10:45 AM - 12:15 PM
[MIS16-P07] Paleotsunami deposits along the Pacific coast of Kyushu region
Keywords:Numerical tsunami simulations, Hyuga-nada
In the Hyuga-nada area, interplate earthquakes with Magnitude around 7 repeatedly occurred, whereas no earthquakes of Magnitude 8 or higher have been recorded in historical time. Since the oldest historical tsunami recorded in this region is associated with the 1662 CE Kanbun Hyuga-nada earthquake, it remains possible that this short 400-year record does not cover the full range of high-magnitude and low-frequency earthquakes. Geological research is therefore important to understand the occurrence of past earthquakes and tsunamis. In this presentation, we report the survey results at three coastal lowlands and the results of numerical tsunami simulations.
In Saiki City, Oita Pref, we obtained the eight sediment cores of 4.5–6.5 m long at two adjacent coastal marshes. The sand layer was observed at almost the same depth between the two marshes. This layer was interpreted as a tsunami deposit because it exhibited graded sub-layers, commonly observed in modern tsunami deposits. The depositional age of this layer is constrained to be 6350–6710 cal. yr BP. Although organic-rich mud is continuously deposited in these marshes, other event layers of possible tsunami deposits were not observed.
In Shimanoura Island, Miyazaki Pref., up to 2.7 m long sediment cores were collected at nine locations in a coastal marsh. The sediment cores are mainly composed of organic-rich mud, and a sandy mud layer is continuously observed on top of the cores. A landward flow may have deposited this sandy mud layer since this layer exhibits a landward fining trend. Other than this layer, no event layers were clearly observed in the marsh sequence, so there may be no tsunami record in this area for the past several thousand years. Numerical simulations showed no inundation for this marsh by the 1662 CE Kanbun Hyuga-nada and the 1707 CE Hoei earthquake models. On the other hand, all eleven models assuming the maximum-size megathrust earthquake (Mw 9.1) along Nankai Trough including Hyuga-nada resulted in widespread inundation for the marsh regardless of large-slip locations.
In Kushima City, Miyazaki Pref., the 4.6 ka tsunami deposit was identified at Nagata lowland (Yamada et al., 2020). In this study, we obtained up to 7 m long three sediment cores at Toi lowland, located 7 km east of Nagata lowland, to confirm the extent of the 4.6 ka tsunami deposit. A sand layer was recognized at almost the same depth as the Nagata lowland in the sediment cores, implying that the 4.6 ka tsunami inundated to these lowlands at least. Numerical simulations showed no inundation for these lowlands by the 1662 CE Kanbun Hyuga-nada, the 1707 CE Hoei, and the 1968 CE Hyuga-nada earthquake models. The eleven megathrust earthquake models resulted in widespread inundation for the Toi lowland, whereas the Nagata lowland was inundated by only two models whose large-slip area was set at Hyuga-nada.
Field surveys revealed that one or two probable tsunami deposits were identified northern part of the Hyuga-nada area (Saiki City and Shimanoura Island) and one in the southern part (Kushima City). In comparison with coastal areas facing Nankai Trough (e.g., Shimada et al., 2019), there is a significant difference in the number of tsunami deposits. In addition, given that tsunamis may not have inundated the coastal marsh in Shimanoura Island for the past several thousand years and that the megathrust earthquake model widely inundated the marsh, such a huge earthquake has probably not occurred in the past. As for the 4.6 ka tsunami deposit found in the southern part, there is no tsunami deposit close to this age in the northern part. Considering that the lowlands in Kushima City were not inundated by the historical earthquakes based on numerical simulations, the 4.6 ka deposit was associated with an unknown earthquake that occurred singularly in the southern part of Hyuga-nada area.
In Saiki City, Oita Pref, we obtained the eight sediment cores of 4.5–6.5 m long at two adjacent coastal marshes. The sand layer was observed at almost the same depth between the two marshes. This layer was interpreted as a tsunami deposit because it exhibited graded sub-layers, commonly observed in modern tsunami deposits. The depositional age of this layer is constrained to be 6350–6710 cal. yr BP. Although organic-rich mud is continuously deposited in these marshes, other event layers of possible tsunami deposits were not observed.
In Shimanoura Island, Miyazaki Pref., up to 2.7 m long sediment cores were collected at nine locations in a coastal marsh. The sediment cores are mainly composed of organic-rich mud, and a sandy mud layer is continuously observed on top of the cores. A landward flow may have deposited this sandy mud layer since this layer exhibits a landward fining trend. Other than this layer, no event layers were clearly observed in the marsh sequence, so there may be no tsunami record in this area for the past several thousand years. Numerical simulations showed no inundation for this marsh by the 1662 CE Kanbun Hyuga-nada and the 1707 CE Hoei earthquake models. On the other hand, all eleven models assuming the maximum-size megathrust earthquake (Mw 9.1) along Nankai Trough including Hyuga-nada resulted in widespread inundation for the marsh regardless of large-slip locations.
In Kushima City, Miyazaki Pref., the 4.6 ka tsunami deposit was identified at Nagata lowland (Yamada et al., 2020). In this study, we obtained up to 7 m long three sediment cores at Toi lowland, located 7 km east of Nagata lowland, to confirm the extent of the 4.6 ka tsunami deposit. A sand layer was recognized at almost the same depth as the Nagata lowland in the sediment cores, implying that the 4.6 ka tsunami inundated to these lowlands at least. Numerical simulations showed no inundation for these lowlands by the 1662 CE Kanbun Hyuga-nada, the 1707 CE Hoei, and the 1968 CE Hyuga-nada earthquake models. The eleven megathrust earthquake models resulted in widespread inundation for the Toi lowland, whereas the Nagata lowland was inundated by only two models whose large-slip area was set at Hyuga-nada.
Field surveys revealed that one or two probable tsunami deposits were identified northern part of the Hyuga-nada area (Saiki City and Shimanoura Island) and one in the southern part (Kushima City). In comparison with coastal areas facing Nankai Trough (e.g., Shimada et al., 2019), there is a significant difference in the number of tsunami deposits. In addition, given that tsunamis may not have inundated the coastal marsh in Shimanoura Island for the past several thousand years and that the megathrust earthquake model widely inundated the marsh, such a huge earthquake has probably not occurred in the past. As for the 4.6 ka tsunami deposit found in the southern part, there is no tsunami deposit close to this age in the northern part. Considering that the lowlands in Kushima City were not inundated by the historical earthquakes based on numerical simulations, the 4.6 ka deposit was associated with an unknown earthquake that occurred singularly in the southern part of Hyuga-nada area.