background image

GEOLOGICA CARPATHICA, DECEMBER 2005, 56, 6, 545—552

www.geologicacarpathica.sk

Paleoclimate record in the Upper Pleistocene loess-paleosol

sequence at Petrovaradin brickyard (Vojvodina, Serbia)

SLOBODAN B. MARKOVIĆ

1

, WILLIAM D. McCOY

2

,

 

ERIC A. OCHES

3

,

 

STEVAN SAVIĆ

1

,

TIVADAR GAUDENYI

1

, MLADJEN JOVANOVIĆ

1

, THOMAS

 

STEVENS

4

, RICHARD

 

WALTHER

2

,

PETAR IVANIŠEVIĆ

5

 and ZORAN GALIĆ

5

1

Quaternary Research Centre, University of Novi Sad, Trg Dositeja Obradovića 3, 21000 Novi Sad, Serbia and Montenegro;

zbir@im.ns.ac.yu

2

Department of Geosciences, University of Massachusetts, MA-01003 Amherst, U.S.A;  wdmccoy@geo.umass.edu

3

Department of Environmental Science & Policy, University of South Florida, 4202 E. Fowler Ave. SCA 528, FL-33620 Tampa,

U.S.A.;  oches@cas.usf.edu

4

Oxford Luminescence Research Group, School of Geography and the Environment, University of Oxford, Mansfield Road, Oxford,

United Kingdom

5

Poplar Institute, Department of Soil science, Antona Čehova 13, 21000 Novi Sad, Serbia and Montenegro

(Manuscript received March 3, 2005; accepted in revised form June 16, 2005)

Abstract: Four loess units and three paleopedological layers are preserved in the 

~ 8  m thick Petrovaradin exposure,

Vojvodina, Serbia. Amino acid geochronology provides stratigraphic correlations between loess units L1 and L2 at
Petrovaradin with loess of glacial cycles B and C, respectively, at other Central European localities. Magnetic suscep-
tibility and sedimentological evidence of the Petrovaradin loess-paleosol sequence are used for correlation with the
SPECMAP paleoclimatic record. Late Pleistocene climate dynamics recorded in the Petrovaradin brickyard loess-
paleosol sequence present temperate humid and warm interglacial and temperate cold glacial climatic conditions. The
last glacial paleoclimatic record provides two main cold and dry stadial periods corresponding to deposition of two
loess layers L1L1 and L1L2, as well as one moderate cold and relatively dry interstadial. Many episodes of alternating
cold-dry and warm-wet paleoclimatic conditions suggest a possible correlation with abrupt paleoclimatic fluctuations
recorded in the North Atlantic region. The results of malacological investigations of the Petrovaradin site demonstrate
significant similarities to the Paleopreillyrian fauna of the southern Transdanubia region in Hungary, which suggests
that the Petrovaradin site has a refugial character during the periods of dust accumulation.

Key words: Late Pleistocene, Serbia, paleoclimate, amino acid geochronology, magnetic susceptibility, loess, grain size.

Introduction

The thick loess-paleosol succession of the Vojvodina re-
gion contains a detailed record of Middle and Late Pleis-
tocene paleoclimatic and paleoenvironmental changes
(Marković et al. 2003). Previous investigations were fo-
cused on well known open loess sections at Stari Slanka-
men, Neštin, Batajnica and Mošorin (Bronger 1976,
2003; Singhvi et al. 1989; Butrym et al. 1991; Kostić &
Protić 2000; Marković et al. 2003) located on the steep
banks of the Danube and Tisza (= Tisa / Theiss) rivers.
However, recent research interests have extended to the
loess-paleosol exposures uncovered in the brickyards at
Ruma, Irig and Petrovaradin (Marković 2001; Marković
et al. 2000, 2004b, in print; Gaudenyi et al. 2003).

The Petrovaradin brickyard loess exposure (45°16’ N Lati-

tude and 19°52’ E Longitude) is situated in the central part of
the north loess slope of the Fruška Gora (Vojvodina, Serbia)
(Fig. 1). Initial investigations focus on the four loess layers
and three paleosols preserved in the ~ 8  m thick exposure.
This study points to the importance of Petrovardin’s loess-
paleosol sequences for understanding paleoclimatic and
paleoenvironmental variations in the southeastern part of
the Carpathian (Panonnian) Basin during the last ~145 ka.

Material and methods

Investigations of the loess-paleosol sequences of the Petro-

varadin brickyard began in 2001. Samples were collected at
5-cm intervals for sedimentological analysis, and at 25-cm
intervals for malacological studies. Grain size (GS) fractions
(<2 , 2—20, 20—200, >2 00 

µm) were measured by sieving

and pipeting, carbonate content was analysed by gas volu-
metrically. The low-field magnetic susceptibility (MS) was
measured in the field using a portable Bartington MS2 sus-
ceptibility meter. Measurements were made at 5-cm intervals
and at each level, ten independent readings were taken and
averaged. 10-kg bulk sediment samples for malacological in-
vestigations were sieved through 0.7 mm mesh. After fossil
gastropod shells were identified, paleoenvironmental classifi-
cation and interpretation was done based on the methods of
Ložek (1964), but also extended with some local variants de-
fined by Krolopp & Sümegi (1995). Gastropod shells were
collected from six levels within L1 (last glacial loess) and the
upper part of L2 (penultimate glacial loess) for amino acid ra-
cemization (AAR) analysis in order to independently corre-
late the stratigraphy with loess-paleosol units elsewhere in
Europe. Details of the sample preparation and analytical
methodology are presented in Oches & McCoy (2001).

background image

546

MARKOVIĆ et al.

Results

Litho- and pedostratigraphy

Stratigraphic studies of loess and paleosols at various ex-

posures in the Vojvodina region have used lithological and
pedogenic criteria and MS variations as the primary bases
for correlation. Marković et al. (2003, 2004a, in print) des-
ignated the Vojvodinian loess-paleosol chronostratigraphic
units by names that follow the Chinese loess stratigraphic
system (Kukla 1987), beginning with the prefix “SL” refer-
ring to the standard section at the Stari Slankamen site. Ac-
cording to the current chronostratigraphic model (Marković
et al. 2004b), the penultimate glacial loess L2 accumulated
during the marine isotope stage (MIS) 6. The last intergla-
cial—early glacial paleosol S1 correlates with MIS 5. This
palaeosol is overlain by composite loess unit L1, correlated
with MIS 4—2. The structure of the last glacial loess L1 var-
ies in different loess localities across the Vojvodina region.
The lower sub-horizon of the upper loess, L1L2, accumulat-
ed above palaeosol S1. The Middle Pleniglacial is repre-
sented in the area by a weakly developed soil complex
L1S1. The youngest loess layer L1L1 accumulated during
the Late Pleniglacial period.

Three paleosols and four loess layers are distinguished

at the Petrovaradin quarry (Fig. 2). Table 1 shows the mor-

phological description of the loess-paleosol sequence.
The oldest pale yellow (5Y 7 / 3, 5 / 4) loess unit L2 is un-
covered only in the lowest 75 cm of the profile. The lower
part of this loess unit is not excavated during the raw ma-
terial exploitation. Many carbonate concretions (1—4 cm
diameter) and humus infiltrations appear near the contact
of the S1 soil complex and the underlying L2 loess. The
reddish-brown pedocomplex S1, 205 cm thick, contains
lower slightly brownized, middle typical and upper weak
developed chernozems. The lower darker transitional A(B)
horizon (10YR 5 / 2—3) is darker than the middle Ah hori-
zon (10YR 6 / 2—4) that contains some preserved carbon-
ate pseudomycelia. The uppermost relatively weakly
developed A horizon is characterized by many krotovinas.

The composite loess unit L1 is 460 cm thick and con-

tains three loess layers (L1L1, L1S1L1 and L1L2) sepa-
rated by two paleopedological subunits (L1S1S1 and
L1S1S2). The 130 cm thick loess layer L1L2 accumulat-
ed above the S1 paleosol. This light yellow gray (5Y 7 / 3
5 / 3) loess layer is porous, loosely cemented and in some
parts finely laminated with thin fine sand beds. The super-
imposed paleopedological horizon L1S1S2 (5Y 6 / 3—4) is
35 cm thick.  It is a weak, incipient pedohorizon with small
and soft carbonate spots.  Loess stratum L1S1L1 is 35 cm
thick with morphological characteristics similar to loess
layer L1L2. The youngest paleosol L1S1S1 is a 75 cm

Fig. 1. Geographical position of the Petrovaradin brickyard exposure and other relevant sites in the Vojvodinian loess area. 1 – loess
plateau, 2 – sandy area, 3 – mountain, 4 – main loess sections.

background image

547

PALEOCLIMATE RECORD IN THE UPPER PLEISTOCENE LOESS-PALEOSOL SEQUENCE (SERBIA)

Fig. 2. Stratigraphy of the Petrovaradin brickyard exposure. Positions
of samples for amino acid analysis and Mammuthus primigenius skele-
tal remains are indicated by the arrows. HYD A / I values are shown for
gastropod genera on which analyses were made. 1 – Loess, 2 – Em-
bryonic pedogenetic layer (incipient soil horizon), 3  – A horizon,
4 – Ah horizon, 5 – Transitional AB horizon, 6 – Krotovinas,
7 – Carbonate concretions, 8 – Humus infiltrations.

thick poorly developed chernozem-like pedological hori-
zon. Loess layer L1L1 is 185 cm thick, very porous and in
some parts intensively bioturbated. Many spherical, rela-
tively soft carbonate nodules and humus infiltrations in
old root channels are found at the contact zone with the
modern soil S0.

The Holocene soils developed on the loess plateau sur-

face in the area surrounding Petrovaradin brickyard are
chernozem eroded and chernozem slightly melanized
(Miljković 2001). At the top of the investigated section,
the modern soil is a 70 cm thick chernozem slightly mela-
nized. The lower Ck horizon contains many CaCO

3

 nod-

ules of 1—5 cm in diameter and numerous krotovinas and
root channels filled with humic material. A transitional
AC horizon (10YR 5 / 1—3 / 3) is a 15 cm thick, very po-
rous, silt loam with fine blocky structure. The A(B) hori-
zon is 35 cm thick, reddish-brown (7.5YR 4 / 2—4 / 4),
and porous with blocky structure. The Ah horizon (10YR
6 / 3—4 / 4) is a 20 cm thick silt loam with granular structure
and some carbonate pseudomycelia.

AAR chronostratigraphy and correlations

Amino acid racemization (AAR) geochronology using

fossil gastropod shells has been successfully applied to
the stratigraphic correlation of loess-paleosol sequences in
different regions of the world (Oches & McCoy 2001). The
Petrovaradin profile is one of the first Serbian loess sites in
which AAR analyses have been carried out. Three differ-
ent genera of terrestrial gastropod shells have been analy-
sed, including samples from six levels within the loess
sequence at Petrovaradin. Sampled genera include Pupil-
la, Succinea and Trichia. Alloisoleucine / Isoleucine total
acid hydrolysate (A / I—HYD) measurements on representa-
tive samples are shown in Fig. 3. Shells of the genus Pu-
pilla and Trichia  were the most abundant and offer the
most direct aminostratigraphic comparison with data from
loess units elsewhere in central and eastern Europe. HYD
A / I values measured in Trichia from the Petrovaradin
brickyard profile can be compared with data from Austri-

Unit   /   subunit 

Thickness (cm) 

Depth (cm) 

Description 

S0 

  70 

   0–70 

Modern soil, 70-cm thick chernozem slightly melanized with weak blocky structure 

L1 

460 

  70–530 

 

 

L1L1  

185 

  70–255 

Very porous, pale yellow loess (10YR 7/4–5/3) with many humic infiltrations and 
carbonate concretions (ø 1–4 cm) 

L1S1S1 

  75 

255–330 

Weakly developed mollic chernozem-like soil (10YR 6/3–4/2) 

L1S1L1 

  35 

330–365 

Light yellow grey (5Y 7/3, 5/3) porous loess with soft and small carbonate concretions 

L1S1S2 

  35 

365–400 

Weakly developed, yellowish brown (5Y 7/3 5/3) incipient (A) horizon with granular 
strucrure 

 

L1L2 

130 400–530 

Light yellow grey (5Y 7/3, 5/3) porous, loosely cemented partly finely laminated 
calcareous loess  

S1 

225 530–735 

Thick transitional chernozem pedocomplex: the lower slightly brownized AB horizon 
has weak platy structure (10YR 5/2–3); the middle typical chernozem part has brighter 
colour (10YR 6/2–4), and the upper weakly developed chernozem has blocky structure 
with many krotovinas in the uppermost level 

L2 

  75 

735–800 

Porous yellow (5Y 1/3, 5/4) loess with many humus infiltrations and carbonate 
concretions 

Table 1: Morphological description of loess-paleosol at Petrovaradin brickyard exposure.

background image

548

MARKOVIĆ et al.

an, Czech, Slovak, Hungarian and German sites (Oches &
McCoy 1995a,b, 2001; Oches et al. 2000) (Fig. 3). Present-
day air and ground temperatures in Hungary and Serbia
are higher than in the more northern sites. This higher tem-
perature explains the more rapid rate of racemization in
Vojvodina loess than in the equivalent age loess units in
other Central European sites. These data show that the

Fig. 3. Aminostratigraphy of the Petrovaradin brickyard section
compared with other Central European localities for glacial cycles
B and C, corresponding to marine oxygen-isotope stages 2—5 and
6—7, for the genus Trichia. H – Hungary, CZ – Czech Republic,
SK – Slovakia, A – Austria, D – Germany.

Fig. 4.  Depth plots of magnetic susceptibility, clay content, carbonate content, and particle size fractions  > 20 

µm and >200 µm at the

Petrovaradin exposure related to SPECMAP paleoclimatic model (Martinson et al. 1987).

method can clearly distinguish between loesses of the last
two glacial cycles.

The  Pupilla data at Petrovaradin come only from loess

of the last glacial cycle. Pupilla HYD A / I values from be-
low L1S1 are only slightly greater than from shells in the
loess above that soil complex, despite being tens of thou-
sand of years older. This can only be possible if the soil
does not represent an interglacial period. A warm intergla-
cial would induce very much more epimerization as is dem-
onstrated by the Trichia HYD A / I ratios from below S1.

AAR geochronology results from the Petrovaradin sec-

tion support the previous chronostratigraphic scheme of
Marković (2000, 2001). According to that chronostrati-
graphic model, loess-paleosol sequences L1 and S1
formed during glacial cycle B (Kukla 1975) and corre-
spond to MIS 2, 3, 4 and 5. The exposed part of the L2
loess horizon was deposited during the youngest part of
glacial cycle C (MIS 6).

Low-field magnetic susceptibility (MS)

According to our investigations, the variations in the

low-field MS are related to the pedostratigraphy in the
Petrovaradin brickyard section. MS values observed in
soils S0 (average 44.7 SI units) and S1 (average 29.5 SI
units) are higher than in the loess units L1 (average 22.4
SI units) and L2 (average 13.3 SI units) (Fig. 4). This
type of MS pattern reflects magnetic enhancement via

background image

549

PALEOCLIMATE RECORD IN THE UPPER PLEISTOCENE LOESS-PALEOSOL SEQUENCE (SERBIA)

pedogenesis and is similar to that in Chinese and Central
Asian loess deposits (e.g. Maher & Thomson 1999).

The basal penultimate glacial loess unit L2 shows the

lowest MS values: less than 15 SI units. Readings in
lower part of the S1 soil complex increase to 37 SI
units. MS increases to 49 SI units in the middle part of
the soil where it is the highest in the S1 paleosol. A
sharp decrease of MS occurs in the base of the overly-
ing lighter horizon with minimal value 21 SI units.
Close to the top of paleosol S1, MS values again in-
crease to nearly 40 SI units and above that continuous-
ly decreased to relatively low values in loess subunit
L1L2:  ~ 1 8 SI units, with one strong peak of 33 SI units
(425 cm depth). The L1S1S2 shows higher, stable val-
ues (range of 21 to 26 SI units). In the thin loess inter
layer, L1S1L1, MS is essentially unchanged ~ 2 2 SI
units. MS values of the interstadial paleosol L1S1S1 in-
crease to 43 SI units. MS of the youngest loess layer
L1L1 is low: ~ 17 SI units. The Holocene soil S0 shows
an increasing trend from ~ 3 0 SI units at the base to
more than ~ 5 0 SI units in its middle part (Fig. 4).

Grain-size (GS) distribution and carbonate content

Measured variations in GS distribution also coincide

well with the pedostratigraphy of the Petrovaradin loess-
paleosol sequence. Generally, the pedogenic horizons
have a lower proportion of coarse material than the loess
layers (e.g. Vandenberghe & Nugteren 2001). Variability
in clay content (<2  

µm) parallels the MS record. The high-

est value of clay content is observed in lower part of
palaeosol S1 (close to 40 %) in contrast with low values
detected in loess layers (ca. 15 %). Variations of coarse
material (>2 0 

µm) content show many abrupt small ampli-

tude changes possibly linked to wind transport intensity.
High values of >20 

µm content are also associated with

several peaks of coarse sand (>2 00 

µm) deposition in both

loess units L1 and L2 (Fig. 4).

High values of carbonate content are detected in loess

units: more than 30 % in L2; L1L1 is greater than 20 %;
L1L3 is characterized by values about 10 %; and L1L2
has more than 7 % carbonate, in contrast to very low val-
ues of carbonate (less than 5 %) in fossil soils. Peaks of the
coarser material deposition are connected with relative
minima of carbonate content inside the loess and paleosol
units (Fig. 4).

Malacology

Shells of 5053 individuals land snails representing 33

species (26 genera) were found in 30 samples taken from the
Petrovaradin section. Generally, the terrestrial malacologi-
cal assemblages reflect humid and relatively cold paleoen-
vironmental conditions with mosaic vegetation (Fig. 5).

The snail assemblage from the upper part (final stage) of

the L2 horizon is characterized by relatively large abun-
dance of shells per sample (more than 200 individuals per
sample) with a dominance of humid-preferring species of
different biotopes. The dominance of Trichia hispida, Vit-

rea crystallina, Punctum pygmaeum, Aegopinella ress-
manni  and Clausilia dubia indicate humid environmental
conditions characterized by woodland steppe like vegeta-
tion. The closed-canopy woodland, humid environment
preferring species Ena montana and Discus ruderatus
have been found with very low frequency in the rest of the
samples. The presence of cold resistant species such as
Vallonia tenuilabris and Succinea oblonga indicates rela-
tively cold climatic conditions.

No land snails were recovered from the lower parts of

paleosol S1. Because of poor preservation and leaching in
the soil, this unit was not valuable for malacological in-
vestigations.

Dominance of aridity-tolerant snails such as Pupilla

triplicata, P. muscorum, Helicopsis striata, Vallonia cos-
tata, Granaria frumentum and Chondrula tridens in loess
layers L1L2 and L1S1L1 and weak pedo-horizon L1S1S2
represents open vegetation and a mostly dry environment
related to steppe-like grassland.

The snail assemblages of the youngest paleosol L1S1S1

and loess L1L1 provided different environmental condi-
tions. The increase in mollusc abundance and the presence
of forest snails such as Aegopinella ressmanni,  Orcula do-
lium  and Clausilidae in the loess layer L1L1 indicates
more humid conditions.

During the last glacial period we note the presence of two

relatively cold episodes. The coldest one was during the ac-
cumulation of loess layer L1L1 indicated by the presence of
the cold-resistant species Vallonia tenuilabris,  Columella
columella, Vertigo parcedentata and Succinea oblonga.
The earlier cold period occurred during the deposition of
loess horizon L1L2 characterized by low frequency of spe-
cies such as Vallonia tenuilabris and V. pulchella.

Discussion

Paleoclimatic and paleoenvironmental interpretations

The multi-proxy data set and identified pedo-stratigra-

phy in the Petrovaradin sequence indicates a sharp con-
trast between glacial and interglacial climatic conditions.
The final part of penultimate glacial loess cycle L2 and
two last glacial loess layers L1L2 and L1L1 accumulated
during temperate, cold, stadial intervals. By contrast, dur-
ing interstadial and interglacial times, the climate was
warmer and generally wetter and this led to enhanced pe-
dogenesis. Whereas interglacial soils tend to be strongly
developed, interstadial paleosols L1S1S2 and L1S1S1 are
weakly expressed. Generally, all paleoclimatic proxies de-
termined from the loess-paleosol sequence of Petrovaradin
correlate well with the SPECMAP paleoclimatic model
(Martinson et al. 1987) (Fig. 4).

Besides climatic changes related to interglacial—glacial

and stadial—interstadial conditions, especially, GS and
carbonate-content variations in the loess units recorded
paleoclimatic instability characterized by many episodes
of cold-dry and warm-wet paleoclimatic conditions
(Fig. 4) indicating possible teleconections with Dans-

background image

5

5

0

MARKOVIĆ et al.

Fig. 5. Relative abundance diagram of the identified mollusc species in the Petrovaradin brickyard loess exposure. The species are clustered in ecological groups, as defined by Ložek
(1964), but also extended with some local variations defined by Krolopp & Sümegi (1995) and Sümegi & Krolopp (2002). 1 – tundra-like, 2 – dry steppe, 3 – grassland, 4 – transi-
tional zone, 5 – forest, 6 – wetland. Star mark shows snail ambudance less than 1 % per sample.

background image

551

PALEOCLIMATE RECORD IN THE UPPER PLEISTOCENE LOESS-PALEOSOL SEQUENCE (SERBIA)

gaard-Oeschger (D / O) cycles (e.g. Dansgaard et al. 1993)
and cold events recorded in the North Atlantic region (e.g.
McManus et al. 1994).

The AAR data show that the interstadial climatic condi-

tions may have been somewhat warmer than during the
stadials, but it was not nearby as warm as during the in-
teglacial represented by paleosol S1.

According to the paleopedological interpretation and

the identified land snail fauna of the Petrovaradin se-
quence, the Late Pleistocene was characterized by differ-
ent environments, ranging from temperate warm interglacial
forest-steppe, across dry lower and Middle Pleniglacial
grassland, to a relatively cold and humid mosaic environ-
ment during the end of the Penultimate glacial and again
during the Late Pleniglacial. The snail assemblage of
Petrovaradin’s loess layers is similar to the nearby loess
site at Mišeluk (Marković et al. 2004b). Both mollusc
records show more humid and relatively colder environ-
ments than in other sites of the Vojvodina region (Mark-
ović et al. 2000). Therefore, we think that the north slopes
of Fruska Gora Mountain had an important role during the
Late Pleistocene. It was perhaps a refugium, that is one of
those rare places in the southeast part of the Carpathian
Basin where the Paleoillyrian snail assemblage (e.g. Mac-
rogastra ventricosa, Aegopinella ressmanni and Trichia
edentula;  Sümegi & Krolopp 2002) survived.

In 1978, during excavation of brickyard raw materials

only 1 km away from the present section, workers found
fragments of a woolly mammoth skeleton (Mammuthus
primigenius) at the base of the L1 loess (L1L2), at a
depth 4.7 m below the present surface (Milić 1978), add-
ing to the paleoecological picture of the Late Pleistocene
environment.

Relation with other loess records in the Carpathian Basin

The investigations of the Petrovaradin brickyard loess-

paleosol sequence provide a reconstruction of the com-
plex Late Pleistocene paleoclimatic and environmental
evolution in this area. These results are comparable with
recent studies of other loess-paleosol sequences in the
Carpathian Basin that also have paleoclimatic and pale-
oenvironmental reconstructions based on luminescence
age estimations, amino-acid geochronology, sedimento-
logical, rock magnetic and malacological evidence from
the last glacial period (Zöller & Wagner 1990; Zöller et al.
1994; Oches & McCoy 1995a,b; Krollopp & Sümegi
1995; Frechen et al. 1997; Sartori et al. 1999; Sümegi &
Krolopp 2002).

The stratigraphic pattern of the Petrovaradin brickyard

site is generally similar to other Upper  Pleistocene loess-
paleosol sequences in Hungary (Pécsi 1993; Horváth
2001). However, all analysed proxies at Petrovaradin and
other loess exposures in the Vojvodina region (Marković
et al. 2004a,b) indicate that Late Pleistocene paleoclimat-
ic and paleoenvironmental evolution was relatively uni-
form compared to Hungarian loess sites (Pécsi 1993;
Horváth 2001). Stratigraphic equivalents of Hungarian
Upper  Pleniglacial humus horizons h1 and h2 are not

found in the Petrovaradin brickyard. In addition, Petrova-
radin’s Middle Pleniglacial paleopedocomplex L1S1 is
weakly developed in comparison with the corresponding
paleosol MF1 in Hungarian loess. However, characteristics
of Lower Pleniglacial loess L1L2 and the last interglacial-
Lower  Pleniglacial soil S1 are similar to Hungarian loess
l2 and paleosol MF2.

Conclusions

The multidisciplinary study of the loess-paleosol se-

quence reveals a very detailed Late Pleistocene record at
the Petrovaradin brickyard section. These results have es-
tablished the importance of this site as a record of the last
interglacial—glacial paleoclimate and paleoenvironment
in the southeastern part of the Carpathian Basin.

Detailed AAR data provide aminostratigraphic correla-

tion between loess units L1 and L2 at Petrovaradin with
loess of glacial cycles B and C, respectively, at other Eu-
ropean loess localities. Sedimentological, paleopedologi-
cal, and paleontological evidence recorded in the
Petrovaradin sequence, all demonstrate different intensi-
ties and scales of climate instability: glacial—interglacial,
stadial—interstadial and abrupt millennial-scale variability.

The identified fossil snail assemblages indicate more

humid environmental conditions in this region than in
other parts of the Vojvodina region (North Serbia) during
the last glacial and the final part of the penultimate gla-
cial. As a result the northern slope of the Fruska Gora was
a biogeographical “island” during the last two glacials;
that is a refugium where the Paleopreillyrian snail assem-
blage survived.

Acknowledgments: 

The authors wish to thank Prof. Tereza

Madeyska and Prof. Pál Sümegi for stimulating comments.
This is part of Marković’s research project supported by a
Humboldt fellowship. Funding for amino acid analysis
was provided by U.S. National Science Foundation grant
ATM-0081754 to Oches and ATM-081115 to McCoy.

References

Bronger A. 1976: Zur quartären Klima – und Landshaftenwick-

lung des Karpatenbeckens auf (paläo-) pedologisher und
bodengeographischer Grundlage. Kieler geographisce schiften
band 45. Im  Selbestvarlag des geographischen Iinstituts der
Universität, Kiel, 1—268.

Bronger A. 2003: Correlation of loess-paleosol sequences in East

and Central Asia with SE Central Europe: Towards a continen-
tal Quaternary pedostratigraphy and paleoclimatic history.
Quat. Int. 106, 107, 11—31.

Butrym J., Maruszcak H. & Zeremski M. 1991: Thermolumines-

cence stratigraphy on Danubian loess in Belgrade environs.
Annales UMCS, sec. B 46, 53—64.

Dansgaard W., Johnsen S.J., Clausen H.B., Dahljensen D., Gunde-

strup N.S., Hammer C.U., Hvidberg C.S., Steffensen J.P.,
Sveinbjornsdottir A.E., Jouzel J. & Bond G. 1993: Evidence
for general instability of past climate from a 250-kyr ice-core
record. Nature 364, 218—220.

background image

552

MARKOVIĆ et al.

Frechen M., Horvath E. & Gabris Gy. 1997: Geochronology of

Middle and Upper Pleistocene loess sections in Hungary.
Quat. Res. 48, 291—312.

Gaudenyi T., Jovanović M., Sümegi P. & Marković S.B. 2003: The

north boundary of the Mediterranean paleoclimate influences
during the late Pleistocene at Southeastern part of Carpathian
Basin based on assemblages of mollusca (Vojvodina, Yugosla-
via). In: Zapata M.B.R., Valino M.D., Rodriguez A.V., Garcia
J.G., Azcarate T.B., de Bustamante Gutierrez I. & Mendizabal
I.M. (Eds.): Quaternary climatic changes and environmental
crises in the Mediterranean Region. Univeridad de Alcala, Al-
cala de Heneras, Madrid, 41—47.

Horváth E. 2001: Marker horizons in the loeses of Carpathian Ba-

sin.  Quat. Int. 76, 77, 157—163.

Kostić N. & Protić N. 2000: Pedology and mineralogy of loess profiles

at Kapela-Batajnica and Stalać, Serbia. Catena 41, 217—227.

Krolopp E. & Sümegi P. 1995: Paleoecological reconstruction of

the Late Pleistocene, based on loess malacofauna in Hungary.
GeoJournal  36, 2, 3, 213—222.

Kukla G.J. 1975: Loess stratigraphy of Central Europe. In: Butzer

K.W. & Isaac L.I. (Eds.): After Australopithecines. Mouton
Publishers,  Hague, 99—187.

Kukla G.J. 1987: Loess stratigraphy in Central China. Quat. Sci.

Rev. 6, 191—219.

Liu T.S. (Ed.) et al. (unnamed) 1985: Loess the environment. Chi-

na Ocean Press, Beijing, 1—251.

Ložek V. 1964: Quartarmollusken der Tschechoslowakei. Rozpr.

Ústř. Úst. Geol., Praha 31, 1—374.

Maher B.A. & Thompson R. (Eds.) 1999: Quaternary climates, en-

vironment and magnetism. Cambridge University Press, Cam-
bridge, 1—390.

Marković S.B. 2000: Paleogeography of Vojvodina region during

the Quaternary. Ph.D. Thesis. Inst. Geogr., Univ. Novi Sad (in
Serbian, unpublished).

Marković S.B. 2001: Paleosols of Srem region. In: Miljković N. &

Marković S.B. (Eds.): Soils of Srem region (Vojvodina, Yugo-
slavia).  Institute of Geography, Novi Sad, 133—155 (in Serbi-
an with English summary).

Marković S.B., Sümegi P., Krolopp E., Miljković L.J., Gaudenyi

T. & Jovanović M. 2000: The last glacial cycle paleoclimat-
ic record of Ruma loess section (Vojvodina, Yugoslavia).
Yearbook Dep. Geogr. 30, 14—19 (in Serbian with English
summary).

Marković S.B., Heller F., Kukla G.J.P., Gaudenyi T., Jovanović M.

& Miljković L.J. 2003: Magnetostratigraphy of the Stari
Slankamen loess-paleosol sequences. Yearbook Dep. Geogr.
32, 20—28 (in Serbian with English summary).

Marković S.B., Kostić N. & Oches E. 2004a: Paleosols in the Ruma

loess section. Revista Mexicana de Ciencias Geológicas 21, 1,
79—87.

Marković S.B., Oches E., Gaudenyi T., Jovanović M., Hambach U.,

Zöller L. & Sümegi P. 2004b: Paleoclimate record in the Late

Pleistocene loess-paleosol sequence at Miseluk (Vojvodina,
Serbia).  Quaternaire  15, 4, 361—368.

Marković S.B., Oches E., Sümegi P., Jovanović M. & Gaudenyi T.

(in print): An introduction to the Upper and Middle Pleis-
tocene loess-paleosol sequences in Ruma section (Vojvodina,
Yugoslavia).  Quat. Int.

Martinson D., Pisias M.G., Hays J.D., Imbrie J., Moore T.C. &

Shackleton N.J. 1987: Age dating and the orbital theory of ice
ages: development of a high-resolution 0 to 300,000-year
chronostratigraphy.  Quat. Res. 27, 1—30.

McManus J.F., Bond G.C., Broecker W.S., Johnsen S., Labeyrie L.

& Higgins S. 1994: High resolution climate records from the
North Atlantic during the last interglacial. Nature 371, 326—329.

Milić R. 1978: Novi nalazi Elephas primigenis Blum. u Sremu,

Vojvodina.  Priroda Vojvodine 4, 49—50.

Miljković N. 2001: Recent soil cover. In: Miljković N. & Marković

S.B. (Eds.): Soils of Srem region. Inst. Geogr., Novi Sad, 39—32
(in Serbian with English summary).

Oches E.A. & McCoy W.D. 1995a: Amino acid geochronology ap-

plied to the correlation and dating of Central European loess
deposits.  Quat. Sci. Rev. 14, 767—782.

Oches E.A & McCoy W.D. 1995b: Aminostratigraphic evaluation

of conflicting age estimates for the “Young Loess” of Hunga-
ry.  Quat. Res. 44, 767—782.

Oches E.A. & McCoy W.D. 2001: Historical developments and re-

cent advances in amino acid geochronology applied to loess
research: examples from North America, Europe and China.
Earth Sci. Rev. 54, 173—192.

Oches E.A., McCoy W.D. & Gniser D. 2000: Aminostratigraphic

correlation of loess-paleosol sequences acros Europe. In:
Goodfried G.A., Collins M.J., Fogel M.L., Macko S.A. & We-
hmiller F. (Eds.): Perspectives in amino acid and protein
geochemistry. Oxford Universsity Press, NY, 331—348.

Pécsi M. 1993: Quaternary and loess research (Loess inform 2).

Akadémia Kiadó, Budapest, 1—82.

Sartori M., Heller F., Forster T., Borkovec M., Hammann J. & Vincent

E. 1999: Magnetic properties of loess grain siye fractions from
section Paks (Hungary). Phys. Earth Planet. Int. 116, 53—64.

Singhvi A.K., Bronger A., Sauer W. & Pant R.K. 1989: Thermolu-

minescence dating of loess-paleosol sequences in the Car-
pathian Basin. Chem. Geol. 73, 307—317.

Sümegi P. & Krolopp E. 2002: Quatermalacological analysis for

modeling the Upper Weichselian palaeoenvironmental changes
in the Carpathian Basin. Quat. Int. 91, 53—63.

Vandenberghe J. & Nugteren G. 2001: Rapid changes in loess suc-

cessions. Glob. Planet. Chan. 28, 1—9.

Zöller L. & Wagner G.A. 1990: Thermoluminescence dating of

loess – recent developments. Quat. Int. 7, 8, 119—128.

Zöller L., Oches E.A. & McCoy W.D. 1994: Towards a revised

chronostratigraphy of loess in Austria with respect to key sec-
tions in the Czech Republic and in Hungary. Quat. Sci. Rev.
13, 465—472.