OCEANOLOGICA ACT A 1983 - VOL.

6 - N' 3 ~----

Sedimentary processes Sonographs

Sediment waves
Eastern boundary current
on .the Iberian continental margin Mediterranean outflow
Submarine canyons
Image acoustique
viewed by long-_range Rides sédimentaires
Courant géostrophique
Écoulement de l'eau méditerranéenne
side-can sonar Canyons sous-marins

Part 1.: Gulf of Cadiz

J.V. Gardnera, R.B. Kidd

Institute of Oceanographie Sciences, Wormley, Godalming, Surrey GUS 5UB, UK.
a Permanent address: US Geological Survey, Menlo Park, California 94025, USA.

Received 19/11/82, in revised form 7/3/83, accepted 24/3/83.

ABSTRACT A long-range side-scan sonar (Gloria) survey of the central and western Gulf of Cadiz
reveals fields of sediment waves in water depths ranging from less than 2 000 to over
4000 m. The sediment waves in water depths less than 2000 rn have an avera·ge
amplitude of 23 rn and an average wavelength of 1 892 m. These bedforms appear to
be directly related to Mediterranean outflow. Sediment waves in water depths greater
than 2 000 rn are different in size from those shallower than 2 000 rn; they have average
wave heights of 8 rn and average wavelengths of 1121 m. These bedforms are probably
not related to Mediterranean outflow, but rather are the result of a proposed eastern
boundary current that appears to flow along the Northwest African margin, into the
Gulf of Cadiz, and along the Western lberian margin.
Submarine c·anyon systems show the development of gulleys and dendritic-like drainage
systems in the upper portions of the canyons. The floors of the canyons appear to have
large blocks of material scattered along axis. The lower regions of the canyons merge
onto the Seine and Horseshoe abyssal plains with no apparent development of fans,
.
distributary channels, levees, etc .
Oceanol. Acta, 1983, 6, 3, 245-254.

RÉSUMÉ Processus sédimentaires dans la marge continentale ibérique observés par

sonar latéral à longue portée. 1re partie : le Golfe de Cadix
Une reconnaissance effectuée dans le Golfe de Cadix à l'aide d'un sonar latéral à longue
portée (Gloria), révèle l'existence de rides sédimentaires de 1000 àplus de 4000 rn de
fond. Au-dessus de 2 000 rn, leur amplitude moyenne est de 23 rn et leur longueur
moyenne de 1 892 m. Ces structures sont dues à l'écoulement dans l'Atlantique des eaux
d'origine méditerranéenne. Au-dessous de 2000 rn, les rides sont moins étendues, avec
une amplitude moyenne de 8 rn et une longueur de 1121 m. Ces figures sédimentaires
profondes sont alors plutôt la conséquence du passage d'un courant géostrophique aux
limites orientales de l'Atlantique. Dirigé vers le Nord, ce courant longe la marge au
nord-ouest de l'Afrique, pénètre dans le Golfe de Cadix, puis remonte le long de la
péninsule ibérique.
A la tête des canyons sous-marins, les images acoustiques mettent en évidence de
nombreux affluents organisés en réseau de drainage dentritique. L'axe des canyons
semble obstrué par de grands blocs effondrés. Dans leur cours inférieur qui rejoint les
plaines abyssales de la Seine et du Fer-à-Cheval, on ne remarque aucun éventail
profond accompagné de chenaux ou de levées.
Océanol. Acta, 1983, 6, 3, 245-254.

0399-1784/1983/245/$ 5.00/ c Gauthier-Villars 245

 J. V. GARDNER. R. B. KIDD

INTRODUCTION

Continental margins, especially the continental slope,

are typically thought of as areas of major down-slope
sediment movement. This province necessarily is the
zone through which sediment must pass as it works its
way to the continental rise and abyssal plains. One
consequence of this view is the impression that ·
continental slopes are replete with ali scales of slumps,
slides and general mass movement features. M uch of
the work on the slope for the past 30 years bas
concentrated on these often spectacular features
(Heezen, Ewing, 1952; Heezen, Drake, 1964; Embley,
1976; Embley, Jacobi, 1977; Malahoff et al., 1980, to
name just a few). These studies have typically relied on
narrow-beam echo sounding, high- and law-resolution
seismic-profiling surveys and more recently surveys by
mid- to short-range side-scan sonar (Belderson, Kenyon,
1976; Kenyan et al., 1978; Ryan, 1982; McGregor et al.,
1982; Twichell, Roberts, 1982). Most ofthese techniques
provide narrow corridors of data often separated by
tens of kilometres between adjacent profiles.

Gloria Mark II is a dual side-scan sonar system that

provides almost the equivalent of a continuous regional
sanie photograph of up to 60-km maximum swath of
the sea-floor perpendicular to the ship's track. A zone Figure 1
about 5 km to either side of the ship's track is typically Location map showing Discovery 90 Gloria sonographs coverage of the
not insonified. This system transmits a linear FM sweep Gulf of Cadiz and southern lberian margin (stippled). The 4,000-m
with a bandwidth of 100Hz and an operating frequency water depth contour is as drawn by Laughton et al. (1975). The area
within box 1 is the central and western Gulf of Cadiz, the subject of this
of 6.5 kHz (Somers et al., 1978). Because of electronics paper. The southwestern lberian margin (box 2) and also the
limitations, the nominal separation of targets down northwestern lberian margin, will be the subjects of subsequent papers
in this journal.
range is 45 rn (at the 60 km total-range setting) and the
nominal resolution parallel to the ship's track is
40 sec. x ship's speed in rn/sec. (200 rn at 10 kts). METHODS USED IN INTERPRETATION

The methodology used in this study was to first

In 1978 Discovery Cruise 90 of the Institute of annotate a transparent Gloria-II track chart with
Oceanographie Sciences collected over 4 000 km of detailed observations from the airgun and 10-kHz
analogue Gloria-II data, seismic-reflection and 10-kHz profiles. Annotations included the position and facing
data along the continental margin of the Iberian directions of ali changes in slope, occurrences of
peninsula (Fig. 1). The survey was run at speeds of 7 to sedimentary bedforms, echo-sounding types, and any
'9 knots and controlled by satellite navigation. The feature suspected of being directly related to underlying
sonograph images shawn in this report are the original structure (diapirs, fault blacks, etc.). Next, the Gloria-
slant-range sonographs anamorphically corrected for II mosaics were interpreted on this same transparent
variations along track (Rusby, Somers, 1977). Prelimi- overlay. The last step involved overlaying the interpre-
nary studies of these data have been reported by tations on a bathymetrie chart to look for relationships
Roberts and Kidd (1982), and Kidd and Roberts (1982). between the Gloria-II features and the bathymetry.
The Gloria-II sonographs commonly show extremely
law-relief lineations that undoubtedly are bedforms.
The present study interprets in detail the data covering Although sorne of these features cao be traced for over
the entire western Iberian margin and the western and 5 km, they are so subtle that, although discernible on
central Gulf of Cadiz in terms of the major sedimentary the original sonographs, we were unable to illustrate
features. Because the area is qui te large and the coverage them adequately in this paper. We feel that these
is dense, the lberian margin has been subdivided into bedforms are either low amplitude sediment waves or
three segments: central and western Gulf of Cadiz; erosional furrows, because they trend in a consistent
southwestern Iberian margin; and northwestern Iberian direction and are often sub-parallel to the larger-scale
margin. This report deals only with the Gulf of Cadiz bedforms described later. This observation illustrates
segment (box labelled 1 in Fig. 1). The two remaining . that law-relief bedforms are quite common on this
segments are at present in preparation for publication section of the sea-floor and that Gloria-II is capable of
in this journal. resolving quite subtle features.

246
 SED IMENTARY PROCESSES IN T H E GULF OF CAD IZ

The 10-kHz profiles taken at the ti me of the Gloria-Il sediment movements. We have concentrated on
survey show features that cannot be matched one-for- identifying features related to sedimentary processes
one with features seen on the sonographs. This is and have attempted, using the above methods, not to
primarily due to the approximately 5-km zone on either confuse these with structurally related features . How-
side of the ship's track that typically is not insonified. ever, the successful identification of sorne features
But in addition, this Jack of resolution by Gloria-Il is, remains unequivocal.
in part, a function of the geometry of the features . The
height resolution of Gloria-Il is a function of frequency
(6.5 kHz) and is theoretically somewhat less than 20 cm.
We believe that the horizontal dimensions, linearity or REGIONAL FRAMEWORK OF THE GULF OF
Jack thereof, the facing slope angles of the features, CADIZ
their orientation relative to the ship's track, and their
distance away from the track ali play a deciding role in The general bathymetry of the Gulf of Cadiz (Fig. 2)
whether or not a feature will backscatter enough energy has been discussed by Heezen and Johnson (1969) and
to be resolved on Gloria-II analog sonographs. For Roberts (1970). The Gulf of Cadiz is an eastward-
instance, ali channels found on the 10-kHz records were projecting cul-de-sac in an otherwise north-south
resolved on Gloria-Il sonographs. It is probable that continental margin. The continental slope within the
many of the bedforms not resolved on Gloria-II Gulf of Cadiz has quite gentle gradients in the central
sonographs are sinuous in shape and short crested. region (about 0.01 along 36°N) and only steepens to
Facing slope angles of bedforms were calculated from about 0.05° along the southem Iberian and northern
the 10-kHz record and it appears that features with Moroccan sectors. The transition from slope to rise is
facing slope angles > 6° are easily resolved and those subtle; gradients of the continental ri se vary from 0.014°
< 6° are not. These calcula ti ons were made on bedforms along 36°N to 0.005° along the southern Iberian margin.
in water depths of about 2 500 m. It should be Studies based on seismic-reflection profiling have
emphasised that in ali areas where the 10-kHz profiles suggested that the northern and eastern Gulf of Cadiz
show bedforms, the Gloria-II sonographs show were tectonically deformed into long, curvilinear ridges
bedforms; it is the number of individual bedforms that and valleys (Roberts, Stride, 1968). Roberts (1970)
differs. proposed that a large olistostrome was emplaced in the
The olistostrome in the Gulf of Cadiz, proposed by Gulf of Cadiz in the Late Miocene. This proposed
Roberts (1970), does present the problem of separating olistostrome could account for the abnormally gentle
features on the sonographs that would be expected to gradients and the rather disrupted mesoscale bathyme-
result from a large, tectonically implaced mass from try. The Gulf of Cadiz is also at present a seismically
those features related to currents and gravity-induced active area (Udias, 1967).

Figure 2a
~ General bathymetry and traces of features
identified on Gloria-II sonographs. The
bathymetry is plotted at a 500-m contour
interval and is generalised from Laughton
et al. ( 1975). The heavier /ines indicate
negative-relief features such as channels
identified on the sonographs. The lighter
/ines indicate features interpreted as
sediment waves. The areas shawn in
Figures 4 and 6 are indicated by boxes.
Figure 2b
Sonograph mosaic of the Gulf of Cadiz.
See Figure 2 a for an interpretation of the
sonographs.
"

247
J. V. GARDNER. R. B. KIDD

Heezen and Johnson (1969) studied the relationship SEDIMENT WAVES FROM THE GULF OF CADIZ
between microtopography and the flow of the APPARENT WAVELENGTH (m) AMPLITUDE (m)
Mediterranean undercurrent in the northern Gulf of
Cadiz. Melières et al. (1970) correlated features from n. 32
"
'•2000m DEPTH

bottom-photographs with the Mediterranean undercur- la23m

O'•t19m
rent in the northern Gulf of Cadiz. Both studies found
a high correlation between the core of Mediterranean 0 000 1000
,.Jll,lia
1500 2000 2500
mg'F'
3000 0102010408010
undercurrent, identified by the 35.5% isohaline, and 2000-3000 m DEPTH
current-swept features such as rock, sand and gravel ".84
i. 983
surfaces, large bedforms, scour, etc. Hydrographie 0 s:t379

studies have suggested that the path of the Mediterra-

nean undercurrent in the northern Gulf of Cadiz may
be influenced by bathymetry, especially submarine
valleys (Madelain, 1970). However, Zenk (1975a) and 3000-4000 m DEPTH
Ambar and Howe (1979 a; 1979 b) have suggested the
Mediterranean water is differentiated into two distinct
cores near the Strait of Gibraltar and that bottom
topography imparts only a secondary influence on the
outflow water.
Kenyon and Belderson (1973) advanced the earlier
geological studies of the central Gulf of Cadiz by using
Gloria-! medium-range (7 and 19-km swath) and shorter
range (1-km swath) side-scan sonar surveys. Their
•4000 m DEPTH
survey is located at about 36°N and ranges from 6°30'W n • 24
i .em
n a2ot
ia1247
to 8°W (north and east of the area reported on here).
]
,.dsm CJ'•l824

Their study refined the distribution and types of m,w,m ... ,w,JJ!ll , ..,
bedform features (sand ribbons, sand waves, large 0 500 1000 aooo
1100 2100

mudwaves, depositional ridges, channels, etc.) that

Figure 3
directly underlie the Mediterranean undercurrent. A
Histograms of sediment wave geometries measured on 10kHz records.
systematic progression of features was related to the The sediment waves are subdivided by water depth into 1,000 m depth
decreasing flow regime of the Mediterranean undercur- intervals.
rent.

The entire central and western Gulf of Cadiz is covered

DISTRIBUTION OF MAJOR FEATURES with scattered fields of sediment waves. They appear on
the sonographs as groups of linear, en échelon positive
Major sedimentary features identified on Gloria-II features, typically about 2-km long (Fig. 2 a).
sonographs in the Gulf of Cadiz, in order of importance, We compiled measurements of wave height and
are: 1) sediment waves and other bedforms; 2) submarine apparent wavelength of ali sediment waves found on
canyon systems; and 3) channels. Ali available the 10-kHz profiles (Fig. 3). Wavelengths are apparent
seismic-reflection records crossing the Gulf of Cadiz because the ship's track crossed bedforms at various
were scanned so that outcrops of acoustic basement angles and individual crests of bedforms commonly do
would not be misinterpreted as sedimentary features. not correlate between the 10kHz profile and the
Gloria-II sonograph. The sediment waves fall into two
classes that each occur in separate bathymetrie
Bedforms provinces. Relatively high-amplitude (mean = 23 rn)
long wavelength (mean = 1 892 rn) sediment waves
The terminology to describe different bedforms from occur between 1 000 and 2 000 rn water depth (Fig. 3).
the deep sea can, at first, be misleading. Terms such as They trend parallel to sub-parallel to bathymetrie
mudwaves and sandwaves are often used, but these contours and are both symmetrical and asymmetrical in
terms imply that the sediment facies is known. Lonsdale cross-section. The compilation by Lonsdale and Spiess
and Spiess (1977) defined mudwaves and sandwaves not (1977) suggests mudwaves are aligned roughly parallel
by facies but rather by their dimensions; mudwaves to the relatively fast currents that formed them.
have wavelengths of several kilometres and amplitudes · Accepting this interpretation, the sediment waves
of tens of metres, whereas sandwaves have wavelengths between 1 000 and 2 000 rn depth appear to be responses
of 10 to 100 m. Ail of the bedfonns described here fall to relatively strong currents within this depth zone.
into the above category of mudwave because of their The region between 2 000 and 4 500 rn water depth has
size as measured on the 10kHz profiles. However, we sediment waves of rather uniform dimensions (Fig. 3).
use the term sediment wave because it is a purely These bedforms have mean amplitudes of about 8 rn
descriptive term and because we have few descriptions and mean wavelengths of about 1121 m. The crests of
of surface sediment that can be usefully related to these sediment waves trend • sub-parallel to the
bedforms seen on the sonographs. bathymetrie contours and appear to be oriented around

248
 SEDIMENTARY PROCESSES IN THE GULF OF CADIZ

Figure 4
Line tracing of the sediment waves in
the western Gulf of Cadiz seen on the
sonograph mosaic. See Figure 2 for
location. Note the "steering" of the
sediment waves around the 1,000-m
high bathymetrie feature. Contour
interval is 200 m. Heavy line is the
track of the 10kHz record shown in
Figure 5.

large bathymetrie highs, such as shown in Figure 4. The 3000

sediment waves in these depths found on a slope are
asymmetrical in cross-section on 10-kHz profiles (Fig. 5) · 3200 E
but they are nearly symmetrical where found on flat ,:z:
...
sea-floor. The smaller dimensions of these bedforms
3400
......
Q
suggests that flow in this deeper region may have been a:
different than the flow above 2000 m. Alternately, these ..."'
sediment waves may be relict features that have been 3600
"'
il:
considerably re-worked and partially destroyed.
3800
0054 0036 0018 oooo;049
SPEED• 7·9KTS COURSE •087• V. E. • 10x

Submarine canyon systems

Figure 5
10-kHz echo-sounder profile over the 1,000-m high bathymetrie feature
Three large submarine canyon systems were surveyed showing the sediment waves that ornament its surface. Figure 4 shows
by Gloria-II in this segment of the lberian margin the ship' s trac k.
(Fig. 2). Saô Vicente Canyon heads at Cabo Saô Vicente
on the inner shelf, incises the shelf, slope and rise, and
leads directly onto Horseshoe abyssal plain at about (Belderson, Kenyon, 1976; Kenyon et al., 1978) and
4 800 rn depth. The second canyon, informally called from the US continental margin (McGregor et al.,
Lagos Canyon here (following Vanney, Mougenot, 1982; Ryan, .1982; ·Twichell, Roberts, 1982).
1981) heads on the upper slope southeast of Cabo Saô Lagos Canyon is almost completely covered by
Vicente at about 1 000 rn depth and, like Saô Vicente Gloria-Il sonographs (Fig. 6a and 6b). The head of
Canyon, incises the slope and rise and leads directly Lagos Canyon is at about 1 800 rn water depth. A
onto Horseshoe abyssal plain. The third canyon, system of gulleys occurs on the northwestern wall of the
informally called Portimao Canyon here (again after upper canyon. These gulleys feed directly into the upper
Vanney, Mougenot, 1981) is located just east of Lagos reaches of the main canyon axis. The canyon proper is
Canyon and appears from the available bathymetry to 2.5 to 3 km wide as it crosses the continental slope.
head on the upper slope and incise only the slope. The Second-order tributaries join the main canyon axis ali
bathymetry suggests that Portimao Canyon joins Lagos along its length from the head to about 3 500 rn water
Canyon at about 3 500 rn depth. depth (Fig. 6).
Portimao Canyon has the least Gloria-Il coverage of Features up to 1 km long, resolved on the floor of the
. the three canyon systems; only the upper reaches of the upper section of Lagos Canyon, appear to be blocks,
canyon were surveyed. It appears on Gloria-II records possibly similar to the large slide blocks described by
as a narrow, steep-walled main canyon axis with McGregor and Bennett (1979). These features are
numerous tributary side-canyons joining the axis along aligned roughly perpendicular to the canyon axis and
its course (Fig. 2). The tributary side-canyons also have they may actually be large masses of coherent material
their own tributary side-canyons (second- and third- that have spalled off the upper canyon walls but have
order tributaries) joining into them. A zone about not been transported too far down axis. The mid-
30 km wide is covered by a dendritic-like drainage canyon area bas features of similar dimensions to those
system feeding into the main axis that suggests head ward in the upper canyon, but here they appear to be aligned
erosion (Fig. 2). Many of the second- and third-order . roughly parallel to the axis, as if following flow lines.
tributaries head on the continental slope, not the shelf. Farther down-axis these features are not present but the
This pattern is similar to that described from side-scan channel fill shows high backscattering (Fig. 6b) that
sonar from the northern margin of the Bay of Biscay suggests either considerable bottom roughness or a

249
J . V. GARDNER, R. B. K\00

SÀO VICENTE CANYON

) - - OEPOSITION.Al L08E

b
Figure 6
Line drawing (6a) and sonograph mosaic (6b) ofthefeatures associated
with Lagos and SaoVicente canyons. See Figure 2for location. Heavy
line is the track of Discovery 126 3.5-kHz profile shown in Figure 1.
a

depositional lobe. This Jack of coverage makes the

coarse sediment facies, or a combination of both . This relationship of the depositional lobe to the canyon
zone may be one of sheet flows or grain flow that have system equivocal. The depositional lobe is either the
by-passed the slower moving large blocks. distal part of the Lagos Canyon system or an interfluve
Lagos Canyon abruptly changes its form at 3 500 rn slump feature, similar to those described by Kenyon et
water depth from an incised canyon to a feature of very al. (1978). The depositional lobe appears roughly
low relief, very high reflectivity, and very irregular semicircular with concentric features having a maximum
outline. The 10-kHz echosounder profiles show a change diameter of about 30 km (Fig. 6).
in acoustic character over this feature. As the zone is
approached, the profile shows a relatively flat, featureless
surface with a strong surface echo and faint, continuous @ E w.- 3225
sub-bottom reflections. The bottom changes within the ~ 3300

feature to a small-scale hummocky surface that produces

..."'~
7km
- 3 375
a diffuse surface echo of small-scale hyperbolae tangent
to the sea-floor and no sub-bottom reflections. The
- 3 450 "':lE
z
-3525 -
sonographs show a very marked increase in reflectivity,
suggesting an increase in bottom roughness, that ~ -3600 ~
outlines this zone. The acoustic characteristics strongly · ~- 3675
suggest that this zone is composed of debris flow ' "' r3760

deposits. The probable debris flow deposits vary from ~

about 8.5 to 14.5 km wide, are about 40 km long, and @E w-3ns
have completely filled the lower portion of Lagos - coso
Canyon. The lower half of the Lagos Canyon system 7km "'"'
has what appears to be meandering channels parallel to
- 4125
..."'
the regional slope that abruptly appear and disappear - ·\200 "':lE
!
(Fig. 6). These may be former tributary channels that - 4275

have been partially blocked off and filled by debris flow

1 .. l4 u; .., r;
«ii .._4350
......
:z:

·"''. :!:
deposits. The path of Lagos Canyon makes a right- - 4425

angle bend from a trend of 70 to 180° just at the - 4500

.!,
~
1 1
beginning of the debris flow. ~ s
0
~0 ~

The only portion of the Lagos Canyon system not

Figure 7
*
~ 0 0 ~

covered by Gloria-II sonographs is the transition

Discovery 126 3.5-kHz profile across the depositional lobe. (A) is
between 4 200 and 4 400 rn water depths from the outside and (B) is within the depositional lobe. See Figure 6a for
probable debris flow deposits to a very low-relief location of the track.

250
 SEDIMENTARY PROCESSES IN THE GULF OF CADIZ

A 3.5-kHz profile across the northernmost part of the Our sonograph coverage does not show the presence of
sonograph (Fig. 7) shows an abrupt change from submarine fans and fan systems along the northern
parallel continuous, strong surface and sub-surface Gulf of Cadiz such as those shown by Vanney and
reflectors outside the depositionallobe to a hummocky, Mougenot (1981). We believe that sediment transported
diffuse surface echo with no sub-bottoms over the to the outer shelf and slope is first intercepted. by the
depositional lobe. The lack of sub-bottom layering strong flow of Mediterranean outflow. Sediment not
within the lobe discounts its interpretation as a entrained in the Mediterranean outflow is re-worked by
submarine fan lobe. A series of meandering tributary the eastern boundary current and eventually transported
channels on the southeastern margin of the lobe appears northward. Consequently, even though fairly sizeable
to have been partially buried by the depositional lobe rivers such as the Guadiana and Guadalquivir rivers
(Fig. 6a). The depositional lobe grades into the flat, flow into the Gulf of Cadiz, the sediment they con tri bute
featureless expanses of Horseshoe abyssal plain without is re-worked along slope by geostrophic flows and not
any surface expressions of channels, levees, distributary down slope by gravity-induced processes.
features, etc.
The third canyon in this segment of the lberian margin
is Saô Vicente Canyon (Fig. 6). This canyon heads at
the shelf break, incises both the slope and rise, and Channel-like features
feeds directly onto Horseshoe abyssal plain. The
Gloria-II coverage includes only the slope and upper Numerous channel-like features, distinguished by
rise portion of the canyon. The trend of Saô Vicente negative relief, appear on the Gloria-II sonographs from
Canyon is toward the south-southwest, transverse to the Gulf of Cadiz but they are not so continuous that
the regional bathymetry. The main axis of the canyon they can be connected into a true channel system
is about 3 km wide and about 45 km long. The canyon (Fig. 2). In general, these channels do not have levées,
walls show a series of linear and curved features in are less than 100 rn deep, are eut traverse to the regional
water depths between about 2 000 and 3 000 rn that may bathymetry and appear to meander. Channels are most
be slump folds or sediment failures. Like Lagos Canyon, common on the upper slope ( < 1 500 rn deep), common
Saô Vicente Canyon has a broad floor with well- to rare on the lower slope (1 500 to 2 500 rn deep), and
developed features that may be large detached masses. rare to absent on the rise ( > 2 500 rn deep). Many of
These features trend roughly perpendicular to the main the channel-like patterns interpreted from the bathy-
axis in the continental si ope section of the canyon axis, metry of Heezen and Johnson (1969), Hydrographie
but trend parallel to the axis in the upper rise section. Department (1969), and Laughton et al. (1975) are not
These features appear to be a succession oflarge blocks, apparent on the Gloria-II sonographs. The Jack of
similar to those described in Lagos Canyon. Gloria-II definition of these bathymetrie interpretations
The Gloria-II sonographs also show an increase in may be a result of either their small relief, or very low
reflectivity (Fig. 6 b) on the floor of Saô Vicente Canyon angles of the walls, or of inaccurate bathymetry. Many
that indicates the presence of either a coarse sediment of the channels seen on the sonographs on the upper
facies or bottom roughness or both. Concentric features slope have second- and third-order tributaries but none
similar to the ones described at the terminus of Lagos of them appear to have distributary systems. The most
Canyon, occur just west of the canyon axis in 3 000-to surprising yet consistent characteristic of the channels
4 200-m water depth (Fig. 6). The pattern adjacent to is their short length and Jack of interconnections. The
Saô Vicente Canyon appears to reflect a series of channels are rarely more than 15 km long and many are
interfluve slumps and possible slump folds generated by only 3 to 4 km long. Sorne may be related to surface
sediment creeping down into the lower reaches of Saô morphology on an olistostrome, as postulated for the
Vicente Canyon. region by Roberts (1970).

@ ® @ ©
0 0 36· 2---
36·30
.---a&·oo-
35·70
500

1000

Figure 8
Two profiles of temperature and
1500

2000

0
.. 1500

500
500
salinity showing the tangue of Medi-
terranean outjlow water as it leaves
1000 1000
contact with the sea-jloor (modified
from Made/ain, 1970). Note the
progressive deepening of the influence 1500 1500
of Mediterranean outjlow water from
east to west. See Figure 9 for locations
-~-
2000
of A-B and C-D.

251
J. V. GARDNER, R. B. KIDD

high eustatic sea levels throughout much of the late

Neogene.
If pelagie sediment requires higher flow regimes to form
DISCUSSION
sediment waves than occur here today, then possibly
the sediment waves could have initially formed during
The Mediterranean outflow has been investigated in bursts of Mediterranean outflow as pulses of dense
sorne detail (Heezen, Johnson, 1969; Madelain, 1970; water spilled into the Gulf of Cadiz just after eustatic
Melières et al., 1970; Kenyon, Belderson, 1973; Zenk, transgressions. Another possibility is that Mediterranean
1975b; Ambar, Howe, 1979a; 1979b and references water was more saline (denser) in the past and bence
therein). The Mediterranean outflow water leaves penetrated deeper with greater speeds than occurs
contact with the· sea floor at 1 000 rn in the eastern today.
region and at 1 400 m in the western region (Fig. 8).
The relatively large-amplitude sediment waves in the However, the smaller amplitude sediment waves found
northern Gulf of Cadiz that occur between 1 000 and between 2 000 and 4 000 m depth probably were not
2 000 rn depth are probably related to warm, salty, produced by Mediterranean outflow water. The
sediment-taden Mediterranean outflow (Kenyon, Bel- orientation of the sediment waves around bathymetrie
derson, 1973; Thorpe, 1972) possibly coupled with contours implies a geostrophic flow and the existence of
effects of internai waves. Kenyon and Belderson (1973) a deep eastern boundary current. The proposed flow
mention maximum current speeds of Mediterranean must be northward because a southward flow would be
outflow of 250 cmfsec. and Ambar and Howe (1979b) deflected towards the west by the Coriolis force and
calculate speeds of 181 cm/sec. The sediment facies bence not follow the general bathymetry of the Gulf of
from 1 000 to 2 000 rn depth of the continental slope Cadiz.
has been described as silty lutite with variable amounts The sediment waves occur considerably below the
of foraminifers and sandy interbeds (Heezen, Johnson, present depth of Mediterranean outflow water, which
1969; Diester-Haass, 1973). Current speeds somewhat is restricted to the top of upper Deep Water at depths
less than 100 cm/sec. are required to erode (and bence less than 2000 m (Newmann, Pierson, 1966; Senk, 1975;
form depositional bedforms) in cohesive clay, and Ambar, Howe, 1979b). The sediment waves found
non-cohesive silt and sand would require even slower between 2000 and 4000 m occur within the deep water
speeds (Southard et al., 1971; Lonsdale, Southard, 1974; mass. Cross-sections of temperature and salinity at 36°N
Young, Southard, 1978). Consequently, we believe the show that the deep water-mass structure of the Gulf of
sediment waves found on the Gloria-II survey between Cadiz is similar to that of the deep eastern North
1 000 and 2 000 m depth are a direct result of the shear Atlantic basin (Fuglister, 1960). Defant (1961) suggests
produced on the surface sediments by Mediterranean that a slow ( < 3 cmjsec.), northward-flowing, geostro-
outflow. · phic, deep water-mass may circulate through the Gulf
It is possible that sediment waves found between of Cadiz. Meincke et al. (1975) and Swallow et al.
1 000 and 2 000 rn deep are relict features, but, if so, (1977) report northward flowing water with speeds less
they must have been produced during former periods of than 3 cmjsec.west of Portugal from depths over
eustatic high sea-level rather than low stands. A drop 2000 m. Lonsdale (1978) reported short-term current-
of sea-level during global glacial conditions of only meter data from the Moroccan rise (-4300 m deep)
130 m (Bloom, 1971) would reduce the cross-sectional and the Saharan rise ( .- 3 950 m) off Northwest Africa.
area of the Gibraltar sill, at present about 200 m deep At both localities, maximum recorded current speeds of
(Curray, 1969) by 65 to 90%, depending on the 9 and 19 cmjsec. respectively were recorded above fields
geometry used for the sill area. This reduction in of bedforms and the current directions were all toward
cross-sectional area suggests that if any outflow did the Northeast.
occur during global low sea levels, then the flux was We believe that the bedforms found deeper than 2000 m
probably greatly reduced over today's. Thunell (1979) are a result of a northward flowing deep eastern
has suggested from paleotemperature estimates that the boundary current. Currents with speeds of about
eastern Mediterranean, the region where Mediterranean 3 cm/sec. may not be capable of forming bedforms of
deep water is produced, has had sea-surface temperatures non-cohesive sediment and are even less likely to be
comparable to today's for much of the last 5 million able to form bedforms in cohesive material (Lonsdale,
years. These estimates of sea-surface temperatures Southard, 1974). The few bottom samples reported
suggests that the salinity of the eastern Mediterranean from 2 000 to 4 000 m depth in the Gulf of Cadiz are all
for the late Neogene was often similar to today's. categorised by Heezen and Johnson (1969) as ••normal
Indeed, conditions similar to today's probably have a pelagie" and ..pelagie ooze"; a sediment that should act
much longer history. Deep-sea drilling into pre- in a relatively cohesive manner. However, sediment
Messinian sequences of both the western and eastern waves do occur, suggesting the sediment facies is
Mediterranean has shown that the pattern of water relatively non-cohesive, or perhaps the current strength
mass exchange with the Atlantic Ocean through a is greater, or both, or possibly the sediment waves are
restricted portal in the west was established in the relict from a time when this flow was much stronger.
Lower Miocene (Hsü et al., 1978). Connection was tost The regional bathymetry of the Gulf of Cadiz forms a
with the Indian Ocean during the Burdigalian. It follows cul-de-sac to the proposed northward-flowing boundary
that the formation ofMediterranean outflow water may current. The orientations of the sediment waves, sub-
have been a common phenomena during periods of parallel to the contours, show the path taken by the

252
 SEDIMENTARY PROCESSES IN THE GULF OF CADIZ

deep eastern boundary undercurrent. Figure 9 shows

our interpretation of this flow through the Gulf of
Cadiz.

SUMMARY

Mapped bedforms interpreted from Gloria-II sono-

graphs and 10-kHz echo-sounder profiles from the
central and western Gulf of Cadiz reveal two distinct
classes of sediment waves that occur in distinct
bathymetrie provinces. Sediment waves in water depths
less than 2 000 rn have average amplitudes of 23 m and
average wavelengths of 1982 m. These sediment waves
appear to be direct responses to the strong flow of
Mediterranean outflow. The sediment waves may either
be presently forming or be relict. However, if they are
relict, they must have formed during periods of eustatic
high sea level because of the constraints imposed by the
sill depth of the Strait of Gibraltar on outflow.
Sediment waves from depths grea ter than 2 000 rn have
average amplitudes of 8 rn and average wavelengths of
1121 m. These bedforms are not related to Mediterra-
nean outflow but are responses to an eastern boundary Figure 9
current that flows along the continental margins of the A highly generalised depiction of the flow of l>!editerranean outjlow
eastern North Atlantic. The flow may have been water (black arrows), salinity 36.5°/00 (shaded), and of the proposed
episodic or continuous and may have been stronger in deep eastern boundary current (open arrows). The heavier patterned
area of Mediterranean outjlow water represents the zone where the
the past than it is today. The flow (or current) was water mass is in contact with the sea-jloor. The lighter pattern represents
competent t<:> form large-scale sediment waves of a the region where this water is not in contact with the sea-jloor.
surface sediment that we presume to be of a relatively Lines A-B and C-D are the tracks for the temperature and salinity
profiles in Figure 8.
cohesive pelagie facies. This implies current speeds of
the order of > 10 cmjsec. Current-meter stations west
of Portugal show a persistent northward flow, but with the Department of Energy-sponsored Continental
speeds less than 3 cmjsec. Margin Programme. •
Three submarine canyons were surveyed with Gloria-II. Peter M. Hunter and Amanda M. Bates at lOS are
The upper canyon regions of Lagos and Saô Vicente thanked for their considerable help in producing charts,
Canyons appear to have a disoriented fill of large mosaics and other illustrations that formed the basis of
blocks of debris. Their mid-canyon channels show a this phase of the Gloria interpretation. Constructive
more ordered arrangement of large blocks, possibly reviews by R.H. Belderson, D.G. Roberts, S.A. Thorpe
along flow lines, and the lower canyon region appears and N.H. Kenyon were greatly appreciated as were
to be filled with sediment that has by-passed the blocks. those by two anonymous reviewers. Ali added
The portion of Lagos Canyon across the lower significantly to the interpretations presented here.
continental rise has either been filled by an interfluve
slump or by a depositional lobe of channel fill. Nowhere
did we observe distributary channel systems or even
submarine fan systems along the continental risejabyssal REFERENCES
plain boundary.
The second segment of this lberian margin study, the
Southwestern lberian margin, will discuss in detail the Ambar 1., Howe M.R., 1979a. Observations of the Mediterranean
outflow-1. Mixing in the Mediterranean outflow, Deep-Sea Res., 26A,
importance of submarine canyons to the sedimentary 535-554.
processes of this margin. Ambar 1., Howe M.R., 1979 b. Observations of the Mediterranean
outflow. II. The deep circulation in the vicinity of the Gulf of Cadiz,
Deep-Sea Res., 26A, 555-568.
Be1derson R.H., Kenyon N.H., 1976. Long-range sonar views of
Acknowledgements submarine canyons, Mar. Geol., 22, M69-M74.
B1oom A.L., 1971. Glacial-eustatic and isostatic controls of sea-level
since the last glaciation, in: The Late Cenozoic glacial ages, edited by
K.K. Turekian, Yale Univ. Press, New Haven, 335-379.
Curray J.R., 1969. History of continental shelves, in: The new concepts
The officers and scientists aboard RRS Discovery of continental margin sedimentation, edited by D.J. Stanley, Am. Geol.
during Croise 90, in particular the Gloria technical Inst., Washington D.C., vol. 1-18.
support team, are acknowledged for their major Defant A., 1961. Physical oceanography, vol. 2, Pergamon Press, New
York, 598 p.
contribution to the findings reported here through
Diester-Haass L., 1973. No current reversai at 10,000 B.P. in the
collection of the data. This was a constituent cruise of ' Straits of Gibraltar, Mar. Geol., 15, M1-M9.

253
 J. V. GARDNER. R. B. KIDD

Embley R.W., 1976. New evidence for the occurrence of debris flow Meincke J., Siedler G., Zenk W., 1975. Sorne current observations
deposits in deep sea, Geology, 4, 371-374. near the continental slope off Portugal, Meteor Forschungsergeb.,
Embley R.W., Jacobi R., 1977. Distribution and morphology oflarge 16A, 15-22.
submarine sediment slides and slumps on Atlantic continental Melières F., NesteroffW.D., Lancelot Y., 1970. Étude photographique
margins, Mar. Geotechnol., 2, 205-228. des fonds du Golfe de Cadix, Cah. Océanogr., 22, 63-72.
Fuglister F.C., 1960. Atlantic Ocean atlas of temperature and salinity Neumann G., Plerson W.J. Jr., 1966. Princip/es. of physica/
profiles and data from the International Geophysical Year of 1957- oceanography, Prentice-Hall, Englewood Cliffs, N.J., 545 p.
1958. Woods Hole Oceanogr. Inst., 209 p. . Roberts D.G., 1970. The Rif-Betic orogen in the Gulf ofCadiz, Mar.
Heezen B.C., Ewing M., 1952. Turbidity currents, submarine slumps Geo/., 9, M31-M37.
and the 1929 Grand Banks turbidity current, Am. J. Sei., 250, 849- Roberts D.G., Stride A.H., 1968. Late Tertiary slumping on the
873. continental slope of Southern Portugal, Nature, 217, 48-50.
Heezen B.C., Drake C.L., 1964. Grand Banks slump, Bull. Am. Assoc. Roberts D.G., Kidd R.B., 1982. Sedimentary and structural patterns
Petrol. Geol., 48, 221-233. of the lberian continental margin; an alternative view of continental
margin sedimentation, Geology (in press).
Heezen B.C., Johnson G.L., 1969. Mediterranean undercurrent and
microphysiography west of Gibraltar, Bull. lnst. Oceanogr. Monaco, Rusby J.S.M., Somers M.L., 1977. The development of the Gloria
67, 1-95. sonar system from 1970-1975, in: A voyage of Discovery, edited by
M. V. Angel, Suppl. to Deep-Sea Res., 24, 1-7.
Hsü K.J., Montadert L., Bernoulli D., Cita M.B., Erickson A., Ryan W.B.F., 1982. Imaging of submarine landslides with wide-swath
Garrison R., Kidd R.B., Melières F., MüUer C., Wright R., 1978. sonar, in: Marine slides and other mass movements, edited by S. Saxov
History of the Mediterranean salinity crisis, in: Initial Reports Deep- and J.K. Nieuwenhuis, Plenum Press, New York, 175-188.
Sea Drilling Project, 42, edited by K.J. Hsü, L. Montadert et al., US
Gov. Printing Office, Washington D.C., 1053-1078. Somers M.L., Carson R.M., Revie J.A., Edge R.H., Barrow B.J.,
Andrews A.G., 1978. Gloria-II: an improved long-range sidescan
Hydrographie Department, 1969. Admiralty Chart C6101, UK. sonar, Ocean. Int. 1978, London, 16-24.
Kenyon N.H., Belderson R.H., 1973. Bedforms of the Mediterranean Southard J.B., Young R.A., HoDister C.D., 1971. Experimental erosion
undercurrent observed with side-scan sonar, Sediment. Geo/., 9, 77-99. of calcareous ooze, J. Geophys. Res., 76, 5903-5909.
Kenyon N.H., Belderson R.H., Stride A.H., 1978. Channels, canyons Swallow J.C., Gould W.J., Saunders P.M., 1977. Evidence for a
and slump folds on the continental slope between south-west Ireland poleward eastern boundary current in the North Atlantic Ocean, /nt.
and Spain, Oceanol. Acta, 1, 3, 369-380. Council Exp/or. Sea, C.32, 1-11.
· Kidd R.B., Roberts D.G., 1982. Long-range sidescan sonar studies of Thorpe S.A., 1972. A sediment cloud below the Mediterranean
large-scale sedimentary features in the North Atlantic, Bull. lnst. geol. outflow, Nature, 239, 326-327.
Bassin Aquitaine, 31, 11-29. Thunnell R.C., 1979. Climatic evolution of the Mediterranean Sea
Laughton A.S., Roberts D.G., Graves R., 1975. Bathymetry of the during the last 5.0 million years, Sediment. Geol., 23, 67-79.
Northeast Atlantic, Deep-Sea Res., 22, 791-810. Twichell D.C., Roberts D.G., 1982. Morphology, distribution, and
Lonsdale P., 1978. Bedforms and benthic boundary layer in the North development of submarine canyons on the United States Atlantic
Atlantic: a croise report of Indomed leg Il, SIO ref. 78-30, 15 p. continental slope between Hudson and Baltimore Canyons, Geology,
10, 408-412.
Lonsdale P., Southard J.B., 1974. Experimental erosion of North Udias A~, 1967. The focal mechanism of earthquakes in the southern
Pacifie red clay, Mar. Geol., 17, M51-M60. coast of the lberian peninsula, Tectonophysics, 4, 229-234.
Lonsdale P., Spiess F.N., 1977. Abyssal bedforms explored with a Vanney J.R., Mougenot D., 1981. La plate-forme continentale du
deeply towed instrument package, Mar. Geol., 23, 57-75. Portugal et des provinces adjacentes. Analyse géomorphologique,
Madelain F., 1970. Influence de la topographie du fond sur Mem. Serv. Géol. Portugal, 28, 145 p.
l'écoulement méditerranéen entre le détroit de Gibraltar et le Cap Young R.N., Southard J.B., 1978. Erosion of fine-grained sediments:
Saint-Vincent, Cah. Océanogr., 22, 43-61. sea-floor and laboratory experiments, Geo/. Soc. Am. Bull., 89, 663-
Malahoff A., Embley R.W., Perry R., Fefe C., 1980. Submarine 672.
mass-wasting of sediments on the continental slope and upper rise Zenk W., l975a. On the origin of the intermediate double maxima in
south of Baltimore Canyon, Earth Planetary Sei. Lett., 49, 1-7. T-S profiles from the North Atlantic, Meteor Forschungsergeb., 16A,
McGregor B.A., Bennett R.H., 1979. Mass movement of sediment on 34-43.
the continental slope and rise seaward of the Baltimore Canyon Zenk W., 1975 b. On the Mediterranean outflow west of Gibraltar,
Trough, Mar. Geo/., 33, 163-174. Meteor Forschungsergeb., 16A, 23-34.

254