Tectonophysics 502 (2011) 175–195
Contents lists available at ScienceDirect
Tectonophysics
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o
The evolution of the Danube gateway between Central and Eastern Paratethys
(SE Europe): Insight from numerical modelling of the causes and effects of
connectivity between basins and its expression in the sedimentary record
K.A. Leever a,⁎, L. Matenco a, D. Garcia-Castellanos b, S.A.P.L. Cloetingh a
a
b
Netherlands Research Centre for Integrated Solid Earth Science (ISES), Faculty of Earth and Life Sciences, VU University, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands
Instituto de Ciencias de la Tierra Jaume Almera (ICTJA-CSIC), Group of Dynamics of the Lithosphere (GDL), Solé i Sabarís s/n, 08028 Barcelona, Spain
a r t i c l e
i n f o
Article history:
Received 15 June 2009
Received in revised form 4 December 2009
Accepted 9 January 2010
Available online 20 January 2010
Keywords:
Connectivity
Sedimentary basins
Numerical modelling
Paleogeography
a b s t r a c t
The Pannonian and Dacic Basins in SE Europe are presently connected by the Danube River across the South
Carpathians, to which they are in a back-arc and foreland position respectively. Part of the Paratethys realm
during the Neogene, open water communication between the basins was interrupted by the Late Miocene uplift
of the Carpathians. Different mechanisms have been proposed for the formation of the Danube gateway: capture
of the upstream lake or an upstream river or incision of an antecedent river. Estimates on its age range from Late
Miocene to Quaternary. A related issue is the effect of the large Mediterranean sea level fall related to the
Messinian Salinity Crisis on the Paratethys subbasins, specifically the “isolated” Pannonian Basin.
In a synthetic numerical modelling study, using a pseudo-3D code integrating tectonics, surface processes and
isostasy, we addressed the causes and effects of changes in connectivity between two large sedimentary basins
separated by an elevated barrier. Specifically, we aimed to find the expression of connectivity events in the
sedimentary record in general and the consequences for the evolution of the Pannonian–Dacic area in particular.
We studied a range of parameters including the geometry and uplift rate of the barrier, downstream sea level
change and lithosphere rigidity.
We found that changes in connectivity are expressed in the sedimentary record through their effect on base level
in the upstream basin and supply in the downstream basin. The most important factors controlling the response
are the elevation difference between the basins and the upstream accommodation space at the time of
reconnection. The most pronounced effect of reconnection through lake capture is predicted for a large elevation
difference and limited upstream accommodation space. Downstream increase in sediment supply is dependent
on the latter rather than the reconnection event itself.
Of the parameters we tested, the rigidity of the lithosphere was found to be of major importance by its control on
sediment loaded subsidence and generation of accommodation space. A downstream sea level change is unlikely
to induce capture, but may affect the upstream lake level by enhancing incision in a pre-existing gateway. In the
Pannonian–Dacic region, the mechanically weak, continuously subsiding Pannonian lithosphere allowed
accommodation of significant volumes of continental sedimentation and as a consequence, transfer of excess
sediment to the downstream Dacic Basin was only gradual. The Messinian sea level fall in the Dacic Basin could
have been recorded in the Pannonian Basin only if a connection between the basins already existed. More detailed
modelling of river incision taking into account lateral differences in erodibility in the South Carpathians will be
required to give better time constraints on the formation of the Danube Gateway.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Paratethys, extending from the foreland of the Alps to the Aral Sea
(Fig. 1a), was a large brackish epicontinental sea that was separated
from the world oceans during the progressive closure of the Tethys
Ocean and associated rising of the Alpine chain. In Southeast Europe,
⁎ Corresponding author. Presently at: Department of Geosciences, University of Oslo,
P.O. Box 1047, Blindern, 0316 Oslo, Norway.
E-mail address: [email protected] (K.A. Leever).
0040-1951/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.tecto.2010.01.003
formation and Late Miocene uplift of the Carpathians disrupted open
water communication between the Central/Western and Eastern
Paratethys, corresponding to the Pannonian Basin and the Dacic Basin
respectively (Fig. 1). The separation of the basins led to the evolution of
separate faunas and different biostratigraphies (e.g. Rögl, 1996) making
stratigraphic correlations problematic.
The Pannonian Basin to the west and the Dacic Basin to the east
(Fig. 1) are in a back-arc and foreland position relative to the Carpathians,
respectively. The Pannonian Basin system formed as a result of back-arc
extension during the Miocene (Horváth et al., 2006) and features the
176
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
weakest continental lithosphere of Europe (Cloetingh et al., 2006). The
subsidence in the Getic Depression (westernmost part of the Dacic
Basin) was due to its transtensional opening during the clockwise
rotation of the orogen along the Moesian platform and later orogenic
loading (Matenco et al., 1997; Fuegenschuh and Schmid, 2005).
Other than by tectonic structuring, the connectivity between the
Paratethys subbasins was influenced by climate through its control on
sea level. Effects of the major sea level fall in the adjacent Mediterranean
Sea related to the Messinian Salinity Crisis (>1000 m, Clauzon, 1978)
have been reported in the Paratethys subbasins, both the Black Sea
(Gillet et al., 2007), the Dacic Basin (Clauzon et al., 2005; Leever et al.,
2009) and the Pannonian Basin (Csató et al., 2007; Leever et al., 2009).
Presently, the Pannonian and Dacic Basins are exposed and
connected by the Danube River, deeply incised into its gorges across
the South Carpathians (Fig. 1b). The morphology of the long-studied
feature has been explained by the contrasting models of either stream
capture (Peters, 1876), lake overflow (Toula, 1896) or incision of an
antecedent river (Cvijic, 1908), while estimates on the age of their
formation range from Sarmatian to Quaternary (see Marovic et al., 1997,
for a review). These models are fundamentally different in the sense that
the “capture model” assumes that the river carved its bed into a preexisting topography by headward erosion, while the “antecedent river
model” assumes that river incision was able to keep pace with uplift of
the mountains—implying that the fluvial connection between the basins
was never lost.
Recent modelling work addressed the evolution of gateways across
elevated barriers by investigating the factors controlling capture
(piracy) of tectonic lakes (Garcia-Castellanos, 2006) and even an
ocean (Loget and Van Den Driessche, 2006): capture of the Atlantic
Ocean across the strait of Gibraltar reputedly allowed flooding of the
Mediterranean, restoring its sea level and ending the Messinian Salinity
Crisis. The time required for capture was found to be dependent mostly
on the precipitation/evaporation budget in the captured drainage basin
(upstream of the barrier), the rate of tectonic uplift at the barrier and the
flexural isostatic response of the lithosphere (Garcia-Castellanos et al.,
2003; Garcia-Castellanos, 2006) and on the base level in the lower
(downstream) basin (Loget and Van Den Driessche, 2006).
In this paper we use forward numerical modelling of coupled surface
processes and lithosphere response to determine the effect of changing
connectivity between two large sedimentary basins on depositional
geometries and sediment partitioning. The aim is to define the signature
of connectivity events on the stratigraphic record, i.e. what to look for in
seismic data. In other words, applied to the study area: are observations
of sedimentary architecture from seismic sequence stratigraphy of any
use to decide which of the models for formation and evolution of the
Danube gateway (capture vs. incision of an antecedent river) is correct?
We address questions such as: what is the influence of the strongly
contrasting lithosphere strength in the Pannonian and Dacic Basins?
Could the effects of the Mediterranean sea level fall of the Messinian
Salinity Crisis have extended to the Pannonian Basin, and how? Did it
influence the evolution of the Danube Gateway and as such the
connectivity between the Pannonian and Dacic basins?
2. Tectonic and paleogeographic setting of Paratethys in SE Europe
2.1. Tertiary tectonic evolution
The Pannonian–Carpathian region evolved during Alpine deformations in the tectonic context of the Mediterranean, which is character-
177
ized by highly arcuate plate boundaries resulting from the roll-back and
steepening of subducted lithosphere into land-locked remnant oceanic
basins (Wortel and Spakman, 2000; Fig. 1a).
The South Carpathian orocline, located between the Balkans in the
south and the South Carpathians in the northeast (Fig. 1), was structured
in a polyphase tectonic history that started in the Middle Cretaceous and
continued well into the Miocene (Sandulescu, 1984; 1988; Matenco and
Schmid, 1999; Fuegenschuh and Schmid, 2005). Paleogene to Early
Miocene rotation of the Cretaceous belt around the western edge of the
Moesian promontory led to orogen-parallel extension which culminated in the Late Eocene and large displacement along curved strike–slip
faults during the Oligocene to Early Miocene (Cerna Jiu and Timok fault,
Fig. 1) Locally, the exhumation induced by the uplift was large enough to
be recorded by apatite fission track data (Fuegenschuh and Schmid,
2005 and references therein).
In the Moesian foreland plate, the transcurrent motions led to the
transtensional opening of the Getic Depression as a dextral pull-apart
basin during the Early Miocene. In the course of the Miocene, the
tectonic regime changed to transpression, with the strike slip motions in
the western Getic Depression gradually changing to thrusting in the east
(Rabagia and Fülöp, 1994; Rabagia and Matenco, 1999). Thrusting of the
belt onto the Moesian Platform led to additional subsidence in the South
Carpathian foredeep. The effective elastic thickness of the Moesian
lithosphere has been estimated at 30 km (e.g. Cloetingh et al., 2006).
The ongoing Neogene E-ward movement of the internal Carpathian
units, driven by slab roll back, led to extension in the intra-Carpathian
area, affecting both Tisza–Dacia in the south and the Alcapa block in the
north (Ustaszewski et al., 2008). The formation of the Pannonian basin
system by rifting and transtension (Horváth et al., 2006) was controlled
by three main tectonic processes, gravitational collapse, subduction
rollback and asthenosphere updoming, the relative importance of which
is still a topic of intense debate (Cloetingh et al., 2006 and references
therein). In particular the latter process led to weakening of the
lithosphere in the Pannonian Basin, its effective elastic thickness being
estimated at 5–10 km by Lankreijer et al. (1997). In the last stage of its
evolution, basin formation and extension in the Intra-Carpathian region
have come to an end. Structural inversion of the Pannonian basin is
currently in progress, driven by the push of the Adriatic plate (e.g. Bada
et al., 1999) and evidenced by significant late-stage uplift and
subsidence anomalies during Late Pliocene through Quaternary times
(Horváth and Cloetingh, 1996).
2.2. Paleogeographic evolution
2.2.1. Biostratigraphic constraints on connectivity
Until the Middle Sarmatian (Bessarabian, Fig. 2), the Central and
Eastern Paratethys were connected over the South Carpathian orocline
(Fig. 1), as evidenced by the brackish water fauna common to Central
and Eastern Paratethys found in the intramontane basins on its west
flank (Marinescu et al., 1981 and references therein; Gagic, 1997).
During the Sarmatian, uplift of the South Carpathian orocline led to
the disconnection of the two basins. Open water communication
ceased and Central Paratethys continued its evolution as the endemic
Lake Pannon, the divergence in the faunal evolution due to the
separation defining the onset of the Pannonian stage. In the Dacic
Basin the Sarmatian stage continued into the Khersonian (Fig. 2).
During the Pannonian stage, corresponding to the Upper Sarmatian
and Meotian of the Dacic Basin (Fig. 2), the fauna in the two basins are
different (e.g. Rögl, 1996; Rögl, 1999). However, in Meotian sediments
Fig. 1. Study area. a. Late Miocene (Messinian) paleogeography of Paratethys and the Mediterranean showing location of gateways and study area. Modified from Popov et al. (2006).
b. DEM of the Pannonian–Carpathian area. Blue lines indicate the extent of Lake Pannon (after Magyar et al., 1999a) and the Dacic Basin (after Saulea et al., 1969) at different time
steps postdating the mid-Sarmatian disconnection of the two domains (M–U Sm, Middle–Upper Sarmatian; Meot, Meotian; U Pont–Dc, Upper Pontian–Dacian; see time scale in
Fig. 2). White lines are drainage divides and indicate the boundaries of the present day Danube drainage basin. Black rectangle represents the modeled area. Within the Danube
drainage basin, the Carpathians and northern Balkans divide the Dacic (200 × 103 km2) and Pannonian (600 × 103 km2) realms. Note the difference in the extent of the drainage
(and sediment source) areas of the modeled parts of the Pannonian and Dacic basins. The Danube crosses the drainage divide between the Pannonian back-arc and Dacic foredeep
realms in the South Carpathian Orocline.
178
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 2. Time scales. a,b. Correlation charts between Central and Eastern Paratethys time scales after Rögl (1999); Steininger et al. (1990); Sacchi et al. (1999) and Vasiliev et al. (2004, 2005). c. Correlation chart of 2nd and 3rd order sedimentary
sequences in the Pannonian Basin (Hungary). Data on sedimentary sequences from (1) Transdanubia (Sacchi et al., 1999); (2) entire Hungary (Juhász et al., 1999); and (3) Eastern Hungary, (Ujszászi and Vakarcs, 1993; Vakarcs et al., 1994).
Sequences were defined from seismic (1, 3) and well data (2); their ages constrained by magnetostratigraphy or by direct correlation to the Haq et al. (1987) eustatic curve (3). For reference the central Paratethys time scale is shown (left hand
panel, ages according to Sacchi et al., 1999; Sacchi and Horváth, 2002). MSC indicates the Messinian Salinity Crisis (5.96–5.33 Ma, Krijgsman et al., 1999). From the lower resolution well data (2), only 2nd order sequences were distinguished,
bounded by significant hiatuses. Sacchi et al. (1999) correlate their 3rd order sequences PAN-1 to PAN-4 with the “Late Miocene sequence” of (2) and sequences 5–8 or IV–VII of Ujszászi and Vakarcs (1993) and Vakarcs et al. (1994) respectively;
the correlated sequences are marked in light grey shading. The large age variations for the base Pannonian unconformity illustrate the correlation problems even within the Pannonian Basin. The 3rd order sequence boundaries (SB) associated
with–according to the authors–the largest base level falls, are indicated with a bold line: SB PAN-3 in Transdanubia at 8.7 My (1) and SB #8 in eastern Hungary (3).
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
in the Dacic Basin, Pannonian-derived Congeria, typical for freshwater
environments, have been found (Olteanu, 1979), suggesting a fluvial
influx from Lake Pannon into the Dacic Basin.
In the Lower Pontian (Eastern Paratethys definition, Fig. 2), the fauna
in the Pannonian, Dacic and Euxinic realms again show common
characteristics. The reconnection has been attributed to the “Odessian
transgression” (Olteanu and Jipa, 2006). Pannonian type fauna are found
over the entire Paratethys. However, no Eastern Paratethys forms have
been found in the Central Paratethys (Magyar et al., 1999b).
2.2.2. Seismic sequence stratigraphic constraints on base level changes
Lake Pannon (Magyar et al., 1999a) inherited its complex bottom
topography from the rifting and transtension stage that led to the
opening of the Pannonian Basin. Subsidence rates and water depth were
therefore different in the individual subbasins. Disconnected from the
world oceans during the Pannonian endemic stage (Fig. 2), the water
level of the lake was controlled by the balance between evaporation and
water supply from contributing rivers and precipitation (GarciaCastellanos, 2006), both controlled by climatic variations (Kázmér,
1990; Juhász et al., 1999). The surrounding mountain chains provided
abundant sedimentary influx. The basin was progressively filled by
fluvial–deltaic and turbiditic sediments dominantly from northwesterly
(palaeo-Danube) and northeasterly directions (palaeo-Tisza) (Vakarcs
et al., 1994).
The variations in rate of progradation to aggradation of the deltafed shelf-slope systems, led to the recognition of third and fourth order
seismic sequences (Vakarcs et al., 1994; Sacchi et al., 1999) and indicate that the relative lake level varied significantly both in space and
time. The most detailed sequence stratigraphic studies have been
carried out in the Hungarian part of the basin, both to the west of the
Danube in Transdanubia (Ujszászi and Vakarcs, 1993; Sacchi et al.,
1999) and to the east in the Great Hungarian Plain (Pogácsás et al., 1990;
Pogácsás et al., 1992; Csató, 1993; Vakarcs et al., 1994; Csató et al., 2007;
Juhász et al., 2007). Estimates of the Pannonian lake level change vary
from “tens of meters” to 200 m. Due to the scarcity of reliable age
constraints, the correlation between the sequences is not straightforward (Fig. 2).
The northwestern part of the Dacic Basin is coincident with the Getic
Depression, its subsidence history controlled by transtension and
subsequent foreland flexure (cf. Section 2.1). From a detailed study of
seismic sequence stratigraphy at the margins of the Getic Depression,
Rabagia and Matenco (1999) concluded that the observed base level
changes until the Late Miocene were predominantly tectonically
controlled, eustatic changes playing only a subordinate role. A regional
seismic sequence stratigraphic interpretation in the western part of the
Dacic Basin focused on the Latest Miocene–Pliocene basin fill evolution
(Leever et al., 2009). In this stage of its evolution, in contrast to the
earlier stage and the Pannonian Basin, control by differential tectonic
subsidence was less important. A sudden base level fall in the lower part
of the Pontian was attributed to the Messinian lowering of the water
level in the Black Sea below the threshold of the Dobrogea barrier. Late
Pontian sediments associated with a subsequent base level rise
transgressively cover the older deposits (Leever et al., 2009).
3. Numerical modelling of the signature of changing connectivity
on the sedimentary record
We use numerical modelling to study the factors influencing
connectivity and determine its signature on the sedimentary record.
The conceptual model considers two adjacent basins separated by an
elevated barrier, with or without a pre-existing gateway. In the latter
case, a fluvial connection is established by capture of the upstream basin
as the result of its higher lake level and the erosion of the barrier.
Coupled forward modelling of tectonics, surface processes and isostasy
is used to determine the response of the system in terms of depositional
geometries to parameters such as tectonic uplift, downstream base level
179
changes and flexural rigidity, the importance of which in controlling
lake capture and the evolution of tectonic lakes has been established in
previous studies (Garcia-Castellanos et al., 2003; Garcia-Castellanos,
2006; Leever et al., 2009).
3.1. The numerical method: TISC
TISC is a pseudo-3D (planform) forward finite difference code
(Garcia-Castellanos, 2002; Garcia-Castellanos et al., 2003) in which
tectonics, surface processes and flexural isostasy are fully coupled. It is
designed to study the interaction between surface mass redistribution
and the lithospheric response by uplift and subsidence on large
temporal (105–106 y) and spatial scales (of an entire sedimentary
basin and/or orogenic belt). The rate of surface uplift (and subsidence) is
thus a function of the rate of tectonic uplift, erosion and sedimentation,
and flexural isostatic rebound. It does not allow–nor do we aim–for
detailed reconstructions of morphological evolution or prediction of
higher than first order sedimentary sequences.
Tectonic uplift (of the barrier, in our case) is kinematically defined.
Resulting surface mass redistribution is calculated at time steps of
1000 yr following the stream power-law formulation by Kooi and
Beaumont (1994) including short-range diffusive and long-range linear
transport functions that represent hillslope and fluvial processes
respectively. Fluvial transport is calculated by explicitly calculating the
drainage network during the topographic evolution, accounting for the
formation of lakes in local topographic minima. These lakes become
closed (endorheic) if evaporation in its surface becomes larger than the
water they collect. The time step for imposed tectonic uplift and
calculated flexural isostatic response is 0.5 Ma. Flexural calculations
follow an elastic thin plate model characterized by the effective elastic
thickness Te, and account for the loading of sediment and water as well
as unloading due to erosion.
3.2. Modelling setup and boundary conditions
The model setup and boundary conditions (Fig. 3) are derived from
the main characteristics of the study area (Fig. 1). The Pannonian and
Dacic basins are in very different stages of their tectono-thermal
evolution, and as a consequence have greatly different lithospheric
rigidities (Te of 5–10 and 30 km respectively; Cloetingh et al., 2006). The
basins are in restricted connection over a topographic high, with only a
small elevation difference between them. The drainage area and hence
the initial sediment supply in the Pannonian basin were much larger
than those of the downstream Dacic basin (Fig. 1).
In the model, where the area of both basins is equal, the larger
drainage area and sediment supply in the upstream basin are
represented by a water and sediment input at the western edge of the
model (Fig. 3). The surrounding topography prevents any sediment
from leaving the model domain. No vertical movements due to flexure
are allowed at the model edges, except on the eastern side. This is based
on the assumption of a steady-state topography in the mountains
surrounding the basins, while the Dacic Basin has a larger extent
eastward than its model counterpart (Fig. 1b). In the model, the
sediment loading is larger than the erosion due to the external sediment
source, and these flexural boundary conditions are required to prevent
excess downward deflection. The parameters used by the surface
process model are listed in Table 1 and have been validated in previous
modelling studies (e.g. Garcia-Castellanos et al., 2003). The models were
run for a period of 15 My. All model runs have a 400 × 200 km model
domain and 2 × 2 km grid cell size (Fig. 3). In our reference model, the
initial maximum elevation of the surrounding topography is 800 m and
the maximum initial depth of the basins is 400 m. Tested variations of
this scenario are summarized in Table 2. An example of model evolution
is shown in Fig. A1.
180
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 3. Model setup. a. 2 km resolution DEM of the modeled area. b. Initial topography and boundary conditions for the conceptual model series: 400 × 200 km, 2 km grid cells.
Topography in this figure corresponds to the initial elevations for Model 2, 5 and 6. The geometry of the barrier separating the two basins is different for the various model runs, see
Table 2. All model boundaries are fixed for deflection, only the eastern side is allowed to move freely. A sediment source on the western side of the model (1.3 103 km3/Ma)
represents the flux from the large Pannonian drainage area. Sediment is allowed to leave the model at all sides. c. Model setup and parameters in cross section. Qs_ext, external
sediment input; P, precipitation rate; E, evaporation rate; S, slope; H, max. elevation of surrounding topography; Hb, elevation of barrier; Hg, elevation of gateway; dH, elevation
difference between base levels at time of capture; U, uplift rate. Parameter values are listed in Table 1 and 2.
3.3. Results
3.3.1. Model 1—reference model
In the reference model, the basins are separated by a threshold at sea
level. The surrounding topography is 800 m (Fig. 3, Table 2). Model
results after 15 My are shown in Fig. 4.
3.3.1.1. Rates of sedimentation and erosion. In the upstream basin,
sedimentation initially occurs at a higher rate than in the downstream
basin (Fig. 4a, b). This is due to the external sediment source that feeds
the upstream basin (Fig. 3), while the downstream basin is initially
only sourced by erosion from the surrounding mountains. Between
6.5 and 9 My, the relative sedimentation rates are inverted: the
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Table 1
Parameters for the surface process model. Transport and diffusion coefficients and erosion
and deposition length scales according to the formulation of Kooi and Beaumont (1994).
Parameters
Values
Model resolution (grid cell size)
Transport coefficient Kf
Diffusion coefficient Ks
Surface processes time step
Tectonic/isostatic time step
Erosion length scale lf
Basement
Sediments
Deposition length scale lf
2 × 2 km
60 kg/m3
0.05 m2/s
0.05 My
0.5 My
120 km
60 km
25 km
upstream basin is filling up and excess sediment is transferred to the
downstream basin. In the upstream basin, the sedimentation rates
decrease to zero as the basin is completely filled (Fig. 4).
The erosion rates show an overall decreasing trend in both basins,
due to the declining topography. Erosion rates are higher in the
upstream basin, which is surrounded by topography on three sides
while the downstream basin is lined by topography only on the N and S
side (Fig. 3). The initial decreasing trend in sedimentation rates in the
upstream basin (0 to 6 My, Fig. 4b) reflects the decrease in supply from
local erosion. The sedimentation rates in the downstream basin,
however, do not reflect the local erosion rates: some sediment is
transferred from the upstream source into the downstream basin also
before the bulk shift in sedimentation between 6.5 and 9 My.
3.3.1.2. Depositional geometries. The increase in sedimentation rates in
the downstream basin is evident from the increasing spacing between
the 0.5 My time lines from 6 My (Fig. 4c). Also the upstream basin shows
locally increased sedimentation rates after 6 My, when looking at the
time lines. The local increase is due to the decreased accommodation
space and is not reflected in the sedimentation rates integrated over the
entire upstream basin (Fig. 4b–c). Sediment loading resulted in flexural
isostatic subsidence in the entire model area: more than 400 m in the
basins (the basin floor was initially at –400 m, Table 2) and some 200 m
on top of the “barrier” (Fig. 4c). In the upstream basin, subsidence led to
a concave shape of the initially flat shelf edge trajectory.
181
3.3.2. Effect of the barrier geometry (models 2-3)
The effect of barrier geometry is studied in different ways. In Model
2, the basins are connected across an existing gateway, the elevation of
which is varied. In the subsequent models, the basins are initially
disconnected and the elevation and slope of the barrier are varied.
3.3.2.1. Model 2: elevation of a pre-existing gateway. Model 2 addresses
the effect of the initial elevation of a pre−existing gateway in the barrier
between the two basins (Table 2, Fig. 3). The results are shown in Fig. 5,
with those of Model 1 for comparison.
3.3.2.1.1. Rates of sedimentation and erosion. During the first 7 My,
the gateway elevation has virtually no effect on the infill of the basins:
the cumulative sedimentation is equal for the three cases (Fig. 5a). Only
in the second model stage, after the shift of the bulk sedimentation to
the downstream basin, do the models show some difference in sediment
volume. The difference with the reference model (in which no barrier is
present) is more pronounced than the difference between the individual
cases: the bulk sedimentation shift occurs earlier for all cases in Model 2
than in the reference model. Moreover, sedimentation continues
afterward, albeit at a lower rate.
An initial difference in erosion rates in the upstream basin (Fig. 5b),
reflecting the initial difference in base level controlled by the elevation
of the outlet of the basin, disappears after 2 My. This is the time required
to bring the gateway to an equal elevation for all three models, by
allowing the river in the gateway to become graded in a dynamic
equilibrium. In the downstream basin, the erosion rates are exactly
equal for all three models during the entire model time (Fig. 5b).
3.3.2.1.2. Depositional geometries. The depositional geometries in
Model 2, i.e. the shelf edge trajectory and the spacing between time
lines, are very similar to the reference model. A major difference is that
sedimentation in the upstream basin continues after the bulk
sedimentation shift, accommodating a large volume of fluvial sediments
(Fig. 5, subparallel subhorizontal time lines). The volume of continental
sediments in the three models is different, however; the largest volume
being accommodated by the model with the initially highest gateway
(Fig. 5c). The fluvial sedimentation and the difference in volumes can be
explained from the river profiles at the final time step. The eroding
barrier between the two basins is uplifting by flexural isostatic rebound.
This is the main difference with the reference model, in which the
“barrier” was initially at sea level, was loaded by sediments and, as a
Table 2
Model setup.
Geometry
Max barrier
uplift rate
U (m/Ma)
Te (km)
W–E
Rate of sea
level change
(m/Ma)
–
2.7
–
–
10
10
–
–
–
10
–
–
0.75
2
3
2.7
10
–
800
200
2.7
100
200
400
–
–
800
800
200
2.7
–
400
400
–
2
–
30
20
10
5
20–30
10–30
5–30
10
Max elevation
H (m)
Barrier elevation
Hb (m)
Gateway elevation
Hg (m)
Max slope
S (degrees)
1 Reference model (Fig. 4)
2 Gateway elevation (Fig. 5)
800
800
0
800
3 Barrier slope (Figs. 6, 7)
400
400
–
0
100
200
–
4 Uplift rate (Figs. 8, 9)
800
0 to 400
5 Te constant (Figs. 10, 11)
800
6 Te varied (Figs. 11, 12)
7 Sea level change (Fig. 13)
–
200
100
66.6
50
182
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
result, subsided. The gateway between the two basins is more resistant
to erosion than the sediments in the basins (Table 1), resulting in a
knickpoint in the river profile. Upstream of this knickpoint, i.e. in the
upstream basin, the river accumulates sediments in order to maintain its
gradient and keep the ability to cross the barrier (see also Snow and
Slingerland, 1990).
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
3.3.2.2. Model 3: slopes of the barrier. In this model, in contrast with the
previous models, there is no pre-existing gateway of reduced elevation
as a way of communication between the basins. The surrounding
topography is 400 m (Table 2). The effect of barrier slope was
addressed in three different models, with a slope on the downstream
side of the barrier of max 0.75, 2 and 3° respectively. The slope on the
upstream side of the barrier is 2°, in order to keep the volume of the
upstream basin constant. Sedimentation and erosion rates are shown
in Fig. 6, profiles in Fig. 7.
3.3.2.2.1. Rates of sedimentation and erosion. The model with the
lowest slopes has the largest exposed area and hence the highest
erosion rates in the downstream basin (Fig. 6). The sedimentation
rates in the downstream basin are initially equal to the local erosion
rates: the only supply to the downstream basin is from the local
erosion. The moment of capture (Fig. 6, marked by x) may be
recognized from the breakdown of this relationship: sedimentation
rates start increasing while the erosion rates keep decreasing. Capture
occurs first for the model with the steepest slope (Fig. 6): due to the
steeper slope the barrier is narrower and less erosion is required to
oversteepen the slopes at the drainage divide. This asymmetry allows
the local backward migration and lowering of the drainage divide and
thereby capture of the upstream lake—of which the level has risen to
the top of the barrier (Fig. 7).
It is important to notice that, after capture of the upstream basin,
the sedimentation rates in the downstream basin initially increase
only gradually. The strongest increase occurs later and is due to the
overfilling of the upstream basin. The capture-induced increase in
sedimentation rates in the downstream basin is not reflected by a
similar decrease in the upstream basin: the water level drop that
results from the capture leads to a larger exposed area and therefore to
increased supply, balancing the sediment outflow to the downstream
basin (Fig. 6). Moreover, the capture time influences the timing of the
bulk sedimentation shift, because capture and subsequent lowering of
the outlet by erosion lead to lowering of the base level in the upstream
basin and reduction of the accommodation space. As such, the sooner
the capture, the sooner the upstream basin is filled, and the sooner the
sedimentation shifts to the downstream basin.
The lacustrine–continental transition in the upstream basin occurs
during or after the strongest decrease in sedimentation rates.
Compared to the previous model series, only a very small amount of
continental sediments are deposited in the upstream basin (compare
Fig. 5c and 7), before sedimentation rates drop to zero or even become
negative (implying erosion of previously deposited sediments).
3.3.2.2.2. Depositional geometries. The implications of lake capture
on depositional geometry are shown in more detail in Fig. 7 for each of
the three scenarios. In the left series of panels the sediment geometry
in the upstream basin before capture and resulting erosion is shown.
The initial progradation, seen from the shelf edge trajectory, is due to
the high sediment input close to the source (Fig. 3). The sediments are
progressively spread over a larger area and consequently develop a
more aggradational character, also due to the larger subsidence in the
center of the basin.
Capture occurs first for steeper slopes (Fig. 7c). Following capture
(time line highlighted in Fig. 7, right hand panels), sediments in the upstream basin are deposited in a downstepping geometry during the
ongoing water level fall that results from erosional lowering of the outlet,
lasting ∼2 My in all cases. The sediment accumulation area in the
upstream basin is reduced by the water level fall, resulting in a local
183
increase in sedimentation rates (see spacing of time lines in Fig. 7, right
hand panels) even though the total rate in the whole basin remains
largely unchanged. The downstream basin shows higher sedimentation
rates (wider spacing of the time lines in cross section, Fig. 7, cf. Fig. 6)
after capture.
The moment of capture determines the total amount of sediments
that can be accommodated in the upstream basin, and therefore the
amount of sediments transferred to the downstream basin. The longer
the duration of the isolated stage of the upstream basin, with its elevated
base level, the more sediment can be accommodated. The duration of
the lacustrine stage in the upstream basin is therefore most extended in
case of the low slope model (upper panels): the upstream basin is not
completely filled even after 15 My. Delay in the moment of capture, i.e. a
longer isolated life time of the lake at local elevated base level, leads to a
higher degree of overfilling. The resulting high river gradient after
capture causes erosion rather than fluvial sedimentation in the
upstream reaches in order to attain equilibrium (compare Figs. 5, 7).
3.3.3. Effect of barrier uplift rate (Model 4)
In this model, the two basins are initially separated by a barrier at
0 m (as in the reference model), which is subsequently uplifted. The rate
of the uplift is varied in the three cases while its magnitude is constant at
400 m (Table 2). The width of the uplifting zone is ∼50 km and has
slopes of ∼3°. The base level or maximum lake level in the upstream
basin is kept at 200 m by an outlet at the N margin of the upstream basin,
simulating the effect of the precipitation–evaporation ratio in the
Pannionian Basin, which has a larger area in nature than in the model.
The surrounding topography is 800 m. Results are shown in Figs. 8 and 9.
3.3.3.1. Rates of sedimentation and erosion. The conditions and initial
model evolution, before the onset of uplift at 2 My, are equal to Model 1.
During this stage, some sediment is transferred from the upstream to
the downstream basin, shown by the sedimentation rates in the
downstream basin that exceed the local supply (Fig. 8b). In the absence
of an initial gradient, the two basins are disconnected the moment the
barrier rises above 200 m, the maximum lake level of the upstream
basin.
Uplift of the barrier leads to an increase of the erosion rates in the
downstream basin (Fig. 8a), while the erosion rates in the upstream
basin decrease. The changes are most pronounced for the models with
the high uplift rates.
The approximately constant sedimentation rates in the downstream
basin during the uplift period for low uplift rates (U = 100 m/Ma,
Fig. 8b) reflect the continuous connection between the basins,
maintained by the barrier erosion that is able to keep up with the uplift.
The sedimentation rates in the downstream basin start increasing
gradually after the end of the uplift stage from 6 My on, due to erosional
lowering of the barrier and increasing sediment transfer from the
upstream basin. For the other models (U = 200, 400 m/Ma), the
sedimentation rates in the downstream basin match the local supply
rates from the moment of disconnection onward (Fig. 8b). Capture in
these models (marked by x in Fig. 8b) directly leads to the onset of the
bulk shift in sedimentation between the basins, as shown by the sudden
and fast increase of sedimentation rates in the downstream basin.
3.3.3.2. Depositional geometries. Sedimentation patterns are shown in
Fig. 9. In the left panels, the time step preceding capture is shown. Note
that the basins in the low uplift rate scenario (U = 100 m/Ma, upper
Fig. 4. Model 1: reference model. a. Cumulative sediment volumes, derived from local erosion and external sediment input. b. Rates of sedimentation (bold lines) and erosion (thin lines).
The sedimentation rate is the rate of variation of the total sediment volume, while the erosion rate represents the total eroded volume per time step, including bedrock and previously
deposited sediments. Note that the initial sediment accumulation rates in the upstream basin greatly exceed the erosion rates due to the external sediment input (see Fig. 3) Grey shading in
a. and b. marks the period of bulk sedimentation shift from the western to the eastern basin, characterized by rapidly changing sediment accumulation rates and related to the lacustrine–
continental transition in the upstream basin (t = 8 My). The (restricted) connection between the two basins allows some sediment transfer from the upstream to the downstream basin
from the start, as seen from the sedimentation rates in the downstream basin that exceed the local supply (erosion) rates. c. W–E cross section at the center of the model (y = 0, cf Figure A1)
at the final time step (15 My) with time lines for each 0.5 My. The onset of increased sedimentation rates in the downstream basin (t = 6.5 My) is marked by a bold line; the final stage of
infill of the upstream basin (t = 8 My) by a heavy broken line. Shelf-slope break indicated by short dash in both basins.
184
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 5. Model 2: effect of gateway elevation. a. Cumulative sediment volumes. For comparison the results of Model 1 are shown. In contrast to Model 1, sedimentation continues in the upstream basin after the bulk sedimentation shift, albeit at
lower rates. The transition from lacustrine to fluvial sedimentation is marked by box (cf Fig. 5c). b. Erosion rates. Rates are equal in the downstream basin for all model runs (dashed lines), independent of the gateway elevation, but different in
the upstream basin during the first 2 My. c. W–E cross sections at t = 15 My. Lacustrine–continental transition is indicated by bold line, shelf-slope break by short dash. The volume of continental sediments deposited in the upstream basin
since 8 My is largest for the model with the initially highest gateway elevation of 200 m (lower panel).
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
185
Fig. 6. Model 3: effect of barrier slope. Sedimentation and erosion rates indicated by bold and thin lines respectively. High slopes correspond to a narrow barrier (see also Figs. 3c, 7),
and lead to faster capture of the upstream basin (marked by X). Capture results in increasing sedimentation rates in the downstream basin, which were initially declining and equal
to the local erosion rates. Increasing erosion rates in the upstream basin after capture reflect the increased exposed area due to the falling base level. The largest increase in
sedimentation in the downstream basin corresponds to the end of lacustrine sedimentation and filling up of the upstream basin (marked by arrows).
panels) were repeatedly connected and disconnected during the period
of uplift, as incision was able to keep up with uplift. The uplifted,
elevated barrier represents a load that causes deflection of the basin
floor adjacent to the barrier (compare Fig. 9, upper left panel and Fig. 7,
lower left). The sedimentation patterns in the upstream and downstream basins subsequent to capture are similar to those in Model 3 (see
Fig. 7). However, in Models 4b and 4c, because the barrier rises to an
elevation exceeding the maximum upstream lake level, and because–
despite the subsidence induced by the barrier uplift–the available
accommodation space preceding capture is smaller (base level at 200 m
instead of 400 m), the basin fill is in a more advanced stage at the
moment of capture than in Model 3. The capture is followed by an
immediate strong increase in sedimentation rates in the downstream
basin (lower panels, see also Fig. 8b), instead of the more gradual
increase in sedimentation rates characterizing the low uplift rate of
Model 4a, and also Model 3 for low slope values.
3.3.4. Effect of flexural rigidity (Models 5–6)
In these models, the effect of flexural isostatic response is studied
by varying the effective elastic thickness Te.
3.3.4.1. Model 5: constant Te. The initial setup of this model is equal to
Model 2 (Fig. 3) with an initial gateway at 200 m elevation in an
800 m surrounding topography. Based on estimates of lithosphere
rigidities (Lankreijer et al., 1997; Cloetingh et al., 2006), we varied Te
values between 5 and 30 km (Table 2), constant over the entire model
area. The results are shown in Fig. 10.
In Fig. 10a the sedimentation rates in the upstream and downstream
basin are compared for different Te. Onset of the (enhanced) decrease of
sedimentation rates in the upstream basin occurs first for the most rigid
plate (Te = 30 km, at ∼3.5 My) and results in a very pronounced and
sudden increase in sedimentation rates in the downstream basin
(Fig. 10a, peak at 5.5 My). The required time to completely fill up the
upstream basin is shortest on the highly rigid plate (Figs. 10 and 11). For
Te= 30 km, even the downstream basin is filled up by 10 My. For the
weaker plates, the onset time of decreasing sedimentation rates in the
upstream basin is later, and the change more gradual. Much more time is
required to completely fill the basin: the low rigidities allow the
generation of significant additional accommodation space by sediment
loading. The onset of fluvial sedimentation in the upstream basin occurs
just before the peak in sedimentation rates in the downstream basin for
all models. Sedimentation rates in the downstream basin start gradually
increasing long before the onset of the fluvial stage in the upstream
basin. In Fig. 10b, the time to the lacustrine/continental transition and
the complete filling of the basin is plotted as a function of Te. In case
of a weak lithosphere, the basin infill not only takes much longer to
complete, but it is also preceded by a protracted stage of fluvial sedimentation (cf. Fig. 11, upper panels).
3.3.4.2. Model 6: lateral transition in Te. These models address the
effect of a lateral (W–E) change in Te. Based on the lithosphere
rigidities in the Pannonian and Dacic Basins, the Te in the east is kept
at 30 km, while its value in the upstream basin is varied between 5
and 30 km (Table 2). The transition is gradual over a distance of 60 km
below the barrier.
The results of this model (Fig. 11, lower panel; Fig. 12) are similar to
those of the previous model. The rigidity of the lithosphere underlying
the upstream basin determines its infill time, and as such the time and
rate of the shift of bulk sedimentation to the downstream basin. Peak
sedimentation in the downstream basin is again most pronounced in
case of a rigid upstream lithosphere. The bulk sedimentation shift occurs
∼0.5 My earlier for all scenarios than in the previous model, because of
reduced subsidence of the downstream margin of the upstream basin.
3.3.5. Effect of downstream base level change (Model 7)
In this model, the basins are initially disconnected by a 400 m high
barrier (Table 2). A sea level fall of 200 m starts at t = 0 My and
proceeds at different rates, until the initial level is reached again at
8 My (Fig. 13a). In case of the fastest sea level fall, this leads to a 6 My
lowstand period. The 8 My time interval was chosen based on the
results of Model 3, which has the same setup except the sea level
change, and where capture is predicted at t = 7.5 My (Fig. 7).
The results are presented in Fig. 13b, which shows only the
sedimentation and erosion rates in the downstream basin. The models
show an initial increase in erosion and sedimentation rates, because the
falling water table leads to the exposure of a larger area. The rate of
change reflects the rate of the sea level fall. The sedimentation and
erosion rates decrease again during the subsequent sea level rise.
186
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 7. Model 3: effect of barrier slope (continued). Cross sections for Model 3. Vertical scale in m, horizontal in km. The left panels show the time step (in My) preceding the capture
of the upstream basin. The corresponding time line is marked by a bold line in the right hand panels (cross sections at the model end time, 15 My). The other highlighted (long dash,
dotted) time lines represent the end of base level fall and the final stage of lacustrine sedimentation in the upstream basin respectively. For the low slope scenario (S = 0.75°), the
western basin is not completely filled after 15 My. The shelf-slope break is indicated by a dashed line in all panels. Note the difference in depth to basement in the upstream basin and
the (apparent) increase in sedimentation rates in both basins after capture. Further discussion in the text.
Capture of the upstream basin (marked by arrows, Fig. 13) occurs only
after the sea level has returned to its initial level and is likely triggered by
the loading effect of the rising sea level which subdues the incised
barrier. From these results it follows that a base level fall does not
accelerate the capture of the upstream basin. To the contrary, the
capture occurs 0.5 to 1 My later than in Model 3 (without sea level fall).
4. Discussion
4.1. Controls on depositional geometry and sediment partitioning
The individually tested parameters in the different model runs lead to
marked differences in the patterns of sedimentary infill in the two basins.
However, we found that the tested parameters are not critical by
themselves, but only to the degree by which they influence the true
parameters that control the downstream sedimentation rates in relation
to changes in connectivity. These are the elevation difference and upstream accommodation space at the time of capture, outlined in Fig. 14.
The difference between the water levels in the two basins before
connection has a large effect on the depositional geometries in the
upstream basin since it determines the magnitude of the capture-induced
upstream base level fall by erosional lowering of the outlet. The elevation
difference is dependent on the barrier geometry and uplift rate: a large
elevation difference at the time of capture is more likely for a wide barrier,
high uplift rates and low erodibility (Model 3, 4). The remaining
accommodation space in the upstream basin at the time of capture
determines the time lag for the bulk shift of sedimentation to the
downstream basin (e.g. Model 1, 2 in Figs. 4 and 5). The remaining
upstream accommodation space is not an independent parameter, it is
dependent on the timing of capture and moreover it correlates with the
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
187
Fig. 8. Model 4: effect of uplift rate. a. Erosion rates. Uplift period is indicated by bars. The vertical scales for the respective basins are offset by 100 m for better distinction.
b. Sedimentation rates. Crosses mark moment (period) of capture. Legend as in 8a.
elevation difference between the basins. Generation of accommodation
space by sediment loading is enhanced by low Te values (Models 5 and
6).
These parameters are discussed in more detail below for different
stages in basin evolution: the disconnected (endorheic) and connected
lacustrine stages and the subsequent (upstream) continental stage.
Though the modelling setup was based on the characteristics of the
Pannonian–Dacic realm, the general results may be extended to other
areas where elevated topography divides two sedimentary basins, such
as the Ebro Basin (foredeep to the Pyrenees) and the adjacent
extensional Valencia trough (Garcia-Castellanos et al., 2003), separated
by the Catalan Coastal Ranges.
4.1.1. Disconnected stage and capture of the upstream endorheic lake
4.1.1.1. Upstream basin. The elevation difference between the basins
after capture (and therefore the magnitude of the base level fall) is
determined by the equilibrium gradient of the river connecting them
which is a function of the erodibility and the width of the barrier. If the
upstream basin is still in the deep lacustrine stage with ample
accommodation space at the time of capture, the capture-induced
base level fall results in the deposition of a strongly progradational series
with a downstepping geometry (e.g. Model 4, Fig. 9). In case of a large
elevation difference and/or limited remaining accommodation space at
the time of capture, the sedimentation rates in the upstream basin are
locally significantly increased because of the reduced sediment
accumulation area (Fig. 9) and capture may lead to accellerated infilling
of the remaining accommodation space, ending the lacustrine stage.
4.1.1.2. Downstream basin. The sedimentary response to capture in the
downstream basin is mainly sensitive to the remaining upstream
accommodation space. If the upstream basin is captured when it still has
ample accommodation space, the response in the downstream basin is
limited regardless of the elevation of the upstream basin. Though the
overall increase in sedimentation rates in the downstream basin after
capture is small in these cases (e.g. Model 3, Fig. 6: low uplift rate in
188
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 9. Model 4: effect of uplift rate (continued). Cross sections. Left panels show the time step preceding capture; right panels represent the final time step (15 My). Sediments
deposited during the uplift stage are shaded. Vertical scale in m, horizontal in km. In the right panels, the highlighted time lines represent the time step preceding capture, the end of
base level fall in the upstream basin, and the final stage of lacustrine deposition (solid, long dash and dotted lines respectively). The capture-triggered increase in sedimentation rates
in the downstream basin is much more pronounced for the higher uplift rates (200, 400 m/Ma): the moment of capture coincides with the last stages of lacustrine infill of the
upstream basin. No fluvial sedimentation is accommodated in any of the models.
Model 4, Fig. 8b), the effect is focused at the inlet and will be locally
significant (Model 3, Fig. 7). In contrast, the bulk sediment shift to the
downstream basin will be triggered directly by capture if the remaining
upstream accommodation space is small at the time of capture, and/or
sufficiently reduced by the capture-induced base level fall (high uplift
rates in Model 4, Fig. 8). The latter requires a large pre-capture elevation
difference between the basins.
In other words, lake capture is expressed in the sedimentary record
by affecting base level in the upstream basin and supply in the
downstream basin. The most significant increase in downstream
sedimentation rates, however, is not necessarily directly linked to the
capture event (Figs. 6 and 14).
4.1.2. Connected stage
The upstream accommodation space is controlled by changes in the
lake level as discussed above and, in the absence of tectonic subsidence,
the subsidence due to sediment loading. The latter is strongly dependent
on the rigidity of the lithosphere below the upstream basin: a weak
lithosphere (low effective elastic thickness, Te) allows more subsidence
and results in the generation of a larger accommodation space (Model
5). A strong lithosphere supporting the upstream basin leads to a sudden
increase in the supply to the downstream basin, while a low Te results in
a gradual increase in downstream sedimentation rates (Model 5 and 6;
Figs. 11–13).
4.1.3. Continental stage in the upstream basin: fluvial incision and deposition
The (post-capture) continental evolution of the upstream basin
and the gateway is determined by fluvial processes which are a
function of the difference between the actual and the equilibrium river
gradient: the river tends to its equilibrium gradient and does so by
either incision or deposition (e.g. Kooi and Beaumont, 1994). The
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
189
Fig. 10. Model 5: effect of lithosphere rigidity. a. Sedimentation rates for different values of effective elastic thickness (Te). Transition from lacustrine to fluvial sedimentation in the
upstream basin is marked by stars, the horizontal bars indicate the preceding period of increasing sedimentation rates in the downstream basin. Erosion rates in the downstream
basin are shown for reference. b. In the upstream basin: transition time from lacustrine to fluvial sedimentation, and fill-up time as a function of Te. t0 is the minimum time required
to fill the basin and is a function of the basin volume and the supply rate. Hatch: period of fluvial sedimentation.
response–incision or deposition–is determined by elevation and the
degree of overfill of the upstream basin at the lacustrine–continental
transition.
If capture occurs in a late stage, when the upstream accommodation
space is nearly filled, the upstream basin will be overfilled and exhumed,
a process much dependent on the elevation difference between the
basins. The larger the elevation difference, the larger the degree of
overfill and the larger the subsequent river incision (Model 3, 4). If
capture happens in an earlier stage, with ample upstream accommodation space, fluvial sedimentation is likely in case of low lithosphere
rigidities (Model 2 and Model 6, 7 in Figs. 11 and 12).
The volume of continental sediments is influenced by the initial
elevation of a preexisting gateway. The results from Model 2 show that,
due to the erosion-resistant barrier, the initial elevation of the gateway
(and as such, upstream base level) controls the upstream accommodation space and corresponding continental sedimentation in the last
stage of basin fill evolution, causing differences in the total deposited
volumes (Fig. 5). This is despite the fact that the same gradient is
established between the upstream and downstream basin in all three
models within the first 2 My (Fig. 5b).
Even after capture and completion of the lacustrine infill of the
upstream basin, the barrier keeps controlling the upstream base level by
forming a knickpoint in the river long profile.
The mechanisms underlying the knickpoint formation promote
upstream fluvial sedimentation as can be seen by comparing the results
from Model 1 and 2. In Model 2, the barrier, uplifting by flexural isostatic
rebound in response to erosion, separates two depocenters in the model
area. In order to maintain its gradient across the erosion-resistant uplifting
barrier, the river responds by upstream deposition (see also Snow and
Slingerland, 1990). The larger the upstream generation of accommodation
space by subsidence, the larger the accumulation of fluvial sediments. This
is well expressed by the results of Model 6 (Figs. 10 and 11), where low Te
values lead to large subsidence in the basins (and higher uplift of the
barrier), accommodating a large volume of continental sediments
compared to the high Te values. In case of Model 1, the initial elevation
of the barrier is at sea level over its entire length. It subsides with the sediment loaded adjacent basins, and is buried by sediment itself. The
wavelength of the subsiding area is different in this case, even though the
rigidity of the supporting lithosphere is the same in both models. The river
is able to maintain its gradient across the former barrier without any
deposition or incision and is consequently at grade over its entire length.
4.2. Implications for the evolution of the Pannonian–Dacic area
Since the Sarmatian, Lake Pannon and the Dacic Basin existed as two
water masses at a relatively small elevation difference (in comparison
190
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 11. Model 5, 6: lithosphere rigidity. Cross sections at 15 My. Vertical line marks change in orientation from W–E to S–N. Bold line indicates onset of fluvial sedimentation in the
upstream basin, short dash marks the shelf-slope break. a. Model 5, Te = 30 km. b. Model 5, Te = 5 km. Note the difference in volume of continental sediments with Fig. 5a. c. Model 6,
Te west = 5 km, Te east = 30 km. Te transition over 60 km across the center of the model is marked schematically at the bottom of the panel.
with e.g. the Ebro Basin and the Mediterranean, see Garcia-Castellanos
et al., 2003), separated by an elevated barrier. The modelling results allow
some inferences on the evolution of the Pannonian and Dacic Basins as a
function of their connectivity. As discussed above (Section 4.1, Fig. 14), the
modelling predicts that in a system such as the Pannonian–Dacic region,
changes in connectivity may be recognized mainly from base level
changes in the upstream basin and supply changes in the downstream
basin, though the two may be separated by a time lag.
4.2.1. Pannonian Basin
Constraints on base level evolution have been derived from
sequence stratigraphy (see Section 2). The rapid retreat of the coastline
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
191
Fig. 12. Model 6: lateral change in lithosphere rigidity. Sedimentation rates for different Te in the western (upstream) basin, Te in eastern basin is constant at 30 km (Table 2).
Transition to continental sedimentation in the upstream basin is marked by x. The lithosphere rigidity below the upstream basin influences the expression of the peak sedimentation
in the downstream basin, even though its Te is equal in all model scenarios.
of Lake Pannon between 6.5 and 4.5 Ma, as reconstructed by Magyar
et al. (1999a) (Fig. 1) closely matches in time the 2nd order sequence
boundary observed in the Hungarian part of the Pannonian Basin (Fig. 2)
and suggests that this sequence boundary is indeed associated with an
absolute lake level fall (rather than reduced tectonic subsidence).
Estimates of its magnitude vary from 10s of meters (Sacchi et al., 1999)
to 200 m (Vakarcs et al., 1994). The hiatus associated with this boundary
has been dated at 5–6 My (Csató et al., 2007).
Both capture (restoring connectivity) and enhanced incision by an
antecedent river will affect the Pannonian lake level. In case of lake
capture, the magnitude of the lake level fall would be determined by the
relative elevation of the Pannonian Lake with respect to the downstream Dacic Basin at the time of capture (cf Model 3). Alternatively, in
case of an existing fluvial connection, the adaptation of the river
gradient in the Danube gateway to the MSC-related lowering of the sea
level in the Dacic Basin would lower the Pannonian lake level as well
(see also Tari et al., 1992). This raises the question, which mechanism
caused the “Messinian” sequence boundary in the Pannonian Basin?
The lake capture model assumes that the Sarmatian uplift defeated
the river crossing the barrier and created an endorheic Pannonian Lake.
Once disconnected, any lake level changes were due to local climatic
factors (i.e., the balance between water influx and evaporation in the
drainage area; see e.g. Juhász et al., 1999). Capture of the Pannonian
Lake, restoring connectivity to the Dacic Basin, would lead to a base level
fall. A capture-induced Pannonian lake level fall needs not be associated
with a dry climate. To the contrary, capture was likely triggered by an
increase in the precipitation/evaporation ratio in the Pannonian Basin,
inducing an initial rise in the lake level. This rise might trigger
reconnection by “overspill and eventually lead to a lake level fall instead
(see also Garcia-Castellanos (2006) for a more detailed discussion on
the parameters influencing lake capture).
Large uplift rates delay or even inhibit the reconnection (Figs. 8 and
9), while the rise of the upstream lake level is a prerequisite (GarciaCastellanos, 2006). Lake level changes in the Pannonian Basin in the
order of 10s of meters (Sacchi et al., 1999) imply that the Sarmatian
uplift of the South Carpathian Orocline was limited, otherwise capture is
impossible.
However, Sarmatian apatite fission track ages along the Danube
gorges which are interpreted to result from a tectonic uplift phase (Bojar
et al., 1998) argue for a much larger uplift, in which case a major rise in
the Pannonian lake level would have been required for capture to occur.
These data are in conflict and contradict the capture model. However,
lateral differences in erodibility in the South Carpathian orocline along
the major strike slip faults (such as the Cerna–Jiu fault, Fig. 1) may have
contributed to keeping the actual “barrier” at reduced elevation by
facilitating river incision despite large tectonic uplift. In addition, rapid
localized incision would reduce the isostatic rebound due to erosional
unloading and thus accelerate capture.
Sea level fall does not accelerate lake capture in our particular
synthetic scenario. Water unloading related to sea level fall produces a
barrier rebound that delays the capture process (Fig. 13), in a way
identical to what Govers et al. (2009) propose for the onset of the MSC
in the Mediterranean. Therefore, even if the lake level fall of the
Pannonian Basin were induced by capture, it is difficult to attribute it
to the sea level fall in the Dacic Basin during the MSC.
Additional modelling giving more precise estimates on the balance
between uplift and incision, and taking into account lateral differences
in erodibility is required before the capture model can be discarded.
According to the antecedent river model, the fluvial connection
between the basins was maintained despite the Sarmatian uplift
because the incision of the Paleo-Danube could keep up with the uplift
in the belt. The Pannonian lake level was controlled by the elevation of
the outlet, and kept at a more or less constant level depending on the
balance between (isostatic) uplift rate of the barrier and river incision.
By calculating the water balance in the Pannonian drainage area
(Fig. 15), the endorheic lake size that would be attained if the Danube
were blocked at the South Carpathian orocline can be predicted. The
calculations show that for the present day range of precipitation and
evaporation values, the predicted lake size exceeds even the maximum
extent of Lake Pannon during the Sarmatian (Fig. 1, Magyar et al.,
1999a). The Early Late Miocene precipitation in Western and Central
Paratethys is estimated at 1200 mm/y (Böhme et al., 2006) and would
lead to an even larger endorheic lake size (Fig. 15). These numbers
suggest that after the Sarmatian uplift closed the open water connections between the Pannonian and Dacic Basins, a fluvial connection
should have evolved in order to maintain the lake size.
The low salinity of Lake Pannon during its endemic Pannonian
stage is an additional argument for a persistent fluvial connection
between the two basins (Kázmér, 1990) even if the outflow occurred
along subsurface karstic channels (Menkovic and Koscal, 1997) and
the surface expression of the Danube Gorges is younger. A pre-Pontian
fluvial connection may explain the occurrence of freshwater Congeria
of Meotian age derived from the Pannonian Basin in the western part
of the Dacic Basin (Olteanu, 1979). The water balance calculations
192
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 13. Model 7: downstream base level change. a. Sea level curve in downstream basin for extreme scenarios (cf Table 2) b. Sedimentation and erosion rates in downstream basin
compared with results from Model 3 (same setup; no sea level fall). Increase in sedimentation and erosion rates reflects increase in exposed area due to base level fall. Arrows mark
capture time for the three presented models.
(Fig. 15) therefore prompt further investigation of the relationship
between endorheism and endemism, but suggest that endorheism is
not a prerequisite for the Pannonian endemism.
A permanent open-water connection would be in conflict with
the well established endemism in the Pannonian basin during the
Pannonian/Meotian stage (Fig. 2) since it would equilibrate the entire
fauna between the two basins. Considering the decreasing salinity also
in the Dacic Basin since the Sarmatian (Popov et al., 2006), even a
fluvial connection between the basins might have led to fauna
equilibration. This does not occur until the Pontian and would be an
argument against the antecedent river model.
If the Pannonian lake level lowering resulted from the deeper
incision of an existing river into its gateway due to the large MSC-related
base level fall in the Dacic Basin (Clauzon et al., 2005; Leever et al., 2009),
the two events of base level fall can be directly correlated. This would
constrain the time of the Pannonian lake level fall to the Middle Pontian
(of the Eastern Paratethys), i.e. at ∼5.5 My (Vasiliev et al., 2004; Vasiliev
et al., 2005).
4.2.2. Dacic Basin
The models predict that the onset of a fluvial connection between
Dacic and Pannonian lakes should be marked by locally increased
sedimentation rates in front of the Danube gateway in the Dacic Basin,
even if the reconnection occurred before the upstream Pannonian Lake
completed its infill and attained its fluvial stage. A much more
pronounced increase in sedimentation rates would be associated with
the overfilling of the Pannonian Basin. The overfilling and bulk sediment
shift can be directly related to capture only in the special case of a small
remaining upstream accommodation space in combination with a large
elevation difference (see Section 4.1, Fig. 14).
No strong indications of enhanced sedimentation in front of the
Danube gorges with respect to other locations are found in published
seismic data in the western Dacic Basin prior to the Pontian (Rabagia
and Matenco, 1999; Tarapoanca et al., 2007; Leever et al., 2009). This
fits with the model predictions (Model 5, 6). The weak Pannonian
lithosphere would allow ongoing generation of accommodation space
in the upstream basin even after the end of the lacustrine stage, while
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
193
Fig. 14. Factors controlling sedimentary response to lake capture.
no pronounced increase in sedimentation rates is to be expected in
the Dacic Basin, even if a connection between the basins exists. Also
the limited elevation difference between the basins would inhibit a
strong signature of a potential reconnection (Fig. 14).
Within the limitations of the 2D seismic data, enhanced sedimentation in front of the Danube gorges may be inferred starting with the
Middle Pontian sea-level drop, coeval with the Messinian event (Leever
et al., 2009). The enhanced sedimentation in the front of the Danube
indicates that the Pannonian basin has reached an overfilled stage and
that a weak, accommodating Pannonian lithosphere (Fig. 11c) no longer
explains the behaviour of the system. A likely mechanism to explain the
change in its “strength”–or its reduced sink capacity–would be the
inversion of the Pannonian basin starting in the latest Miocene (∼6 Ma,
Vrabec and Fodor, 2004; Leever et al., 2009). Modelling results from
Jarosinski et al. (this volume) predict uplift in the entire Pannonian basin
and show concentration of strain and differential vertical movements at
the limit between the basin and the Dinarides, in close proximity with
the Danube basin outlet.
To determine the relative importance of sea level fall (in the Dacic
basin) and basin inversion (leading to changes in lithosphere
response to loading in the Pannonian basin and differential vertical
motions around the South Carpathian orocline), additional data are
required, such as estimates of sediment volumes for both basins and
more detailed exhumation analysis of the gateway area in the South
Carpathian orocline. Further modelling studies of river incision and
gateway formation should take into account the effect of intraplate
stresses (Cloetingh et al., 1985) and lateral changes in erodibility.
5. Conclusions
Numerical modelling showed that changes in connectivity between
two adjacent sedimentary basins are expressed in the sedimentary
record through its effect on lake level (accommodation space) in the
upstream basin and sediment supply in the downstream basin. The key
factors controlling the sedimentary response are the upstream
accommodation space and the elevation of the upstream basin at the
time of capture. We found that lithosphere rigidity, through its control
on accommodation space, is of major influence on sediment partitioning
and depositional geometries. In the configuration of our model,
downstream sea level changes will only affect the upstream basin if a
gateway already exists.
Lake level changes in the upstream basin, affecting depositional
geometries by changing accommodation space, are most pronounced for
a large pre-capture elevation difference between the basins. Depending
on both the elevation difference and the remaining accommodation
space, the capture-induced lake level fall may lead to locally strongly
increased sedimentation rates and final infilling of the upstream basin.
In addition to the elevation difference between the basins and
assuming that the climatic conditions are steady and well constrained,
the rigidity of the lithosphere is the most important parameter
influencing the post-capture sedimentary evolution. It controls both
the timing and the rate of the bulk sediment shift from the upstream to
the downstream basin. A strong upstream lithosphere will lead to a fast
and sudden shift while for a weak lithosphere the response is more
gradual. The capture event itself will not lead to a strong downstream
change in sedimentation rates, unless it triggers the sediment overfill of
the upstream basin. This is more likely to occur for a large elevation
difference between the basins.
Sea level fall does not accelerate lake capture in our particular
synthetic scenario. Water unloading related to sea level fall produces a
barrier rebound that delays the capture process (see also Govers et al.,
2009).
As for the Pannonian and Dacic Basins, the modelling results suggest
that, after the Sarmatian uplift of the South Carpathians, the fluvial
connection between the basins was never lost. Calculations of a water
balance for the Pannonian drainage area and the low salinity of Lake
Pannon during the endemic Pannonian stage (Kázmér, 1990) also argue
in favor of the antecedent river model. This model may or may not be in
disagreement with the overall endemism of Pannonian and Meotian
faunas. Alternatively, occasional incursions of Pannonian faunas in the
194
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Fig. 15. Equilibrium endorheic surface for Lake Pannon. a. Water balance within a drainage area. b. Equilibrium surface (Aw) of a hypothetical endorheic Lake Pannon for different
precipitation (P) and evaporation rates (at land, El), as a function of evaporation at the lake (Ew). Curves (a) and (b) are based on the average present day outflow Qout of the Danube
at the Iron Gates of 5450 m3/s (http://www.grdc.sr.unh.edu/). If zero evaporation on land is assumed, this yields an average annual precipitation of 287 mm/y over the Pannonian
drainage area (a). The actual present-day annual precipitation is 600 mm/y in Hungary (Vituki, 2002), corresponding to an El of 313 mm/y (b). Assuming the average Miocene
precipitation values (1200 mm/y, Böhme et al., 2006) leads to a much larger estimate of endorheic lake size (c, d). Heavy horizontal dashed line represents maximum Middle
Miocene surface area of Lake Pannon (Magyar et al., 1999a); the surface area in the present-day Pannonian drainage area below 100 and 200 m (derived from SRTM DEM) is shown
for reference. These equations show that, within the range of present day evaporation rates at Lake Balaton, an endorheic lake is unlikely for any of the considered P/El ratios and
suggest that a permanent outflow must have existed.
Dacic basin may be explained by overspill events rather than by a
permanent connection. The Messinian sea level fall in the Dacic Basin
resulted in deeper fluvial incision into the Danube gateway, which
caused the lowering of the Pannonian lake level; the connection
between the basins was permanent from then on. The lowering of the
Pannonian lake level marked the end of the 2nd order “Late Miocene
sequence” (sensu Juhász et al., 1999), enhanced the final infilling of the
basin and increased the sedimentation rates into the Dacic Basin.
Combined with constraints from basin-scale seismic sequence
stratigraphy, the modelling technique used here is a promising tool
for reconstructing paleogeographic evolution, aiding the interpretation
of the existing biostratigraphic database and outlining the need for
additional (field) constraints on the vertical motions in the barrier and
its flanking basins.
Acknowledgements
This research was part of KAL's PhD thesis sponsored by ISES. Two
anonymous reviewers provided useful comments on an earlier
version of the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tecto.2010.01.003.
References
Bada, G., Horvath, F., Gerner, P., Fejes, I., 1999. Review of the present-day geodynamics of
the Pannonian Basin; progress and problems. Journal of Geodynamics 27, 501–527.
K.A. Leever et al. / Tectonophysics 502 (2011) 175–195
Böhme, M., Ilg, A., Ossig, A., Küchenhoff, H., 2006. New method to estimate
paleoprecipitation using fossil amphibians and reptiles and the middle and late
Miocene precipitation gradients in Europe. Geology 425–428.
Bojar, A.-V., Neubauer, F., Fritz, H., 1998. Cretaceous to Cenozoic thermal evolution of
the southwestern South Carpathians: evidence from fission-track thermochronology. Tectonophysics 297, 229–249.
Clauzon, G., 1978. The Messinian Var canyon (Provence, southern France)—paleogeographic implications. Marine Geology 27 (3–4), 231–246.
Clauzon, G., et al., 2005. Influence of Mediterranean sea-level changes on the Dacic
Basin (Eastern Paratethys) during the late Neogene: the Mediterranean Lago Mare
facies deciphered. Basin Research 17 (3), 437–462.
Cloetingh, S., et al., 2006. Models of Basin (de)formation, lithospheric strength and
vertical motions in the Pannonian–Carpathian system: inferences from thermomechanical modelling. In: Gee, D.G., Stephenson, R. (Eds.), European Lithosphere
Dynamics, Memoirs. Geological Society, London, pp. 207–221.
Cloetingh, S., McQueen, H., Lambeck, K., 1985. On a tectonic mechanism for regional
sealevel variations. Earth and Planetary Science Letters 75, 157–166.
Csató, I., 1993. Neogene sequences in the Pannonian basin, Hungary. Tectonophysics
226 (1–4), 377–400.
Csató, I., Kendall, C., Moore, P.D., 2007. The Messinian problem in the Pannonian basin,
eastern Hungary—insights from stratigraphic simulations. Sedimentary Geology
201 (1–2), 111–140.
Cvijic, J., 1908. Entwicklungsgeschichte des Eisernen Tores. Peterm, Mitteilungen
Ergänzungsheft. 160.
Fuegenschuh, B., Schmid, S., 2005. Age and significance of core complex formation in a
very curved orogen: evidence from fission track studies in the South Carpathians
(Romania). Tectonophysics 404 (1–2), 33–53.
Gagic, N., 1997. Neogene brachyhaline and freshwater microfossil association in
Zagubica Basin. In: Grubic, A., Berza, T. (Eds.), International Symposium on Geology
in the Danube Gorges. Geoinstitut, Donji Milanovac/Orsova, pp. 61–64.
Garcia-Castellanos, D., 2002. Interplay between lithospheric flexure and river transport
in foreland basins. Basin Research 14, 89–104.
Garcia-Castellanos, D., 2006. Long-term evolution of tectonic lakes: climatic controls on
the development of internally drained basins. GSA Special Paper 398, 283–294.
Garcia-Castellanos, D., Verges, J., Gaspar-Escribano, J., Cloetingh, S.A.P.L., 2003. Interplay
between tectonics, climate, and fluvial transport during the Cenozoic evolution of
the Ebro Basin (NE Iberia). Journal of Geophysical Research 108 (B7), 2347.
Gillet, H., Lericolais, G., Rehault, J.P., 2007. Messinian event in the Black Sea: evidence of
a Messinian erosional surface. Marine Geology 244 (1–4), 142–165.
Govers, R., Meijer, P., Krijgsman, W., 2009. Regional isostatic response to Messinian
Salinity Crisis events. Tectonophysics 463 (1–4), 109–129.
Haq, B.U., Hardenbol, J., Vail, P.R., 1987. Chronology of fluctuating sea levels since the
Triassic. Science 235 (4793), 1156–1167.
Horváth, F., et al., 2006. Formation and deformation of the Pannonian Basin: constraints
from observational data. In: Gee, D.G., Stephenson, R. (Eds.), European Lithosphere
Dynamics, Memoirs. Geological Society, London, pp. 191–206.
Horváth, F., Cloetingh, S.A.P.L., 1996. Stress-induced late-stage subsidence anomalies in
the Pannonian Basin. Tectonophysics 266, 287–300.
Juhász, E., et al., 1999. Late Neogene sedimentary facies and sequences in the Pannonian
Basin, Hungary. In: Durand, B., Jolivet, L., Horváth, F., Séranne, M. (Eds.), The
Mediterranean Basins: Tertiary Extension Within the Alpine Orogen. Special
Publications. Geological Society, London, pp. 335–356.
Juhász, G., Pogácsás, G., Magyar, I., Vakarcs, G., 2007. Tectonic versus climatic control on
the evolution of fluvio-deltaic systems in a lake basin, Eastern Pannonian Basin.
Sedimentary Geology 202 (1–2), 72–95.
Kázmér, M., 1990. Birth, life and death of the Pannonian Lake. Palaeogeography,
Palaeoclimatology, Palaeoecology 79, 171–188.
Kooi, H., Beaumont, C., 1994. Escarpment evolution on high-elevation rifted margins:
insights derived from a surface processes model that combines diffusion, advection,
and reaction. Journal of Geophysical Research 99 (B6), 12191–12209.
Krijgsman, W., Hilgen, F.J., Raffi, I., Sierro, F.J., Wilson, D.S., 1999. Chronology, causes and
progression of the Messinian salinity crisis. Nature 400 (6745), 652–655.
Lankreijer, A., Mocanu, V., Cloetingh, S.A.P.L., 1997. Lateral variations in lithosphere
strength in the Romanian Carpathians: constraints on basin evolution. Tectonophysics 272 (2–4), 269–290.
Leever, K.A., Matenco, L., Rabagia, T., Cloetingh, S., Krijgsman, W., Stoica, M., 2009.
Messinian sea level fall in the Dacic Basin (Eastern Paratethys): palaeogeographical
implications from seismic sequence stratigraphy. Terra Nova 22 (1), 12–17.
Loget, N., Van Den Driessche, J., 2006. On the origin of the Strait of Gibraltar.
Sedimentary Geology—The Messinian Salinity Crisis Revisited 188–189, 341–356.
Magyar, I., Geary, D.H., Müller, P., 1999a. Paleogeographic evolution of the Late Miocene
Lake Pannon in Central Europe. Palaeogeography, Palaeoclimatology, Palaeoecology 147, 151–167.
Magyar, I., Geary, D.H., Suto-Szentai, M., Lantos, M., Muller, P., 1999b. Integrated
biostratigraphic, magnetostratigraphic and chronostratigraphic correlations of the
Late Miocene Lake Pannon deposits. Acta Geologica Hungarica 42, 5–31.
Marinescu, F., Ghenea, C., Papaianopol, I., 1981. Stratigraphy of the Neogene and the
Pleistocene boundary. Guide to Excursion A6, Carpatho-Balkan Geological Association
XII Congress. Guidebook Series. Institute of Geology and Geophysics, Bucharest.
Marovic, M., Grubic, A., Djokovic, I., Toljic, M., Vojvodic, V., 1997. The genesis of Djerdap
gorge. In: Grubic, A., Berza, T. (Eds.), International Symposium on Geology in the
Danube Gorges. Geoinstitut, Donji Milanovac/Orsova, pp. 99–104.
Matenco, L., Bertotti, G., Dinu, C., Cloetingh, S.A.P.L., 1997. Tertiary tectonic evolution of
the external South Carpathians and the adjacent Moesian platform (Romania).
Tectonics 16 (6), 896–911.
195
Matenco, L., Schmid, S., 1999. Exhumation of the Danubian nappes system (south
Carpathians) during the early Tertiary: inferences from kinematic and paleostress
analysis at the Getic/Danubian nappes contact. Tectonophysics 314 (4), 401–422.
Menkovic, L., Koscal, M., 1997. Geomorphological features and morphogenesis of
Djerdap gorge. In: Grubic, A., Berza, T. (Eds.), International Symposium on Geology
in the Danube Gorges. Geoinstitut, Donji Milanovac/Orsova, pp. 89–98.
Olteanu, R., 1979. Signification biostratigraphique des ostracodes Méotiens et Pontiens
du Bassin Dacique. Rev. Roum. Géol., Geophys. et Geogr., Géologie, 23(1): 77–87
Olteanu, R., Jipa, D., 2006. Dacian Basin environmental evolution during Upper Neogene
within the Paratethys domain. Geo-Eco-Marina 12, 91–105.
Peters, K.F., 1876. Die Donau und ihr Gebiet: Eine geologische Skizze. F.A. Brockhaus,
Leipzig. 375 pp.
Pogácsás, G., et al., 1990. Chronostratigraphic relations of Neogene formations of the
Great Hungarian Plain, based on interpretation of seismic and paleomagnetic data.
International Geology Review 32 (5), 449–467.
Pogácsás, G., Szabó, A., Szalay, J., 1992. Chronostratigraphic relations of the progradational
delta sequence of the Great Hungarian Plain. Acta Geologica Hungarica 35, 311–327.
Popov, S.V., et al., 2006. Late Miocene to Pliocene palaeogeography of the Paratethys
and its relation to the Mediterranean. Palaeogeography, Palaeoclimatology,
Palaeoecology 238, 91–106.
Rabagia, T. and Fülöp, A., 1994. Syntectonic sedimentation history in the Southern
Carpathians foredeep. In: T. Berza (Editor), ALPACA: II. Geological evolution of the
Alps–Carpathian–Pannonian System. Rom. J. Tecton. Reg. Geol., pp. 48.
Rabagia, T., Matenco, L., 1999. Tertiary tectonic and sedimentological evolution of the
South Carpathians foredeep: tectonic versus eustatic control. Marine and Petroleum
Geology 16 (7), 719–740.
Rögl, F., 1996. Stratigraphic correlation of the Paratethys Oligocene and Miocene.
Mitteilungen Gesellschaft Geologie- und Bergbaustudien Österreich 41, 65–73.
Rögl, F., 1999. Mediterranean and Paratethys. Facts and hypotheses of an Oligocene to
Miocene paleogeography. Geologica Carpathica 50, 339–349.
Sacchi, M., Horváth, F., 2002. Towards a new time scale for the Upper Miocene continental
series of the Pannonian Basin (Central Paratethys). In: Cloetingh, S.A.P.L., Horváth, F.,
Bada, G., Lankreijer, A. (Eds.), Neotectonics and Surface Processes: the Pannonian
Basin and Alpine/Carpathian System. EGU Stephan Mueller Special Publication Series.
Copernicus GMBH, Katlenburg-Lindau, pp. 79–204.
Sacchi, M., Horváth, F., Magyari, O., 1999. Role of unconformity-bounded units in
stratigraphy of continental record: a case study from the Late Miocene of western
Pannonian Basin, Hungary. In: Durand, B., Jolivet, L., Horváth, F., Séranne, M. (Eds.),
The Mediterranean Basins: Tertiary Extension Within the Alpine Orogen.
Geological Society Special Publication, London, pp. 357–390.
Sandulescu, M., 1984. Geotectonica României (translated title: GeoTectonics of
Romania). Ed. Tehnica, Bucharest, 450 pp.
Sandulescu, M., 1988. Cenozoic tectonic history of the Carpathians. In: Royden, L.H.,
Horváth, F. (Eds.), AAPG Memoir, pp. 17–25.
Saulea, E., Popescu, I., Sandulescu, J., 1969. Lithofacies Atlas. Institutul de Geologie si
Geofizica, Bucharest.
Snow, R.S., Slingerland, R., 1990. Stream profile adjustment to crustal warping:
nonlinear results from a simple model. Jounal of Geology 98, 699–708.
Steininger, F.F., Bernor, R.L., Fahlbusch, V., Mein, P., 1990. European Neogene marine/
continental chronologic relations. In: Lindsay, E.H., Fahlbusch, V., Mein, P. (Eds.),
European Neogene Mammal Chronology. Plenum Press, New York, pp. 15–46.
Tarapoanca, M., Tambrea, D., Avram, V., Popescu, B., 2007. The geometry of the South
Carpathian sole thrust and the Moesia boundary: the role of inherited structures in
establishing a transcurrent contact on the concave side of the Carpathians. In:
Lacombe, O., Lavé, J., Roure, F., Verges, J. (Eds.), Thrust Belts and Foreland Basins:
From Fold Kinematics to Hydrocarbon Systems. Frontiers in Earth Sciences. Springer,
pp. 369–384.
Tari, G., Horváth, F., Rumpler, J., 1992. Styles of extension in the Pannonian Basin.
Tectonophysics 208 (1–3), 203–219.
Toula, F., 1896. Uber den Durchbruch der Donau durch das Banategebirge. Vort. Ver. zur
Verbreitung der natur., Kennt. Wien.
Ujszászi, K., Vakarcs, G., 1993. Sequence stratigraphic analysis in the south Transdanubian
region, Hungary. Geophysical Transactions 38 (2–3), 69–87.
Ustaszewski, K., et al., 2008. A map-view restoration of the Alpine–Carpathian–
Dinaridic system for the Early Miocene. Swiss Journal of Geosciences 101, 273–294.
Vakarcs, G., et al., 1994. Third order Middle Miocene–Early Pliocene depositional
sequences in the prograding delta complex of the Pannonian Basin. Tectonophysics
240 (1–4), 81–106.
Vasiliev, I., et al., 2004. Towards an astrochronological framework for the eastern
Paratethys Mio-Pliocene sedimentary sequences of the Focsani Basin (Romania).
Earth and Planetary Science Letters 227, 231–247.
Vasiliev, I., Krijgsman, W., Stoica, M., Langereis, C., 2005. Mio-Pliocene magnetostratigraphy in the southern Carpathian foredeep and Mediterranean–Paratethys
correlations. Terra Nova 17 (4), 376–385.
Vituki, 2002. A Balatoni vízpótlás lehetöségeinek vizsgálata. Vituki Rt, Budapest.
Vrabec, M., Fodor, L., 2004. Late Cenozoic tectonics of Slovenia: structural styles at the
northeastern corner of the Adriatic microplate. In: Pinter, N., Grenerczy, G., Weber, J.,
Stein, S., Medak, D. (Eds.), NATO Advanced Research Workshop on Adria Microplate—
GPS Geodesy, Tectonics and Hazards, Veszprem, HUNGARY, pp. 151–168.
Wortel, M.J.R., Spakman, W., 2000. Subduction and slab detachment in the
Mediterranean–Carpathian region. Science 290, 1910–1917.