Marine Geology 220 (2005) 181 – 204
www.elsevier.com/locate/margeo
Tidally incised valleys on tropical carbonate shelves: An example
from the northern Great Barrier Reef, Australia
Peter T. Harris a,*, Andrew Heap a, Vicki Passlow a, Michael Hughes b,
James Daniell a, Mark Hemer a, Ole Anderson c
a
Marine and Coastal Environment Group, Geoscience Australia, GPO Box 378, Canberra ACT 2601, Australia
b
School of Geosciences, University of Sydney, Sydney NSW 2006, Australia
c
Kort and Matrikelstyrelsen, Geodetic Division, Rentemestervej 8, DK-2400 Copenhagen, Denmark
Received 23 November 2004; received in revised form 19 May 2005; accepted 20 June 2005
Abstract
The formation of incised valleys on continental shelves is generally attributed to fluvial erosion under low sea level
conditions. However, there are exceptions. A multibeam sonar survey at the northern end of Australia’s Great Barrier Reef,
adjacent to the southern edge of the Gulf of Papua, mapped a shelf valley system up to 220 m deep that extends for more than
90 km across the continental shelf. This is the deepest shelf valley yet found in the Great Barrier Reef and is well below the
maximum depth of fluvial incision that could have occurred under a 120 m, eustatic sea level low-stand, as what occurred on
this margin during the last ice age. These valleys appear to have formed by a combination of reef growth and tidal current scour,
probably in relation to a sea level at around 30–50 m below its present position.
Tidally incised depressions in the valley floor exhibit closed bathymetric contours at both ends. Valley floor sediments are
mainly calcareous muddy, gravelly sand on the middle shelf, giving way to well-sorted, gravely sand containing a large relict
fraction on the outer shelf. The valley extends between broad platform reefs and framework coral growth, which accumulated
through the late Quaternary, coincides with tidal current scour to produce steep-sided (locally vertical) valley walls. The deepest
segments of the valley were probably the sites of lakes during the last ice age, when Torres Strait formed an emergent landbridge between Australia and Papua New Guinea. Numerical modeling predicts that the strongest tidal currents occur over the
deepest, outer-shelf segment of the valley when sea level is about 40–50 m below its present position. These results are
consistent with a Pleistocene age and relict origin of the valley.
Based on these observations, we propose a new conceptual model for the formation of tidally incised shelf valleys. Tidal
erosion on meso- to macro-tidal, rimmed carbonate shelves is enhanced during sea level rise and fall when a tidal, hydraulic
pressure gradient is established between the shelf-lagoon and the adjacent ocean basin. Tidal flows attain a maximum, and
channel incision is greatest, when a large hydraulic pressure gradient coincides with small channel cross sections. Our tidalincision model may explain the observation of other workers, that sediment is exported from the Great Barrier Reef shelf to the
* Corresponding author.
E-mail address: [email protected] (P.T. Harris).
0025-3227/$ - see front matter. Crown Copyright D 2005 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.margeo.2005.06.019
182
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
adjacent ocean basins during intermediate (rather than last glacial maximum) low-stand, sea level positions. The model may
apply to other rimmed shelves, both modern and ancient.
Crown Copyright D 2005 Published by Elsevier B.V. All rights reserved.
Keywords: Great Barrier Reef; Gulf of Papua; tidal currents; continental shelf; submarine valleys; sea level change; Quaternary
1. Introduction
The creation of continental shelf submarine valleys and the accommodation space of most transgressive estuarine systems is generally attributed to the
erosion and incision of the shelf by rivers (or by
glaciers on high latitude shelves) under low sea level
conditions (e.g., Dalrymple et al., 1994). However,
in some cases, shelf valleys are formed by processes
unrelated to rivers. Examples include valleys cut by
shelf-edge slumps and mass flows (e.g., Jansen et al.,
1987) and those eroded by the action of strong tidal
currents (Dalrymple et al., 1994). Valleys cut by
shelf-edge slumps and mass flows are generally confined to the outer shelf and upper slope, although the
heads of some submarine canyons may extend a
considerable distance landward. Examples are the
Scripps and Monterey Canyons in California (Shepard, 1963) and the bNo GroundQ Channel incised
into the Ganges–Brahmaputra Delta (Kuehl et al.,
1997). These valleys generally become both deeper
and wider in a seaward direction.
Valleys formed purely by tidal erosion processes
are not well documented in the literature, presumably
because the initial valley is usually inferred to have
formed by fluvial or glacial processes during low sea
level stands, followed by further erosion by tidal
currents under transgressive, high-stand conditions.
For example, in his review of the seabed morphology
of the tidally dominated shelf around Great Britain,
Pantin (1991) attributed tidal current scour to the
origin of only a few shelf valleys, such as St. Catherine Deep in the English Channel. Similarly, most
submarine valleys on the macrotidal Korean shelf
are heterogeneous, with tidal current scour overprinting features that were at least partially formed by
glacial or fluvial processes during the Quaternary
glaciations (Chough et al., 2000, 2002). Tidally
scoured sections of such shelf valleys generally exhibit local depressions with closed bathymetric contours
(see also Harris et al., 1995).
In the case of many continental shelves, significant portions of the valleys formed during low sea
level were infilled with transgressive sediment as
sea level rose (e.g., Dalrymple et al., 1994). On
the northeast Australian continental shelf, a combination of factors appears to have conspired to limit
the overall effect of low sea level fluvial erosion.
These factors include (1) the mostly arid climate
(and thus, low river discharge) that has persisted in
Australia throughout the Quaternary; (2) the continent’s relatively low relief and, thus, low river
energy (e.g., Woolfe et al., 1998); and (3) although
alpine glaciers formed in southeast Australia and
Tasmania during the Pleistocene ice ages (Williams,
1984), they did not extend down to sea level and
glaciers have not formed in NE Australia at any
time during the Quaternary. Consequently, shelf valleys formed by tidal current scour are more easily
recognised on the Australian shelf (because we can
discount glacial processes) and a number are found
where tidal ranges are meso- to macro-tidal (2–4 m
and N 4 m, respectively). Examples are the tidally
scoured deeps formed between offshore islands and
the coast within Banks Strait off northeastern Tasmania and Van Dieman’s Gulf in the Northern
Territory (Harris, 1994a).
In the Torres Strait region off northeastern Australia, the occurrence of tidally scoured valleys was
described by Harris (1994b, 2001). In particular, the
erosion and formation of the deeply incised section
of Missionary Passage (up to 60 m deep locally;
Fig. 1A) was attributed to modern tidal current
scour (Harris, 2001). Further to the east, along the
southern part of the Gulf of Papua, Harris et al.
(1996) describe an bincised valley zoneQ that exhibits a complex, offshore–onshore trending seafloor
topography on the middle to outer shelf between 50
m and 120 m water depth (Fig. 1B). The data
available at that time indicated the presence of a
well-developed shelf valley southeast of Bramble
Cay (Fig. 1A), along the northern margin of the
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Great Barrier Reef (GBR), forming a system extending bover a 100 km east–west distance between the
60 m and 120 m isobathsQ (Harris et al., 1996, p.
802). In that study, the shelf valleys were attributed
largely to low sea level fluvial incision and the
143
183
deposition of regressive, fluvial–deltaic complexes
during the late Quaternary. Tidal current scour was
thought to have played a secondary role, by overdeepening some sections of the shelf valleys and
reworking older deposits.
o
144
o
o
145 E
A
10
o
8 S
10
20 30
50 40
20
30
40 50
10
100
10
20
10
20
500 0
60
80
100
90
30 20
100
900 800
30
o
1B
20
1000
50
50
9
1000
700
200
700
900
2 0
0
1
00
40
3
40
0
50
10
50
1 00 00
700
40
1
50
0 00 0
3 45 0
1 0
0
1000
1
600
10
1 0 0
5 0
20
900
10
30
40
10
4 0
2 0
30
Australia
10
1 0
3 0
200
3
10
30
5 0
20
20
20
20
0
20
0
2000
50
40
10
700
10
20
10
Holocene Delta
Reef
Shelf valley
30
2010
10
1
20 0
3 0
2
10
10
10
30
20
20
0
30
1 0
10
1
500
30
30
20
20
30
30
0
2
1000
20
40
1
1 0
30
7
2
10
10
10
o
800
0
50
0
20
40
10
20
10
50 10
30 10
10
10
500
500 0
60
10
40
10
30
1 0
10
200
0
50
20
10 20
10
3
500
30
20
20
10
10
11
0
40
10
10
10
30
20
20
1
10
10
10
20
2 100
10
1
10
10
10
50
30
20
10
10
30
10
1 0
2 0
10 20
10
20
50
200
o
90
20
700
10
30
10
10
100
1900000
40
20
30
2000
10
20
10
10
10
30
10
10
40
40
10
10
30
50
100
70
30
30
20
20
100
30
10
50
10
10
East
Cay
40
10
150
40
10
60
20
20
10
1
0
2
1
1
1 0 0
1 0
Fig. 1. (A) Bathymetric contour map of the Gulf of Papua, Torres Strait and northern Great Barrier Reef based on a new compilation of digital
data provided by the CSIRO Division of Marine Research and the Royal Australian Navy Hydrographic Office. The location of the area shown
in Fig. 2B is indicated. (B) Enlarged section of (A) showing the locations of the Great North East Channel (GNEC), Darnley Valley and Bramble
Valley. Shelf valleys are highlighted by dark grey shading; reefs are shaded light grey and the area less than 30 m in depth is indicated by
vertical bars. The positions of Station 21 and 22, listed in Table 2, are shown. (C) Three-dimensional, shaded relief image showing the same
bathymetry data as in (A), but with superimposed reef outlines extracted from LANDSAT images of the area. The dashed-line boxes marked bAQ
and bBQ are the locations of study areas shown in Figs. 2, 5 and 6). Locations shown include Darnley Island (DI), Missionary Passage (MI) and
Bramble Cay (BC).
184
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Fig. 1 (continued).
Here we present new detailed multibeam sonar data
to show that one shelf valley located along the northern
edge of the GBR (southern Gulf of Papua) forms a
complex network, extending at least 93 km across the
shelf that contains isolated depressions that are as much
as 220 m deep locally (Fig. 1A and B). Our new data
indicate that large sections of the valleys were eroded
during the late Quaternary by tidal currents and in many
cases, appear never to have contained significant river
flows. The existence of these valleys raises important
questions regarding the creation of transgressive
estuarine embayments in such carbonate provinces, as
well as the interaction and coexistence of the large
Papua New Guinea rivers with the northern Great
Barrier Reef throughout the late Quaternary. The question we pose is: what factors have controlled the for-
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
mation of over-deepened (N 200 m deep) shelf valleys
in the northern Great Barrier Reef?
1.1. Study area: the Gulf of Papua and northern Great
Barrier Reef
The Gulf of Papua continental shelf forms a lowgradient platform typically 50–90 m deep and can be
divided into four geomorphic zones: (1) a low-relief,
inner-shelf, Holocene deltaic zone; (2) a high-relief,
mid- to outer-shelf incised valley zone; (3) a highrelief, southern reef zone; and (4) a moderate to lowrelief mid- to outer shelf zone in the east and northeast
(Harris et al., 1996). The deltaic zone is an extensive,
flat, shallow surface in 5–30 m water depth (Fig. 1A).
It is bordered by a relatively steep prodelta region in
20–50 m of water that extends along the coast
between the tide-dominated Fly and wave-dominated
Purari River mouths (Fig. 1A; Harris et al., 1993).
Seismic profiles and radiometrically dated core samples suggest that the Fly Delta is prograding seawards
at an average rate of about 6 m/a (Harris et al., 1993).
Progradation is indicated also by the occurrence of
beach ridges aligned parallel to the coastline (Baker,
1999) and by seismic data that show prodelta muds
downlapping progradationally onto the modern shelf
carbonates that mark the maximum flooding surface
(Harris et al., 1993, 1996; Walsh and Nittrouer, 2003).
Collectively, the Gulf rivers deliver approximately
350 million tonnes/year of sediment and discharge
approximately 13,000 m3/s of water into the coastal
environment (Harris et al., 1996). The Gulf coastline
is dominated by mangrove-forested deltaic islands
having an onshore–offshore orientation (Fig. 1A)
that reflects the influence of tidal processes.
The high-relief, mid- to outer-shelf incised valley
zone has a complex, offshore–onshore-trending seafloor topography on the middle to outer shelf between
50 m and 100 m water depth. Two separate, large
shelf-valleys are recognised in this paper; the northern
valley is called the bBramble Valley,Q named for
Bramble Cay located nearby (Fig. 1B). The southern
valley is called the bDarnley ValleyQ in this paper
(named for Darnley Island, located nearby; Fig. 1B).
Available bathymetric data suggest that the Darnley
Valley extends east–west for about 93 km across the
northern limit of the Great Barrier Reef before turning
south-westwards where it extends at least a further 20
185
km into the Great North East Channel (Fig. 1A–C).
Details of the morphology, seismic character and formative processes of these two valleys, based on newly
collected data as part of this study, are presented and
discussed further below.
The high-relief, southern reef zone is a complex of
barrier and patch reefs in the southern part of the study
area, south of 9830V. The bathymetry is rugged near
steep-sided coral reefs which locally may have vertical
sides. Between the reefs, depths are mostly in the range
of 30–70 m, with some broad shallows less than 20 m
and local depressions up to 100 m deep. The available
bathymetry data compiled for this study shows a branch
of the Darnley Valley extends southwards into the reef
complex east of Darnley Island (Fig. 1B). Water depths
increase rapidly to the east of the barrier reef, which is
located on the shelf margin of the Coral Sea basin (Fig.
1A) in water 120–140 m deep.
The moderate- to low-relief, mid- to outer-shelf
zone lies north of the incised valleys and offshore
from the deltaic zone. Closely spaced, 20–50-m isobaths, trending parallel to the coast, delimit the steep
prograding edge of the Holocene deltaic sediments and
contrast with the onshore–offshore trending ridges and
swales in 50–80-m depth. These ridges are flat-topped
and have been described as bmesaQ-like in profile
(Harris et al., 1993). Based on their morphology, seismic character and core samples, including one radiocarbon date of 33,850 years BP taken from deltaicmarine sediments (massively bedded gray mud with
scattered shell debris throughout) on the top of one of
these banks, they are interpreted to be relict Pleistocene
deltaic deposits (Fig. 1A). Generally, this zone forms a
low-relief plain, gently dipping towards the shelf break
which is in about 140 m of water.
2. Methods
Bathymetry, seabed sediment samples, seismic,
water column and seabed video data were collected
using the Australian National Research Facility, R/V
Franklin during a cruise carried out in January–February, 2002 (Geoscience Australia Survey 234,
National Facility Cruise 01/02; see Harris et al.,
2002). We used a Reson model 8101, 240 kHz swath
system for multibeam sonar mapping together with a
Datasonics 3–7 kHz Chirp Sub-bottom profiler (model
186
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
no. DSP 661/66) and model TTV170S tow fish. Survey
data were collected along evenly spaced track lines,
over a total track length of around 1400 km spread
between two survey areas on the middle and outer shelf
(Fig. 1). Navigation was by differential global positioning system (DGPS).
Sampling stations occupied during the cruise
included the collection of water column profile data
using a Seabird SBE911 CTD, calibrated by surface
and bottom water samples. Seabed sediment samples
were collected using a Smith Macintyre grab and a 1tonne gravity corer, also rigged to work as a piston
corer. An underwater video camera was also deployed
at each station to document the substrate character and
benthic biota.
Surface sediment grab samples were analysed for
percentage gravel, sand and mud content by the sieve
method, using nested 2-mm and 63-Am analytical
sieves. Carbonate content was determined on the
dried gravel sand and mud fractions separately using
a carbonate bomb (Muller and Gastner, 1971).
3. Results
3.1. Acoustic survey data
3.1.1. Survey Area A (middle shelf)
The survey in the middle-shelf area confirmed the
presence of the Darnley Valley submarine valley sys-
Fig. 2. Hill-shaded relief images based on a 10-m grid of multibeam sonar data showing (A) mid-shelf submarine valleys, platforms (P1, P2 and
P3) and submerged reefs; and (B) outer-shelf submarine valleys, platforms (P4 and P5) and submerged reefs. See Fig. 1B and C for the locations
of the images.
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
tem, trending east–west across the shelf (Figs. 1A–C,
2A and B), as described by Harris et al. (1996).
Multibeam survey results collected in this study
show that the Darnley Valley comprises two separate
limbs, which branch around an unnamed rocky pinnacle reef, located in the west-central part of survey
area A (Fig. 2A). The Darnley Valley is as deep as 130
m below sea level south of the pinnacle reef, and shoal
reefs were found in the southernmost section of the
survey area, reaching to within 9 m of the sea surface.
Smaller, low-relief (F 10 m) gullies drain into the
Darnley Valley from shallow platforms located
along the valley margins (Fig. 2A).
187
The seismic survey data collected demonstrates
that the floor of the Darnley Valley is commonly
featureless and acoustically opaque. The two limbs
of the Darnley Valley separate three elevated platforms (P1, P2 and P3, Fig. 2A) each having its own
distinct assemblage of acoustic features (Fig. 3A). The
southernmost platform (P1) is dominated by rocky
(coral) reefs that locally rise up towards the sea surface (Fig. 3A). The largest of these reefs (having the
greatest bathymetric relief) is located adjacent to the
southern walls of the Darnley Valley. Platform P1 also
exhibits localised, infilled channels, erosional notches
and incisions (Fig. 3A). The seabed is acoustically
Fig. 3. Chirper-seismic sections (A–D). The locations of the sections, including start/stop points of section parts, are shown in Fig. 2A and B.
Seismic sections show (A) a profile of sediment onlapping onto the southern side of platform P2 (inset A1), rocky reefs adjacent to the unnamed
reef (Fig. 2A), incised valley floor with sediment wedges (inset A2) and incised surface of Platform P1; (B) a profile starting over the smooth,
sub-horizontally bedded sediment body of Platform P3, exhibiting lateral accretion surfaces and possible slumping on its southern margin (Inset
B1), and an undulating, irregular surface of Platform P2, containing rocky-reef outcrops (Inset B2) and local incisions; (C) high-relief, rockyreef surface adjacent to East Cay, with small sediment mounds in local depressions (Inset C1), incised valley floor with localized sediment
mounds (Inset C2), and reef-studded surface of Platform 4; and (D) low-relief valley floor in ~140 m depth with sediment-filled depressions
(Inset D1), smooth, subhorizontally bedded sediment body (Inset D2), which crop out on its southern margin, and reef-studded surface of
Platform 5.
188
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Fig. 3 (continued).
highly reflective and there is no evidence of unconsolidated sediment deposits associated with platform P1.
The central platform (P2) includes a large (N2
km across), unnamed patch reef on its western
margin (Fig. 2A). The elevated, middle part of platform P2 exhibits widespread erosional notches and
incisions (Fig. 3A and B), with some rocky reefs
along the eastern part of the surveyed area. Subsurface reflectors, in the form of lateral accretion surfaces, occur on both sides of platform P2 and
suggest sediment deposition associated with the lateral migration of former river channels or deltaic
distributary channels. These lateral accretion surfaces are commonly of low relief, b 10 m in overall
amplitude. On the southern flank of platform P2,
foresets dip gently towards the south (Fig. 3, Inset
A1). To the south of the unnamed pinnacle reef
(Fig. 2A), the sediment reaches a maximum thickness of around 20 m above acoustic basement. A
dune field occurs at a depth of between 43 and 70
m on the southern flank of platform P2 (Fig. 2A).
These large-scale, submarine dunes (terminology of
Ashley, 1990) rise about 4–6 m above the level of
surrounding seabed and appear to have their steeper
faces oriented towards the west. They tend to have
sharp crests and no internal stratification was visible
in our seismic records.
The northern platform (P3) exhibits gullies,
notches and incisions, but the most prominent feature of the northern platform is the widespread
occurrence of subsurface reflectors in the Chirp
seismic data. These are most common in the west
and comprise lateral accretion surfaces, gently dipping towards the south on the southern flank of the
platform (Fig. 3A).
3.1.2. Survey Area B (outer shelf)
On the outer shelf, our survey discovered a seaward
extension of the Darnley submarine valley complex,
which attains a maximum depth of 220 m in the area we
surveyed (Fig. 2B). This is the deepest shelf valley yet
found on the NE Australian shelf. The deepest sections
of the Darnley Valley floor form a series of isolated
depressions having closed contours (Fig. 2B). The
valley is flanked on its southern and eastern margins
by rough, rocky-reef seabed that is acoustically opaque,
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
189
Fig. 3 (continued).
apart from isolated sediment lenses up to about 10 m in
thickness (Fig. 3C and D).
On the northern side of East Cay is another shelf
valley, called the bBramble ValleyQ in this paper.
Available bathymetric data suggests that the Bramble
Valley extends westwards to the north of Bramble Cay
and as far landwards as the prograding edge of the Fly
River Delta. In the seaward part of the Bramble Valley
mapped in our study (Fig. 2B) seismic sections reveal
only isolated sediment lenses, b 5 m in thickness (Fig.
3D). The Bramble Valley attains a maximum depth of
around 120 m north of East Cay (Fig. 2B) and seawards of this it shoals to b100 m and becomes indistinct, where it merges with a broad shelf area, mantled
with large-scale, submarine dunes (terminology of
Ashley, 1990). The dunes rise about 4–6 m above
the level of surrounding seabed and appear to have
their steep faces oriented towards the southeast. They
tend to have well-rounded crests and no internal stratification was visible in our chirp seismic records.
The major geomorphic features of the outer shelf
are submerged platform reefs flanking the Darnley
Valley (P4) and extending eastwards from East Cay
(P5; Fig. 2B). The largest platform (P5) is cut by two
channels and exhibits a sharp eastern face, aligned
north–south over a distance of about 15 km. The
surface of both platforms P4 and P5 is characterised
by a strong acoustic reflector, suggesting a hard, rocky
seabed (Fig. 3C and D).
A hypsometric analysis of the bathymetry data for
the two survey areas (Fig. 2A and B) indicates that the
platforms (P1–P5) are consistently located at a depth
of around 45–55 m below sea level (Fig. 4). Peaks in
the hypsometric curves, corresponding with the broad
platform surfaces, can be correlated between the two
survey areas. Other peaks correspond with the relatively flat and broad valley floors at around 85 m
depth in the mid-shelf area (Figs. 2A and 4) and
with broad outer shelf areas where dune fields occur
at a depth of around 95 m (Figs. 2B and 4).
190
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Fig. 3 (continued).
P2, P3
3.2. Sediment sample and video-observation data
8000
Cumultaive surface area (km2)
Mid-shelf area
Outer-shelf area
P1
6000
Broad shelf area
(dunes)
4000
P4, P5
Valley floor
2000
0
0
50
100
150
Water Depth (m)
200
250
Fig. 4. Hypsometric curves for the two mutibeam survey areas shown in
Fig. 2A and B. The depths of submerged platforms P1 to P5 and broad
shelf and valley floor areas are identified as peaks in the two curves.
Regional surface sediment distribution maps were
published by Harris et al. (1996). The information
collected in our study includes underwater video
observations of the seabed, which provides details of
the surface sediments found on the deep valley floors
and some insight into the modern sedimentary processes in this setting.
Darnley Valley floor sediments from the middle
shelf (Area A) are relatively muddy having a mean
mud content of 20.7% (Stations 32–36, 42–45), by
comparison with sediments deposited on the elevated platforms where samples contain an average
of 9.2% mud (Fig. 5A). This higher mud content is
reflected in the underwater video observations
(Table 1), in which the valley floor is generally
observed to be muddy, flat and featureless with
sparse biota (i.e., little if any biota was observed
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
9°
22'
S
% Mud content
191
47
5
>20%
10-20
<10%
60
48
7
46
11
80
23'
45
18
44
20
43
20
60
39
13
80
24'
42
22
60
25'
60
60
36
22
100
34
17
2
37
80
100
33
12
32
26
120
29
3
40
31
10
60
80
60
30
9
28
29
30'
2
38
35
20
80
120
26'
17
41
14
58'
143° 57'
59'
A
144° 00'
9°
22'
S
54
7
01'
02' E
80
100
140
59
34
24'
58
3
55
52
3
57
56
R
80
25'
68
10
120
60
9
60
60
80
6
61
67
7
62
13
26'
66
11
100
120
160
30'
100
74
18
69
14
73
16
60
72
1
120
6
63
65
22
120
100
9
70
71
32'
60
R
160
64
9
9 180
75
60
100
B
4
16
18'
20'
144° 22' E
Fig. 5. Maps showing distribution of surficial sediment mud content (percentage dry weight) for (A) the mid-shelf and (B) outer shelf study
areas. The station numbers shown correspond with information listed in Table 1. See Fig. 1 for the locations of the maps.
192
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Table 1
Presence/absence ( = present) of molluscs, crustacea, polychaetes, fish, sponges, corals, echinoids, soft corals, Halimeda, burrows and ripple
marks observed at underwater video camera stations
Station
Water
depth
Middle shelf
28
55
29
46
30
54
31
52
32
131.5
33
50.5
34
90
35
103
36
85.5
37
50
38
55.5
39
60
40
57.5
41
59.5
42
80.5
43
80
44
87
45
91.5
46
63.5
47
57.5
48
57
Outer shelf
54
119
55
64.5
56
29
57
50
58
104.5
59
76
60
96
61
92
62
103.5
63
90.5
64
117
65
52.5
66
82.5
67
79
68
89
69
109
70
50.5
72
171.5
73
116
74
90
75
220
Molluscs
Crustacea
Polychaetes
Fish
Sponges
Corals
Echinoids
Halimeda
Burrows
Ripple
marks
Soft
coral
For locations of stations, see Fig. 5A (middle shelf stations) and B (outer shelf stations).
at stations in greater than 80 m water depth; Table
1). In contrast, the platforms contain a greater diversity of substrate types, including coral bommies,
limestone rubble and sandy sediments with ripple
marks. Carbonate content did not vary between the
valley floor and platform environments; valley floor
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
which is higher than all other areas sampled in this
study. For example, surface sediments at Station 75
in 220 m water depth comprised gravelly (25%
gravel) calcareous (81% carbonate) coarse olive
sand (5Y 5/3) with abundant granule-pebble shell
hash. The sample included a live sponge, hydroid,
crab, mysid shrimp and polychaete worms (Table 1).
Underwater video at this station showed the seafloor
was relatively flat with rare burrows and pits ~5 cm
in diameter, and sparse biota, apart from some small
rocky outcrops where soft corals, hydroids and
sponges were seen (Table 1).
stations (n = 8) contain 62.1% carbonate mud and
have a 59.6% total carbonate content, which is
similar to platform stations (n = 13) which contain
64.2% carbonate mud and have a 61.2% total carbonate content.
Outer shelf (Area B) sediments are generally
more coarse (only 2 out of 20 stations contained
more than 20% mud) and generally higher in carbonate mud content than the middle shelf sediments
(Fig. 5A and B). Outer shelf sediment samples
taken from the Darnley Valley floor (54, 58, 72–
75; Fig. 5B) have a mean gravel content of 22.1%,
144° 02'
144° 00'
143° 58'
143° 56'
9° 20'
193
N
X-ray of Core 50
47
3 km
48
50 cm
5
46
25
50
4
C1
75
43
100
55
45
44
4
C1
125
Depth (m)
P2
60
4 49
C1
42
41
Un-named
9° 25'
reef
65
51
50
0
52
53
Holocene
Carbonate
1
28
Transgressive
terrigenous
mud
2m
9° 30'
Fig. 6. Map showing the location of core samples taken on the middle shelf, and illustrating the occurrence of the Holocene high-stand facies
and transgressive fluviodeltaic facies. Inset: core X-radiograph (positive) showing laminated sediments comprising the transgressive-aged
sediments located on the middle shelf in core from station 50.
194
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
The other main difference between outer shelf and
middle shelf surface sediments is that the mean carbonate mud content is 82.6 F 3.9% (n = 22) on the
outer shelf, compared with 63.4 F 13.8% (n = 21) on
the middle shelf. This difference probably reflects the
greater distance between the major terrigenous sediment source (i.e., The Fly River) between the outer
shelf area, compared with the middle shelf area.
3.3. Sediment cores and radiocarbon dates
Sediment cores from the middle shelf contained a
0.5–1-m-thick bed of massive, Holocene, carbonate,
poorly sorted, muddy gravelly sand sediment which
overlies a fine-grained, terrigenous, laminated to
finely bedded fluvial–deltaic facies. Laminations
and thin beds (1–5 cm thick) of fine sand occur
within the terrigenous mud and are clearly observed
in X-ray photographs (Fig. 6). Radiocarbon dates
from peat layers extracted from within the finegrained, laminated to bedded, terrigenous sediment
facies (Table 2) are from around 8000 to about
19,000 years old, which indicates they were deposited mostly during the post-glacial sea level transgression of the Gulf of Papua shelf. The corrected
peat dates are consistent with the published sea
Table 2
Radiocarbon ages of peat samples recovered in the Gulf of Papua on R/V Franklin Survey 234 (2002) and FR93 (after Harris et al., 1996)
Radiocarbon
age (years BP)
Error
(Fyears BP)
Calendar age
(years BP)
Water depth
(m)
Core number
Lab reference
number
8684
8969
8902
8697
11,733
10,399
8747
8778
8687
8921
8878
8597
9421
9024
10,216
10,785
11,050
11,117
10,865
10,992
10,898
10,625
13,430
11,070
7290
16,750
9280
9280
8600
60
50
60
55
60
60
50
50
55
50
55
60
50
60
55
55
60
55
65
55
55
70
530
330
350
340
90
100
29
9711–9547
10,204–10,143
9711–9547
10,204–10,143
13,870–13,487
12,385–12,091
9827–9624
9909–9692
9710–9549
10,080–9919
10,153–9899
9556–9528
10,703–10,566
10,226–10,174
12,129–11,719
12,940–12,819
13,135–12,925
13,163–13,008
12,983–12,856
13,105–12,899
13,027–12,872
12,888–12,628
16,113–14,970
12,940–11,953
8099–7434
19,864–18,893
10,150–9811
10,148–9814
9079–8948
35
35
35
35
81
80
30
30
30
30
30
30
30
30
59
59
59
59
59
59
59
59
77
47
64
93
40
40
29
234/22PC06
234/22 PC06
234/22PC06
234/22PC06
234/43GC04
234/49GC10
234/21PC05
234/21PC05
234/21PC05
234/21PC05
234.21PC05
234/21PC05
234/21PC05
234/21PC05
234/41GC02
234/41GC02
234/41GC02
234/41GC02
234/41GC02
234/41GC02
234/41GC02
234/41GC02
FR93/15GC4
FR93/24PC5
FR93/38PC12
FR93/54PC20
FR93/13VC8
FR93/13VC8
FR93/20VC15
NZA16780
NZA16781
NZA16782
NZA16783
NZA16784
NZA16785
NZA16832
NZA16833
NZA16834
NZA16835
NZA16836
NZA16837
NZA16838
NZA16839
NZA16840
NZA16841
NZA16842
NZA16843
NZA16844
NZA16845
NZA16846
NZA16847
SUA 3070
SUA 3072
SUA 3073
SUA 3074
SUA 2837
SUA 2838
SUA 3075
Laboratory reference numbers refer to the University of Sydney (SUA) and Rafter Radiocarbon Laboratory (NZA), New Zealand. Radiocarbon
ages are uncorrected in years before present (BP). The uncorrected radiocarbon ages were converted to upper and lower estimates of calendar
ages using the INTCAL-98 calibration curve (Stuiver et al., 1998; see Fig. 7). The locations of cores 234/21PC05 234/22PC06 are shown in Fig.
1C (labelled with underlined station numbers only, i.e., 21 and 22, respectively) and locations of cores 234/41GC02, 234/43GC04 and 234/
49GC10 are shown in Fig. 6 (labelled with underlined station numbers only, i.e., 41, 43 and 49, respectively).
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
195
20
0
-40
Depth of highest tidal energy
Time of highest
tidal energy
-60
-80
Sea level (m)
-20
-100
-120
Peat (error bars), this paper
Shell data from Larcombe et al. (1995)
-140
14
12
10
8
6
4
2
0
Corrected calendar age kyr (BP)
Fig. 7. Plot of radiocarbon ages determined from mangrove peat samples in relation to water depth for samples taken in the Gulf of Papua in the
present study plus those reported by Harris et al. (1996). Dates for shell samples are from Larcombe et al. (1995). The sea level curve shown is
from Edwards et al. (1993). The uncorrected radiocarbon ages listed in Table 2 were converted to upper and lower estimates of calendar ages
using the INTCAL-98 calibration curve (Stuiver et al., 1998), prior to plotting on this graph. The timing of meltwater pulses is from Bard et al.
(1996). The hypsometric curve for the Gulf of Papua illustrates the mean depths of the modern coastal plain (CP), Holocene deltas (D) and delta
front (DF) systems and plateaus (PL; probably Pleistocene age; Harris et al., 1996) located on the mid-shelf. During the post-glacial
transgression, shoreface retreat would have been rapid over the broad, low-relief plateaus, limiting the preservation potential for coastal
mangrove peat deposits.
level curves of other workers (e.g., Edwards et al.,
1993; Bard et al., 1996; Fig. 7).
4. Discussion
4.1. Origins of shelf valleys
The Darnley and Bramble Valleys mapped in this
study appear to have been formed by at least two
different processes operating over glacial–interglacial
timescales. The Bramble Valley is less than 120 m in
depth over its entire length. Lateral accretion surfaces
are observed in seismic sections along its walls, which
are documented in sediment cores as transgressiveaged, terrigenous mud deposits (Harris et al., 1996).
These transgressive-aged deposits are similar in composition and primary sedimentary structures to modern deltaic deposits that have been described from the
nearby Fly River delta (Dalrymple et al., 2003), and
we infer a similar (deltaic) depositional environment
for this facies. Mangrove peat beds are commonly
associated with deltaic islands and intertidal mud
flats in the modern Fly Delta (e.g., Baker et al.,
1995; Baker, 1999; Dalrymple et al., 2003). We conclude that the Bramble Valley was probably one of the
main channels (if not the main channel) for the Fly
River during the last glacial maximum, and that it was
formed by fluvial erosion during Quaternary glacial
maxima and modified by fluvial–deltaic sedimentation during the post-glacial transgression, between
18,000 and 6500 years BP.
The Darnley Valley, located further to the south,
has a more complex morphology and was probably
formed by several different processes operating over
different temporal and spatial scales. Lateral accretion
surfaces are observed in seismic sections which are
also documented in sediment cores as transgressiveaged, terrigenous mud deposits. Hence, the Darnley
Valley was influenced by fluvial–deltaic depositional
processes during the post-glacial sea level rise. This is
not surprising, given that the modern Gulf of Papua
coast is comprised entirely of mangrove-forested, deltaic islands that migrate laterally (Harris et al., 1993,
2004; Walsh and Nittrouer, 2003). However, lateral
channel accretion deposits are found along the mar-
196
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
gins only of the northern limb of the Darnley Valley
and in patches along the northern wall of the southern
limb. None were found along its southern margin on
the middle shelf (Fig. 2A) or at any location on the
outer shelf (Fig. 2B).
The Darnley Valley is over-deepened locally, to
depths of over 200 m. Erosion of the valley floor to
this depth cannot be explained by fluvial erosion
during late Quaternary, glacial, low sea level periods
even taking into account an inferred subsidence rate of
about 10 cm per 1000 years (Pigram et al., 1989),
because sea level did not drop more than 120–130 m
below its present position. Sedimentary strata underlying the Darnley Valley are thought to be comprised
of Cenozoic, interbedded carbonates and fluvio-deltaic deposits (Pigram et al., 1989) and there is no
seismic profiling evidence for a major basement
fault to explain the position of the valley.
Isolated depressions 150–220 m deep with closed
bathymetric contours at both ends, on the floor of the
Darnley Valley demand an explanation other than
fluvial erosion during glacial maxima. Given their
depth and the closed contours of their morphology,
it is possible that depressions along the length of the
Darnley Valley were the sites of lakes during the last
ice age when Torres Strait formed a land-bridge
between Australia and Papua New Guinea. Given
that the valleys are incised into limestone, karst processes may have played a role in forming the depressions. However, there is no indication of former lake
deposits in our cores from the Darnley Valley floor,
nor were any thick, unconsolidated sediment deposits
observed anywhere on the floor of the Darnley Valley
in our seismic records. Karst processes may have
played a role in creating localised depressions in the
study area, perhaps as much as 10–20 m in relief.
However, it seems unlikely that the elongate orientation of the depressions (that are aligned with the axis
of the valley) or the shear vertical scale of relief
observed for the Darnley Valley could be explained
by Karst processes alone. We conclude that processes
other than (or in addition to) Karst must have formed
the deepest sections of the Darnley Valley.
To account for the absence of lacustrine sediment
deposits within isolated depressions of the Darnley
Valley, we infer that either (1) there were no lakes here
during the ice ages, and hence, no lake deposits exist;
(2) there are lake deposits but they are too thin to be
resolved by our Chirp sub-bottom profiler; or (3) there
were lake deposits but the small volumes of sediment
deposited were exhumed by tidal currents during the
post-glacial sea level rise (and flooding of the shelf).
In the latter case, we might expect to find strong tidal
currents flowing today within the scoured sections of
the valleys. The available current and oceanographic
data, however, indicate that this is not so.
Recent tidal and wind-driven current modelling
work carried out for the region by Hemer et al.
(2004) indicates that tidal and wind-driven currents
over the mid- to outer shelf valleys are relatively
weak. Strong tidal currents do occur in central Torres
Strait and in the 60 m deep valley (Missionary Passage; Fig. 1) located north of the Warrior Reefs on the
inner shelf (Harris, 2001), but tidal flows are generally
weaker in the eastern patch reef complex. Strong tidal
flows must occur locally to explain the sharp-crested,
active dunes on Platform P2 (Fig. 2A); however, the
large area of dunes in deep water in Area B having
rounded crests (Fig. 2B) that we assume to be moribund coincides with an area where modeling predicts
weaker currents. Low current energy is also consistent
with the accumulation of muddy sediments on the
Darnley Valley floor on the middle shelf (Fig. 5A).
Oceanographic observations indicate that the Darnley
Valley provides a conduit onto the shelf for cool and
saline, upwelled Coral Sea water (Harris et al., 2004).
Temperature and salinity depth profiles taken within
the valley during the waning spring tide phase, suggest that the water column is stratified above 50 m
depth, although any density-driven circulation is
likely to be sluggish (speeds of b 0.1 m s 1). It
appears, therefore, that the modern current regime
could not have created the deep valleys observed on
the mid- to outer shelf. The question is, therefore, how
(and when) were the valleys formed?
4.2. Tidal current modelling of low sea level scenarios
To answer this question, we have carried out a
simple tidal modelling exercise. A brief description
of the tidal model follows; details are documented
elsewhere (Harris et al., 2000; Porter-Smith et al.,
2004). The tidal model has a resolution of 0.0678 in
both latitude and longitude, equal to about 7.4 km.
The bathymetry input was not varied to account for
sediment deposition, reef growth or erosion, which are
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
considered small within the area of interest (incised
valley zone) in comparison with the model resolution.
The model uses major eight constituents M2, S2, N2,
K2, K1, O1, Q1, P1 and was run for eight different sea
levels at 10-m increments below the present sea level.
The solution was obtained using time stepping on an
Arakawa C grid by applying periodic forcing and time
stepping from homogenous initial conditions. The
solution was achieved in 10,000 time steps using a
step length of 15 s. This corresponds to running the
model for roughly 2 days. Along the open boundaries
around the edges of the model a global ocean tide
model by Andersen et al. (1995; version AG95.1) was
used to provide boundary elevations. The model forcing was not varied for the different sea level increments, which is a potential source of error, but this
must be accepted since we do not have any data to
avoid making the assumption that the tidal forcing
parameters do not vary with changing sea level.
The results (Fig. 8) for the modern sea level are
consistent with available observational data; bottom
stress due to maximum spring tidal currents is generally low apart from a few localized areas around the
central Darnley Valley (where active dunes also occur;
see above), near Missionary Passage and among the
distributary channels of the Fly Delta. With sea level
at 10 m, the model suggests that bottom stress
maxima occur over the eastern Missionary Passage
area and in small patches around Darnley Island and
other locations north of the island. With sea level at
20 m, Torres Strait is exposed and the Australian
mainland is joined to Papua New Guinea; the model
suggests that bottom stress maxima increase in overall
intensity and occur in the Great North East Channel,
around the patch reefs adjacent to Darnley Island and
around patch reefs on the outer shelf flanking the
Darnley Valley. With sea level at 30 m, the model
suggests that intense bottom stress maxima occur;
tides are apparently amplified within a broad embayment extending eastwards from Missionary Passage.
Bottom stress bhot spotsQ are also centred over the
Darnley Valley and part of the Bramble Valley. The
pattern is virtually the same for sea level at 40 m,
albeit shifted slightly eastwards. With sea level at 50
m, the Missionary Valley is mostly exposed and the
model suggests that three bottom stress bhot spotsQ
occur, centred over the Darnley Valley the Bramble
Valley and an area north of the Bramble Valley. As sea
197
level is lowered further too 60 and 70 m, the
bottom stress bhot spotsQ become smaller in size as
more shelf area becomes emergent. Bottom stress bhot
spotsQ are predicted to occur in different places along
the Darnley Valley at all depths between 30 and 60 m
(see Fig. 8).
In summary, the model (Fig. 8) predicts that maximum tidal bottom current stress reached their highest
values over the site of the deepest shelf valleys when
sea level was between 30 and 50 m below its present
position. At these depths, the geomorphology of the
shelf manifests itself as a chain of elongate islands (the
modern barrier reef) separating the open sea from a
shallow, protected lagoon. Currents are accelerated as
they are forced to flow through the constricted channels
between the shelf-edge islands, and in particular as they
flow around the northern promontory of the island
chain. The strongest tidal flows are predicted to occur
at the northern end of the shelf edge barrier (Fig. 8), and
although the 7.4-km resolution of the model is too
coarse to resolve the details of flow over the complex
bathymetry in the area, this general area coincides with
the location of the deepest (N 200 m deep) segment of
the Darnley Valley south of East Cay.
The flat-top surfaces of submerged reef platforms
adjacent to the valleys (Fig. 2A and B) also occur in
this 40–50-m depth range, suggesting that these relict
reefs (bgive-up reefsQ using the terminology of Neumann and Macintyre, 1985) developed at times when
sea level was at about 40 m below its present position.
These submerged reefs occur at similar depths to
those reported from other locations in the GBR (Harris and Davies, 1989). When sea level drops below
~40 m, these reefs are emergent and hydraulic tidal
flow is further enhanced.
The overall shape of the Darnley Valley itself
points to processes unrelated to erosion by the Fly
River. It has tributary branches extending southwards
into the northern Great Barrier Reef patch reef province, and its western end turns southwestwards, into
the Great North East Channel (and not northwest
towards the Fly River catchment). These features are
more easily explained as being created by tidal water
exchange between the Coral Sea and the protected
lagoon that is formed when sea level drops below
around 30 m (Figs. 1B and 8).
When considering the integrated effect of the
changes in tidal flow associated with different sea
198
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Fig. 8. Plot of peak (spring) tidal bed stress (N m 2) derived from a tidal model, with outputs derived for sea levels at 0, 10, 20, 30, 40, 50, 60
and 70 m below present position. The locations of modern land (in black) reefs (dark gray), and the 50 m isobath outlines of the Missionary,
Darnley and Bramble Valleys are shown. Locations of Missionary passage (MP), the Great North East Channel (GNEC), Darnley Island (DI),
Bramble Cay (BC) and East Cay (EC) are also shown.
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
level scenarios, it is important to keep in mind that
the duration of sea level at each point has not been
equal over the last glacial cycle (past 120,000 years;
Fig. 9A and B). The eustatic sea level curve (e.g.,
Chappell and Shackleton, 1986; Fig. 9A) indicates
that over the past 120,000 years, sea level has been
between 30 and 50 m below its present position for
about 38% of the time. Prolonged periods (N 10,000
years duration) of sea level within this depth range
occurred during isotope stages 3, 4, 5a and 5c (Fig.
9A). Isotope stage 3, in particular, was a time span
of around 22,000 years during which sea level was
positioned at between 30 and 50 m below present.
It is not just the depth where maximum tidal bed
stress occurs that is important, but also the duration
A
0m
1
2
Oxygen isotope stage
3
4 5a b c d e
6
Depth
Below 50
Present
Sea-level 100
150
0
B
25
50
100
Kyr before present
150
20-60 m
Percent of time
(last 120 kyr)
30-50 m
20
15
10
199
of sea level within the 30–50-m depth range (equal
to 38% of the last 120 ky; Fig. 9B) that must be
considered in considering the development of the
shelf’s geomorphology. It is logical that, all things
being equal, current erosion features would form at
depths corresponding with the long-lived and strong
tidal flow conditions related to sea level at 30–50 m
below present. Although it is difficult to accurately
determine the age of any geomorphic feature
formed by erosion, we conclude that the over-deepened Darnley Valley is a relict feature, whose
origin is related mainly to tidal current scour during
the late Pleistocene when sea level was ~30–50 m
below its present position.
An important consequence of these valley-erosion
processes is to halt the southward advance of terrigenous sediment, supplied to the Gulf of Papua
from the island of New Guinea, into the Great
Barrier Reef province. During interglacial, high sea
level stands, such as at present, fluvial deltaic
deposition is focused along the coast with the prograding delta extending offshore into waters less
than around 40 m deep (Harris et al., 1993). The
Darnley Valley tends to trap only small amounts of
fine-grained terrigenous sediment (Fig. 5A and B).
During intermediate sea level positions (isotope
stages 3, 4, 5a and 5c; Fig. 9) strong tidal flows
scour the valley clear of unconsolidated sediment.
The integrated result, over several glacial/interglacial
sea level cycles, is that the valley system is maintained as it is seen today.
4.3. Valley formation on tropical carbonate shelves by
tidal scour and reef growth
5
0
10 20 30 40 50 60 70 80 90 100 110 120 130 140
Depth below present sea level (m)
Fig. 9. (A) Global eustatic sea level curve for the last 150,000 years,
modified from Chappell and Shackleton (1986) to illustrate times of
sea level positioned at between 30 and 50 m below present (greyshaded areas). Isotope stages are after Martinson et al. (1987). (B)
Histogram showing percentage of time that sea level has been
within 10-m depth bands (i.e., 0–10 m, 10–20 m, etc.) over that
past 120,000 years (isotope stages 1 to 5d, inclusive), based on the
curve shown in bAQ. The graphs show that sea level was within the
30–50-m depth range for approximately 38% of the time (46,400
years) over the past 120,000 years. For comparison, sea level has
been within the 20–60-m depth range for approximately 60% of the
time (74,500 years) and in the 0–10-m range for only 12.8% of the
time (15,500 years).
The formation of shelf valleys by tidal current
scour and sea level change in the northern Great
Barrier Reef implies a conceptual model for understanding the evolution of tidally incised shelf valley
systems (Fig. 10), which can be contrasted with the
conventional model of fluvially incised valley formation. For both types, the changes caused by a rise
or fall of relative sea level are of fundamental importance. However, whereas a lowering of sea level
results in fluvial erosion, entrenchment and the formation of fluvially incised valleys (a sequence
boundary; e.g., Vail et al., 1977; Van Waggoner et
al., 1990), the formation of a tidally incised valley
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Calcareous marine
} sediments
Erosional unconformity
} Lacustrine sediments
Erosional unconformity
} Incised older shelf deposits
C. Coastal
progradation
shelf deposition
reef growth
Va
ll
ey
s
Reefs
A. Tidal erosion
reef growth
sediment export
1
2
Oxygen isotope stage
3
4 5a b c d e
6
0m
50
100km
F
me ly Rive
and
er z r
on
100
e
B. Fluvial erosion
shelf-perched
lakes
150
0
50
100
Kyr before present
Depth Below Present Sea-level
200
150
Reefs
100km
Fig. 10. Conceptual model showing the evolution of tidally incised shelf valleys and inferred facies succession. (A) bIntermediateQ sea level
conditions dominate the late Quaternary and give rise to valleys that are tidally incised into broad platform reefs on rimmed, tropical carbonate
shelves. The occurrence of a brimmedQ outer shelf establishes a tidally varying hydraulic pressure gradient between the shelf-lagoon and the
adjacent ocean basin, which reaches a maximum tidal energy under certain sea level conditions. Tidally incised channels may be further
increased in relief by framework coral growth along the valley margins, akin to levy banks, through the late Quaternary to produce steep-sided
(locally vertical) valley walls. Tidal erosion liberates shelf sediment for export to the adjacent slope and forms a regional erosional
unconformity. (B) During glacial, low sea level conditions, the shelf valleys may be transformed into lakes perched on the outer shelf, trapping
sediments and limiting sediment export to the adjacent slope. Post-glacial sea level transgression of the shelf results in erosion and localized
exhumation of the valleys. (C) High-stand valleys (e.g., Fig. 1) contain weak currents and may be locally depositional, containing muddy
calcareous marine sediments. Elsewhere, deposits of relict carbonate gravels dominate valley floor deposits.
relies on the occurrence of a brimmedQ outer shelf to
establish a hydraulic pressure gradient between the
shelf-lagoon and the adjacent ocean basin (Fig. 10).
A key factor is the relative tidal range, which must
be large enough to generate the strong tidal flows
necessary to erode the seabed and form the shelf
valleys. The tidal flows attain a maximum value
when the largest hydraulic pressure gradient is forced
through the smallest channel cross section, or as in
the case of the Darnley Valley, around the end of a
promontory (Fig. 10).
The generalised stratigraphic succession produced
by the transgression of tidally incised shelf valleys
will also differ from the conventional, fluvially
incised valley-fill model. The facies succession of
an infilled estuary, in relation to a rise and stabilisation
of relative sea level, consists of an erosional unconformity overlain by alluvial channel deposits (lowstand systems tract, LST) followed by estuarine/marine units (transgressive systems tract, TST), which
vary in character depending upon sediment supply
and the relative influence of tides and waves (Dalrymple et al., 1994). High-stand systems tracts (HST)
marine sediments drape over the TST deposits, completing the succession.
The facies succession within tidally incised shelf
valleys mapped in the present study includes an
erosional unconformity that is not necessarily over-
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
lain by any unconsolidated deposits. LST deposits
may contain lacustrine facies (NB the valleys in our
study were scoured clear of any such deposits). The
sequence boundary may be overlain by thin beds of
coarse-grained, transgressive marine sediments,
grading upward into lower energy, muddy gravelly
sands (HST marine sediments; Fig. 10). A likely
scenario is that, as infilling of the lagoon leads to
reduced tidal flows in the valleys (due to the diminishing size of the tidal prism), the preservation
potential of LST lacustrine and TST marine deposits
is enhanced.
A common feature of the channels in the northern Great Barrier Reef is the growth of coral reefs,
like river levy banks, on the margins of channels.
Current control over reef growth and reef morphology is well established in the literature. For example, the modern barrier reef in the study area was
described by Veron (1978) who noted that some
shelf-edge, barrier reefs appear to have a bdeltaicQ
shape in plan view (this term was first used to
describe a specific reef type by Maxwell, 1968).
Strong tidal flows up to 3.8 m/s characterize the
narrow, inter-reef channels, which are about 33 m in
water depth. Veron (1978) explained the deltaic
morphology as a reef-growth response to the strong
tidal flows characteristic of the Torres Strait region.
The elongate reefs in central Torres Strait have the
appearance of tidal current sand banks in plan view,
as first noted by Off (1963) and later studied by
Jones (1995) using seismic profiling.
Reefs growing preferentially along the margins
of tidally scoured channels act as a positive feedback, by further constricting the channel width and
accelerating the tidal flow. The vertical scour and
over-deepening of channels by tidal currents reaches
a maximum in the most narrow part of the channel,
at the point of greatest flow constriction, and tidal
energy drops away along the channel axis in both a
landward and seaward direction, which explains the
closed bathymetric contours of tidally scoured channels. Missionary Passage at the northern end of the
Warrior Reefs (Fig. 1) is a modern example of a
tidally scoured channel that exhibits these morphological features. Tectonic subsidence of the Gulf of
Papua foreland basin at rates of around 10 cm per
1000 years (Pigram et al., 1989) may also be a
factor for the evolution of channels.
201
Our model of tidally incised valley formation
may be applied to other rimmed shelves in Australia
and elsewhere. In particular, the modern valleys
adjacent to the Sahul Shoals in the Timor Sea
(e.g., the Malita Shelf valley) have a similar geomorphology, exhibiting closed isobaths and a significant basin (Bonaparte Depression) located
landward of the valley complex (Lavering, 1993;
Fig. 11). These incised channels have closed bathymetric contours, contain silty clays and molluscan
debris (Lavering, 1993) and currents measured in
the channels are mainly weak and variable (Marshall
et al., 1994); it may be concluded that they were not
created by modern processes. Further tidal current
modelling, under different sea level scenarios, is
needed to determine the depth range at which tidal
currents are at a maximum, when the emergent
Sahul banks produced accelerated tidal flows
through interbank, incised channels. Other arid, tropical carbonate shelves may exhibit similar, outershelf incised valley systems, both in the modern
world and in the rock record.
4.4. GBR transgressive sediment export to slope
explained?
In a study of sediment mass accumulation rates on
the slope adjacent to the Great Barrier Reef, Page et
Fig. 11. False colour bathymetyric map of the Timor Sea region of
northern Australia showing shelf valleys (tidally?) incised into an
800-km-long, shelf-edge, raised carbonate rim.
202
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
al. (2003) have noted that peaks in sediment accumulation rate do not conform to the prediction of the
conventional model, which calls for maximum sediment export during sea level low-stands. Over the last
glacial cycle, the conventional model predicts that
maximum export to the slope would have occurred
around 18 ky BP, but instead Page et al. (2003)
measured peak export to occur from between 12 and
7 ky BP, during the rising phase of sea level. Page et
al. (2003) suggest that the reason for this low sediment discharge during the last glacial maximum was
because sediment is deposited on the low-gradient
shelf behind the rimmed profile. Flooding of the
shelf during sea level rise liberates the sediment stored
behind this rim resulting in an increased sediment
supply to the adjacent slope.
Our model for tidal incision of rimmed shelves
provides a mechanism to explain how the sediment
is mobilized during sea level rise. Our model predicts a phase of strong, tidally induced, bed stress
on the outer shelf when sea level is between 30 and
50 m below its present position. Available sea level
curves indicate this depth range corresponds to a
time of from between 11 and 9 ky BP (Fig. 7),
which is within the time range of 12–7 ky BP
specified by Page et al. (2003) for maximum sediment export from shelf to slope. The ebb-orientation
of the large field of moribund dunes on the outer
Gulf of Papua shelf (Fig. 2B) also conforms with
the concept of sediment export from the shelf
caused by a pre-Holocene phase of enhanced tidal
current activity.
5. Conclusions
Shelf valleys at the northern end of the Great
Barrier Reef may be divided into types having two
different origins. The Bramble Valley in the north is
clearly a relict fluvial valley, exhibiting lateral accretion surfaces observed in seismic profiles and
incised margins that intersect and truncate underlying strata. The over-deepened Darnley Valley,
located among reefs in the northernmost GBR, however, appears to have formed by a combination of
reef growth and tidal current scour. During sea level
transgression of the shelf, the Bramble Valley was
an estuary where fluvial deltaic and marine sedi-
ments were deposited, as documented by cores and
seismic data. The Darnley Valley, however, would
have formed an estuarine embayment, whose origin
is not reliant upon fluvial erosion processes, and
where cores and seismic data indicate minimal terrigenous deposition occurred.
The tidally incised Darnley Valley is up to 220 m
deep locally and extends for 93 km across the continental shelf. Depressions on the valley floor exhibit
closed bathymetric contours at both ends and are
floored with well-sorted carbonate gravely sand containing a large relict fraction and up to 20% mud,
locally. The deepest segments of the Darnley Valley
were probably the sites of lakes during the last ice
age, when Torres Strait formed an emergent landbridge between Australia and Papua New Guinea.
Current modeling and observations suggest that modern tidal flows are weak and insufficient to have
eroded the valley. Numerical modelling further predicts that the strongest tidal currents occur over the
deepest, outer-shelf segment of the valleys when sea
level is about 30–50 m below its present position.
This depth range corresponds to a brief period of
around 9–7 ky BP during the post-glacial sea level
rise, but more importantly also to several prolonged
phases of sea level during oxygen isotope stages 3, 4,
5a and 5c (Fig. 9). These results are consistent with a
Pleistocene age and relict origin of the over-deepened
portions of the valley floor. Tidal erosion and mobilisation of sediments could explain the peak in export
of shelf sediments to the adjacent slope during the
post-glacial sea level rise as reported by Page et al.
(2003).
We propose a conceptual model to explain the
evolution of shelf valleys that are tidally incised into
broad platform reefs on rimmed, tropical carbonate
shelves. The occurrence of a rim on the outer shelf
establishes a tidally varying hydraulic pressure gradient between the shelf-lagoon and the adjacent ocean
basin, which reaches a maximum under certain sea
level conditions (Fig. 10). Tidally incised channels
may be further increased in relief by framework
coral growth along the valley margins, akin to levy
banks, through the late Quaternary to produce steepsided (locally vertical) valley walls. This model may
explain the evolution of valleys incised into the Sahul
shelf of northern Australia as well as other rimmed
carbonate shelves.
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Acknowledgements
Ship time aboard R/V Franklin was granted to
PTH by the Australian National Facility Steering
Committee. Additional bathymetric images of the
reefs and seabed photographs can be viewed online at this address:www.ga.gov.au/oceans/projects/
20010917_12.jsp. The paper was improved by the
critical reviews of Drs. Jim Colwell, Phil O’Brien,
John Marshall, Gavin Dunbar and an anonymous
reviewer. Published with permission of the Executive
Director, Geoscience Australia.
References
Andersen, O.B., Woodworth, P.L., Flather, R.A., 1995. Intercomparison of recent global ocean tide models. J. Geophys. Res. 100
(C12), 25261 – 25282.
Ashley, G.M., 1990. Classification of large-scale subaqueous bedforms: a new look at an old problem. J. Sediment. Petrol. 60,
160 – 172.
Baker, E.K., 1999. Sedimentation in a tidally dominated delta and
the impact of upstream mining: Fly River, Papua New Guinea.
PhD thesis, Sydney.
Baker, E.K., Harris, P.T., Short, S.A., Keene, J.B., 1995. Patterns of
sedimentation in the Fly River Delta. In: Flemming, B., Bartholoma, W. (Eds.), Tidal Signatures in Modern and Ancient Environments, International Association of Sedimentologists Special
Publication, vol. 24, pp. 193 – 211.
Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G.,
Faure, G., Rougerie, F., 1996. Deglacial sea-level record from
Tahiti corals and the timing of global meltwater discharge.
Nature 382, 241 – 244.
Chappell, J., Shackleton, N.J., 1986. Oxygen isotopes and sea level.
Nature 324, 137 – 140.
Chough, S.K., Lee, H.J., Yoon, S.H., 2000. Marine Geology of
Korean Seas. Elsevier, Amsterdam. 313 pp.
Chough, S.K., Kim, J.W., Lee, S.H., Shinn, Y.J., Jin, J.H., Suh, M.C.,
Lee, J.S., 2002. High-resolution acoustic characteristics of epicontinental sea deposits, central-eastern Yellow Sea. Mar. Geol.
188, 317 – 331.
Dalrymple, R.W., Boyd, R., Zaitlin, B.A., 1994. History of research,
types and internal organisation of incised valley systems: introduction to the volume. In: Dalrymple, R.W., Boyd, R., Zaitlin,
B.A. (Eds.), Incised Valley Systems: Origin and Sedimentary
Sequences, SEPM Special Publication, vol. 51. Society for
Sedimentary Geology, Tulsa, OK, pp. 3 – 10.
Dalrymple, R.W., Baker, E.K., Harris, P.T., Hughes, M.G., 2003.
Sedimentology and stratigraphy of a tide-dominated, forelandbasin delta (Fly River, Papua New Guinea). In: Sidi, F.H.,
Posamentier, H.W., Darman, H., Nummedal, D., Imbert, P.
(Eds.), Tropical Deltas of Southeast Asia and VicinitySedimentology, Stratigraphy, and Petroleum Geology, SEPM Special
203
Publication, vol. 76. Society for Sedimentary Geology, Tulsa,
OK, pp. 147 – 173.
Edwards, R.L., Beck, J.W., Burr, G.S., Donahue, D.J., Chappell,
J.M.A, Bloom, A.L., Druffel, E.R.M., Taylor, F.W., 1993. A
large drop in atmospheric 14C/12C and reduced melting in the
Younger Dryas, documented with 230Th ages of corals. Science
260, 962 – 968.
Harris, P.T., 1994a. Incised valleys and backstepping deltaic
deposits in a foreland-basin setting, Torres Strait and Gulf
of Papua, Australia. In: Dalrymple, R.W., Boyd, R., Zaitlin, B.
(Eds.), Incised Valley Systems: Origin and Sedimentary
Sequences, SEPM Special Publication, vol. 51. SEPM, Tulsa,
OK, pp. 97 – 108.
Harris, P.T., 1994b. Comparison of tropical, carbonate and temperate, siliciclastic tidally dominated sedimentary deposits: examples from the Australian continental shelf. Aust. J. Earth Sci. 41,
241 – 254.
Harris, P.T., 2001. Environmental management of Torres Strait: a
marine geologist’s perspective. In: Gostin, V.A. (Ed.), Gondwana to Greenhouse: Environmental Geoscience—An Australian Perspective, Adelaide. Geological Society of Australia
Special Publication, vol. 21, pp. 317 – 328.
Harris, P.T., Davies, P.J., 1989. Submerged reefs and terraces on the
shelf edge of the Great Barrier Reef, Australia: morphology,
occurrence and implications for reef evolution. Coral Reefs 8,
87 – 98.
Harris, P.T., Baker, E.K., Cole, A.R., Short, S.A., 1993. A preliminary study of sedimentation in the tidally dominated Fly
River Delta, Gulf of Papua. Cont. Shelf Res. 13 (4), 441 – 472.
Harris, P.T., Pattiaratchi, C.B., Collins, M.B., Dalrymple, R.W., 1995.
What is a bedload parting? In: Flemming, B.W., Bartholoma, A.
(Eds.), Tidal Signatures in Modern and Ancient Sediments , IAS
Special Publication, vol. 24. Blackwell, Oxford, pp. 1 – 18.
Harris, P.T., Pattiaratchi, C.B., Keene, J.B., Dalrymple, R.W., Gardner, J.V., Baker, E.K., Cole, A.R., Mitchell, D., Gibbs, P.,
Schroeder, W.W., 1996. Late quaternary deltaic and carbonate
sedimentation in the Gulf of Papua foreland basin: response to
sea-level change. J. Sediment. Res. 66 (4), 801 – 819.
Harris, P.T., Smith, R., Anderson, O., Coleman, R., Greenslade, D.,
2000. GEOMAT—Modelling of Continental Shelf Sediment
Mobility in Support of Australia’s Regional Marine Planning
Process. 2000/41. Australian Geological Survey Organisation.
Harris, P.T., Heap, A., Passlow, V., Hughes, M., Porter-Smith, R.,
Beaman, R.J., Hemer, M., Daniell, J., Buchanan, C., Watson, T.,
Collins, D., Bleakley, N., Anderson, O., 2002. Cross-shelf
sediment transport in the Torres Strait—Gulf of Papua Region.
Geoscience Australia Survey 234, Post-Cruise Report: RV
Franklin Cruise 01/-02, January–February, 2002. 2002/26,
Geoscience Australia, Canberra.
Harris, P.T., Hughes, M.G., Baker, E.K., Dalrymple, R.W., Keene,
J.B., 2004. Sediment transport in distributary channels and its
export to the pro-deltaic environment in a tidally-dominated
delta: Fly River, Papua New Guinea. Cont. Shelf Res. 24
(19), 2431 – 2454.
Hemer, M.A., Harris, P.T., Coleman, R., Hunter, J., 2004. Sediment
mobility due to currents and waves in the Torres Strait–Gulf of
Papua region. Cont. Shelf Res. 24 (19), 2297 – 2316.
204
P.T. Harris et al. / Marine Geology 220 (2005) 181–204
Jansen, E., Befring, S., Bugge, T., Eidvin, T., Sejrup, H., 1987.
Large submarine slides in the Norwegian continental margin:
sediments, transport and timing. Mar. Geol. 78, 77 – 107.
Jones, M.R., 1995. The Torres Reefs, North Queensland, Australia—strong tidal flows a modern control on their growth. Coral
Reefs 14, 63 – 69.
Kuehl, S.A., Levy, B.M., Moore, W.S., Allison, M.A., 1997. Subaqueous delta of the Ganges–Brahmaputra river system. Mar.
Geol. 144, 81 – 96.
Larcombe, P., Carter, R.M., Dye, J., Gagan, M.K., Johnson, D.P.,
1995. New evidence for episodic post-glacial sea level rise,
central Great Barrier Reef, Australia. Mar. Geol. 127, 1 – 44.
Lavering, I.H., 1993. Quaternary and modern environments of the
Van Dieman Rise, Timor Sea and potential effects of additional
petroleum exploration activity. BMR J. Aust. Geol. Geophys.
13, 281 – 292.
Marshall, J., Davies, P.J., Mihut, I., Troedson, A., Bergerson, D.,
Haddad, D., 1994. Sahul Shoals neotectonics and Cainozoic
Environments: Rig Seismic Cruise 122 Post Cruise Report.
1994/33. Australian Geological Survey Organisation, Canberra,
Australia.
Martinson, D.G., Pisias, N.G., Hays, J.D., Imbrie, J., Moore Jr.,
T.C., Shackleton, N.J., 1987. Age dating and the orbital theory
of ice ages: development of a high-resolution 0 to 300,000 years
chronostratigraphy. Quat. Res. 27, 1 – 29.
Maxwell, W.G.H., 1968. Atlas of the Great Barrier Reef. Elsevier,
Amsterdam.
Muller, G., Gastner, M., 1971. The bkarbonate bombeQ a simple
device for the determination of the carbonate content in sediments, soils and other materials. Neus Fahrb. Mineral., Monatsh. 10, 466 – 469.
Neumann, A.C., Macintyre, I.G., 1985. Reef response to sea
level rise: keep-up, catch-up or give-up. 5th International
Coral Reef Congress. International Coral Reef Society, Tahiti,
pp. 105 – 110.
Off, T., 1963. Rythmic linear sand bodies caused by tidal currents.
Bull. Am. Assoc. Pet. Geol. 47 (2), 324 – 341.
Page, M.C., Dickens, G.R., Dunbar, G.B., 2003. Tropical view of
Quaternary sequence stratigraphy: siliciclastic accumulation on
slopes east of the Great Barrier Reef since the last glacial
maximum. Geology 31 (11), 1013 – 1016.
Pantin, H.M., 1991. The sea-bed sediments around the United
Kingdom. Research Report SB/90/1, British Geological Survey.
47 pp.
Pigram, C.J., Davies, P.J., Feary, D.A., Symonds, P.A., 1989. Tectonic controls on carbonate platform evolution in southern
Papua New Guinea: passive margin to foreland basin. Geology
17, 199 – 202.
Porter-Smith, R.P., Harris, P.T., Anderson, O., Coleman, R., Greenslade, D., 2004. Classification of the Australian continental shelf
based on predicted sediment threshold exceedance due to tidal
currents and swell waves. Mar. Geol. 211, 1 – 20.
Shepard, F.P., 1963. Submarine Geology. Harper & Row, New
York. 557 pp.
Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen,
K.A., Kromer, B., McCormack, G., van der Plicht, J., Spurk, M.,
1998. INTCAL98 radiocarbon age calibration, 24,000-0 cal BP.
Radiocarbon 40 (3), 1041 – 1083.
Vail, P.R., Mitchum, R.M., Thompson III, S., 1977. Seismic stratigraphy and global changes of sea level from coastal onlap. In:
Payton, E. (Ed.), Seismic Stratigraphy—Applications to Hydrocarbon Exploration, Memoir, vol. 26. American Association of
Petroleum Geologists, Tulsa, OK, pp. 63 – 81.
Van Waggoner, J.C., Mitchum, R.M., Campion, K.M., Rahmanian,
V.D., 1990. Siliciclastic Sequence Stratigraphy in Well Logs,
Cores and Outcrops: Concepts for High Resolution Correlation
of Time and Facies. American Association of Petroleum Geologists, Tulsa, OK.
Veron, J.E.N., 1978. Deltaic and dissected reefs of the far northern
region. Philos. Trans. R. Soc. Lond. 284, 23 – 37.
Walsh, J.P., Nittrouer, C.A., 2003. Contrasting styles of offshelf sediment accumulation in New Guinea. Mar. Geol. 196,
105 – 125.
Williams, M.A.J., 1984. Palaeoclimates and palaeoenvironments:
Quaternary environments. In: Veevers, J.J. (Ed.), Phanerozoic
Earth History of Australia. Clarendon Press, Oxford, pp. 42 – 47.
Woolfe, K.J., Larcombe, P., Naish, T., Purdon, R.G., 1998. Lowstand rivers need not incise the shelf: an example from the Great
Barrier Reef, Australia, with implications for sequence stratigraphic models. Geology 26, 75 – 78.