Comets and Asteroids - Tsunami - page 174

unidirectional, pointing away from the impact site for sev-
eral minutes under the crest. This would have been followed
by a longer period of reverse flow that decreased in mag-
nitude until the next wave in the tsunami wave train
approached. Thus, sandy sediment once deposited in the
deep ocean could have been reworked by other waves in the
tsunami wave train and by the seiches that followed. In
addition, ejecta in the form of spherules are absent or sparse
at the base of some deposits. Ejecta present at the base of
deposits in the ocean did not necessarily fall there in situ
from the atmosphere. Because ejecta were thrown high into
the atmosphere, it took time for it to settle back to the
Earth's surface. During this time the Earth rotated east-
wards. Thus most ejecta tended to fall on the shelf to the
west of the impact site, forming a layer 1 m thick. Strong
tsunami backwash swept these spherules seaward and
deposited them first below the subsequent turbidites. Shelf
sands then covered the turbidites, leading to a reversal in
stratigraphy.
Both the hydrodynamics of the tsunami and the relative
abundance of ejecta are reflected in the deposits. At Beloc,
on the Island of Haiti, which was in water about 2 km deep
at the time of the impact, deposits are thinnest because they
are furthest from sources of sand on the shelf. A turbidity
current first deposited a layer of ejecta 15-70 cm thick
(Smit et al. 1996 ). This grades upwards into a deposit of fine
sand and silt that is 20 cm thick with low-angled cross-
bedding (Fig. 9.8 ). Thin iridium-rich layers of sandy silt
1-2 mm thick occur at the top of the sequence. This upper
segment in many locations is disturbed, implying that a
second tsunami may have affected the area afterwards. This
could have been generated by volcanism triggered by the
impact or by subsequent landslides on an unstable seabed.
In slightly shallower water, but still offshore of any shelf,
west of the impact site at Mimbral where still water con-
ditions existed on the seabed, fine limestone-rich clays
called marls were overlain by 1 m of ripped-up limestone,
mixed together with the ejecta debris from the impact zone
(Behor 1996 ). These sediments are best preserved in the
channels cut by tsunami backwash and are overlain by at
least 2 m of sand derived from the distant shoreline and
brought to the seabed by turbidity currents. This unit can be
traced over a distance of 2,000 km, from Alabama through
Texas to the southern border of Mexico (Smit et al. 1996 ;
Bourgeois et al. 1988 ). This unit is overlain by a meter of
crosscurrent beds of sequential rippled sand and fine clay. In
many respects, the deposit has all of the characteristics of a
Bouma sequence as described in Chap. 3 . Finally, a thin
layer, several centimeters thick, caps the sequence. This
uppermost layer contains the iridium-rich dust fallout that
settled out of the atmosphere over several weeks following
the impact. Closer to shore, about 100 km landward of the
shelf edge at a site like the Brazos River in Texas, the

tsunami wave swept over the muddy, flat shelf, scouring out
swales with a relief of 0.7 m. Backwash dominates these
sites. The bottom part of the sand deposits is about 1.3 m
thick and consists of rounded calcareous cobbles, shell, fish
teeth, terrestrial wood debris, and angular pieces of mud-
stone. At Moscow Landing, Alabama, the shelf was only
30 m deep at the time of impact. Seismic waves preceded
the tsunami creating normal faults in a north-south direc-
tion. These are paralleled by grooves, flute casts, scour
features, and lineations created by the passage of the wave
over undular topography molded into soft bottom muds.
Here, backwash in the form of undertow wedged the bottom
sediments against fault blocks. Boulders are also present in
the deposit. The bottom unit fines upwards into paral-
lel-laminated and symmetrically rippled sandstone, silt-
stone, and mudstone. In places, the sequence is repeated up
to three times, indicating the passage of more than one
wave. The upward part of the sequence, which is irid-
ium-rich, shows evidence of seiching.
The size of mud clasts at the Brazos River site gives
some indication of the bottom shear velocities that must
have been generated over the shelf (Bourgeois et al. 1988 ).
These values are as high as 1 m s -1 , far exceeding those
that could be generated by storms on a shelf in 50-100 m
depth of water. The velocities allude to a tsunami wave
between 50 m and 100 m high with a period of 30-60 min
at this location. Modeling suggests that the waves rolled
back and forth across the proto-landscape at a periodicity of
1-2 h (Matsui et al. 2002 ). Boulder deposits indicative of a
mega-tsunami are rare; however, they have been found
close to the shoreline of the proto-Gulf at Parras Basin in
northeastern Mexico and in the mid-south of the United
States. At the latter location, anomalous sandstone boulders
up to 15 m in diameter have been found 80 m above
floodplains on hills that would have been coastal headlands
at the time (Patterson 1998 ).

9.4.3 Other Events

Chicxulub tends to dominate the public's perception of what
a large asteroid or comet impact can do. While this is the
biggest event in the last 225 million years, it is not the only
one to have generated an impressive tsunami. For example,
an event known as the Eltanin Asteroid Event occurred
during the Pliocene 2.15 million years ago (Gersonde et al.
1997 ). This asteroid, estimated to have been 4 km in
diameter, plunged into the Pacific Ocean 700 km off the
southwest corner of South America and exploded, sending
ejecta into the atmosphere. If the object had a density of
3.6 g cm -3 and struck at a speed of 20 km s -1 , it potentially
generated a mega-tsunami—according to Eq. ( 9.6 )—at least
30 m in height along nearby coasts in South America and