[M-GI25] Holocene paleoenvironment, paleoclimate, and paleohazards in the Pacific Islands

[MGI25-02] Prehistoric settlement of a remote atoll, Pukapuka, and tropical cyclones

★Invited Papers

*Toru Yamaguchi¹, Hiroya Yamano² (1.Keio University, 2.National Institute for Environmental Studies)

Keywords:atoll, geoarchaeology, tropical cyclones, pit and swamp agriculture of taro, Northern Cook Islands, Polynesia

There are more than 170 atolls in the Pacific, and their low and flat landform of unconsolidated calcareous sand and gravel can support only a small terrestrial biota; moreover, many atolls lie far from volcanic high islands blessed with natural resources. Even in such a harsh atoll environment, Pacific islanders have lived for nearly 2000 years in Eastern Micronesia and at least 600 years in Polynesia. Wet taro cultivation would have been the most important element of their subsistence. Pukapuka is one such remote atoll in the northern Cook Islands, located between Samoa and Tahiti in Polynesia. Located on the eastern edge of the South Pacific Convergence Zone (SPCZ), Pukapuka has been subject to the occasional severe impacts of tropical cyclones (TCs) that have formed around Samoa since the early 20th century. Some TCs, especially under El Niño conditions, passed close to the atoll along the SPCZ and their high waves washed over the islets and flooded the wet taro fields. The long-lasting devastation of the food production system should have been caused by TCs during the prehistoric period. Therefore, we try to discuss data from our geoarchaeological surveys in the atoll in the context of prehistoric human settlement, exploitation of taro fields, and the physical impacts of and recovery from TCs. We will first look at coral sediments to consider the geomorphologic formation of an islet of Pukapuka, and then focus on coral samples attached to storm rocks washed onto the reef to estimate the timings of prehistoric TCs. Next, we will discuss trench excavations of banks surrounding taro fields. Because most remote atolls are generally unaffected by volcanic ash or alluvial deposits, except for wave-induced deposition, earlier cultural layers are often disturbed by later human activities. It is, therefore, difficult for archaeologists to have chronological control over excavated materials. However, we have found good stratigraphic conditions in the excavation of sand and gravel embankments dug up for the construction of taro fields, and the excavation of these embankments has yielded reliable ¹⁴C ages related to human activities. Finally, we will discuss the chronological correspondence between geoscientific data of prehistoric TCs and archaeological data of taro field construction, with reference to paleoclimatological data of ENSO activities reconstructed from Palmira coral delta¹⁸O records (Cobb, et al. 2003). The following is the summary of several analyses; (1) the oldest age of the islet was ca. 5500 yr BP, and its major part was formed by 2000 yr BP, (2) nine ¹⁴C ages from six storm rocks show five prehistoric cyclones at least, (3) archaeological excavations shows some pits constructed around 400 cal. yr. BP and probably re-excavated around 300 cal. yr. BP.