ArticlePDF Available


Content may be subject to copyright.
R. H. Inglis1,2, A. G. Sinclair3, A. M. Alsharekh4, D. N. Barfod5, H. C. Chang2,
P. C. Fanning2 D. T. Al Othaibi6, H. K. Robson1, A. Shuttleworth7, A. Stone8,
and G. N. Bailey1
1 Department of Archaeology, University of York, York, UK
2 Department of Environmental Sciences, Macquarie University, Sydney, Australia
3 Department of Archaeology, Classics and Egyptology, University of Liverpool, UK
4 Department of Archaeology, King Saud University, Riyadh, Saudi Arabia
5 Scottish Universities Environmental Research Centre, East Kilbride, Glasgow, UK.
6 Saudi Commission for Tourism and National Heritage, Riyadh, KSA.
7 Department of Anthropology, Durham University, UK.
8 Department of Geography, University of Manchester, UK.
January/February 2017 Fieldwork at Wadi Dabsa
1. Wadi Dabsa and the Palaeolithic of SW Saudi Arabia
1.1 Introduction
1.2 Geology and Archaeology of the Harrat Al Birk
1.3 Wadi Dabsa
1.4 Aims and Objectives
2. Methods
2.1 Regional Geomorphological and Surface Condition Mapping
2.2 Local Landform and Surface Condition Mapping
2.3 Artefact Collection
2.4 Archaeological Excavation
2.5 Lithic Artefact Analysis
2.6 Chronometric Sampling
2.6.1 Basalt Sampling
2.6.2 Tufa Sampling
3. Results
3.1 Landscape Evolution
3.1.1 Remote Sensing Observations
3.1.2 Field Observations
3.2 Basalt Morphology in the Wadi Dabsa Basin
3.3 Tufa Development in the Wadi Dabsa Basin
3.4 Local Landform Mapping and Surface Condition Recording
3.5 Spatial Distribution of Surface Artefacts
3.6 Excavation and Artefact Condition
3.7 Lithic Analysis
3.7.1 ESA/Acheulean Elements at L0106
3.7.2 MSA/Prepared Core Technology at L0106
3.7.3 Potential for Further Analysis
4. Discussion
4.1. Evolution of the Wadi Dabsa basin
4.2 Geomorphological Context of the L0106 & L0130 Artefacts
4.3 The Wadi Dabsa Lithic Assemblage in Techno-Typological Context.
5. Summary and Conclusions
6. Official Meetings
7. Acknowledgements
8. References
Appendix 1: 2017 Field Team
Appendix 2: List of Basalt Samples Collected in 2017
Appendix 3: List of Tufa Samples Collected in 2017
January/February 2017 Fieldwork at Wadi Dabsa
1. Wadi Dabsa and the Palaeolithic of SW Saudi Arabia
1.1 Introduction
Understanding the ways in which hominin populations used the landscapes they inhabited is
crucial to understanding their spread across the globe. At the crossroads between Africa and
Eurasia, Arabia’s Palaeolithic record is central to these debates (Petraglia, 2003; Petraglia &
Alsharekh, 2003; Groucutt et al., 2015). Spanning potentially over 1 million years, this
record holds great potential for interpreting patterns of hominin landscape use and the role
this played in the ways hominins dispersed. Yet the interpretation of artefact distributions
within a landscape requires a thorough understanding of the geomorphological processes that
control artefact preservation and visibility, and shape landscapes over time (Holdaway &
Fanning, 2014). Combined geomorphological and archaeological methodologies for
recording, analysing, and interpreting the archaeological record in its landscape context must
therefore be applied to the Arabian Palaeolithic record to fully realise its potential to inform
on early hominin landscape use.
Figure 1: Location of the 2017 study area and Harrat Al Birk, southwestern Saudi Arabia.
January/February 2017 Fieldwork at Wadi Dabsa
The Red Sea coast of southwestern Arabia has a rich, but not comprehensively understood,
record of Early and Middle Stone Age (ESA and MSA) archaeology. Remaining relatively
humid during the Pleistocene, it may have been one of the Peninsula's most persistently
attractive regions (Bailey, 2009; Bailey et al., 2015). It was easily accessible to populations
migrating across both the Southern Route (Bab el Mandab/Hanish Sill) from East Africa, and
the Northern Route via the Nile/Sinai (Lambeck et al., 2011), making the region crucial to
understanding hominin dispersals. A newly-discovered major concentration of ESA/MSA
artefacts at Wadi Dabsa, in the Harrat Al Birk, Asir, southwestern Saudi Arabia (Foulds et
al., In Press; Inglis et al., 2015) therefore presents a rare and exciting opportunity to apply
interdisciplinary approaches to a significant lithic assemblage to understand the
environmental and behavioural context of the artefacts and implications for early hominin
dispersals out of Africa.
1.2 Geology and Archaeology of the Harrat Al Birk
The Harrat Al Birk, Asir province, consists of over 1800 km2 of lava flows and cinder cones
stretching between the modern-day coastline and the foothills of the Arabian escarpment
between approximately N18.76°–17.79° latitude. Remote sensing of the region shows flows
of varying apparent age and degree of erosional degradation. Numerous younger flows
emanating from dozens of extant cinder cones, overlie extensive dissected fields of older
flow deposits. Erupted products consist of flows and scoriaceous deposits, dominated by
alkali olivine basalts and hawaiites.
The age and duration of volcanism in the Al Birk field is poorly constrained. The timing of
activity derives exclusively from 17 K-Ar whole-rock dates ranging from 0.2 to 2.6 Ma
(Coleman et al., 1983; Dabbagh et al., 1984; Zarins et al., 1981). The most recently obtained
dates were measured on a single basalt flow 8 km south of the town of Al Birk, yielding K-Ar
whole rock ages of 1.37 ± 0.02 and 1.25 ± 0.02 Ma (Bailey et al., 2007). Although
representing a considerable analytical effort, the inherent limitations of whole-rock K-Ar
ages and the lack of concordant results for some replicated analyses, call into question the
accuracy and precision of the existing temporal framework and highlight the need for an
extensive geochronological study using modern 40Ar/39Ar methods. Nevertheless, post-rifting
volcanism in the region appears to be dominated by Pleistocene or younger ages, generally
<1.5 Ma, postdating the first known dispersals of Homo out of Africa and indicating that
volcanic eruptions may have altered a landscape already inhabited by hominins.
January/February 2017 Fieldwork at Wadi Dabsa
The driving factors that generated the Al Birk volcanism are enigmatic and may not directly
relate to rifting of the Red Sea. Lying to the northeast of the Al Birk field, regionally
extensive Oligocene to Miocene (ca. 20–25 Ma, Bosworth & Stockli, 2016) mafic dikes trend
parallel to the Red Sea coast, marking the opening of the Red Sea rift, significantly predating
Al Birk volcanism. Coleman et al. (1983) note that the overall trend of eruptive centres
shows a NNE alignment, distinct from the NNW trend of the Red Sea margin and spreading
centre. Since the beginning of the Pleistocene, the Harrat al Birk was more than 100 km
distant from the Red Sea axial valley, suggesting that the origin of the Al Birk volcanism is
unrelated to seafloor spreading and more likely akin to the intraplate volcanism seen in the
other harrats of western Arabia (Bosworth & Stockli, 2016).
The Harrat al Birk is drained by a series of wadis that flow west to the Red Sea. Some, such
as Wadi Najla, have become deeply incised, exposing the bedrock of shales and schists
underlying the basalt flows, while others, such as Wadi Dabsa, have undergone less incision.
This variation is probably related to the age of the basalt flows that shape each area, as well
as variations in catchment size. In many of these wadis it is clear that past lava flow
emplacement has altered wadi courses, as well as filled pre-existing valleys (Inglis et al.,
Our understanding of the archaeological record of the Harrat al Birk is emerging. The
Comprehensive Archaeological Survey Project (CASP) in the 1970s and 1980s identified
numerous locations where Palaeolithic stone tools were preserved (Zarins et al., 1980, 1981).
Following preliminary work in 2004 and 2006 (Bailey et al., 2007), a series of surveys from
2012–2015 by the UK-Saudi team, under the umbrella of the DISPERSE project, expanded
this understanding of the distribution of Palaeolithic artefacts across the harrat (Bailey et al.,
2012, 2015; Foulds et al., In Press; Inglis et al., 2013, 2014a, 2014b, 2015). During these
surveys, discovery of a large number of Palaeolithic artefacts at Wadi Dabsa provided the
largest assemblage of ESA and MSA artefacts yet recorded from the Harrat al Birk, and
indeed the wider southwestern Saudi Arabia region. Most critically, it was observed during
these surveys that the rapid pace of development in the Harrat Al Birk region threatens its
abundant archaeological record, rendering the survey and recording of the artefacts
particularly timely.
January/February 2017 Fieldwork at Wadi Dabsa
1.3 Wadi Dabsa
In 2015, numerous ESA/MSA artefacts were recorded in the headwaters of Wadi Dabsa,
located on the surface of, and apparently embedded within, a 1 km2 area of tufa carbonate
deposits (Inglis et al., 2015). More than 900 Palaeolithic basalt artefacts were recovered from
a 40 x 50 m grid, with the deposit/artefacts extending for at least this same area again. Initial
analysis indicated that the assemblage had potential for refitting flakes to their parent cores
and identifying patterns of learning in stone tool manufacture. Formed under past wetter
conditions, the tufa potentially contained a dateable isotopic record of past environments. The
association of a major artefact assemblage with a palaeoenvironmental archive thus provided,
potentially, a unique opportunity to examine Palaeolithic activity in its environmental
1.4 Aims and Objectives
The 2017 UK-Saudi fieldwork aimed to examine the behavioural and environmental context
of the Palaeolithic artefacts at Wadi Dabsa, Asir Province, to further understand early
hominin landscape use in a region key to global dispersals.
1. Use remote sensing and satellite data to map geomorphological units surrounding the Wadi
Dabsa site to target archaeological and geomorphological survey and excavation.
2. Conduct geomorphological survey of the locality and its environs to provide a robust
model of landscape evolution within which to situate the archaeological
3. Record the distribution of, and collect, surface artefacts.
4. Analyse collected artefacts to illuminate hominin behaviour at the locality, and its
implications for the Arabian record.
5. Excavate the tufa deposits at targeted locations to confirm relationships between the
artefacts and tufa deposition.
6. Sample the tufa deposits for 234U–230Th dating and palaeoenvironmental analysis using
stable isotopes to provide both a chronological framework for Palaeolithic activity
and a palaeoenvironmental archive in this semi-arid region where preservation of
archives is rare.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 2: Overview of the Wadi Dabsa basin showing the 2017 area of investigation, and the
location of the 2015 artefact collection grid, L0106. Imagery © CNES/Airbus, Imagery date
15/11/2016, accessed through Google Earth.
The fieldwork took place over four weeks in January/February 2017 (three weeks of
survey/excavation and one of post-excavation analysis of the lithic artefacts). During this
time, the team achieved all of the above objectives, and produced a detailed record of a large
assemblage of ESA/MSA artefacts, further analysis of which will provide a rare coherent
insight into the technology and behaviour of early hominin populations in their landscape
context. Following analysis, the artefacts were packed for long-term storage in the Asir
Regional Museum, Abha. The geological samples were exported to the UK for specialist
January/February 2017 Fieldwork at Wadi Dabsa
Figure 3: (a) Overview of grid location in landscape, base map © CNES/Airbus, Imagery
date 15/11/2016, accessed through Google Earth; (b) schematic summary of date of survey,
collection methods, and quadrants recorded across grids.
January/February 2017 Fieldwork at Wadi Dabsa
2. Methods
A multi-scalar geoarchaeological approach to understanding the locality, its landscape and
the archaeology was undertaken, working in from the wider landscape to the area where the
highest number of basalt artefacts had been recorded in 2015. The artefacts in this area had
been sampled by collecting all lithic artefacts within a 40 x 50 m grid, (designated L0106,
Inglis et al., 2015). In 2017, this grid was extended to the northeast by 20 m, increasing the
area collected over the two seasons to 50 x 60 m (i.e., 3,000 m2). In addition, a second 50 x
60 m grid (designated L0130) was set up to the west of L0106, offset 5m to the west to avoid
a vehicle track that runs northeast/southwest through the concentration of artefacts (Figure 3).
As with to L0106, L0130 was divided into 10 x 10 m squares, each of which was further
divided into four 5 x 5 m quadrants (labelled w, x, y and z). The location of each quadrant
corner across the entirety of both the 2015 and 2017 grids was recorded in ArcPad 10 using a
Trimble Geo7X with Zephyr Model 2 external antenna, and its relative height recorded using
a dumpy level and staff to ensure maximum precision over the gently sloping area.
2.1 Regional Geomorphological and Surface Condition Mapping
Desktop analysis of satellite imagery was undertaken to map the landform units within the
Wadi Dabsa basin, to: (a) build a working landscape stratigraphy that, after groundtruthing,
would elucidate landscape development as well as guiding sampling for 40Ar/39Ar dating of
basalts and 234U–230Th dating of the tufa; and (b) assess these landforms in terms of their
potential for the exposure, preservation and visibility of Palaeolithic surface artefacts.
The units were defined by visual examination of Pléiades 1 Satellite panchromatic (50 cm
resolution) and 4-band (2 m resolution; blue, green, red, near infra-red) multi-spectral
images, dated 29/2/2016, in QGIS 2.0.1 Dufor, with reference to the ASTER Global Digital
Elevation Model (30 m resolution) as well as true-colour imagery (red, green, blue) accessed
through Google Earth (dated 19/01/2014). Units were defined by their assumed geological
and geomorphological origins and the processes that had shaped them, e.g. basalt flows,
basalt terraces, tufa, etc.
More detailed mapping of surface condition, influencing artefact visibility across the tufa
deposits (the focus of the 2017 fieldwork), was carried out using automatic classification of
Google Earth imagery. Unsupervised classification of the imagery into ten classes was
carried out using the ISOData method in ENVI 5.2 (output image cleaned by smoothing to 3
pixels, aggregation to 5 pixels) The resulting classification of the tufa surface condition was
January/February 2017 Fieldwork at Wadi Dabsa
groundtruthed through visits to the various mapped areas on the tufa surface to verify the
surface condition classes and their boundaries. This data will be used in future work to
further refine the surface classifications derived from satellite data.
2.2 Local Landform and Surface Condition Mapping
Landform units, defined as areas of relatively uniform surface morphology, were mapped
across the two grids by recording their boundaries as polylines in ArcPad10 using the
Trimble and external antennae. The polylines were later used to generate polygons of each
landform unit in QGIS 2.1 Las Palmas. Six landform units were identified and mapped:
Ridge, Crest, Upper Slope, Lower Slope, Drainage Depression, Disturbed.
Surface condition within each of the defined landform units is variable, and affects the
visibility of artefacts. Three surface condition types were recorded within each quadrant
across the grid where artefacts were collected or recorded: % of bare tufa, % sediment cover
and nature of sedimentation, and % vegetation cover.
2.3 Artefact Collection
After extending the L0106 grid from its 2015 boundaries, all surface lithic artefacts were
collected from the new area (total 20 x 50 m) in 5 x 5 m quadrants, following the
methodology employed in 2015. In L0130, given the already large sample of artefacts
collected from L0106 and due to time constraints, artefacts were recorded and photographed
but were (bar a single handaxe) left in situ. In L0130, only artefacts in the x and y quadrants
were recorded in alternating squares in a chequerboard pattern, apart from Rows 1 and 2
where, due to the low density of the artefacts, it was possible to record the artefacts within x
and y quadrants of all the squares (Figure 4).
2.4 Archaeological Excavation
In addition to the surface collections, two small excavations were carried out in L0106 to
investigate the depth and nature of the sediment overlying the tufa and relationships between
the surface artefacts, sediments and tufa. Two 100 x 50 cm pits were excavated in quadrants
5Ax and 6Az/By to examine the stratigraphy of the sediments in the ‘Crest’ landform unit,
reaching depths of up to ~30cm. The ‘Crest’ landform unit (Figure 11) is where sediment is
most abundant, and the relationship between the artefacts, the sediments and the tufa can be
most readily investigated.
January/February 2017 Fieldwork at Wadi Dabsa
2.5 Lithic Artefact Analysis
Collected lithic artefacts (from L0106) have been described according to a system of lithic
analysis developed for the examination of lithic artefacts as part of a previous landscape
survey in southern Africa (McNabb et al., 2003; Sinclair & McNabb, 2005). The use of a
combined typological and techno-typological recording system allows us to expand upon the
purely typological information collected previously by the CASP (Zarins et al., 1980, 1981).
By including technological information, we can begin to identify different aspects of
manufacturing activity happening at varied locations by following a chain of reduction from
nodule to final discarded fragment. In so doing, this information will also help us recognise
the processes of hominin artefact transport across the study area.
In typological terms, the system records the type of lithic blank (flake, blade, core, etc.) as
well as the specific type of tool if the piece has been modified by retouch. For the artefacts
collected and recorded from Wadi Dabsa these retouched types include types appropriate to
the ESA/Acheulean as well as the MSA/prepared-core periods and following standard
typologies (Bordes, 1979; Debenath & Dibble, 1994).
In technological terms, the recording system identifies the type of lithic material from which
the artefact has been made, the percentage of cortex present, the state of weathering and
roundedness, as well as the metric dimensions. For cores, the technological recording system
also identifies any separately identifiable episodes of debitage, and the form of working that
occurred. For flakes and other products of debitage, the recording system also records the
pattern of prior removals identifiable through the patterning of negative flake scars on the
dorsal surface, the form and location of any retouch, and any form of macroscopic edge
damage. For the analysis of prepared-core working, which comprises much of the assemblage
recovered from Wadi Dabsa, the system of technological recording also identifies whether
flakes derive from the preparation of the upper surface or the lower surface of prepared cores
following the scheme originally proposed by Boëda (1995) in which prepared core
technology proceeds through the preparation and maintenance of upper and lower ‘volumes’
above and below a central level within the core.
Finally, all collected artefacts greater than 20 mm in length have been digitally photographed
for both the dorsal and ventral surface against a 10 mm squared grid. This allows for the
same square grid to be superimposed as a new layer on top of the artefact to facilitate the
creation of line-drawn illustrations of selected lithic artefacts for publication at a later stage.
January/February 2017 Fieldwork at Wadi Dabsa
In the case of the artefacts observed but not collected in L0130, a brief typological and
technological description was assigned to artefacts in the field, and the ventral and dorsal
surfaces of each were photographed against a plain background with a standard photographic
scale prepared for the survey.
2.6 Chronometric Sampling
2.6.1 Basalt Sampling
The basalt sampling strategy aimed to acquire the freshest possible samples from all of the
morphologically distinct basalt units within and surrounding the basin and in the artefact
recording grids, in order to chronologically constrain the emplacement of the lava flows
through 40Ar/39Ar dating. Sample collection sites were chosen to ensure, to the greatest extent
possible, that the sampled rocks were in situ with respect to initial deposition as a lava flow.
Each sampled basalt consisted of ~ 1 kg of material extracted from the interiors of larger
boulders, typically 40–60 cm in diameter. At the end of fieldwork, the samples were
transported to the UK for analysis in the NERC Argon Isotope Facility at the Scottish
Universities Environmental Research Centre, East Kilbride.
2.6.2 Tufa Sampling
Samples of tufa were collected from locations across the basin, within wadi channels and
within the artefact recording grids, to provide an understanding of the timing (through
forthcoming 234U–230Th dating) of their deposition, related to past humidity. Stable isotopes
(δ18O and δ13C) from the same samples will be analysed to provide a palaeoenvironmental
reconstruction relating to source moisture and palaeotemperature. Samples were selected on
the basis of their appearance in hand section, targeting carbonate that was densely cemented
and looked clear of non-carbonate detritus, following the experience of Stone et al. (2010), in
order to minimise potential problems of detrital contamination and open-systems behaviour
for 234U–230Th dating. At the end of fieldwork, the samples were transported to the UK for
analysis in the NERC Stable Isotope Facility at the British Geological Survey, Keyworth, and
the Department of Geography, Royal Holloway University of London.
January/February 2017 Fieldwork at Wadi Dabsa
3. Results
3.1 Landscape Evolution
3.1.1 Remote Sensing Observations
From observation of the satellite imagery, it is clear that the landscape of the Wadi Dabsa
basin is dominated by basalt flows (Figure 4), laid down by successive eruptions from the
various cinder cones that dot the landscape. These flows and cinder cones exhibit the effects
of varying degrees of erosion, suggesting an extended history of magmatism, likely spanning
millions of years (Bosworth & Stockli, 2016; Coleman et al., 1983).
Figure 4: Landform mapping of Wadi Dabsa basin showing location of samples of basalt
(green) and tufa (orange) collected in 2017. Purple rectangle in centre of tufa unit locates
position of artefact recording grids.
Mapping of the flows from satellite imagery was straightforward, given the lack of vegetation
on the surface. Flows to the east of the study area were markedly more defined on the
satellite imagery than those to the west and southwest, possibly indicating a younger age for
these flows in comparison with those in the southwest. Superimposition of the flows based on
remote sensing observations allowed the development of a phasing matrix of flow
January/February 2017 Fieldwork at Wadi Dabsa
emplacement to be used to both understand the development of the basin as well as guide the
sampling (Figure 4).
One particular area of interest highlighted by the landform mapping was the relationship of
flow F4, one of the youngest inferred flows, and the tufa at the eastern edge of the basin. The
shape of this flow, especially the lobate form of its boundary with the tufa, suggests that it
may have flowed along a pre-existing drainage depression in the older basalt, and onto the
tufa. Such a disruption of drainage at the upstream end of the basin may have altered the
hydrological conditions in the basin, perhaps reducing water flow across the floor of the basin
and hence reducing the attractiveness of this landscape to hominins and their prey.
Groundtruthing of the relationship between the basalt and tufa at this location was therefore a
high priority (see below).
Unsupervised landscape classification of the Wadi Dabsa basin and its immediate
surroundings using Google Earth imagery (RBG) produced three land surface classes related
to the sedimentation conditions on the surface of the tufa, and seven classes corresponding to
differences in the basalt flow surfaces (Figure 5). From satellite observations, these three tufa
surface classifications were interpreted as corresponding to:
Red – bare surfaces composed of tufa;
Teal – bare tufa surfaces, but potentially more heavily weathered;
Bright Green – unconsolidated sediment, e.g. fluvial sediment along drainage channels;
aeolian sediment on crests and slopes.
Field observations of these three classes resulted in the following additional information
being recorded:
Red – tufa surfaces with very little or no sediment cover;
Teal – areas of tufa or sediment (but most commonly tufa) with a significant amount of basalt
clasts scattered across the surface;
Bright Green – unconsolidated sediment, ranging from fined grained aeolian deposits, sandy
to gravelly lag deposits overlying tufa, outcrops of grey silty material that has a buff-coloured
wind-blown dust veneer, and alluvial sediment in shallow drainage channels and depressions.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 5: Map of surface sediment classes identified through the unsupervised classification
of Google Earth imagery (Imagery © CNES/Airbus, Imagery date 19/01/2014, accessed
through Google Earth) and photographs showing variability in surface sediment at locations
used for groundtruthing.
The unsupervised classification of remotely sensed imagery therefore produced a largely
robust distinction between areas of sediment cover on the tufa and areas of bare tufa, as well
January/February 2017 Fieldwork at Wadi Dabsa
as identifying areas of tufa exposure partially covered by basalt clasts that had been
interpreted from the satellite imagery as solely heavily-weathered tufa. These broad-scale
distinctions were observed in particular on the landforms in the direct vicinity of grids L0106
and L0130, providing a robust method of medium-scale surface-cover mapping across the
tufa, which will, after refinement through the field observations, guide future surveys of the
tufa surface.
3.1.2 Field Observations
In the field, the lava flow surfaces were predominantly low-relief boulder fields. Basalt clasts
ranged from cm- to metre-scale, showed signs of aeolian abrasion and thermal expansion
weathering and were typically embedded silt and sand, deposited by aeolian or fluvial
processes. On the basin floor, where many of these flows have been engulfed by later water-
lain tufa deposits, only the uppermost surfaces were exposed as patches. The patches
contained basaltic clasts within a loose matrix of yellowish brown silts and fine sands, and
sparse coarser fragments (cm-scale) of siliciclastics (e.g. quartz granules) and tufa. The
majority of basaltic clasts were vesicle-poor, with rare vesiculated clasts interspersed.
Basaltic clasts of varying texture interspersed with other coarse grain clastic materials or
isolated basaltic clasts distributed across the tufa deposits are suggestive of relatively high-
energy sedimentation events affecting significant portions of the generally flat basin floor,
e.g. sheet floods or debris flows.
Figure 6: General views across (a) tufa surface with variable sediment cover, looking east
from the western edge of tufa exposure; and b) basalt fields with small sediment patches,
looking north-east from basalt fields at the western edge of the study area.
January/February 2017 Fieldwork at Wadi Dabsa
Channels within the central/eastern portions of the basin exposed a number of basaltic units,
clastic deposits and tufa structures. The basaltic units hinted at original, coherent lava flow
structures such as colonnades, but in general showed a similar amount of erosion as basaltic
material exposed on the basin floor. Further downstream, to the west of the main basin
(Figure 4), weathered basalt with well-developed columns (~ 1 m diameter) and overlain by
coarse clastics including metre-scale clasts of tufa and basalt, was exposed on the floor
of the wadi.
A full assessment of the stratigraphic relationship between the F4 basalt flow and the tufa
was hampered by aeolian sedimentation that covered much of the boundary between the tufa
and basalt. However, at one location (WP5000) on the southwestern lobe of the F4 basalt
flow (Figure 4), a small (c. 10 x 15 m) patch of tufa was observed directly overlying this flow
(Figure 7), demonstrating that, at least at this location, tufa deposition occurred after basalt
flow emplacement. Nowhere around the basin was tufa found underlying an in situ basalt
flow, although the assessment of stratigraphic relationships around the margins of the basin
was hampered by erosion of the edges of the tufa at the contact with the basalt, and the filling
of these depressions with younger sediments of unknown provenance.
Figure 7: Tufa deposits overlying F4 basalt flow at WP5000.
3.2 Basalt Morphology in the Wadi Dabsa Basin
Seventeen samples of basalt were removed from around and within the Wadi Dabsa basin
(Figure 4). In addition, eight further samples were collected from locations further afield in
the Harrat al Birk, with a view to their dating providing a wider and broader understanding of
the timing of development of Harrat al Birk (Appendix 2).
The basalts observed and sampled (Figure 8) were mostly fresh, with thin weathering rinds
typically less than a few millimeters thickness and, in many samples, almost non-existent.
January/February 2017 Fieldwork at Wadi Dabsa
The groundmass phases of most samples were relatively fresh, with a micro-phaneritic
texture and plagioclase rich. Vesicularity varied from 0% to >20% by volume, with most
samples <5% vesicles and vesicle sizes of no more than a few millimetres. Phenocrysts were
sparse, accounting for <2–3% by volume of most samples. Only two samples appeared to be
aphyric. Olivine was the most prevalent phenocryst, generally 1–2 mm in diameter and
displaying varying states of alteration, from minor (slight discolouration) to more extensive
alteration to hydroxides/iddingsite, etc. Black, equant clinopyroxene was also present as a
phenocryst phase, ranging up to 1 mm diameter but less common than olivine. One sample
contained xenolithic and antecrystic fragments. The xenolith in this case (~ 1 cm long)
appeared to be both silica- and volatile-rich, having partially melted to form a grey glass,
strongly vesiculated within the basalt host. This same sample also hosted pyroxene+olivine
glomerocrysts of a few millimetres diameter. A number of samples displayed formation of
some alteration or secondary phases, typically lining vesicle walls with whitish or grey/blue
minerals. In many of these types of samples, not all vesicles showed secondary mineral
linings, suggesting that groundwater percolation through the samples was limited in extent.
Figure 8: Examples of basalt samples collected for 40Ar/39Ar dating at locations in Wadi
Dabsa and the Harrat al Birk.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 9: Tufa facies, where (a) is massive intact sheet, (b) is broken and dissected sheet, (c)
pool-type with clear phytoherm framework, (d) heavily weathered former small barrage, (e)
barrage cuspate forms, partly eroded, (f) tufa barrage in narrowing of channel, (g) cascade-
type form, (h) crust or ‘flowstone’-type.
3.3 Tufa Development in the Wadi Dabsa Basin
Tufa in the Wadi Dabsa basin comprises two sub-basins in the upper part of the valley and a
fan-shaped deposit downstream of the basin in the west (Figure 4). At all identified locations,
the tufa appears to overly, and therefore post-date, the emplacement of the basalt flows that
January/February 2017 Fieldwork at Wadi Dabsa
have shaped the basin. The variety of morphologies present can be broadly categorised as
massive sheets greater than 1 m thick that are sometimes intact and sometimes highly broken-
up, pool-type deposits, fan-type deposits, fluvial barrages (both heavily weathered, low relief
forms and larger preserved features in more constrained channels), fluvial cascades at larger
breaks in the underlying topography, and crusts (or tufa flowstone) coating a variety of larger
features (Figure 9).
The Wadi Dabsa basin floor, underlain by tufa, declines in elevation from ~124 m in the east
to ~89 m at its westernmost extent. Drainage is consequently from east to west. A 220 m long
incised section of the wadi marks the western boundary of the two main sub-basins, where
the wadi flows through to the fan-shaped tufa deposits downstream of the main basin (Figure
5). However, the spatial pattern of tufa morphologies suggests that, in the past, water also
flowed from north to south and/or NNE-SSW across the basin. This is demonstrated by:
- patches of sheet tufa overlying basalt at the highest elevations within the north-east
alcoves in the two main sub-basins of the wadi;
- the northeast to southwest direction of slope of the broken massive sheet tufa at the
southern end of the eastern sub-basin towards the incised wadi channel that skirts the
southern edge of the basin; and
- thick (2 to 4 m) tufa-cascade morphologies on the northern banks of the west to east
flowing wadi and its tributaries, most likely resulting from water flowing from north to
south over a break of slope, possibly initiated by fluvial incision of the wadi.
Thirteen samples were collected from ten locations around the Wadi Dabsa basin, including
three from within the artefact sampling grids (Appendix 3). The tufa upon which the
L0106/130 artefact collection grids were situated are heavily eroded barrage morphologies,
some of which provided material potentially suitable for dating. From the northwestern edge
of L0130 grid, a tufa barrage complex extends in a NW direction down into an incised
channel (Figure 10a–c). This complex is composed of six discrete former barrages separated
by flat regions of gravel pavement, which are eroded to different degrees and range in size.
The smallest at the top (Barrage 1) represents vertical break of slope of 5 cm, whilst the
largest is Barrage 6, which itself is a complex of multiple prograding cuspate-shaped forms
(Figure 10a). Sample WD Tufa 9 was taken from Barrage 4 (Figure 10e) and WD Tufa 11
from Barrage 6 (Figure 10d). Samples WD 12–14 were taken from locations on
the L0106 grid.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 10: Tufa barrage complex, locates at NW edge of L0130 grid, illustrating location of
samples WD9 and WD11.
3.4 Local Landform Mapping and Surface Condition Recording
The artefact survey grids are located on a small rise in the centre of the eastern part of the
Wadi Dabsa basin, one of a number of tufa ‘terraces’ that rise in elevation to the north and
east (Figure 3). The survey area is bordered to the south and southwest by a sandy drainage
Six landform units were identified and mapped (Figure 11):
Ridge: This unit forms the highest part of the survey area, in the northwestern section of
L0130, where it is characterised by the presence of a basalt outcrop consisting of rounded to
sub-angular basalt boulders, cobbles and gravel. It also occurs as minor outcrops of basalt at
the southeastern end of the spur in L0106 within the ‘Crest’ unit. In L0106, the surface
between the basalt rocks is covered with fine to coarse, angular to sub-rounded, basalt gravel
mixed with what looks like fine-grained pedogenic carbonate nodules (i.e., not tufa). In
L0130, the surface between the rocks varies from this same carbonate gravel to silty sand.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 11: Map and photos of landform units across L0106/130.
January/February 2017 Fieldwork at Wadi Dabsa
Crest: This unit comprises the broad, flat surface of the northwest to southeast trending spur
on which the L0106 and L0130 grids are located. The surface is mostly bedrock
(predominantly tufa in L0106 and basalt in L0130), with patches of buff-coloured silty sand
becoming more extensive to the northeast.
Upper Slope: This unit consists of a stepped or terraced surface across tufa bedrock
outcropping on the southwest-facing slope of the survey area. The surface is mostly bedrock
(tufa in L0106; tufa and basalt cobbles in L0130), with patches of sandy to gravelly sediment
on the flat treads of the terraces.
Lower Slope: This unit comprises a low-angled footslope towards base of the southwest-
facing slope of the survey area. The surface consists of tufa and basalt gravel and cobbles
lying on sandy to gravelly sediment. Basalt gravel and cobbles are more abundant towards
the northwestern boundary of L0130. Plant growth was present at the time of survey, but it
was very sparse.
Drainage Depression: This unit is characterised by a relatively low-angled drainage
depression bordering the survey area to the east, south and southwest. In the northeastern
section of L0130, it takes the form of an eroding drainage line leading downslope from the
crest to northwest. In the eastern corner of the grid in L0106, this unit consists of numerous
shallow rills through sandy, gravelly sediment. Plant growth is more abundant in this unit
than any of the others, indicating occasional water flow. At the south/southwest boundary of
the grids, this landform unit has flatter topography and marks the distal margin of the
drainage depression at the base of the southwest-facing slope. The surface here is covered
with sediment, mostly buff-coloured silty fine sand. There are rare gravels and cobbles on the
surface, and sparse plant growth. Shallow distributary rill channels indicate that water flows
across this surface from time to time.
Disturbed: This unit comprises narrow strips of land on either side of the unsealed track that
separates L0106 from L0130, where track construction has disturbed the original surface, and
inverted and redeposited large boulders of tufa.
Surface condition observed in the surveyed quadrants (Figure 12), were, in general,
consistent with the characteristics of the wider landform units: sandy sediments are
ubiquitous in all quadrants within the Drainage Depression landform unit; the Lower Slope
unit quadrants consisted of varying amounts of tufa and basalt clasts overlying
unconsolidated sediment, with the occasional in situ tufa exposure; and quadrants within
January/February 2017 Fieldwork at Wadi Dabsa
Figure 12: Examples of varying surface conditions across the L0106 and L0130 grids,
showing in some cases the variability of surface conditions within landform units (Ridge,
Upper Slope, Drainage Depression).
the Crest landform unit were predominantly sediments covered with a deflated lag of clasts
comprising artefacts, tufa and basalt, with quadrants close to the Disturbed landform unit
displaying an increase in sandy sediments. The greatest variation in surface condition was
observed in the Upper Slope landform unit, where quadrants contained widely varying
January/February 2017 Fieldwork at Wadi Dabsa
percentages of bare, in situ tufa and sediment pockets with similar characteristics to those
observed in the Crest unit. These observations therefore illustrate that, within landform units,
there is still significant variation in lithology and active geomorphological processes that
affect the preservation, exposure and visibility of surface artefacts.
3.5 Spatial Distribution of Surface Artefacts
Over two seasons of fieldwork, 2847 lithic artefacts have been collected from grid L0106. In
2015, 909 artefacts were collected from Rows 1 to 4; a further 1938 artefacts collected from
Rows 5 and 6 in 2017. A further 399 lithic artefacts have been recorded but not collected
from the accompanying grid L0130. Given the time constraints for the survey and recording
of material in grid L0130, this assemblage is likely to represent a small proportion of the
artefacts that might be present in this grid.
Figure 13: Spatial distribution of artefacts counts and landforms across L0106/130 grids.
In L0106, the mean number of artefacts per quadrant recorded across the surveyed quadrants
is 23.7 (0.94 artefacts / m2), and there is a clear increase in numbers of artefacts per quadrant
towards the northeast edge of the grid (Figure 13), with some quadrants in Rows 6 and 5
containing over 80 artefacts each (>3.2 artefacts / m2). These quadrants are located on the
January/February 2017 Fieldwork at Wadi Dabsa
flattest part of the Crest landform unit, although 5Dz, the quadrant with the second highest
observed lithic count, is within the Upper Slope landform unit. As shown in Figure 13 the
highest artefact densities are found on the landform units with the highest visibility surfaces
(i.e., Crest and Upper Slope) while the lowest densities are found on the landform unit with
the lowest visibility surface (Drainage Depression).
3.6 Excavation and Artefact Condition
Two small test pits were excavated on the Crest unit to determine the relationships between
the surface artefacts and the underlying sediments and bedrock. Two sediment layers were
observed, a buff-coloured silty fine sand layer of variable thickness (~10 cm) overlying, with
a clear boundary, a layer of grey silty fine sand, with an abrupt lower boundary with the tufa
bedrock beneath (Figure 14a). Both units contain clasts of both tufa and basalt, and lithic
artefacts made on basalt. The similarity of the field texture of the two layers, but difference in
colour, suggests that the buff layer may be a layer of active accumulation by aeolian
deposition, while the grey layer may have had the same origin (i.e., aeolian deposition) but
with accumulation of organic matter perhaps producing the grey colour. Samples of both the
buff and grey layers were collected for further analysis in the lab.
Figure 14: a) Test Pit 1, Quadrant 6By, showing sharp lithic artefact within grey unit; b)
stratigraphy in Test Pit 2, 5Az, showing grey silty sand layer overlain by buff layer
In one pit, an artefact was found lying on top of the tufa, within the lower of the two sediment
units (Figure 14b). Artefacts within this unit overly, and therefore postdate, the tufa. Artefacts
from the grey layer exhibited a greyish patina and very sharp edges, suggesting rapid burial
following manufacture and discard, while those in the buff layer exhibited weathering states
ranging from fresh (i.e., grey and sharp) to a dark brown colour close to the colour of the
basalt outcrops in the grids, with a slightly polished patina and subangular to subrounded
January/February 2017 Fieldwork at Wadi Dabsa
edges (Figure 15). The latter characteristics suggest a period of exposure at the surface after
manufacture and discard.
Figure 15: Different patination states for basalt artefacts at L0106: (a) grey, sharp artefacts
from lower ‘grey layer’ of test pit; (b) mixed grey/sharp and brown/polished artefacts from
‘buff layer’ of test pit; (c) surface artefact (core) exposed on previously-surveyed 2015 grid
showing differential patination/weathering.
This variation in patina shown in the buried artefacts is the same as that observed in the
surface assemblage: the vast majority of the artefacts collected and observed in the 2015 and
2017 surveys were dark brown, some with a slightly polished appearance, although with still
distinct flake scars. Others were either completely grey with sharp edges, or exhibited both
patinas and grades between them (Figure 15).
The observations from the test pits suggest that artefacts contained within the Crest unit may,
after manufacture from the local basalt, have developed a grey patina during burial in the
grey unit. Exposure of these artefacts on the surface by deflation of the fine-grained
sediments in which they were buried then allowed the grey patina to be weathered and
removed, either physically through the polishing effect of aeolian erosion or chemically
through oxidation, or a combination of both processes. It is therefore possible that the surface
artefacts were originally all buried within the grey sediment unit. During the 2017 fieldwork,
numerous artefacts were observed on the surface of the quadrants that had been surveyed and
fully collected in 2015. While it is possible that these artefacts were missed in the original
survey, given that most of these artefacts exhibit variable patina conditions it is also possible
that they have become exposed by deflation since the 2015 survey. In addition, some may
have moved downslope from the as yet uncollected Rows 5 and 6 between 2015 and 2017.
January/February 2017 Fieldwork at Wadi Dabsa
Figure 16: Tufa-encased handaxe found on L0130 grid showing a) location of the findspot in
L0130, Quadrant 5Jw, where the handaxe was found as a loose clast on the Drainage
Depression landform. Handaxe is just in front of scale, which is 30cm; b) tufa-encased
handaxe, dorsal view.
The probable burial of artefacts within the grey unit identifies their deposition as post-dating
tufa formation, as the grey unit overlies them, yet it cannot be dismissed that some artefacts
may never have been buried in this grey layer, but had a different relationship to the timing of
tufa deposition in the basin. A single find from the L0130 grid suggests a different
relationship between the artefacts and tufa deposition: in Quadrant 5Jw, a basalt handaxe
partially encased in tufa was found on the edge of a small rill that drains the NW corner of
L0130 (Figure 16). While the find was a loose clast, and removed from the geomorphological
context in which it developed the tufa coating, it illustrates that a period of tufa deposition
occurred after manufacture of at least some of the artefacts. The tufa therefore has the
potential to provide a minimum age for the handaxe. Detailed analysis of samples of the tufa
will be undertaken at the NERC Stable Isotope Facility at the British Geological Survey,
Keyworth, and the Department of Geography, Royal Holloway University of London.
3.7 Lithic Analysis
The artefact assemblage from Wadi Dabsa is the largest assemblage collected and recorded
from southwestern Saudi Arabia by the UK-Saudi fieldwork team; it is more than twice as
large as the other major collection from the Wadi Dhahaban (Inglis et al., 2014a, b).
January/February 2017 Fieldwork at Wadi Dabsa
In broad terms, the assemblage comprises artefacts and manufacturing debris that can be
typologically assigned both to the ESA and MSA on typological and technological criteria,
with some limited potential evidence of artefacts from a later time period as noted below. As
might be expected from the abundantly available basalt deposits surrounding the site, the
artefacts at Wadi Dabsa are overwhelmingly made on the locally available raw material either
sources as naturally exfoliated basalt flakes or collected in the form of rolled cobbles. The
quality of the material used varies from a fine-grained, dense and homogeneous basalt
through to a less dense material with a greater proportion of vesicules. Just five artefacts
collected from Grid L0106 are not made from locally-available volcanic materials; these
include two pieces each of chert and of quartz and one piece of indurated shale. The size,
presence of retouch and typological form of these exotic pieces suggests that they may date to
the LSA in this region and represent tools brought from outside the basin to the site.
The assemblage from Grid L0106 contains artefacts from all stages of lithic reduction (Table
1) including tested clasts, cortical and semi-cortical flakes, cores and core-preparation
materials, retouched pieces and a large number of shatter fragments attesting to the flaking of
basalt at this location. A large number of cores show working that is both carefully prepared
as well as relatively expedient and simple (Table 2). The assemblage even includes a small
number of probable hammer stones. In addition to the evidence of lithic manufacture, there
are more than 280 retouched pieces which include a range of classic retouched forms from
the ESA/Acheulean as well as the MSA whose forms indicates a range of activities that might
have been undertaken at the site, and in certain cases the preparation of armatures for use
elsewhere (Table 3). A selection of exemplar artefacts is presented below (Figure 17).
The assemblage recorded but not collected from L0130 is, as noted above, smaller in number
and almost certainly missing the not easily spotted pieces. Despite this, the range of
technological pieces as well as retouched pieces suggests that the assemblage in this grid
closest mirrors that from L0106 both in chronological periods and the range of technological
activity; there is evidence for a range of ESA and MSA tool forms as well as a similar range
of manufacturing activities.
January/February 2017 Fieldwork at Wadi Dabsa
Technological Type
L0106 Frequency
L0130 Frequency
Cortical Flakes (Cortex = 100%)
Semi-cortical Flakes (Cortex > 50%)
Semi-cortical Flakes (Cortex < 50%)
Semi-cortical Flakes (Cortex < 20%)
Prepared Core Flakes
lateral & directional
lateral & mixed-directional
flake blades
True Blades
Biface Trimming flakes
Burin Spalls
Tranchet Flakes
Retouched pieces
Cores (whole)
Core (fragments)
Core rejuvenation flakes
Prepared-core preparation flakes
Indeterminate (including tested clasts)
Hammer Stones
Table 1. The frequency of different technological types from Wadi Dabsa (L0106 & L0130).
Core Form
L0130 Frequency
Early Stone Age
Discoidal cores
Middle Stone Age
Centripetal-prepared core
Convergent-prepared core
Prepared-core preform
Single-platform cores
2-platform cores
Globular cores
Cores on flake
Table 2. The frequency of core types within the assemblages recorded at Wadi Dabsa (L0106
& L0130).
January/February 2017 Fieldwork at Wadi Dabsa
Figure 17. The lithic assemblage collected from Quadrant 6Bz (L0106). The top rows of
artefacts include large ESA flakes, as well as early-stage decortification flakes for the
preparation of prepared cores. The second and third row include prepared core flakes,
including a convergent flake and a flake blade, as well as prepared core preparation flakes
for both the upper and lower surfaces. The final two rows include small platform preparation
flakes as well as examples of shatter pieces that detach from the core during working.
3.7.1 ESA/Acheulean Elements at L0106
Approximately 25% of the retouched tool collection comprises pieces that can be ascribed to
the ESA/Acheulean on typological grounds. This includes a number of bifaces (42
examples), cleavers (ten examples), choppers (six examples) and a collection of large cutting
tools (38 examples). Most of these tools have been relatively simply retouched using the
minimum number of retouch flakes to create a functional tool shape. There is evidence to
suggest that at least the final stages of biface making took place at this site through the
presence of a collection of clearly identifiable biface trimming flakes (22 examples) within
the assemblage, and a small number of tranchet flakes (six examples) used to sharpen the
edge of these tools.
3.7.2 MSA/Prepared Core Technology at L0106
The clearest evidence for the use of this locality in the MSA takes the form of the extensive
collection of prepared core forms (both centripetal- and convergent-prepared: 69 examples
and 43 examples respectively), a number of core preparation flakes (346 examples), and the
January/February 2017 Fieldwork at Wadi Dabsa
intended prepared core flakes themselves, again in both centripetal and convergent forms
(119/97 examples respectively). In addition to these forms, there is also a large number of
flake-blades produced either as lateral blades on centripetal cores or as flake clades with clear
central ridges (119 examples). Finally, there are a small number of true blades and bladelets
deriving from prismatic blade cores, possibly of MSA age, or younger (7/4 examples).
Within the extensive collection of convergent prepared core flakes, some flakes have been
retouched into point forms (11 examples), and a small number of fragmented convergent
flakes and points suggested that some had been hafted and then broken in use, with the
fragments removed from the hafts at the site and then refitted with new point forms. A small
number of the convergent flakes and point forms also show evidence of probable impact
fractures at the tip, further supporting their use as weapons. Finally, a number of borers or
piercers (33 examples), both large and small, as well as burins (13 examples) indicate the
likely working of softer materials such as wood or skin.
Retouched Type
Core Tool
Large Cutting Tool
Backed Knife
Bitruncated Piece
Borer / Piercer
Splintered Piece / wedge
Table 3. The frequency of different retouched pieces recorded at Wadi Dabsa (L0106 &
January/February 2017 Fieldwork at Wadi Dabsa
3.7.3 Potential for Further Analysis
Despite the predominantly surface-collected nature of the assemblage, and the attendant
problems of palimpsest formation, in contrast to all the sites previously encountered through
field survey, the Wadi Dabsa assemblage presents a great opportunity to understand in detail
the processes of lithic manufacture for both ESA and MSA time periods. Specifically, the
abundance of locally available raw materials, along with the presence of manufacturing
debris from the earliest stages of material acquisition, through core preparation, use and
maintenance of both cores and finished products offers a, so-far, unique opportunity in the
region to reconstruct in detail the process of manufacture and the range of expertise present
at the site.
Taking the cores as an example, there are 247 cores so far collected from the site,
approximately 10% of the collected assemblage (Table 3). Many of these cores display
similar patterns of working to cores and tools encountered at other sites: short episodes of
working of through either one or two episodes of parallel flaking (single platform and 2-
platform cores – 78 examples). There are also a significant number, however, of cores
designed for longer episodes of exploitation either as ESA discoidal cores (26 examples), or
as MSA prepared cores (104 examples) in wither centripetal or convergent forms with the
former producing flakes suitable for later working into scrapers, and knives and the latter
producing flakes suitable for hafting as weapon points. The abundance of prepared core
forms makes it possible to explore whether the prepared-core technology of Wadi Dabsa
conforms to the Nubian technique of prepared core manufacture as seen at other sites in
Arabia and identified as a form of technology that may help in the identification of the
processes of hominin dispersal (see Rose et al., 2011; Hilbert et al., 2017
and references therein).
Moreover, there are signs of a clear disparity in the quality of preparation amongst the
prepared cores, possibly reflecting different levels of lithic manufacturing expertise amongst
hominins at the site. As noted above, prepared cores require an understanding of the balance
between the volume and shape of raw material above and below a central level within the
core (Boëda, 1995). These volumes must be carefully prepared in the working of the raw
material from initial ‘block’ through to core, hence the term prepared-core. This requires a
sequence of flakes to be removed as both upper- and lower-surface preparation flakes. With
the removal of each preparation flake, there a chance that the intended flake will not detach
smoothly leaving a future ‘problem’ in the core (for example a step-fracture on a flaking
January/February 2017 Fieldwork at Wadi Dabsa
Figure 18. An example of two prepared core flakes that refit at the butt end. The upper flake
in this picture has been further retouched around the distal end and no longer fills its
negative flake scar
surface that prevents further preparation of the flaking surface unless removed) that must be
worked around if the core is to be prepared for the successful removal of the intended flake
form. Amongst the prepared cores at Wadi Dabsa, there are examples of cores that have been
successfully prepared (both centripetal and convergent), and re-prepared (maintained) for the
detachment of flakes of a designed shape. There are also cores that demonstrate a clear
understanding of the appropriate strategy for creating centripetally-prepared cores whilst at
the same time demonstrating a lack of the practical skills necessary to work around the errors
produced in core preparation. The opportunity of reconstructing some of this process of
working is suggested by the presence of two refitting flakes identified by chance in the
analysis of the lithic assemblage collected in 2015 (Figure 18).
January/February 2017 Fieldwork at Wadi Dabsa
4. Discussion
4.1. Evolution of the Wadi Dabsa basin
Fieldwork at Wadi Dabsa clarified the preliminary desktop mapping of landscape units using
remote sensing data, and provided a solid framework to begin to interpret the artefacts we
recorded from the basin within a dynamic landscape context.
The regional landscape mapping had suggested that the basalt flow F4 at the eastern end of
the basin may have post-dated formation of the highest elevation tufa, thus potentially
providing a minimum age for tufa deposition, and also a potential mechanism by which tufa
formation may have been interrupted by blockage by basalt of the pre-existing drainage
system. The observation we made of tufa overlying part of the northwestern boundary of this
basalt flow, however, shows that tufa deposition post-dated the emplacement of the flow and,
in the absence of any further field evidence showing basalt emplacement over tufa, indicates
that tufa formation, at least in this part of the basin, occurred more recently than the youngest
basalt flow. This sequence, of tufa overlying basalt, is repeated over the entire basin,
especially in the incised wadis, and indicates that the tufa was forming in, and filling, a basin
whose underlying morphology was controlled by the basalt flows. It is likely, however, that
there was more than one phase of tufa deposition in the basin, and that deposition of tufa and
basalt may well have overlapped. The absolute dating of the tufa and basalt flows should
test this hypothesis.
It seems likely that the input of water and dissolved CaCO3 into the Wadi Dabsa basin has
both a surface water and a groundwater component, although the balance between the two
requires additional investigation within the region. It is also likely that the CaCO3 is
ultimately derived from the limestones and shales of the Asir Mountains, approximately 60 to
80 km to the east of the survey location, and that Ca might also be derived from the
weathering of the mafic basalt at, and surrounding, the site. Clarification of carbonate source
should be resolvable in future using 87Sr/86Sr isotopic ratios of the carbonate. The possibility
of a groundwater component in tufa formation in the Wadi Dabsa basin is supported by
the following observations:
(i) It was observed that the east-west flowing wadi to the south of the artefact survey area has
no visible tufa deposits, whereas the deeply incised wadi channel between the basin and the
downstream fan contains extensive, thick tufa deposits;
January/February 2017 Fieldwork at Wadi Dabsa
(ii) Many of the tributary wadis on the eastern margin of the basin do not
contain tufa deposits;
(iii) The higher elevation tufa deposits in the northeast sections of the two sub-basins occur
where there is no obvious fluvial channel input. It is possible, however, that the basalt lava
flows in these areas have infilled and cut off former fluvial channels, so that drainage is now
via shallow groundwater and perhaps even in the unsaturated zone through the basalt flows.
Channels may also have incised below the level of the uppermost tufa. Vertical boreholes
drilled through the floor of the basin would help to clarify these issues.
4.2 Geomorphological Context of the L0106 & L0130 Artefacts
Observations from the 2017 field season have revealed a complex relationship between the
artefacts and tufa formation. The tufa-coated handaxe discovered on the northeastern side of
Grid L0130 shows that a period of tufa deposition occurred after manufacture and deposition
of the artefact. Yet in the test pits excavated in Grid L0106, artefacts appear to lie within a
sediment layer directly overlying the tufa, suggesting that deposition of the artefacts post-
dates the formation of the tufa. Such variability in tufa-artefact relationships is, however,
unsurprising given the potentially long time span of activity represented by the presence of
ESA, MSA and LSA artefacts at the site, and the evidence we have so far for complexity of
tufa formation within the basin. The lithic artefact assemblages and the tufa are probably both
the product of multiple phases of human activity and tufa deposition across the basin, phases
that potentially occurred over extended periods of time.
The stratigraphy of the test pits, and the observation of new artefacts found exposed on the
previously collected areas of the L0106 grid, suggest that at least some of the artefacts at
L0106 were buried within a silty sand layer that overlies the tufa that was subsequently
deflated by wind and water erosion, exposing these artefacts to weathering and erosion at the
surface. The origin of this layer is currently unknown it may be a remnant of a soil that
formed on the surface of the tufa. Elsewhere in the basin, a greyish silty sand material was
observed underlying tufa structures, again highlighting the complexity of tufa formation and
its subsequent weathering and erosion.
The presence in 2017 of numerous artefacts on the area of L0106 that underwent full
collection in 2015 may be the result of one of two possibilities. Either, they were missed in
the original survey, or they have become exposed since that survey. A number, but not all, of
these artefacts are patinated, indicating their recent exposure. Late 2016 saw heavy rainfall in
January/February 2017 Fieldwork at Wadi Dabsa
the region which may have intensified deflation of the artefact-containing sediments, primed
perhaps by the 2015 team members disturbing a previously stable surface by trampling it
during survey. A similar situation was observed by one of us (PCF) at semi-arid field sites in
western New South Wales, Australia (Fanning et al., 2009). The increase, however, is of an
order of magnitude, and the distribution of artefacts across the landforms units surveyed in
2017 still shows patterning consistent with the rest of the grid, indicating that this variation is
not simply an artefact of observing the distributions before and after a two-year hiatus.
Geomorphological mapping and observations of the artefact distributions across the grids has
allowed the examination of whether the artefact distribution at the grid scale is influenced by
geomorphic processes that control the preservation, exposure and visibility of artefacts. There
are low counts of artefacts on the landform units most likely to exhibit surface conditions that
inhibit the visibility of artefacts, such as the Drainage Depression unit and, to some extent,
the Lower Slope unit, with its mix of fine sediment and large clasts.
The relatively high density of artefacts recorded within the Crest landform unit which has
abundant sediment cover and therefore, theoretically, low artefact visibility, is an example of
how the interplay between preservation and exposure of artefacts is key to where artefacts are
observed. The Crest is the landform unit that preserves the greatest extent of the artefact-
bearing sediment layer. The low relief of the landform means that, while there is active
deflation occurring across the surface, it is the fine-grained sediments that are being eroded.
The larger clasts, including the artefacts, are not moving very far laterally, if at all, but
instead form a ‘pavement’ or ‘lag’ of clasts on the land surface in which the visibility of
artefacts is high. The stepped nature of the ‘Upper Slope’ unit mirrors this situation, with the
flatter areas behind the bare tufa ridges and barrages exhibiting a similar ‘pavement’ where
sediment is preserved, and containing slightly higher counts of artefacts than the quadrants
with a predominance of bare tufa. The tufa ridges, in contrast, are environments in which
artefacts on the surface, are susceptible to downslope movement because of their exposure on
irregular or sloping surfaces. This mixture in artefact preservation and exposure across the
two different types of surface condition within the ‘Upper Slope’ landform unit explains why
the artefact counts within this landform lie between those of the Lower Slope unit and the
Crest unit. Testing of these models in the L0130 grid is not really possible given the rapid
survey that was undertaken there in comparison to the full collection carried out in L0106. In
addition, the surveyed squares did not intersect the Crest landform unit in this grid.
January/February 2017 Fieldwork at Wadi Dabsa
4.3 The Wadi Dabsa Lithic Assemblage in Techno-Typological Context.
The assemblage from the Wadi Dabsa comprises a mixture of ESA/Acheulean pieces
including a collection of bifaces, large cutting tools and cleavers, as well as an extensive
range of MSA/Middle Palaeolithic pieces including both an extensive collection of
manufacturing debitage as well as retouched pieces. All of these materials appear to be made
on locally available volcanic rocks, that can be accessed, at present, both at the site itself and
throughout the Wadi Dabsa basin. In a similar fashion to sites examined elsewhere in
southwestern Saudi Arabia, the abundance of locally available raw materials would appear to
have created a situation in which hominins did not need to maximise the efficiency with
which they used materials. As a result, many of the pieces are simply retouched with just the
working tips of bifaces, for example, finely retouched to produce rectilinear edges and the
buts remaining often either cortical or very simply retouched. There are a few examples of
lithic materials that are not obviously occurring in the local basin (including quartz, chert and
indurated shale) and these pieces also show greater preparation and retouch. It seems likely
at the moment that these pieces represent the discard of later, post MSA hominins. Likewise,
there are a few examples of tool types (endscrapers and burins) that are normally associated
with Upper Palaeolithic assemblages in Europe, although such tools have been found in MSA
assemblages in Africa.
Looking beyond the Wadi Dabsa basin, the lithic assemblage itself contains pieces that are
similar to examples found during the course of the fieldwork by the UK-Saudi team since
2013, but the quantity and, specifically, the quality of the technological information present
at Wadi Dabsa is more extensive and complete in sequence than at any other site located so
far. The variability in evident skill among the examples of prepared cores also suggests that
the assemblage represents a cross-section of ages and experience in lithic manufacture; this is
unique to this site.
5. Summary and Conclusions
The January/February 2017 field season in Wadi Dabsa, Asir Province, Saudi Arabia,
achieved the stated objectives and provided a wealth of information on the Palaeolithic
artefacts and their geomorphological and geological setting.
Objectives achieved
1. Remote sensing and satellite data, groundtruthed by field observations, proved
invaluable in the definition of landscape units within the wider basin, the development
January/February 2017 Fieldwork at Wadi Dabsa
of sampling strategies, and the understanding of sediment cover across the basin and
its potential to impact on artefact visibility.
2. Geomorphological survey of the locality and its direct environs, with specific focus
on the tufa deposition and basalt, allowed the development of hypotheses regarding
the landscape setting of L0106/130, and geomorphological controls on the
preservation, exposure and visibility of the artefacts in the wider landscape. Detailed
landform and surface mapping of the collection grids themselves furthered the
understanding of artefact distribution at the locality and the geomorphological
controls acting upon it.
3. A total of 1938 Palaeolithic artefacts were collected from the 20 x 50 m extension of
L0106, with a further 399 lithic artefacts recorded in situ across L0130, comprising a
recorded assemblage over the two seasons of 3226 artefacts, many of them with ESA
or MSA affinities, one of the richest Palaeolithic assemblages so far recorded in SW
Saudi Arabia.
4. Comprehensive post-excavation analysis of the artefacts collected from L0106,
coupled with the in situ recording of the L0130 artefacts, has allowed the assemblage
to be considered in relation to those from neighbouring regions.
5. Excavations of two small test pits and the discovery of a handaxe encased in tufa have
indicated a complex relationship between the artefacts and tufa deposition, perhaps
resulting from the extended period of time spanned by the activity at the site
(suggested by the techno-typological characteristics of the artefacts) and/or multiple
phases of tufa deposition.
6. Samples for both 234U230Th!dating of tufa and 40Ar/39Ar dating of basalt lava flows
were collected from locations in the Wadi Dabsa basin and further afield to provide a
future chronological framework for the development of the basin and the archaeology
within it. The tufa samples will also potentially provide, through isotopic analyses,
high-resolution snapshots of palaeoenvironments in the basin.
The 2017 field season, whilst very successful in achieving its stated objectives, has only just
begun to realise the potential of the Wadi Dabsa basin for informing on ESA and MSA
hominin behaviour. The geomorphological work at the site, particularly the mapping and
sampling of the tufa, has proven its complexity, and indicated that there is much more to be
done in terms of survey to understand the geomorphological controls and landscape context
of the archaeology of this region.
January/February 2017 Fieldwork at Wadi Dabsa
Future work in the Dabsa basin must focus on two main issues. The first is the further
understanding of the timing and conditions of tufa deposition within the basin. Detailed
mapping using high-resolution remote sensing and field observations, coupled with detailed
stratigraphic and microscopic analysis of the tufa facies should be used to unravel the nature
and sequence of the depositional environments. This stratigraphic framework, along with the
data from the pilot isotopic analyses and 234U–230Th dating programme from the 2017
samples, would then be used to target further palaeoenvironmental investigation, and
chronological constraint of, the tufa deposition in the basin.
The second area is the expansion of the surveyed areas to further understand the L0106/130
assemblage in the context of the artefact distributions across the basin is the density of
artefacts at this location unusual, and if so, why might there be a concentration of artefacts
deposited or preserved here? The 2015 transects, considered along with the low density of
artefacts observed during geomorphological investigations across the basin, indicate that the
number of artefacts at L0106/130 is unusually high. Further archaeological investigation of
similar geomorphological settings to this assemblage around the basin, targeted through
detailed geomorphological and tufa mapping, however, should be carried out to confirm that
it represents an unusually dense concentration of artefacts, in order to understand the
potential for the assemblage to inform on human activity at this location.
6. Official Meetings
During our visit to Saudi Arabia, members of the team carried out a number of official
meetings. Professor Geoff Bailey presented the work of the team and future plans to HRH
Prince Sultan Bin Salman bin Abdal Aziz at a meeting at the SCTH in Riyadh on 16th
January. Members of the team attended an audience with HRH Prince Faisal bin Khalid bin
Abdul Aziz Al Saud in his offices in Abha, Asir province on 1st February, where they
presented information about the project which was received with great interest. In addition,
members of the team met the Mayor of Al Birk, in his offices in Al Birk to provide
information on our work and highlight the cultural potential of the area to thank him and the
various government departments for their extensive logistical support towards our research
activities in the region.
7. Acknowledgements
We thank HRH Prince Sultan bin Salman bin Abdul Aziz, President of the Saudi
Commission for Tourism and National Heritage (SCTH), KSA, Professor Ali Al-Ghabban,
January/February 2017 Fieldwork at Wadi Dabsa
Consultant to the President, Dr Hussein Abu Al Hassan, Vice President for Antiquities and
Museums, and Dr Abdullah Al Saud, Director General, Dr Abdullah Al Zahrani, General
Manager of Research and Archaeological Studies for granting fieldwork permission and for
their support of our work in Saudi Arabia. Grateful thanks are also extended to Mr Saeed Al
Karni, Director of Antiquities in Asir and his staff at the Abha Regional Museum. Mr
DhaifAllah bin Tha’ar Al-Otaibi, National Museum lead the Saudi team in the field.
The 2017 fieldwork was funded by generous grants from the British Academy (Arthur
Reckitt Fund), the Gerald Averay Wainwright Fund for Near Eastern Archaeology at the
University of Oxford, and the British Foundation for the Study of Arabia, with additional
funding from the European Union’s Horizon 2020 research and innovation programme under
the Marie Skłodowska-Curie grant agreement No. 660343, “SURFACE: Human-Landscape-
Interactions and Global Dispersals: The Surface Record of Palaeolithic Arabia”.
8. References
Bailey, G. N., 2009. The Red Sea, coastal landscapes, and hominin dispersals. In: Petraglia,
M. D, Rose, J. I. (Eds), The Evolution of Human Populations in Arabia. Amsterdam,
Springer: 15–37.
Bailey, G.N., Alsharekh, A., Flemming, N.C., Lambeck, K., Momber, G., Sinclair, A., Vita-
Finzi, C., 2007. Coastal prehistory in the southern Red Sea basin, underwater
archaeology, and the Farasan Islands, Proceedings of the Seminar for Arabian Studies
37, 1–16.
Bailey, G.N., Inglis, R.H., Meredith-Williams, M.G., Hausmann, N., Alsharekh, A.M., Al
Ghamdi, S., 2012. Preliminary Report on Fieldwork in the Farasan Islands and Jizan
Province by the DISPERSE project, NovemberDecember 2012, Report to the Saudi
Comission for Tourism and Antiquities.
Bailey, G.N., Devès, M.H., Inglis, R.H., Meredith-Williams, M.G., Momber, G., Sakellariou,
D., Sinclair, A., Rousakis, G., Al Ghamdi, S., Alsharekh, A., 2015. Blue Arabia:
Palaeolithic and underwater survey in SW Saudi Arabia and the role of coasts in
Pleistocene dispersals, Quaternary International, 382, 42–57.
Boëda, E., 1995. Levallois: A volumetric construction, methods, a technique. In: Dibble,
H.L., Bar-Yosef, O. (Eds.), The Definition and Interpretation of Levallois
Technology. Prehistory Press, Madison (Wisconsin). 41–68.
Bordes, F., 1979. Typologie du Palaeolithique, Ancien et Moyen. Paris, Centre National de la
Recherche Scientifique.
Bosworth, W, Stockli, D. 2016. Early magmatism in the greater Red Sea rift: timing and
significance. Canadian Journal of Earth Science,s 53(11): 1158–1176.
Coleman, R.G., Gregory, R.T., Brown , G.F., 1983. Cenozoic Volcanic Rocks of Saudi
Arabia, United States Geological Survey Open-File Report 83-788.
January/February 2017 Fieldwork at Wadi Dabsa
Dabbagh, A., Emmermann, R., Hötzl, H., Jado, A.R., Lippolt, H.J., Kollman, W., Moser, H.,
Rauert, W., Zötl, J.G., 1984. The development of Tihamat Asir during the Quaternary.
In: Jado, A.R., Zötl, J.G. (Eds), Quaternary Period in Saudi Arabia Volume 2:
Sedimentological, Hydrogeological, Hydrochemical, Geomorphological,
Geochronological and Climatological Investigations in Western Saudi Arabia,
Springer-Verlag, Vienna. 150–173.
Debenath, A., Dibble, H.L., 1994. Handbook of Palaeolithic Typology. Philadelphia,
University of Pensylvania Museum Press.
Fanning, P.C., Holdaway, S.J., Rhodes, E.J., Bryant, T.G. 2009. The surface archaeological
record in arid Australia: geomorphic controls on preservation, exposure and visibility.
Geoarchaeology, 24: 121–146.
Foulds, F., Shuttleworth, A., Sinclair, A., Inglis, R., Alsharekh, A., Al Ghamdi, S., Bailey,
G.N., In Press. A Preliminary Report of a Large Handaxe from Wadi Dabsa, Saudi
Arabia: Implications for Hominin Adaptations to the Arabian Peninsula, Antiquity.
Groucutt, H.S., Petraglia, M.D., Bailey, G.N., Scerri, E.M., Parton, A., Clark-Balzan, L.,
Jennings, R.P., Lewis, L., Blinkhorn, J., Drake, N.A., Breeze, P.S., Inglis, R.H.,
Deves, M.H., Meredith-Williams, M.G., Boivin, N., Thomas, M.G., Scally, A. 2015.
Rethinking the dispersal of Homo sapiens out of Africa. Evolutionary Anthropology,
24: 149–164.
Hilbert, Y.H., Crassard, R., Charloux, G, and R. Loreto, 2017. Nubian technology in northern
Arabia: impact on interregional variability of Middle Palaeoltihic Industries.
Quaternary International, 435: 77–93.
Holdaway, S.J., Fanning, P.C. 2014. Geoarchaeology of Aboriginal Landscapes in Semi-arid
Australia. CSIRO Publishing, Melbourne.
Inglis, R.H., Foulds, F., Shuttleworth, A., Alsharekh, A.M., Al Ghamdi, S., Sinclair, A.G.,
Bailey, G.N., 2015. The Palaeolithic Occupation of the Harrat Al Birk: Preliminary
Report on the 2015 Fieldwork in Asir Province, Southwest Saudi Arabia. Report to
the Saudi Comission for Tourism and Antiquities.
Inglis, R.H., Sinclair, A., Shuttleworth, A., Alsharekh, A., Al Ghamdi, S., Devès, M.,
Meredith-Williams, M.G., Bailey, G.N., 2014a. Investigating the Palaeolithic
Landscapes and Archaeology of the Jizan and Asir Regions, Southwest Saudi Arabia,
Proceedings of the Seminar for Arabian Studies, 44: 193–212.
Inglis, R.H., Sinclair, A.G.M., Shuttleworth, A., Meredith-Williams, M.G., Hausmann, N.,
Budd, W., Alsharekh, A., Al Ghamdi, S., Bailey, G.N., 2014b. Preliminary Report on
2014 Fieldwork in Southwest Saudi Arabia by the DISPERSE project: (1) Jizan and
Asir Provinces. Report to the Saudi Comission for Tourism and Antiquities.
Inglis, R.H., Sinclair, A.G.M., Shuttleworth, A., Alsharekh, A.M., Al Ghamdi, S., 2013.
Preliminary Report on 2013 Fieldwork in Southwest Saudi Arabia by the DISPERSE
Project: (2) Jizan and Asir Provinces, February-March 2013. Report to the Saudi
Comission for Tourism and Antiquities.
Lambeck, K., Purcell,A., Flemming, N.C., Vita-Finzi, C., Alsharekh, A.M., Bailey, G.N.
2011. Sea level and shoreline reconstructions for the Red Sea: isostatic and tectonic
considerations and implications for hominin migration out of Africa. Quaternary
Science Reviews, 30(25–26): 3542–3574.
January/February 2017 Fieldwork at Wadi Dabsa
McNabb, J., Sinclair, A., Quinney, P. 2003. Recent Investigations into the Later Acheulean
of the Makapansgat Region, Northern Province, South Africa. In: Moloney, N., Shott,
M., (Eds). Lithic Analysis at the Millenium. London, Institute of Archaeology. 3–16.
Petraglia, M. D. 2003. The Lower Palaeolithic of Arabian Peninsula: occupations,
adaptations, and dispersals. Journal of World Prehistory, 17(2): 141–179.
Petraglia, M. D., Alsharekh, A. 2003. The Middle Palaeolithic of Arabia: implications for
modern human origins, behaviour and dispersals. Antiquity, 77: 671–684.
Rose, J.I., Usik, V.I., Marks, A.E., Hilbert, Y.H., Galletti, C.S., Parton, A., Geiling, J.M.,
Cerny, V., Morley, M.W. , Roberts, R.G. 2011. The Nubian complex of Dhofar: an
African Middle Stone Age industry in southern Arabia. PLoS One, 6: e28239
Sinclair, A., McNabb, J. 2005. All in a day's work: Middle Pleistocene individuals,
materiality and lifespace at Makapansgat, South Africa. In: Gamble, C., Porr, M.
(Eds). The Hominid Individual in Context: the Archaeological Investigations of Lower
and Middle Palaeolithic Landscapes. Locales and Artefacts. London, Routledge. Pp.
Stone, A. E. C., Viles, H. A., Thomas, L., van Calsteren, P. 2010. Can 234U230Th"dating be
used to date large semi-arid tufas? Challenges from a study in the Naukluft
Mountains, Namibia. Journal of Quaternary Science, 25(8): 1360–1372.
Zarins, J., Whalen, N.M., Ibrahim, I., Morad, A., Khan, M., 1980. Comprehensive
Archaeological Survey Program: preliminary report on the Central Southwestern
Provinces survey. Atlal, 4: 9–117.
Zarins, J., Murad, A., Al-Yaish, K., 1981. The Second Preliminary Report on the
Southwestern Province. Atlal, 5: 9–42.
Appendix 1: Field Team
Dr Robyn Inglis, University of York, UK
Prof. Anthony Sinclair, University of Liverpool, UK
Prof. Geoff Bailey, University of York, UK
Dr Harry Robson, University of York, UK
Dr Patricia Fanning, Macquarie University, Australia
Dr Abi Stone, University of Manchester, UK
Dr Dan Barfod, Scottish Universities Environmental Research Centre, Glasgow, UK
Dr Abdullah Alsharekh, King Saud University, KSA
Mr DhaifAllah Tha’ar Al Othaibi, SCTH, Riyadh
Mr Bandar Al Wabr, SCTH, Riyadh
Mr Abdulrahman Al Hammad, SCTH, Asir
Mr Saeed Abu Mater, SCTH Asir
Mr Saleh bin Hashwal Al Qatani, SCTH, Asir
Mr Faia’ Essa Assiri, SCTH, Asir
Mr Ahmed Al Sharef, SCTH, Riyadh
January/February 2017 Fieldwork at Wadi Dabsa
Appendix 2: List of Basalt Samples Collected in 2017
Sample No.
Samples from Wadi Dabsa Basin
N 18.306903
E 41.570055
F4 basalt flow snout, central lobe, 123 m elevation. 1–2 kg mined from larger block. Sparse olivine and clinopyroxene phenocrysts, up to 1 mm diameter, no vesicles.
Groundmass is fresh. Flow surface is boulder and seems to have undergone aeolian erosion.
N 18.309157
E 41.571555
F4 basalt flow snout, 136 m elevation. Sparsely phyric, similar to sample AB1 (possibly same flow).
N 18.311031
E 41.569493
F6 basalt flow, 128 m elevation. Olivine phyric, vesicular basalt.
N 18.310359
E 41.563783
F6 basalt flow, 116 m elevation. Olivine and clinopyroxene phyric, vesicular basalt. Very sparse xenoliths present as olivine-clinopyroxene glomerocrysts and partially
melted vesicle-rich silicic inclusions.
N 18.309636
E 41.556531
Basalt of flow protruding through tufa. 93 m elevation. Olivine phyric, vesicular basalt.
N 18.302100
E 41.569500
F10 basalt flow. Olivine & clinopyroxene phyric vesicular basalt. Olivine shows slight brownish discoloration. Phenocrysts range up to 1 mm diameter. Groundmass is
fresh and dense but has pervasive irregular microvesicles. Flow top is relatively flat. Sample taken from site above stream.
N 18.301836
E 41.568121
Scoriaceous composite spatter clast (~15 cm) from river channel cut. Sample from chaotic, matrix supported deposit adjacent to a megablock (~ 3 m diameter) oxidised
pyroclastic material, The pyroclastics of the megablock appear to be spatter, i.e., irregular cm-scale clasts. Weathering obscures contacts, but the scoria bearing diamicton
appears to encase the megablock. Megablock is also overlain by a thin lava flow, a few tens of cm thick, and relatively flat lying that is overlain in turn by basaltic cobbles.
May be debris avalanche deposit overlain by subsequent flows.
N 18.301883
E 41.556533
Olivine phyric basalt. The elongate features of the ‘C5’ cluster are mostly likely a set of eroded dikes. There was no scoria deposit there, only dense lavas. The sample was
taken from the foot of one of these ridges, above the surrounding plain. Basalts have a hackly, but weathered appearance here. Abundant olivine phenocrysts, up to 1 mm,
light brownish alteration. Dense, moderately fresh groundmass.
N 18.315283
E 41.555300
Sample from F6 basalt flow, top of flow snout. Olivine phyric, vesicular basalt. Some vesicles show secondary minerals and caliche. Sparse olivine phenocrysts, up to 1
mm, are lightly altered to a brown color.
N 18.314583
E 41.559967
Sample from F6 basalt flow snout. Olivine phyric, basalt. Groundmass is fresh, dense and vesicle free. Conspicuous spheroidally weathered zones, ~ 0.5 m diameter.
N 18.307367
E 41.560533
Sample from basalt of unknown flow protruding through tufa. Sparsely olivine phyric, vesicular basalt. Sparse, large vesicles. Bouldery snout, clasts (?) up to 1 meter
diameter. Maybe debris flow snout? low surface relief (1020 cm) and small clasts of tufa distributed across the surface suggest at least some sedimentation.
N 18.308167
E 41.564383
Sample from basalt of unknown flow protruding through tufa. Olivine phyric, vesicular basalt. Olivine is lightly altered and some vesicles have secondary minerals. ~30
meters distant from L0130 grid. Sparse phenocrysts and vesicles.
N 18.307650
E 41.564583
Sample from basalt of unknown flow protruding through tufa under L0130 grid. Olivine phyric, vesicular basalt. Sample from 0.5 m boulder. Vesicles have secondary
minerals, caliche and light alteration. Sparse olivine phenocrysts range up to 3 mm diameter and show variable light alteration.
N 18.305467
E 41.556783
Basalt from flow F10. Olivine phyric, vesicular basalt. Pervasive alteration. Olivine and vesicle rich. Olivine phenocrysts are moderately altered to brownish minerals and
vesicles are mostly lined with secondary minerals.
N 18.307100
E 41.554917
Basalt from flow F18. Olivine phyric, vesicular basalt. Abundant olivine shows variable alteration, from light discoloration to complete replacement. Minor vesicles are
lined or filled with secondary minerals. 87 m elevation, above tufa barrage.
N 18.303183
E 41.539350
Basalt from cinder cone C6. Aphyric basalt. Dense, black aphyric basalt. Single exception is a large, anhedral plagioclase megacryst (~10 mm) with smooth, sharp contacts
against the host basalt. Megacryst contains oxide inclusions, up to 1 mm diameter. This sample is from the foot of the cone, above the surrounding plain.
January/February 2017 Fieldwork at Wadi Dabsa
Samples from Harrat Al Birk Region
N 18.071567
E 41.624333
Sample from Dhahaban Quarry. Olivine phyric basalt. 14 m elevation. Cone near highway. Abundant phenocrysts of olivine, up to 1 mm diameter, are altered to a
brownish color. Groundmass is dense and fresh.
N 18.223900
E 41.542850
Sample from columnar basalt flow to NE of centre of Al Birk. Olivine phyric basalt. Sparse olivine is lightly altered. Groundmass is reasonably fresh. Colonnade is
overlain by cinder cone deposits.
N 18.223900
E 41.542850
Sample from scoria bomb to NE of centre of Al Birk. Aphyric, variably oxidized, vesicular basalt. 47 m elevation.
N 18.262817
E 41.594633
Columnar basalt along Al Birk/Muhayil Road. Olivine phyric, vesicular basalt. Sparse olivine is discoloured to light brown. Vesicles are abundant but very small,
<0.5mm. Groundmass is fresh. 192 m elevation
N 18.281867
E 41.617733
Basaltic scoria bomb along Al Birk/Muhayil Road. Olivine phyric, vesicular basalt, very fresh. Minor, large plagioclase phenocrysts. Groundmass is variably oxidized
259 m elevation.
N 18.297700
E 41.645117
Sample from basalt along Al Birk/Muhayil Road. Olivine phyric basalt. Sparse olivine is slightly discoloured. Weathered, hackly appearance of basalt outcrop surface. 279
m elevation.
N 18.421050
E 41.639717
Sample from basalt along Al Birk/Muhayil Road. Olivine phyric basalt. Olivine is altered and discoloured. Groundmass is reasonable fresh. Sample has sparse vesicles.
Flow surface is eroded and deflated, low relief cobble field. 267 m elevation.
N 18.089167
E 41.697683
Columnar basalt flow in Wadi Najla. Prominent colonnade in river channel. 55 m elevation. Basalt is altered, minor vesicles are lined, but dense groundmass is of
moderate quality. Olivine is discoloured and lightly altered. Basement outcrops of sandstone make up the north side of channel, which are in turn capped by basalts many
10’s of meters above the channel floor.
N 18.198300
E 41.572233
Sample from vesicular basaltic scoria bomb in cinder cone, approx.. 0.5 m diameter. Vesicular olivine basalt. Olivine is lightly discoloured. 81 m elevation. Deposit is
combination of scoria and spatter.
January/February 2017 Fieldwork at Wadi Dabsa
Appendix 3: List of Tufa Samples Collected in 2017
Sample No.
Samples from Wadi Dabsa basin environs
N 18.308746
E 41.556010
Within (but not in situ) a dry channel running E-W in the western sub-basin of Wadi Dabsa. WD1 is dense carbonate. WD2 contains wavy thin bands.
N 18.312167
E 41.540917
Extracted from a large intact boulder within E-W tufa ridge preserved in fan area. Massive tufa without banding and with some porosity.
N 18.311083
E 41.542611
Sampled at 1.6 m height in a south-facing section (6 m thick) into a large, near-horizontal-topped tufa cascade feature (140 m wide, 200 m long, 7 to 10 m thick) located
at the mouth of the fan area. Thin, densely cemented tufa, with thin bands sampled within a vugh (airspace).
N 18.306528
E 41.561139
Sampled from ~ 2m height within a south-facing section of tufa (~3 m thick), within a channel that skirts the southern edge of the Wadi Dabsa basin. Both samples are
densely-cemented and banded, sampled from the top of a unit of massive phytoherm framework tufa. The overall 3 m section contains repeated massive phytoherm tufa
and tufa-cemented gravel-to-boulder units.
N 18.307722
E 41.568111
A small former barrage a few hundred meters northwest of the 2015 L0106 grid. Sample removed from the cuspate vertical face. It is 5 cm thick and contains repeated sets
of narrow-bands, separated by phytoherm framework material.
N 18.306139
E 41.561472
Located in the south at the centre of the basin. Sampled from a 14 cm thick front of a former prograding barrage that has been eroded down to ground level. The narrow-
bands are orientated vertically and are slightly diffuse in places.
Samples associated with L0106/130 collection grids
~7 m W of L0130
Quadrant 5Jx
The outermost coating on the vertical front of Barrage 4, it is 1.5 cm thick and contains well-defined near-parallel narrow bands.
Quadrant 6Jz
Taken from the front-portion of an eroded barrage (the banding is orientated vertically, so that this sample represented the downward face of the barrage step). It is a 2.5
cm thick densely-cemented unit containing well-defined wavy narrow bands.
~10 m W of L0130
Quadrant 5Jx
~ 5 cm thick unit taken from the vertical front of the first of the complex at Barrage 6. It contains wavy narrow bands that become diffuse and grade into a phytoherm
L0106 Quadrant 5Dx
2 cm thick densely cemented unit with near-horizontal diffuse narrow bands, taken from the front edge of barrage
L0106 Quadrant 2Ay
~ 1 cm thick densely cemented unit, nearly crust/flowstone, coating the front edge of a heavily eroded barrage.
L0106 Quadrant 1Bz
~ 0.5 cm thick crust/flowstone, coating the front edge of a heavily eroded barrage.
... Middle Stone Age (MSA) in character, associated with extensive tufa deposits at Wadi Dabsa in the volcanic Harrat al Birk, Asir Province (Foulds et al., 2017;Inglis et al., 2015Inglis et al., , 2017 represents the richest Paleolithic locality recorded to date in southwestern Saudi Arabia (Sinclair et al., 2018), offering a rare chance to examine ESA and MSA activity at different scales within a landscape that may have proved persistently attractive to hominins over an extended period of time. ...
Surface artifacts dominate the archaeological record of arid landscapes, particularly the Saharo‐Arabian belt, a pivotal region in dispersals out of Africa. Discarded by hominins, these artifacts are key to understanding past landscape use and dispersals, yet behavioral interpretation of present‐day artifact distributions cannot be carried out without understanding how geomorphological processes have controlled, and continue to control, artifact preservation, exposure and visibility at multiple scales. We employ a geoarchaeological approach to unraveling the formation of a surface assemblage of 2,970 Palaeolithic and later lithic artifacts at Wadi Dabsa, Saudi Arabia, the richest locality recorded to date in the southwestern Red Sea coastal region. Wadi Dabsa basin, within the volcanic Harrat Al Birk, contains extensive tufa deposits formed during wetter conditions. We employ regional landscape mapping and automatic classification of surface conditions using satellite imagery, field observations, local landform mapping, archaeological survey, excavation, and sedimentological analyses to develop a multiscalar model of landscape evolution and geomorphological controls acting on artifact distributions in the basin. The main artifact assemblage is identified as a palimpsest of activity, actively forming on a deflating surface, a model with significant implications for future study and interpretation of this, and other, surface artifact assemblages.
... Most of this material is in the Arabian hinterland and much of it in the Arabian deserts, associated with periods of climate change when hydrological corridors with lakes or wetlands and grasslands were extensively distributed across the deserts of the Rub al 'Khali and the Nefud (Breeze et al. , 2016Petraglia et al., this volume). None of this material can be described as coastal except in the very broad sense that some of the sites are in coastal regions broadly defined, such as the recently discovered site of Wadi Dabsa in the Harrat Al Birk region of southwest Saudi Arabia (Foulds et al. 2017;Inglis et al. 2017;Sinclair et al. this volume), with an extensive palimpsest of material of many different periods in an inland basin about 10 km from the present-day coastline. ...
This chapter examines the different sources of evidence—phylogenetic, palaeoclimatic and archaeological—that have been used to investigate the hypothesis that early human dispersals from Africa during the late Pleistocene were facilitated by exploitation of marine resources and seafaring abilities and followed a predominantly coastal route including a crossing of the southern end of the Red Sea. We examine critically the current evidence and arguments for and against such a hypothesis and highlight the need for a more sophisticated understanding of the taphonomic factors that determine the formation, preservation and distribution of coastal archaeological deposits such as shell mounds. We present new data on the mid-Holocene shell mounds of the Farasan Islands and examine their spatial and temporal distribution in relation to a coastal environment that has been subject to rapid changes of sea level, geomorphology and ecological potential. We demonstrate that substantial shell mound deposits can accumulate rapidly over a matter of decades, even in a dynamic shoreline environment undergoing changes in relative sea level, that the ecological conditions that provide an abundant supply of marine molluscs as food are highly episodic in time and space, and that the resulting archaeological record is extremely patchy. We highlight the problem of dealing with negative evidence in the archaeological record and the need for a more detailed investigation and understanding of the various factors that determine the survival and visibility of archaeological deposits.
... Middle Stone Age (MSA) in character, associated with extensive tufa deposits at Wadi Dabsa in the volcanic Harrat al Birk, Asir Province (Foulds et al., 2017;Inglis et al., 2015Inglis et al., , 2017 represents the richest Paleolithic locality recorded to date in southwestern Saudi Arabia (Sinclair et al., 2018), offering a rare chance to examine ESA and MSA activity at different scales within a landscape that may have proved persistently attractive to hominins over an extended period of time. ...
Preprint download available via EarthArxiv at: Surface artefacts dominate the archaeological record of arid landscapes, particularly the Saharo-Arabian belt, a pivotal region in dispersals out of Africa. Discarded by hominins, these artefacts are key to understanding past landscape use and dispersals, yet behavioural interpretation of present-day artefact distributions cannot be carried out without understanding how geomorphological processes have controlled, and continue to control, artefact preservation, exposure and visibility at multiple scales.We employ a geoarchaeological approach to unravelling the formation of a surface assemblage of 2,970 Early and Middle Stone Age lithic artefacts at Wadi Dabsa, Saudi Arabia, the richest locality recorded to date in the southwestern Red Sea coastal region. Wadi Dabsa basin, within the volcanic Harrat Al Birk, contains extensive tufa deposits formed during wetter conditions.We employ regional landscape mapping and automatic classification of surface conditions using satellite imagery, field observations, local landform mapping, archaeological survey, excavation, and sedimentological analyses to develop a multi-scalar model of landscape evolution and geomorphological controls acting on artefact distributions in the basin. The main artefact assemblage is identified as a palimpsest of activity, actively forming on a deflatingsurface, a model with significant implications for future study and interpretation of this, and other, artefact surface assemblages.
Since 2012, a new phase of landscape survey for archaeological remains from the Palaeolithic has been undertaken in the provinces of Jizan and Asir in Southwestern Saudi Arabia. This is the first Palaeolithic landscape survey in this area since the Comprehensive Survey of the Kingdom undertaken between 1977 and 1982. More than 100 Palaeolithic sites have been identified from the Early Stone Age to the Late Stone Age, evidencing a regular association between archaeological remains and the Harrat deposits of basalt. The analysis of two major newly discovered sites, Dhahaban Quarry and Wadi Dabsa, has demonstrated the quality of archaeological and behavioural information that can still be recovered through landscape surveys in this region. At the site of Dhahaban Quarry, the survey has confirmed that Middle Stone Age lithic artefacts can be found in situ in the preserved beach deposits of ancient shorelines suggesting the use of marine resources. At Wadi Dabsa the technological study of a large assemblage of lithic artefacts suggests variations in expertise in lithic technology, and possibilities for understanding the process of learning the skills of lithic technology.
Full-text available
Throughout the greater Red Sea rift system the initial late Cenozoic syn-rift strata and extensional faulting are closely associated with alkali basaltic volcanism. Older stratigraphic units are either pre-rift or deposited during pre-rupture mechanical weakening of the lithosphere. The East African superplume appeared in northeast Africa ~46 Ma but was not accompanied by any significant extensional faulting. Continental rifting began in the eastern and central Gulf of Aden at ~31–30 Ma coeval with the onset of continental flood volcanism in northern Ethiopia, Eritrea, and western Yemen. Volcanism appeared soon after at Derudeb in southern Sudan and at Harrats Hadan and As Sirat in Saudi Arabia. From ~26.5 to 25 Ma a new phase of volcanism began with the intrusion of a dike field reaching southeast of Afar into the Ogaden. At 24–23 Ma dikes were emplaced nearly simultaneously north of Afar and reached over 2000 km into northern Egypt. The dike event linked Afar to the smaller Cairo mini-plume and corresponds to initiation of lithospheric extension and rupture in the central and northern Red Sea and Gulf of Suez. By ~21 Ma the dike intrusions along the entire length of the Red Sea were completed. Each episodic enlargement of the greater Red Sea rift system was triggered and facilitated by breakthrough of mantle-derived plumes. However, the absence of any volumetrically significant rift-related volcanism during the main phase of Miocene central and northern Red Sea – Gulf of Suez rifting supports the interpretation that plate–boundary forces likely drove overall separation of Arabia from Africa.
Full-text available
Since 2013, the authors have conducted archaeological surveys across the Al-Jawf province in northern Saudi Arabia. In the past two seasons, 48 sites were mapped and characterized by the presence of Levallois technology and, therefore, attributed to the Middle Paleolithic of Arabia. Preferential Levallois reduction using different methods of dorsal core preparation have been found at these sites. The technological variability includes Nubian Levallois methods, preferential Levallois with centripetal preparation, as well as recurrent centripetal reduction methods. In Arabia, sites with Nubian Levallois reduction are known from southern Oman, eastern Yemen, and central Saudi Arabia, while in Africa this reduction method has been identified across much of the northeastern continent. Preferential Levallois with centripetal preparation and recurrent centripetal Levallois methods have been found across Saudi Arabia, Yemen, Oman, and the United Arab Emirates. Outside of Arabia, these methods have been found in many regions across the Old World. In this paper, we present the results from technological analyses on the Middle Paleolithic assemblages from the newly discovered Al-Jawf sites. The technological data are used to place these sites into a wider regional framework, assessing whether connections with known lithic industries from across the Near East and northeastern Africa can be surmised.
This book presents the major tool types of European Lower and Middle Paleolithic. Building on the typelist of the late Francois Bordes, with many forms that have been recognized since, it presents working definitions of the types with illustrations and discussions of the variability inherent to lithic typologies. The authors combine classic typological views with current notions of lithic typological variation. This handbook represents not only an important reference source for gaining a practical understanding of how Lower and Middle Paleolithic typology is applied but of the nature of lithic variability in other kinds of assemblages as well. © 1994 University Museum University of Pennsylvania Philadelphia. All rights reserved.
This chapter provides a critical assessment of environment, landscape and resources in the Red Sea region over the past five million years in relation to archaeological evidence of hominin settlement, and of current hypotheses about the role of the region as a pathway or obstacle to population dispersals between Africa and Asia and the possible significance of coastal colonization. The discussion assesses the impact of factors such as topography and the distribution of resources on land and on the seacoast, taking account of geographical variation and changes in geology, sea levels and palaeoclimate. The merits of northern and southern routes of movement at either end of the Red Sea are compared. All the evidence indicates that there has been no land connection at the southern end since the beginning of the Pliocene period, but that short sea crossings would have been possible at lowest sea-level stands with little or no technical aids. More important than the possibilities of crossing the southern channel is the nature of the resources available in the adjacent coastal zones. There were many climatic episodes wetter than today, and during these periods water draining from the Arabian escarpment provided productive conditions for large mammals and human populations in coastal regions and eastwards into the desert. During drier episodes the coastal region would have provided important refugia both in upland areas and on the emerged shelves exposed by lowered sea level, especially in the southern sector and on both sides of the Red Sea. Marine resources may have offered an added advantage in coastal areas, but evidence for their exploitation is very limited, and their role has been over-exaggerated in hypotheses of coastal colonization.