- Email: [email protected]
Re-estimated fault model of the 17th century great earthquake off Hokkaido using tsunami deposit data
Earth and Planetary Science Letters 433 (2016) 133–138
Contents lists available at ScienceDirect
Earth and Planetary Science Letters www.elsevier.com/locate/epsl
Re-estimated fault model of the 17th century great earthquake off Hokkaido using tsunami deposit data Kei Ioki ∗ , Yuichiro Tanioka Institute of Seismology and Volcanology, Hokkaido University, Japan
a r t i c l e
i n f o
Article history: Received 12 May 2015 Received in revised form 1 October 2015 Accepted 1 October 2015 Available online 7 November 2015 Editor: P. Shearer Keywords: tsunami great earthquake Hokkaido Kurile trench
a b s t r a c t Paleotsunami researches revealed that a great earthquake occurred off eastern Hokkaido, Japan and generated a large tsunami in the 17th century. Tsunami deposits from this event have been found at far inland from the Pacific coast in eastern Hokkaido. Previous study estimated the fault model of the 17th century great earthquake by comparing locations of lowland tsunami deposits and computed tsunami inundation areas. Tsunami deposits were also traced at high cliff near the coast as high as 18 m above the sea level. Recent paleotsunami study also traced tsunami deposits at other high cliffs along the Pacific coast. The fault model estimated from previous study cannot explain the tsunami deposit data at high cliffs near the coast. In this study, we estimated the fault model of the 17th century great earthquake to explain both lowland widespread tsunami deposit areas and tsunami deposit data at high cliffs near the coast. We found that distributions of lowland tsunami deposits were mainly explained by wide rupture area at the plate interface in Tokachi-Oki segment and Nemuro-Oki segment. Tsunami deposits at high cliff near the coast were mainly explained by very large slip of 25 m at the shallow part of the plate interface near the trench in those segments. The total seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The 2011 great Tohoku earthquake ruptured large area off Tohoku and very large slip amount was found at the shallow part of the plate interface near the trench. The 17th century great earthquake had the same characteristics as the 2011 great Tohoku earthquake. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Historically many great underthrust earthquakes occurred off Hokkaido in the Pacific Ocean. In 1843, 1952, and 2003, the Tokachi-Oki earthquakes occurred off eastern Hokkaido (Hatori, 1984; Hirata et al., 2003; Tanioka et al., 2004, respectively). In 1894 and 1973, the Nemuro-Oki earthquakes occurred off eastern Hokkaido at northeast of Tokachi-Oki earthquakes (Tanioka et al., 2007) (Fig. 1). In the southernmost Kuril Trench, large earthquakes have variability in spatial extent in both along-trench direction and trench-normal direction (Hirata et al., 2009). Tsunami deposits due to prehistoric tsunamis have been found along the coast of eastern Hokkaido (Sawai, 2002; Nanayama et al., 2003, 2007). These tsunami deposits were found at far inland from the coast where tsunamis generated by above historic earthquakes did not reach. The elevations of the locations where tsunami deposits were found at cliffs near the coast were also much higher than estimated tsunami heights by historical earth-
*
Corresponding author. E-mail address: [email protected] (K. Ioki).
http://dx.doi.org/10.1016/j.epsl.2015.10.009 0012-821X/© 2015 Elsevier B.V. All rights reserved.
quakes (Hirakawa et al., 2005). Nanayama et al. (2003) found that the latest event which left those tsunami deposits occurred in early 17th century. Satake et al. (2005) found that the 17th century great earthquake ruptured the plate interface at Tokachi-Oki, Akkeshi-Oki and Nemuro-Oki segments in Fig. 1 from tsunami simulation. Furthermore, Satake et al. (2008) estimated the fault model which reproduce the distribution of lowland tsunami deposits at five areas. However, the elevations of the locations where tsunami deposits were found near the coast as high as 18 m above the sea level (Hirakawa et al., 2005) were not completely explained by this fault model. In this paper, we estimate the fault model of the 17th century great earthquake to explain both locations of lowland tsunami deposits and highland elevations where tsunami deposits were found near the coast. 2. Previous studies of the 17th century tsunami Several marsh and lakes exist along the coast of eastern Hokkaido, so peat layers are common in those places. Tsunami deposits and volcanic ash layers are also seen within those peat layers. Nanayama et al. (2007) found 9 prehistoric tsunami sand beds in those peat layers among the past about 4000 yrs.
134
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
3. Data 3.1. Tsunami deposit data Lowland tsunami deposit data used by Satake et al. (2008) at Oikamanai pond in Taiki town, Pashukuru pond in Kushiro city, Tokotan pond in Akkeshi town, Kiritappu marsh in Hamanaka town, and Nambu pond in Nemuro city are used in this study. New lowland tsunami deposit data found by Nakamura et al. (2011, 2012) at Kinashibetsu marsh and Onbetsu in Kushiro city, and at Fureshima marsh in Nemuro city are also used in this study. Lowland tsunami deposit data found by Hirakawa et al. (2000b) at Rekifune river in Taiki town, Yudou pond and Choubushi lake in Toyokoro town are also used in this study. The locations of those ponds and marsh are shown in Fig. 1. The eight elevations of tsunami deposits surveyed at highland cliffs near the coast by Hirakawa et al. (2005) are also used as data. 3.2. Fault model parameter
Fig. 1. The fault model of the 17th century great earthquake is shown (T: TokachiOki segment, N: Nemuro-Oki segment, and S: Shallow part of the plate interface). Space–time diagram of great earthquakes along the southern Kurile trench from the 17th through 21st centuries is shown (modified from Satake et al., 2005). Green rectangles show the slip distribution of the 1952 Tokachi-Oki earthquake (Hirata et al., 2003) and purple rectangles show the slip distribution of the 1973 NemuroOki earthquake (Tanioka et al., 2007). Black rectangles show the slip distribution of the 2011 Tohoku earthquake (Gusman et al., 2012). Black dots show locations where tsunami inundation area and tsunami heights were calculated; (R: Rekifune, O: Oikamanai, Y: Yudou, C: Choubushi, K: Kinashibetsu, O: Onbetsu, P: Pashukuru, T: Tokotan, K: Kiritappu, F: Fureshima, N: Nemuro-nambu). The depth contour interval is 1000 m.
Sawai et al. (2009) also found sandy deposits by 15 tsunamis in about 6000 yrs in those places. The recurrence interval of those events is about 400∼600 yrs (Hirakawa et al., 2000a). The latest event occurred early 17th century because the latest tsunami deposits can be seen just under the volcanic ash caused by the 1667 Tarumae eruption (Nanayama et al., 2007). Those tsunami deposits are found about 1∼4 km inland from the coast of eastern Hokkaido. At Kiritappu marsh, tsunami deposits can be seen far inland from observed tsunami inundation area by the 1952 Tokachi-Oki earthquake (Satake et al., 2005). On the other hand, tsunami deposits were also observed at high cliffs near the coast (Hirakawa et al., 2000b, 2005; Hirakawa, 2006). The elevations of the locations where tsunami deposits were found are more than 10 m. At Oikamanai pond in Taiki town, the elevation was 18 m at high cliff near the coast (Hirakawa et al., 2005). At Fureshima marsh, the elevation of the locations where tsunami deposit was found was 15 m (Nakamura et al., 2011). Observed tsunami heights by historical earthquakes that occurred off Hokkaido were less than 4 m from Erimo Cape to Kushiro and less than 7 m from Kushiro to Nosappu Cape along the coast of eastern Hokkaido. The prehistoric tsunami of the 17th century were much larger than those historical tsunamis. Satake et al. (2008) estimated the fault model of the 17th century great earthquake by comparing locations of lowland tsunami deposits and computed tsunami inundations at five areas. The estimated slip amounts were 10 m in Tokachi-Oki segment and 5 m in Nemuro-Oki segment which includes Akkeshi-Oki segment in Fig. 1. The total seismic moment was calculated to be 0.8 × 1022 N m (Mw 8.5) with the rigidity of 4 × 1010 N/m2 . However, the tsunami heights along the coast of eastern Hokkaido calculated by this fault model (Satake et al., 2008) are smaller than the elevations of the locations where tsunami deposits were found near the coast in Tokachi (Hirakawa et al., 2005).
As we have already discussed before, the fault models of Tokachi-Oki segment and Nemuro-Oki segment are not enough to explain the elevations of the locations where tsunami deposits were found at high cliffs near the coast by Hirakawa et al. (2005). We need to find the fault model that can explain both lowland tsunami deposit data and the elevations of the locations where tsunami deposits were found near the coast. Satake et al. (2008) tested tsunami earthquake model which rupture the shallow part of the plate interface. The tsunami earthquake model yielded locally variable coastal heights on the coast of eastern Hokkaido and yielded little tsunami inundation to lowland marshes where tsunami deposits were found. Therefore, we hypothesize that both widespread inland tsunami inundations and locally large tsunami heights near the coast are potentially explained by the fault model which rupture not only the plate interface at Tokachi-Oki segment and Nemuro-Oki segment but also the shallow part of the plate interface near the trench. This hypothesis is come from the similarity of the 2011 great Tohoku earthquake (Mw 9.1) occurred along the Japan Trench and the 17th century great earthquake. Both tsunamis of the 17th century great earthquake and 2011 great Tohoku earthquake have characteristics that high tsunami heights were observed along the coast and large inundation areas were spread over wide areas. Slip distributions of the 2011 great Tohoku earthquake were estimated by several previous studies using seismic waves (Hayes, 2011), GPS data (Pollitz et al., 2011) and tsunami waveform data (Fujii et al., 2011). Gusman et al. (2012) also estimated the slip distribution of the earthquake using tsunami waveforms, GPS data, and seafloor crustal deformation data. The result shows that the largest slip amount of about 44 m ruptured the plate interface near the trench (Fig. 1). Satake et al. (2013) also estimated the slip distribution of the earthquake shown the huge slip on the shallow part of the plate interface near the trench. The characteristics of very large slip at the plate interface near the trench is common in all of the above previous studies. The largest slip near the trench was one of reasons to generate devastating tsunami along the coast in Tohoku. Before the 2011 great Tohoku earthquake occurred, it was believed that the shallow part of the plate interface near the trench is aseismic or rupture with a slow velocity. However, the 2011 great Tohoku earthquake indicates the plate interface can be ruptured with very large slip amount. Therefore, the fault model of the 17th century great earthquake can be similar to that of the 2011 great Tohoku earthquake. We assumed that the 17th century great earthquake ruptured the shallow part of the plate interface near the Kurile trench with a large slip amount, similar to the 2011 great Tohoku earthquake.
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Additional fault model, S model, was allocated at the shallow part of the plate interface near the Kurile trench to explain high tsunami heights along the coast of eastern Hokkaido (Fig. 1). Fault parameters of S model have length of 300 km, width of 30 km, and shallowest depth of the fault of 6.7 km. Slip amount of S model was changed to 15 m, 25 m, and 35 m. The mechanism of the fault model of the 17th century great earthquake has strike of 228◦ (parallel to the Kurile trench), dip of 15◦ (average angle of subducting Pacific plate in the Kurile trench), and rake of 90◦ (pure thrust earthquake). The fault length, fault width, and slip amount of the Tokachi-Oki segments and Nemuro-Oki segments in Fig. 1 are the same as those used by Satake et al. (2008). Those fault parameters of the Tokachi-Oki segment are length = 100 km, width = 100 km, and slip = 10 m. Those fault parameters of the Nemuro-Oki segment are length = 200 km, width = 100 km, and slip = 5 m. The shallowest depth of the fault of 14 km was used in both segments. Parameters of the fault model of the 17th century great earthquake used in the tsunami numerical simulation are summarized in Table 1. 4. Tsunami numerical simulation The bathymetry used for tsunami numerical simulation was created by interpolating M7000 digital bathymetric chart published by Japan Hydrographic Association. The topography was created by interpolating SRTM 90 m Digital Elevation Data published by Table 1 Fault models and their parameters of the 17th century great earthquake.
135
CSI-CGIAR and Digital map 50 m Grid published by Geospatial Information Authority of Japan. A coarse grid size for tsunami numerical simulation that covers all computational area, 142◦ –149◦ E and 39◦ –45◦ N, is 10 arcsecond. A fine grid size for tsunami inundation simulation near the coast of eastern Hokkaido is 1 arc-second. To compute tsunami, first, coseismic deformation on the seafloor was computed using Okada’s formula (Okada, 1985). Crosssections of the calculated ocean bottom deformations are shown in Fig. 2. We assumed that the initial water deformation on the sea surface is equal to the coseismic deformation on the seafloor. Linier long-wave equations were solved to compute tsunami in the coarse grid system. A finite-difference scheme with a staggered grid system was used to solve the equations (Satake, 2007). The open boundary condition was used at the end of the computed area. The time step of 1 s was chosen to satisfy a stability condition. Non-linear long-wave equations were solved to compute tsunami inundation (Imamura, 1996) in the fine grid system. Tsunami heights computed using linear long-wave equations are interpolated at the boundary of fine grid area. These interpolated tsunami waveforms are from the boundary into the fine grid system where tsunami inundation were computed using non-linear long-wave equations (Goto et al., 1997). To satisfy a stability condition, the time step was chosen to be 0.05 s. Manning’s roughness coefficient of 0.03 m−1/3 s was assumed in this simulation. The coseismic deformation was included into the bathymetry data in computed area. 5. Results
Fault model
Length (km)
Width (km)
Depth (km)
Strike (deg)
Dip (deg)
Rake (deg)
Slip (m)
T N S
100 200 300
100 100 30
14 14 6 .7
228 228 228
15 15 15
90 90 90
10 5 0–35
Slip amount of S model that located the plate interface were estimated by comparing tsunami deposits were found and computed areas and also by comparing the elevations
shallow part of the the locations where tsunami inundation where tsunami de-
Fig. 2. Cross-sections of ocean bottom deformation along the line in A through A computed from the fault model of the 17th century great earthquake (red line: without S model; blue line: with S model) and along the line B through B computed from the fault model of the 2011 Tohoku earthquake (Gusman et al., 2012).
136
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Table 2 Comparisons of measured and calculated tsunami heights (m) with changing slip amounts of S model. Measureda Rekifune river (R1) Rekifune river (R2) Rekifune river (R3) Oikamanai pond (O1) Oikamanai pond (O2) Yudou pond (Y1) Yudou pond (Y2) Kinashibetsu marsh (K1) Fureshima marsh (F1) a
15
>12 16 14 18 >10 15 12 15
Without S model
Slip = 15 (m)
Slip = 25 (m)
Slip = 35 (m)
5 7 6 11 11 2 12 11 7
5 7 11 12 17 3 16 13 13
6 11 17 16 24 6 20 15 17
8 13 21 18 29 9 27 17 21
The elevations of the locations where tsunami deposits were found (Hirakawa et al., 2005; Nakamura et al., 2011).
Fig. 3. (a) Computed tsunami inundation and heights estimated from the fault model of the 17th century great earthquake at Oikamanai pond (left side: without S model, right side: with S model). Red circles describe the locations where tsunami deposits of the 17th century great earthquake were found by Satake et al. (2008) and Hirakawa et al. (2005). Black crosses describe the locations where tsunami deposits were not found. The numbers show computed tsunami heights and the numbers in the bracket show the elevations of the locations where tsunami deposits were found. (b) Comparison of computed coastal tsunami heights (red: without S model, blue: with S model). Black rectangles show the elevations of the locations where tsunami deposits were found.
posits were found and computed tsunami heights near the coast of eastern Hokkaido. Computed tsunami heights at highland near the coast with slip amount of 0 m (without S model), 15 m, 25 m, and 35 m in S model are shown in Table 2. Computed tsunami heights without S model and with S model having a slip amount of 15 m were much smaller than the measured tsunami heights. Computed tsunami heights with S model having a slip amount of 25 m and 35 m explain most of the elevations of the locations where tsunami deposits were found. The elevations of the locations where tsunami deposits were found indicate the minimum tsunami height at the location. Therefore, computed tsunami heights should be larger than the elevations of the locations where tsunami deposits were found. As a result, at least, slip amount of 25 m in S model was needed to explain measured tsunami heights at high cliff. Computed tsunami inundation areas of lowland and computed tsunami heights of highland at eleven areas are shown in Fig. 3 and Fig. S1. Computed tsunami inundation areas should cover the locations where tsunami deposits were found. Computed tsunami inundation areas with S model having a slip amount of 25 m best fit to the locations where tsunami deposits were found. Slip amount of S model was chosen to fit the locations where tsunami deposits were not found, black crosses in Fig. 3 and Fig. S1, and the end of computed tsunami inundation areas in this study. Namegaya and Satake (2014) indicate that the tsunami inundation are extend farther inland than the sandy tsunami deposits and the flow depth are approximately 1 m at the most inland sand deposit sites. It is possible that the slip amount of the fault model in this study is
slightly underestimated. As shown in Fig. 3, red circles describe the locations where tsunami deposits were found. Computed tsunami inundation areas cover the locations of red circles at eleven areas. Black crosses describe the locations where tsunami deposits were not found. Computed tsunami inundation areas do not reach the locations of black crosses at Oikamanai, Pashukuru, Tokotan, and Nambu. At Onbetsu and Kiritappu, computed tsunami inundation reach the locations of several black crosses. The seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The rigidity was assumed to be 4.0 × 1010 N/m2 was used to calculate the seismic moment. This result is larger than the seismic moment of 0.8 × 1022 N m (Mw 8.5) estimated using the fault model of Satake et al. (2008). 6. Discussion Satake et al. (2008) indicates that if large tsunami came to the Sanriku coast in the 17th century, the tsunami would have caused damage and should have been recorded in historical document. Lack of historical records implies that the tsunami of the 17th century was not large enough to cause damage along the Sanriku coast. We examined the effect of tsunami heights of the 17th century great earthquake along the coast of Pacific Ocean in Tohoku region. Tsunami heights along the coast of Pacific Ocean in Tohoku region were numerically computed using a grid size of 1 arc-min by solving linear-long wave equations. Comparisons of computed tsunami heights with S model and without S model along the coast of Pacific Ocean in Tohoku region are shown in Fig. 4. The result shows that computed tsunami heights with S model along
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Fig. 4. Computed tsunami heights of the 17th century great earthquake along the coast of Pacific Ocean in Tohoku region are shown. Circles describe tsunami heights estimated from the fault model of the 17th century great earthquake without S model (red) and with S model (blue).
137
Tohoku is similar to that off Hokkaido. Yomogida et al. (2011) studied historical large earthquake distribution off Tohoku area. Large earthquakes of M 7.5 to 7.8 occur in the deep part of the plate interface in Miyagi region repeatedly. They also studied seismicity in 1980–2001 at depths shallower than 60 km in the Tohoku subduction zone, and found no seismicity in shallow segments close to the trench. Off Hokkaido, interplate earthquakes of M 7.4 to 8.2 occur in Tokachi-Oki and Nemuro-Oki segments in the deep part of seismic couple zone of the plate interface repeatedly. Takahashi and Kasahara (2007) investigated the spatiotemporal relationship between background seismic activity, coseismic asperities and aftershock activity in the southern Kurile Islands in 1952–2003 at depths shallower than 80 km. They found that the low seismicity area near the trench axis astride the Tokachi-Oki and Nemuro-Oki segments. The seismicity in the source area of the 2011 great Tohoku earthquake and the 17th century great earthquake also have similar characteristics. 7. Conclusions
Fig. 5. The fault model of the 17th century great earthquake is shown (T: TokachiOki segment, N: Nemuro-Oki segment, and S: Shallow part of the plate interface). Green rectangles show the slip distribution of the 1952 Tokachi-Oki earthquake estimated from tsunami waveform analysis (Hirata et al., 2003) and purple rectangles show the slip distribution of the 1973 Nemuro-Oki earthquake estimated from tsunami waveform analysis (Tanioka et al., 2007).
the coast of Sanriku, southern part of Tohoku region, are less than 4 m. Imamura (1949) and Iida (1958) indicate that tsunami height of about 4–6 m cause destruction of some houses and considerable loss of life. Computed tsunami heights of less than 4 m at Sanriku coast is consistent with lack of historical records in this region. Also, computed tsunami heights at Sanriku coast with S model and without S model are almost the same. It is due to an effect of directivity with the narrow width of S model. Calculated tsunami propagations of the 17th century great earthquake show that large tsunami was propagated toward the coast of eastern Hokkaido and much smaller tsunami was propagated toward the coast of Tohoku (Fig. S2). The locations of the Tokachi-Oki segment and Nemuro-Oki segment of the 17th century great earthquake at the deep part of the plate interface is almost same as the locations of slip distributions of the 1952 Tokachi-Oki earthquake and the 1973 Nemuro-Oki earthquake. Moreover, the 17th century great earthquake ruptured the shallow part of the plate interface near the trench (Fig. 5). The 2011 great Tohoku earthquake ruptured large area off Tohoku and the largest slip amount was fond at the shallow part of the plate interface near the trench. The fault model of the 17th century great earthquake also have the large slip amount at the large area and very large slip amount of 25 m at the shallow part of the plate interface near the trench. Additionally, seismicity off
The fault model of the 17th century great earthquake off eastern Hokkaido was estimated using tsunami deposit data. Lowland tsunami deposit data were compared to the calculated tsunami inundation areas. The elevations of the locations where tsunami deposits were found at highland near the coast were compared to the calculated tsunami heights. The result shows that a large slip amount of 25 m is necessary at the shallow part of the plate interface near the trench in Tokachi-Oki segment and NemuroOki segment. Lowland tsunami deposit data were explained by wide rupture area at the plate interface in Tokachi-Oki segment and Nemuro-Oki segment. High elevations of the locations where tsunami deposits were found near the coast were explained by very large slip at the shallow part of the plate interface near the trench in those segments. The total seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The 2011 great Tohoku earthquake ruptured large area off Tohoku and very large slip amount was found at the shallow part of the plate interface near the trench. The 17th century great earthquake have the same characteristics as the 2011 great Tohoku earthquake. Acknowledgements We thank Peter M. Shearer and two anonymous reviewers for their helpful comments and suggestions. We thank Aditya R. Gusman for help on slip distribution of the 2011 Tohoku earthquake. This study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, under its Earthquake and Volcano Hazards Observation and Research Program. This work was supported by JSPS KAKENHI Grant Number 24241060. Appendix A. Supplementary material Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.epsl.2015.10.009. References Fujii, Y., Satake, K., Sakai, S., Shinohara, M., Kanazawa, T., 2011. Tsunami source of the 2011 off Pacific coast of Tohoku Earthquake. Earth Planets Space 63, 815–820. Goto, C., Ogawa, Y., Shuto, N., Imamura, F., 1997. Numerical method of tsunami simulation with the leap-flog scheme. In: IUGG/IOC TIME Project. In: IOC Manual and Guides, vol. 35. UNESCO, pp. 1–126.
138
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Gusman, A.R., Tanioka, Y., Sakai, S., Tsushima, H., 2012. Source model of the great 2011 Tohoku earthquake estimated from tsunami waveforms and crustal deformation data. Earth Planet. Sci. Lett. 341–344, 234–242. Hatori, T., 1984. Source area of the east Hokkaido tsunami generated in April, 1843. Bull. Earthq. Res. Inst. Univ. Tokyo 59, 423–431 (in Japanese with English abstract). Hayes, G., 2011. Finite Fault Model. Updated Result of the Mar 11, 2011 Mw 9.0 Earthquake Offshore Honshu, Japan. http://earthquake.usgs.gov/earthquakes/ eqinthenews/2011/usc0001xgp/finite_fault.php. Hirakawa, K., 2006. Empirical knowledge on the significance of landforms in the Paleo-Tsunami study. In: Special Issue. Chikyu Monthly 54, 36–41 (in Japanese). Hirakawa, K., Nakamura, Y., Haraguchi, T., 2000a. Mega tsunami and recurrence interval at Tokachi coast in Hokkaido. In: Special Issue. Chikyu Monthly 28, 154–161 (in Japanese). Hirakawa, K., Nakamura, Y., Echigo, T., 2000b. Mega tsunami of Tokachi region at the coast of Pacific Ocean. In: Special Issue. Chikyu Monthly 31, 92–98 (in Japanese). Hirakawa, K., Nakamura, Y., Nishimura, Y., 2005. Mega-Tsunamis since last 6500 years along the Pacific Coast of Hokkaido. In: Special Issue. Chikyu Monthly 49, 173–180 (in Japanese). Hirata, K., Geist, E., Satake, K., Tanioka, Y., Yamaki, S., 2003. Slip distribution of the 1952 Tokachi-Oki earthquake (M 8.1) along the Kurile Trench deduced from tsunami waveform inversion. J. Geophys. Res. 108, 2196. http://dx.doi.org/ 10.1029/2002JB001976. Hirata, K., Satake, K., Tanioka, Y., Hasegawa, Y., 2009. Variable tsunami sources and seismic gaps in the southernmost Kuril trench: a review. Pure Appl. Geophys. 166, 77–96. Iida, K., 1958. Magnitude and energy of earthquakes accompanied by tsunami, and tsunami energy. J. Earth Sci. 6 (2), 101–112. Institute of Earth Sciences, Faculty of Science, Nagoya University. Imamura, A., 1949. List of Tsunamis in Japan. Zisin 2nd Ser. 2, 23–28. Imamura, F., 1996. Review of tsunami simulation with a finite difference method. In: Yeh, H., Liu, P., Synolakis, C. (Eds.), Long-Wave Runup Models. World Scientific, ISBN 981-02-2909-7, pp. 25–42. Nakamura, Y., Nishimura, Y., Moore, A.L., 2011. Correlation of tsunami deposits based on temporal change in coastal environment, eastern Hokkaido, Japan Geoscience Union Meeting 2011, HCG36-07. Nakamura, Y., Nishimura, Y., Sulastya, P.P., Moore, A.L., 2012. Correlation of paleotsunami layers based on grain size and sediment composition, eastern Hokkaido, Japan Geoscience Union Meeting 2012, MIS25-P17. Namegaya, Y., Satake, K., 2014. Reexamination of the A.D. 869 Jogan earthquake size from tsunami deposit distribution, simulated flow depth, and velocity. Geophys. Res. Lett. 41, 2297–2303.
Nanayama, F., Satake, K., Furukawa, R., Shimokawa, K., Atwater, B.F., Shigeno, K., Yamaki, S., 2003. Unusually large earthquakes inferred from tsunami deposits along the Kurile trench. Nature 424, 660–663. Nanayama, F., Furukawa, R., Shigeno, K., Makino, A., Soeda, Y., Igarashi, Y., 2007. Nine unusually large tsunami deposits from the past 4000 years at Kiritappu marsh along the southern Kurile trench. Sediment. Geol. 200, 275–294. Okada, Y., 1985. Surface deformation due to shear and tensile faults in a half-space. Bull. Seismol. Soc. Am. 75, 1135–1154. Pollitz, F.F., Burgmann, R., Banerjee, P., 2011. Geodetic slip model of the 2011 M9.0 Tohoku earthquake. Geophys. Res. Lett. 38, L00G08. http://dx.doi.org/10.1029/ 2011GL048632. Satake, K., 2007. Tsunamis. In: Earthquake Treatise on Geophysics Seismology. In: Treatise on Geophysics, vol. 4, pp. 483–512. Satake, K., Nanayama, S., Yamaki, S., Tanioka, Y., Hirata, K., 2005. Variability along tsunami source in the 17th–21st centuries along the southern Kurile trench. In: Tsunamis: Case Studies and Recent Developments, pp. 157–170. Satake, K., Nanayama, F., Yamaki, S., 2008. Fault models of unusual tsunami in the 17th century along the Kurile trench. Earth Planets Space 60, 925–935. Satake, K., Fujii, Y., Harada, T., Namegaya, Y., 2013. Time and Space Distribution of Coseismic Slip of the 2011 Tohoku Earthquake as Inferred from Tsunami Waveform Data. Bull. Seismol. Soc. Am. 103 (2B), 1473–1492. Sawai, Y., 2002. Evidence for 17th-century tsunamis generated on the Kurile– Kamchatka subduction zone, Lake Tokotan, Hokkaido, Japan. J. Asian Earth Sci. 20, 903–911. Sawai, Y., Kamataki, T., Shishikura, M., Nasu, H., Okamura, Y., Satake, K., Thomson, K.H., Matsumoto, D., Fujii, Y., Komasubara, J., Aung, T.T., 2009. Aperiodic recurrence of geologically recorded tsunamis during the past 5500 years in eastern Hokkaido, Japan. J. Geophys. Res. 114, B01319. http://dx.doi.org/ 10.1029/2007JB005503. Takahashi, H., Kasahara, M., 2007. Spatial relationship between interseismic seismicity, coseismic asperities and aftershock activity in the Southwestern Kurile Islands. In: Geophysical Monograph Series, vol. 172, pp. 153–164. Tanioka, Y., Hirata, K., Hino, R., Kanazawa, T., 2004. Slip distribution of the 2003 Tokachi-oki earthquake estimated from tsunami waveform inversion. Earth Planets Space 56, 373–376. Tanioka, Y., Satake, K., Hirata, K., 2007. Recurrence of recent large earthquakes along the southernmost Kurile–Kamchatka subduction zone. In: Geophysical Monograph, vol. 172, pp. 145–152. Yomogida, K., Yoshizawa, K., Koyama, J., Tsuzuki, M., 2011. Along-dip segmentation of the 2011 off the Pacific coast of Tohoku Earthquake and comparison with other megathrust earthquakes. Earth Planets Space 63, 697–701.
Contents lists available at ScienceDirect
Earth and Planetary Science Letters www.elsevier.com/locate/epsl
Re-estimated fault model of the 17th century great earthquake off Hokkaido using tsunami deposit data Kei Ioki ∗ , Yuichiro Tanioka Institute of Seismology and Volcanology, Hokkaido University, Japan
a r t i c l e
i n f o
Article history: Received 12 May 2015 Received in revised form 1 October 2015 Accepted 1 October 2015 Available online 7 November 2015 Editor: P. Shearer Keywords: tsunami great earthquake Hokkaido Kurile trench
a b s t r a c t Paleotsunami researches revealed that a great earthquake occurred off eastern Hokkaido, Japan and generated a large tsunami in the 17th century. Tsunami deposits from this event have been found at far inland from the Pacific coast in eastern Hokkaido. Previous study estimated the fault model of the 17th century great earthquake by comparing locations of lowland tsunami deposits and computed tsunami inundation areas. Tsunami deposits were also traced at high cliff near the coast as high as 18 m above the sea level. Recent paleotsunami study also traced tsunami deposits at other high cliffs along the Pacific coast. The fault model estimated from previous study cannot explain the tsunami deposit data at high cliffs near the coast. In this study, we estimated the fault model of the 17th century great earthquake to explain both lowland widespread tsunami deposit areas and tsunami deposit data at high cliffs near the coast. We found that distributions of lowland tsunami deposits were mainly explained by wide rupture area at the plate interface in Tokachi-Oki segment and Nemuro-Oki segment. Tsunami deposits at high cliff near the coast were mainly explained by very large slip of 25 m at the shallow part of the plate interface near the trench in those segments. The total seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The 2011 great Tohoku earthquake ruptured large area off Tohoku and very large slip amount was found at the shallow part of the plate interface near the trench. The 17th century great earthquake had the same characteristics as the 2011 great Tohoku earthquake. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Historically many great underthrust earthquakes occurred off Hokkaido in the Pacific Ocean. In 1843, 1952, and 2003, the Tokachi-Oki earthquakes occurred off eastern Hokkaido (Hatori, 1984; Hirata et al., 2003; Tanioka et al., 2004, respectively). In 1894 and 1973, the Nemuro-Oki earthquakes occurred off eastern Hokkaido at northeast of Tokachi-Oki earthquakes (Tanioka et al., 2007) (Fig. 1). In the southernmost Kuril Trench, large earthquakes have variability in spatial extent in both along-trench direction and trench-normal direction (Hirata et al., 2009). Tsunami deposits due to prehistoric tsunamis have been found along the coast of eastern Hokkaido (Sawai, 2002; Nanayama et al., 2003, 2007). These tsunami deposits were found at far inland from the coast where tsunamis generated by above historic earthquakes did not reach. The elevations of the locations where tsunami deposits were found at cliffs near the coast were also much higher than estimated tsunami heights by historical earth-
*
Corresponding author. E-mail address: [email protected] (K. Ioki).
http://dx.doi.org/10.1016/j.epsl.2015.10.009 0012-821X/© 2015 Elsevier B.V. All rights reserved.
quakes (Hirakawa et al., 2005). Nanayama et al. (2003) found that the latest event which left those tsunami deposits occurred in early 17th century. Satake et al. (2005) found that the 17th century great earthquake ruptured the plate interface at Tokachi-Oki, Akkeshi-Oki and Nemuro-Oki segments in Fig. 1 from tsunami simulation. Furthermore, Satake et al. (2008) estimated the fault model which reproduce the distribution of lowland tsunami deposits at five areas. However, the elevations of the locations where tsunami deposits were found near the coast as high as 18 m above the sea level (Hirakawa et al., 2005) were not completely explained by this fault model. In this paper, we estimate the fault model of the 17th century great earthquake to explain both locations of lowland tsunami deposits and highland elevations where tsunami deposits were found near the coast. 2. Previous studies of the 17th century tsunami Several marsh and lakes exist along the coast of eastern Hokkaido, so peat layers are common in those places. Tsunami deposits and volcanic ash layers are also seen within those peat layers. Nanayama et al. (2007) found 9 prehistoric tsunami sand beds in those peat layers among the past about 4000 yrs.
134
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
3. Data 3.1. Tsunami deposit data Lowland tsunami deposit data used by Satake et al. (2008) at Oikamanai pond in Taiki town, Pashukuru pond in Kushiro city, Tokotan pond in Akkeshi town, Kiritappu marsh in Hamanaka town, and Nambu pond in Nemuro city are used in this study. New lowland tsunami deposit data found by Nakamura et al. (2011, 2012) at Kinashibetsu marsh and Onbetsu in Kushiro city, and at Fureshima marsh in Nemuro city are also used in this study. Lowland tsunami deposit data found by Hirakawa et al. (2000b) at Rekifune river in Taiki town, Yudou pond and Choubushi lake in Toyokoro town are also used in this study. The locations of those ponds and marsh are shown in Fig. 1. The eight elevations of tsunami deposits surveyed at highland cliffs near the coast by Hirakawa et al. (2005) are also used as data. 3.2. Fault model parameter
Fig. 1. The fault model of the 17th century great earthquake is shown (T: TokachiOki segment, N: Nemuro-Oki segment, and S: Shallow part of the plate interface). Space–time diagram of great earthquakes along the southern Kurile trench from the 17th through 21st centuries is shown (modified from Satake et al., 2005). Green rectangles show the slip distribution of the 1952 Tokachi-Oki earthquake (Hirata et al., 2003) and purple rectangles show the slip distribution of the 1973 NemuroOki earthquake (Tanioka et al., 2007). Black rectangles show the slip distribution of the 2011 Tohoku earthquake (Gusman et al., 2012). Black dots show locations where tsunami inundation area and tsunami heights were calculated; (R: Rekifune, O: Oikamanai, Y: Yudou, C: Choubushi, K: Kinashibetsu, O: Onbetsu, P: Pashukuru, T: Tokotan, K: Kiritappu, F: Fureshima, N: Nemuro-nambu). The depth contour interval is 1000 m.
Sawai et al. (2009) also found sandy deposits by 15 tsunamis in about 6000 yrs in those places. The recurrence interval of those events is about 400∼600 yrs (Hirakawa et al., 2000a). The latest event occurred early 17th century because the latest tsunami deposits can be seen just under the volcanic ash caused by the 1667 Tarumae eruption (Nanayama et al., 2007). Those tsunami deposits are found about 1∼4 km inland from the coast of eastern Hokkaido. At Kiritappu marsh, tsunami deposits can be seen far inland from observed tsunami inundation area by the 1952 Tokachi-Oki earthquake (Satake et al., 2005). On the other hand, tsunami deposits were also observed at high cliffs near the coast (Hirakawa et al., 2000b, 2005; Hirakawa, 2006). The elevations of the locations where tsunami deposits were found are more than 10 m. At Oikamanai pond in Taiki town, the elevation was 18 m at high cliff near the coast (Hirakawa et al., 2005). At Fureshima marsh, the elevation of the locations where tsunami deposit was found was 15 m (Nakamura et al., 2011). Observed tsunami heights by historical earthquakes that occurred off Hokkaido were less than 4 m from Erimo Cape to Kushiro and less than 7 m from Kushiro to Nosappu Cape along the coast of eastern Hokkaido. The prehistoric tsunami of the 17th century were much larger than those historical tsunamis. Satake et al. (2008) estimated the fault model of the 17th century great earthquake by comparing locations of lowland tsunami deposits and computed tsunami inundations at five areas. The estimated slip amounts were 10 m in Tokachi-Oki segment and 5 m in Nemuro-Oki segment which includes Akkeshi-Oki segment in Fig. 1. The total seismic moment was calculated to be 0.8 × 1022 N m (Mw 8.5) with the rigidity of 4 × 1010 N/m2 . However, the tsunami heights along the coast of eastern Hokkaido calculated by this fault model (Satake et al., 2008) are smaller than the elevations of the locations where tsunami deposits were found near the coast in Tokachi (Hirakawa et al., 2005).
As we have already discussed before, the fault models of Tokachi-Oki segment and Nemuro-Oki segment are not enough to explain the elevations of the locations where tsunami deposits were found at high cliffs near the coast by Hirakawa et al. (2005). We need to find the fault model that can explain both lowland tsunami deposit data and the elevations of the locations where tsunami deposits were found near the coast. Satake et al. (2008) tested tsunami earthquake model which rupture the shallow part of the plate interface. The tsunami earthquake model yielded locally variable coastal heights on the coast of eastern Hokkaido and yielded little tsunami inundation to lowland marshes where tsunami deposits were found. Therefore, we hypothesize that both widespread inland tsunami inundations and locally large tsunami heights near the coast are potentially explained by the fault model which rupture not only the plate interface at Tokachi-Oki segment and Nemuro-Oki segment but also the shallow part of the plate interface near the trench. This hypothesis is come from the similarity of the 2011 great Tohoku earthquake (Mw 9.1) occurred along the Japan Trench and the 17th century great earthquake. Both tsunamis of the 17th century great earthquake and 2011 great Tohoku earthquake have characteristics that high tsunami heights were observed along the coast and large inundation areas were spread over wide areas. Slip distributions of the 2011 great Tohoku earthquake were estimated by several previous studies using seismic waves (Hayes, 2011), GPS data (Pollitz et al., 2011) and tsunami waveform data (Fujii et al., 2011). Gusman et al. (2012) also estimated the slip distribution of the earthquake using tsunami waveforms, GPS data, and seafloor crustal deformation data. The result shows that the largest slip amount of about 44 m ruptured the plate interface near the trench (Fig. 1). Satake et al. (2013) also estimated the slip distribution of the earthquake shown the huge slip on the shallow part of the plate interface near the trench. The characteristics of very large slip at the plate interface near the trench is common in all of the above previous studies. The largest slip near the trench was one of reasons to generate devastating tsunami along the coast in Tohoku. Before the 2011 great Tohoku earthquake occurred, it was believed that the shallow part of the plate interface near the trench is aseismic or rupture with a slow velocity. However, the 2011 great Tohoku earthquake indicates the plate interface can be ruptured with very large slip amount. Therefore, the fault model of the 17th century great earthquake can be similar to that of the 2011 great Tohoku earthquake. We assumed that the 17th century great earthquake ruptured the shallow part of the plate interface near the Kurile trench with a large slip amount, similar to the 2011 great Tohoku earthquake.
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Additional fault model, S model, was allocated at the shallow part of the plate interface near the Kurile trench to explain high tsunami heights along the coast of eastern Hokkaido (Fig. 1). Fault parameters of S model have length of 300 km, width of 30 km, and shallowest depth of the fault of 6.7 km. Slip amount of S model was changed to 15 m, 25 m, and 35 m. The mechanism of the fault model of the 17th century great earthquake has strike of 228◦ (parallel to the Kurile trench), dip of 15◦ (average angle of subducting Pacific plate in the Kurile trench), and rake of 90◦ (pure thrust earthquake). The fault length, fault width, and slip amount of the Tokachi-Oki segments and Nemuro-Oki segments in Fig. 1 are the same as those used by Satake et al. (2008). Those fault parameters of the Tokachi-Oki segment are length = 100 km, width = 100 km, and slip = 10 m. Those fault parameters of the Nemuro-Oki segment are length = 200 km, width = 100 km, and slip = 5 m. The shallowest depth of the fault of 14 km was used in both segments. Parameters of the fault model of the 17th century great earthquake used in the tsunami numerical simulation are summarized in Table 1. 4. Tsunami numerical simulation The bathymetry used for tsunami numerical simulation was created by interpolating M7000 digital bathymetric chart published by Japan Hydrographic Association. The topography was created by interpolating SRTM 90 m Digital Elevation Data published by Table 1 Fault models and their parameters of the 17th century great earthquake.
135
CSI-CGIAR and Digital map 50 m Grid published by Geospatial Information Authority of Japan. A coarse grid size for tsunami numerical simulation that covers all computational area, 142◦ –149◦ E and 39◦ –45◦ N, is 10 arcsecond. A fine grid size for tsunami inundation simulation near the coast of eastern Hokkaido is 1 arc-second. To compute tsunami, first, coseismic deformation on the seafloor was computed using Okada’s formula (Okada, 1985). Crosssections of the calculated ocean bottom deformations are shown in Fig. 2. We assumed that the initial water deformation on the sea surface is equal to the coseismic deformation on the seafloor. Linier long-wave equations were solved to compute tsunami in the coarse grid system. A finite-difference scheme with a staggered grid system was used to solve the equations (Satake, 2007). The open boundary condition was used at the end of the computed area. The time step of 1 s was chosen to satisfy a stability condition. Non-linear long-wave equations were solved to compute tsunami inundation (Imamura, 1996) in the fine grid system. Tsunami heights computed using linear long-wave equations are interpolated at the boundary of fine grid area. These interpolated tsunami waveforms are from the boundary into the fine grid system where tsunami inundation were computed using non-linear long-wave equations (Goto et al., 1997). To satisfy a stability condition, the time step was chosen to be 0.05 s. Manning’s roughness coefficient of 0.03 m−1/3 s was assumed in this simulation. The coseismic deformation was included into the bathymetry data in computed area. 5. Results
Fault model
Length (km)
Width (km)
Depth (km)
Strike (deg)
Dip (deg)
Rake (deg)
Slip (m)
T N S
100 200 300
100 100 30
14 14 6 .7
228 228 228
15 15 15
90 90 90
10 5 0–35
Slip amount of S model that located the plate interface were estimated by comparing tsunami deposits were found and computed areas and also by comparing the elevations
shallow part of the the locations where tsunami inundation where tsunami de-
Fig. 2. Cross-sections of ocean bottom deformation along the line in A through A computed from the fault model of the 17th century great earthquake (red line: without S model; blue line: with S model) and along the line B through B computed from the fault model of the 2011 Tohoku earthquake (Gusman et al., 2012).
136
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Table 2 Comparisons of measured and calculated tsunami heights (m) with changing slip amounts of S model. Measureda Rekifune river (R1) Rekifune river (R2) Rekifune river (R3) Oikamanai pond (O1) Oikamanai pond (O2) Yudou pond (Y1) Yudou pond (Y2) Kinashibetsu marsh (K1) Fureshima marsh (F1) a
15
>12 16 14 18 >10 15 12 15
Without S model
Slip = 15 (m)
Slip = 25 (m)
Slip = 35 (m)
5 7 6 11 11 2 12 11 7
5 7 11 12 17 3 16 13 13
6 11 17 16 24 6 20 15 17
8 13 21 18 29 9 27 17 21
The elevations of the locations where tsunami deposits were found (Hirakawa et al., 2005; Nakamura et al., 2011).
Fig. 3. (a) Computed tsunami inundation and heights estimated from the fault model of the 17th century great earthquake at Oikamanai pond (left side: without S model, right side: with S model). Red circles describe the locations where tsunami deposits of the 17th century great earthquake were found by Satake et al. (2008) and Hirakawa et al. (2005). Black crosses describe the locations where tsunami deposits were not found. The numbers show computed tsunami heights and the numbers in the bracket show the elevations of the locations where tsunami deposits were found. (b) Comparison of computed coastal tsunami heights (red: without S model, blue: with S model). Black rectangles show the elevations of the locations where tsunami deposits were found.
posits were found and computed tsunami heights near the coast of eastern Hokkaido. Computed tsunami heights at highland near the coast with slip amount of 0 m (without S model), 15 m, 25 m, and 35 m in S model are shown in Table 2. Computed tsunami heights without S model and with S model having a slip amount of 15 m were much smaller than the measured tsunami heights. Computed tsunami heights with S model having a slip amount of 25 m and 35 m explain most of the elevations of the locations where tsunami deposits were found. The elevations of the locations where tsunami deposits were found indicate the minimum tsunami height at the location. Therefore, computed tsunami heights should be larger than the elevations of the locations where tsunami deposits were found. As a result, at least, slip amount of 25 m in S model was needed to explain measured tsunami heights at high cliff. Computed tsunami inundation areas of lowland and computed tsunami heights of highland at eleven areas are shown in Fig. 3 and Fig. S1. Computed tsunami inundation areas should cover the locations where tsunami deposits were found. Computed tsunami inundation areas with S model having a slip amount of 25 m best fit to the locations where tsunami deposits were found. Slip amount of S model was chosen to fit the locations where tsunami deposits were not found, black crosses in Fig. 3 and Fig. S1, and the end of computed tsunami inundation areas in this study. Namegaya and Satake (2014) indicate that the tsunami inundation are extend farther inland than the sandy tsunami deposits and the flow depth are approximately 1 m at the most inland sand deposit sites. It is possible that the slip amount of the fault model in this study is
slightly underestimated. As shown in Fig. 3, red circles describe the locations where tsunami deposits were found. Computed tsunami inundation areas cover the locations of red circles at eleven areas. Black crosses describe the locations where tsunami deposits were not found. Computed tsunami inundation areas do not reach the locations of black crosses at Oikamanai, Pashukuru, Tokotan, and Nambu. At Onbetsu and Kiritappu, computed tsunami inundation reach the locations of several black crosses. The seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The rigidity was assumed to be 4.0 × 1010 N/m2 was used to calculate the seismic moment. This result is larger than the seismic moment of 0.8 × 1022 N m (Mw 8.5) estimated using the fault model of Satake et al. (2008). 6. Discussion Satake et al. (2008) indicates that if large tsunami came to the Sanriku coast in the 17th century, the tsunami would have caused damage and should have been recorded in historical document. Lack of historical records implies that the tsunami of the 17th century was not large enough to cause damage along the Sanriku coast. We examined the effect of tsunami heights of the 17th century great earthquake along the coast of Pacific Ocean in Tohoku region. Tsunami heights along the coast of Pacific Ocean in Tohoku region were numerically computed using a grid size of 1 arc-min by solving linear-long wave equations. Comparisons of computed tsunami heights with S model and without S model along the coast of Pacific Ocean in Tohoku region are shown in Fig. 4. The result shows that computed tsunami heights with S model along
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Fig. 4. Computed tsunami heights of the 17th century great earthquake along the coast of Pacific Ocean in Tohoku region are shown. Circles describe tsunami heights estimated from the fault model of the 17th century great earthquake without S model (red) and with S model (blue).
137
Tohoku is similar to that off Hokkaido. Yomogida et al. (2011) studied historical large earthquake distribution off Tohoku area. Large earthquakes of M 7.5 to 7.8 occur in the deep part of the plate interface in Miyagi region repeatedly. They also studied seismicity in 1980–2001 at depths shallower than 60 km in the Tohoku subduction zone, and found no seismicity in shallow segments close to the trench. Off Hokkaido, interplate earthquakes of M 7.4 to 8.2 occur in Tokachi-Oki and Nemuro-Oki segments in the deep part of seismic couple zone of the plate interface repeatedly. Takahashi and Kasahara (2007) investigated the spatiotemporal relationship between background seismic activity, coseismic asperities and aftershock activity in the southern Kurile Islands in 1952–2003 at depths shallower than 80 km. They found that the low seismicity area near the trench axis astride the Tokachi-Oki and Nemuro-Oki segments. The seismicity in the source area of the 2011 great Tohoku earthquake and the 17th century great earthquake also have similar characteristics. 7. Conclusions
Fig. 5. The fault model of the 17th century great earthquake is shown (T: TokachiOki segment, N: Nemuro-Oki segment, and S: Shallow part of the plate interface). Green rectangles show the slip distribution of the 1952 Tokachi-Oki earthquake estimated from tsunami waveform analysis (Hirata et al., 2003) and purple rectangles show the slip distribution of the 1973 Nemuro-Oki earthquake estimated from tsunami waveform analysis (Tanioka et al., 2007).
the coast of Sanriku, southern part of Tohoku region, are less than 4 m. Imamura (1949) and Iida (1958) indicate that tsunami height of about 4–6 m cause destruction of some houses and considerable loss of life. Computed tsunami heights of less than 4 m at Sanriku coast is consistent with lack of historical records in this region. Also, computed tsunami heights at Sanriku coast with S model and without S model are almost the same. It is due to an effect of directivity with the narrow width of S model. Calculated tsunami propagations of the 17th century great earthquake show that large tsunami was propagated toward the coast of eastern Hokkaido and much smaller tsunami was propagated toward the coast of Tohoku (Fig. S2). The locations of the Tokachi-Oki segment and Nemuro-Oki segment of the 17th century great earthquake at the deep part of the plate interface is almost same as the locations of slip distributions of the 1952 Tokachi-Oki earthquake and the 1973 Nemuro-Oki earthquake. Moreover, the 17th century great earthquake ruptured the shallow part of the plate interface near the trench (Fig. 5). The 2011 great Tohoku earthquake ruptured large area off Tohoku and the largest slip amount was fond at the shallow part of the plate interface near the trench. The fault model of the 17th century great earthquake also have the large slip amount at the large area and very large slip amount of 25 m at the shallow part of the plate interface near the trench. Additionally, seismicity off
The fault model of the 17th century great earthquake off eastern Hokkaido was estimated using tsunami deposit data. Lowland tsunami deposit data were compared to the calculated tsunami inundation areas. The elevations of the locations where tsunami deposits were found at highland near the coast were compared to the calculated tsunami heights. The result shows that a large slip amount of 25 m is necessary at the shallow part of the plate interface near the trench in Tokachi-Oki segment and NemuroOki segment. Lowland tsunami deposit data were explained by wide rupture area at the plate interface in Tokachi-Oki segment and Nemuro-Oki segment. High elevations of the locations where tsunami deposits were found near the coast were explained by very large slip at the shallow part of the plate interface near the trench in those segments. The total seismic moment of the 17th century great earthquake was calculated to be 1.7 × 1022 N m (Mw 8.8). The 2011 great Tohoku earthquake ruptured large area off Tohoku and very large slip amount was found at the shallow part of the plate interface near the trench. The 17th century great earthquake have the same characteristics as the 2011 great Tohoku earthquake. Acknowledgements We thank Peter M. Shearer and two anonymous reviewers for their helpful comments and suggestions. We thank Aditya R. Gusman for help on slip distribution of the 2011 Tohoku earthquake. This study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, under its Earthquake and Volcano Hazards Observation and Research Program. This work was supported by JSPS KAKENHI Grant Number 24241060. Appendix A. Supplementary material Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.epsl.2015.10.009. References Fujii, Y., Satake, K., Sakai, S., Shinohara, M., Kanazawa, T., 2011. Tsunami source of the 2011 off Pacific coast of Tohoku Earthquake. Earth Planets Space 63, 815–820. Goto, C., Ogawa, Y., Shuto, N., Imamura, F., 1997. Numerical method of tsunami simulation with the leap-flog scheme. In: IUGG/IOC TIME Project. In: IOC Manual and Guides, vol. 35. UNESCO, pp. 1–126.
138
K. Ioki, Y. Tanioka / Earth and Planetary Science Letters 433 (2016) 133–138
Gusman, A.R., Tanioka, Y., Sakai, S., Tsushima, H., 2012. Source model of the great 2011 Tohoku earthquake estimated from tsunami waveforms and crustal deformation data. Earth Planet. Sci. Lett. 341–344, 234–242. Hatori, T., 1984. Source area of the east Hokkaido tsunami generated in April, 1843. Bull. Earthq. Res. Inst. Univ. Tokyo 59, 423–431 (in Japanese with English abstract). Hayes, G., 2011. Finite Fault Model. Updated Result of the Mar 11, 2011 Mw 9.0 Earthquake Offshore Honshu, Japan. http://earthquake.usgs.gov/earthquakes/ eqinthenews/2011/usc0001xgp/finite_fault.php. Hirakawa, K., 2006. Empirical knowledge on the significance of landforms in the Paleo-Tsunami study. In: Special Issue. Chikyu Monthly 54, 36–41 (in Japanese). Hirakawa, K., Nakamura, Y., Haraguchi, T., 2000a. Mega tsunami and recurrence interval at Tokachi coast in Hokkaido. In: Special Issue. Chikyu Monthly 28, 154–161 (in Japanese). Hirakawa, K., Nakamura, Y., Echigo, T., 2000b. Mega tsunami of Tokachi region at the coast of Pacific Ocean. In: Special Issue. Chikyu Monthly 31, 92–98 (in Japanese). Hirakawa, K., Nakamura, Y., Nishimura, Y., 2005. Mega-Tsunamis since last 6500 years along the Pacific Coast of Hokkaido. In: Special Issue. Chikyu Monthly 49, 173–180 (in Japanese). Hirata, K., Geist, E., Satake, K., Tanioka, Y., Yamaki, S., 2003. Slip distribution of the 1952 Tokachi-Oki earthquake (M 8.1) along the Kurile Trench deduced from tsunami waveform inversion. J. Geophys. Res. 108, 2196. http://dx.doi.org/ 10.1029/2002JB001976. Hirata, K., Satake, K., Tanioka, Y., Hasegawa, Y., 2009. Variable tsunami sources and seismic gaps in the southernmost Kuril trench: a review. Pure Appl. Geophys. 166, 77–96. Iida, K., 1958. Magnitude and energy of earthquakes accompanied by tsunami, and tsunami energy. J. Earth Sci. 6 (2), 101–112. Institute of Earth Sciences, Faculty of Science, Nagoya University. Imamura, A., 1949. List of Tsunamis in Japan. Zisin 2nd Ser. 2, 23–28. Imamura, F., 1996. Review of tsunami simulation with a finite difference method. In: Yeh, H., Liu, P., Synolakis, C. (Eds.), Long-Wave Runup Models. World Scientific, ISBN 981-02-2909-7, pp. 25–42. Nakamura, Y., Nishimura, Y., Moore, A.L., 2011. Correlation of tsunami deposits based on temporal change in coastal environment, eastern Hokkaido, Japan Geoscience Union Meeting 2011, HCG36-07. Nakamura, Y., Nishimura, Y., Sulastya, P.P., Moore, A.L., 2012. Correlation of paleotsunami layers based on grain size and sediment composition, eastern Hokkaido, Japan Geoscience Union Meeting 2012, MIS25-P17. Namegaya, Y., Satake, K., 2014. Reexamination of the A.D. 869 Jogan earthquake size from tsunami deposit distribution, simulated flow depth, and velocity. Geophys. Res. Lett. 41, 2297–2303.
Nanayama, F., Satake, K., Furukawa, R., Shimokawa, K., Atwater, B.F., Shigeno, K., Yamaki, S., 2003. Unusually large earthquakes inferred from tsunami deposits along the Kurile trench. Nature 424, 660–663. Nanayama, F., Furukawa, R., Shigeno, K., Makino, A., Soeda, Y., Igarashi, Y., 2007. Nine unusually large tsunami deposits from the past 4000 years at Kiritappu marsh along the southern Kurile trench. Sediment. Geol. 200, 275–294. Okada, Y., 1985. Surface deformation due to shear and tensile faults in a half-space. Bull. Seismol. Soc. Am. 75, 1135–1154. Pollitz, F.F., Burgmann, R., Banerjee, P., 2011. Geodetic slip model of the 2011 M9.0 Tohoku earthquake. Geophys. Res. Lett. 38, L00G08. http://dx.doi.org/10.1029/ 2011GL048632. Satake, K., 2007. Tsunamis. In: Earthquake Treatise on Geophysics Seismology. In: Treatise on Geophysics, vol. 4, pp. 483–512. Satake, K., Nanayama, S., Yamaki, S., Tanioka, Y., Hirata, K., 2005. Variability along tsunami source in the 17th–21st centuries along the southern Kurile trench. In: Tsunamis: Case Studies and Recent Developments, pp. 157–170. Satake, K., Nanayama, F., Yamaki, S., 2008. Fault models of unusual tsunami in the 17th century along the Kurile trench. Earth Planets Space 60, 925–935. Satake, K., Fujii, Y., Harada, T., Namegaya, Y., 2013. Time and Space Distribution of Coseismic Slip of the 2011 Tohoku Earthquake as Inferred from Tsunami Waveform Data. Bull. Seismol. Soc. Am. 103 (2B), 1473–1492. Sawai, Y., 2002. Evidence for 17th-century tsunamis generated on the Kurile– Kamchatka subduction zone, Lake Tokotan, Hokkaido, Japan. J. Asian Earth Sci. 20, 903–911. Sawai, Y., Kamataki, T., Shishikura, M., Nasu, H., Okamura, Y., Satake, K., Thomson, K.H., Matsumoto, D., Fujii, Y., Komasubara, J., Aung, T.T., 2009. Aperiodic recurrence of geologically recorded tsunamis during the past 5500 years in eastern Hokkaido, Japan. J. Geophys. Res. 114, B01319. http://dx.doi.org/ 10.1029/2007JB005503. Takahashi, H., Kasahara, M., 2007. Spatial relationship between interseismic seismicity, coseismic asperities and aftershock activity in the Southwestern Kurile Islands. In: Geophysical Monograph Series, vol. 172, pp. 153–164. Tanioka, Y., Hirata, K., Hino, R., Kanazawa, T., 2004. Slip distribution of the 2003 Tokachi-oki earthquake estimated from tsunami waveform inversion. Earth Planets Space 56, 373–376. Tanioka, Y., Satake, K., Hirata, K., 2007. Recurrence of recent large earthquakes along the southernmost Kurile–Kamchatka subduction zone. In: Geophysical Monograph, vol. 172, pp. 145–152. Yomogida, K., Yoshizawa, K., Koyama, J., Tsuzuki, M., 2011. Along-dip segmentation of the 2011 off the Pacific coast of Tohoku Earthquake and comparison with other megathrust earthquakes. Earth Planets Space 63, 697–701.