A roughly 1,500-km survey found ancient lake deposits but no visible occupation evidence, leading the authors to argue that freshwater alone did not sustain settlement.
Researchers surveying about 1,500 km through the Rub’ al Khali desert in the southern Arabian Peninsula found ancient lake deposits but no stone artefacts or other cultural material at any of 125 localities. In a study published in Nature on 28 October 2026, the authors argue that access to both freshwater and stone suitable for tools constrained settlement—not water alone.
Earlier models of human dispersal used reconstructed lakes, rivers and climate to identify the desert interior as a likely place for occupation. Periodic humid phases had created lakes, wetlands and grasslands, and dated archaeological sites around the desert’s margins supported that expectation. The authors say the predictions had not previously been tested by systematic ground survey in the interior or against the distribution of toolmaking stone.
| Evidence | Present | Absent |
|---|---|---|
| Ancient lake deposits | 55 | 70 |
| Stone suitable for toolmaking | 10, all at the basin margins | 115 |
Lake sediments and freshwater mollusc shells supported the identification of former lakes, or palaeolakes. The authors interpret these waters as fresh when lake levels were high and increasingly saline as they dried. Even the ten marginal localities with toolmaking stone yielded no artefacts.
The survey targeted deposits correlated with lake sections independently dated in earlier work to about 122,000–67,000 years ago. Additional deposits at equivalent elevations were interpreted as remnants of the same lake systems. Those dates therefore provide a regional framework, not a new, direct age for every surveyed locality. Previously dated occupation at nearby Mundafan, about 96,000–77,000 years ago, suggests that people lived around the desert margins during part of that interval.
Analysis
The study focuses on the Middle Palaeolithic and Middle Stone Age, broadly contemporary toolmaking traditions in western Eurasia and Africa. The authors use the joint label MP/MSA because Arabian tools show affinities with both. These traditions commonly involved preparing a stone core before removing flakes.
For these toolmakers, knapping—striking suitable stone to remove flakes—required rock that fractures predictably, producing sharp edges. Not every rock has the necessary curved, shell-like fracture, known as conchoidal fracture. Making and resharpening tools also leaves durable stone debris, which archaeologists can recognize.
The nearest known suitable outcrops lie more than 200 km from the surveyed interior basins. Across roughly 800 km of the central basin, the team found no suitable surface stone. The authors interpret the lakes as groundwater-fed because their fine sediments lack coarse material brought by runoff. Unlike rivers, groundwater would not have carried toolmaking stone into the basins from distant outcrops.
Analysis
The researchers built a spatial model using mapped surface geology and reconstructed drainage, then tested it against 301 MP/MSA findspots in Oman. Sites clustered where water and suitable stone occurred together. A wider compilation of 363 Arabian sites showed the same pattern; 362 fell within the area covered by the statistical model.
Compared with 999 simulations placing sites randomly, documented sites were unusually concentrated in the highest-potential zones and unusually scarce where neither resource occurred. Both results fell outside the simulations’ 95% range. The model also classified 57% of the area within its water-access thresholds as lacking suitable stone. These associations support the authors’ explanation, rather than directly proving why people chose particular places.
The authors draw two useful comparisons. At Jubbah in northern Saudi Arabia, an isolated ancient lake had locally abundant quartzite and evidence of Middle Palaeolithic shoreline occupation. Around Duqm in central Oman, chert is abundant but MP/MSA sites are rare—a pattern attributed to scarce freshwater. Together, these examples support the argument that neither resource alone was sufficient.
The authors report ground visibility above 95% on wind-eroded surfaces between dunes. Finding centimetre-scale shells and eggshell fragments, they argue, suggests exposed stone artefacts would also have been conspicuous. That visibility applies to those exposed surfaces: dune sand covers underlying rock, and the observations do not establish that nobody ever crossed or briefly visited the basins.
The wider model has another limit. Its geological map, at a scale of 1:2,000,000, can miss small occurrences of suitable stone. The authors use that limitation to explain some documented sites classified as having low settlement potential.
Future fieldwork could still uncover small sites used during short trips or resource gathering, the authors acknowledge. Against the absence of occupation evidence at all 55 surveyed palaeolake localities, they would interpret such discoveries as low-intensity activity rather than sustained settlement. Their conclusion concerns where people left recognizable evidence of occupation, not every route people may have travelled.
The NIH–NASA competition seeks better measurements in laboratory drug-safety models, with submissions scheduled to open on 15 December 2026.
2 min
The programme brings together national laboratories, universities and industry partners, with plans for a sensor-equipped test bed and self-adjusting pilot production lines.
2 min
Previously seen only through its infrared heat, the dusty feature is visible in reflected starlight, allowing researchers to test a possible collision origin.
4 min
