OSL dating of the AD 869 Jogan tsunami deposit, northeastern Japan

Abstract

Sediments deposited by the AD 869 Jogan tsunami offer an opportunity to test the reliability of optically stimulated luminescence (OSL) dating of relatively old historical tsunami deposits. We collected a geoslicer sample from sand deposited on the Sendai Plain, northeastern Japan, by the Jogan tsunami and applied quartz OSL dating to it. We then compared the OSL ages with the known age of the tsunami event. In ascending order, the sedimentary sequence in the geoslicer sample consists of the beach–dune sand, lower peat, Jogan tsunami deposit, upper peat, pre-2011 paddy soil, and the 2011 tsunami deposit. To obtain equivalent dose (De,bulk), a standard single-aliquot renegerative-dose (SAR) protocol was applied to large aliquots of the 180–250 μm fraction of two samples from the beach–dune sand, and four samples from differing levels of the Jogan tsunami deposit. The OSL decay curves were dominated by the medium component; thus, for two samples from the Jogan deposit the fast-component OSL signal was isolated and used to determine the equivalent dose (De,fast). Using De,bulk, OSL ages of the tsunami deposit were underestimated by ∼40%, and even the beach–dune sand was dated younger than AD 869. In contrast, De,fast provided a robust age estimate with only slight underestimation. A pulse annealing test showed that the bulk and medium-component OSL signals were thermally unstable. The medium component in the natural OSL was clearly truncated in comparison to the regenerated OSL; the medium component is thus considered to be the main cause of the underestimated ages. Similar effects of a dominant medium-component OSL have been reported in tectonically active regions, which are also prone to tsunamis. The effect of this dominance should be carefully considered in quartz OSL dating of tsunami deposits.