Anlage D am Göbekli Tepe, 10. Jt. v. Chr.. Foto: Klaus Schmidt ^© Oktober 2009

Any interpretation of the radiocarbon dates of Göbekli Tepe can be only preliminary because the excavations are still in progress. 

The tell of Göbekli Tepe lies 15 km northeast of the modern city of Şanlıurfa in southeastern Anatolia. It is located at the highest point of the Germuş range overlooking the Harran plain (Schmidt 2006, 2011). The excavations conducted by Klaus Schmidt since 1995 have revealed no traces of domestic structures but rather a cult-architecture, the monumentality, the antiquity (early and middle Pre-Pottery Neolithic), and the iconographic richness of which have exceeded all expectations.  

The stratigraphy of the ruin hill includes an older layer (III) dated archaeologically to the PPNA and which produced monumental architecture with huge T-shaped pillars arranged in circle-like enclosures around two taller central pillars. A younger phase (layer II, early and middle PPNB) consists of smaller rectangular buildings often containing only two small central pillars or none at all. Layer I comprises mixed sediments derived from agricultural activities and containing material of the Middle Ages and of modern date, but no architectural remains. 

At least the big enclosures of layer III were intentionally and rapidly backfilled at the end of their period of use. This poses a severe problem for the radiocarbon dating of this layer, as the sediments could either be considerably younger than the enclosures, or roughly contemporary with them.  Given this basic problem, the first thing is to attempt to date the architecture directly.  Pedogenic carbonates have formed on the lower surfaces of the limestone of Göbeklis´ architecture, and this can be radiocarbon dated (Pustovoytov 2002, Pustovoytov 2006, Pustovoytov, Taubald 2003, Pustovoytov et al. 2007a, Pustovoytov et al. 2007b).  Unfortunately the pedogenic carbonate layers accumulate at a variable rate over a long period of time so a sample comprising a whole layer yields only an average value. This problem can be avoided by sampling only the oldest calcium carbonate layer in a thin section:  the result should be a date near the beginning of soil formation (burying) around the stone (Pustovoytov 2002). 

Radiocarbon data are available from the architectonic structures of layers III and II. The radiocarbon sample age sequence (Figure 1) confirms the archaeological strata, though absolute ages are clearly too young, with layer III being dated to the 9th millennium and layer II to the 8th or even 7th millennia. The data thus do not provide absolute chronological points of reference for the architecture and strata, but only a terminus ante quem for the refilling of the enclosures (layer III) and the abandonment of the site (layer II).  

Better suited for the direct dating of the architectonic structures is the wall plaster of the enclosures, which consists of loam that fortunately contains small amounts of charcoal (Figure 2; Dietrich and Schmidt 2010). A sample (KIA 44149) taken from the wall plaster of Enclosure D (Area L9-68, Loc. 782.3) gives a date of 9745-9314 calBC (95.4%), placing the circle in the PPNA. 

Two approaches were taken with the filling material. Carbonized plant remains have been scarce at Göbekli, thus limiting the possibilities for dating charcoal. Nevertheless three charcoal samples (Figure 2) are available for Enclosure A. The two samples Hd-20025 and Hd-20036 are derived from short-lived plants (pistacia sp. / amygdalus sp.) from the filling of Enclosure A (Kromer and Schmidt 1998). Both date to the late 10th / earliest 9th millennium. This fixes the filling material of Enclosure A firmly in the PPNA, though, as was explained above, the possibility of older material used for burying the structure can not be completely excluded. A third charcoal sample (KIA-28407) was found in Enclosure A under a fallen fragment of a pillar:  it dates the final refilling of the enclosure to approximately the middle of the 9th millennium. 

The archaeological appraisal of a recently acquired series of 20 data obtained from bone samples (Figure 3) is more complicated, as they pose some methodological problems (Dietrich 2011). At least within the group of samples chosen, collagen conservation is poor and the carbonate-rich sediments at Göbekli Tepe may cause problems with the dating of apatite fractions. Only one sample (KIA-38007) yielded sufficient collagen for an AMS dating:  it comes from Enclosure B and resulted in a date of 8306-8236 calBC (95.4%), which does not meet the expected use time of the monumental enclosures. Recent reviews on radiocarbon dates of bone samples seem to show that such dates are systematically younger than charcoal dates (Denaire 2009; Zazzo and Saliège in press).^[1] On the other hand it could be valid for the moment Enclosure B was finally buried, thus confirming KIA-28407. 

This general picture is further supported by a date measured on humic acids from a buried humus horizon (IGAS-2658). It gives the late 9th / early 8th millennium for layer II over the fill of Enclosure D. Another date on humic acids (KIA-28407) was measured on the earth sample which yielded the charcoal for KIA-28407; but its calibrated date (5559-5373 calBC, 94.6%) lies far outside the expected age range. 

In conclusion, charcoal samples suggest that the refilling and “burial” of the big enclosures began in the late 10th and early 9th millennium, while KIA 44149 from the wall plaster of enclosure D indicates building activities in the early PPNA. The final intrusions in the big enclosures are dated by KIA-28407 roughly to the middle of the 9th millennium, and this is supported by the data on pedogenic carbonates. 

The analysis of charcoal samples, which are now available in a considerable numbers from recent excavation work, will be pursued preferentially in the future. Also further analysis is necessary to show whether or not poor conservation of collagen is a general problem at the site.

Oliver Dietrich
Deutsches Archäologisches Institut
Orientabteilung

[1] A similar observation was made for the radiocarbon dates of Abu Hureyra.