Journal of Applied Geophysics 111 (2014) 299–311

Contents lists available at ScienceDirect

Journal of Applied Geophysics

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

Gravity analysis of the Precambrian basement topography associated

with the northern boundary of Ghadames Basin (southern Tunisia)
Mohamed Dhaoui a,b,f, Hakim Gabtni b, Chokri Jallouli c,f, Ali Jleilia d,
Kevin Lee Mickus e, Mohamed Moncef Turki f
a
Office National des Mines, ONM, Tunisia
b
Georessources Laboratory, Centre de Recherches et des Technologies des Eaux, BO 273, 8020, Soliman, CERTE, Tunisia
c
Department of Geology and Geophysics, College of Sciences, King Saud University, P.O. Box. 2455, Riyadh 11451, Saudi Arabia
d
Entreprise Tunisienne des Activités Pétrolières, ETAP, Tunisia
e
Department of Geosciences, Missouri State University, Springfield, MO 65897, United States
f
Geological Department, Faculty of Sciences, University Tunis El Manar, Campus Universitaire, 2092 Manar 2, Tunis, Tunisia

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

Article history: Gravity data were analyzed to determine the structural development of the northern boundary of the Ghadames
Received 30 June 2014 Basin in southern Tunisia. The Ghadames Basin which also occurs in eastern Algeria and northwestern Libya is
Accepted 14 October 2014 one of the most prolific hydrocarbon producers in North Africa with several of the largest oil fields occurring
Available online 28 October 2014
along its northern boundary. The Ghadames Basin was formed during a series of tectonic events ranging from
the Early Paleozoic to the Early Cenozoic. These tectonic events produced a basin in southern Tunisia that has a
Keywords:
Gravity analysis
complex basement configuration which is not completely known. A residual gravity anomaly map constructed
Precambrian basement using polynomial trend surfaces, and vertical and horizontal gravity derivative maps indicate that the northern
Ghadames Basin boundary contains a series of maxima and minima anomalies that trend in two prominent directions:
Tunisia northeast–southwest and east–west. The horizontal and vertical derivative gravity anomaly maps indicate that
the width of the basement structures range between 10 and 20 km in width. Three-dimensional (3D) Euler
deconvolution and 3D forward modeling constrained by well data, one seismic reflection profile and
remote sensing data confirm the width of the basement structures and indicates that the depth of basin varies
between 1.5 and 5 km, with deeper sections in general more numerous in the southern sections of the boundary.
The gravity analysis constrained by the seismic reflection profile and well data implies that the basement topog-
raphy may have been formed during the Pan African and/or late Mesozoic rifting. However, additional seismic
reflection and well data are needed to confirm this conclusion. The discovery of the numerous basement
structures suggests that there may exist additional hydrocarbon traps within the northern boundary of the
Ghadames Basin.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction
The Ghadames Basin is a large sedimentary encompassing over
350,000 km2 located in southern Tunisia, eastern Algeria and northwest-
ern Libya (Fig. 1) (Echikh, 1998; Gauthier et al., 2003). The Ghadames
Basin is one of a series of sedimentary basins formed on the early Paleo-
zoic passive margin (Saharan Platform) that stretches from Morocco to
Egypt (Boote et al., 1998). While the Ghadames Basin has previously
been considered to be an intracratonic sag basin (Echikh, 1998; Klett,
2000),recent detailed studies of the structures within the basin suggest
that the basin has a far more complex formation history involving a
series of compressional and extensional tectonic events (Gauthier
et al., 2003).
The earliest formation of the basin involved the Pan African orogeny
in the Late Precambrian where Cambrian and Ordovican sediments
were deposited within small-scale rift basins formed along a suture
zone between formed during the Pan African orogeny. During the Mid-
dle to Late Paleozoic, a series of collisions (Taconic, Caledonian and
Hercynian orogenies) occurred between Africa and Laurentia that influ-
enced the structural development of the basin (Echikh, 1998; Gauthier
et al., 2003). The most important of these tectonic events were two
phases of the Herycynian orogeny during the late Paleozoic where
large scale thrusting aided in deepening the basin through flexural de-
pression (Bishop, 1975; Echikh, 1998). During the early Mesozoic,
rifting was associated with the opening of the Tethyan Ocean and this
led to the development of a series of en echelon normal faults and tilted
blocks with the subsequent deepening of portions of the basin (Echikh,
1998). During early Aptian time, deformation caused localized uplifts
and strike-slip faulting (Rossi et al., 2002). Renewed subsidence oc-
curred in the late Aptian that resulted in the deposition of carbonates
and evaporates (Boote et al., 1998). The Alpine orogeny which caused
large scale deformation to the north of the Ghadames Basin, only

http://dx.doi.org/10.1016/j.jappgeo.2014.10.014
0926-9851/© 2014 Elsevier B.V. All rights reserved.
300 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Fig. 1. Tectonic elements within Tunisia, eastern Algeria and northwestern Libya (adapted
from Bishop, 1975). Also shown is the location of the Ghadames Basin and the study area.
produced subtle folding of the sediments within the Ghadames Basin
(Rossi et al., 2002). The combination of the above tectonic events in
the Ghadames Basin region resulted in a series of fault-bounded
Fig. 2. Geological map of the study area superimposed over a shaded elevation relief map (adapted from Ben Ferjani et al., 1990). Major cities are shown and these cities are shown in
subsequent maps. The inset shows the location of the study area within southern Tunisia. The remaining maps cover the same study area as the inset. The coordinate system is UTM
zone 32N using the WGS84 datum and is used for all the maps.
structural highs (e.g., Telemzane High (Fig. 1)) surrounding a central
depression that contains over 6000 m of Paleozoic and Mesozoic sedi-
ments (Klett, 2000).
This study concerns the basement structure of the northern sections
of the Ghadames Basin in southern Tunisia and its relationship to petro-
leum exploration. The northern boundary of the Ghadames Basin in
Tunisia consists of the Telemzane High (Fig. 1) which is a basement up-
lift. The Telemzane High is distinguished by gently dipping Paleozoic
strata truncated by the Hercynian unconformity overlain by Mesozoic
strata (Fig. 2) (Ben Ferjani et al., 1990). Previous studies using well
and magnetic data within the northern boundary have shown this re-
gion to consist of a series of normal faults that controlled the subsidence
of the basin (Acheche et al., 2001; Gabtni et al., 2006; Gabtni et al.,
2012).
The goal of this work is to use detailed gravity data constrained by
geological and remote sensing data to determine the basement struc-
ture and to locate Precambrian basement topography irregularities.
These features are then related to petroleum exploration in terms of
petroleum migration paths and the location of petroleum traps. A pre-
liminary gravity analysis showed that all the major oil and gas fields
along the northern boundary of the Ghadames Basin are located on pos-
itive gravity anomalies (Gabtni et al., 2012). A variety of enhancement
techniques including upward continuation, polynomial trend surfaces,
3D Euler deconvolution and 3D forward modeling will be used to ac-
complish the above goals.
2. Geopetroleum setting
The Ghadames Basin is one of the largest petroleum producing ba-
sins in Tunisia, Algeria and Libya where over 3.5 billion bbl of recover-
able oil has been discovered (Acheche et al., 2001; Galeazzi et al.,
2010). Within Tunisia, the largest oil and/or gas fields include the El
Borma, Debbech, Larich, Makhrougha, and Chouich es Saida (Fig. 3)
where the recoverable reserves are located mainly in middle Triassic
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311 301

Fig. 3. Topographic map of the study area showing the location of the major hydrocarbon fields (green) and deep wells. Also shown are the location of profiles (A–B, C–D yellow dots) that
shows a cross section from a 3D gravity model. The orange profile (SP) is the location of a seismic reflection profile shown in Fig. 13. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)

Fig. 4. Geologic cross section along profile A–B (Fig. 3) based on wells OS1 and SN1 along the northern border of Ghadames Basin.
302 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
sandstones but additional reserves are known in Devonian and
Silurian sandstones (Acheche et al., 2001; Echikh, 1998; Troudi
et al., 2003). Additionally, there are Carboniferous producing units
within the fluvial channel sandstones in Libya (Frohlich et al.,
2010) but to date, the petroleum potential of these Carboniferous
units is unknown in Tunisia.
The general stratigraphy of the northern border of Ghadames Basin
is shown in Fig. 4 (Ben Ferjani et al., 1990; Bishop, 1975). The Cambrian
and Ordovician units represent mainly fluvial sandstones that are
interfingered with transgressive marine shales. The Taconian unconfor-
mity separates the Ordovocian units from the early Silurian units, with
the early Silurian being represented by the thick marine black shales
of the Tannezuft Formation. The Devonian units above the Caledonian
unconformity include continental and shallow-marine sandstones and
shales. The Triassic units above the Hercynian unconformity consist
of basal Triassic fluvial–deltaic deposits (TAGI Formation), and older
carbonates and evaporates. The Triassic formations overlie Cambrian–
Ordovician formations on the Telemzan High and Devonian formations
in the central basin area. Jurassic to lower Cretaceous rocks contain
mainly carbonates and fluvial sediments capped by the regional Alpine
unconformity. The upper Cretaceous units include carbonates and
dolomites.
To date, the majority of the oil production has been from Triassic
reservoirs which are dominated by the fluvial sandstones of the TAGI
Formation. The TAGI reservoirs are located in the northern Ghadames
Basin within the El Borma, Makrougha and Larich oil fields (Echikh,
1998). The TAGI Formation is widespread within the Ghadames Basin
with producing oil fields are also located in eastern Algeria where they
are found in a linear trend southwest from the El Borma field in
Tunisia (Rossi et al., 2002). These oil fields in eastern Algeria represent
the most important recently discovered petroleum systems in North
Africa (Rossi et al., 2002). The TAGI Formation outcrops along the north-
ern boundary of the Ghadames Basin but the outcrops are discontinuous
and are generally covered by sand. An optical and radar remote sensing
analysis indicated that the TAGI Formation contains numerous structur-
al features (folds and faults) that might be related to trapping mecha-
nisms in the subsurface and thus these surface structures could be
useful when analyzing geophysical data in locating future exploratory
wells (Pena and Abdelsalam, 2006).
Lower Silurian and upper Devonian shales and mudstones are
thought to be the primary source rocks for the all the producing units
within the Ghadames Basin (Echikh, 1998; Rossi et al., 2002; Van de
Weerd and Ware, 1994). In many parts of the basin, the source rocks
are in direct contact with the Triassic reservoir units (Rossi et al.,
2002) and thus are thought to be the source of the oil fields. Basin
modeling has indicated that in the western sections of the basin most
of petroleum generation occurred within the lower Silurian shales and
the petroleum was formed and migrated during the Hercynian orogeny
(Underdown and Redfern, 2007). Within the central sections of the
basin, the upper Devonian mudstones are the main source rocks and
the petroleum was generated in Cretaceous time (Underdown and
Redfern, 2007).
Basin modeling has been important in determining the timing of
hydrocarbon maturation and expulsion of the Silurian source rocks
(Underdown et al., 2007). Underdown et al. (2007) showed that the
Hercynian and Alpine orogenies had the biggest influence in the gener-
ation and expulsion of the hydrocarbons. The migration of the hydrocar-
bons during these events required petroleum traps which in the
northern Ghadames Basin are frequently structural traps including gen-
tle dipping anticlines with normal and reverse faults, and uplifted base-
ment blocks (Echikh, 1998). The basement involved traps within the
northern Ghadames Basin involves two different lithologies and
uplifted blocks. Precambrian granite has been encountered in the SN1
well (Fig. 4) at a depth of 3966.4 m and Precambrian quartzite has
been encountered in wells Ra-1 and LA-2 (Fig. 4) at depths of 3796 m
and 4024 m depths respectively.
3. Gravity analysis
3.1. Gravity data
The gravity data, obtained from the Entreprise Tunisienne d'Activités
Pétrolières (ETAP) who has collected gravity data within the study area
since the 1970's, are shown in Fig. 5. The data were tied to the Interna-
tional Gravity Standardization Net (IGSN71) using the absolute base
station in Sfax, Tunisia. The gravity data which totaled 1500 stations
have an average station spacing of 2 km and are evenly distributed
allowing for a detailed analysis of the deeper basement structures
along the northern boundary of the Ghadames Basin. All the data were
merged and reduced using the 1967 International Gravity formula
(Morelli, 1996). Free Air and Bouguer gravity corrections were made
using sea level as a datum and 2.67 g/cm3 as a reduction density. The
Bouguer gravity anomaly data were gridded at a 1 km spacing and
contoured to produce a Bouguer gravity anomaly map (Fig. 6).
3.2. Bouguer and residual gravity analysis
The purpose of this study is to analyze the gravity data to determine
the basement structure along the northern boundary of the Ghadames
Basin and relate these structures to potential hydrocarbon systems. At
first glance, the Bouguer gravity anomaly map (Fig. 6) indicates a grad-
ual decrease in gravity values toward the south and southwest from the
Telezman High. The source for this gradual decrease in the gravity field
is probably the thickening of the Mesozoic and Paleozoic sediments as
indicated by a structural contour map on top of the late Ordovician sed-
iments (Echikh, 1998; Gabtni et al., 2006). Additionally, a portion of this
gravity decrease is caused by the thickening of the crust from north to
south in Tunisia (Jallouli and Mickus, 2000). However, this gravity gra-
dient has several short wavelength gravity maxima superimposed on
it implying that, assuming the anomalies may be caused by basement
structures, the basin has a variable thickness and is not a simple sag
basin. Such an anomaly pattern is expected given the complicated tec-
tonic history of basin formation (Echikh, 1998).
The Bouguer gravity field is relatively as evidenced by the nine grav-
ity minima representing possible subbasins or grabens (labeled AN1 to
AN9 in Fig. 6) and 11 maxima representing possible uplifts or horsts
(labeled AP1 to AP11 in Fig. 6). The gravity maxima and minima are
expressed by different wavelength anomalies superimposed on a longer
wavelength anomaly caused by either the crustal thickening and/or
sediment thickening to the south and southwest. Since we are interest-
ed in gravity anomalies caused by structures within the basin, a
regional-residual gravity anomaly separation should be attempted to
isolate the gravity anomaly due to the basin and the structures within
the basin which is traditionally called the residual gravity anomaly
(Jacobsen, 1987). In formal terms, the observed field, f0, is the sum of
the regional field, freg, the residual field, fres, and aggregate noise fnoise,
f 0 ðr Þ ¼ freg ðr Þ þ fresðr Þ þ f noiseðrÞ ð1Þ
were r is the location of gravity field.
There are a number of techniques that can accomplish the regional-
residual anomaly separation including spectral wavelength filtering,
upward continuation and polynomial trend surfaces. While these tech-
niques will produce regional (e.g., longer wavelengths) and residual
(e.g., shorter wavelengths) anomalies, the spectral wavelength methods
usually cannot completely separate out the regional and residual com-
ponents due to spectral overlap (Beltrao et al., 1991; Ulrych, 1968).
The polynomial trend surface and continuation methods are mathemat-
ical methods that may not produce the “true” regional gravity anomaly
but they are more likely to produce regional and residual anomalies that
are supported by geological and geophysical evidence (Beltrao et al.,
1991; Jacobsen, 1987). For our analysis, we used polynomial trend sur-
faces and upward continuation, to approximate a regional gravity
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311 303

Fig. 5. Gravity station locations within the study area showing the range of Bouguer gravity values for each station.

anomaly which was then removed from the Bouguer gravity anomalies the general form of the regional gravity field. Since we are interested
to produce residual gravity anomalies. in removing gravity fields produced by density contrasts deeper than
To use polynomial trend surfaces or upward continuation to deter- the basement/basin interface, usually the only constraints available
mine a regional gravity field, one must have constraints to estimate are deep seismic studies. In the Ghadames Basin region, Jallouli and

Fig. 6. Bouguer gravity anomaly map of the study area. Also shown are solid white lines that trend through the center of gravity maxima, dashed white lines trending through the center of
gravity minima, and well location. AP and AN are labels for the solid and dashed white lines, respectively. Contour interval is 0.5 mGal.
 304
M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Fig. 7. A. Upward continuation of the Bouguer gravity field to 30 km. B, C, and D are first, second and third polynomial trend surfaces of the Bouguer gravity field. These maps represent regional gravity fields.
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Mickus (2000) based on gravity modeling and one north-south seismic
refraction profile shows that the upper and lower crusts thicken to-
ward the south, so a regional gravity anomaly that gradually de-
creases toward the south would be expected. The3D configuration
of the crustal thickness under the Ghadames Basin is unknown due
to the lack of deep seismic studies (Jallouli and Mickus, 2000).
Additionally, since the study area covers a small region, the upper
and lower crustal configurations should be approximated by a low
order polynomial. Thus, to estimate the regional gravity field, first-,
second- and third-order polynomial trend surfaces were calculated
(Fig. 7b, c and d). Additionally, an upward continuation of the Bouguer
gravity field was calculated to a height of 30 km above the observation
surface (Fig. 7a). Jacobsen (1987) showed that such a height may
represent density variations at a depth of 15 km. Continuation
heights corresponding to the crustal thickness depths (~35 km) were
attempted but did not produce anomalies that decreased toward the
south or southwest.
The second-order polynomial trend surface (Fig. 7c) and the 30 km
upward continuation gravity field (Fig. 7a) roughly agree in shape and
amplitude. This agreement suggests that either regional gravity field
may represent a reasonable regional gravity field for the study area. To
determine the residual gravity field, both of the above regional fields
were subtracted from the Bouguer gravity field and Fig. 8 shows the re-
sidual gravity anomaly map obtained using a second-order polynomial
surface. The residual gravity anomaly field is similar to the Bouguer
gravity field (Fig. 6), however a detailed investigation of this map
indicates that the gravity minima and maxima are more pronounced
and the amplitudes are more indicative of the gravity anomalies caused
by density variations within the basin. The residual gravity anomaly
map outlines, like Bouguer gravity map, 9 gravity minima (named
AN1 to AN9 in Fig. 8) and 11 gravity maxima (named AP1 to AP11 in
Fig. 8).
Fig. 8. Residual gravity anomaly map of the study area based on the removal of a second order polynomial trend surface (Fig. 7c). Also shown are solid white lines that trend through the
center of gravity maxima, dashed white lines trending through the center of gravity minima, and well locations. AP and AN are labels for the solid and dashed white lines, respectively. Also
shown are wells within the study area. Contour interval is 0.5 mGal.
305
3.3. Vertical/horizontal gravity derivative analysis
Another common method to analyze potential field data are deriva-
tives where either vertical derivatives, horizontal derivatives or some
combination of both (e.g., analytic signals) is applied to map data to
better locate density contrasts in the subsurface (Verduzco et al.,
2004). Derivative methods are most effective when analyzing shallow
depth sources and geologic bodies with large density contrasts
(Cooper and Cowan, 2004). Both vertical and horizontal derivatives
are useful but can provide different information about the density
sources as horizontal derivatives are better suited to located or enhance
the edge of a body (Blakely and Simpson, 1986; Cordell, 1979), while
vertical derivatives are useful in narrowing the width of the anomalies
so that the source body can be better located (Cooper and Cowan,
2004; Marson and Klingele, 1993). Additionally, vertical derivatives
may enhance shallower sources better than the horizontal derivative
(Marson and Klingele, 1993). However, to obtain depth estimates to a
source body, one has to apply some type of two-dimensional (2D) or
three-dimensional (3D) derivative method such as analytic signal or
Euler deconvolution. Additionally, the vertical gravity derivative has
been shown to provide information on the location of source bodies in
the case of interfering gravity anomalies (Marson and Klingele, 1993)
and this method works best when the density boundaries are vertical.
Grauch and Cordell (1987) showed the limitations of the horizontal deriv-
ative method especially when the source boundary is dipping, where the
horizontal derivative maximum is offset from the boundary which may
also be caused by other factors including station spacing. So, in any case,
the interpretation of the vertical and horizontal gravity derivative maps
is not straightforward and one needs outside constraints (e.g., geology,
seismic data) to obtain geological meaningful interpretations.
Fig. 9 shows the vertical gravity derivative of the residual gravity
anomaly map of the study area. To aid in the interpretation of the
306 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311

Fig. 9. First vertical derivative of the residual gravity anomalies (Fig. 8). Also shown are solid white lines that trend through the center of vertical gravity derivative maxima, dashed white
lines trending through the center of vertical gravity derivative minima, and well locations. AP and AN are labels for the solid and dashed white lines, respectively. Contour interval is 0.001
mGal/km.

Fig. 10. Magnitude of the horizontal derivatives of the residual gravity anomalies (Fig. 8). Solid black lines are drawn through the center of the horizontal derivative maxima. Also shown
are solid white lines that trend through the center of vertical gravity derivative maxima, dashed white lines trending through the center of vertical gravity derivative minima, and well
locations. Contour interval is 0.001 mGal/km.
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
anomalies, solid and dashed white lines were drawn through maxima
and minima anomalies that were roughly linear. The main advantage
of vertical gravity derivative map is to outline the location of density
sources within the regions (AN1 to AN9, AP1 to AP11) that were outline
on Fig. 8.
The magnitude of the horizontal gravity derivative using residual
gravity anomalies (Fig. 8) was calculated using north–south and east–
west directional derivatives. The magnitude of the horizontal gravity
gradient is given by (Blakely and Simpson, 1986):
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
   
∂gðx; yÞ 2 ∂gðx; yÞ 2
HGGMðx; yÞ ¼ þ 2
∂x ∂y
   
∂gðx;yÞ 2 ∂gðx;yÞ 2
where ∂x
þ ∂y
are the north–south and east–west direction
derivatives, respectively. Fig. 10 shows magnitude of the horizontal
gravity derivative of the residual gravity anomaly map of the study
area. The maxima are highlighted with solid black lines and to compare
with the results from the vertical gravity derivative map.
The derivative analysis indicates that the majority of the high ampli-
tude anomalies trend either northeast–southwest or east–west (Figs. 9
and 10). As shown below in the modeling section, we believe that these
anomalies are caused by Precambrian basement topography. The deriv-
ative maps (Figs. 9 and 10) indicate that the widths of the source bodies
are relatively narrow (between 10 and 20 km). The vertical derivative
map indicates that a majority of the basement blocks are not continuous
as a positive derivative trend is actually composed of a series of linear
maxima. This can be seen especially on anomalies AP5, AP7 and AP6
on Fig. 9. Now this feature may be caused by noise in the data, however,
as shown below, 3D modeling of the residual gravity anomalies indi-
cates that this may be a true feature of the basement structure.
3.4. Power spectrum and 3D Euler deconvolution analysis
Spectral methods have been commonly used in the analysis of po-
tential field data to provide first-order depth estimates to density and
magnetic susceptibility interfaces (e.g., Blakely, 1988; Dorman and
Lewis, 1970; Shuey et al., 1977; Tselentis et al., 1988). They have been
most commonly used in the analysis of magnetic data to estimate the
depth to the Curie isothermal point (e.g., Bouligand et al., 2009; Shuey
et al., 1977; Tanaka et al., 1999) to provide estimates of the regional
heat flow regime. The spectral method assumes that the density or mag-
netic susceptibility boundaries are essentially horizontal interfaces. To
estimate spectrally determined depths, a radially averaged 2D power
density spectrum is calculated over a region and in our case we used
the entire region shown in Fig. 6. The depths that are determined de-
pend on the size of the region that is analyzed. The size must be at
Fig. 11. Example of the gravity power spectrum analysis. Dashed lines represent the best fit line tangent to the linear segments.
307
least three to four times greater than the depths one wants to deter-
mined (Hussein et al., 2012). The radially average power density spec-
trum is then plotted on log plot (Fig. 11) and best-fit straight lines are
estimated through “obvious” linear segments on the plot (Tanaka
et al., 1999) (Fig. 11). The slope of these line segments provides
estimates of the depths to the interface with lower wavenumbers
representing deeper interfaces.
Fig. 11 shows an example of the power spectrum analysis. One prob-
lem with using the spectral methods is determining what wavenumbers
to use in fitting a straight line as slight variations in the slope can lead to
different depths (Ravat et al., 2007). To determine the variation in the
depths to density interfaces, we measured the wavenumber slopes at
three different locations for each segment. By performing this analysis,
we estimated that the largest error in the depths was less than 0.2 km
for the deepest interface. Fig. 11 shows that there are three obvious
linear segments with depths at 4, 2 and 1 km. Based on well data, we es-
timate that the 4 km depths represent the deeper basement regions,
2 km represents shallower basement depths and 1 km represents the
Mesozoic/Paleozoic sediment interface.
Euler deconvolution is a commonly applied method to analyze grav-
ity and magnetic for either 2D (Thompson, 1982) or 3D (Reid et al.,
1990) source geometries. Euler deconvolution is based on Euler's homo-
geneity equation and can be used like the power spectrum method, as a
fast approach to estimate the depth and geometry of the density sources
(Reid et al., 1990). We used the method of Reid et al. (1990) on the re-
sidual gravity anomalies to delineate the depths and geometries of the
source bodies. In order to obtain depth solutions, two parameters had
to be considered for calculations: the size of a moving window and
the structural or shape index. The size of the moving window is similar
to methods applying moving averages (Gotze and Krause, 2007) and to
the size of the window which determines the size of the source that
could be analyzed. The structural index varies from a point source, line
source, plane-shaped or a block (3D source). When applying the Euler
deconvolution method, one usually uses a variety of window sizes and
structure indexes to determine which combination of window size
and structural index produces results that agree with the available con-
straints. The method produces a series of points that estimate the loca-
tion of density sources. There are a number of variations on the 3D Euler
deconvolution method such as the use of multiple vertical sources
(Hansen and Suciu, 2002) and derivatives of the gravity field
(e.g., Debeglia and Corpel, 1997). However, we used the most common
method (Reid et al., 1990) where the 3D Euler's equation is defined as
∂gðx; y; zÞ ∂gðx; y; zÞ ∂gðx; y; zÞ
x þy þz þ ŋgðx; y; zÞ ð3Þ
∂x ∂y ∂z
∂gðx; y; zÞ ∂gðx; y; zÞ ∂gðx; y; zÞ
¼ xo þ yo þ zo þ ŋb
∂x ∂y ∂z
308 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
where ∂gð∂x
x;y;zÞ ∂gðx;y;zÞ
, ∂y and ∂gðx;y;z
∂z
Þ
are the derivatives of the gravity field in
the x, y and z directions, (xo, yo, and zo) is the position of the gravity
source, b is the regional gravity value to be estimated and ŋ is the structur-
al index. By considering four or more neighboring gravity values on a grid,
the source location (xo, yo, and zo) and b can be computed by solving a lin-
ear system of equations. Then by moving the window to the next grid lo-
cations, multiple solutions for the same source are can be obtained.
For the northern boundary of the Ghadames Basin, we found that the
best fit to the residual gravity anomaly data occurred using a window
size of 10 grid points. This grid size was determined by trial and error
with smaller window sizes emphasizing shallow, local density sources
and larger window sizes not providing depth information on density
sources within the basin. The structural index was chosen after deter-
mining an optimum window size. All structural indexes were attempted
and a structural index of 0 (3D bodies) was chosen as the results were
similar to the power spectrum results.
Fig. 11 shows the results of the 3D Euler deconvolution analysis. The
depths to the density sources varied between 1 and 5 km. The majority
of the depths were greater than 2 km implying that most of the depth
sources are related to the top of the Precambrian basement. Well data
indicate that the shallowest depth to the basement is approximately
1.5 km. Even though there are deeper sources throughout the basin,
the deeper portions are concentrated in the southern and in the
north–central portions of the study area. Within the central portions
of the study area, there are large regions with shallower basement
depths (2–3 km-green color in Fig. 12). This analysis shows that the
basin has a complex basement with no region that can be considered
the basin's deepest portion as the basement depths occur in basically
north–northeast and east–west trending regions.
4. Gravity modeling and discussion
To quantify the interpretations of the Bouguer and residual gravity
anomalies discussed above, we constructed a 3D gravity model of the
Fig. 12. Locations of the 3D Euler deconvolution solutions within the Ghadames Basin in Tunisia and surrounding regions. The solutions were found using a window size of 10 grid cells, and
a structure index of 0. Circles with plus signs-well locations. The black, white and dashed-white lines are the same as shown in Figs. 7 and 8.
northern boundary of the Ghadames Basin. The model was constructed
using a 3D forward modeling algorithm that calculates the gravitational
attraction due to an ensemble of right, rectangular prisms (Nagy, 1966).
The model was constructed using the residual gravity anomalies
(Fig. 8). Since gravity modeling is nonunique, constraints must be used
to obtain geologically reasonable models. Rock densities of the Mesozoic,
Paleozoic and Precambrian lithologies, and depths to the Paleozoic and
Precambrian lithologies were estimated from well logs. The depth of the
various units was estimated from one seismic reflection profile (Figs. 3
and 13). Additionally, the location of structural lineaments determined
from remote sensing data (Pena and Abdelsalam, 2006) was used to con-
strain the locations of structural elements. The model was broken up into
three layers: 1) Mesozoic sediments (2.60 g/cc), 2) Paleozoic sediments
(2.57–2.72 g/cc) and Precambrian lithologies (2.76 g/cc). The gravity
model was calculated for the region shown in Fig. 6.
Pena and Abdelsalam (2006) analyzed Landsat Enhanced Thematic
Mapper (EM+), Advanced Spaceborne Thermal Emission and Reflection
Radiometer (ASTER), radar (RADARSAT), and Digital Elevation Models
(DEMs) derived from the Shuttle Radar Topography Mission (SRTM) to
determine the spatial location of surface geology units and lineaments
possibly related to near surface structural features. EM+ and ASTER
data were fused to produce band and band-ratio images that were used
to identify lithological units ground truthed from surface geological
studies (Ali et al., 1985; Bishop, 1975). The RADARSAT data were used
to trace these geological units under the regions covered by thin
(b 1 m) sand. These remote sensing-based geological units were used
to constrain our gravity models. The lineament analysis was performed
by analyzing the DEMs using hill-shading techniques (Sabins, 1997) to
determine linear features. Additionally, the RADARSAT was analyzed
for linear features under thin (b1 m) sands. While the lineament anal-
ysis produced a multitude of linear features, those that correlated
with geophysically and/or geologically derived structural features
(Montgomery, 1994) were interpreted to be structural lineaments.
These lineaments were used to help constrain near surface structural
elements in our gravity models.
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Fig. 13. Interpreted seismic reflection profile (SP) whose location is shown on Fig. 3. Well
PGA1 is shown on the profile.
Adapted from Howell and Shoemaker (1981)
The final model was obtained through a trial-and-error process until
the predicted gravity values matched the residual gravity anomalies
using the above constraints. The final model is not unique but is a rea-
sonable model given the available constraints. The model was con-
structed to confirm the derivative, power spectrum and 3D Euler
deconvolution analysis which suggested that the basin was structurally
complex.
Two cross sections are shown and they both indicate that the base-
ment depth is highly variable ranging in depth 1.8 to 4.3 km (Figs. 14
and 15) which essentially agrees with the 3D Euler deconvolution re-
sults (Fig. 12) that indicate that the basin has a complex and highly var-
iable basement structure. The seismic reflection profile, while short in
length, indicates that there are significant variations (at least 0.25 km
based on a seismic velocity of 2.5 km/sec) in the depth to basement
Fig. 14. A north–south cross-section (A–B on Fig. 3) of the 3D gravity model. Wells OS1 and SN1 are shown on the cross section. The cross-section consists of three layers: Mesozoic sed-
iments, Paleozoic sediments and Precambrian lithologies.
309
(Fig. 13). The uplift of the Precambrian basement shown on the seismic
reflection profile is on the northwest edge of large amplitude gravity
maximum (Fig. 8) which implies that the amount of uplift is greater
over this anomaly than is shown on Fig. 8. The gravity model (Fig. 15)
through this region indicates at least 0.5 km of offset. This complex
basement topography is probably caused by deep faults offsetting the
basement and probably the overlying sediments. However, the one seis-
mic reflection profile that was available does not indicate any offset on
the Triassic and younger sediments. At best, there is possible doming of
the overlying Paleozoic and lower Triassic units. To determine if this
scenario is repeated throughout the basin, additional seismic reflection
and well data are needed as gravity data alone cannot determine these
features.
The analysis of the residual and Bouguer gravity field of the northern
boundary of the Ghadames Basin in Tunisia by a variety of interpreta-
tional methods aided in delineating its basement structure. From
these analyses, the depth to the basement is highly variable ranging
from approximately 1.5 to 5 km with the depths being controlled by
faulting probably in the Precambrian units. The deepest sections occur
both in the southern and northern portions of the study area (Fig. 13).
The trend of the basement structures is highly variable as determined
by the derivative analyses (Figs. 9, 10 and 12), however these trends
agree with those found by a remote sensing analysis (Pena and
Abdelsalam, 2006). The major trends follow two prominent directions:
northeast–southwest and east–west. This finding is not surprising given
the complex tectonic basin development history.
The faults controlling the basement structure were formed by one or
more the tectonic events listed in the introduction. While the gravity
data alone cannot distinguish which events caused the structures, we
can speculate on their origin based on the tectonic structures formed
by these events in surrounding regions. The gravity analysis has im-
proved our knowledge about the basement topography and suggests
that the northern flank of Ghadames complex basin is controlled by un-
known deep faults related to Pan African orogeny or Mesozoic rifting
during the opening of the Tethys Sea. This analysis updates the geome-
try of the Ghadames Basin and their hydrocarbon habitat and provides
new ideas regarding all elements responsible for the oil and gas
distribution. The development of the northeast–southwest trending
structures could have been caused by one or more tectonic events
310 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Fig. 15. A north–south cross-section (C–D on Fig. 3) of the 3D gravity model. Wells PGA1 and BZ1 are shown on the cross section. The cross-section consists of three layers: Mesozoic sed-
iments, Paleozoic sediments and Precambrian lithologies.
(Badalini et al., 2002; Coward and Ries, 2003). Throughout north–north-
west Africa, northeast-trending rift structures are found that is related
to Pan African rifting during the Ordovician. Additionally, northeast-
trending thrust faults associated with the Late Paleozoic Hercynian
orogeny are found throughout northwest Africa (Coward and Ries,
2003). These thrust faults emplaced large amounts of crustal material
onto the African crust causing flexure within the Ghadames Basin
region. This flexure produced a series of northeast-trending normal
faults that controlled the thickness of the basin. Lastly, Mesozoic rifting
that caused the opening of the Tethyan Ocean produced a series of
northeast-trending grabens in the Ghadames Basin region (Coward
and Ries, 2003). From the gravity analysis we cannot determine which
one (or more) of these events was responsible for the formation of the
northeast-trending gravity anomalies, especially since we are speculat-
ing that the gravity anomalies are caused mainly by varying basement
topography. The seismic reflection profile indicates that the uplift of
the Precambrian may have affected the overlying Paleozoic and possibly
the Triassic sediments (Fig. 13). If this is true, the major movement of
the Precambrian faults may have occurred in the late Mesozoic. To con-
firm this model, additional well or seismic reflection data is needed to
show if there is a variable thickness of the Mesozoic formations. The
north–south structures may have been formed by north–south trending
strike-slip faulting that occurred during the Late Apian time (Coward
and Ries, 2003).
The northern boundary of the Ghadames Basin contains several oil
fields within the Triassic TAGI Formation (Fig. 3) and these are mainly
found in structural traps (Echikh, 1998). The gravity analysis indicated
there are several faults that control the formation of the basement to-
pography. These structures should be the focus of exploration especially
if they involve the Mesozoic units. Even if they do not, hydrocarbon pro-
duction is known from Ordovician, Silurian and Carboniferous units in
other portions of the Ghadames Basin (Acheche et al., 2001; Echikh,
1998; Frohlich et al., 2010; Troudi et al., 2003) and the hydrocarbon
potential of these units in the northern boundary is unknown. The
gravity analysis indicated several locations to concentrate additional ex-
ploration to test the potential of these Paleozoic units in the Ghadames
Basin.
5. Conclusions
The analysis of gravity data combined with well data and one seis-
mic reflection profile along the northern boundary of the Ghadames
Basin in southern Tunisia indicated that the boundary consists of a com-
plex of basement highs and lows. Residual gravity anomaly, and vertical
and horizontal derivative maps indicated that these highs and lows are
between 10 and 20 km wide and trend in two major directions: north-
east–southwest and east–west. A 3D Euler deconvolution and 3D for-
ward modeling of the residual gravity anomalies indicated that the
thickness of the basin varies between 1.5 and 5 km with thicker regions
in the northern and southern portions of the northern boundary of the
basin. Even though the basin was formed by a series of tectonic events
ranging from the Early Paleozoic to the Early Cenozoic, the gravity anal-
ysis without additional constraints could not determine which of these
events were mainly responsible for the complex basement topography.
However based on the one seismic reflection profile, the basement to-
pography may have formed by either the Pan African orogeny and/or
late Mesozoic rifting. The gravity analysis has important petroleum ex-
ploration implications as several major oil fields in Triassic sediments
occur in the northern boundary mainly within structural traps and
there are several additional structural features (blocks and faults) with-
in the northern boundary that may contain additional hydrocarbon
accumulations.
References
Acheche, M.H., M'rabet, A., Ghariani, H., Ouahchi, A., Montgomery, S.L., 2001. Ghadames
Basin, southern Tunisia: a reappraisal of Triassic reservoirs and future prospectivity.
AAPG Bull. 85, 765–780.
Ali, M.B., Jedoui, Y., Dali, T., Salem, H.B., Memmi, L., 1985. Cartes geologiques au 500000 de
la Tunisie, Office de la Topographie et de la Cartographie, Cite Olympique, Tunis, 1o et
2o ed.
Badalini, G., Redfern, J., Carr, I., 2002. A synthesis of current understanding of the structur-
al evolution of North Africa. J. Petrol. Geol. 25, 249–258.
Beltrao, J., Silva, J., Costa, J., 1991. Robust polynomial fitting for regional gravity estimation.
Geophysics 56, 80–89.
Ben Ferjani, A., Burollet, P., Mejri, F., 1990. Petroleum geology of Tunisia. Entreprise
Tunisienne des Activités Pétrolières, Tunis.
 M. Dhaoui et al. / Journal of Applied Geophysics 111 (2014) 299–311
Bishop, W.F., 1975. Geology of Tunisia and adjacent parts of Algeria and Libya. AAPG Bull.
59, 413–450.
Blakely, R., 1988. Curie temperature isotherm analysis and tectonic implications of aero-
magnetic data from Nevada. J. Geophys. Res. 93, 11817–11832.
Blakely, R.J., Simpson, R.W., 1986. Approximating edges of source bodies from magnetic
or gravity anomalies. Geophysics 51, 1494–1498.
Boote, D., Clark-Lowes, D., Traut, M., 1998. Paleozoic petroleum systems of North Africa.
In: MacGregor, D., Moody, R., Clark-Lowes, D. (Eds.), Petroleum Geology of North
Africa. Geological Society of London Special Publication 132, pp. 7–68.
Bouligand, C., Glen, J., Blakely, R., 2009. Mapping Curie temperature depth in the western
United States with a fractal model for crustal magnetization. J. Geophys. Res. 114.
http://dx.doi.org/10.1029/2009JB006494.
Cooper, G., Cowan, D., 2004. Filtering using variable order vertical derivatives. Compt.
Rendus Geosci. 30, 455–459.
Cordell, L., 1979. Gravimetric expression of graben faulting in Santa Fe country and the
Espanola Basin, New Mexico. New Mexico Geol. Soc. Guidebook, 30th Field Conf,
pp. 59–64.
Coward, M., Ries, A., 2003. Tectonic development of North African basins. In: Arthur, T.,
MacGregor, D., Cameron, N. (Eds.), Petroleum Geology of Africa: New Themes and
Developing Technologies. Geological Society of London Special Publication 207, pp.
61–83.
Debeglia, N., Corpel, J., 1997. Automatic 3-D interpretation of potential field data using an-
alytic signal derivatives. Geophysics 62, 87–96.
Dorman, L., Lewis, B., 1970. Experimental isostasy, 1, Theory of the determination of the
Earth's isostatic response to a concentrated load. J. Geophys. Res. 75, 3357–3365.
Echikh, K., 1998. Geology and hydrocarbon occurrences in the Ghadames Basin, Algeria,
Tunisia, Libya. In: MacGregor, D., Moody, R., Clark-Lowes, D. (Eds.), Petroleum
Geology of North Africa. Geological Society of London Special Publication 132,
pp. 109–129.
Frohlich, S., Redfern, J., Petitpierre, L., Marshall, J., Power, M., Grech, P., 2010. Diagenetic
evolution of incised channel sandstones: implications for reservoir characterization
of the lower Carboniferous Marar Formation, Ghadames Basin, western Libya. J. Petrol.
Geol. 33, 3–18.
Gabtni, H., Jallouli, C., Mickus, K., Zouari, H., Turki, M.M., 2006. The location and nature of
the Telemzan High — Ghadames basin boundary in southern Tunisia based on gravity
and magnetic anomalies. J. Afr. Earth Sci. 44, 303–313.
Gabtni, H., Jallouli, C., Mickus, K., Turki, M.M., Jaffal, M., Keating, P., 2012. Basement struc-
ture of southern Tunisia as determined from the analysis of gravity data: implications
for petroleum exploration. Petrol. Geosci. 18, 143–152.
Galeazzi, S., Point, O., Haddadi, N., Mather, J., Druesne, D., 2010. Regional geology and
petroleum systems of the Illizi-Berkine area of the Algerian Saharan platform: an
overview. Mar. Petrol. Geol. 27, 143–178.
Gauthier, F., Hedley, R., Mckenna, S., 2003. The Structural and Tectonic Evolution of the
Berkine-Ghadames basin. 1st North African/Mediterranean Petroleum & Geoscience
Conference, October 6–9, 2003, Tunis, Tunisia.
Gotze, H., Krause, S., 2007. The central Andean gravity high, a relic of an old subduction
complex? J. S. Am. Earth Sci. 14, 799–811.
Grauch, T., Cordell, L., 1987. Limitations of determining density and magnetic boundaries
from the horizontal gradient or pseudogravity data. Geophysics 52, 118–121.
Hansen, R., Suciu, L., 2002. Multiple-source Euler deconvolution. Geophysics 67, 525–535.
Howell, J.C., Shoemaker, D.W., 1981. Guel Ahmar 1 (PGA1), Final Drilling Report. Pecten
Tunisia Company, Tunis, Tunisia (91 pp.).
311
Hussein, M., Mickus, K., Serpa, L., 2012. Curie point depth estimates from Death Valley and
surrounding regions, southern California. Pure Appl. Geophys. 170, 617–632.
Jacobsen, B.H., 1987. A case for upward continuation as a standard separation filter for
potential-field maps. Geophysics 52, 1138–1148.
Jallouli, C., Mickus, K., 2000. Regional gravity analysis of Tunisia. J. Afr. Earth Sci. 30, 63–78.
Klett, T., 2000. Total petroleum systems of the Trias/Ghadames province, Algeria, Tunisia
and Libya — The Tanezzuft-Oued, Tanezzuft-Melrhir, and Tanezzuft-Ghadames. US
Geol. Surv. Bull. 2202-C.
Marson, I., Klingele, E., 1993. Advantages of using the vertical gradient of gravity for 3-D
interpretation. Geophysics 58, 1588–1595.
Montgomery, S., 1994. Ghadames Basin and Surrounding Areas: Structural in Tectonics,
Geochemistry and Field Summaries. Pet. Front., 78 pp.
Morelli, C., 1996. Modern standards for gravity surveys. Geophysics 41, 1051.
Nagy, D., 1966. The gravitational attraction of a right rectangular prism. Geophysics 31,
362–371.
Pena, S., Abdelsalam, M., 2006. Orbital remote sensing for geological mapping in southern
Tunisia: implication for oil and gas exploration. J. Afr. Earth Sci. 44, 213–219.
Ravat, D., Pignatelli, A., Nicolosi, I., 2007. A study of spectral methods of estimating the
depth to the bottom of magnetic sources from near-surface magnetic anomaly
data. Geophys. J. Int. 169, 421–434.
Reid, A.B., Allsop, J.M., Granser, H., Millett, A.J., Somerton, I.W., 1990. Magnetic interpreta-
tion in three dimensions using Euler deconvolution. Geophysics 55, 80–91.
Rossi, C., Kalin, O., Arribas, J., Tortosa, A., 2002. Diagenesis, provenance and reservoir qual-
ity TAGI sandstones from Ourhoud field, Berkine (Ghadames) Basin, Algeria. Mar.
Petrol. Geol. 19, 117–142.
Sabins, F., 1997. Remote Sensing Principles and Interpretation. W.H. Freeman Company,
(494 pp.).
Shuey, R., Schellinger, D., Tripp, A., Alley, L., 1977. Curie depth determination from aero-
magnetic spectra. Geophys. J. R. Astron. Soc. 50, 75–101.
Tanaka, A., Okubo, Y., Matsubayashi, O., 1999. CPD based on spectrum analysis of the
magnetic anomaly data in East and Southeast Asia. Tectonophys 306, 461–470.
Thompson, D.T., 1982. EULDPH: a new technique for making computer-assisted depth
estimates from magnetic data. Geophysics 47, 31–37.
Troudi, H., Ferjaoui, M., Kebaier, D., Achechei, M., 2003. Paleozoic petroleum systems in
Ghadames Basin, southern Tunisia. 1st North African/Mediterranean Petroleum &
Geoscience Conference, October 6–9, 2003, Tunis, Tunisia.
Tselentis, G., Drakopoulos, J., Dimitriadias, K., 1988. A spectral approach to Moho depths
estimation from gravity measurements in Epirus (NW Greece). J. Phys. Earth 36,
255–266.
Ulrych, T., 1968. Effect of wavelength filtering on the shape of the residual anomaly.
Geophysics 33, 1015–1018.
Underdown, R., Redfern, J., 2007. The importance of constraining regional exhumation in
basin modeling: a hydrocarbon maturation history of the Ghadames Basin, North
Africa. Pet. Geosci. 13, 253–270.
Underdown, R., Redfern, J., Lisker, F., 2007. Constraining the burial history of the Ghadames
Basin, North Africa: an integrated analysis using sonic velocities, vitrinite reflectance
data and fission track ages. Basin Res. 19, 557–578.
Van de weerd, A.A., Ware, P.L.G., 1994. A review of the east Algerian Sahara oil and gas
province (Triassic, Ghadames and Illizi basins). First Break 12, 363–373.
Verduzco, B., Fairhead, J., Green, C., MacKenzie, C., 2004. New insights into magnetic
derivatives for structural mapping. Lead. Edge 23, 116–119.