International Journal of Geosciences, 2012, 3, 25-36
http://dx.doi.org/10.4236/ijg.2012.31004 Published Online February 2012 (http://www.SciRP.org/journal/ijg)
Hydromorphological Mapping and Analysis for
Characterizing Darfur Paleolake, NW Sudan Using
Remote Sensing and GIS
Samy Ismail Elmahdy
Department of Civil Engineering, Faculty of Engineering, Universii Putra Malaysia (UPM), Serdang, Malaysia
Email: Samy903@yahoo.com
Received August 8, 2011; revised October 16, 2011; accepted November 15, 2011
ABSTRACT
The north-western part of Sudan, which is the driest region on earth has revealed newly surface and near surface pa-
leodrainage network underneath sand sheets ind icating the possibilities for econo mic groundwater reservoirs. Advanced
Space-born Thermal Radiometer (ASTER), the Shuttle Radar Topography Mission (SRTM ~90 m) DEMs and Quick-
bird images corroborate the presence of surface and near surface paleodrainage network. Bivariate quadratic surfaces
with moving window size of 3 × 3 were fitted to the SRTM DEM. The second derivative surface curvature was calcu-
lated to reveal landform classes that may receive most of fossil water. The results showed that the new unnamed de-
pression which recharges by a longitudinal paleodrainage network may receive vast amount of groundwater during hu-
mid phases. The results demonstrate that the D8 and curvature algorithms are very efficient tools for revealing and
characterizing hydrological elements in arid and semi-arid regions and they provide information for hydrological ex-
ploration in remote deserts over large scale prior to geophysical survey.
Keywords: Darfur; Groundwater; Sudan; Paleodrainage Network
1. Introduction
North-western Sudan is sandy desert and shallow reser-
voirs are drying up and groundwater levels are declining
due to over pumping and exploitation to meet the de-
mand of an ever increasing drought affected population
built up [1]. About 80% of the inhabitants of Sudan de-
pend on groundwater for their livelihood almost all the
year round.
Nowadays, the groundwater could greatly affect the
urban and reclamation development [2-5]. However, dur-
ing the late Quaternary, the region was flooded as indi-
cated by lacustrine deposits and playa carbonate deposits,
dried extinct drainage patterns, flora and signs of prehis-
toric hum an oc cup at i on [2,6-13].
Due to the study area being extremely dry and partly
inaccessible, it is practically impossible to reveal and
explore economic amounts of groundwater on a large
scale through traditional methods. Digital Elevation
Models (DEMs) (NASA Shuttle Radar Topography Mis-
sion—SRTM) enable more reliable analysis in many
applications than before, especially in remote arid and
semi-arid regions such as the study are SRTM data is
widely used as a main topographic dataset to evaluate
regional scale hydrology, providing accurate and consis-
tent elevation data on a worldwide basis [14]. Prelimi-
nary analyses done by [15] suggests that SRTM DEM in
low relief areas over large scale is likely more accurate
than the mission specifications. For this reason, SRTM
DEM has a great and partly unrealized potential for hy-
drologic applications.
The eastern Sahara represents an ideal location for re-
vealing near surface geological and geomorphological
features and groundwater exploration over large scale
using SAR data.
The depth of penetration near surface paleodrainages
using C-band radar is up to 0.5 m [16,17]. This is be-
cause of the scarcity of lakes (water bodies), rainfall,
vegetation cover, low topography and low soil moisture
content. Few studies have been applied for mapping pa-
leodrainage network in the Northern Darfur depression,
Sudan. [18] has mapped paleolake shorelines and re-
gional buried paleodrainage network over a wide area in
Northern Darfur using SRTM DEM and RADARSAT-1
images. Paleolake deposits [2,8], drainage network de-
lineated from Shuttle Imagine Radar (SIR-A) images us-
ing visual interpretation [2,8], mapping the Northern
Darfur basin using GPS and field observation and
LANDSAT image [13].
According to them, these features have developed dur-
ing the late Quaternar y when the clima te wa s wetter phases
C
opyright © 2012 SciRes. IJG
S. I. ELMAHDY
26
than it is today as indicted by lacustrine deposits and
playa carbonate deposits, dried extinct drainage patterns,
flora and signs of prehistoric human occupat ion.
The use of digital elevation models (DEMs) through
geographical information systems (GIS) is a powerful
approach in this matter, since automatic methods to ana-
lyze topographic features are allowed, with both opera-
tional and quality advantages. While GIS-based feature
extraction are processed with all advantages of digital
resources (speed, repeatability and computer integration
with other databases), the reduction of manual interve-
netions (and thus subjectivity) and the possibility of pa-
rametric approach represent qualitative distinctions from
interpretative-based methods [19-22].
In watershed studies, GIS modeling of DEMs is often
applied to erosion estimates [23], in which the automatic
extraction of its related variables consists as important
field of research [24]. Directly concerned to the compre-
hension of watershed drainage structure, watershed parti-
tion [25] and the identification of terrain units [26] are
important tasks to be supported by DEM analysis.
This study presents different approach based on auto-
matic and semiautomatic algorithms with verification of
visual interpretation in high resolution satellite images.
More specifically, this study used D8 algorithm, terrain
category by elevation and Wood algorithms with two
different sensors, SRTM DEM of spatial resolution ~90
m and Quick Bird image of spatial resolution 0.6 m, for
delineation and analysis the surface and near surface pa-
leodrainage and depressions that drained and veined the
western part of Sudan. Hence, the main goals of this
study are to:
1) Extract of the surface and near surface paleodrain-
age network in the north-western part of Sudan from 90
m DEM using D8 algorithm over large scale.
2) Precise mapping of paleo-shorelines of the Northern
Darfur depressions from 90 m DEM using terrain cate-
gory algorithm.
3) Classify the landform classes (morphometric features)
from ~90 m DEM using maximum curvature algorithm.
2. Study Area
The Darfur region occupies one fifth of the area of Sudan
comprising approximately 500,000 square kilometers. The
region borders Libya to the northwest, Chad to the west,
and the Central African Republic to the south-west. Ex-
tending from north to south, the climate in Darfur varies
from extremely arid, semi-arid to the wet semi-tropic
climate in the southern part [1].
The rainfall varies from zero in the north and gradually
increases southwards reaching 800 mm annually and rises
up to 1000 mm Jebel Marra highlands (Figure 1). The
water table lies between 10 m to 1000 m deep [2]. The
amount of groundwater recharge in Sahara Nubian and
Baggara basin have been estimated to be 20.6 × 106 m3
and 155 × 106 m3 respectively (Table 1).
The groundwater aquifer thickness varies between 250
m and 550 m [1,5,27]. The main lithology varies from
unconsolidated rocks such as Nubian sandstone with high
to low hydrological importance to fractured rocks with
medium to low hydrological importance (Figure 1).
3. Material and Methods
3.1. Data Sources and Pre-Processing
Three remotely sensed data were used in this study: a
SRTM DEM of ~90 m spatial resolution; the Quick bird
image of spatial resolution 0.6 m. The SRTM data prod-
uct is one of the most valuable global resources of to-
pographic data to date [28]. However, the original [29]
DEMs contain holes that represent no data or attenuated
measurement of elevation. These holes affect the accu-
racy of hydrological modeling which requires a continu-
ous flow surface [28]. The new versions of SRTM DEMs
[30,31] are the result of interp olation efforts of the uned-
ited SRTM DEM version 2 (~90 m). These SRTM data
has been chosen because the depth of penetrating near
surface paleodrainages is up to 0.5 m [16-18]. Near sur-
face features (palaeolake and paleodrainages) appear
dark in SRTM data. The second set was Quickbird im-
ages of spatial resolution 0.6 m and available as a geo-
graphic (long/lat) projection , with the WGS84 horizontal
datum currently available from US department of State
Geographer and Google earth browser [32]. These im-
ages were used to compare visually the textural features
(paleodrainage networks) evident in SRTM and to decide
if patterns expressed by high-resolution optical data are
different. The third data set was The Advanced Space-
borne Thermal Emission and Reflection Radiometer (AS-
TER) DEM of spatial resolution 30 m is distributed as a
geographic (long/lat) projection, with the WGS84 hori-
zontal datum currently available from the Earth Remote
Sensing Data Analysis Center (ERSDAC) database [33].
The auxiliary data in terms of hydrological map of
Sudan [34], Quickbird images and ASTER DEM was used
to compare visually the textural features (paleodrainage
networks) evident in STRM and to decide if patterns ex-
pressed by high-resolution remotely sensed data are dif-
ferent.
3.2. Delineation of the Northern Darfur
Paleolake and Its Depressions
Four sets of terrain category by elevation and shaded
relief maps were generated from DEM at sun azimuth
(compass direction to the sun) of 335˚ and sun elevation
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY
Copyright © 2012 SciRes. IJG
27
(a) (b)
Figure 1. Location of study area (a) and hydrological map (b) in the north west of Sudan (after hydrological map of Sudan,
1989).
Table 1. Lithological formation and their hydrological characteristic (Omer, 2002).
Basin Aquifer thickness (m) Water level (m) Amount of water recharge (106 m3)
Sahara Nubian 300 - 500 10 - 50 20.6
Central Darfur 250 - 550 25 - 100 47.6
Western Kordofan 300 - 500 50 - 70 15
Baggara 300 - 500 10 - 75 155
(elevation angle for the sun) of 45˚. For the boundary and
shoreline delineation of northern Darfur paleolake, 35
times vertical exaggeration was applied and paleolake
boundaries and shorelines were represented in different
colors to facilitate discrimination and visualization.
tures quantitative, numerical attribute of topography [36].
At a more accuracy of hydrological modelling, it is
importantly to eliminate the sinks (faulty values) in DEM
by means of filtration through fill in algorithm so there
no runoff interruption [37,38].
Boundaries and shorelines of the Northern Darfur pa-
leolake were manually screen digitized in each thematic
map of terrain category and shaded relief and then the
resulting boundaries and shorelines maps were merged
into one. The total area of each boundary and shoreline
(polygon) was then calculated using Envi v.4.5 and the
open source MicroDEM version 10 software [35].
To extract the surface and near surface paleodrainage
network, the deterministic-eight node (D8) single flow
direction algorithm [39,40] was used. The D8 algorithm
works well to mimic the flow of rivers and streams, and
flow convergence in valleys [29]. The algorithm deter-
mines into which neighboring pixel any water in a central
pixel will flow naturally. Flow direction in a DEM is
calculated for every central pixel of input blocks of a 3 ×
3 window, all the time comparing the value of the central
pixel in the window with the value of its eight neighbo rs.
Then, find the steepest slope and downhill slope of a
central pixel to one of eight neighboring pixels.
3.3. Extraction of Surface and Near Surface
Paleodrainage Networks
For a semi-automated method, it is necessary to imple-
ment algorithms which reveal li near geomorphom etric fea-
S. I. ELMAHDY
28
The threshold definition (stream definition) is per-
formed after flow direction and flow accumulation are
determined. Selection of a threshold value depends on de-
sired performance [41]. In this study, a threshold of 5000
cells was chosen as the resulting drainage was found to
correspond with those d e lineated in the STRM DEM.
Extraction and delineation of surface and near surface
paleodrainage network is implemented in the commercial
ArcGIS v9.2 software [41]. The new resultant paleodrai-
nage network maps were mosaiced using Envi. v. 4.5 soft-
ware. For better analysis and perspective viewing, the
paleodrainage network was draped over shaded relief re-
flectance generated from the DEM.
3.4. Classifications of Landform and Relief
Shape
Identification of landform express paleoflow accumula-
tion and paleoflow dissipation zones can be realized by
curvatures maps of Wood’s algorithm, which is basically
based on terrain parametrization analysis of DEM. In
terrain parameterization analysis, the second derivatives
of DEM (Table 2) are the basic parameters of mor-
phometric analysis and landform classification [42-44].
To calculate landform classes, a 3 × 3 local window is
passed over the DEM and slope change of a central pixel
in relation to its neighbors is derived by a bivariate
quadratic function as follows:
z = ax2 + by2+ cxy + dx + ey + f (1)
where x and y are the Cartesian coordinates, and the let-
ters a through f are the co efficien ts of the polyno mial that
contain information on relief attributes [42,43]. Curvature
parameters can be separated into two orthogonal compo-
nents, profile and plan curvatures [34].
Wood (1996) used two alternative parameters, maxi-
mum and minimum curvatures if the slope normal is in-
clined and an aspect can be calculated. For example, plan
convexity or planc (intersecting with XY plane), mini-
mum curvature or minic and maximum curvature maxic
(intersecting in any plane). Wood (1996) identified geo-
morphometric features based on unique sets of slope steep-
ness, cross sectional maximum and minimum curvatures.
Wood (1996) defined set of rules to classify DEMs into
landform classes (Table 3).
In the results, valleys and channels have the most ne-
gative values for minimum curvature and ridges have the
most positive value for maximum curvature. In other words,
the curvature has negative values mean convex shapes
and paleoflow dissipation and curvature has positive
values mean concave ones and paleoflow accumulation.
Identification and classification of landform is imple-
mented in t he o pen source MicroDEM v.10 softw are.
To evaluate the hydrol ogical setting and areas of ground -
water recharge for each depression, the percentage of land-
form classes (channels, pits and/or micro-depressions)
per total surface area of each depression were calculated.
3.5. Validation of the Extracted Paleodrainage
Network
Three validation methods were constructed to assess the
accuracy of revealed features. In the first validation; the
revealed paleodrainage network from SRTM DEM and
those from Quickbird images [32] and ASTER DEM [33]
were correlated visually. The validation was performed
by comparing the textural features (surface and near sur-
face paleodrainage network) and deciding if these pat-
terns are different. In the second validation method, the
Table 2. Morphometric parameters (Evans, 1972; Wood, 1996; Shary et al., 2002).
Morphometric parameter Formula Description
Slope steepness (degrees) arctan (sqrt(d2 + e2)) Magnitude of steepest gradient in both X and Y
directions.
Maximum Curvature (1/m) n × g × ( a b + sqrt((a b) 2 + c2)) In any plane
MinimumCurvature (1/m) n × g × ( a b sqrt((a b)2 + c2)) In any plane
Profile Curvature (1 / m) n × g × (a × d2+b × e2+c × d × e) /(d2 + e2) (1 +
(d2 + e2))1.5
Vertical component in direction of aspect.
(Intersecting with the plane of z axis and aspect
direction).
Plan curvature (1/m) n × g × (b × d2 + a × e2 c × d × e)/(d2 + e2)1.5 Horizontal component in direction of aspect
(Intersecting with the X, Y plane).
Table 3. Morphometric features classification criteria (modified from Wood (1996)).
Morphometric features Slope (˚) Maximum curvature Minimum curvature
Ridge 0, +va +va 0
pass 0 +va –va
Channel 0, +va –va –va
Pit (micro-depression) 0 –va –va
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY 29
revealed surface and near surface paleodrainage network
were compared to those in the previous study by [18]. In
the third validation method, the resultant maps were com-
pared with hydrological map of Sudan.
4. Results and Discussions
4.1. Delineation of the Northern Darfur
Paleolake Shorelines and Its Depressions
The results of DEM processing and analysis (Figures 2-4)
show the Northern Darfur paleolake is a multi-elevation
depression (regional basin) and consists of seven depres-
sions. These depressions are the unnamed depression,
Al-Natrun, Nukhiela, Hammra, Oyo, Hussien and Wadi
Fesh Fesh. The unnamed depression is located at an ele-
vation ranges from 500 m to 540 m and occupies an area
of about 2000 km2, Al Natrun depression is located at
elevation 530 m and occupies an area of about 895.57
km2, the Nukhiela depression is located at an elevation of
540 m and occupies an area of about 713.8 km2, the Hus-
sien depression is located at an elevation ranges from
470 m to 540 m and occupy an area of about 433.4 km2,
the Fesh Fesh corridor is located at an elevation range
from 530 m to 550 m and has flow path of 51.7 km in
length, Hamra depression is located at an elevation
ranges from 515 m to 535 m and occupies an area of
about 252 km2, and the Oyo depression is located at an
elevation range from 478 m to 530 m and occupies an
area of about 286 km2.
These results are in contradiction wit h [18] who mapped
the paleoshorelines of the Northern Darfur and its de-
pressions from shaded relief maps and RADARSAT-1
images using visual interpretation. Geologically, the Nor-
thern Darfur paleolake and its depressions are probably
results from intersections northeast-southwest and north-
west-south east of regional fault zones, which serve as
underground channels for upward transport deep seated
fossil water to the land surface and downward transport
to the deep aquifers [37]. So, some portions of the sur-
face and near surface paleodrainage network routs can
coincide with discharges of deep-seated fossil water path-
ways [1,4,5,20,38,39].
4.2. Extraction of Surface and Near Surface
Paleodrainage Networks of North Western
Sudan
Figure 5 shows the resultant surface and near surface
paleodrainage network map derived from DEM using D8
algorithm where the major paleodrainage network is rep-
resented by white and blue lines to facilitate the visual
interpretation. The paleodrainage network map shows four
tributaries (Wadis) drain the Northern Darfur paleolake.
These features are the Uweinat tributaries (north) drain
from Libyan and Egypt borders and end at Wadi Fesh
Fesh, Hammra and Oyo depressions with bearing of
N13E, the Erdi Ma tributaries (east) drain from the Er-
diMa Mountain (Chad) and end at the unnamed depres-
sion with bearing of N60W. The Ennedi tributaries drain
from Ennedi Mountain and end at the unnamed depres-
sion with bearing of S42W. They are very longitudinal
and parallel in shape of about 350 km in length, and the
Figure 2. The Northern Darfur Basin (left) and its muti-elevation depressions are represented in different colors (right).
These were derived from DEM using terrain category by elevation algorithm.
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY
30
Figure 3. Topographic profiles perfor me d from DEM of the Northern Darfur Depressions, Sudan.
Figure 4. 3D view s visualize (a) Hammra and Oyo depressions; (b) Hussien and Nukhilea depressions; (c ) The unnamed de-
pression; and (d) Al-Natrun depression.
Marra tributaries drain from Marra Mountain and end at
Wadi Howar wi t h bea ring S65W.
Other tributaries drain from eastern flank of the Nor-
thern Darfur depression and en d at Al Natru n depression ,
Nukheila depression and Hamrah depression with bear-
ing of N72E. They are disconnected and dendritic in shape
of about 71 km in length. They have well developed in
late Quaternary (pre-Holocene period) of high fluvial ac-
tivity, intensive precipitation, surface runoff and a long-
live process that completed during wet phases.
Some of these paleodrainage networks were delineated
from RADARSAT-1 C band, Shuttle Imagine Radar
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY 31
Baggara Basin
M. Erdi Ma
M. Marra
W. Howar
Chad
Libya Egypt
M. Ennedi
28°3'15"E
28°3'15"E
25°2'20"E
25°2'20"E
22°1'25"E
22°1'25"E
18°33 ' 0" N
18°33 ' 0" N
15°32'5"N
15°32'5"N
12°31 '10 "N
12°31 '10 "N
Northern Darfur
Depressions
Mourdi
Depression
M. Erdi Ma
M. Marra
W. Howar
Libya Egypt
M. Ennedi
27°12'15"E
27°12'15"E
25°20'45"E
25°20'45"E
23°29'15"E
23°29'15"E
20°34 '15 "N
20°34 '15 "N
18°4 2' 4 5" N
18°4 2' 4 5" N
16°51 ' 15 " N
16°51 ' 15 " N
14°59'45"N
14°59'45"N
(a) (b)
Figure 5. (a) Regional surface and near surface paleodrainage network extracted from DEM using D8 algorithm for the
north-western apart of Sudan; and (b) Nor t he rn Darfur Paleolake.
SIR-A images and SRTM DEM using visual interpreta-
tion [1,6-8,18], and field observation [2,13]. Hydrologi-
cally, the longitudinal and parallel p aleodrainag e netwo rk
in shape clearly reflects flatter to gently slope for terrain
and high rate of infiltration and dendritic paleodrainage
network reflects steep slopes for terrain. Therefore, high
rate of groundwater infiltration and accumulation are
greater in flat areas [40]. Some of these paleodrainage
networks have been related regional fault zones and joint
systems and are extremely significant as they relate to
preferential flow zones in the Nubian Sandstone [45-47].
Verification of the extracted paleodrianage network
was made by comparing them with 0.6 m resolution
Quickbird images (Figure 6). The paleodrainage network
delineated from DEM using D8 algorithm coincide with
paleodrainage network identified by visual interpretation
in the high resolution satellite images and much more
than revealed previously. The analysis of the paleodrai-
nage network map and corresponding high resolution
optical satellite images shows the powerful of the method
to delineate and map the surface and near surface pa-
leodrainage netwo rk in the we stern si de of S uda n.
4.3. Landform and Relief Shapes and
Groundwater Flow Accumulation
Figures 7 and 8 shows the resultant classification map
where the major landform classes (e.g. ridges, channels,
passes and pits) are represented in different colours to
facilitate the interpretation. The most positive curvatures
(>0) are represented by red color defines ridges, flanks of
the channels and potential levees and over bank deposits
(foot-wall of fault zones) with steep slopes and paleoflow
dissipation zones, while the most negative curvature (<0)
are represented by blue color highlights axis of channels,
valleys, thalwegs and paleoflow accumulation zones.
Zero values for curvature correspond to planner area and
paleoflow transition zones [36,44].
Curvature maps display a great help in bringing out the
definition of channels, valleys and depressions, thus zones
of groundwater accumulation [9,42-44,48]. These fea-
tures are attributes which often highlight the regions
where the rocks are broken and fractured and zones of
groundwater accumulation. This is a regular result sub-
stantiating facts that paleod rainage network and accumu-
lation zones often relate to fault intersections [45,49]. In
general, in reality, classes with flat and high value for
negative curvature and/or classes of zero values for cur-
vature (e.g. pits, micro-depressions and channels) support
low discharge of overland flow and high rate of infiltra-
tion. Thus, groundwater potentiality is much more in the
flat and high negative curvature [49]. Flatter classes with
zero value for curvature then will give more chance for
groundwater accumulation.
Results of landform classification revealed that from
total 2000 km2 coverage of the unnamed depression, about
40.2% are classified as channels, valleys and 9.4% are
classified as pits and depressions with recharge area of
about 993 km2 (Table 4).
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY
32
Table 4. The most groundw ater recharge classes per total area in km2 of the Northern Darfur depressions.
Depression Channels (%) Pit (%) Total area (%) Total area (km2) Recharge area (km2)
Un-named 40.2 9.4 49.6 2000 922
Al Natrun 36.2 7.2 47.4 895.6 424.5
Nukheila 35 7 37 713.8 246.1
W. Fesh Fesh 34.3 7 41.3 461 170.5
Hussein 35.4 6.8 42.2 433.4 182.8
Hamra 35.7 7.1 37.3 252 95.2
Oyo 32.2 5.1 37.3 286 106.6
22°4 2 '9 " E
22°4 2 '9 " E
14°41'33"N
23°2 6 '4 2 "E
14°41'33"N
13°57'0"N
13°57'0"N
23°2 6 '4 2 "E
(a) (b)
Figure 6. Two zooms of surface paleodrainage network derived from DEM using D8 algorithm (a) and corresponding 60 m
resolution quickbird images (b).
(a) (b)
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY 33
(c) (d)
Figure 7. Maximum curvature maps derived from DEM using Wood’s algorithm for (a) Al-Natrun; (b) Hammra; (c) Hussien
depressions; and (d) Nukhiela de pr e ssion.
This area has the highest percent of zones of accumula-
tion, while the Wadi Fesh Fesh and Oyo depression have
the smallest revealed area of zones of accumulation of
approximately 34.314% and 32.272% respectively (Fig-
ure 9 and Tab le 4).
A multi-disciplinary approach using an integration of
different datasets and algorithms is a powerful approach
in the revealing and analysis surface and near surface
paleodrainage network and flow accumulation zones that
are closely associated with groundwater accumulation
where there is a lack of hydrological and geological in-
formation.
The proposed approach can be used to extract and
simulate several paleodrainage network conceal sand sheets.
The analysis of the results and corresponding high reso-
lution satellite images and hydrological map of Sudan
(1989) shows the eff ectiveness of the method to map and
analyze the overall pattern of the hydromorphological
elements in the north western part of Sudan. The orienta-
tions of these features were found to be in the NW-SE,
NE-SW, NNW-SSE, NNE-SSW and WNW-ESE direc-
tions and end in the seven revealed depressions. These
were confirmed by patterns of streams and regional fault
and hydrological unites that are reported in the geologi-
cal and hydrological maps (Figure 1).
Future work will involve the application of both SAR
data (e.g. ALOSPALSAR L band), and LANDSAT En-
hanced Thematic Map (ETM+) images (e.g. thermal band)
which will be applied in the future in order to delineate
near surface paleodrainage network, fractured zones and
soil moisture and enable a more hydrological model of
the Northern Darfur paleolake to be constructed.
5. Conclusions
A set of automated algorithms were used successfully to
reveal several new surface and near surface paleodrain-
age stream networks in the north western part of Sudan
and northern Darfur paleolake (mutli-elevation paleolake),
estimate its elevation and total area and classify land-
forms. The D8 algorithm was used to reveal surface and
near surface paleodrainage networks much more than
revealed previously. D8 algorithm allows revealing suc-
cessfully geomorphometric features over large scale for
the first time. These geomorphometric features are terms
correspond to physical entities (surface and near surface
paleodrainage networks). The supplementary data such
as 30 m DEM produced from ASTER and 0.6 m Quick-
bird satellite images corroborate the presence o f new sur-
face and near surface paleolake and paleodrainage net-
work. This is an important finding, confirming that the
methods are reliable and produce consistent results. Re-
sults of landform classification revealed that from total
2000 km2 coverage of the unnamed depression, about
40.2% are classified as valleys and 9.4% are classified as
pits and depressions with recharge area of about 992 km2.
The results show the efficiency of the methods to identify,
characterize and analyze hydromorphological elements
(paleodrsainage network and landform classes) in the
north western part of Sudan. The obtained results dem-
onstrate that integrated methods, which are based on two
different sensors and semi-automatic algorithms, are ef-
ficient for hydrological application in arid and semi-arid
regions where inaccessible locations and scarcity of hy-
drological information is available.
The obtained maps can be used as a reference in
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY
34
(a) (b)
(c)
Figure 8. Maximum curvature maps derived from DEM using Wood’s algorithm for (a) Oyo depression; (b) W Fesh Fesh;
and (c) The un-named depression.
Figure 9. Total area and percentage of recharge area of each depression in the Northern Darfur Paleolake.
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY 35
groundwater exploration prior to geophysical survey and
agriculture designs.
REFERENCES
[1] A. M. Omer, “Water Sources and Environment in Sudan,”
Network for Water and Sanitation, Vol. 4, No. 3, 1997,
pp. 5-7.
[2] C. V. Haynes, “Oyo: A Lost Oasis of the Southern Libyan
Desert,” The Geographical Journal, Vol. 155, No. 2, 1989,
pp. 189-195. doi:10.2307/635060
[3] R. K. Farnsworth, E. C. Brrett and M. S. Dhanju, “Appli-
cation of Remote Sensing to Hy drology Including Ground-
water,” United Nations Educational Scientific and Cul-
tural Organization, Paris, 1984.
[4] A. Ho, “Gravity and Seismic Refraction Measurements
for Deep Ground Water Search in Southern Darfur Re-
gion, Sudan,” NWWA Conference on Surface and Bore-
hole Geophysical Methods in Ground Water Investigations,
Fort Worth, 12-14 February 1985, pp. 106-120.
[5] A. M. Omer, “Focus on Groundwater in Sudan,” Envi-
ronmental Geology, Vol. 41, No. 8, 2002, pp. 972-976.
doi:10.1007/s00254-001-0476-9
[6] F. Wendorf, A. Close, R. Schild, R. Said, C. V. Haynes,
A. Gautier and N. Hadidi, “Late Pleistocene and Recent
Climatic Changes in the Egyptian Sahara,” Geographical
Journal, Vol. 143, No. 2, 1977, pp. 211-234.
doi:10.2307/1795874
[7] J. McCauley, G. Schaber, C. Breed, C. V. Haynes, J.
Grolier, B. Issawi, C. Elachi and R. Blom, “Subsurface
Valleys and Geoarchaeology of Egypt and Sudan Revealed
by RADAR,” Science, Vol. 218, 1982, pp. 1004-1020.
doi:10.1126/science.218.4576.1004
[8] H. Pachur and F. Rottinger, “Evidence for a Large Ex-
tended Paleolake in the Eastern Sahara as Revealed by
Space Borne Radar Lab Images,” Remote Sensing of En-
vironment, Vol. 61, No. 3, 1997, pp. 437-440.
doi:10.1016/S0034-4257(96)00210-6
[9] I. Florinsky and G. A. Kuryakova, “Influence of Topog-
raphy on Some Vegetation Cover Properties,” Catena,
Vol. 27, No. 2, 1996, pp. 123-141.
doi:10.1016/0341-8162(96)00005-7
[10] J. Ritchie and C. Haynes, “Holocene Vegetation Zonation
in the Eastern Sahara,” Nature, Vol. 330, No. 6194, 1987,
pp. 645-647. doi:10.1038/330645a0
[11] H. Pachur and P. Hoelzmann, “Paleoclimatic Implications
of late Quaternary Lacustrine Sediments in Western Nu-
bia, Sudan,” Quaternary Research, Vol. 36, No. 3, 1991,
pp. 257-276. doi:10.1016/0033-5894(91)90002-M
[12] H. J. Pachur and F. Rottinger, “Evidence for a Large Ex-
tended Paleolake in the Eastern Sahara as Revealed by
Spaceborne Radar Lab Images,” Remote Sensing of En-
vironment, Vol. 61, No. 3, 1997, pp. 437-440.
doi:10.1016/S0034-4257(96)00210-6
[13] P. Hoelzmann, B. Keding, H. Berke, S. Kropelin and H.
Kruse, “Environmental Change and Archaeology: Lake
Evolution and Human Occupation in the Eastern Sahara
during the Holocene,” Palaeogeography, Palaeoclima-
tology, Palaeoecology, Vol. 169, 2001, pp. 193-217.
doi:10.1016/S0031-0182(01)00211-5
[14] K. J. Bhang, F. W. Schwartz and A. Braun, “Verification
of the Vertical Error in C Band Srtm Dem Using Icesat
and Landsat-7, Otter Tail County Mn,” IEEE Transaction
of Geoscience & Remote Sensing, Vol. 45, No. 1, 2007,
pp. 36-44. doi:10.1109/TGRS.2006.885401
[15] K. J. Bhang and F. W. Schwartz, “Limitations in the Hy-
drologic Application of C Band Srtm Dems in Low Relief
Settings,” IEEE Geoscience and Remote Rensing Letters,
Vol. 5, No. 3, 2008, pp. 497-501.
doi:10.1109/LGRS.2008.920712
[16] L. E. Roth and C. Elachi, “Coherent Electromagnetic
Losses by Scattering from Volume in Homogeneities,”
IEEE Transactions of the Antennas and Propagation, Vol.
23, 1975, pp. 674-675. doi:10.1109/TAP.1975.1141170
[17] G. Schaber, J. Mccauley and C. Breed, “The Use of Mul-
tifrequency and Polarimetric SIR-C/X-SAR Data in Geo-
logic Studies of Bir Safsaf, Egypt,” Remote Sensing of
Environment, Vol. 59, No. 2, 1997, pp. 337-363.
doi:10.1016/S0034-4257(96)00143-5
[18] E. Ghoneim and F. El-Baz, “The Application of Radar
Topographic Data to Mapping of a Mega-Paleodrainage
in the Eastern Sahara,” Journal of Arid Environments,
Vol. 69, No. 4, 2007, pp. 658-675.
doi:10.1016/j.jaridenv.2006.11.018
[19] I. D. Moore, P. E. Gessler, G. A. Nielsen and G. A. Pe-
terson, “Soil Attribute Prediction Using Terrain Analysis,”
Soil Science Society of America Journal, Vol. 57, No. 2,
1993, pp. 443-452.
doi:10.2136/sssaj1993.03615995005700020026x
[20] I. Florinsky and G. A. Kuryakova, “Influence of Topog-
raphy on Some Vegetation Cover Properties,” Catena,
Vol. 27, No. 2, 1996, pp. 123-141.
doi:10.1016/0341-8162(96)00005-7
[21] P. A. Goovaerts, “Geostatistical Approaches for Incorpo-
rating Elevation into the Spatial Interpolation of Rain-
fall,” Journal of Hydrology, Vol. 228, No. 1-2, 2000, pp.
113-129. doi:10.1016/S0022-1694(00)00144-X
[22] M. M. Valeriano and O. A. Carvalho Jr., “Geoproces-
Samento de Modelos Digitais de Elevac-a˜o Para
Mapeamento da Curvature Horizontal em Microbacias,”
Revista Brasileira de Geomorfologia, Vol. 4, No. 1, 2003,
pp. 17-29.
[23] D. K. Molna´r and P. Y. Julien, “Estimation of Upland
Erosion Using GIS,” Computers & Geosciences, Vol. 24,
No. 2, 1998, pp. 183-192.
doi:10.1016/S0098-3004(97)00100-3
[24] P. J. J. Desmet and G. Govers, “A GIS Procedure for
Automatically Calculating the USLE LS Factor on To-
pographically Complex Landscape Units,” Journal of Soil
and Water Conservation, Vol. 51, No. 5, 1996, pp. 427-
433.
[25] L. E. Band, “Topographic Partition of Watersheds with
Digital Elevation Models,” Water Resources Research,
Vol. 22, No. 1, 1986, pp. 15-24.
doi:10.1029/WR022i001p00015
[26] G. Miliaresis, “Geomorphometric Mapping of Zagros
Copyright © 2012 SciRes. IJG
S. I. ELMAHDY
36
Ranges at Regional Scale,” Computers & Geosciences, Vol.
27, No. 7, 2001, pp. 715-728.
doi:10.1016/S0098-3004(00)00168-0
[27] W. Iskander, “An Appraisal of Ground Water Resources
of Zalingei Area—Darfur Province—Sudan,” M.Sc. The-
sis, University of Arizona, Tucson, 1969.
[28] B. Rabus, M. Eineder, A. Roth and R. Bamler, “The
Shuttle Radar Topography Mission—A New Class of Di-
gital Elevation Models Acquired by Space Borne Radar,”
ISPRS Journal of Photogrammetry and Remote Sensing,
Vol. 57, 2003, pp. 241-262.
doi:10.1016/S0924-2716(02)00124-7
[29] ftp://e0srp01u.ecs.nasa.gov/srtm
[30] A. Jarvis, H. Reuter, A. Nelson and E. Guevara, “Hole-
Filled Seamlen SRTM Data v3,” 2006.
ftp://e0srp01u.ecs.nasa.gov/srtm
[31] http://srtm.csi.cgiar.org/Index.asp
[32] www.Googleearth.com
[33] http://www.gdem.aster.ersdac.or.jp/search.jsp
[34] Hydrological Map of Sudan, “Scale 1:2000000, Ministry
of Energy and Mines,” Khartoum, 1989.
[35] P. L. Guth, “Geomorphometry in MICRODEM,” In: T.
Hengl and H. I. Reuter, Eds., Geomorphometry: Concepts,
Software, Applications. Developments in Soil Science Se-
ries, Elsevier, Amsterdam, 2008, pp. 351-366.
[36] A. H. Ehsani and F. Quiel, “Application of Self Organiz-
ing Map and SRTM Data to Characterize Yardangs in the
Lut Desert, Iran,” Remote Sensing of Environment, Vol.
112, 2008, pp. 3284-3294. doi:10.1016/j.rse.2008.04.007
[37] D. G. Tarboton, “A New Method for the Determination of
Flow Directions and Contributing Areas in Grid Digital
Elevation Models,” Water Resources Research, Vol. 33,
1997, pp. 309-319. doi:10.1029/96WR03137
[38] M. Klimanek, “Geoinformation Support of Derived Map-
ping Based on Digital Terrain Model,” Ph.D. Thesis,
University of Agriculture and Forestry Brno, Brno, 2010.
[39] S. Jenson and J. Domingue, “Extracting Topographic Struc-
ture from Digital Elevation Model Data for Geographic
Information System Analysis,” Photogrammetric Enginee-
ring and Remote Sensing, Vol. 54, No. 11, 1988, pp. 1593-
1600.
[40] J. F. O’callaghan and D. M. Mark, “The Extraction of
Drainage Networks from Digital Elevation Data,” Com-
puter Vision, Graphics, and Image Processing, Vol. 28,
No. 3, 1984, pp. 323-344.
doi:10.1016/S0734-189X(84)80011-0
[41] Environmental Research Institute, “Using ArcGIS Geo-
statistical Analyst Arc GIS software Manual,” Redlands,
2007.
[42] I. S. Evans, “General Geomorphology, Derivatives of
Altitude and Descriptive Statistics,” In: R. J. Chorley, Ed.,
Spatial Analysis in Geomorphology, Methuen & Co. Ltd,
London, 1972, pp. 17-90.
[43] J. Wood, “The Geomorphologic Characterization of Di-
gital Elevation Models,” Ph.D. Thesis, University of Lei-
cester, Leicester, 1996.
[44] P. A. Shary, S. L. Sharaya and A. V. Mitusov, “Funda-
mental Quantitative Methods F Land Surface Analysis,”
Geoderma, Vol. 107, No. 1-2, 2002, pp. 1-32.
doi:10.1016/S0016-7061(01)00136-7
[45] A. Shatta, “Remarks on the Regional Geological Struc-
ture of the Ground Water Reservoir of E1-Kharga and
Dakhla Oases,” Bulletin Society Geography, Vol. 34, 1961,
pp. 177-186.
[46] J. Ball, “Problems of the Libyan Desert,” The Geogra-
phical Journal, Vol. 70, No. 1, 1927, pp. 21-38.
[47] G. Caton-Thompson, “Kharga Oases in Prehistory with A
Physiographic Introduction by E. W. Gardner,” Univer-
sity of London, Athlone Press, London, 1952.
[48] I. Samy, “Remote sening for Hydrological Application in
the Vacinity of AlDhied,” M.Sc. Thesis, University Tech-
nology, Malaysia (UTM), Kuala Lumpur, 2006.
[49] R. N. Subba, “Groundwater Potential Index in a Crystal-
line Terrain Using Remote Sensing Data,” Environmental
Geology, Vol. 50, No. 7, 2006, pp. 1067-1076.
doi:10.1007/s00254-006-0280-7
Copyright © 2012 SciRes. IJG