Journal of Applied Geophysics 88 (2013) 105–113

Contents lists available at SciVerse ScienceDirect

Journal of Applied Geophysics

journal homepage: www.elsevier.com/locate/jappgeo

Structural interpretation of the Erzurum Basin, eastern Turkey, using curvature

gravity gradient tensor and gravity inversion of basement relief
B. Oruç a,⁎, İ. Sertçelik a, Ö. Kafadar b, H.H. Selim c
a
Kocaeli University, Engineering Faculty, Department of Geophysical Engineering, Umuttepe Campus, 41380 İzmit/Kocaeli, Turkey
b
Kocaeli University, Kosekoy Vocational School, Department of Computer Programming, 41135 Kartepe/Kocaeli, Turkey
c
Istanbul Commerce University, Faculty of Engineering and Design, Department of Jewellery Engineering, 34840 Küçükyalı, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The Erzurum Basin has received more attention in petroleum potential research because of its particularity in
Received 28 May 2012 geographic and tectonic position. There remains debate on the basement structure of the basin since igneous
Accepted 7 October 2012 rocks and faults make the structure and stratigraphy more complicated. We utilize gravity data to understand
Available online 5 November 2012
the structure of the Erzurum Basin. This study describes an edge enhancement technique based on the eigen-
values and determinant obtained from the curvature gravity gradient tensor (CGGT). The main goal of this
Keywords:
Gravity anomalies
technique is to delineate structural boundaries in complex geology and tectonic environment using CGGT.
Curvature gravity gradient tensor The results obtained from theoretical data, with and without Gaussian random noise, have been analyzed
Gravity inversion in determining the locations of the edges of the vertical-sided prism models. The zero contours of the
Structural boundaries smallest eigenvalue delineate the spatial location of the edges of the anomalous sources. In addition, 3-D
gravity inversion of Bouguer anomalies has been used with purpose to estimate the structure of the substrata
to allow modeling of the basement undulation in the Erzurum basin. For this reason, the Parker–Oldenburg
algorithm helped to investigate this undulation and to evidence the main linear features. This algorithm re-
veals presence of basement depths between 3.45 and 9.06 km in the region bounded by NE–SW and E–W
trending lineaments. We have also compared the smallest eigenvalue zero contours with the HGM images
and Tilt derivative (TDR) of Bouguer anomaly map of the study area. All techniques have agreed closely in
detecting the horizontal locations of geological features in the subsurface with good precision.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction over a complex model representing a horizontally and vertically

faulted basement surface. Roberts (2001) showed the more negative
An important objective in the interpretation of gravity data is to eigenvalue is equivalent to the most negative curvature obtained
enhance the data in order to bring out important features. Gravity from conventional potential field anomalies and is useful in estimat-
prospecting delineates lateral change in density contrasting and pro- ing source depth. Roberts (2001) has introduced curvature attributes
vides information not only on lithological changes but also on struc- to seismic systematically, and defined many different curvature attri-
tural trends. However, gradients themselves may be used directly to butes. Murphy (2004) has used the horizontal gravity gradient tensor
infer properties of subsurface structures. Horizontal-gradient analysis to locate geologic contacts. Hansen and deRidder (2006) has de-
of gravity data delineate the vertical and lateral locations of the edges scribed a linear feature analysis using eigenvalues of Hessian matrix
of the geological features. Cordell (1979) and Cordell and Grauch defined by horizontal gravity gradients observed or computed from
(1985) have used the maximum amplitudes of the horizontal gradi- magnetic data. They have shown that the two eigenvalues of this ma-
ents to locate near-vertical geologic boundaries from gravity or trix which are just the most positive and most negative curvatures
pseudogravity anomalies. They have shown the HGM maxima indi- and the more positive eigenvalue can be used for linear feature anal-
cate abrupt lateral density contrasts. ysis and the more negative eigenvalue is useful in estimating source
In recent years, gradiometers allow the measurement of all the depth. Murphy (2007) has used invariants obtained from horizontal
tensor components. Many techniques for mapping have been devel- gravity gradient components to image subsurface geology. Murphy
oped to delineate structural features from potential field gradient and Brewster (2007) have described a procedure for working with
tensor data. Foss (2001) has computed gravity tensor invariants gravity gradient tensor. Phillips et al. (2007) have defined the qua-
dratic surface, a curvature matrix constructed from the coefficients
of the second order terms of conventional potential field anomalies.
⁎ Corresponding author. Fax: +90 262 3352812. Beiki and Pedersen (2010) showed that the strike direction of quasi
E-mail address: [email protected] (B. Oruç). 2D bodies can be estimated from the eigenvector corresponding to

0926-9851/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jappgeo.2012.10.006
106 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113

the smallest eigenvalue of full tensor gravity gradient matrix. Mataragio The solutions of this characteristic equation are the eigenvalues of Γ:
and Kieley (2009) have shown the rotational and horizontal invariants vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi1
0 u !2
of the gravity gradient tensor are imaged sub-vertical plugs, dikes, or  2
1 @∂g x ∂g y ut ∂gx ∂g y ∂g x A
diatremes associated with alkaline intrusions. Murphy and Dickinson λ1 ¼ þ þ − þ4 ð7Þ
2 ∂x ∂y ∂x ∂y ∂y
(2010) have combined the individual tensor components into invari-
ants and defined geologic contact and body shape information. vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi1
0 u !2
We examine the eigenvalues of CGGT in determining the edges of  2
1 @∂g x ∂g y u ∂g ∂g ∂g
−t A
x y x
anomalous sources. The method has been applied to CGGT computed λ2 ¼ þ − þ4 ð8Þ
2 ∂x ∂y ∂x ∂y ∂y
from the Bouguer gravity data compiled from Erzurum area in the east-
ern region of Turkey. In addition, we have estimated a basement undula-
tion model from 3D gravity inversion of Bouguer anomalies and It is well known that the product of the eigenvalues of a square
compared with the zero contours of smallest eigenvalue map. matrix is equal to the determinant of that matrix. It is

2. Outline of the CGGT theory detðΓ Þ ¼ λ1 λ2 : ð9Þ

Hansen and deRidder (2006) made a different derivation, through a

2×2 Hessian matrix of the horizontal vector gradients for magnetic appli- 3. Theoretical examples
cations. The Hessian matrix is also called curvature which describes how
bent a curve or surface is at a particular point on a geometric curve or sur- In this section, the robustness of the CGGT used for the edge en-
face. The curvature gradient matrix of the gravity field is defined as hancement is tested with gravity field anomaly map caused by four
0 1 vertical-sided prisms at a depth to the top of 0.5 km (labeled 1),
∂gx ∂g x 1.5 km (labeled 2), 2 km (labeled 3), and 1 km (labeled 4) (Fig. 1a).
B ∂x ∂y C
B
Γ ¼ CGGT ¼ @ C; ð1Þ
∂gy ∂g y A
∂x ∂y

where gx and gy are the horizontal gravity vectors. ∂gx/∂x and ∂gx/∂x are
first order derivatives of gravity vector components with respect to x and
y. Eq. (1) is equivalent to the magnetic data defined as a Hessian matrix of
the anomalous magnetic field developed by Hansen and deRidder
(2006). From Eq. (2), because the CGGT tensor is symmetric, off-diagonal
elements relate, as follows:

∂g x =∂y ¼ ∂g y =∂x: ð2Þ

Off-diagonal elements tend to emphasize symmetries in the x and

y direction. Boring (1998) has discussed detailed properties of a real
symmetric tensor. When the coordinate system was rotated, the ele-
ments of the tensor reduce to the diagonal form eigenvectors of a ten-
sor have the property that when the inner product of the original
tensor and an eigenvector are taken the result is a vector that is a sca-
lar multiple of the original eigenvector:

Γx ¼ Λx ð3Þ

The solutions Λ to Eq.(3) are the eigenvalues of Γ. The columns of x

are the eigenvectors of Γ and the eigenvalues are arranged:
 
λ1 0
Λ¼ : ð4Þ
0 λ2

It is clear that the diagonal form of the tensor Γ is a simpler repre-

sentation. The vectors x associated with each eigenvalue are the
eigenvectors of Γ. Eq. (3) is written as

ðΓ−ΛIÞx ¼ 0: ð5Þ

Omitting the null vector (x = 0), Eq. (5) implies the matrix Γ − ΛI
is singular and its determinant zero. From the determinant of Γ − ΛI,
it can be constructed the homogeneous characteristic equation for a
2D tensor Γ, expands to

 
 ∂gx ∂g x 
 
 ∂x −λ ∂y  Fig. 1. (a) Shematic representation of 3D vertical-sided prism models used for theoretical
detðΓ−ΛIÞ ¼   ¼ 0:
 ð6Þ examples. (b) Gravity anomaly of vertical-sided prism models at the depths of 0.5 km
 ∂gy ∂g y
−λ 
 (prism 1), 1.5 km (prism 2), 2.0 km (prism 3) and 1 km (prism 4). Density contrast is
∂x ∂y 1 g/cm3.
 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113 107

Fig. 2. The curvature gradient tensor maps over vertical prism models in the upper panel and eigenvalues and determinant maps in the lower panel. Dashed lines are zero contours.
Solid lines represent boundaries in plain view of the prisms.

Fig. 1b shows the theoretical gravity anomaly map due to these prisms. vertical-sided prisms. It is understood that these maps will not show
All prisms are defined with a density contrast of 1 g/cm3. The theoretical the significant increase in the ambiguity even if the noise level is high.
gravity anomalies are calculated using the formula given by Banerjee
and Das Gupta (1977) on a regular grid with a spacing of 0.5 km. In 4. Field example
Fig. 2, the CGGT components (gxx, gxy and gyy) of Bouguer anomaly
map have been calculated using the method based on the FFT by 4.1. General geology and tectonics of the Erzurum area
Mickus and Hinojosa (2001). Mickus and Hinojosa (2001) computed
the FFT of the vertical gravity component and from this determined The surface geology of the study area is largely covered by the
the spectra of all elements of the gravity gradient tensor. Subsequently, Oligocene-Quaternary volcano-sedimentary sequence (Fig. 5). The
the gradients are computed through an inverse FFT. Then the eigen-
values (λ1, λ2) and determinant have been calculated using Eqs. (7),
(8) and (9). It is clear that the zero contours of the smallest eigenvalue
λ2 delineate the spatial location of the edges of the models, responding
well to the edge locations, whereas the zero contours of eigenvalue λ1
are placed outside the source region (Fig. 2). In addition, the zero con-
tours of determinant map have been traced at near edge, as well as ob-
served outside the source region. Note that “inside” zero contours of
determinant map have the same results than those of λ2 in determining
the horizontal locations of the edges of the prism models.

3.1. Noise analysis

Fig. 3 illustrates the gravity anomalies of vertical prisms in Fig. 1a in

case of the presence of noise. The data were contaminated with additive
pseudorandom Gaussian noise with zero mean and a standard deviation
of 0.3 mgal. The CGGT components were then calculated from the grav-
ity anomalies with noise, and were therefore obtained noisy. We have
illustrated CGGT components, and eigenvalues and determinant maps
from these components in Fig. 4. Fig. 4 contains a series of images that
illustrates the CGGT maps computed from Fig. 3 in case of the presence
of noise. The eigenvalues and determinant maps were obtained from
CGGT maps with noise. It is clear that λ1, λ2 and determinant maps Fig. 3. Gravity anomalies with Gaussian noise, generated from distributed random
are least susceptible to noise, and work well in imaging the edges of numbers with a standard deviation of 0.3 mgal.
108 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113

Fig. 4. The curvature gradient tensor maps (gxx, gxy, gyy), eigenvalues (λ1, λ2) and determinant maps computed from gravity anomalies with noise in Fig. 3. Dashed lines are zero
contours.

crystalline basement in the central part of the study area is a complex reconnaissance gravity surveys in the study area were conducted
high-grade metamorphic rocks. The Quaternary-aged alluvial de- using a station spacing of 0.5 km with an accuracy of 0.01 mgal.
posits cover all units unconformably. The sediment thickness, which Fig. 6 shows the Bouguer anomaly map covering a portion of the
is the sum of all six stratigraphic units, is about 6.5 km in the region Erzurum area. The Bouguer anomaly map of the study area is charac-
(Şaroğlu and Güner, 1981). terized by a broad negative gravity low indicating basin structure
The main structural grain of the Erzurum area is NE–SW trending and low density distribution. The gravity anomalies have generally
structures which are mainly widespread in the area as a result of different wavelengths and amplitudes, denoting the different density
phases of strike-slip faulting. As shown in Fig. 5, NE–SW trending contrasts of the Erzurum basin infill. Thus, the Bouguer anomalies are
faults or zones have mapped in Erzurum area and they are associated well correlated with the grabens and structural uplifts filling the sed-
with neotectonic deformation after Plio-Quaternary (Koçyiğit et al., iment units. The large negative Bouguer anomaly (~ 185 mgal) in the
2001). The main fault zones are Çobandede Fault Zone (ÇFZ) and central portion of the map is located where the sediments are the
Dumlu Fault Zone (DFZ) which are lateral strike-slip faults (Koçyiğit thickest.
et al., 2001).
The ÇFZ includes both right and left lateral strike-slip conjugate 4.2.1. CGGT components of Bouguer anomaly data
faults sets. The left lateral strike-slip faults included in the fault The gravity anomaly map was pre-processed to prepare for analy-
zone are dominant (Koçyiğit et al., 1985). The smallest left lateral sis and interpretation. The reason for this pre-processing is that the
strike-slip fault basin, which is the contemporaneous with the initia- major features in the data which may be defined by the edges and
tion of faulting, has been developed on the southern block of the ÇFZ minor features that are not easily seen in the original image are to
in the study area. The basement of this basin has been underlain by detect. In order to calculate CGGT components from the Bouguer
the Lower Cretaceous aged ophiolitic mélange (Koçyiğit et al., anomaly map, the method developed by Mickus and Hinojosa
1985). The old basement in this basin is overlain by a thick continen- (2001) was used. Fig. 7 shows the CGGT components from the
tal sequence alternated with the volcanic. The DFZ cuts and deforms Bouguer anomaly map. The CGGT components allow assembling of
Plio-Quaternary basin fill and volcanic rocks of both the NW part an overview of all buried structures and provide a general knowledge
of the east Anatolian plateau, and the E–SE parts of the east Pontides of the structural frame-work of the study area. Generally,
and has sinistral strike-slip faults formed after Late Miocene, or high-gradient values were observed around the low gravity of the
during the Plio-Quaternary neotectonic period (Koçyiğit et al., 2001). Erzurum-Pasinler-Horasan basin. The pattern of the high-gradient
anomalies in CGGT maps is sharp anomalies which may be produced
4.2. Gravity data by buried boundaries. The NE–SW trending high-gradient zones con-
firm these variations and attribute to the presence of discontinuities
The acquisition of gravity data in the study area was carried out by in the anomaly pattern. As shown in Fig. 8, the zero of smallest eigen-
the Turkish Petroleum Company (TPAO) in order to assist with map- value proves identification of the curvilinear trends within the buried
ping and delineation of anomalous sources with the occurrence of oil basement uplifts and contains a well-defined gravity maximum that
and gas formation in the Erzurum-Pasinler-Horasan basin. Detailed continues these trends.
 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113 109

Fig. 5. Simplified geological map showing major compressional and extensional structures (modified from Koçyiğit et al., 2001) and location map of the study area.
KÇFZ, Kelkit-Çoruh fault zone; TAFZ, Tercan-Aşkale fault zone; KBF, Kavakbaşı fault; HKB, Hasankale basin; HB, Horasan basin; ÇFZ, Çobandede fault zone; DFZ, Dumlu fault zone;
MF, Malazgirt fault; KYF, Karayazı fault; TF, Tutak fault; KGF, Kağızman fault; HF, Horasan fault; LDF, Leninakan-Digor fault; SF, Süphan fault; EF, Erciş fault; HTF, Hasantimur Lake
fault; BGF, Balık Lake fault; DF, Doğubeyazıt fault; IF, Iğdır fault; AF, Aras fault; EFZ, Erevan fault zone; PSFZ, Pambak-Seven fault zone.

4.2.2. Comparison of smallest eigenvalue with TDR and HGM of the

gravity anomalies
In order to compare with the locations obtained from the smallest
eigenvalue zeros of the study area, we have used the TDR and HGM
images. The HGM is a commonly used edge detection filter, the rate
of change of the gravity field in the x and y directions and is given
by Cordell and Grauch (1985):
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
s   2
∂g 2 ∂g
HGM ¼ þ ð8Þ
∂x ∂y

The maxima of HGM highlight high-gradient locations, such as struc-

tural boundaries. However, Cordell and Grauch (1985) discussed the
limitations of the HGM for gravity anomaly data and concluded that,
the maxima of HGM can be offset from a horizontal location directly
over the edges if the edges are not near-vertical and close to each other.
Verduzco et al. (2004) developed the tilt angle filter or tilt deriva-
tive (TDR) for potential fields. For gravity application, this filter is de-
fined as

∂g
!
−1 ∂z
TDR ¼ tan ð9Þ
HGM
Fig. 6. Bouguer gravity anomaly map of the Erzurum-Pasinler-Horasan area.
110 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113

Fig. 7. The CGGT components computed from Bouguer gravity anomaly of study area.

where ∂ g/∂ x, ∂ g/∂ y and ∂ g/∂ z are the first derivatives of the field f in agreement between the observed (Fig. 11a) and calculated (Fig. 11b)
the x, y and z directions. Thus, a map of TDR can be considered an anomalies from basement undulation (Fig. 11c). It is clear that the in-
image of the tangent of the angle. ferred basement configuration shows a general depression of the base-
The first horizontal gradients and vertical gradient of the Bouguer ment in the northern part of the study area. This configuration has also
anomaly have been calculated by using the FFT algorithm developed illustrated thick sediments which are volcano-sedimentary sequence
by Gunn (1975). The first vertical gradient map is illustrated in which consists of andesitic-basaltic volcanic rocks and marine sedimen-
Fig. 9a. As shown in Fig. 9b, mapping local maxima of the HGM of tary clastics. In large part, the gravity anomalies arise principally from
Bouguer anomaly can be used to locate structural boundaries. In variations in basement topography. Thus, one can easily conclude that
Fig. 6, we observe that the pattern of the high-gradient anomalies is gravity anomalies are correlated with basement topography. A major
broad, not like sharp ones of vertical boundaries. Thus, it is concluded uplift of Erzurum-Horasan-Pasinler basin is placed at the northeastern
that the study area represents anomalies caused by a range with of the study area. The bedrock topography shallows southwards and
boundaries, are not vertical and relatively deep . The local maxima northwards, and reaches a 3.45 km. The most significant feature of
on HGM map coincides with zero contours of TDR (Fig. 9c). Fig. 10a the map is the ENE–WSW trending basement depression bounded by
and b shows the comparison of smallest eigenvalue zero contours NW-SW lineament. In general, note that boundaries between uplifts
with HGM maxima and TDR zero contours, respectively. It should be and grabens in the basement undulation have been generated by evolu-
noted that all techniques provide similiar results in determining the tion of fault zones from structural disorder towards geometrical
edges of anomalous sources by producing only minor changes. simplicity. Basement topographic high is associated with the anticlines
and roughly NW-striking basement trough plunges into the series of
4.2.3. Three-dimensional inversion of Bouguer anomaly data basins.
The knowledge of the basement in continental areas is of crucial In order to better interpret the edge positions related to the base-
interest in many applications of geosciences. The structural elements ment undulations, we compare the results obtained from the zero
of the basement can be useful in investigating the petroleum explora- contours of smallest eigenvalue and basement modeling from 3D
tion and regional geodynamical studies. In this study, we used inversion. Fig. 12 shows some detail in the area, supporting the struc-
Parker–Oldenburg inversion method (Gomez and Agarval, 2005; tural complexity. The performance of two techniques is almost the
Oldenburg, 1974; Parker, 1973) to image a sediment-basement inter- same, containing the effects in determining the undulated bedrock
face from gridded Bouguer anomaly data. The method allows us to es-
timate basement depths in regional tectonics. The technique is based
on FFT algorithm. Given the mean depth of the density interface and
the density contrast between the two media, the three-dimensional
geometry of the interface is iteratively calculated. The iterative
process is terminated when a certain number of iterations has been
accomplished or when the difference between two successive
approximations to the topography is lower than a pre-assigned
value as the convergence criteria (Gomez and Agarval, 2005). Once
the topographic relief is computed from the inversion procedure, it
is computed the gravity anomaly produced by topographic relief. In
general, the computed anomaly must be very similar to observed
anomaly. Reaching up to 6.3 km depth of crystalline basement of
Horasan basin studies based on geological insight (Pelin et al., 1980;
Şaroğlu and Güner, 1981) revealed that this region is hopeful in
terms of hydrocarbon.
In the inversion scheme, the modeling has been constrained by
specifying 6 km as the depth to the middle reference surface. An
average density contrast of 0.4 g/cm 3 between crystalline basement
(~ 2.8 g/cm 3) and volcano-sedimentary rocks (~ 2.4 g/cm 3) has been
selected. The smallest and greater cut-off frequency parameters are
chosen as 0.011 and 0.020 km −1, respectively. The inversion process
is iterated until a satisfactory agreement between observed and Fig. 8. The smallest eigenvalue map (λ2) of Bouguer anomaly map in Fig. 6. Dashed
calculated gravity is obtained. Fig. 11 shows a reasonably good lines show the zero contour which corresponds to boundary curves.
 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113 111

Fig. 9. (a) The first vertical gradient, (b) HGM, and (c) TDR images of the Erzurum area. Dashed lines show the zero contour of the TDR.

topography (Fig. 12a). It should be noted that the zero contours only topography. In Fig. 12b, the faults are superimposed on the depths
correspond to the edges along uplifted bedrock topography. Thus, to basement undulation. It is interesting that good correlations have
it is therefore understood that the zero contours of smallest eigen- been obtained between the basement undulation from 3D gravity in-
value have imaged by tracking of the uplifted edges of basement version and known faults in the study area. In general, it is clear that

Fig. 10. Comparison of smallest eigenvalue zero contours with (a) HGM computed from Bouguer anomaly data and (b) TDR zero contours of the study area.
112 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113

structural features may be located in the pathway of oil and gas,

which migrates to maximize its trapping potential. On the regional
scale, a structural high must be in close proximity to a structural
low (Pratsch, 1986). That's why structural highs near structural
lows are the preferred targets. There is currently renewed intense
geological and geophysical interest in the exploration for oil and gas
in the Erzurum-Pasinler-Horasan basin. We suggest that future re-
search explore on the basement relief for a new oil and gas reservoirs.
The occurrence of deep oil and gas formation in the subsurface
(>3.45 km) can be expected to find preserved accumulations.

5. Conclusions

Our main contribution in this study has been to present the effec-
tiveness of the CGGT and 3D inversion results for the interpretation of
gravity data. The results coming out of this study facilitate the identi-
fication of new features as well as the mapping of known main trends
that represents Erzurum-Pasinler-Horasan basin. The CGGT compo-
nents provide information about subsurface features that cannot be
inferred from the only vertical component of the gravity field since cur-
vature of the vertical component has higher frequency content than the
field itself and can potentially provide better contact locations.
Thus, the CGGT data contribute to a detailed model of edges of base-
ment undulation. The smallest eigenvalue obtained from CGGT matrix
has clear indications of utility to infer the location of the boundaries of
lithologies since the analysis of this eigenvalue is only based on the
zero contour. So we can say from this example that zero contour of
λ2 is an elegant solution to extract image features from gravity anom-
alies such as edge enhancement which require knowledge of only the
zero contours of smallest eigenvalue of a CGGT matrix.
The fault pattern in the basement beneath a sedimentary basin
can be unraveled by analyzing the basement undulation obtained
Fig. 11. Observed Bouguer anomaly (a), Bouguer anomaly (b) computed from depths by inverting gravity data. In Erzurum area, significant hydrocarbon
to the basement (bottom) derived from inversion of Bouguer anomalies using finding in stratigraphic sequence has not been established largely be-
Parker–Oldenburg's algorithm (c). Note that considerable similarity exists between cause substantial thickness of sediments exists. Application of the
the observed and calculated gravity.
CGGT data and 3D gravity inversion to the gravity data of the Erzurum
area has showed that the basement highs, lows, and the major faults
structural deformation of basement has well correlated with known in the direction NE–SW and E–W fracture systems of the basement. It
faults. However, ÇFZ, DFZ and HF fault have partly shifted under the is possible to conclude that the potential new oil and gas reservoirs in
indentation pressure of the basement. the study area of the Erzurum basin will be likely associated with
As is well known, the basins became complicated geologic units these features. Prior to the comparatively very expensive exercise of
with a series of basement highs, and lows or associated faults. Such acquisition and analysis of seismic data in the exploration for oil

Fig. 12. Comparison of depths to the basement with zero contours (dashed lines) of smallest eigenvalue λ2 (a) and tectonic lineaments (b). Dashed lines represent lineaments
interpreted from the zero contours.
 B. Oruç et al. / Journal of Applied Geophysics 88 (2013) 105–113 113

and gas in sedimentary basins it is common recommendation to con- Koçyiğit, A., Yılmaz, A., Adamia, A., Kuloshvili, S., 2001. Neotectonics of East Anatolian
Plateau (Turkey) and Lesser Caucasus: implication for transition from thrusting
duct preliminary investigations using CGGT data and 3D inversion. to strike-slip faulting. Geodinamica Acta 14, 177–195.
Mataragio, J., Kieley, J., 2009. Application of full tensor gradient invariants in detection of
intrusion-hosted sulphide mineralization: implications for deposition mechanisms.
Acknowledgements
First Break 27, 95–98.
Mickus, K.L., Hinojosa, J.H., 2001. The complete gravity gradient tensor derived from
The authors are thankful to the Turkish Petroleum Company vertical component of gravity: a Fourier transform technique. Journal of Applied
(TPAO) for supplying gravity data. Thanks are also due to two anony- Geophysics 46, 159–174.
Murphy, C.A., 2004. The Air-FTG airborne gravity gradiometer system. In: Lane, R.J.L.
mous reviewers for their valuable comments and suggestions. (Ed.), Airborne Gravity 2004-Abstracts from the ASEG-PESA Airborne Gravity
2004 Workshop: Geoscience Australian Record, 18, pp. 7–14.
Murphy, C.A., 2007. Interpreting FTG Gravity data using horizontal Tensor components:
References
EGM 2007 International Workshop-Innovation. EM, Grav. and Mag. methods: new
Perspective for Exploration.
Banerjee, B., Das Gupta, S.P., 1977. Gravitational attraction of rectangular parallelepiped.
Murphy, C.A., Brewster, J., 2007. Target delineation using full tensor gravity gradiometry
Geophysics 42, 1053–1055.
data. ASEG, Extented abstracts.
Beiki, M., Pedersen, L.B., 2010. Eigenvector analysis of gravity gradient tensor to locate
Murphy, C.A., Dickinson, J.L., 2010. Geological mapping and targeting using invariant
geologic bodies. Geophysics 75, I37–I49.
tensor analysis on full tensor. EGM 2010 International Workshop, Adding new
Boring, E., 1998. Visualization of Tensor Fields. Master's thesis, University of California,
value to electromagnetic, gravity and magnetic methods for exploration.
Santa Cruz, USA.
Oldenburg, D.W., 1974. The inversion and interpretation of gravity anomalies. Geophysics
Cordell, L., 1979. Gravimetric expression of graben faulting in Santa Fe country and the
39, 526–536.
Espanola Basin, New Mexico. New Mexico Geol. Soc. Guidebook, 30th Field Conf,
Parker, R.L., 1973. The rapid calculation of potential anomalies. Geophysical Journal of
pp. 59–64.
the Royal Astronomical Society 31, 447–455.
Cordell, L., Grauch, V.J.S., 1985. Mapping basement magnetization zones from aeromag-
Pelin, S., Özsayar, T., Gedik, İ., Tokel, S., 1980. Geological ınvestigation of pasinler (Erzurum)
netic data in the San Juan Basin, New Mexico. In: Hinze, W.J. (Ed.), The utility of re-
basin for oil. MTA 729, 25–40.
gional gravity and magnetic anomaly maps: Soc. Explor. Geophys., pp. 181–197.
Pratsch, J., 1986. The distribution of oil and gas reserves in major basin structures: an ex-
Foss, C.A., 2001. Mapping basement relief with airborne gravity gradiometri. ASEG 15th
ample from the Powder River basin, Wyoming, USA. Journal of Petroleum Geology 9,
Geophysical Conference and Exhibition, Brisbane.
393–412.
Gomez, D., Agarval, B.N.P., 2005. 3DINVER.M: a MATLAB program to invert the gravity
Phillips, J.D., Hansen, R.O., Blakely, R.J., 2007. The use of curvature in potential-field
anomaly over a 3D horizontal density interface by Parker–Oldenburg's algorithm.
interpretation. Exploration Geophysics 38, 111–119.
Computers and Geosciences 31, 513–520.
Roberts, A., 2001. Curvature attributes and their application to 3D interpreted horizons.
Gunn, P.J., 1975. Linear transformations of gravity and magnetic fields. Geophysical
First Break 19, 85–99.
Prospecting 23, 300–312.
Şaroğlu, F., Güner, Y., 1981. Doğu Anadolu'nun jeomorfolojik gelişimine etki eden
Hansen, R.O., deRidder, E., 2006. Linear feature analysis for aeromagnetic data. Geophysics
öğeler: Jeomorfoloji, volkanizma ilişkileri. TJK Bülteni 24, 39–50 (in Turkish).
71, L61–L67.
Verduzco, B., Fairhead, J.D., Green, C.M., Mackenzie, C., 2004. New insights into magnet-
Koçyiğit, A., Öztürk, A., İnan, S., Gürsoy, H., 1985. Karasu havzasının (Erzurum)
ic derivatives for structural mapping. The Leading Edge 23, 116–119.
tektonomorfolojisini ve mekanik yorumu. Cumhuriyet Univ., J. Engineer. Fac., Seri
A-Earth Sci. 2, 3–15.