Radiocarbon Reservoir Effect

During a rescue operation in the upper basin of the Loa River (northern Chile), archaeologists discovered a prehispanic burial
of a newborn associated with various offerings and the partial skeletal remains of a teenager. We conducted stable isotope
analysis to establish palaeodiet (δ13C and δ15N) and mobility (δ18O and 87Sr/86Sr). In addition, we carried out paired radiocarbon
dating of bioanthropological and cultural remains in order to date the burial event and evaluate a possible marine
reservoir effect. The results of these analyses suggest that (1) the individuals’ diets were composed primarily of terrestrial
resources, supplemented by moderate consumption of marine resources; and (2) that the individuals studied were of isotopically
local origin. Radiocarbon dates showed an unexpected difference between the ages of the offerings and the human
remains, likely due to the marine reservoir effect. These results shed new light on the consumption of marine foods during
the Early Formative in the region. At the same time, this study is an example of the importance of carrying out paired
dates of human/cultural remains on burials.

The prehistoric site of Palamari – located on the northeastern coast of Skyros (38˚ 57΄ 52΄΄ N Lat., 24˚ 30΄ 35΄΄ E. Long) – is a fortified settlement of the Early and Middle Bronze Age (3rd and 2ndmillenniumBC) that has been excavated since 1981 (Parlama 2007, 2009; Parlama et al. 2010; in press). During the excavation, pieces of marine and terrestrial materials suitable for radiocarbon dating were unearthed. This paper reports on the absolute dating of the various inhabitation phases of the site and the calculation of the “marine reservoir effect” for this region of the Aegean by dating contemporaneous pairs of samples consisting of marine mollusc shells-charcoals/or animal bones.
This work represents a case study of a larger-scale project whose objective is to establish a marine reservoir correction curve for the entire Aegean Sea region. This curve will have wide application to the dating of archaeological strata throughout the region and will be particularly useful where marine shells are the only datable material for a site. Although the project emphasizes archaeological applications, the results will be of importance also to palaeo-climatic and palaeo-oceanographic studies.
In total 42 samples of terrestrial andmarine origin were collected during two in situ sampling campaigns, in December 2000 and October 2005. During the first campaign 6 samples
were selected and dated at the radiocarbon unit of the laboratory of Archaeometry of NCSR “Demokritos” (Facorellis 1996; Facorellis et al. 1998),whereas during the second one, 36 samples, were dated at the Keck Carbon Cycle AMS facility, University of California at Irvine. The latter set of 36 samples were analyzed for species identification prior to dating (Table 1). The sea shells belonged to the following species: Monodonta turbinata (Von BORN, 1778), Patella
rustica (L., 1758), Patella caerulea (L., 1758), Glycymeris sp., Arca noae (L.,1758), Stramonita haemastoma (L., 1766), Monodonta articulata (LMCK, 1822), Patella ulyssiponensis
(GMELIN, 1791),Hexaplex trunculus (L., 1758) and OperculumBolma rugosa (L., 1767) (Sabelli et al. 1990, SoHelME 2005, Delamotte and Vardala-Theodorou 2007, SoHelFI 2007).
The charcoals belong to the following species: Quercus type evergreen, Olea europaea, cf. Erica sp. and Angiosperm (Schweingruber 1990). The animals bones where fragments from different skeletal parts of Caprinae and one part from Bovidae and stable isotope analyses (δ15N and δ13C: Table 2) indicate that their diet was based on C3 terrestrial plants (Papathanasiou 2000).
The results of the radiocarbon dating are presented in Table 1 and are listed from later to earlier. The conventional radiocarbon ages were converted to calendrical dates based on the international calibration curves IntCal09 for the charcoal samples andMarine09 for the shell samples (Reimer et al. 2009) and using the calibration programCALIB Rev. 6.0.1. (Stuiver and
Reimer 1993). In total 13marine reservoir correction values ΔR (Table 3) were determined. A detailed description of the calculation of the ΔR value can be found in the literature (Stuiver et al. 1986, Stuiver and Braziunas 1993); Facorellis et al. 1998; Facorellis 2011).
Figure 1 shows the fluctuation of ΔR values versus time with a third degree polynomial fit. The variationsmay be due to changes in climate (temperature, freshwater input, etc.) over this period and are probably associated with changes in themovement of sea watermasses in this region of the Aegean. It is worth noting that Skyros lies in the region where the cold, low-salinity sea watermasses coming fromthe Black Sea through the straits of the Hellespont, meet with the warm, high salinitymasses fromthe southern Aegean Sea (Aksu et al. 1995). A more detailed discussion of this variation will be reported elsewhere.
Figure 2 depicts the calibrated calendrical dates in years BC of charcoal samples and animal bones (black bars), and of marine mollusc shells (gray bars) for standard deviations 1σ (closed bars) and 2σ (open bars). These results indicate that Palamari on Skyros was inhabited for about 1200 years (ca 2900-1700 BC), i.e. from PE I till theME II period.
In conclusion, the analysis of samples of terrestrial and marine origin from Palamari has provided information about the plant and animal species used by the residents of the settlement.
The combined absolute dating with the 14C method has allowed firstly the precise determination of the inhabitation, which was found to be between 2900-1700 BC and secondly
the variation of the local ΔR. The variation over this period can now be used for a more accurate and reliable calibration of the conventional radiocarbon dates of marine samples from
the same region and time interval.
The authors wish to thank the archaeologists and directors of the excavation of Palamari site, Mrs. Liana Parlama, Maria Theocharis and Elissavet Chatzipouliou-Kalliri who entrusted
us the study of these samples, as well as the archaeologists, Mrs. Christina Romanou and Mr. Stamatis Bonatsos for providing substantial help and information. We greatly acknowledge the Institute for Aegean Prehistory (INSTAP) and the Curtiss T. & Mary G. Brennan Foundation for funding the first and the second sampling campaigns and samples analyses,
respectively. We also thank Ms. Rachel Moore and Drs. Guarciara dos Santos and Dachun Zhang at the Keck Carbon Cycle AMS facility for their help and stimulating discussions.

ABSTRACT
Objectives

The island cemetery site of Ostorf (Germany) consists of individual human graves containing Funnel Beaker ceramics dating to the Early or Middle Neolithic. However, previous isotope and radiocarbon analysis demonstrated that the Ostorf individuals had a diet rich in freshwater fish. The present study was undertaken to quantitatively reconstruct the diet of the Ostorf population and establish if dietary habits are consistent with the traditional characterization of a Neolithic diet.
Methods

Quantitative diet reconstruction was achieved through a novel approach consisting of the use of the Bayesian mixing model Food Reconstruction Using Isotopic Transferred Signals (FRUITS) to model isotope measurements from multiple dietary proxies (δ13Ccollagen, δ15Ncollagen, δ13Cbioapatite, δ34Smethione, 14Ccollagen). The accuracy of model estimates was verified by comparing the agreement between observed and estimated human dietary radiocarbon reservoir effects.
Results

Quantitative diet reconstruction estimates confirm that the Ostorf individuals had a high protein intake due to the consumption of fish and terrestrial animal products. However, FRUITS estimates also show that plant foods represented a significant source of calories. Observed and estimated human dietary radiocarbon reservoir effects are in good agreement provided that the aquatic reservoir effect at Lake Ostorf is taken as reference.
Conclusions

The Ostorf population apparently adopted elements associated with a Neolithic culture but adapted to available local food resources and implemented a subsistence strategy that involved a large proportion of fish and terrestrial meat consumption. This case study exemplifies the diversity of subsistence strategies followed during the Neolithic.

A review of current research reveals multiple lines of evidence suggesting that no single freshwater reservoir offset (FRO) correction can be applied to accelerator mass spectrometer (AMS) ages obtained on carbonized food residue from cooking vessels. Systematically evaluating the regional presence, magnitude, and effects of a freshwater reservoir effect (FRE) is a demonstrably difficult analytic problem given the variation of ancient carbon reservoirs in both space and time within water bodies, and which should be performed in advance of AMS assays. In coastal and estuarine contexts, a priori partitioning FRE from known marine reservoir effects (MRE) is also necessary to eliminate potential mixed effects. Likewise, any FRE varies based on the proportional mix of resources producing the residues and the ancient carbon uptake of those products. Processing techniques are a significant component of assessing potential FRE, and each pot/cooking vessel is therefore an independent context requiring analytic evaluation. In northeastern North America, there is little ethnohistoric/ ethnographic evidence for fish boiling/stewing in ceramic cooking vessels; rather, fish were more often dried, smoked, or cooked for immediate consumption on open fires. Assays of fatty acids extracted from prehistoric vessel fabrics even on known fishing sites reveals no evidence for fish in the food mix. These observations suggest that the likelihoods of FRE in carbonized food residue in northeastern North America is therefore low, and that assays potentially suffering from FRO are minimal. In turn, this suggests that AMS ages from carbonized food residues are reliable unless analytically demonstrated otherwise for specific cases, and should take primacy over ages on other associated materials that have historically been employed for critical threshold chronological events.

Accelerator Mass Spectrometry (AMS) is increasingly employed to date encrusted carbonized food residues on prehistoric pottery sherds, particularly in regions where other datable material is absent or scarce, or where such materials lack good association with objects of chronological interest. The accuracy of AMS dating of residues has recently been questioned in Europe and North America, with skepticism directly or indirectly attributed to the presence of carbonate rich bedrock, a freshwater carbonate reservoir resulting from association wiui such bedrock or the cooking of aquatic resources such as mollusk or fish engaged in the uptake of older carbonates. It is argued that carbonized food residues from such contexts or resources are expected to display older apparent ages than dates on other materials. We evaluate this proposition through hypotheses assessing the accuracy and internal consistency of a broad range of data from the northeastern and midwestern United States. We statistically assess dates from 25 site components where either both dates on residue and context dates on other materials have been obtained. Of the 70 dates on residue tested, only 5.7 percent are considered inconsistent with their respective context dates. We also test for consistency 46 dates on residues from 14 site components lacking context dates. Of these, only 4.5 percent are considered internally inconsistent with expectations at the site or regional level. At present, we conclude that AMS dating of residues provides results consistent with those of other datable materials.

Fischer and Heinemeier (2003) present a hypothesis that the freshwater reservoir effect produces old apparent ages for radiocarbon dates run on charred cooking residues in regions where fossil carbon is present in groundwater. The hypothesis is based in part on their analysis of dates on charred cooking residues from 3 inland archaeological sites in Denmark in relation to contextual dates from those sites on other materials. A critical assessment of the dates from these sites suggests that rather than a pattern of old apparent dates, there is a single outlying date—not sufficient evidence on which to build a case for the freshwater reservoir effect.

Obtaining radiocarbon assays on objects of chronological interest is always preferable to obtaining assays on spatially associated charcoal. The development of Accelerator Mass Spectrometer (AMS) dating has expanded the number of objects that can be directly assayed because it requires only a few milligrams of material. Pottery can be directly assayed when charred cooking residues adhering to the interior walls are present. The accuracy of AMS ages derived from residues has been questioned in cases where cooking freshwater aquatic organisms may have introduced carbon from ancient carbon reservoirs into residues. Here we provide analytic protocols for examination of this phenomenon and the results of systematic modeling of age estimates on residues formed from fish and maize with varying percentages of dead carbon. We present a regional case study using a large series of AMS age estimates on residues from the Finger Lakes region of northeastern United States to demonstrate how the paleolimnological record and lipid analysis of residues can help to determine if dates on residues from a given region are likely to have been affected by the presence of ancient carbon. In the case of the Finger Lakes, there is no evidence that ancient carbon affected the age estimates.""

Ancient carbon reservoirs in freshwater bodies have the potential to introduce ancient carbon into charred cooking residues adhering to pottery wall interiors when aquatic organisms are parts of cooked resource mixes. This ancient carbon results in old apparent ages when these cooking residues are subjected to accelerator mass spectrometry (AMS) dating, the so-called freshwater reservoir effect (FRE). Roper’s (2013) assessment of the FRE on 14C ages from cooking residue in the Central Plains is only the second such peer-reviewed regional assessment in eastern North America. Roper suggests that 13 of 23 14C ages on residue are too old as a result of ancient carbon from fish or leached from shell temper or old carbon introduced via maize nixtamalization. Here, we reassess Roper’s (2013) data set of 14C ages on cooking residues and annual plants and conclude that she is mistaken in her assessment of the accuracies of 14C ages from residues. This outcome is placed in the context of the larger FRE literature.