestrial input of carbonate into the lake.
Figure 6 also shows d18
Oc records from Lakes Prespa (Leng
et al., 2010b) and Lake Ioannina (Pamvotis) (Frogley et al.,
2001), compared with the Lake Shkodra record. The three
records show evident arid/wet/arid oscillations between
ca. 2000 and 1000 cal a BP, although differences in the
resolution and chronology do not provide sufficient evidence
for correlating these oscillations. In addition, the Prespa record
shows a clear peak with higher d18
Oc, which is nearly
coincident with the similar event at ca. 4100 cal a BP in
Shkodra.
Although a recent synthesis of studies suggests the existence
of a regional pattern of hydrological variability (e.g. Roberts
et al., 2008, 2011b) it is clear that a fine-scale of resolution and
excellent chronological control are necessary for improving our
knowledge of climatic changes over the Mediterranean Basin,
including the Balkans. However, these correlations may
indicate that the Shkodra d18
O record is not representative
for the Balkans in general because more agreement is found
with records outside of the study area (e.g. Go¨lhisar and Zon˜ar
lakes).
Conclusion
This study has shown that Lake Shkodra can potentially help
with reconstruction of the palaeohydrological evolution of
the region thanks to the high sedimentation rate, with high
authigenic carbonate component, its rapid residence time and
low evaporative effect on the water isotopic composition. In
particular, the Shkodra d18
Oc record indicates the presence of
two main wet phases: one centred at ca. 4300 cal a BP, and one
between ca. 2500 and 2000 cal a BP. The latter phase is well
expressed in southern Spain, Central Italy and, tentatively, in
Turkey, and possibly represents a prominent event in the
Mediterranean. Four minor wetter intervals have occurred over
the last 2000 years: between ca. 1800 and 1500 cal a BP (150–
450 AD), 1350–1250 (600–700 AD), 1100–800 (850–1150
AD), and, less so, at ca. 90 cal a BP. Between ca. 4100 and
2500 cal a BP the d18
Oc values are relatively high, with three
prominent peaks indicating drier conditions at ca. 4100–4000,
ca. 3500 and centred at ca. 3300 cal a BP. The event found at
Shkodra at 4100–4000 cal a BP confirms this event as a
prominent one in Mediterranean records. Four additional drier
events were identified at ca. 1850 (100 AD), 1400 (550 AD),
1150 (800 AD) and ca. 750 cal a BP (1200 AD). The MCA may
have been wetter, but is not particularly prominent when
compared with the last 2000 years. On the other hand, part of
the Little Ice Age seems to have been drier, with wetter
conditions towards the end, but better resolution and dating for
this interval is necessary to substantiate this. Since ca. 900 cal a
BP, human impact becomes more apparent in the pollen data,
and may also have affected evolution of the lake.
Acknowledgements. Shkodra lake coring was funded by a NATO
Science for Peace project (SFP 977 993), which was initiated and
managed by F. Jouanne. A.W. and C.B. thank their colleagues who
made the coring possible, especially J. L. Mugnier, R. Koci and S.
Bushati. This research was also partially supported by INGV-DPC
2005–2007 projects (V-3 subproject on explosive activity of SommaVesuvius
– Tephra dispersion, Leaders R. Santacroce and G. Zanchetta).
We thank M. Guidi and L. Dallai (IGG-CNR) for assistance during the
analytical work. C. Martin-Puertas is acknowledged for Zon˜ar lake data
and for useful discussion. We thank M. Magny and an anonymous
reviewer for useful suggestions, which greatly improved the quality of
the manuscript.
Abbreviations. AP, arboreal pollen; DIC, dissolved inorganic
carbon; HSG, haematite-stained grain; MCA, Medieval Climate
Anomaly; NAP, non-arboreal pollen; SEM, scanning electron
microscopy; TC, total carbon; TN, total nitrogen; TOC, total
organic carbon; XRD, X-ray diffraction.
References
Anovski T (ed). 2001. Progress in the Study of Prespa Lake using
Nuclear and Related Techniques, Project Report, IAEA Regional
Project RER/8/008. Skopje, Macedonia.
Arnaud F. 2005. Discriminating bio-induced and detrital sedimentary
processes from particle size distribution of carbonates and noncarbonates
in hard water lake sediments. Journal of Paleolimnology
34: 519–526.
Bard E, Delaygue G, Rostek F, et al. 2002. Hydrological conditions in
the western Mediterranean basin during the deposition of Sapropel 6
(ca. 175 kyr). Earth and Planetary Science Letters 202: 481–494.
Baroni C, Zanchetta G, Fallick AE, et al. 2006. Mollusca stable isotope
record of a core from Lake Frassino, northern Italy: hydrological and
climatic changes during the last 14 ka. The Holocene 16: 827–837.
Blockley SPE, Bronk Ramsey C, Pyle DM. 2008. Improved age modelling
and high-precision age estimates of late Quaternary tephras, for
accurate palaeoclimate reconstruction. Journal of Volcanology and
Geothermal Research 177: 251–262.
Bond G, Kromer B, Beer J, et al. 2001. Persistent solar influence on
North Atlantic climate during the Holocene. Science 294: 2130–
2136.
Bradley RS, Hughes MK, Diaz HF. 2003. Climate in medieval Time.
Science 302: 404–405.
Bretscheider J, Van Lerbeghe K. 2008. Tell Tweini, ancient Gibala,
between 2600 BCE and 333 BCE. In In search of Gibala, an archaeological
and historical study based on eight seasons of excavations at
Tell Tweini (1999–2007) in A and C fields, In Bretscheider J, Van
Lerbeghe K 2008. (Eds.) Aula Orientalis: Barcelona; pp. 12–66.
Celle-Jeanton H, Travi Y, Blavoux B. 2001. Isotopic typology of the
precipitation in the Western Mediterranean region at three different
time scales. Gophysical Research Letters 28: 1215–1218.
Coltelli M, Del Carlo P, Vezzoli L. 2000. Stratigraphic constraints for
explosive activity in the past 100 ka at Etna volcano, Italy. International
Journal of Earth Science 89: 665–677.
Di Rita F, Magri D. 2009. Holocene drought, deforestation and evergreen
vegetation development in the central Mediterranean: a 5500
year record from Lago Alimini Piccolo, Abulia, southeast Italy. The
Holocene 19: 295–306.
Drysdale R, Zanchetta G, Hellstrom J, et al. 2006. Late Holocene
drought responsible for the collapse of Old World civilizations is
recorded in an Italian cave flowstone. Geology 34: 101–104.
Eastwood WJ, Leng MJ, Roberts N, et al. 2007. Holocene climate
change in the eastern Mediterranean region: a comparison of stable
isotope and pollen data from Lake Go¨lhisar, southwest Turkey.
Journal of Quaternary Science 22: 327–341.
Fouache E, Desruelles S, Magny M, et al. 2010. Palaeogeographical
reconstructions of Lake Maliq (Korc¸a Basin, Albania) between
14,000 BP and 2000 BP. Journal of Archaeological Science 37:
525–535.
Frogley MR, Griffiths HI, Heaton THE. 2001. Historical biogeography
and Late Quaternary environmental change of Lake Pamvotis, Ioannina
(north-western Greece): evidence from ostracods. Journal of
Biogeography 28: 745–756.
Hammarlund D, Buchardt B. 1996. Composite isotope records
from Late Weichselian lacustrine sequence at Grænge, Lolland,
Denemark: evidence of Allerør and Younger Dryas environments.
Boreas 25: 8–22.
Hollander DJ, McKenzie JA. 1991. CO2 control on carbon-isotope
fractionation during aqueous photosynthesis: a paleo-pCO2 barometer.
Geology 19: 929–932.
Hollander DJ, Smith MA. 2001. Microbially mediated carbon cycling as
a control on the d13
C of sedimentary carbon in eutrophic Lake
Mendota (USA): new models for inpterpreting isotopic excursions
in the sedimentary record. Geochimica et Cosmochimica Acta 65:
4321–4337.
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
788 JOURNAL OF QUATERNARY SCIENCE
Kaniewski D, Paulissen E, Van Campo E, et al. 2010. Late second-early
fist millennium BC abrupt climate changes in coastal Syria and their
possible significance for the history of the Eastern Mediterranean.
Quaternary Research 74: 207–215.
Karaman GS, Beeton A. 1981. The biota and limnology of Lake Skadar.
Smithonian Institution: Washington DC; and Univerzitet Veljko
Vlahvic, Titograd.
Kelts K, Hsu¨ KJ. 1978. Freshwater carbonate sedimentation. In Lakes,
Chemistry, Geology, Physics, Lerman A (ed.). Springer-Verlag:
Berlin; 295–323.
Keukelaar F, de Goffau A, Pradhan T, et al. 2006. Lake Shkoder
Transboundary Diagnostics Analysis (Albania & Montenegro), World
Bank (IBRD) Final Report, 9P6515.
Kim ST, O’Neil JR. 1997. Equilibrium and non-equilibrium oxygen
isotope effects in synthetic carbonates. Geochimica et Cosmochimica
Acta 61: 3461–3475.
Jimenez-Espejo FJ, Martınez-Ruiz F, Rogerson M, et al. 2008. Detrital
input, productivity fluctuations, and water mass circulation in the
westernmost Mediterranean Sea since the Last Glacial Maximum.
Geochemistry Geophysics Geosystems 9: Q11U02. DOI: 10.1029/
2008GC002096.
Lasca NP, Radulovic V, Ristic RJ, et al. 1981. Geology, hydrology,
climate and bathymetry of Lake Skadar. In: Karaman GS, Beeton AM
(eds). The Biota and Limnology of Lake Skadar. Smithonian Institution:
Washington, DC; 17–38.
Leng MJ, Marshall JD. 2004. Palaeoclimate interpretation of stable
isotope data from lake sediment archives. Quaternary Science
Reviews 23: 811–831.
Leng MJ, Jones MD, Frogley MR, et al. 2010a. Detrital carbonate
influences on bulk oxygen and carbon isotope composition of
lacustrine sediments from the Mediterranean. Global Planetary
Change 71: 141–149.
Leng MJ, Baneschi I, Zanchetta G, et al. 2010b. Late Quaternary
palaeoenvironmental reconstruction from Lakes Ohrid and Prespa
(Macedonia/Albania border) using stable isotopes. Biogeosciences 7:
3109–3122.
Lionello P, Malanotte-Rizzoli P, Boscolo R. 2006. Mediterranean
Climate Variability, 4. Elsevier: The Netherlands.
Longinelli A, Selmo E. 2003. Isotopic composition of precipitation in
Italy: a first overall map. Journal of Hydrology 270: 75–88.
Magny M, de Beaulieu JL, Drescher-Schneider R, et al. 2007. Holocene
climate changes in the central Mediterranean as recorded by lakelevel
fluctuations at Lake Accesa (Tuscany, Italy). Quaternary
Science Reviews 26: 1736–1758.
Magny M, Peyron O, Sadori L, et al. 2012. Contrasting patterns of
precipitation seasonality during the Holocene in the south- and
north-central Mediterranean. Journal of Quaternary Science 27:
290–296.
Magny M, Vannie`re B, Zanchetta G, et al. 2009. Possible complexity of
the climatic event around 4300–3800 cal. BP in the central and
western Mediterranean. The Holocene 19: 1–11.
Martı´n-Puertas C, Jimenez-Espejo F, Martine-Ruiz F, et al. 2010. Late
Holocene climate varibility in the southwestern Mediterranean
region: an integrated marine and terrestrial geochemical approach.
Climate of the Past 6: 807–816.
Mayer B, Schwark L. 1999. A 15,000-year stable isotope record
from sediments of Lake Steisslingen, Southwest Germany. Chemical
Geology 161: 315–337.
Meyers PA, Lallier-Verge`s E. 1999. Lacustrine sedimentary organic
matter records of Late Quaternary paleoclimates. Journal of Paleolimnology
21: 345–372.
McMillan EA, Fairchild IJ, Frisia S, et al. 2005. Annual trace elements
cycle in calcite-aragonite speleothems: evidence of drought in the
western Mediterranean1200–1100 years BP. Journal of Quaternary
Science 20: 423–433.
Mook WG, Tan FC. 1991. Stable carbon isotopes in rivers and estuaries.
In: Biogeochemistry of Major World Rivers, Degens ET, Kempe S,
Richey E (eds). John Wiley: New York; 245–264.
Reimer PJ, Baillie ML, Bard E, et al. 2004. IntCal04 terrestrial radiocarbon
age calibration, 0–26cal kyr BP. Radiocarbon 46: 1029–1058.
Roberts N, Brayshaw D, Kuzucuog˘lu C, et al. 2011a. The mid-Holocene
climatic transition in the Mediterranean: Causes and consequences.
The Holocene 21: 3–13.
Roberts N, Eastwood WJ, Kuzucuog˘lu C, et al. 2011b. Climatic,
vegetation and cultural change in the eastern Mediterranean during
the mid- Holocene environmental transition. The Holocene 21: 147–
162.
Roberts N, Jones MD, Benkaddour A, et al. 2008. Stable isotope records
of Late Quaternary climate and hydrology from Mediterranean lakes:
the ISOMED synthesis. Quaternary Science Reviews 27: 2426–2441.
Roberts N, Stevenson AC, Davis B, et al. 2004. Holocene climate,
environment and cultural change in the circum-Mediterranean
region. In: Past climate variability through Europe and Africa,
Battarbee RW, Gasse F, Stickley C (eds). Kluwer: Dordrecht;
pp. 343–362.
Roberts N, Zanchetta G, Jones MD. 2010. Oxygen isotopes as tracers
of Mediterranean climate variability: an introduction. Global and
Planetary Change 71: 135–140.
Rohling EJ, Mayewsky PA, Hayes A, et al. 2002. Holocene atmosphereocean
interactions: records from Greenland and the Aegean Sea.
Climate Dynamic 18: 587–593.
Rosi M, Santacroce R. 1983. The A.D. 472 ‘Pollena’ eruption, volcanological
and petrological data for this poorly know Plinian-type
event at Vesuvius. Journal of Volcanology and Geothermal Research
17: 249–271.
Sadori L, Giraudi C, Petitti P, et al. 2004. Human impact at Lago di
Mezzano (central Italy) during the Bronze Age: a multidisciplinary
approach. Quaternary International 113: 5–17.
Sadori L, Jahns S, Peyron O. 2011. Mid-Holocene vegetation history of
the central Mediterranean. The Holocene 21: 117–129.
Sulpizio R, van Welden A, Caron B, et al. 2010. The Holocene
tephrostratigraphic record of Lake Shkodra (Albania and Montenegro).
Journal of Quaternary Science 25: 633–650.
Thorndycraft VR, Benito G. 2006. Late Holocene fluvial chronology of
Spain: the role of climatic variability and human impact. Catena 66:
34–41.
van Welden A, Beck C, Reyss JL, et al. 2008. The last 500 year of
sedimentation in Shkodra Lake (Albania/Montenegro): paleoenvironmental
evolution and potential for paleoseismicity studies. Journal
of Paleolimnology 40: 619–633.
Wanner H, Beer J, Bu¨tikofer J, et al. 2008. Mid- to Late Holocene
climate change: an overview. Quaternary Science Reviews 27:
1791–1828.
Wierzbowski H. 2007. Effects of pre-treatments and organic matter on
oxygen and carbon isotope analyses of skeletal and inorganic
calcium carbonate. International Journal of Mass Spectrometry
268: 16–29.
Zanchetta G, Borghini A, Fallick AE, et al. 2007. Late Quaternary
palaeohydrology of Lake Pergusa (Sicily, southern Italy) as inferred by
stable isotopes of lacustrine carbonates. Journal of Paleolimnology
38: 227–239.
Zanchetta G, Sulpizio R, Roberts N, et al. 2011. Tephrostratigraphy,
chronology and climatic events of the Mediterranean basin during
the Holocene: an overview. The Holocene 21: 33–52.
Zielhofer C, Faust D. 2008. Mid- and Late Holocene fluvial chronology
of Tunisia. Quaternary Science Reviews 27: 580–588.
Zoto J. 2006. Environmental Isotope Ratios of Hydrogen and Oxygen in
the Water Systems and Information Data System. Water Observation
and Information System for Decision Support. Ohrid, Republic of
Macedonia.
Copyright ß 2012 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 27(8) 780–789 (2012)
MULTIPROXY RECORD FROM LAKE SHKODRA 789