The Nile Valley is the lifeline between two of the world's driest deserts; it is maintained by an exotic river that derives its waters from tropical East Africa. The Nile floodplain provides a long, narrow oasis, and the ancient Egyptians were vividly aware of the stark contrast between the “Red Land” of the desert uplands and the “Black Land” of its seasonally inundated alluvium. In terms of aridity and lack of vegetation, the Egyptian deserts rival Death Valley in the United States; only the Atacama Desert of Chile is drier. Rainfall in an “average” year decreases rapidly from 200 millimeters (8 inches) on the northwest coast of the Nile Delta to 30 millimeters (1 inch) near Cairo; most of the Western Desert (Libyan Desert) statistically receives less than 5 millimeters (0.2 inch) equivalent to a modest shower every few generations. The Eastern Desert or more mountainous Red Sea Hills may receive 50 millimeters (2 inches) in places, but even when collected in wadis or in larger, dry river courses, flood events occur there only once or twice a century.

The potential concentration of occasional rainwater in wadis, in local depressions, or in shallow aquifers is critical to the regional distribution of vegetation and game. In the north, the desert shrub cover thickens from the Faiyum Depression toward the shores of the Mediterranean Sea. To the south, the Libyan Desert is essentially lifeless, with the exception of some scattered, spring-fed oases. Occasional rains are a little more frequent in the highland areas, making possible the ephemeral blooming of a sparse cover of grasses. In the Red Sea Hills, there are scattered thorn trees, with some shrubs and grasses along many of the wadis; otherwise there is barren desert. As a consequence, very small groups of pastoralists move through the wadis of the Eastern Desert, while for millennia sedentary populations have persisted in several oases of the Western Desert. Yet, until at least the 1920s, southern Sahara pastoral families and their cattle sporadically visited the Gilf Kebir highlands of southwest Egypt.

Although the eastern Sahara is parched Red Land, small numbers of resourceful people and their hardy animals have managed to use its occasional opportunities to advantage—complementing the caravans of traders and camels that once traversed this forbidding desert, moving from one watering place to another. The Sahara was and is no more a barrier than the oceans, one that could and can be crossed by experienced pastoralists and merchants. The Sahara became a wasteland only in the last three or four millennia, since in late prehistoric times, the environmental conditions were modestly wetter; for those skilled in surviving the desert, the Sahara was then a world of opportunity. Consequently, the prehistory of the Nile Valley was actively interlinked with that of the eastern Sahara.

The first clue that the prehistoric Red Land had not always been barren was provided by the rock engravings and paintings found deep within the desert; they first came to public attention during the 1920s and 1930s, depicting game animals now found in the Sahel, as well as livestock and peoples that were ethnographically distinctive. Systematic archaeological work began in the 1970s, and it has enhanced these once ambiguous impressions. A substantial and complex archaeological record is now linked with a robust body of geo-archaeological data that documents the environmental change. Precisely because desert resources are and were localized and point-specific, the overall picture is composed of a collage of microstudies—all of which differ in detail.

Today, occasional summer rains of southerly, monsoonal origin rarely stray north of the Tropic of Cancer; the winter showers, of northerly origin in the westerlies, very seldom penetrate south of the Tropic of Cancer. In areas of mountainous topography, however, occasional rains of both kinds may overlap. Whether water is collected and stored in a deep or a shallow aquifer is important, as is whether surface runoff is a major component. Such factors will affect the threshold conditions when water becomes available at or near the surface, as well as the persistance of that water for months, years, or even centuries. Similarly, the presence of animals and fish depends on a cyclic evolution, from smaller to larger life forms, at the point when water first becomes available until it finally gives out. Points of water surrounded by a complex biota will support more people than those with limited biotic diversity; whether an optimal association will develop, with the capacity to support repeated and protracted human settlement, will depend on the constellation of environmental variables and the predictability and productivity of resources.

In sum, the Egyptian deserts are very different, discontinuous environmental mosaics from the dependable, riverine oasis that forms the Nile corridor. At times, overall, some degree of co-variance occurred in environmental trends, but not sufficiently to allow normative inferences. Equally important to the region is that the productivity of the desert oases was never great—certainly not when compared with the Nile floodplain—and that their resources were ecologically fragile. Yet the Red Land was never a void, and the emergence of Egyptian civilization in the Nile Valley was based on both the human experience and the cultural roots of the diverse prehistoric adaptations to the desert.

Floodplain Margins.

Representations of animals and, more occasionally, vegetation are fairly common until Middle Kingdom times; these include tomb paintings and carved tomb reliefs. Tombs were usually located on the desert edge and rock engravings were placed on cliffs bordering the Nile Valley, mainly in southern Egypt. There are also a variety of art pieces with such portrayals—slate palettes, ivory carvings, and the decorations on pottery, primarily those from late Predynastic and Early Dynastic times. Such representations are to some degree equivocal as to their ecological interpretations. Do they provide an authentic record of locally familiar biota or are they mainly symbolic? Do they attempt to represent “nature” in such areas, or are they an elite contrivance similar to the hunting enclosures stocked with game captured elsewhere? Do they reflect desert environments rather than the riverine oasis watered by the Nile? No categorical answer is possible, but independent evidence as well as a number of ecological arguments suggest that such representations are informative.

The artistic record of the fourth millennium BCE suggests a fauna in and around the Nile floodplain that resembles the present-day dry savanna fauna of central Sudan, as known during the 1800s. There are numerous representations of such floodplain-dependent life forms as the elephant, giraffe, and hartebeest, with less frequent wild cattle, cheetah or leopard, rhinoceros, and possibly fallow deer. Semidesert “runners,” such as the oryx and gazelle, also are common, as are the “climbers” of dry rocky environments, the ibex and Barbary sheep. Desertedge species include the lion, jackal, hyena, and ostrich. During Old Kingdom times, the large savanna forms were no longer shown, while lions and Barbary sheep became rare. The main animals shown were the large and small antilopines—oryx, gazelle, addax, and hartebeest (in that order of frequency)—while ibex representations remained fairly common. For the Middle and New Kingdoms, there was a further shift to showing desert-adapted forms—with gazelle the most common, oryx and ibex declining, hartebeest increasing, and addax no longer shown.

From the animal portrayals, then, a progressive aridification of the environment beyond the floodplain is suggested, in conjunction with partial or complete elimination of small populations of the larger animals—the elephant, giraffe, and lion—by hunting. That this array of game and predators was once present in the region is plausible from the consistent depiction of diagnostic features in the representations. Further, elephant bone has been recovered from the Faiyum Neolithic, and leopard skeletons have been found in a prehistoric cave in the Red Sea Hills. Addax and oryx were still hunted in the coastal steppes of Egypt in the 1890s. Barbary sheep only became extinct in the Eastern Desert in historical times; and ibex is still present there but very rare. Hartebeest and gazelle were standard forms in Paleolithic times, verified in the Faiyum Neolithic and at el-Omari, with gazelle still present in the coastal areas and in Sudan. The high probability is that the animals shown were directly familiar to the artists and, in fact, all had established names in the Old or Middle Egyptian language.

The question of the local versus the regional presence of the animals depicted is more difficult to answer. Egypt's trade and diplomatic contacts with the Sudan increased during the Old Kingdom and, more important, by the fifth dynasty the animals are shown being hunted within fenced enclosures. Therefore, the animals shown on Old Kingdom reliefs would probably not have been familiar to the average floodplain farmer. By that time, the animals were probably trapped in the Red Sea Hills or on the coastal steppes, at some distance from the Nile Valley, and herded to elite private game parks. Since that is consistent with the ecological adaptations of the animals depicted, it implies that the simplification of the recorded faunal assemblage was due more to hunting pressures than to progressive aridification.

The relatively sparse representations of desert vegetation nonetheless argue for a parallel deterioration of desert productivity. The fifth dynasty sun temple of Newoserre Any at Abusir attempted to show the course of the seasons for the floodplain and for the adjacent desert, depicting a range of wild animals giving birth—gazelle, addax, oryx, wild cattle, ostrich, and cheetah. The animal scenes have a gently undulating surface, stippled to suggest sand, which supports an elaborate flora that, significantly, is labeled “plants of the Western Desert.” The tree types, as conventionally drawn in ancient Egypt, include acacia and sycomore fig. Although that fig is a floodplain genus, demanding considerable water, thick roots of both types of tree were found in a wadi fill under a twelfth dynasty building at Armant; found in 1.6 meters (4.5 feet) of fill, resting on Badarian potsherds, the trees grew before the Predynastic fill was cut into them. The lower vegetation tier shows a variety of shrubs, some suggesting succulents, as well as distinctive bunchgrasses. The last probably represent halfa grass, today common on desertedge sand surfaces, but halfa was also recorded at Neolithic el-Omari, together with acacia, tamarisk, and several chenopods.

The Newoserre Any reliefs, confirmed by other evidence, provide an authentic representation of the edge of the Western Desert as explicitly perceived by Egyptians c.2500 BCE. Other reliefs in fifth and twelfth dynasty private tombs show animal enclosures on a similar rolling sandy surface, highlighted by red stipples, although Middle Kingdom counterparts are devoid of vegetation. This would argue that the small wadis of the Western Desert margin once had scattered trees, with a groundcover of halfa grass and semidesert forbs and shrubs, much of which is vestigially preserved in the larger wadis of the Eastern Desert even today. What modest vegetation existed had apparently disappeared by the Middle Kingdom, much like the bulk of the desert fauna. Yet a tree root dated c.1150 BCE, under drift sand near the Neolithic site at Merimde, suggests that degradation of the Western Desert edge to its current barren condition was not abrupt. Yet the evidence for floodplain fodder plants that were cut for the domesticated livestock of the desert-edge Predynastic Naqada peoples cautions against the assumption that the desert edge provided substantial grazing resources.

Climatic Fluctuations.

The geological record confirms some minor climatic fluctuations on the margins of the Nile Valley. In the South, in Egyptian Nubia, mediumsized eastern-bank wadis were periodically active, with repeated flood events from about 10,000 to 7800 BCE, when a conspicuous fossil red soil developed; such wadi activity took place when the floodplain was dry. Since the Nile flood rose in midsummer, and the river in Nubia receded into its channel by mid-October, the local desert rains came during the winter half of the year and were not monsoonal. Subsequent soil formation, with abundant snail shells, implies a protracted period of weathering with fairly frequent gentle rains. Some large wadis debouch on the edge of the floodplain east of Kom Ombo. They were quite active during that time span, and the sandy composition and bedding indicate repeated strong floods. Abundant vegetation is indicated by proliferations of snail shells and common calcareous root casts (vertical and horizontal) that record some shrub and tree vegetation. These thick accumulations were terminated by the same red soil that developed in Nubia. A later episode of wadi activation records strong but sporadic periods of rain, which eroded slope materials, but with little organic evidence. By 3100 BCE, this had ended, prior to occupation of late prehistoric surface sites.

Nonetheless, beyond the floodplain margin, opposite Qena, tree roots of tamarisk and acacia on the desert are extensive in a level with six dates of 5570 to 3650 BCE. Well into the desert, north of Qena, sites within a colluvium dated 5280 to 3970 BCE have charcoal of acacia and two other thorny, tropical trees. Earlier deposits near Qena began to accumulate before 8700 BCE. A last pulse of activity in the great Wadi Qena deposited clays on the channel floor in Islamic times (eleventh century CE)—an episode that was recorded in various smaller eastern-bank wadis, by more modest sandy fills, in part spreading out over Roman ruins.

From the margins of the Holocene lakes of the Faiyum there is fragmentary evidence of what appears to have been a complex sequence of modest, local climatic changes. Between 4000 and 1700 BCE, there were various episodes of colluvial deposition, minor gullying, and the brief development of semidesert soils. Earlier, about 5900 to 5700 BCE, there was a stronger pulse of stream activity. Possibly, some sort of groundcover inhibited dune movement for much, but not all of the time, from before 7000 to 1700 BCE. The Faiyum evidence indicates the possible amplitude of these late prehistoric to early historical climatic fluctuations and cautions against assumptions of extended duration.

A further perspective is provided by the Tree Shelter and Sodmein Cave, both near Quseir, in an area decidedly affected by winter rains today. Adjacent torrential alluviation before 7100 BCE points to sporadic, very heavy rains. Then, until 3600 BCE, stream deposition was less torrential, suggesting more frequent but less intense rains, apparently closing with humic soil formation. Thereafter, hyperarid conditions like those of today prevailed, and there was no more mobilization of rock on the valley slopes. Alluviation was unusually rapid about 7000 BCE and again about 5950 to 5250 BCE. Cave deposits of the latter age include at least six genera of trees, with acacia, tamarisk, and wild olive among them.

The sum of this geological and botanical evidence implies that from somewhat before 7000 to about 3500 BCE, the desert margins of the Nile Valley and the Red Sea Hills were at times wetter than today, with some tree growth typical along the wadi floors. Since then, the adjacent deserts have been hyperarid, but with brief episodes of higher rainfall continuing until perhaps 1700 BCE. The thinning of the wadi tree stands and the reduction of biotic diversity has therefore not been entirely a result of human impact. The fifth dynasty desert reliefs are substantially younger than the termination of the modest “wet phase” of mid-Predynastic times, however, which suggests that the former vegetation of the desert margins was also degraded by human use, just as the riverine and desert fauna was progressively eliminated by hunting during dynastic times.

Late Prehistory and the Libyan Desert.

The wadis that occasionally spilled out onto the edge of the flood-plain emphasize the effect of concentrating a little water in and along a stream channel—most important where igneous or metamorphic rocks of low porosity act almost like an impervious cover. Even if floods only wash down such a channel once in fifty years or more, water collects and is stored in the gravelly sands below the stream bed, accessible to the deep-rooted trees and shrubs adapted to an arid environment. This water concentration effect is magnified in the Libyan Desert, where there are no linear channels, and water instead collects in central points or in deep and extensive aquifers, especially in porous sandstones and at the base of dune sands. In a dry climate, aquifers provide an additional advantage—that they collect water during wet intervals to store for millennia—later emerging in wells or springs.

The ecological opportunities provided by the Libyan Desert for plants, animals, and people during late prehistoric times were made possible by only modest climatic changes. The Libyan Desert has many shallow depressions—essentially hollows in a fairly flat but undulating surface—that are almost undiscernible to the human eye. Created over long time spans by wind erosion, others formed between dunes or where a dry valley was blocked by an encroaching dune. Surface runoff from a single heavy rain on a rock surface can form a sheet of standing water that may persist for months. If the frequency of good rains increases to several times a decade, clays and silts accumulate in the depression and seal the floor against water percolation into the soil, so that waters remain for years, with moisture retention in the subsoil allowing colonization by reeds. If persistent enough, such a mud pan or ephemeral playa lake may acquire a ring of shrubs and trees, which attract insects, waterfowl, animals, herders, and their flocks.

Sandstones or thick sand covers absorb and retain groundwater, and recurrent heavy rains every few years can recharge small or large aquifers to raise the water table by tens of meters, until a deep and persistent lake may form. Such groundwater may initially be brackish or saline from accumulated minerals, but as the volume of water in the aquifer expands, the calcium, magnesium, and sodium salts are diluted and the lake may turn fresh. Microorganisms and mollusca will be introduced by visiting birds that carry traces of mud and leave wastes with spores, seeds, and more. A diversified biotic community evolves in and around such a micro-oasis, with a ring of aquatic plants screening out blowing sand but trapping wind-borne dust. As the freshwater lake first deepens and then eventually shrinks, its water chemistry is transformed, becoming a salt lake and then a dry, salt flat. The Selima Oasis in northwestern Sudan is a prime example of such a groundwater lake. The aquifer was recharged before 9000 BCE, and by 8200 BCE, a shallow brackish water body was transformed into a deeper freshwater lake. It was a modest salt lake from 2500 to 2000 BCE, and it finally remained a salt flat, although residual savanna trees survived in the area for another millennium.

From a dozen or so key locations of the Libyan Desert broad parallels become apparent in the environmental response to climatic change, with first a weak pulse of wetter conditions (c.8000–7600 BCE), followed by three stronger peaks (7300–6300, 5800–5200, and 4500–3400 BCE). This composite picture, which parallels that of the Nile Valley margins and the Red Sea Hills, nonetheless masks a lack of synchroneity in detail between centers. At Bir Kiseiba, improved water conditions are apparent two thousand years before the Kharga Oasis. Optimal moisture conditions at Selima, Nabta Playa, Kiseiba, Kharga, and the Tibesti Mountains of Libya were dated to about 7000 BCE, while those of the Gilf Kebir, the Great Sand Sea, and the Dahkla Oasis were delayed until about 4500 BCE. Nabta Playa was last abandoned in 4700 BCE, twenty-five hundred years before the Gilf Kebir. Modest later improvements of moisture, evidenced by tamarisk trees fixing dune sands, are apparent on the northern foothills of the Tibesti (1600–350 BCE and 90–640 CE) and in the Siwa Oasis (1210–1110 BCE and 65 BCE–560 CE). There are similar shifts in the time spans of the largest array of Sudanese tree types and game animals.

These discordances probably reflect the different thresholds at which surface waters became available or reliable through geohydrological conditions, as well as a complex interplay of augmented monsoonal and westerly rains. Groundwaters can be dated by the radiocarbon dating technique. Histograms of dated fossil water for the region show nonsynchronic patterns of short-term recharge in the northerly and southerly aquifers. In fact, the viable oases of the Libyan Desert today draw on fossil waters.

Until about 5000 BCE, the mobile peoples of the eastern Sahara were hunter-gatherers (despite the mainly unaccepted claims for domesticated cattle at Nabta Playa two millennia earlier, but sheep or goats may have been present a little earlier in places). Both the stone tools and the economy that are sometimes labeled Early Neolithic in the Libyan Desert were Epi-Paleolithic. A pastoral Neolithic, with some bifacial tools, was only established near Kharga and Dakhla Oases about 5500 BCE. These are the two oases that remained important throughout historical times, and the ones that show the closest archaeological affinities with the emerging Neolithic of the Nile Valley. Elsewhere, the northern Libyan Desert was only lightly and sporadically utilized by small groups after 3500 BCE, and the Libyan invaders of Old Kingdom Egypt probably came from the coastal semidesert.

The late prehistory and changing environments of the Libyan Desert remain of paramount interest for emerging desert adaptations in Africa, as well as for the incorporation of livestock into nomadic economies. Still, the sum of all the late prehistoric oases will have supported hundreds, not thousands, of people at the best of times—and they never formed a demographic reservoir that later drained into the Nile Valley. Occupation was seasonal or episodic, and whatever the patterns of periodic movements to better watered areas, they probably ran north and south, from and to better watered areas in the Sahel.



  • Banks, Kimball M. Climates, Cultures and Cattle: The Holocene Archaeology of the Eastern Sahara. Dallas, 1984. A good presentation, with primary data on Nabta Playa.
  • Butzer, Karl W. “Wüste, Wüstentiere” (Desert and Desert Animals, English text). In Lexikon der Ägyptologie, 4: 1291–1297. Wiesbaden, 1986.
  • Butzer, Karl W. “Late Quaternary Problems of the Egyptian Nile: Stratigraphy, Environments, Prehistory.” Paleorient 23.2 (1998), 151–173.
  • Friedman, Renée, and Barbara Adams, eds. The Followers of Horus. Oxford, 1992. A diverse and useful collection, with several papers on rock art.
  • Haynes, C. Vance, C. H. Eyles, L. A. Pavlish, J. C. Ritchie, and M. Rybak. “Holocene Palaeoecology of the Eastern Sahara: Selima Oasis.” Quaternary Science Reviews 8 (1989), 109–136.
  • Keding, Birgit. “Prehistoric Investigations in the Wadi Howar Region.” Kush 17 (1997), 37–46.
  • Kröpelin, Stefan. “Palaeoclimatic Evidence from Early to Mid-Holocene Playas in the Gilf Kebir (Southwest Egypt).” Palaeoecology of Africa 18 (1987), 189–208.
  • Kuper, Rudolph, ed. Forschungen zur Umweltgeschichte der Ostsahara. Acta Praehistorica, 2. Cologne, 1989. A major compendium of recent research, mainly in German.
  • Moeyersons, Jan, P. M. Vermeersch, H. Beekman, and P. Van Peer. “Holocene Environmental Changes in the Gebel Umm Hamad, Eastern Desert, Egypt.” Geomorphology 26 (1999), 297–312.
  • Neumann, Katharina. “Holocene Vegetation of the Eastern Sahara.” African Archaeological Review 7 (1989), 97–116.
  • Pachur, Hans-Joachim and S. Kröpelin. “Wadi Howar: Paleoclimatic Evidence from an Extinct River System in the Southeastern Sahara.” Science 237 (1987), 298–300.
  • Pachur, H. J., H. P. Roper, S. Kröpelin, and M. Goschin. “Late Quaternary Hydrography of the Eastern Sahara.” Berliner Geowissenschaftliche Abhandlungen (A) 75.2 (1987), 331–384.

Karl W. Butzer