In This Issue
Technologies for Clean Water
September 1, 2008 Volume 38 Issue 3
The Bridge, Volume 38, Number 2 - Fall 2008. The papers in this issue of The Bridge describe some recent advances in the search for water, in the distribution and treatment of water and wastewater, and in the modeling of complex water systems.

Remote Sensing of the Earth: Implications for Groundwater in Darfur

Wednesday, December 3, 2008

Author: Farouk El-Baz

Satellite images and sophisticated mapping techniques are revealing new sources of fresh
water in surprising places.

Images of Earth from space have improved steadily during the past 40 years. In the mid-1960s, photographs taken by astronauts of the Gemini, Apollo, Skylab, and Apollo-Soyuz missions using hand-held cameras with color film indicated the nature and composition of salient topographical features. Ancient rocks that contained a lot of iron and other dark elements appeared brown; limestone sediments looked bright; sands appeared golden yellow; and ocean currents were discernible. With those images, we were able to begin mapping hard-to-reach regions based solely on views from space.

In 1972, NASA initiated the next generation of images, which were relayed by digital sensors. From altitudes up to 920 kilometers above the Earth, sensing instruments looked down at rows of tiny spots, measured reflected sunlight from each of them, translated the light intensity into numbers, and beamed the numbers to receiving stations on the ground for study and analysis by researchers worldwide (Lillesand et al., 2004).

Detail in space images depends on two major factors: (1) the altitude of the spacecraft (the lower the orbit, the higher the resolution) and (2) the focal length of the camera lens (the longer the focal length, the greater the detail). However, ground resolution of space images was also determined by how much detail was allowable in images used for civilian purposes, as compared to images used for intelligence and military purposes. As the latter achieved more and more detail, the rules for civilian images were regularly relaxed. In the first satellite images, whole towns appeared as dots. Today a single car can be clearly identified in a high-resolution image.

One advantage of digital imaging from space is that a filter in front of a sensor’s lens can separate reflected light into various wavelengths. When certain bands of Landsat are used, for example, they become equivalent to visible light (Figure 1). They can then be combined with an infrared band or a thermal band that measures differences in the temperatures of rock, soil, and sand.

FIGURE 1 The electromagnetic spectrum, from gamma rays to radio waves, is shown on satellite images of particular wavelengths. Source: Adapted from

Another significant advantage of digital remote sensing is that it provides repeat coverage of the same area from an equal height by the same instrument. The repetitions provide comparisons that can show changes over time, and when images are overlapped using computer software, we can produce very accurate “change-detection maps,” which are essential to evaluating environmental changes due to natural processes and human activities.

This article includes a review of procedures for processing and analyzing satellite images followed by an example of how imaging data have been used to locate badly needed groundwater resources in the dry and troubled region of Darfur.

Image Processing
Satellite images are used to generate maps of drainage systems, geologic structures, thermal anomalies, geomorphologic features, and the distribution of vegetation. All of these factors are important to the understanding of a region, its environment, and its resources, particularly groundwater. Images must be preprocessed using radiometric and geometric corrections before data can be analyzed.

In radiometric corrections, images collected at different dates and times, and by different sensors, are normalized to each other so they can be directly compared. Geometric corrections are used to counteract sensor irregularities, terrain relief, and the effects of the curvature and rotation of the Earth. In some cases, geo-referencing involves transferring ground-control points.

Image Transformation
In image transformation, several multispectral bands (Figure 1) are used to generate a single image that highlights a particular feature or property of the land surface. Examples of transformations include image subtractions and ratios. Image subtractions are used to identify differences or changes among images of the same area acquired at different times. Image ratios are used to enhance particular information about the status of the land surface. For example, the normalized-difference vegetation index (NDVI) indicates the amount of green vegetation present in each picture element (pixel).

Enhancement procedures make it easier to interpret images by changing digital pixel values. These procedures are always the last step in the preprocessing of images. Enhancements can be either stretches (used with the image histogram) or spatial filters (used to highlight or suppress features based on pixel frequency). Density slicing, another type of enhancement, is used to select data ranges and colors for highlighting areas in gray-scale images.

Mosaicking of individual satellite scenes provides coverage of an entire region. The purpose of a mosaic is to create a seamless image from a group of individual scenes that may vary in brightness. Mosaicking involves three steps: (1) resampling images to refine their resolution; (2) matching the brightness of images; and (3) blending overlapping areas (Figure 2).

FIGURE 2 Digital mosaic composed of 65 Landsat images covering the total area of Egypt (more than one million square kilometers). Bright, parallel striations in the Western Desert are sand dunes shaped by prevailing winds from the north.

The classification of image data is used to produce thematic maps. This procedure involves using information in a multispectral image to classify each pixel. Unsupervised classification is useful for preliminary discrimination of spectral classes. Supervised classification involves using a priori knowledge of data to “train” computer software to identify categories in an image (Lillesand et al., 2004).

Change Detection
The process of change detection is used to identify differences in the state of an object or a phenomenon by observing it at different times (Singh, 1989). Change-detection maps are particularly important in monitoring the types, stages, and distribution of vegetation.

Image Analysis

Drainage Mapping
Satellite images are ideal for studying the movement of water on the Earth’s surface. Drainage maps, which are essential for flood control, searching for groundwater, and other water-related studies, can also be also used for mapping and interpreting regional drainage systems and individual stream courses for drainage-basin analyses. Such analyses include the study of drainage channels, as revealed in satellite images that reflect the influence of the fabric and structure of the underlying rocks.

Surface rocks may control the development of drainage systems by affecting the texture (shape) and density (spacing) of drainage. Both primary (in the rock fabric) and secondary (fracture-influenced) permeability of the surface rock are important. In areas where surface rocks are relatively impervious and easily eroded, a fine-textured surface drainage network of closely spaced channels develops. In areas of pervious rock, surface erosion is minimal because of infiltration, and coarse-textured drainage with widely spaced channels results.

In most cases, the uniformity of drainage patterns is an indicator of rock types. For example, a branching pattern implies homogeneous rock with little structural control. Deviations from this pattern (e.g., an increase in angularity, parallelism, or angle of confluence) may indicate a change in rock type or an increase in structural control.

The climatic conditions at the time of channel formation and the amount of erosion also significantly affect the development of drainage patterns. Higher rates of precipitation increase erosion and result in finer textures and more completely integrated patterns than in areas with lower precipitation where the rock fabric is more uniform. Relatively young, underdeveloped patterns imply fewer groundwater resources.

Surface drainage patterns are mapped using Landsat image composites, but in places covered by dry, fine-grained sand, radar data can be used to map ancient drainage patterns (Figure 3). The use of radar data for detecting old, sand-covered channels is well documented (e.g., El-Baz, 1988; McCauley et al., 1986; Robinson et al., 2000).

FIGURE 3 A Shuttle imaging radar (SIR-A) strip superposed on Landsat data. The radar waves penetrate the desert sand cover to reveal courses of ancient rivers and streams in an area of North Darfur in northwestern Sudan. Source: Adapted from

Structural Analysis
Fractures induce secondary porosity in any type of rock, and fracture zones store large amounts of water, usually in a network (e.g., NRC, 1996). Fracture zones may (1) drain large areas and extend for tens of kilometers in length; (2) act as conduits for water from mountainous regions, where the recharge potential from rainfall is high, to areas of lower elevation; (3) connect several horizontal groundwater aquifers, thus increasing the volume of water; or (4) represent areas of potentially high artesian pressure where water drained from higher elevations accumulates beneath the surface.

Thermal Anomalies
Anomalous cool areas in thermal images may represent water at or near the surface. This is because the latent heat content of water slows the process of absorption and the emission of radiation, thus, at a given time in the diurnal heating cycle, slowing the warming of moist soil (Pratt and Ellyett, 1979). Similarly, cooling during the night is also slowed. Thus moist soils have higher thermal inertia, which shows up as an anomalous cold area in the thermal data collected during daylight hours (Figure 4). Freshwater seeps into the ocean can also be detected by temperature differences.

Geomorphic Classification
Determining water accumulation requires an understanding of the geology and geomorphology of an area, because the amount of accumulation depends on the infiltration rates of surface water and the nature of the host rock, which can affect groundwater chemistry. The processes described in the preceding section can be combined to ensure that as much information as possible is extracted for the finished product. The geomorphologic classification of satellite images is based on interpreting their spectral information—the higher the spectral resolution the more information. 

Vegetation Mapping
The presence of vegetation in a region indicates that water is also present, either through irrigation or shallow, near-surface water. Mapping and monitoring the spatial distribution, type, and stage of vegetation over time can help determine (1) evaporation-transpiration rates and (2) the amount and type of water used in agriculture. Vegetation mapping can also help in locating potential water-bearing structures or buried channels that may act as preferential flow paths for subsurface water. Vegetation associated with fault zones may also indicate near-surface water. When vegetation-distribution maps are correlated with structural maps, they might lead to the identification of possible sites of ground-water resources. 

FIGURE 4 A dark patch (cool anomaly) in a thermal image of a sandy region in the Emirate of Sharjah (U.A.E.). The anomaly, which developed after rainfall on much higher topography further east, indicates water accumulation at or near the surface. Source: Center for Remote Sensing, Boston University.

Groundwater in Darfur
Water is essential for survival and for sustainable economic development. Water shortages already plague half the world’s population, and the United Nations (UN) estimates that 1.8 million people die every year because of unsafe water. Thus two key targets of the UN Millennium Development Goals are access to safe drinking water and adequate sanitation. Furthermore, one of the goals in the NAE “Grand Challenges for Engineering” project is to “provide access to clean water.” It is incumbent upon us as engineers to ensure that those who need it most have access to clean water.

Nowhere is the need more apparent than in the Darfur province of Sudan. The northern region of Darfur is part of the eastern Sahara of North Africa—the driest desert belt in the world—and the UN has declared that a shortage of water there during the past few decades is a major cause of the turmoil in the region (UNEP, 2008). Competition for meager water resources between sedentary farmers and nomadic populations has resulted in untenable violence and a major humanitarian crisis.

General Setting
Darfur is divided into three governorates. In Northern Darfur, which lies in the driest region on the planet, solar radiation is capable of evaporating 200 times the amount of rain the region receives (Henning and Flohn, 1977). Because of this hyper-aridity, human consumption and agriculture are completely dependent on groundwater resources. Growing populations, and the attendant increase in food and fiber requirements, have exacerbated the situation.
Recently, severe droughts have led to years of unrest and a vicious war in Darfur. Since 1968, the region has experienced seven-year cycles of dryness followed by cycles of meager rainfall (El-Baz, 1988). Because of water shortages during the dry cycles, sedentary farmers have encroached on wells that were usually used by nomads, which initiated many conflicts.

As described above, satellite images are an ideal tool for searching for groundwater resources in this region. The effectiveness of satellite images is enhanced by elevation data recently acquired by the Shuttle Radar Topography Mission (SRTM), which provides three-dimensional views.

Although the Sahara today is dry and subject to the erosive action of strong winds from the north, geological and archaeological data indicate that the climate was much wetter in the past. Surface water during moist periods formed lakes in topographic depressions, and much of the water from these basins may be stored in the underlying porous sandstone rocks. When the climate dried up, the wind covered these land features with sheets of sand.

The wind regime in the eastern Sahara traces a pattern that emanates from the coastal zone of the Mediterranean Sea (Figure 2). This pattern changes from southward in the northern areas to southwestward along the borders with the Sahel (El-Baz, 2000). Erosion scars throughout the desert suggest that this wind regime has been in effect for much of the last one million years. Careful observation reveals that sand accumulations in the eastern Sahara occur within or near topographic depressions, a characteristic that must be taken into account in any theory of the origin of the sand and the evolution of dune forms in space and time.

Radiocarbon dating and geo-archaeological investigations show that the eastern Sahara experienced a period of greater moisture from 10,000 to 5,000 years ago, as is evidenced by the numerous remains of human occupation throughout the Western Desert of Egypt and the neighboring region in northern Sudan (Figure 5). When a uranium-series technique was used to date lake carbonates from the Western Desert of Egypt and Northern Darfur, the results indicated that there have been five wet, paleo-lake-forming episodes in the past 320,000 years (Szabo et al., 1995). These wet episodes, which correlate with major interglacial stages, were separated by dry periods like the current one.

FIGURE 5 Hand axes and knives fashioned by humans from hard rock. These tools were dated by association to be from 6,000 years old (the smallest three) to more than 200,000 years old (the objects on either end). Such artifacts abound near ancient lake boundaries in southwestern Egypt and northwestern Sudan. Source: Photograph by the author.

Two dynamic forces are at play in the relationship between sand and water in the eastern Sahara. First, surface water systems worked from south to north during humid phases of climate, just as the Nile River does today. Rivers were responsible for transporting particulate materials and depositing them at the mouths of river channels. Second, the prevailing wind system during dry episodes worked in the opposite direction, from north to south. As the wind became the principal agent of modification, water deposits dried up, and sand was shaped into the dunes and sheets that now cover the desert surface.

This scenario implies that sand must have been borne by water and then sculpted by wind. During wet episodes, water percolated into the substrate through the porous layers of sandstone and was stored as groundwater (El-Baz, 2000). Thus we may infer from present sand dunes that there are groundwater resources in the area.

Southwestern Egypt
A flat, round, sand-covered area, some 300 kilo-meters in diameter, straddles the border between Egypt and Sudan (Figure 6). Named the Great Selima Sand Sheet, after an oasis on its eastern border, this area is morphologically a depressed basin covered by sand deposits with a few exposures of solid rock.

Field research conducted in 1978 was interpreted to indicate a groundwater accumulation in the lowest, eastern area in Egypt (El-Baz, 1988). Radar images obtained by the Space Shuttle and SRTM revealed the courses of rivers and streams leading to the area from highlands to the west and southwest (Figure 6). Hence, in 1995, the government of Egypt drilled exploratory wells, which were monitored for five years to ensure the presence of large amounts of groundwater.

FIGURE 6 Numerous channels emanate from the Gilf Kebir plateau and neighboring highlands, as shown by SRTM data. All channels lead to the low area toward the east where wheat and other crops are being profitably grown. Source: Adapted from Robinson et al., 2000.

Since 2000, 10,000-acre plots have been offered for agricultural development by the private sector in Egypt. Today wheat, chickpeas, peanuts, and other crops are being profitably raised in the region, irrigated by pervasive groundwater from the underlying porous sandstone (El-Baz, 2000). The salinity of this water is only 200 parts per million, which is lower than the salinity of Nile River water. Proven resources of this sweet, “fossil” water in the investigated area are large enough to support agriculture on 150,000 acres for 100 years.

Alluvial fans at the mouths of radar-revealed channels coincide with gentle slopes in the SRTM data, suggesting a long “stay time” of surface water and a high probability of finding groundwater. Indeed, groundwater wells in these regions are now producing low-salinity water. This is an example of how heterogeneous data from different sources can be used for the exploration for groundwater.

“1000 Wells for Darfur”
Interpretations of space-borne data for Northern Darfur suggest that water accumulated there in a lake-like expanse of 30,750 square kilometers, about the size of Lake Erie (Ghoneim and El-Baz, 2007). Horizontal sedimentary layers occur at 573 meters above sea level, the highest level of terraces formed at the shorelines of the lake water (Figure 7). Based on topographic information from SRTM data, the area of that lake would have been approximately 2,530 square kilometers.

During the residence time of ancient water in the Northern Darfur depression, for thousands of years before the lake dried up, much of it would have seeped into the substrate. This seepage probably occurred through the primary porosity of the underlying sandstone and/or secondary porosity caused by fractures in the rock, particularly the north-south-trending fault in the eastern part of the lake area.

FIGURE 6 Topographic rendition of the ancient lake in North Darfur based on SRTM data. Black dashes mark the locations of former lake terraces, as revealed by radar images (Radarsat 1), and shown in the enlargement of the box at lower right; younger streams were held back by the terraces, except for the one at the bottom of the radar image. The sketch at upper left shows the location of the lake relative to North, West, and South Darfur in western Sudan. Source: Adapted from Ghoneim and El-Baz, 2007.

Once the lake boundary had been completely mapped, based on space data, I conveyed this information to Omar Al-Bashir, president of Sudan, in the presence of Minister of Irrigation and Water Resources Kamal Ali, an engineer. President Al-Bashir stated that he recognized the importance of water shortages in the recurring crises in Darfur and the potential benefits of the discovery of this new water resource. He then announced an initiative called “1,000 Wells for Darfur.” News of the initiative was well received in Darfur. Upon reviewing the data, Governor of North Darfur Osman Kebir declared that “this brings hope for a better future. I have seen smiles on faces in Darfur.”

Shortly thereafter, the Egyptian Ministry of Water Resources and Irrigation offered to drill 20 wells to satisfy the urgent needs of the people of Darfur. Experts in this ministry have had a great deal of experience in drilling water wells in the nearly identical environment just north of Darfur.

In addition to the needs of the people of Darfur, the water will be used to meet the needs of a 26,000-strong UN-African Union peacekeeping force that will be deployed there. When I briefed UN Secretary General Ban Ki Moon, he immediately recognized the significance of the initiative and was amenable to placing it under the auspices of the UN, which will ensure both the expediency of the work and accountability for the spending of contributed funds. Efforts are under way to select the best sites for drilling. The well-drilling program will be a tangible example of how advanced space technology can be used to address a major humanitarian crisis.

As the Darfur example shows, using space-borne data and innovative approaches can lead to a better understanding of the potential for groundwater resources in dry lands and can increase the chances of locating groundwater for people in dire need of it. In the meantime, we must pursue research on innovative engineering techniques for better site selection, the drilling and pumping of water, the use of renewable energy in remote locations, and efficient water transport and delivery systems. I appeal to the engineering community to contribute as much as possible to meeting the challenge of “providing access to clean water” in an effort to save the people of Darfur and similar dry regions of the Earth.

El-Baz, F. 1988. Sand accumulation and groundwater in the eastern Sahara. Episodes 21(3): 147–151.
El-Baz, F. 2000. Satellite observations of the interplay between wind and water processes in the Great Sahara. Photogrammetric Engineering and Remote Sensing 66(6): 777–782.
Ghoneim, E., and F. El-Baz. 2007. DEM-optical-radar data integration for palaeohydrological mapping in the northern Darfur, Sudan: implications for groundwater exploration. International Journal of Remote Sensing 28(22): 5001–5018.
Henning, D., and H. Flohn. 1977. Climate Aridity Index Map. UN Conference on Desertification. United Nations Environment Program, Nairobi, Kenya, August 29–September 9, 1977.
Lillesand, T.M., R.W. Kiefer, and J.W. Chipman. 2004. Remote Sensing and Image Interpretation, 5th ed. New York: John Wiley and Sons.
McCauley, J.F., C.S. Breed, and G.G. Schaber. 1986. The megageomorphology of the radar rivers of the eastern Sahara. Pp. 25–35 in Publication 86-26. Pasadena, Calif.: Jet Propulsion Laboratory.
NRC (National Research Council). 1996. Rock Fractures and Fluid Flow. Washington, D.C.: National Academy Press.
Pratt, D., and C.D. Ellyett. 1979. The thermal inertia approach to mapping of soil moisture and geology. Remote Sensing of Environment 8(2): 151–168.
Robinson, C.A., F. El-Baz, M. Ozdogan, M. Ledwith, D. Blanco, S. Oakley, and J. Inzana. 2000. Use of radar data to delineate palaeodrainage flow directions in the Selima Sand Sheet, Eastern Sahara. Photogrammetric Engineering and Remote Sensing 66(6): 745–753.
Singh, A. 1989. Digital change detection techniques using remotely-sensed data. International Journal of Remote Sensing 10(6): 989–1003.
Szabo, B.J., C.V. Haynes, and T. Maxwell. 1995. Ages of Quaternary pluvial episodes determined by uranium-series and radiocarbon dating of lacustrine deposits of Eastern Sahara. Paleogeography, Paleoclimatology, Paleontology 113(2–4): 227–242.
UNEP (United Nations Environment Program). 2008. Sudan Post-Conflict Environmental Assessment–2008. Nairobi, Kenya: United Nations Environment Programme.

About the Author:Farouk El-Baz, research professor and director of the Center for Remote Sensing at Boston University, is a geologist, NAE member, and veteran of NASA's Apollo Program.