Archaeological marine carbonates in northern Hokkaido, Japan : methodology, chronology and palaeothermometry

The site of Hamanaka 2, Rebun Island (Hokkaido), contains an archaeological record including well-preserved faunal remains, including clam shells and cod otoliths, from hunter-gatherer groups across the Epi-Jomon to Historical Ainu periods. As demonstrated by sclerochronologists worldwide, incremental stable isotope analysis of marine carbonates can provide an archaeologically-relevant and directly-datable proxy record of oceanic palaeotemperature, which is especially important for coastal sites such as Hamanaka 2. Despite this, marine carbonates from the site have not yet been used for this purpose, partly due to concerns over a variable regional marine reservoir effect. This, in combination with the possibility of 'old' shells being incorporated into younger stratigraphic units, means that any climatic information gained can be difficult to associate conclusively with the stratigraphic unit from which it was excavated. This thesis aims to address the methodological and interpretational obstacles relating to the archaeological use of this material at Hamanaka 2, and considers the local palaeotemperature record these materials can provide and their implications. The first paper examines the candidacy of the Sakhalin surf clam as a sub-annual temperature record, using stable isotope analysis of modern shells to compare carbonate-derived temperature estimates with observed water temperature. The results confirm that, as with most large, slow-growing marine bivalves, this species is a reliable recorder or water temperature, although seasonal extremes are likely to be dampened in the shell isotopic record. The second paper focuses on the complications arising from high resolution incremental sampling of marine carbonates, specifically the issue of contamination from embedding resin, and the relative merits of hand drilling versus micromilling. Comparison between two series of directly comparable samples shows that embedding may produce implausibly high δ13C values in the embedded, micromilled samples. This is interpreted as evidence of the contamination effect of resin on δ13C, whilst δ18O appears unaffected. The results indicate that hand drilling fast-growing species can provide a sub-annual palaeotemperature record comparable to micromilling, minimising sampling time and maximising sample size. The final paper addresses the main archaeological questions of the thesis. It assesses the palaeotemperature record of shells and otoliths from Hamanaka 2 over the Epi-Jomon-Final Okhotsk period in the context of new radiocarbon dates, which are paired with existing terrestrial dates. This allows calculation of new marine reservoir corrections, useful not only for this site but all others in the region which contain similar marine evidence. The integration of marine material into Bayesian chronological modelling at Hamanaka 2 constrains the timing of the Final Okhotsk period more tightly compared to previous models, showing a revised end date of 1107-958 cal BP. δ18O results show no clear temperature trend across the study period, although the single Early Okhotsk otolith sampled is consistent with the existing suggestion that this period saw a short-lived period of cooler temperatures. Overall, stable isotope analysis and radiocarbon dating combine to show the potential of marine carbonate material from Hamanaka 2 as a climatic and chronological record.