Tsunamis - Extreme Events: A Physical Reconstruction and Risk Assessment - page 125

Darienzo and Petersen (1990), Clague and Brobrowsky (1994)and Clague et al .
(1999), identified six separate tsunami events along the northwest USA and south-
west Canadian coasts from buried sand layers separated by fine-grained sedi-
ments and peat dating between 300 and 3200 years BP. Twelve layers of sand
generated from separate tsunami events have also been identified in a peat bog in
Crescent City, California. These tsunamis would have been generated from large
earthquakes in the Cascadia subduction zone. Nanayama et al .(2003)identified
17 tsunami sand layers separated by peat and volcanic ash on the Japanese island
of Hokkaido. These sand sheets extend inland for up to 3 km and were deposited
over the last 2000--7000 years giving an average return interval between events of
500 years. Comparisons between the spatial extent of these sheets and those of
historically recorded tsunami over the last two centuries, showed that the earth-
quakes responsible were very large, involving multiple segment ruptures along
many hundreds of kilometres of the Kuril Trench where the Pacific tectonic
plate converges with Eurasia. Historical tsunamis here have only been generated
by earthquakes from single segment ruptures of 100--200 km length, which was
deemed to be the normal style of earthquake and tsunami inundation for the
region.
Fewstudies have reported clearly identifiable sedimentary structures within
tsunami sand sheets as the majority of deposits display massive sedimenta-
tion. However, surveys of the sand sheet deposited by the 1998 Aitape, Papua
NewGuinea tsunami near Sissano Lagoon reported faint horizontal stratifica-
tion towards the top of the deposit (USGS, 1998)(Fig.5.1), and Sato et al .(1995)
reported cross bedding in the 1993 southwest Hokkaido deposit. Lamination bed-
ding was also reported in other Japanese tsunami deposits (Fujiwara et al . 2000).
Sometimes load structures (where depressions are made within an underlying
deformable unit of sediments) are observed at the base of the deposit (Goff et al .,
1998).
Goff et al .'s (1998)study of the AD 1855 tsunami deposit near Wellington,
New Zealand examined the roundness and sphericity of lithic clasts, as well as
the fabricofclasts(i.e.alignment of a axes), and the presence and nature of
pumice and marine shells. Marine lithic clasts tend to have a flatter disk shape
than the rounded gravels of rivers, because marine clasts experience continuous
movement up and down the beach with the swash and backwash associated with
normal fair weather waves. Clast axis alignment tends to result in the longest
axis ( a axis) aligning parallel or subparallel to the shore or perpendicular to
thedirection of wave transport. This is because the clast can be more easily
transported in a fluid flow by rolling over its shortest axis ( c axis) which is
perpendicular to the a axis. Clasts can also become imbricated in fluid flows
and in the case of marine inundations the clasts will normally dip seaward or