Tsunami-induced sediment transport in the
abyssal Mediterranean Sea
KIM A. KASTENS* University of California, San Diego, Marine Physical Laboratory of the Scripps Institution of Oceanography,
La Jolla, California 92093
MARIA B. CITA Instituto di Geologia, University of Milan, Milan, Italy
ABSTRACT
An unusual stratigraphic unit (nicknamed "homogenite") fills
topographic lows in the complex ridge and trough bathymetry at
two survey sites on the Western Mediterranean Ridge and the Cal-
abrian Ridge. On near-bottom 4-kHz seismic-reflection profiles,
this unit us an acoustically transparent, near-surface, flat-lying
layer, whereas in cores, it is a homogeneous gray marl as much as 7
m thick. Grain size decreases upcore within the unit, implying that
it was deposited in a single event controlled by gravitational set-
tling. The stratigraphic position of the homogenite relative to a
firmly dated sapropel bed suggests emplacement between 4400 and
3100 yr B.P. The source of the homogenite is inferred to be the
nearby basin walls. Farther east, two other sites with similar rugged
topography lack homogenite entirely.
A triggering mechanism is required which is capable of initiat-
ing massive sediment transport simultaneously in many separate
basins at the western two survey sites, but which is not effective at
the eastern sites. A large archeologically recorded earthquake of the
correct age is considered and rejected because its epicenter is closer
to the nonhomogenite-bearing sites than to the sites where this
sediment type was observed, and because several other earthquakes
of comparable magnitude have since been recorded in the area,
whereas the homogenite is unique. The 3,500 yr
B.
P. collapse of the
caldera of the volcano of Santorini caused a huge tsunami which is
recorded archeologically and geologically around the eastern Medi-
terranean. Because of refraction of the tsunami by the bathymetry,
and because the caldera collapsed in its southwest corner, a
disproportionate amount of tsunami energy was directed toward
the western area where homogenite is observed. In contrast, the
homogenite-free sites were relatively sheltered. An order-of-
magnitude calculation shows that the near-bottom oscillating cur-
rents accompanying the Santorini tsunami were at or above the
threshold erosion velocity at the homogenite-bearing sites. In addi-
tion, the near-bottom pressure pulse under the tsunami at the
homogenite-bearing sites was sufficient to cause liquefaction of sed-
iments. Neither mechanism was adequate to cause sediment trans-
port or slope failure at the homogenite-free sites.
•Present address: Lamont-Doherty Geological Observatory, Palisades,
New York 10964.
INTRODUCTION
During the summer of 1978, two oceanographic cruises
explored areas of the eastern Mediterranean (Fig. 1) characterized
by a peculiar hummocky relief that is referred to as "cobblestone
topography." First, the Deep Tow Instrument Package of the
Marine Physical Laboratory of Scripps Institution of Oceano-
graphy (Spiess and Tyce, 1973) was used to obtain bathymetric,
side-looking sonar, and 4-kHz seismic-reflection data. After preli-
minary data reduction, R/V Eastward obtained 40 cores precisely
located with respect to the geophysical data by means of the same
acoustic transponders used on the first cruise.
The highly irregular topography of basins and domes and
ridges and troughs which makes up the "cobblestone" terrain is the
setting for a complex system of small-scale sediment erosion, trans-
port, and depositional processes. This paper describes an unusual
sediment type that we suggest was deposited from sediment trans-
port induced by the tsunami which followed the 3500 yr B.P. erup-
tion and collapse of the caldera of Santorini. A description of the
sedimentology and acoustic nature of the sediment, its distribution,
age, and provenance, will be followed by a detailed discussion of the
plausibility of this unorthodox sediment transport triggering
mechanism.
NATURE OF' THE SEDIMENT
Description
During the near-bottom seismic-reflection survey, a distinct
acoustically transparent unit was found occupying the uppermost
part of the sediment column in certain areas (Fig. 2). It fills local
depressions and shows a flat upper surface and a slightly concave
upward contact with the acoustically layered sediment below. In
thickness it ranges from barely discernible to >10 m sec two-way
travel time (>7.5 m).
Eleven cores were recovered from six separate basins contain-
ing the transparent layer. In each case, the transparent layer corre-
lated with a thick, gray, hemipelagic marl unit. Quartz and clays
dominate the terrigenous component of this marl, and the biogenic
fraction is predominantly nannofossils. The basal contact is sharp
and perhaps erosional. The unit starts with a half-centimetre-thick
Geological Society of America Bulletin, Part I, v. 92, p. 845-857, 10 figs., 1 table, November 1981.
845
on February 24, 2015gsabulletin.gsapubs.orgDownloaded from
Figure 1. Location map showing the survey area on the Mediterranean Ridge and the Calabrian Ridge. All four sites were surveyed with
the Deep Tow and an intensive coring program was conducted at the western two sites. CR is the Calabrian Ridge Site; WMR, CMR, and
EMR are western, central, and eastern Mediterranean Ridge sites.
CD
>
O
I
o
o
o
a>
CO
LU
O
en
LU
TRANSPARENT LAYER
r
= HOMOGENITE
Figure 2. Near-botton 4-kHz seismic-reflection record showing the acoustically transparent "homogenite" layer. This sediment type is
flat-lying and basin filling, never found on local bathymetric highs or slopes.
on February 24, 2015gsabulletin.gsapubs.orgDownloaded from
TSUNAMI-INDUCED SEDIMENT TRANSPORT 847
%
IN
SIZE FRACTION
Figure 3. Grain size distribution in homogenite from core 42.
The monotonie increase in fine fraction upcore is taken as evidence
that the entire unit was deposited in one event dominated by gravi-
tational settling of sediment from suspension.
sandy layer (primarily planktonic forminiferal tests), and fines
rapidly upward (Fig. 3). The graded layering implies that the depo-
sition process involved gravitational settling.
The most remarkable feature of this unit is that, except for the
grain size gradation, there are absolutely no sedimentary structures
discernible in the split cores or in X-radiographs throughout a
thickness of as much as 7.7 m. In this respect, this sediment type is
very different fro sediments previously recovered from the eastern
Mediterranean (Hsu and others, 1978; Ryan and others, 1973). The
more typical cores contain numerous tephra and sapropel layers,
and have changes in lithology every few centimetres. Because of its
structureless nature and because initially we had no idea what
genetic name to give this sediment type, we nicknamed it "homo-
genite."
Distribution
Of the four survey areas, the western two sites contain the
acoustically transparent layer diagnostic of homogenite, whereas
the eastern two sites show no trace of this sediment.
Within each of the homogenite-bearing survey areas, homo-
genite was found only on the basin floors. Figures 4 and 5 show the
bathymetry of the Calabrian Ridge site, the distribution of the
acoustic transparent layer, and the location of homogenite-bearing
cores. Within the 100-km
2
survey area, 18 separate homogenite-
filled basins were located. Homogenite is conspicuously absent
from the steep basin walls and from the plateaus which separate the
basins. Instead, sediments on the plateaus contain a normal eastern
Mediterranean sequence of sapropels and tephras in cores (Blech-
schmidt and others, in press) and are seen seismically as closely
spaced, subparallel reflectors. Cores from the walls are similar to
plateau cores, but are frequently missing the top of the section.
Age
In every core, including those separated by 300 km, the
homogenite occupies an identical stratigraphic position (Fig. 6). It
is overlain by 16 to 36 cm of apparently normal pelagic ooze. It is
underlain by 15 to 32 additional centimetres of normal pelagic sed-
iments and then by the SI sapropel layer. The SI sapropel, an
organic-carbon rich layer, is a widespread marker bed firmly dated
at 8,000 yr B.P. (Kidd and others, 1978). At the stratigraphic
position where homogenite would be expected, plateau cores con-
tain instead a few centimetres of fine silt-clay laminae, implying a
redeposition events of less magnitude but of the same age as the
homogenite event. If the post-homogenite and the pre-homogenite-
post-Sl sedimentation rates are assumed equal and uniform, and
the top of the SI sapropel is assumed to be 8,000 yr old, one can
calculate a date for the redeposition event from the ratio of pre- to
post-homogenite sediment thicknesses (Fig. 6). With one exception,
these dates fall between 3100 and 4400 yr B.P.
Provenance
Two conflicting hypotheses might be suggested for the source
of the homogenite. The first involves sediment transport from a
distant source area, perhaps the continental shelf, by some large-
scale process such as a major turbidity current. This hypothesis
explains the simultaneous deposition of homogenite in many b
asins by a single sediment transporting agent. The alternative
hypothesis is that homogenite comes from an adjoining source area,
and that the transport distance was less than a kilometre, from
basin walls to basin floor. This hypothesis requires some wides-
pread event capable of initiating local sediment transport in many
basins simultaneously.
We prefer the adjoining source hypothesis for the following
reasons:
1. The homogenite contains no allochthonous elements, such
as quartz sand or unusually abundant terrigenous clay, which
require a source outside the survey area. In fact, both the mineral-
ogy and micropaleontology are a weighted average of the properties
of the varying lithologies in the uppermost part of the normal pla-
teau pelagic sequence. For example, the average organic carbon
content of 43 homogenite samples from 3 cores at the Calabrian
Ridge site was 0.36%, which is intermediate between the organic
carbon content of marls (0.29%; Cita and others, in press) and of S1
sapropel (2.8%; this study).
on February 24, 2015gsabulletin.gsapubs.orgDownloaded from
17"40'E 17°4l'E 17°42'E 17'43'E ^7°44*E 17°45'E 17°46' 17°47'E
CALABRIAN RIDGE
CONTOUR INTERVAL:
\0
FATHOMS,
UNCORRECTED
j
I I I I L
36°
20 ' N
36°
19'N
36°
<8'N
36°
17'N
36°
)6'N
36°
15'N
36°
14'N
36"
\ 3'N
Figure 4. Bathymetric map of the Calabrian Ridge survey area. Elongate basins run subparallel to the regional trend of the Calabrian
Ridge. These troughs have flat bottoms and steep walls (up to 45°) and are separated by flat or gently dipping plateaus. Notice the small
scale; each basin is approximately 1 km across.
on February 24, 2015gsabulletin.gsapubs.orgDownloaded from
T
^
| /
EROSIONAL AREA
© CORE
^ TRANSPONDER
"PELAGIC* AREA
"HOMOGENITE" AREA
ISOPACHS
AT 0, 5,
7.5
B 10 MILLISECONDS,
2-WAY TRAVEL TIME
36°
19'N
36°
18'N
36°
I7'N
36°
16'N
36°
15'N
36®
H'N
36°
( 3'N
17°40E 17°41'E Í 7"42'E 17*43 *E 17'44'E
f7°45'E
17'46'E ^7
0
47'E
Figure 5. Distribution of the acoustically transparent layer and of homogenite-bearing cores at the Calabrian Ridge site. Homogenite is
found only in the basins. The "pelagic" facies, found on plateaus, is characterized by closely spaced subparallel reflectors on 4-kHz records
and a normal Mediterranean sequence of sapropels and tephras in cores. The "erosional" facies, found on steep slopes, is reverberant on
seismic-reflection records. Piston cores from the "erosional" facies lack the top of the section.
on February 24, 2015gsabulletin.gsapubs.orgDownloaded from