Long-shelf sediment transport and storm-bed formation by Cyclone Winifred, central Great Barrier Reef, Australia

Marine Geology 267 (2009) 101–113

Contents lists available at ScienceDirect

Marine Geology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m a r g e o

Long-shelf sediment transport and storm-bed formation by Cyclone Winifred, central Great Barrier Reef, Australia R.M. Carter a,⁎, P. Larcombe b, J.E. Dye a, M.K. Gagan c, D.P. Johnson a a b c

Marine Geophysical Laboratory, James Cook University, Townsville, Queensland, Australia Centre for Environment, Fisheries and Aquaculture Science, Lowestoft, UK Research School of Earth Sciences, Australian National University, Canberra, Australia

a r t i c l e

i n f o

Article history: Received 23 October 2007 Received in revised form 27 August 2009 Accepted 30 August 2009 Available online 11 September 2009 Communicated by J.T. Wells Keywords: cyclone Cyclone Winifred hurricane storm bed Great Barrier Reef shelf mid-cycle shellbed sand ribbons palimpsest shelf

a b s t r a c t Three major sediment facies belts occur on the Great Barrier Reef shelf: an inner shelf prism of Holocene terrigenous sand and mud, a patchy middle shelf veneer of palimpsest muddy shelly calcsand (<1 m thick) above a weathered Pleistocene erosion surface, and an apron of modern carbonate mud and sand around reefs of the main reef tract. These sediments were mobilized and strongly affected by the passage of severe tropical Cyclone Winifred (central pressure 958 hPa), which traversed the central Great Barrier Reef shelf on February 1, 1986. Re-surveys after the passage of Cyclone Winifred showed that (i) a storm-induced shell lag and/or a normally graded bed of terrigenous fine sand-mud had been deposited at depths down to 20 m; (ii) middle shelf longitudinal bedforms were widespread at depths of 28–35 m, and comprised a furrowed substrate of palimpsest muddy shell gravel surmounted by 40–150 m wide ribbons of quartzose and bioclastic sand up to 15 cm thick; (iii) fields of northward-facing, 1–2 m wavelength megaripples were present adjacent to zones of ribbons, and in the alleys between ribbons; (iv) five days after the cyclone, mud was still settling from suspension, and large parts of the inner shelf were bathed in muddy hypopycnal river plumes, some of which reached 30 km seaward to the inner edge of the reef tract; and (v) suspended mud, derived in part by unmixing of the seabed, was present throughout the shelf water column, and a seaward-thinning mud drape up to 40 cm thick had accumulated on and seawards of the inshore sediment prism, tapering seawards to a few mm thick only over most of the middle shelf. It is inferred that storm-waves and currents associated with the passage of Cyclone Winifred caused widespread unmixing of bottom sediments. Offshore, long-shelf transport of bedload sand ribbons and megaripples was affected by powerful shelf-parallel currents of velocity 1–3 m/s. As the storm passed, the graded, seaward-thinning sand–mud bed was deposited over the inner-middle shelf. On the middle shelf, bioturbation and downward mixing of the mud drape started immediately after the passage of the cyclone, and was well advanced after 3 months. In this way, the ephemeral seabed features produced by passage of a cyclone are gradually degraded, and the storm-layer stratigraphy becomes incorporated within a thin middle shelf veneer of muddy shelly calcsand. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Cyclone Winifred crossed the central Great Barrier Reef coast at Innisfail, near latitude 17° 30′ S, on February 1, 1986 (Fig. 1). The cyclone ranked as a category 3 tropical cyclone, with a 50–70 yr recurrence interval (Puotinen et al., 1997). Minimum central barometric pressure was 957 hPa, maximum measured wind gusts were 55 m/s, and the associated storm surge reached 1.6 m in height (Gagan et al., 1990). Modelling predicts deep water waves up to 9.0 m high and of 11.6 s period for a cyclone of this intensity (Bretschneider tropical cyclone

⁎ Corresponding author. E-mail address: [email protected] (R.M. Carter). 0025-3227/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.margeo.2009.08.009

model; Young, 1985). Landfall occurred at 7 p.m. on February 1st, 1986, and was followed by intense rainfall as the cyclone degenerated into a rain depression, causing the North and South catchments of the Johnstone River to flood and produce its highest recorded instantaneous flow of 5715 m3/s (Table 1). Serendipitously, the area of shelf traversed by the cyclone was at the time subject to detailed sedimentological study, funded by the Australian Marine Science and Technologies Grants Scheme. A moderately extensive pre-cyclone database therefore existed, which included two cross-shelf transects of sidescan sonar and 3.5 kHz seismic, 40 short cores taken from undisturbed van Veen framesupported-grab samples (most along two the cross-shelf transects), and 8 vibrocores at other significant locations. Shortly after Cyclone Winifred had passed, a rapid response dataset was collected that

102

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

Fig. 1. Locality map of the Innisfail area, central Great Barrier Reef. The path of the eye of Cyclone Winifred (bold, solid line) and the adjacent swathe of shelf waters that were affected by strong winds (bold, dotted lines) are indicated.

included replicate profiles along the sidescan and 3.5 kHz lines, 6 new sidescan and 3.5 kHz lines, 19 repeat grab samples, 6 new grab samples, 9 vibrocores, 14 camera stations, 6 video-camera stations, salinity profiles and water samples. These samples and observations were collected on r.v. James Kirby cruises KG 846, 854, 861, 862 863, 865 in 1984–86; cruise reports and samples are archived at James Cook University, Townsville. In addition, the locations of sedimentladen flood plumes were mapped along the coast by aerial observation on February 4, two days after Cyclone Winifred made landfall. Based on this dataset, it has been shown that Cyclone Winifred caused major sedimentary effects (Gagan et al., 1987, 1988, 1990). Extensive sediment remobilisation occurred from the inshore sediment prism, from the middle shelf substrate and from the offshore reef tract, and new sediment was also delivered to the shelf system via settling of fine-grained material from river flood plumes and the reef tract. As the storm subsided, an extensive fining-upward storm bed up

to 15 cm thick was deposited in waters up to 43 m or more deep, and extended at least 30 km offshore and 50 km along-shelf (Gagan et al., 1988). In total, the Winifred storm bed probably covered an area greater than 3000 km2 and contained an approximate volume of 140 × 106 m3 sediment, of which perhaps only 40 × 106 m3 represented new terrigenous material. Carbon isotope studies of pre- and post-Winifred samples (Gagan et al., 1987) indicate that most of the new riverine sediment contributed after the cyclone was deposited near river mouths, with little moving more than 15 km offshore. Erosion by wave and wind-forced currents eroded more than 14.4 cm from the middle shelf seabed, and significant portions of the resuspended sediment were advected alongshelf and at least 15 km shorewards before being deposited within the storm bed. Published observations of cyclone-induced currents are rare from Australian seas, though Church and Forbes (1983) recorded windforced residual currents of up to 0.5 knots (25 cm/s) from the Gulf of Carpentaria. Further west, on the Northwest Shelf, Hearn and

Table 1 Estimated water motions for cyclones of different intensity on the Great Barrier Reef shelf off Townsville, including storm surge and shallow and deep water wave-orbital velocities (after Young, 1985). Saffir–Simpson

NQ return

Pressure

Max. wind gust

WD current

Surge

Wave Hs

Wave Ts

Wave L

Umax

Umax

Scale

Magnitude

(Yr)

(hPa)

(Knots)

(m/s)

(cm/s)

(m)

(m)

(s)

(m)

(cm/s (deep))

(cm/s (shallow))

1 2 3 3 4 5

Mild Moderate Severe Severe Very severe Catastrophic

5 10 50 100 500 1000

>990 970–985 950–965 945–950 930–945 <925

40–60 70–90 100–120 120–130 130–150 160–180

20–30 35–45 50–60 60–65 65–75 80–90

40–60 70–90 100–110 120–130 130–150 160–180

0.0–1.0 1.5–2.5 3.0–4.0 4.0–4.5 4.5–5.5 6.0–7.0

8.3 8.5 9.0 9.2 9.7 10.0

11.2 11.3 11.6 11.8 12.1 12.2

193 197 207 214 225 229

204 212 234 245 269 280

75 77 81 83 88 90

The wind-driven current is also shown, calculated at 2% of the wind strength.

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

Holloway (1990) reported currents of up to 2 knots (100 cm/s) associated with Cyclone Jane (1983), and demonstrated amplification of the shelf current from 0.4 knots (20 cm/s) to 0.8 knots (40 cm/s) adjacent to the coast for Cyclone Ian (1982). Cyclone Joy (category 3) passed across the GBR shelf in 12 m of water off Cairns and alongshore flows occurred at sustained speeds of 60 cm/s and instantaneous speeds up to 140 cm/s near the bed (residual of 130 cm/s) (Larcombe and Carter, 2004). Given the dearth of in situ field measurements, documentation of storm-induced deposits such as those of Cyclone Winifred form a valuable archive with which to help hindcast the magnitude of near-bed flows during cyclones. Here, we add to the work of Gagan et al. (1987, 1988, 1990) and Larcombe and Carter (2004) by analyzing in particular the sediment transport and deposition implications of Cyclone Winifred. The earlier papers established the characteristics of the Winifred storm bed and the regional pattern of sediment movement that was involved, and contain information on the location of seismic and sidescan lines, grab samples, cores and bottom camera stations that, for reasons of length, we do not replicate here. In this new study, we demonstrate that during and after the passage of Cyclone Winifred the middle shelf seabed was subjected to extensive longitudinal currents with velocities up to 155 cm/s (3 knots) or faster, and that the microbathymetry of the middle shelf bedforms as seen in the immediate aftermath of the storm can be used to infer the type and direction of sediment transport. We also explore briefly the stratigraphic significance of these findings.

103

nearshore thickness is 4–5 m and its offshore extent 8–10 km. Offshore, the prism tapers to be only a few cm thick in depths greater than 20 m. Cores reveal that bedding is virtually never preserved within the coastal prism, because of the presence of pervasive bioturbation. Grain size analysis shows the sediment to be poorly sorted and trimodal, with modes in the fine sand, coarse silt, and fine silt ranges (e.g. Woolfe et al., 2000). The middle shelf (24–40 m) is sediment starved, and comprises a thin veneer (<1 m) of poorly sorted muddy, shelly, calcsand (Ohlenbusch, 1991) which locally thins out to expose the underlying weathered Pleistocene clay at the seabed. The general surface of the outer shelf (40–80 m) is also sediment starved in many places (Scoffin and Tudhope, 1985), though areas of framework and detrital carbonate sediment occur at and nearby to modern reefs (e.g. Maxwell, 1968), or where carbonate muds or Halimeda banks accumulate between the reefs (e.g. Dye, 1988; Orme, 1985). The boundary between the shore-connected sediment prism and the middle shelf shelly calcsand is shore-parallel, generally lies between the 20 and 25 m isobaths, and corresponds to an increase in carbonate content from <30% to >45% (e.g., Gagan et al., 1990). In contrast, the boundary between the reef carbonates (>95%) and the surrounding outer shelf sediment veneer is irregular. Generally, however, an apron of detrital carbonate surrounds each reef pedestal, narrow on the windward side but reaching up to several km from the reef on the leeward side. This apron is fed by off-reef transport of carbonate sand and gravel during both fair weather and storms (cf. Scoffin, 1993; Hughes, 1999).

2. Regional oceanography The central GBR shelf is bathed by tropical water masses and subjected to a seasonal summer-monsoonal climate. During winter months, the shelf water mass is under the influence of 15–25 knot southeasterly trade winds, which drive a regional northward flow of up to 40 cm/s. This flow is opposed by that of the East Australian Current (EAC) which flows southward along the shelf edge and slope, but occasionally penetrates far across the shelf to cause southwardflowing currents of about 20 cm/s during periods of calm weather (Burrage et al., 1996). In summer, winds are more variable and blow from the north and northeast, the weather is punctuated by occasional cyclones, and shelf waters are then periodically strongly influenced by the discharge of mud-laden river flood plumes, especially during and after cyclonic events (Devlin et al., 2001). Except during the summer months, then, the prevailing southeasterly trade winds on the central GBR shelf cause a steady drift of coastal and shelf waters to the north, at speeds up to 50 cm/s (e.g., Maxwell, 1968; Belperio, 1983; Wolanski and Spagnol, 2000). This regional drift is an important control on both bedload and suspended sediment transport and distribution. For instance, sediment mineralogical patterns show northward trends of increasing sediment maturity (Lambeck and Woolfe, 2000); offshore islands exhibit submergent, northward tapering, lee sediment banks; and, along the coastline, longshore drift builds similar northward-hooking, and mostly cape-attached, sediment bodies. Importantly, none of these low-moderate energy fair weather processes has the capacity to create the longitudinal bedforms and associated sedimentary features that were observed in the aftermath of Cyclone Winifred, and which we describe in this paper. 3. Regional facies distribution The central Great Barrier Reef shelf can be divided into three distinct sedimentation zones (Belperio, 1983) (Fig. 1). The inner shelf (down to depths of c.24 m) comprises a terrigenous, shore-connected prism of sand, muddy sand and sandy mud. In coastal embayments near river mouths, the terrigenous prism may exceed 20 m in thickness (Johnson and Searle, 1984), but more typically its maximum

4. Shelf bedforms observed prior to Cyclone Winifred Regional sidescan sonar surveys over wide areas reveal that during fair weather the surface of the inner shelf terrigenous prism is featureless at depths greater than 8 m. Updrift (north) of river mouths, sand is carried longshore in the form of low amplitude (<1 m high) foreshore bars, spaced a few tens of metres apart and oriented at an oblique angle to the adjacent beach. Within some embayments, the nearshore, sandy part of the sediment prism, down to depths of c. 10 m, sometimes carries fields of symmetric, wave-induced ripples with a wavelength of c.1 m. The situation on the middle shelf plain is markedly different. In many areas longitudinal bedforms are apparent on sidescan images, similar to the sand ribbons described from the tidally-influenced waters of the North Sea, as summarised by Kenyon (1970). Individual ribbons are diffuse and generally 50–150 m wide, at least many hundred metres long, and have a maximum thickness of 1–2 m. Between ribbons, the sidescan image reveals an essentially featureless but moderately reflective seabed, and 3.5 kHz profiles and cores demonstrate that this intra-ribbon seabed corresponds to the top of the Pleistocene (mapped locally as reflector A, after Johnson and Searle, 1984), which is intermittently covered with a thin veneer (<20 cm) of muddy shell hash. Significant to the impact of the cyclone-induced near-bed water motions inferred below, no megaripples were observed in these earlier, pre-cyclone surveys. 5. Post-cyclone bedforms and sediment distribution At a regional scale, the sidescan and 3.5 kHz lines run after the passage of Cyclone Winifred show that no major shift occurred in lithology nor the position of the borders of the three main shelf sediment facies belts (as reported also by Mearns et al. (1988) for Hurricane Diana). However, within the main middle shelf sand ribbon belt, at depths of 26–30 m, the individual ribbons were more sharply delineated than before the storm, suggesting their mobilization and showing that, at smaller scales, the currents induced by the cyclone had a major effect on the seabed.

104

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

On the inner shelf, slumping occurred on the south side of the natural channel at the mouth of Mourilyan Harbour (Fig. 2a), probably caused by undercutting of the bank by cyclone-induced waves. Waveinduced symmetrical ripples of <1 m wavelength were observed on sidescan images at around −10 m on the top of the terrigenous sediment prism off Mourilyan Harbour, and isobath-parallel sediment streaks, perhaps incipient sand ribbons, occurred in several places deeper than this on the seaward mud slopes of the inshore sediment prism. Geochemical evidence, including both carbon isotope and Pb210 profiles, indicates that during the passage of the cyclone the inner shelf seabed was eroded to depths up to 15 cm or more over wide areas (Gagan et al., 1987; 1990). The most dramatic seabed change, however, occurred on the middle shelf plain. Here, the pre-Winifred belt of diffuse sand ribbons was replaced by newly-formed, sharply delineated sand ribbons, between and around which the seabed surface was widely covered with 1.2 m wavelength megaripples (Fig. 3c). The new sand ribbons were of two main types. First, major ribbons up to 2 m thick, as seen previously in the 26–30 m belt, but now occurring widely across the shelf to depths as great as 38 m (Figs. 2d, 3). Second, faint, wispy ribbons which show no relief on 3.5 kHz profiles, and are therefore inferred to represent very thin skiffs of sand at particularly sediment starved locations. A third longitudinal bedform comprised gutters incised up to 1 m into Pleistocene reflector A (Fig. 2c). The megaripples occurred predominantly in belts <50 m wide between ribbons or gutters, with an average wavelength of 1.2 m, and diver observations indicate that those inspected faced north. Similar megaripples occurred as more extensive fields in areas where sediment starvation precluded the formation of sand ribbons, or occasionally as isolated, single trains on a reflective (presumed Pleistocene) seabed. 6. Other post-cyclone data 6.1. Aerial survey An aerial survey was performed to track the location of the buoyant sediment-laden plumes off river mouths, between 12 noon and 2 p.m. on February 4, two days after Cyclone Winifred made landfall. Not unexpectedly, major flood plumes were observed off the mouth of the Herbert, Tully and Johnston rivers, whose catchments received most of the post-cyclone rainfall (Fig. 4). Plume-fronts were located up to 30 km offshore, and those for the Johnston and Tully comprised up to four successive, tide-controlled pulses. Plumes were mostly located directly offshore, or slightly south (Tully) or north (Johnston and Herbert) of their point of origin. The Johnstone River plume had travelled across the shelf almost as far as the inner edge of the reef tract. Observations of other cyclones, as summarised by Devlin et al. (2001), show that the Cyclone Winifred pattern conformed to that normally seen after other major floods, whereby only the largest flood plumes extend beyond the inner shelf to impact directly upon the reef tract, the majority being confined to inner shelf waters.

105

6.2. Bottom photographs Extensive observations of seabed conditions were made between February 5 and February 9, 4–8 days after Cyclone Winifred made landfall. Photographic observations include both still and video coverage, mainly at locations at which repeat short cores were taken from frame-supported grab samples that capture a 15–20 cm deep, undisturbed section of the seabed substrate. Some stations were then reoccupied again between May 7 and 12, about 3 months after the cyclone (cruise KG-865). Four days after the cyclone, all bottom photographs were obscured by relatively high levels of turbidity at all stations on the middle shelf between the outer edge of the inshore sediment prism (c. 25 m depth) and the inner edge of the reef tract (c. 38 m depth). From a camera distance of c. 2 m, the seabed was only just visible, mostly by larger white shells or coral fragments (near the reef) that contrasted against the background gloom. Vibrocores and short cores collected at this time exhibited a mm thick skin of mud layered above a well sorted shelly sand, consistent with the start of settlement of a post-storm mud drape (Fig. 5b) (cf. Gagan et al., 1988). In contrast to the turbid bottom water immediately post-storm, 3 months after Winifred, bottom waters across the middle shelf were clear, allowing sharply focused images of the seabed to be captured. These images show a texture which ranged from shelly sand exposed at the seabed (with a small admixture of mud, as indicated around the bounce of the trigger weight; Fig. 5a), to a slightly hummocky completely mud-draped seabed (Fig. 5c). Near the reef, some larger coral fragments remained scattered across the muddy seabed. Bioturbation of the seabed is indicated by the hummocky texture of the surviving mud drape, and by the presence in the photos of discrete burrows and trails. Just 12 months after Cyclone Winifred, burrows and bioturbation were well developed at all stations photographed, and the integration into the shelly sand substrate of the post-storm mud drape was far advanced. 6.3. Salinity and turbidity data Two cross-shelf water-sample transects were completed 4 and 5 days after Cyclone Winifred made landfall, and were analyzed respectively for salinity and suspended sediment concentration (SSC) in mg/l (Fig. 7). Surface-to-near-bottom profiles of temperature and salinity measurements were obtained at 17 stations along the two transects. Water samples were collected from the surface and at selected depths through the water column using acid-cleaned Niskin bottles. Temperature and salinity (conductivity) were measured onboard using a laboratory salinometer (Hytech 6220). A further two 1000 ml aliquots of water were taken from each Niskin bottle and stored in sealed glass containers. In the laboratory, the 1000 ml water samples were filtered through oven-dried and pre-weighed glass fiber filters (Whatman GF/F, 25 mm) to determine total particulate matter. All samples were run in duplicate, dried and re-weighed to determine particulate weight. Measurements are recorded in mg liter− 1, and the averaged value is reported here.

Fig. 2. Selected segments of 3.5 kHz profiles between the shoreline and reef tract, east of Mourilyan Harbour, all except (a) directed in a SW–NE direction. (a) SE–NW profile across the edge of the dredged Mourilyan shipping channel immediately seawards of the harbour mouth; the 4 m thick inshore sedimentary prism is prograding NE into the channel across Reflector A (RA); after the passage of Cyclone Winifred a substantial failure occurred in the southerly wall and part of the prism slid into the channel. (b) Profile across the edge of the prism just offshore from Mourilyan Harbour; the presence of substantial longshore sand bars and currents at this location results in a seaward disjunct in the usually smoothly tapering edge of the sediment prism; seawards of this disjunct, the upper part of the prism of sediment above RA may have been emplaced during Cyclone Winifred; below this, the channel-incised RA merges with the transgressive surface to form the substrate for the middle shelf plain. (c–e) Successive profiles across the sand ribbon terrain of the middle shelf (cf. Fig. 3); all variations exist between (c) RA lying almost at a seabed mantled with a few cm thick veneer of tough shelly clay, and incised by erosive furrows; to (d) RA lying a few tens of cm below the crests of fields of mobile sand ribbons separated by furrows; to (e) RA lying 1–2 m below a semi-continuous Holocene sediment blanket that may have formed by the coalescence of successive generations of sand ribbons. The sand ribbons and erosive gutters between them are created during the passage of cyclones only; between cyclones they remain on the seafloor as relict features which, with increasing time and bioturbation, become intermixed to form a regional, middle shelf palimpsest deposit. (f) Profile across the edge of a typical peri-reef bioclastic sediment apron at Howie Reef. Note the wide variation in sedimentation rates across the shelf, from 0.2 cm/ky in the middle shelf lag veneer, to 2 cm/ky in the sand ribbon terrain, to 250 cm/ky in peri-reef apron locations, to 500 cm/ky or more in bay depocentres of the shore-connected sediment prism. The 3.5 kHz profiles were typically run at a ship speed of 5 knots, and the seabed scale therefore, varies slightly between different lines according to weather and currents. The horizontal scale shown on (f) is indicative for all profiles (b–f); profile (a) has its own scale.

106

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

Fig. 3. Longitudinal bedforms on the middle shelf of the Great Barrier Reef lagoon. (a) Seabed morphology in 28–32 m water depth as mapped by LADS measurements, 50 km northeast of Cape Cleveland and 15 km southeast of Broadhurst Reef (data courtesy Royal Australian Navy). Note the presence of widespread sand ribbons 100–200 m wide, an area of diffuse (“sediment starved”) dunes, and the presence of erosive scour pits, the orientations of which are consistent with westerly and northwesterly flowing currents. (b) 3.5 kHz profile and sidescan images of middle shelf sand ribbons mobilized during Cyclone Winifred; location, 32 m water depth, 10 km WSW of the western edge of Howie Reef. (c) Detail of sidescan image nearby, showing the development of fields of 1.2–1.5 m wavelength megaripples in shell gravel in the gutters between adjacent sand ribbons. Similar ripples were intermittently but widely present across the middle shelf after Winifred, indicating slightly more easterly current orientations than do the adjacent ribbons, and may represent transport and deposition from post-cyclone storm-waves. Panels a–c (also, Fig. 2) demonstrate that the Holocene sediment veneer is generally less than 1 m thick, and that the top of the Pleistocene clay that marks the LGM lowstand land surface (Reflector A) lies at or close below the seabed over the entire middle shelf.

The transect off Mourilyan Harbour mouth delineated a surface plume of low salinity water (<34‰) which is <10 m deep and extends 15 km offshore. Most of the middle shelf was bathed in water of intermediate salinity, 34–35‰, with fully marine values of 36‰ characteristic of the inner edge of the reef tract (Fig. 7). Thus the

salinity profile shows a close correlation with the distribution of turbid flood plumes that were observed during the aerial overflight (cf. Fig. 4). Subsequent sampling of other flood plumes in the Great Barrier Reef has shown that though these plumes are visually spectacular, volumetrically they carry very little sediment, having SSCs of only 3–10 mg/l (Taylor,

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

107

Fig. 4. Seaward extent of flood plumes from the Herbert, Tully and Johnstone Rivers, observed from aerial overflight two days after Cyclone Winifred. Successive plume-fronts represent tidal pulses. The strong northerly winds that ensued after passage of the cyclone inhibited the normal northerly-drift for such flood plumes (e.g. Wolanski, 1994), causing them to spread more directly eastward across the shelf. Note that the majority of sediment settlement from these dilute plumes is likely occurring across the middle shelf (cf. Fig. 5). A–D mark the measurement locations for the current meter results that are plotted in Fig. 8.

1996), and therefore equating to a sediment layer around 10 µm thick (Orpin et al., 1999). Three features characterise the distribution of suspended sediment on the cross-shelf line between Ella Bay and Howie Reef (Fig. 7). Higher than average SSCs are associated with: • the fresh-water flood plume issuing from the mouth of the Johnson River (>5 mg/l); • a half-bull's-eye pattern of enhanced turbidity centred at the seabed c.28 m deep on the middle shelf, and extending laterally between c.25 m and 34 m depth (>6 mg/l); and • a strong vertical gradient of turbidity at the inner edge of the reef tract (>5 mg/l). These patterns of enhanced turbidity are consistent with Cyclone Winifred providing suspended sediment to the shelf system in three separate ways: first, by the delivery of new terrestrial material in flood 364 plumes at the river mouth (Fig. 6b); second, by unmixing of seafloor material from the middle shelf sediment veneer (Fig. 6a), together with erosion of new material from Pleistocene clay where it becomes exposed at the seabed, and; third, by erosion and resuspension of carbonate material from the reef tract. The strong, near-bed pattern of enhanced SSCs on the middle shelf is consistent with derivation of the suspended sediment from the seabed, and also with its transport by the strong along-shelf currents inferred to have been driven by the cyclone. The Cyclone Winifred observations provide no evidence for the surface (hypopycnal) sediment plume giving rise to seawarddirected turbid hyperpycnal flows by convective fingering from above, as simulated, for example, by Parsons et al. (2001) in laboratory experiments.

6.4. Current measurements No oceanographic instrumentation was deployed in the direct path traversed by Cyclone Winifred across the shelf. However, Bode and Wolanski (1986) and Wolanski and Ridd (1990) reported on a cross-shelf instrumented transect from east of Hinchinbrook Island, 80 km south of Winifred's path (Fig. 8). This transect shows that a surficial fresh-water plume that extended seawards from the Herbert River arrived on the middle shelf on February 2nd, about 1 day after the cyclone made landfall (cf. Fig. 4), and persisted for 6 days until February 8th. Subsequently, the plume exhibited pulsatory tidal rhythms with an amplitude of 24 h, and recovered towards full marine salinities, presumably by mixing with ambient water, especially at station D (42 m water depth, immediately west of Britomart Reef). The data also showed that current direction prior to landfall was uniformly to the NW, at speeds greater than 50 cm/s across the width of the shelf, peaking at 80 cm/s at inshore station A. After a 24 hour lull associated with the eye of the storm, the wind reversed to drive a post-cyclone current SE-SSE at 30–60 cm/s. Though this current data was collected at sites 80 km from the point of storm landfall, importantly it shows that: (i) the entire shelf water mass was entrained in long-shelf flow, with little difference in velocities between the inner and middle shelf, and with little vertical stratification between surface and bottom current meters; (ii) currents were directed predominantly along-shelf both immediately before and after the passage of the cyclone; and (iii) northwardflowing currents peaked immediately prior to the cyclone making landfall. Data from the Hinchinbrook shelf-transect show, first, that the river-fed turbid underflow model (Goldring and Bridges, 1973;

108

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

of linear bedforms which also includes current lineations (McKinney, 1974) and longitudinal furrows (Dyer, 1970; Flood, 1981). 7.1. Longitudinal bedforms on the Great Barrier Reef shelf Data collected before and after Cyclone Winifred, and images based on high-resolution Laser Airborne Depth Sounder (LADS) mapping (Larcombe and Carter, 2004), show that longitudinal bedforms are widespread on the low-gradient middle shelf seabed of the Great Barrier Reef lagoon. Both sand ribbons and longitudinal furrows occur, often nearby to fields of megaripples. Theoretical calculations and limited observations indicate that fair weather waveinduced and tidal currents on the inner and middle GBR shelf generally have velocities less than 0.5 m/s (Church et al., 1985; Orpin et al., 1999). However, erosion of the Pleistocene substrate, unmixing of Holocene transgressive sediments (Fig. 6a), and movement of sand ribbons require velocities of at least 1 m/s (Kenyon, 1970). Therefore, it is probable that the inferred erosion of Pleistocene and Holocene sediment, and its longitudinal transport in sand ribbons, takes place only during highly energetic events, which on the GBR shelf equates with the passage of tropical cyclones. Within the shelf area traversed by Cyclone Winifred, 3.5 kHz profiles and sidescan data collected before the cyclone show that the longitudinal-furrowed terrain in part predates the cyclone. Nonetheless, the widespread fields of fresh megaripples, some furrowed terrain, and at least the upper parts of the sand ribbons, were demonstrably the result of Cyclone Winifred-induced water motions, because:

Fig. 5. (a, c) Bottom photographs across a middle shelf field of relict sand ribbons at station 865-72, 32 m water depth, 10 km WSW of Feather Reef three months after the passage of Cyclone Winifred. Between ribbons, the seabed still carries an almost continuous drape of post-cyclone mud a few mm thick, albeit with some bioturbatory modification (a); bioturbation is more advanced in other areas, probably on ribbon crests, causing downward mixing of mud into the subjacent sand ribbon, leaving only discontinuous patches of modified mud drape at the surface (c). This process of mixing the post-cyclone drape (and even younger sediment and shell) into the underlying relict sand ribbon produces the poorly sorted, poorly stratified, palimpsest, muddy shell sand that covers wide areas of the modern middle shelf. (b) Field photograph of the top of a freshly collected core at station 863-4, 32 m water depth 8 km WSW of Feather Reef, 7 days after Winifred made landfall. The underlying shelly sand ribbon is mantled with a 1–2 mm thick post-cyclone mud drape, similar to that being mixed down from the seabed in photos (a) and (c).

Wright et al., 1986; Higgs, 1990) did not apply to the deposition of the Winifred storm bed. Second, the long-shelf sediment transport which occurred during the cyclone was a response to direct wind-forcing of the shelf water mass within the cyclone corridor, and not caused by either geostrophic flow (Hart et al., 1990) or by a semi-permanent oceanic current (Leckie and Krystinik, 1990). Furthermore, during GBR cyclones, strong along-shelf flows operate right into nearshore shallow water (Larcombe and Carter, 2004). In contrast with observations on other modern (Niederoda et al., 1984) and ancient (Leckie and Krystinik, 1989; Hart et al., 1990) shelves, the Cyclone Winifred data provide no evidence for a significant storm-associated nearshore belt of offshore-directed bedload sediment transport driven by storm surge relaxation currents. 7. Significance of longitudinal bedforms Large-scale bedforms parallel to the prevailing current orientation were first identified in the marine environment by Stride (1959), using a developmental sidescan sonar system. These features, later termed sand ribbons (Stride, 1963, 1982), are one member of the class

• only very faint ribbons, and no megaripples, are visible on the precyclone sidescan line; • most of the middle shelf, including fields of megaripples and sand ribbons, was covered by a mud drape up to several mm thick 4 days after the cyclone, implying that currents generated by the strong northerly 16 m/s winds after the cyclone were too weak to modify the megaripples and ribbons created during passage of the cyclone; currents generated by normal (strong) fair weather southeasterly wind velocities of up to 12.5 m/s are therefore even less likely to affect the middle shelf seafloor (cf. Orpin et al., 1999); • bottom photographs and grabs taken up to 3 months after the cyclone show that the mud drape was being gradually incorporated into the substrate by bioturbation; • bottom photographs and video coverage collected 3 months after the cyclone at shelf stations at depths between 24 m and 40 m show no newly-formed megaripples nor any other current-induced features. 8. Intensity of cyclone-generated currents Cyclone Winifred attained maximum wind speeds of 55–60 m/s where it crossed the coast (Walker and Reardon, 1986), but the near coincidence of cyclone landfall and low tide restricted the storm surge to 1.8 m (Beach Protection Authority, 1986). Lawford and Veley (1956) and Caston (1976) have shown that, given sufficient time, near-surface, wind-driven flows in storms may reach 2–4% of the wind speed. Thus, Cyclone Winifred may have produced surface currents of up to 2.5 m/s in speed (Table 1). In an unstratified water column, and at depths of 20–40 m, these currents would diminish to c.1.5–2.0 m/s at 2 m above the seabed. Storm-forced long-shelf currents of similar strength to this have been measured elsewhere, for example currents up to 1.6 m/s were measured as Hurricanes Camille (Murray, 1970) and Delia (Forristall et al., 1977) traversed the Gulf of Mexico shelf, and velocities up to 1.4 m/s were recorded from the GBR shelf during the passage of Cyclone Joy (Larcombe and Carter, 2004). Broadly consistent with this, a numerical current model for Winifred predicted that northerly, depth-averaged, long-shelf currents peaked at

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

109

Fig. 6. (a) Pre- (left) and 4 day post-Winifred (right) short cores from station 863-74, 25 m water depth, 20 km east of Mourilyan Harbour. Prior to the cyclone the substrate comprised bioturbated, structureless, very poorly sorted muddy, shelly sand. The post-Winifred core consists of well sorted sand that has been unmixed from the palimpsest substrate by storm erosion (cf. Gagan et al., 1990), after which insufficient time had elapsed for a mud drape to accumulate. (b) Post-Winifred short core from station 863-20, 8 m water depth on the sandy mouth bar of the Johnstone River. The post-Winifred mud drape rests on well sorted nearshore sand and, having accumulated to a thickness of ∼3 cm in 8 days, indicates a post-storm period of calm weather.

1.3 m/s during the 3 h before landfall, or up to 1.5 m/s when augmented by tidal flow (after Harper et al., 1977; Lance Bode, pers. comm.). The model uses meteorological data (central pressure, speed and direction

of storm-track, radius of maximum winds) and known astronomical tidal elevations as the basis for forcing, and predicts that the response of the north Queensland shelf watermass is relatively insensitive to the direction of approach of a cyclone; consistent, shelf-wide, northerly, long-shelf currents are generated irrespective of whether the cyclone moves parallel, oblique or normal to the shoreline. For the Winifred model, subtraction of the tidal fields yielded the net storm-induced response, and the storm surge predicted at coastal locations matched well with shore-based observations (Bode and Wolanski, 1986; Beach Protection Authority, 1986). Testament to the sustained magnitude of the current flows associated with Cyclone Winifred was an elegant wind-drogue experiment whereby a floating vibrocore platform — lost at Green Island, Cairns during the passage of the storm — was transported to John Brewer reef, a distance of 235 km of southward transport in 4 days that indicates an average surface current speed of 0.7 m/s (Peter Walbran, pers. comm.). Thus, despite the absence of groundtruth current speed measurements under the direct path of Cyclone Winifred, the available data, supported by computer model projections, suggests the operation of current speeds of up to 150 cm/s, i.e., almost 3 knots. In addition to the directional currents produced by the cyclone, substantial wave-orbital motions will have affected bottom sediments. Despite the limited fetch imposed by the presence of the offshore reef tract, and wave attenuation caused by energy absorption of the seabed in shallow water (e.g. Forristall and Reece, 1985; Kraft et al., 1990), it is likely that waves during the passage of Cyclone Winifred reached at least 4.5 m in height with a 10–12 s period (cf. data measured for Cyclones Keith, Nancy and Otto in 1977; Beach Protection Authority, 1984). Thus near-bed orbital velocities, as calculated from linear Airy wave theory for middle shelf depths of 25–35 m, may have been as high as 0.7 m/s. Clearly, storm-generated wave motions will have added to the wind-forced current in eroding and mobilizing bottom sediment (cf. Orpin et al., 1999), thus allowing coarser-grained sediment to be entrained within directional storm currents. 9. Cyclone-induced sediment transport mechanisms

Fig. 7. Cross-shelf, water column transects east of Mourilyan 4–5 days after the landfall of Cyclone Winifred (for location, see Fig. 1). (a) Suspended sediment concentration (mg/l); (b) salinity.

The evidence presented in the first part of this study, especially the widespread and enhanced post-cyclone presence of longitudinal

110

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

Fig. 8. Oceanographic data from GBR shelf-transect between Hinchinbrook Island and Britomart Reef, 80 km south of the landfall of Cyclone Winifred (location, Fig. 4) (after Wolanski and Ridd, 1990). (a) Instruments deployed; (b) surface salinity profiles for stations C and D between January 31st and February 10th (1986); (c) current vector diagrams for stations A–D between January 25th and February 8th (1986); note the reversal of wind direction as the storm made landfall and the continuation of strong northerly winds afterwards for several days.

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

bedforms, and the theoretical models of water motion, establish that Cyclone Winifred had a widespread effect on the seabed. Perhaps surprisingly, therefore, a study by Mearns et al. (1988) off the coast of North Carolina found that hurricane Diana (central pressure 949 mb) produced no measurable seafloor changes in sediment distribution. Our analysis shows that this conclusion applies also to the passage of Cyclone Winifred at the level of the distribution of the major sediment facies belts. However, within each facies belt, Cyclone Winifred exerted major effects on the distribution of sediment, as manifested by an abundance of new, mesoscale sedimentary bedforms and the presence of a shelf-wide storm bed after its passage. During and immediately after the passage of Winifred, cyclonegenerated currents operated from the shoreline at least out to the inner edge of the reef tract, 40 km offshore and rising from depths of 40–50 m. Theoretically, four cyclone-induced current mechanisms might be involved: • direct northerly, wind-driven, along-shelf currents through the whole cross-shelf water column; • easterly to southeasterly, wind- and pressure-forced, bottom return flows generated at the peak of the storm; • easterly to northeasterly, post-storm, relaxation currents generated by the decay of the storm surge; and • oscillatory bottom currents generated by storm-wave activity. Away from the equator, both storm surge relaxation currents and wind-forced three-dimensional bottom currents are affected by Coriolis forces, in the Southern Hemisphere turning left longshore as they traverse the shelf and merge with the geostrophic flow (cf. Swift and Rice, 1984). Although it is possible that the passage of Winifred induced a short phase of easterly, turning northerly, threedimensional bottom currents on the innermost shelf, at latitude 18° S such current components are weak and were therefore probably overwhelmed by the strong, northerly wind-forced flow which occurred for several days once the cyclone eye had made landfall (Fig. 8). Therefore, the graded bed across the inner and middle shelf was most likely deposited from the waning stages of the northerly storm-forced long-shelf bottom current (cf. Morton, 1981, for Hurricane Carla), rather than from offshore-moving relaxation motions (cf. Hayes, 1967, for Hurricane Carla). At 80 km south of the location of landfall, the relative magnitudes of the measured prelandfall northward storm-forced flow and the post-landfall southerly flow are more equal (Fig. 8) than would be expected near the location of landfall, due to pre-landfall weaker northward components of wind, the occurrence of post-landfall northerly winds, and, possibly in this case, local effects of coastal orientation. The sedimentary effects of the currents generated by Cyclone Winifred manifestly included: • erosion, undercutting and slumping of sediment on the inner shelf (Fig. 2a, b); • erosion, unmixing or winnowing of pre-existing bottom sediment, and of the top of the Pleistocene, at all depths across the innermiddle shelf to the reef tract; • extensive northward long-shelf transport of sand ribbons and megaripples in depths between about 24 m and 40 m (Figs. 2c, 3), as manifested by sidescan images and by the deposition of a poorly graded carbonate sand bed up to 15 cm thick within the middle shelf zone of ribbons; • coral breakage and bottom sediment resuspension from some reefs near the path of the cyclone (cf. Fig. 2d); and • as the storm-forced, northward, shelf-wide current waned, deposition of a terrigenous or shelly sand bed occurred over a wide area of the ISP to seabed depths of about 24 m; this was closely followed by the deposition of a seaward-thinning mud drape, greater than 10 cm and possibly up to 44 cm thick on the inner shelf, a few mm

111

thick on the middle shelf, and locally more than 1 cm thick (due to resettled carbonate mud) near the reef (Gagan et al., 1988); and • strong advection of suspended sediment from the middle shelf onto the outer parts of the ISP (Gagan et al., 1987, 1990). 10. Stratigraphic significance Although sand ribbons have been described from modern locations which are consistent with their movement by storm-forced currents (Carter et al., 1985), this is the first full description of a major occurrence of longitudinal bedforms, including both ribbons and furrows, in a demonstrably storm-controlled sedimentary environment (cf. Larcombe and Carter, 2004). The offshore muddy calcsands which represent longitudinal storm transport on the GBR shelf rest directly on the lowstand Pleistocene substrate, i.e. occur in the same stratigraphic position as the basal shell hash or transgressive “sand sheet” described from the eastern USA shelf, and there inferred to have been deposited by erosional shoreface retreat (e.g. Swift et al., 1972; Figueiredo et al., 1982). Similarly, Saito et al. (1989; their Unit 2) show that the middle-outer shelf off Sendai, Japan is covered by a transgressive sheet of coarse pebbly, shelly sand of lag and palimpsest origin. Similar mid-outer shelf sediments again are described from the Brazilian shelf by Testa and Bosence (1999) and the South African shelf by Flemming (1980), who stresses the relict and carbonate-rich aspect of these sediment starved facies. Considering ancient sediments, the GBR muddy calcsands are similar to the carbonate-rich inferred storm beds described from the Jurassic of Wyoming (Brenner and Davies, 1973) and France (Seguret et al., 2001), the early Carboniferous of Morocco (Kelling and Mullin, 1975), and the Cenozoic of Egypt (Nebelsick and Kroh, 2002). However, in these older carbonate successions, and in almost all other storm-dominated sequences which have been described, the individual storm beds have erosive, or at least sharp, bases, and possess conspicuous normal size grading. The GBR cyclone beds are apparently unique in having gradational lower boundaries (representing seabed unmixing during passage of the storm), and in having their depositional grading largely obliterated by subsequent bioturbation. In ancient successions, similar storm-bed facies can be expected to be present as a condensed succession of slightly muddy, flaggy, shelly calcsand. Understanding the stratigraphic setting will, however, be crucial to their correct interpretation, for the modern GBR condensed storm-bed facies occupy the position of the mid-cycle shellbed in the sequence stratigraphic model (Abbott and Carter, 1994; Larcombe and Carter, 1998). In consequence, such storm-bed limestones will (i) either lie on or just above a transgressive ravinement surface, and there correspond to a composite shellbed in the sense of Naish and Kamp (1997); or (ii) be separated from the ravinement surface by a greater or lesser thickness of transgressive systems tract deposits (i.e., usually a terrigenous shoreface prism); and, (iii) in both cases will be overlain by progradational highstand and regressive systems tracts, again usually terrigenous in nature. The still-forming, widespread condensed storm-bed veneer on the middle GBR shelf corresponds to circumstance (i) above. Potentially, circumstance (ii) will apply to some earlier Holocene examples of the GBR mid-cycle storm bed that are now enclosed at a depth of several metres within the nearshore sediment prism, though no cores have yet been described that demonstrate this. 11. Conclusions During the passage of Cyclone Winifred, reef and bottom sediments were eroded and resuspended across the shelf by storm-waves. Concomitantly, the onset of strong, northward, long-shelf currents of between 1.0 and 3.0 m/s resulted in extensive seabed erosion and sediment transport. Along the inner shelf sediment prism, this transport and winnowing was in places marked by the production of

112

R.M. Carter et al. / Marine Geology 267 (2009) 101–113

shell-lags. As currents waned, fields of 1–2 m wavelength megaripples developed where mobile bedload sediment was available on the middle shelf, whilst, nearshore, deposition of the sandy-silt base of the graded bed commenced. Several hours after the storm made landfall, the wind-driven long-shelf currents reversed, were perhaps briefly augmented by storm surge relaxation, and the upper part of the graded storm bed and an ensuing shelf-wide mud drape settled from suspension as sea conditions returned to normal after the passage of the cyclone. As inferred also by Bentley et al. (2002) for Hurricane Camille and Larcombe and Carter (2004) for Cyclone Joy, it is likely that the strong wind-forced currents which caused widespread sediment mobilization occurred only for the few peak hours of storm passage of Cyclone Winifred. Subsequently, under less energetic fair weather conditions, burrowing and bioturbation becomes the dominant process affecting bottom sediments, bedforms become degraded and are eventually destroyed, and the storm layer is mixed downwards and integrated into an essentially structureless, composite and palimpsest sedimentary unit. As Keen et al. (2004) have similarly concluded for hurricanes in the Gulf of Mexico “physical and biological post-depositional processes have reworked the event layers, producing regional discontinuities and localized truncation, and resulting in an imperfect and biased record of sedimentary processes during the storms”. Acknowledgements We thank the captains and crew of the RV James Kirby for their expert help and support during field operations, Kevin Hooper and Neil Mockett for their technical assistance, Alistair Birtles for diving observations and Lance Bode for the information regarding computer model runs of Cyclone Winifred. Financial support for this work was provided by the Australian Marine Science and Technologies grant no 83/1147, by the Australian Research Council, and by James Cook University. References Abbott, S.T., Carter, R.M., 1994. The sequence architecture of mid-Pleistocene (0.35– 0.95 Ma) cyclothems from New Zealand: facies development during a period of known orbital control on sea-level cyclicity. Orbital Forcing and Cyclic Sequences: In: de Boer, P.L., Smith, D.G. (Eds.), Int. Assoc. Sedimentol., Spec.Publ., 19, pp. 367–394. Beach Protection Authority (Queensland), 1984. Mulgrave Shire Northern Beaches. Beach Protection Authority, Queensland. 366 pp. Beach Protection Authority (Queensland), 1986. Cyclone ‘Winifred’ hits North Queensland. BPA Newsletter. Belperio, A.P., 1983. Terrigenous sedimentation in the central Great Barrier Reef lagoon: a model from the Burdekin region. BMR J. Aust. Geol. Geophys. 8, 179–190. Bentley, S.J., Keen, T.R., Blain, C.A., Vaughan, W.C., 2002. The origin and preservation of a major hurricane event bed in the northern Gulf of Mexico: Hurricane Camille, 1969. Mar. Geol. 186, 423–446. Bode, L., Wolanski, E., 1986. Some comparisons between observed and modelled oceanographic response to tropical cyclone Winifred. In: Dutton, I.M. (Ed.), The Offshore Effects of Cyclone Winifred. Great Barrier Reef Marine Park Authority Workshop, Series 7, p. 43 and figs. Brenner, R.L., Davies, D.K., 1973. Storm-generated coquinoid sandstone: genesis of high-energy marine sediments from the Upper Jurassic of Wyoming and Montana. Bull. Geol. Soc. Amer. 84, 1685–1698. Burrage, D.M., Steinberg, C.R., Skirving, W.J., Kleypas, J.A., 1996. Mesoscale circulation features of the Great Barrier Reef region inferred from NOAA satellite imagery. Remote Sens. Environ. 56, 21–41. Carter, R.M., Carter, L., Williams, J., Landis, C.A., 1985. Modern and relict sedimentation on the Otago continental shelf. N.Z. Oceanogr., Mem. 93 43 pp. Caston, V.N.D., 1976. A wind-driven near-bottom current in the Southern North Sea. Estuar. Coast. Mar. Sci. 4, 23–32. Church, J.A., Forbes, A.M.G., 1983. Circulation in the Gulf of Carpentaria. I. Direct observations of currents in the south-east corner of the Gulf of Carpentaria. Aus. J. Mar. Freshw. Res. 34, 1–10. Church, J.A., Andrews, J.C., Boland, F.M., 1985. Tidal currents in the central Great Barrier Reef. Continent. Shelf Res. 4, 515–531. Devlin, M., Waterhouse, J., Taylor, J., Brodie, J., 2001. Flood plumes in the Great Barrier Reef: spatial and temporal patterns in composition and distribution. Great Barrier Reef Marine Park Authority Rep., 114 pp. Dye, J.E., 1988. Late Pleistocene and Holocene development of the Maori and Hedley reef areas, central Great Barrier Reef. In: Global Sea-level Change and the

Stratigraphic Record, (ESSO Distinguished Lecture Programme, Econ. Geol. Res. Unit, James Cook Univ., Townsville), extended abstract, 29–35. Dyer, K.R., 1970. Linear erosional furrows in Southampton Water. Nature 225, 56–58. Figueiredo, A.G., Sanders, J.E., Swift, D.J.P., 1982. Storm-graded layers on inner continental shelves: examples from Southern Brazil and the Atlantic coast of the central United States. Sedim. Geol. 31, 171–190. Flemming, B.W., 1980. Sand transport and bedform patterns on the continental shelf between Durban and Port Elizabeth (southeast African continental margin). Sedim. Geol. 26, 179–205. Flood, R.D., 1981. Distribution, morphology and origin of sedimentary furrows in cohesive sediments, Southampton Water. Sedimentology 28, 511–529. Forristall, G.Z., Reece, A.M., 1985. Measurements of wave attenuation due to a soft bottom: the SWAMP experiment. J. Geophys. Res. 90, 3367–3380. Forristall, G.Z., Hamilton, R.C., Cardone, V.J., 1977. Continental shelf currents in Tropical Storm Delia: observations and theory. J. Phys. Oceanog. 7, 532–546. Gagan, M.K., Sandstrom, M.W., Chivas, A.R., 1987. Restricted terrestrial carbon input to the continental shelf during Cyclone Winifred: implications for terrestrial runoff to the Great Barrier Reef Province. Coral Reefs 6, 113–119. Gagan, M., Johnson, D.P., Carter, R.M., 1988. The Cyclone Winifred storm bed, central Great Barrier Reef shelf. Australia. J. Sedim. Petrol. 58, 845–856. Gagan, M.K., Chivas, A.R., Herczeg, A.L., 1990. Shelf-wide erosion, deposition, and suspended sediment transport during cyclone Winifred, central Great Barrier Reef. Australia. J. Sedim. Petrol. 60, 456–470. Goldring, R., Bridges, P., 1973. Sublittoral sheet sandstones. J. Sedimen. Petrol. 43, 736–747. Harper, B.A., Sobey, R.J., Stark, K.P., 1977. Numerical simulation of tropical cyclone storm surge along the Queensland coast. Part IV. Innisfail. Department of Civil and Systems Engineering Report. James Cook University, Townsville. Hart, B.S., Vantfoort, R.M., Plint, A.G., 1990. Is there evidence for geostrophic currents preserved in the sedimentary record of inner to middle-shelf deposits? — A discussion. J. Sedim. Petrol. 60, 633–635. Hayes, M.O., 1967. Hurricanes as geologic agents: case studies of Hurricanes Carla, 1961, and Cindy, 1963. Rept. Inv. No. 61. Bur. Econ. Geol., Texas Univ. Hearn, C.J., Holloway, P.E., 1990. A three-dimensional barotropic model of the response of the Australian North West Shelf to tropical cyclones. J. Phys. Oceanog. 20, 60–80. Higgs, R., 1990. Is there evidence for geostrophic currents preserved in the sedimentary record of inner to middle shelf deposits? — Discussion. J. Sedim. Petrol. 60, 630–632. Hughes, T.P., 1999. Off-reef transport of coral fragments at Lizard Island. Australia. Mar. Geol. 157, 1–6. Johnson, D.P., Searle, D.E., 1984. Post-glacial seismic stratigraphy, central Great Barrier Reef, Australia. Sedimentol. 31, 335–352. Keen, T.R., Bentley, S.J., Vaughan, W.C., Blain, C.A., 2004. The generation and preservation of multiple hurricane beds in the northern Gulf of Mexico. Mar. Geol. 210, 79–105. Kelling, G., Mullin, P.R., 1975. Graded limestones and limestone-quartzite couplets: possible storm deposits from the Moroccan Carboniferous. Sedim. Geol. 13, 161–190. Kenyon, N.H., 1970. Sand ribbons of European tidal seas. Mar. Geol. 9, 25–39. Kraft, L.M., Suhayda, J.N., Helfrich, S.C., Marin, J.E., 1990. Ocean wave attenuation due to soft seafloor sediments. Mar. Geotech. 9, 227–242. Lambeck, A., Woolfe, K.J., 2000. Composition and textural variability along the 10 m isobath, Great Barrier Reef: evidence for pervasive northward sediment transport. Aus. J. Earth Sci. 47, 327–335. Larcombe, P., Carter, R.M., 1998. Sequence architecture during the Holocene transgression: an example from the Great Barrier Reef shelf. Australia. Sedimen. Geol. 117, 97–121. Larcombe, P., Carter, R.M., 2004. Cyclone pumping, sediment partitioning and the development of the Great Barrier Reef shelf system: a review. Quat. Sci. Rev. 23, 107–135. Lawford, A.L., Veley, V.F.C., 1956. Change in the relationship between wind and surface water movement at higher wind speeds. Trans. Amer. Geophys. Union 37, 691–693. Leckie, D.A., Krystinik, L.F., 1989. Is there evidence for geostrophic currents preserved in the sedimentary record of inner to middle shelf deposits? J. Sedim. Petrol. 59, 862–870. Leckie, D.A., Krystinik, L.F., 1990. Is there evidence for geostrophic currents preserved in the sedimentary record of inner to middle shelf deposits — reply. J. Sedim. Petrol. 60, 636–637. Maxwell, W.G.H., 1968. Atlas of the Great Barrier Reef. Elsevier, Amsterdam. 258 pp. McKinney, T.F., 1974. Large-scale current lineations on the central New Jersey shelf: investigations by side-scan sonar. Mar. Geol. 17, 79–102. Mearns, D.L., Hine, A.C., Riggs, S.R., 1988. Comparison of sonographs taken before and after Hurricane Diana, Onslow Bay, North Carolina. Geology 16, 267–270. Morton, R.A., 1981. Formation of storm deposits by wind-forced currents in the Gulf of Mexico and the North Sea. Int. Assn. Sedimentol. Spec. Publ. 5, 385–396. Murray, S.P., 1970. Bottom currents near the coast during Hurricane Camille. J. Geophys. Res. 75, 4579–4582. Naish, T., Kamp, P.J.J., 1997. Foraminiferal depth palaeoecology of Late Pliocene shelf sequences and systems tracts, Wanganui Basin, New Zealand. Sedim. Geol. 110, 237–255. Nebelsick, J.H., Kroh, A., 2002. The stormy path from life to death assemblages: the formation and preservation of mass accumulations of fossil sand dollars. Palaios 17, 378–393. Niederoda, A., Swift, D.J.P., Hopkins, T.S., Ma, Chen-Mean, 1984. Shoreface morphodynamics on wave-dominated coasts. Mar. Geol. 60, 331–354. Ohlenbusch, R., 1991. Post-glacial sequence stratigraphy and sedimentary development of the continental shelf off Townsville, central Great Barrier Reef province. Unpub. BSc Hons thesis, James Cook Univ., 94 pp.

R.M. Carter et al. / Marine Geology 267 (2009) 101–113 Orme, G.R., 1985. The sedimentological importance of Halimeda in the development of back reef lithofacies, northern Great Barrier Reef (Australia). Proc. 5th Int. Coral Reef Congr., Tahiti, v.5, 31–37. Orpin, A.R., Ridd, P.V., Stewart, L.K., 1999. Assessment of the relative importance of major sediment-transport mechanisms in the central Great Barrier Reef lagoon. Aust. J. Earth Sci. 46, 883–896. Parsons, J.D., Bush, J.M., Syvitski, P.M., 2001. Hyperpycnal plume formation from riverine outflows with small sediment concentrations. Sedimentology 48, 465–478. Puotinen, M.L., Done, T.J., Skelly, W.C., 1997. An Atlas of Tropical Cyclones in the Great Barrier Reef Region, 1969–1997. Tech. Rept. 19. CRC Reef Research Centre, Townsville. 201 pp. Saito, Y., Nishimura, A., Matsumoto, E., 1989. Transgressive sand sheet covering the shelf and the upper slope off Sendai, northeast Japan. Mar. Geol. 89, 245–258. Scoffin, T.P., 1993. The geological effects of hurricanes on coral reefs and the interpretation of storm deposits. Coral Reefs 12, 203–221. Scoffin, T.P., Tudhope, A.W., 1985. Sedimentary environments of the central region of the Great Barrier Reef of Australia. Coral Reefs 4, 81–93. Seguret, M., Moussine-Pouchkine, A., Gabaglia, G.R., Bouchette, F., 2001. Storm deposits and storm-generated coarse carbonate breccias on a pelagic outer shelf (South-East Basin, France). Sedimentology 48, 231–254. Stride, A.H., 1959. A linear pattern on the sea floor and its interpretation. J. Mar. Biol. Assn. U.K. 38, 315–318. Stride, A.H., 1963. Current-swept sea floors near the southern half of Great Britain. Quart. J. Geological Soc. Lond. 119, 175–199. Stride, A.H. (Ed.), 1982. Offshore Tidal Sands: Processes and Deposits. Chapman and Hall, London. 222 pp. Swift, D.J.P., Rice, D.R., 1984. Sand bodies on muddy shelves: a model for sedimentation in the western interior Cretaceous seaway, North America. In: Tillman, R.W., Siemers, C.T. (Eds.), Siliciclastic Sediments: Soc. Econ. Paleont. Mineral. Spec. Publ., 34, pp. 43–62.

113

Swift, D.J.P., Kofoed, J.W., Salisbury, F.P., Sears, P., 1972. Holocene evolution of the shelf surface, south and central Atlantic shelf of North America. In: Swift, D.J.P., et al. (Ed.), Shelf Sediment Transport: Process and Pattern. Dowden, Hutchinson and Ross, Stroudsburg Pa, pp. 499–574. Taylor, J., 1996. Sediment input to the Great Barrier Reef lagoon via river discharge — the Barron River. Great Barrier Reef: Terrigenous Sediment Flux and Human Impacts: In: Larcombe, P., et al. (Ed.), JCU Dept. Earth Sci. & CRC Reef Res., pp. 152–154. Testa, V., Bosence, D.W.J., 1999. Physical and biological controls on the formation of carbonate and siliclastic bedforms on the north-east Brazilian shelf. Sedimentology 46, 279–301. Walker, G.R., Reardon, G.F., 1986. Wind speeds in the Great Barrier Reef region from Cyclone Winifred and their effect on buildings. In: Dutton, I.M. (Ed.), Workshop on the Offshore Effects of Cyclone Winifred. GBR Marine Park Auth, Townsville, pp. 25–39. Wolanski, E., 1994. Physical Oceanographic Processes of the Great Barrier Reef. CRC Press. 194 pp. Wolanski, E., Ridd, P., 1990. Mixing and trapping in Australian coastal waters. In: Cheng, R.T. (Ed.), Residual Currents and Long-term Transport: Coast. Estuar. Studies, vol. 38, pp. 165–183. Wolanski, E., Spagnol, S., 2000. Pollution by mud of Great Barrier Reef coastal waters. J. Coastal Res. 4, 1151–1156. Woolfe, K.J., Larcombe, P., Stewart, L.K., 2000. Shelf sediments adjacent to the Herbert River delta, Great Barrier Reef. Australia. Aus. J. Earth Sci. 47, 301–308. Wright, L.D., Yang, Z.-S., Bornhold, B.D., Keller, G.H., Prior, D.B., Wiseman, W.J., 1986. Hyperpycnal plumes and plume fronts over the Huanghe (Yellow River) delta front. Geo-Mar. Lett. 6, 97–105. Young, I. R., 1985. Wave and wind data for John Brewer Reef. In: Det Norske Veritas, the Reef Link Floating Hotel Resort on John Brewer Reef in the Great Barrier Reef Marine Park, Environmental Impact Statement, Appendix 7.