Fig 1.
Distribution map of the so-called transitional technocomplexes.
CP: Châtelperronian; LRJ: Lincombian-Ranisian-Jerzmanowician; ULU: Uluzzian. Base cartographic data obtained from the European Environment Agency (free of use, downloadable at https://www.eea.europa.eu/ds_resolveuid/070F2DAD-1AED-4B9B-950F-0047E5ADDF35. Rivers and bathimetry obtained from Natural Earth (free for use). DEM built up with QGIS 2.18. Infographics made with Inkscape.
Fig 2.
A) Digital Elevation Model showing the position of Aranbaltza site; B) Excavation plan, with the area of the excavated Châtelperronian deposits in red. Base cartographic data obtained from National Geographic Institute of Spain (IGN) under the Creative Commons License CC-BY 4.0. DEM built up with ArcGis 10.3. Infographics made with Inkscape.
Fig 3.
A) Detailed map of the excavated Châtelperronian deposits with the position of the sedimentological cores and the coordinated archaeological remains; B) North and east section of Area 3 excavation with the projected position of the archaeological materials and the positions of the pollen samples and the dated OSL samples.
Fig 4.
Stratigraphic section between Area 1 and 3 (section C) and synthetic stratigraphic section with the position of the dated OSL samples Stratigraphy.
A detail photograph of the sequence in corner C/D is presented in S6 Fig in S1 File.
Fig 5.
Châtelperronian lithic assemblage.
Aranbaltza II, US4b (Châtelperronian) archaeological materials. 1) Complete thin Châtelperronian point; 2–4) Proximal fragments of thin Châtelperronian points; 5) Thick Châtelperronian point, with the point made in the proximal part; 6) Distal fragment of wide Châtelperronian point; 7–10) Distal fragments of Châtelperronian points: #9 exhibits axial diagnostic impact fracture; 11–13) proximal fragments of backed blades, possible Châtelperronian points; 14) Distal fragment of backed blade; 15,16,20) Fragments of marginally backed blades; 17–19) Distal fragments of marginally backed points; 21–22) Marginally backed points; 23) Nucleiform burin; 24–28) Backed bladelets; 29) Bidirectional blade core for the production of pointed blades; 30–31) Unidirectional crested blades; 32) Endscraper on natural fragment; 33) Endscraper on blade; 34–35) Blades; 36) Overshot blade dragging an opposed platform; 37) Platform rejuvenation flake. Figure made by Aixa San Emeterio and Joseba Rios-Garaizar.
Fig 6.
Châtelperronian retouched tools.
Selected retouched tools from US4b. 1) Complete thin Châtelperronian point; 2) Thick Châtelperronian point, with the point made in the proximal part; 3) Distal fragment of wide Châtelperronian point; 4–7) Distal fragments of Châtelperronian points: #5 exhibits axial diagnostic impact fracture; 8–12) Proximal fragments of backed blades, possible Châtelperronian points; 10–12) Proximal fragments of Châtelperronian points; 13) Distal fragment of backed blade; 14–15) Marginally backed points; 16–18) Fragments of marginally backed blades; 19–21) Distal fragments of marginally backed points; 22–26) Backed bladelets: #24 presents typical Dufour ventral semi-abrupt retouch. Drawings made by Joseba Rios-Garaizar.
Fig 7.
Bayesian modelling results for the timing of the Châtelperronian culture in southwest Europe.
The modelled durations of the Châtelperronian culture (shown in grey) for sites that contain multiple dating results have been derived from site-specific Phase models using the Date command (see S39 Fig in S1 File for details). The site-specific modelled durations have been combined with the new OSL age for the Aranbaltza Châtelperronian culture (unmodelled and modelled OSL age distributions shown as light blue and dark blue, respectively) as part of the regional Châtelperronian Phase model shown here. The combined age range for the Châtelperronian culture of southwest Europe (“Duration Châtelperronian”) has been calculated from the modelled posterior probabilities of the start and end boundaries of this Phase model (also shown in grey) using the Date function. Bayesian modelling has been undertaken in OxCal v.4.4.4 [70], assuming each likelihood has a 5% prior probability of being an outlier within a general t-type outlier model [71]. The median ages and 1σ uncertainty ranges are shown for the modelled probability distributions. The 95.4% ranges of the highest posterior probabilities are also indicated by the broader horizontal bars underneath the probability density functions. The modelled durations are compared against the NGRIP GICC05 δ18O record, with interstadials (milder climatic periods) numbered accordingly.