GEOLOGICA CARPATHICA, APRIL 2008, 59, 2, 147—158
www.geologicacarpathica.sk
Gravity modelling along seismic reflection profiles in the
Krško basin (SE Slovenia)
ANDREJ GOSAR
University of Ljubljana, Faculty of Natural Sciences and Engineering, Aškerčeva 12, SI-1000 Ljubljana, Slovenia
Environmental Agency of the Republic of Slovenia, Dunajska 47, SI-1000 Ljubljana, Slovenia; andrej.gosar@gov.si
(Manuscript received February 2, 2006; accepted in revised form November 5, 2007)
Abstract: Two-dimensional gravity modelling was applied along six seismic reflection profiles measured across the
Krško basin within the framework of seismic hazard assessment studies at the location of Krško nuclear power plant and
for oil exploration. The aim was to provide additional insight into structural interpretation of the pre-Tertiary bedrock of
the basin, because the quality of the seismic profiles acquired with various techniques over an extended time span is fairly
heterogeneous. The Krško basin, located on the rim of the Pannonian basin, is filled with up to 2 km of Neogene to
Quaternary molasse sediments. The geometry of the gravity models was based on interpretation of seismic reflection
profiles from three different surveys. A density of 2.5 g/cm
3
was used for the pre-Tertiary carbonate bedrock and values
from 2.3 g/cm
3
to 2.0 g/cm
3
for the sequence of Neogene to Quaternary deposits, composed of marl, sand, silt and
limestone. Calculated anomalies were compared with observed gravity anomalies extracted from the two detailed surveys
of the area. In general, a good fit was obtained; some exceptions may originate in undetected density variations or in
violation of the 2D approximation near the basin margins. The findings of the gravity modelling are consistent with the
interpretation of seismic reflection data, which considers the Krško basin to be a fairly regular syncline without normal
faults at the northern and southern margin, assumed in previous works. The modelling showed that gravity data are
useful for interpolation of the shape of the basin’s bedrock between seismic profiles. The structural map of the pre-
Tertiary basement was constructed showing two depressions with a maximum depth of 1600 and 2100 m.
Key words: Krško basin, molasse basin, gravity, seismic reflection, gravity modelling.
Introduction
Several geological and geophysical investigations have
been performed in the Krško basin (Fig. 1), which is filled
with up to 2 km of Neogene to Quaternary molasse
sediments, with a wide range of objectives: for oil and gas
prospecting, for exploitation of geothermal energy, for
underground gas storage in aquifers and for assessment of
earthquake hazard at the location of the Krško Nuclear
Power Plant (NPP). According to the prevailing hypothe-
sis, the Krško basin was considered to be a tectonic graben
structure (Pleničar & Premru 1977; Šikić et al. 1979; Pol-
jak & Živčić 1995), although no proof was available for
presumed normal border faults at the northern and south-
ern margins of the basin. Since most of the previous geo-
physical investigations were limited to the flat central part
of the basin, it was not possible to prove this hypothesis
before multi-fold seismic reflection profiling was complet-
ed (Gosar 1998; Persoglia et al. 2000). On the
basis of these data the Krško basin is now con-
sidered to be a folded syncline with no border
faults, at least in the eastern part of the basin
(Gosar 1998; Verbič et al. 2000; Poljak & Gos-
ar 2001).
Two-dimensional gravity modelling was ap-
plied to provide additional insight into a struc-
tural interpretation of seismic reflection pro-
files acquired in the Krško basin, to support the
application of gravity data for the interpolation
of data between seismic profiles in contouring
the structural maps of seismic horizons and to
extend some structural information to the area
outside the grid of seismic profiles. This was
motivated by the fact that the quality of avail-
able seismic reflection profiles recorded with
various techniques over an extended time span
is highly variable and the accuracy of their in-
terpretation is therefore fairly uneven. In old
Fig. 1. Location map of the Krško basin with indicated study area.
148
GOSAR
analogue profiles, it was possible to interpret with limited
accuracy only two main horizons, whereas the more recent
digital profiles allow a much more detailed view of the in-
ternal structure of the basin. The second reason for using
gravity data was because the density of the grid of reflec-
tion seismic profiles in the Krško basin is still relatively
low. Modelled gravity anomalies were compared with ob-
served anomalies extracted from two detailed gravity sur-
veys performed in the area. A structural map of the pre-
Tertiary
basement
was
constructed
showing
two
sub-basins (Raka and Globoko depressions) with a maxi-
mum depth of 1600 and 2100 m.
Geological setting
The Krško basin (Fig. 2) lies within the Sava folds
(Placer 1998) in the transition zone from the Pannonian
basin to the Southern Alps and Dinarides. It is filled with
up to 2100 m of Neogene to Quaternary molasse
sediments (Gosar et al. 2005). In the past, the Krško basin
was considered to be a Quaternary tectonic depression
with prominent faults on its northern and southern rims
(Pleničar & Premru 1977; Poljak & Živčić 2005). Geo-
physical investigations later revealed a structure of Neo-
gene folds with no signs of bounding faults (Gosar 1998;
Accaino et al. 2003).
Structurally, the Krško basin exhibits a heterogeneous
pattern of Dinaric and Alpine features. The pre-Tertiary
basement (Mesozoic carbonates) was mainly deformed
according to a Dinaric pattern consisting of faults and
folds in a NW-SE direction, as well as transverse faults in
a NE-SW direction (Accaino et al. 2003). Neogene sedi-
ments exhibit Alpine structures, folds and faults that
stretch in a general E-W direction. During the Alpine
folding,
some
Paleogene
Dinaric
structures
were
reactivated as Alpine ones (Poljak & Gosar 2001). The
tectonic activity culminated in the Middle Miocene.
Quaternary sediments are characterized by E-W oriented
compressional structures. According to Šikić et al.
(1979), faults in the NW-SE direction were activated at
the beginning of the Holocene, which opened a path for
the Sava river to flow from the Krško basin to the Sava
depression in the southeast.
Fig. 2. Simplified geological map of the Krško basin (after Pleničar & Premru 1977; Šikić et al. 1979) with Bouguer anomaly map (af-
ter Urh 1955) after removal of regional trend, and location of seismic reflection profiles.
149
GRAVITY MODELLING ALONG SEISMIC REFLECTION PROFILES (SLOVENIA)
The area is characterized by increased heat-flow densi-
ty and increased seismic activity compared to neigh-
bouring areas. The strongest earthquake in recorded his-
tory was magnitude 5.7, which occurred in 1917 at
Brežice (Poljak & Živčić 1995). The recent regional
stress field is presumably compressional with a N-S ori-
ented principle stress (Poljak et al. 2000).
Seismic reflection profiling in the Krško basin
Single-fold analogue reflection profiling in 1959
The first seismic reflection investigation, composed of
four profiles (Fig. 2) with single fold analogue recording
was performed in 1959 for oil and gas prospecting
(Kaloper 1984). In comparison with modern digital
profiles, the quality of these sections is rather poor
(Fig. 3b). In general, deeper parts of the basin are better
imaged, whereas noise dominates in the shallow parts.
The profiles (Figs. 5, 7 and 9) were reinterpreted using
new structural and velocity data (Gosar 1996). It was
possible to interpret only two horizons (Fig. 3b): the top
of Badenian limestone (horizon B in Table 1) and the
pre-Tertiary basement (horizon C).
Multi-fold reflection profiling in 1994/95
To improve the structural model for assessment of
earthquake hazard at the location of Krško NPP, the
high-resolution seismic reflection method was applied
(Poljak et al. 1996; Gosar 1998) in a profile (P-3/94 and
P-4/95) of intermediate depth penetration, using 15 m
group spacing and 12-fold coverage.
The most prominent reflections were obtained from
near the top of the Badenian limestone (horizon B),
Fig. 3. Comparison of 12-fold digital P3/95 profile (a) and the single-fold analogue P86/59
profile (b) at their crossing point. Marked seismic horizons are described in Table 1.
Table1: Generalized geological column for the Krško basin with unit thicknesses (Poljak et
al. 1996) and horizon depths in Drn-1/89 borehole (Kranjc et al. 1990; Gosar 1998).
Horizons interpreted in seismic profiles and used in gravity modelling are marked.
while the Mesozoic basement
(horizon C) was less pronounced
(Fig. 3a). In the shallower part, a
clear image of the boundary
between Pontian sandy marl and
Pannonian marl (horizon A) was
also obtained. A folded structure
is clearly visible (Fig. 8), with a
maximum depth to the pre-
Tertiary basement of 1500 m.
Multi-fold reflection profiling
in 1999/2000
Reflection
profiling
in
the
Krško basin continued in an
international project of the EU
PHARE program (Persoglia et al.
2000), comprising three regional
reflection profiles in a total
length of 41 km (Fig. 2). They
were measured with an explosive
source fired in 5—10 m deep
boreholes, using 15 m group
spacing and 18-fold coverage
(Accaino et al. 2003).
Six horizons were identified in
all reflection profiles (Table 1).
They
include
the
Mesozoic
basement of the Krško basin and
a sufficient number of horizons
within the Neogene sequence to
capture
its
internal
structure
(Figs. 6 and 10).
Seismic horizons
In this section, the lithostrati-
graphic description of the inter-
preted seismic horizons (Ta-
ble 1)
is
summarized
after
150
GOSAR
Persoglia et al. (2000), Poljak et al. (2002) and Gosar et
al. (2005).
The deepest horizon mapped is the top of the
Cretaceous flysch or Triassic dolomite (horizon C). The
lowermost Neogene sequence between horizons C and B
represents Ottnangian sediments transgressively deposited
over the Mesozoic basement. These have relatively weak
and diffuse seismic signals, which is probably caused by
their heterogeneous lithological content.
In the Badenian, Lithothamnion limestone (horizon B)
was transgressively deposited over the Ottnangian
sequence. Upwards it transits into sandy and marly
limestone, and sandy marl of Badenian to Sarmatian age.
This sequence has a relatively uniform thickness of
200 m, except in the eastern part of the Krško basin
(Globoko depression), where it shows a slight increase.
The next sequence between horizons M and A is Pan-
nonian marl, which has an almost uniform thickness of
about 100 m. Upwards, this marl transits into sandy marl
of Early Pontian age between horizons A and P2.
The Upper Pontian (above P2
seismic horizon) is repre-
sented by sand with rare lenses of gravel. The main charac-
teristic of this unit is a variable thickness, from 100 m in the
western part to up to 500 m in the Globoko depression in
the east. The uppermost horizon P1 is related to no clear
lithological change within the Upper Pontian.
Gravity data
The Krško basin area has to date been investigated with
the gravity method in two regional and two detailed
surveys, firstly for hydrocarbons exploration and secondly
for underground gas storage in aquifers. The regional
studies were published as gravity maps of Slovenia (Čibej
1967) and Yugoslavia (Federal Geological Survey, 1972).
The gravity study performed for oil and gas explora-
tion in 1955 (Urh 1955) covered an area of 258 km
2
(Fig. 2). Altogether, 751 points were measured, giving an
average density of 3 points/km
2
. Bouguer anomalies
were computed using a density of 2.5 g/cm
3
, derived
from laboratory measurements on samples taken from
outcrops or from shallow boreholes and with the Nettle-
ton method (Dobrin & Savit 1988). The shape of the ba-
sin with two depressions is clearly visible in the Bouguer
anomaly map (Fig. 2). Bouguer anomalies are in the
range between + 11 and + 32 mGal. The minimum values
in both depressions are + 11 mGal.
The gravity investigations performed for a feasibility
study of underground gas storage in aquifers (Starčević et
al. 1989) surveyed only the central part of the Krško basin.
In an area covering 150 km
2
, a total of 1162 points were
measured, with an average density of 13.5 points/km
2
in
the central part and 7.7 points/km
2
in the border area. The
Bouguer anomalies were computed with a density of
2.5 g/cm
3
. Some indications of antiform structures in the
pre-Tertiary rocks were indicated in the structural height
between the two depressions, which could be useful for
the storage of gas in porous rocks (aquifers).
Comparison of the two maps showed good correspon-
dence, but a bulk shift between Bouguer anomalies
amounts to approximately 30 mGal, because the two
maps were tied to a different datum. The regional gravity
trend, which is relatively small in the area, was removed
by polynomial fitting.
Two-dimensional gravity modelling
Two-dimensional gravity modelling was performed
along six transverse seismic reflection profiles acquired
in
the
described
projects.
The
modelled
gravity
anomalies were compared with observed anomalies
extracted from two detailed gravity surveys (Urh 1955;
Starčević et al. 1989) after removal of the regional trend.
The main goals of the gravity modelling were to support
the structural interpretation of seismic profiles and to
facilitate the contouring of the structural maps of seismic
horizons by interpolation of the data between seismic
profiles.
The computer program MODEL2D (Roach 1993) was
used in modelling. This program uses the algorithms
developed by Talwani et al. (1959) for calculation of the
gravity anomaly resulting from bodies with polygonal
cross-sections and infinite extension in a direction
perpendicular to the modelled cross-section. In the Krško
basin, only profiles recorded in a transverse direction (N-S
or NNW-SSE) with respect to the axis of the syncline
(WSW-ENE) fulfilled this condition.
Methodology
Quantitative interpretation of gravity data commonly
involves both inverse and forward methods (modelling).
The forward method involves the direct calculation of
the gravity effect of an assumed density distribution, and
its comparison with the observed anomaly. The model
parameters are then adjusted in order to improve the fit
between the two anomalies (Dobrin & Savit 1988). In
this three-step procedure, constraints imposed by all
available geological, geophysical and other independent
information are used. Models based on the interpretation
of seismic reflection profiles in particular are commonly
used as starting models.
Initially, bodies of simple geometrical shape were used
in gravity modelling for approximation of bodies or
structures with anomalous density. Later, a number of
methods were developed for the forward calculation of
the gravity effect of bodies of arbitrary shape. The most
useful way is to approximate the irregular cross-sectional
shape of a two-dimensional geological structure with
simplified multisided polygons. This method stems from
Hubbert (1948), but Talwani et al. (1959) developed it in
a way suitable for adaptation to computer algorithms.
The gravity attraction of a two-dimensional body,
therefore, depends on the position of the corners of the
polygon. This method is the most widely used technique
in potential field interpretation today (Blakely 1995;
151
GRAVITY MODELLING ALONG SEISMIC REFLECTION PROFILES (SLOVENIA)
Benson & Floyd 2000), implemented in many software
packages (Roach 1993; Martín-Atienza & García-Ab-
deslem 1999; García-Abdeslem 2000).
Geological structures are often longer than they are
wide (two dimensional or linear). The rather subjective
criterion of “sufficiently linear” to apply a two-
dimensional approach means, according to different
authors, that the body is between three and twenty times
longer than it is wide and depends on the shape of the
structure (Blakely 1995).
For cases in which structures do not have infinite
extension, algorithms that calculate the gravity attraction
from bodies of finite-length have been developed
(Rasmusen & Pedersen 1979; Cady 1980; Hunter et al.
1998). Three-dimensional modelling algorithms are today
available (Goetze & Lahmeyer 1988; Fellner 2007) and
applied in favourable conditions, but 2D modelling is still
widely used. The main reason for this is no longer the lack
of availability of suitable algorithms or computer power
but usually the lack of input data, which makes a 3D
solution too ambiguous (Reynolds 1997).
Examples of gravity modelling in sedimentary basins
similar to the Krško basin (thickness of sediments of up
to 2 km) can be found in Birch (1982) for the
Albuquerque basin, Kieniewicz & Luyendyk (1986) for
the Santa Maria basin, Campos-Enriques et al. (1997) for
the Mexico City basin, Nemes et al. (1997) for the
Klagenfurt basin and Gabriel et al. (2003) for basins in
northern Germany.
Geometry of the models
The main horizons as interpreted in the seismic
profiles were used to separate layers of different density
in the gravity model. In the modelling process firstly
only densities were varied. If necessary the geometry of
the layers was later also slightly
adjusted. Constant densities were used
inside each layer, because no surface or
borehole data were available to support
density
variations
within
individual
layers related to different depths of
burial. This approximation resulted in a
less good fit between calculated and
observed gravity values at the model
margins, where the layers outcrop.
In the single-fold analogue profiles
recorded in 1959 it was possible to
interpret
only
two
horizons,
corresponding
to
the
pre-Tertiary
bedrock and to Badenian limestone (C
and B in Table 1). Three layers of
different
density
were
therefore
distinguished
for
gravity
modelling
(Figs. 5, 7 and 9). In the multi-fold
profiles P-3/94 and P-4/95, three main
horizons were interpreted; in addition to
C and B, also horizon A (Table 1), which
corresponds to the boundary between
Pontian sandy marl and Pannonian marl. Four layers of
different density were considered in this model (Fig. 8). In
two profiles recorded in 1999, the data quality allows
identification of six horizons, including the pre-Tertiary
basement and five horizons within the Neogene sequence
(Table 1). In addition to C, B and A, these are two shallow-
er horizons (P2 between the Lower and Upper Pontian and
P1 within the Pliocene) and horizon M between the Sar-
matian and Pannonian. In gravity modelling, seven layers
were therefore used (Figs. 6 and 10), although the density
contrast between adjacent layers was sometimes very
small.
Observed gravity anomalies
Gravity anomalies computed from two-dimensional
models were compared with observed anomalies con-
structed from both gravity survey maps after removal of
the regional trend by polynomial fitting. Since the earlier
survey covers a larger area to the east and north, gravity
profiles taken from the map of Urh (1955) were used for
most (four) of the seismic lines. Gravity data taken from
the map of Starčević et al. (1989) were used for the pro-
files KK-02/99 and P-85/59 profiles, because of a higher
density of measurements.
The observed gravity anomaly profiles along all six
transverse seismic lines considered in the gravity model-
ling are shown in Fig. 4. All the profiles are plotted
against distance, so that the axis of syncline is in the
same position. The regular, almost symmetrical shape of
the Krško syncline is clearly shown in these graphs. The
Bouguer anomalies are in a range from + 12 mGal to
+ 33 mGal. The highest gradients of Bouguer anomalies
are in the northern limb of the syncline, in a range from
5.0 mGal/km in the KK-03/99 profile to 6.9 mGal/km in
the KK-02/99 profile. In the southern limb, they are in a
Fig. 4. Observed gravity anomaly profiles along six transversal seismic lines across the
Krško basin considered in gravity modelling. All the profiles are plotted against the
distance so that the axis of the syncline is at the same position. The profiles are con-
structed from the Bouguer anomaly map shown in Fig. 2.
152
GOSAR
range from 4.8 mGal/km in the KK-02/99 profile to
6.7 mGal/km in the P-84/59 profile.
Rock densities
The rock density range used in gravity modelling is
discussed in this section. Very little data about the rock
density of the lithological units found in the wider Krško
basin area is available and even less data for values at
depth, where the rocks are exposed to higher pressure.
The single source for data in depth was a compensated
density log from the Drn-1/89 borehole (Kranjc et al.
1990). Average densities were estimated from the well
logging diagram, taking into account larger intervals
corresponding to various lithological units (Table 2).
The bulk density values range from 2.1 g/cm
3
in
Pliocene sand, gravel and clay and Upper Miocene marl
to 2.3 g/cm
3
for Middle Miocene sandy marl, marly lime-
stone, silt and sand. The approximately 40 m thick layer
of Badenian limestone at a depth of 650 m has a bulk
density of 2.4 g/cm
3
and the Cretaceous marly limestone
in the bedrock has values between 2.55 and 2.6 g/cm
3
.
These values are in agreement with densities known from
geologically similar areas in Slovenia for the same litho-
logical units.
In gravity modelling, a density of 2.5 g/cm
3
was select-
ed for the pre-Tertiary bedrock, in order to be consistent
with the value used for the calculation of Bouguer anoma-
lies. For the Neogene to Quaternary sequence of sedi-
ments, values from 2.0 g/cm
3
to 2.3 g/cm
3
were taken.
Since the upper and lower boundary of the Badenian lime-
stone, which has a very variable thickness in the Krško ba-
sin, cannot be distinguished in the seismic profiles, it can-
not be inserted as a separate layer of higher density (2.4 g/
cm
3
) but appears together with the adjacent lower density
sandy marl and marly limestone. Some indications of lat-
eral variations in the lithology and thickness of the Neo-
gene layers precluded the use of a consistent density dis-
tribution in all the profiles. The most prominent is the
thickening of the Ottnangian and Lower Badenian layers
between horizons C and B from 300 m in the Globoko de-
pression to up to 1000 m in the Raka depression. On the
other hand, the thickness of Upper Pontian sand, gravel
and clay increases from 100 m in the western part to up to
500 m in the eastern Globoko depression. Because I decid-
ed to adjust the model geometry only slightly in the mod-
elling process, I kept open the possibility of varying the
density distribution within the basin, but respecting the
constraints discussed above. For instance, the density in
the very thick first layer in the old analogue seismic pro-
files was between 2.12 and 2.15 g/cm
3
in the eastern part
(P-83/59 and P-84/59), whereas it was only 2.0 g/cm
3
in
the western part (P-85/59). However, the first layer extends
down to a maximum of 700 m here, whereas it extends
down to 1450 m in the first two profiles.
In the case of three multi-fold reflection profiles (KK-
02/99, KK-03/99 and P3/94-P4/95) the results of seismic
velocity analysis were used to reveal eventual lateral
density variations. In all three profiles very smooth 2D
velocity fields were obtained. Significant lateral density
variations inside the same profile are therefore not likely.
Gravity models
Fairly good agreements between observed and calculat-
ed anomalies were obtained for most of the models. They
were obtained mainly by varying the densities and by
only small adjustments to the initial geometry of the lay-
ers. This also means that some differences still present in
the models may be caused by relatively small errors in ge-
ometry. Several tests performed during modelling showed
that individual vertexes were rarely moved for more than
300 m in horizontal direction from their initial position.
Undetected density variations in a horizontal direction are
therefore a more likely reason for some observed misfits
between calculated and observed anomalies, particularly
close to the border of the basin.
The fit of calculated and observed gravity anomalies
in the P-83/59 profile (Fig. 5) is good in the central part
of the profile and diminishes in the border area, after
outcropping of the first layer.
The best fit between calculated and observed anoma-
lies among all considered profiles was obtained in the
KK-03/99 profile (Fig. 6). It was obtained with a very
low vertical density gradient in the upper part. From the
surface down to a depth of 1500 m, the density increases
only from 2.15 to 2.21 g/cm
3
.
A good fit of calculated and observed anomalies was
obtained in the P-84/59 profile (Fig. 7). The relatively
large thickening of the second layer at a depth of 1500—
2000 m has no impact on the anomaly, due to the small
density contrast between the first and the second layers,
being 2.15 and 2.23 g/cm
3
respectively.
The fit between calculated and observed gravities in the
P-3/94 and P-4/95 profile (Fig. 8) is good in the central
part of the basin and becomes worse in the border area. A
Table 2: Average densities of the lithological units drilled in Drn-1/89 borehole from compensated density log. Lithology and well
logging data after Kranjc et al. (1990).
153
GRAVITY MODELLING ALONG SEISMIC REFLECTION PROFILES (SLOVENIA)
Fig. 5. 2D gravity modelling along
P-83/59 seismic reflection profile.
Densities are in g/cm
3
.
Fig. 6. 2D gravity modelling along
KK-03/99 seismic reflection profile.
Densities are in g/cm
3
.
154
GOSAR
Fig. 7. 2D gravity modelling along
P-84/59 seismic reflection profile.
Densities are in g/cm
3
.
possible reason for the misfit close to the border of the ba-
sin may lie in a possible violation of 2D conditions.
The P-85/59 profile (Fig. 9) crosses the Krško basin at
the location of the saddle separating the western and
eastern depressions. Although the model is relatively
simple, composed of only three layers, a good fit was
obtained between observed and calculated gravities.
In the KK-02/99 profile (Fig. 10), a good fit between cal-
culated and observed anomalies was obtained after slight
adjustment of the geometry of the model, particularly in the
northern part, where a more prominent change in the dip of
the bedrock in relation to the position of the presumed
Orlica fault was observed in the reflection profile. Lower
densities in the upper part of the model with respect to the
KK-03/99 model can be explained by thinner layers.
Structural map of the pre-Tertiary basement
A structural map of the pre-Tertiary basin (Fig. 11) was
prepared based on eleven seismic reflection profiles re-
corded in three surveys performed so far in the Krško ba-
sin. The support of gravity modelling in the interpreta-
tion of the pre-Tertiary basement was important, because
this is not the strongest horizon observed in seismic pro-
files. Moreover in some of older profiles the horizon C
related to the basement is less clear in some parts due to
presence of more prominent horizon B above it. Al-
though some computer contouring algorithms were test-
ed, the final interpolation of the contours of the Mesozo-
ic basement was done by hand, considering the shape of
the gravity contours.
The structural map clearly shows the shape of the ba-
sin, which is elongated in a WSW-ENE direction. In the
cross direction, the syncline is fairly symmetrical. The
average dip of the basement towards the central part is
20°. This dip is similar to the average dip of Neogene
sediments (Poljak et al. 1996), which is an indication of
post-sedimentary folding. The larger Globoko depression
in the eastern part has a very regular elongated shape and
reaches a maximum depth of 2100 m close to the
intersection of KK-01/99 and KK-03/99 profiles. The
Raka depression in the western part is only partly seen in
this map, which is limited to the extent of gravity data.
This depression is smaller than the Globoko depression
155
GRAVITY MODELLING ALONG SEISMIC REFLECTION PROFILES (SLOVENIA)
Fig. 8. 2D gravity modelling
along P-3/94 and P4/95 seismic
reflection profiles. Densities are
in g/cm
3
.
Fig. 9. 2D gravity modelling
along P-85/59 seismic reflection
profile. Densities are in g/cm
3
.
156
GOSAR
Fig. 10. 2D gravity modelling
along KK-02/99 seismic reflection
profile. Densities are in g/cm
3
.
Fig. 11. Structural map of the pre-Tertiary basement in the Krško basin. Depth contours from the mean elavation of the surface (150 m a.s.l.).
157
GRAVITY MODELLING ALONG SEISMIC REFLECTION PROFILES (SLOVENIA)
and reaches a maximum depth of 1600 m. The two de-
pressions are separated by a wide saddle at a depth of
1150 m, interpreted as the top of the Dinaric thrust.
Discussion and conclusions
Two-dimensional
gravity
modelling
along
six
transverse seismic reflection profiles in general supports
the interpretation of seismic reflection data, which
considers the Krško basin to be a fairly regular syncline
without faults at the northern and southern margins, as
concluded in previous works, before multi-fold reflection
profiling was conducted (Gosar 1998; Persoglia et al.
2000). The support of gravity data was useful, because
the quality of the seismic reflection profiles recorded
over an extended time span is variable and has resulted
in a heterogeneous interpretation, and because the
density of the grid of reflection seismic profiles in the
Krško basin is still relatively low.
Although relatively simple models were used (three to
seven layers of constant densities), a good fit between
observed and calculated gravities was obtained in most
profiles. Some exceptions may originate in undetected
density
variations
or
in
violation
of
the
2D
approximation near the basin’s margins. Because no
surface or borehole data is available to support density
variations within individual layers related to the
different depths of burial, we prefer to keep constant
values, taking into account the lower accuracy at the
model margins, but avoiding ambiguity in the transition
zone, where variations in density may apply. A possible
solution to this problem is the application of an algo-
rithm assuming a vertical density gradient inside an indi-
vidual layer. A less appropriate solution is to use several
layers for the simulation of the vertical gradient. The ef-
fect of model geometry adjustment was also tested,
which showed that differences in the geometry of the
bedrock do not exceed 300 m in the horizontal direction.
The best results were obtained using relatively small
density contrasts. Since a value of 2.5 g/cm
3
is taken for
the pre-Tertiary carbonate bedrock (used also for the
calculation of Bouguer anomalies), corresponding values
for the sequence of Neogene to Quaternary deposits
composed of marl, sand, silt and marly limestone are in
the range 2.0—2.3 g/cm
3
. These values are consistent with
densities obtained from the density log in the Drn1/89
borehole, which is the only deep borehole in the central
part of the Krško basin.
Acknowledgments: The Slovenian administration for
nuclear safety is acknowledged for the permission to use
seismic reflection data. The Geological Survey of
Slovenia and Geoinženiring are acknowledged for the
provision of data from their archives. The author is grate-
ful to collaborators of the project Geophysical research
in the surroundings of the Krško NPP for the most recent
seismic reflection data. Special thanks are due to Karl
Millahn, Rinaldo Nicolich, Marijan Poljak and Sergio
Persoglia for many valuable discussions during data in-
terpretation.
References
Accaino F., Gosar A., Millahn K., Nicolich R., Poljak M., Rossi
G. & Zgur F. 2003: Regional and high-resolution seismic
reflection investigations in the Krško Basin (SE Slovenia).
Ann. Univ. Sci. Bp. Rolando Eötvös Nomin., Sect. Geol. 35,
116—117.
Benson A.K. & Floyd A.R. 2000: Application of gravity and
magnetic methods to assess geological hazards and natural
resource potential in the Mosida Hills, Utah County, Utah.
Geophysics 65, 5, 1514—1526.
Birch F.S. 1982: Gravity models of the Albuquerque basin, Rio
Grande rift, New Mexico. Geophysics 47, 1185—1197.
Blakely R.J. 1995: Potential theory in gravity and magnetic appli-
cations. Cambridge University Press, 1—441.
Cady J.W. 1980: Calculation of gravity and magnetic anomalies
of finite-length right polygonal prisms. Geophysics 45,
1507—1212.
Campos-Enríquez J.O., Delgado-Rodríguez O., Chávez-Segura R.,
Gómez-Contreras P., Flores-Márquez E.L. & Birch F.S. 1997:
The subsurface structure of the Chalco sub-basin (Mexico
City) inferred from geophysical data. Geophysics 62, 23—35.
Čibej B. 1967: Regional gravity map of Slovenia. Geološki Zavod
Slovenije 1—63 (internal report, in Slovenian).
Dobrin M.B. & Savit C.H. 1988: Introduction to geophysical
prospecting. McGraw-Hill, 1—867.
Federal Geological Survey 1972: Gravity map of Yugoslavia
1 : 500,000. Zvezni Geološki Zavod.
Fellner J. 2007: Development of a software package for visualiza-
tion of three-dimensional mass distributions and forward
gravity modelling. J. Geophys. and Engineering 4, 1, 31—39.
Gabriel G., Kirsch R., Siemon B. & Wiederhold H. 2003: Geo-
physical investigation of buried Pleistocene subglacial valleys
in Northern Germany. J. Appl. Geophys. 53, 4, 159—180.
García-Abdeslem J. 2000: 2-D inversion of gravity data using
sources laterally bounded by continuous surfaces and depth-
dependent density. Geophysics 65, 1128—1141.
Gosar A. 1996: Seismic reflection method in structural investiga-
tions for assessment of earthquake hazard in the Krško basin.
Doktorska disertacija, Univerza v Ljubljani, 1—288.
Gosar A. 1998: Seismic-reflection surveys of the Krško basin struc-
ture: Implications for earthquake hazard at the Krško nuclear
power plant, SE Slovenia. J. Appl. Geophys. 39, 131—153.
Gosar A., Komac M. & Poljak M. 2005: Structural model of the
pre-Tertiary basement in the Krško basin. Geologija 48, 1,
23—32 (in Slovenian).
Götze H.J. & Lahmeyer B. 1988: Application of three-dimension-
al interactive modeling in gravity and magnetics. Geophysics
53, 1096—1108.
Hubbert M.K. 1948: A line-integral method of computing the
gravimetric effects of two-dimensional masses. Geophysics
13, 215—225.
Hunter H., Green E.F. & Aiken G.L.V. 1998: An integrated 3-D
gravity and seismic study of the Ouachita frontal thrust belt,
Pittsburg and Latimer Counties, Oklahoma. In: Gibson R.I. &
Millegan P.S. (Eds.): Geological applications of gravity and
magnetics: case histories. SEG, 75—78.
Kaloper D. 1984: Krško basin, processing of seismic data. Geo-
loški Zavod Slovenije 1—9 (internal report, in Slovenian).
Kieniewicz P.M. & Luyendyk B.P. 1986: A gravity model of the
basement structure in the Santa Maria Basin, California. Geo-
158
GOSAR
physics 51, 1127—1140.
Kranjc S., Božović M. & Matoz T. 1990: Report on geological in-
vestigations in the Krško basin for underground storage of
gas, Drn-1/89 borehole. Geološki Zavod Slovenije, 1—19 (in-
ternal report, in Slovenian).
Martín-Atienza B. & García-Abdeslem J. 1999: 2-D gravity mod-
eling with analytically defined geometry and quadratic poly-
nomial density functions. Geophysics 64, 1730—1734.
Nemes F., Neubauer F., Cloetingh S. & Genser J. 1997: The
Klagenfurt Basin in the Eastern Alps: an intra-orogenic de-
coupled flexural basin? Tectonophysics 282, 1, 189—203.
Persoglia S., Gosar A., Millahn K., Nicolich R., Nieto D., Poljak
M., Vesnaver A. & Wardell N. 2000: Geophysical research in
the surroundings of the Krško NPP. Final report. European
Commission – PHARE, 1—68 (unpublished report).
Placer L. 1998: Structural meaning of Sava folds. Geologija 41,
191—221.
Pleničar M. & Premru U. 1977: Basic geological map, sheet Novo
mesto 1 : 100,000. Zvezni Geološki Zavod.
Poljak M. & Gosar A. 2001: Structural characteristics of the Krško
basin from geophysical data acquired in years 1994—2000.
Geološki Zbornik 16, 79—82 (in Slovenian).
Poljak M. & Živčić M. 1995: Tectonics and seismicity of the Krš-
ko basin. 1. Hrv. Geol. Kongr., Zbor. Rad. 475—479.
Poljak M., Verbič T., Gosar A., Živčić M. & Ribičič M. 1996:
Neotectonic investigations in the vicinity of NPP Krško. Geo-
loški Zavod Slovenije, 1—70 (internal report, in Slovenian).
Poljak M., Živčić M. & Zupančič P. 2000: The seismotectonic
characteristics of Slovenia. Pure Appl. Geophys. 157, 37—55.
Poljak M., Rižnar I. & Verbič T. 2002: Geological structure of the
Krško basin. 1. Slov. Geol. Kongr. 73—74 (in Slovenian).
Rasmusen R. & Pedersen L.B. 1979: End corrections in potential
field modelling. Geophys. Prospecting 27, 749—760.
Reynolds J.M. 1997: An introduction to applied and environmen-
tal geophysics. John Wiley & Sons, 1—796.
Roach M. 1993: Model 2D, Ver. 3.0, User Manual. Univ. Tasma-
nia, 1—32.
Starčević M., Stopar R. & Rihtar B. 1989: Report on gravimetric
investigations in the Krško basin in year 1989. Geološki
Zavod Slovenije, 1—8 (internal report, in Slovenian).
Šikić K., Basch O. & Šimunić A. 1979: Basic geological map,
sheet Zagreb 1 : 100,000. Zvezni Geološki Zavod.
Talwani M., Worzel J.L. & Landisman M. 1959: Rapid gravity
computations for two-dimensional bodies with application to
the Mendocino submarine fracture zone. J. Geophys. Res.
64, 1, 49—59.
Urh I. 1955: Report on gravimetric investigations in the Krško ba-
sin in years 1953—1954. Geološki Zavod Slovenije, 1—7 (in-
ternal report, in Slovenian).
Verbič T., Rižnar I., Poljak M., Toman M. & Demšar M. 2000:
Quaternary sediments of the Krško basin. 2. Hrv. Geol. Kon-
gr., Zbor. Rad. 451—457.