International Journal of Geosciences, 2012, 3, 357-364 Published Online May 2012 (
Interpretation of Groundwater Flow into
Fractured Aquifer
Sameh W. Al-Muqdadi, Broder J. Merkel
Geology Department, Technische Universität Bergakademie Freiberg, Freiberg, Germany
Received February 11, 2012; revised March 21, 2012; accepted April 16, 2012
The region of investigation is part of the western desert of Iraq covering an area of about 12,400 km2, this region in-
cludes several large wadis discharging to the Euphrates River. Since the Tectonic features in particular fault zones play
a significant role with respect to groundwater flow in hard rock terrains. The present research is focus on investigate
lineaments that have been classified as suspected faults by means of remote sensing techniques and digital terrain
evaluation in combination with interpolating groundwater heads and MLU pumping tests model in a fractured rock aq-
uifer, Lineaments extraction approach is illustrated a fare matching with suspected faults, moreover these lineaments
conducted an elevated permeability zone.
Keywords: Fault Interpretation; Lineaments Extraction; Remote Sensing; Digital Terrain Model; Analytical Pumping
Test Evaluation
1. Introduction
Tectonic features and in particular faults, fault zones, and
fracture zones play a significant role with respect to
groundwater flow in hard rock terrains and show sub-
stantial impact at multiple scales. Faults and fracture
zones may be areas of preferential flow but, may act as
barriers as well. Whether a fault zone or fractures are
water bearing or not is mainly controlled by the tectonic
stress and strain and secondary fracture fillings. The
permeability of fractures is assumed to be greater in the
direction parallel to the principal stress field. Stress-relief
fracturing might be as well a reason for increased per-
Numerous studies have been performed in order to de-
termine fractured zones in aquifers and its impact onto
groundwater flow and heads [1] examined the hydroge-
ology of regional flow system in carbonate terrain by
large-scale pumping for irrigation. Water-level decline in
corresponding wells was used as a proof to identify hy-
draulic connection between several wells. The study
showed that the orientation of water bearing fractures or
conduits inferred is consistent with the major orientation
of local and regional structural features.
A conceptual model of groundwater flow was imple-
mented by Mayo and Koontz, 2000 [2] for fractured
zones associated with faulting in sedimentary rocks. The
model is based on the results of field and laboratory in-
vestigations, groundwater and methane gas inflows from
fault-fracture systems, showing that groundwater stored
in fractured sandstone is confined above and below by
clayey layers.
Electromagnetic surveys by VLF-WADI resistivity
sounding was used by Sharma and Baranwal, 2005 [3] to
interpret the geological structures and groundwater move-
ment through fractures rock. Laboratory experiments
were carried out by Qian et al., 2005 [4] to study ground-
water flow in a single fracture with different surface
roughness and apertures. Results show that the gradient
of the Reynolds number versus the average velocity in a
single fracture was almost independent of the change of
fracture surface roughness, and it decreased when the
aperture decreased under the same surface roughness.
Studying lineaments from remote sensed data is an al-
ternative because high production areas in fractured aqui-
fers are often associated with visible lineaments at the
ground surface. An effective technique for delineation of
fracture zones is based on lineament indices extracted
from air-photos and from satellite imagery. In combina-
tion with structural and tectonic analysis information for
understanding groundwater flow and occurrence in hard-
rock aquifers may be provided. Several techniques has
been developed and applied to extract lineaments from
air-photos, satellite images, and digital terrain models. A
comprehensive toolkit was created by Raghavan et al.,
1995 [5] for extracting lineaments from digital images
using segment tracing and rotation transformation
(START). This algorithm was coded in FORTRAN 77
opyright © 2012 SciRes. IJG
and implemented on a UNIX-based workstation and is
executed in two stages: at first a binary line element im-
age is generated using the segment tracing algorithm;
then in a second step the rotation transformation algo-
rithm is used to extract lineaments from the line element
image. This algorithm may work with different data set
such as LANDSAT MSS, LANDSAT thematic mapper
(TM), DEM’s as well as shaded aeromagnetic images.
Kim et al., 2004 [6] Developed for ArcView SHP files
Avenue scripts to extract lineaments and lineament den-
sity from satellite images. The scripts may facilitate to-
gether with borehole date the analysis of the relationship
between groundwater characteristics and lineament dis-
tribution. Accuracy of extracted lineaments depends str-
ongly on the spatial resolution of images: the higher the
resolution the better will be the quality of lineament map
2. Motivations
The aim of this paper is to investigate lineaments that
have been classified as suspected faults by means of re-
mote sensing techniques and find out the impact of these
lineaments into the groundwater flow.
3. Regional Geological Setting
The western desert of Iraq (south-west of Euphrates river)
covers nearly 32% from the whole Iraq (437.072 km²)
and habits a population of about 1.3 million [8], UNEP
has adopted an index of aridity, defined as: I = P/PET
where PET is the potential evapotranspiration and P is
the average annual precipitation [9] Hype-arid: <0.05,
arid: 0.05 - 0.2, semi-arid: 0.2 - 0.5, dry sub-humid: 0.5 -
0.65. According to this definition the region has an I-
value of 0.01 and is thus classifies as Hype-arid.
The rather flat terrain is sloping gently towards the
Euphrates River. Rainfall is not sufficient to maintain a
continuous plant cover. Ground water is found in several
horizons in different depths. The majority of the ground
water is nonrenewable flowing in confined aquifers. Re-
charge occurs locally by limited flood events only which
happens immediately after rapid and intense rain events.
The region of investigation is part of the western de-
sert of Iraq (41.14˚E - 32.59˚N and 42.78˚E - 31.86˚N)
covering an area of about 12,400 km² including several
large wadis, such as Ubaiydh, Amij, Ghadaf, Tubal, and
Hauran discharging to the Euphrates River [10]. The re-
gion (Figure 1) was chosen for two reasons: on the one
hand it was classified by Consortium-Yugoslavia, 1977
[11] as a promising groundwater abstraction zone and on
the other hand offering a sufficient number of wells for
carrying out a thorough ground water study. However,
recent political events have made investigations in this
area of Iraq rather difficult after 2008.
An elevation contour map was created by using the 90
m SRTM data. With respect to the ROI the highest value
(609 m) is in the west while the lowest one (233 m) is in
the east.
Figure 1. Western desert-Iraq.
Copyright © 2012 SciRes. IJG
The region of interest is considered as a part from sta-
ble shelf zone/Rutba-Jazira sub zone [12]. The Rutba-
Jazira zone is an inverted Paleozoic basin. The inversion
started in the late Permian. Its basement was relatively
stable during Mesozoic-Neogene time and more mobile
during Infracambrian and Paleozoic times. The basement
depth ranges from 5 km in Jazira area up to 11 km south
of Rutba. Jazira area was part of the Rutba uplift domain
in late Permian to early Cretaceous time, following the
Cretaceous Jazira area subsided while Rutba remained
uplifted these two areas are thus differentiated as sepa-
rate sub zones.
Stratigraphic cross section for the Western Desert re-
gion has shown respectively the following formations
Euphrates-L. Miocene, Dammam-M. Eocene, Umm Er-
Rad-huma-M./U. Palaeocene, Tayarat-U. Cretaceous [13]
where Limestone and dolomitic limestone are the most
dominated units for these formations.
Based on (Buday, 1984) three local faults can be dis-
tinguished in the middle and eastern part of the ROI. The
main directions for these faults are NS & NE-SW; these
directions are related to the Najd-Hejaz origin move-
ment which belongs to the Precambrian-Palaeozoic [14].
Fault 1 has been described as a normal fault while
faults 2 and 3 are classified as suspected faults. The Geo-
logical cross sections were generated by four boreholes
[11] (Figure 2): showing the main formation, faults and
moreover a thin layer ~20 m. E recedes to ~6 m. W from
marl located in between the two formations Umm Er-
Radhuma and Tayarat working as an aquiclude while
lenses from marl are imbedded between Dammam and
Umm Er-Radhuma formations. The general tectonic set-
ting indicates that the maximum horizontal stress is ori-
entated SW-NE.
Groundwater flow is suspected to be from W to E to-
wards the Euphrates River with 102 pumping wells scat-
tered in the area tapping the three main aquifers.
4. Methodology
A lineament is usually defined as a straight or curve lin-
ear feature to be seen on the ground surface. Lineaments
can be manmade structures such as roads and canals or
geological structures such as faults/fractures, folds, and
unconformities, differences in vegetation and soil mois-
ture, or drainage networks (rivers). Lineaments can be
mapped during a field survey, or by using air photos and
remote sensing data either manually or by means of pat-
tern recognition algorithms [6]. On contrary lineaments
can be traced from remote sensed data or digital terrain
models by mathematical algorithms.
Sharpening tools are used for automatic merging a low-
resolution color, multi-, or hyper-spectral image with a
Figure 2. Cross sections.
Copyright © 2012 SciRes. IJG
high-resolution grey scale image and resampling to the
high-resolution pixel size. The Gram-Schmidt Spectral
Sharpening algorithm [15] may sharpen multispectral
data using high spatial resolution data and simulating a
panchromatic image for the corresponding sensor ETM.
Gram-Schmidt transformation is performed on the simu-
lated panchromatic band and the spectral bands, using the
simulated panchromatic as the first band by swapping the
high spatial resolution panchromatic band with the first
Gram-Schmidt band.
A directional algorithm implemented by ENVI is a
first derivative edge enhancement filter that selectively
enhances image features having specific direction com-
ponents (gradients). The sum of the directional filter
kernel elements is zero. The result is that areas with uni-
form pixel values are zeroed in the output image, while
those that are variable are presented as bright edges.
The lineament extraction algorithm implemented by
Geomatica is based on three fundamental steps to extract
linear features from an image: at first an edge detection
operator is used to produce a gradient image from the
original image. Then a threshold value is applied on the
gradient image to create a binary edge image and finally
linear features are extracted from the binary edge image.
The last step contains several sub steps such as edge
thinning, curve pruning, recursive curve segmentation,
and proximity curve linking.
If lineaments are faults and deep reaching fracture
zones they may impact the groundwater flow either by
being a zone of preferential flow or acting as barriers.
Such zones can be recognized from groundwater contour
lines for both unconfined and confined aquifers. Several
spatial interpolation algorithms may be used for plotting
groundwater contour lines.
Furthermore pumping test might provide additional
information. A great variety of equations and procedures
are known for pumping tests with observations wells.
However, if as in the case of this study no observation
wells are available the evaluation of such pumping test
data is awkward. Analytical and numerical models may
offer a viable approach. In this study the analytical model
MLU was used for drawdown calculations and inverse
modeling of transient well flow. MLU estimates selected
aquifer parameters based on a best fit analytical solution
to measured time-distance-drawdown data. The software
includes an automatic curve-fitting algorithm computing
optimized aquifer parameters and fitted drawdown re-
sults [16].
5. Result and Discussion
5.1. Lineaments Interpretation
SRTM data with 30 m resolution and 15 m Landsat ETM
have been used to determine lineaments in the region of
interest based on both extraction lineaments algorithm by
Geomatics, 2001 [17] and directional filter algorithm by
Haralick et al., 1987 [18]. Figure 3 shows that the
lineaments F1 and F2 based on field survey [14] match-
ing rather well with the automatically extracted linea-
ments with only a small difference between field data
and remote sensed data. On contrary no evidence was
found by lineament interpretation for the suspected F3
5.2. Groundwater Flow Direction
Static water levels from 102 wells were used to deter-
mine the groundwater flow direction for these aquifers
Figure 3. Extracted lineaments.
Copyright © 2012 SciRes. IJG
by using different spatial interpolation algorithms. (Fig-
ure 4) shows the groundwater flow net for the three aq-
uifers using minimum curvature algorithm [19]. Kriging
[20] and radial basis function [21] did not result in sig-
nificant different contour lines. The main flow direction
is W-E with the highest value in the NW (~487.5 masl./
well 5) while the lowest was in the NE (~189.8 masl./
well 31).
Between the fault lines F1 and F2 the groundwater
gradient is less steep than in the total area and the flow
direction is diverted slightly from the general W-E direc-
tion to ESE. This can be explained by a better permeabi-
lity in the area between F1 and F2. The main stress direc-
tion is ~40˚ and thus almost parallel to F1 and F2; per-
pendicular to this an anticline is assumed [14]. Thus it
can be speculated that the entire area between F1 and F2
is showing a denser fracturing than the rest of the region
of interest. Evidence for any impact on the groundwater
hydraulics by the assumed F3 lineament was not found.
Therefore the existence of this lineament F3 has to be
5.3. Pumping Test by Means of MLU Model
Two pumping test were performed close to the lineament
F2 in the unconfined aquifer Dammam using the wells 9
and 17, the location for both wells is shown in the
groundwater flow net (Figure 4). Table 1 shows that
there is a slight difference in depth to the groundwater
between databank and field records. No adjustment of the
pumping rate was possible and observation wells were
Figure 4. Groundwater flow net.
Copyright © 2012 SciRes. IJG
not available. Pumping rate was determined by means of
a stop watch filling a gradient container. Drawdown was
monitored with a depth sounder in the pumping wells.
Pumping test in well no 9 was performed for 420 min-
utes to reach steady state, and then pump was switched
off to monitor recovery reaching the former static level
after 72 min (Table 2). Pumping test in well no 17 was
performed for 360 min to reach steady state and then
pump was switched off to monitor recovery which was
reached after 300 minutes (Table 3). Results of pumping
test and recovery evaluated with the analytical model
MLU for Windows are presented in Figures 5 and 6,
Results shown that well 17 has given a higher Transiti-
vity value 0.1048 m2/min in compare with well 9 where
T = 0.0832 m2/min, this results support the assumption of
unique elevated permeability zone occurred in between
of the F1 and F2 because of the tectonic stress and the
anticline structure.
6. Conclusion
The lineaments F1 and F2 recognized by field survey
could be confirmed by automated lineament extraction
from 30 m SRTM data and 15 m Landsat ETM Interpo-
lated groundwater heads of the uppermost Dammam aq-
uifer display a different flow pattern between F1 and F2
which correlates with the assumption of an elevated per-
meability between these two lineaments due to tectonic
stress and the anticline structure. Furthermore the results
of the pumping tests support the hypothesis that the area
between F1 and F2 is characterized by increased perme-
7. Acknowledgements
The authors wish to acknowledge to University of Frei-
berg/Hydrogeology institute—Germany for providing the
appropriate field instruments. A special thanks goes for
Table 1. Pumping test comparison wells.
Well no. Coordination Elevation Depth to the GW. Data Bank Depth to the GW. Field Difference %
9 42.28 / 31.95 296.7 61.3 59.1 2.2 2
17 42.32 / 32.09 297.9 49.7 46.3 3.4 3.5
Average 2.8 2.75
Table 2. Recovery test data well 9. MLU aquifer test analysis—for unsteady-state flow.
------------------ Drawdown (m) ----------------
Observation well Aquifer Time (day) Calculated Observed Cal-Obs
Observation well 9 1 0.29167 14.581 14.170 0.411
Observation well 9 1 0.29236 4.165 4.530 0.365
Observation well 9 1 0.29306 2.251 3.150 0.899
Observation well 9 1 0.29375 1.767 2.350 0.583
Observation well 9 1 0.29444 1.581 1.700 0.119
Observation well 9 1 0.29514 1.477 1.550 0.073
Observation well 9 1 0.29861 1.217 1.190 0.027
Observation well 9 1 0.30208 1.079 1.080 0.001
Observation well 9 1 0.30556 0.985 1.070 0.085
Observation well 9 1 0.31250 0.859 0.910 0.051
Observation well 9 1 0.32292 0.738 0.800 0.062
Observation well 9 1 0.33958 0.617 0.690 0.073
Observation well 9 1 0.34375 0.595 0.580 0.015
Sum of squares: 1.4896E+00 m2; K = 0.00208 m/min; T = 0.0832 m2/min.
Copyright © 2012 SciRes. IJG
Table 3. Recovery test data well 17. MLU aquifer test analysis—for unsteady-state flow.
------------------ Drawdown (m) ----------------
Observation well Aquifer Time (day) Calculated Observed Cal-Obs
Observation well 17 1 0.25000 12.497 13.370 0.873
Observation well 17 1 0.25069 12.498 11.000 1.498
Observation well 17 1 0.25139 12.499 9.920 2.579
Observation well 17 1 0.25208 12.500 8.840 3.660
Observation well 17 1 0.25278 12.501 7.960 4.541
Observation well 17 1 0.25347 12.502 7.070 5.432
Observation well 17 1 0.26042 12.511 5.450 7.061
Observation well 17 1 0.27083 12.524 3.660 8.864
Observation well 17 1 0.28472 12.541 2.680 9.861
Observation well 17 1 0.30556 1.059 1.790 0.731
Observation well 17 1 0.33681 0.682 0.790 0.108
Observation well 17 1 0.37847 0.498 0.600 0.102
Observation well 17 1 0.43056 0.383 0.490 0.107
Observation well 17 1 0.49306 0.303 0.380 0.077
Observation well 17 1 0.57639 0.238 0.230 0.008
Observation well 17 1 0.70139 0.181 0.100 0.081
Observation well 17 1 0.86806 0.138 0.040 0.098
Observation well 17 1 1.07639 0.107 0.020 0.087
Sum of squares: 2.9944E+02 m2; K = 0.00262 m/min; T = 0.1048 m2/min.
Figure 5. Recovery test well 9.
Figure 6. Recovery test well 17.
logistic support during the fieldwork to the following
colleagues (Khalid Hoshi, Amar Falah, Yasser Kassim
and the guide Abo Duraid).
[1] A. Braathen, “Kinematics of Post-Caledonian Polyphase
Brittle Faulting in the Sunnfjord Region, Western Nor-
way,” Tectonophysics, Vol. 302, No. 1-2, 1999, pp. 99-
121. doi:10.1016/S0040-1951(98)00281-9
[2] A. L. Mayo and W. Koontz, “Fracture Flow and Ground-
water Compartmentalization in the Rollins Sandstone,
Lower Mesaverde Group, Colorado, USA,” Hydrogeol-
ogy Journal, Vol. 8, No. 4, 2000, pp. 430-446.
[3] S. P. Sharma and V. C. Baranwal, “Delineation of Ground-
water-Bearing Fracture Zones in a Hard Rock Area Inte-
grating Very Low Frequency Electromagnetic and Resis-
tivity Data,” Journal of Applied Geophysics, Vol. 57, No.
2, 2005, pp. 155-166. doi:10.1016/j.jappgeo.2004.10.003
[4] J. Qian, H. Zhan, W. Zhao and F. Sun, “Experimental
Study of Turbulent Unconfined Groundwater Flow in a
Single Fracture,” Journal of Hydrology, Vol. 311, No. 1-4,
2005, pp. 134-142. doi:10.1016/j.jhydrol.2005.01.013
[5] V. Raghavan, S. Masumoto, K. Koike and S. Nagano,
“Automatic Lineament Extraction from Digital Images
Using a Segment Tracing and Rotation Transformation
Approach,” Computers & Geosciences, Vol. 21, No. 4,
1995, pp. 555-591. doi:10.1016/0098-3004(94)00097-E
[6] G. B. Kim, J. Y. Lee and K. K. Lee, “Construction of
Lineament Maps Related to Groundwater Occurrence
with ArcView and Avenue™ Scripts,” Computers and
Geosciences, Vol. 30, No. 9-10, 2004, pp. 1117-1126.
Copyright © 2012 SciRes. IJG
[7] L. Q. Hung, O. Batelaan and F. De Smedt, “Lineament
Extraction and Analysis, Comparison of LANDSAT
ETM and ASTER Imagery,” Case Study: Suoimuoi
Tropical Karst Catchment, 2005, p. 59830T-1.
[8] NIC, “Republic of Iraq National Investment Commis-
sion,” 2004.
[9] V. Unep, “World Atlas of Desertification,” 1992.
[10] N. I. Division, “Western Desert of Iraq,” University of
Texas Libraries, Austin, 1944.
[11] Consortium-Yugoslavia, “Hydrogeological Explorations
and Hydrotechnical Work-Western Desert of IRAQ.
Blook7,” Directorate of Western Desert Development
Projects, Anbar Province, 1977.
[12] S. Z. Jassim, J. C. Goff and M. Muzeum v Brn, “The
Geology of Iraq,” Geological Society, 2008.
[13] T. Buday and J. Hak, “The Geological Survey of the
Western Part of the Western Desert. No.1000,” The State
Company of Geological Survey and Mining, IRAQ, 1980.
[14] T. Buday and S. Z. Jassim, “Tectonic Map of IRAQ,”
Geosurv. Iraq., Baghdad, 1984.
[15] C. A. Laben and B. V. Brower, “Process for Enhancing
the Spatial Resolution of Multispectral Imagery Using
Pan-Sharpening,” Google Patents, 2000.
[16] C. J. Hemker, “Transient Well Flow in Layered Aquifer
Systems: The Uniform Well-Face Drawdown Solution,”
Journal of Hydrology, Vol. 225, No. 1-2, 1999, pp. 19-44.
[17] P. C. I. Geomatics, “Geomatica Softwave Manual,” 2001.
[18] R. M. Haralick, S. R. Sternberg and X. Zhuang, “Image
Analysis Using Mathematical Morphology,” IEEE Trans-
actions on Pattern Analysis and Machine Intelligence,
Vol. 9, No. 4, 1987, pp. 532-550.
[19] W. H. F. Smith and P. Wessel, “Gridding with Continu-
ous Curvature Splines in Tension,” Geophysics, Vol. 55,
1990, p. 293. doi:10.1190/1.1442837
[20] N. Cressie, “The Origins of Kriging,” Mathematical Ge-
ology, Vol. 22, No. 3, 1990, pp. 239-252.
[21] R. E. Carlson and T. A. Foley, “Interpolation of Track
Data with Radial Basis Methods,” Computers & Mathe-
matics with Applications, Vol. 24, No. 12, 1992, pp. 27-
34. doi:10.1016/0898-1221(92)90169-I
Copyright © 2012 SciRes. IJG