The Marion Plateau and its carbonate platforms, located between 18°S and 23°S, seaward of the south-central Great Barrier Reef on the northeastern Australian continental margin, provides an ideal case study to address the causes, magnitudes, and effects of sea level change on continental margin sediments. This plateau is the most southerly of the northeast Australian marginal plateaus, forming a deeper extension of the Queensland continental shelf. The plateau is bounded by the Townsville Trough along its northern margin, by the Cato Trough along its eastern margin, and by the south-central Great Barrier Reef to the west (Fig. F1). The Marion Plateau is part of a slowly subsiding margin. It is believed that the plateau top remained exposed throughout much of the Paleogene and formed a gently northeastward-dipping, relatively smooth plateau surface (Pigram et al., 1992).

The western Coral Sea has been affected by two distinct tectonic events. The earlier event, Late Jurassic-Early Cretaceous in age, was responsible for the formation of the Queensland and Townsville basins that underlie the present-day bathymetric features of the Queensland and Townsville Troughs (Fig. F1). These basins formed as a result of oblique extension along preexisting Paleozoic structural trends (Struckmeyer and Symonds, 1997). The Queensland and Townsville Basins do not appear to have been affected by the later tectonism responsible for seafloor spreading in the Tasman and Coral Sea basins (Struckmeyer and Symonds, 1997). In the Late Cretaceous to early Paleocene, rifting in the Coral Sea Basin created numerous continental fragments that are now capped by carbonate platforms such as the Marion Plateau (Figs. F1, F2). This rifting in the Coral Sea was an extension of Late Cretaceous (80 Ma) seafloor spreading in the Tasman Basin, which extended to the north to form the Cato Trough and the Coral Sea Basin by 64 Ma (Weissel and Hayes, 1971; Hayes and Ringis, 1973; Shaw, 1978). Spreading is believed to have ceased along the length of this system by the earliest Eocene (52 Ma) (Gaina et al., 1999). Thus, the main physical elements of the western Coral Sea were in place by the early Tertiary (Davies et al., 1989). Although the exact structural style and development history of the rift system are still not completely understood, it is clear that the Late Jurassic-Early Cretaceous rifting event controlled the gross architecture of the margin and the shape and distribution of the high-standing structural elements on which the carbonate platforms are located. The Marion Plateau is an internally largely undeformed basement block with structural elements existing only on its margins (Fig. F3). Basement along the northern margin of the plateau consists of gently dipping ramps that gradually deepen toward the Townsville Trough until a fault is encountered. Normal extensional faults along this northern margin are restricted to the edge of the plateau (Symonds et al., 1988). The eastern margin of the plateau is free of major structural offsets (Mutter and Karner, 1980) and the slope is apparently simple and continuous. Faults along this margin are steeply dipping to vertical from the margin of the plateau into the Cato Trough. A southeasterly plunging, gently arched basement high forms the southern part of the plateau. The top of the arch is unstructured, and faults are confined to the flanks of the arch (Pigram, 1993).

The tectonic histories of the Marion and Queensland Plateaus are well constrained by Ocean Drilling Program (ODP) Leg 133 drill holes and extensive multichannel seismic data. Subsidence curves for these plateaus have been produced using both benthic foraminifers (Katz and Miller, 1993) and geohistory modeling (Fig. F4) (Müller et al., 2000). Geohistory models were calculated using integrated geophysical logs, biostratigraphic and lithologic information, and seismic reflection data (Müller et al., 2000). These models predict post-9-Ma subsidence of 1300 Trough and 650 subsidence of 500 northern margin of the Marion Plateau (Fig. F4) (Müller et al., 2000). Although the Marion and Queensland Plateaus are located on a passive margin, ~1000 km south of the Pacific/Australian plate boundary, geohistory models record a greater amount of post-9-Ma subsidence than simple elastic models can predict. This subsidence occurred in pulses between 9 and 5 Ma on both plateaus. It is difficult to account for this observed subsidence either by means of thrust loading in Papua New Guinea or by a combination of the latter and in-plane stresses originating from collision along the Pacific/Australian plate boundary (Müller et al., 2000). Despite the occurrence of post-9-Ma subsidence events on the Marion Plateau, the observed rates of tectonic subsidence are much slower than those of third-order sea level changes and can thus be differentiated from glacial eustasy. In addition, because of the methodology proposed here to investigate sea level change, any unaccounted subsidence should be similar for all sites. Although we will attempt to quantify in detail the additional water depth added to all sites as a result of tectonic subsidence, this increase is likely to be <10 m between Zones N12 and N14 and thus will not greatly affect our attempts to quantify sea level variations.

Stratigraphies for the Marion Plateau were obtained during Leg 133 (Figs. F1, F5) (Davies, McKenzie, Palmer-Julson, et al., 1991), and these data supplement extensive seismic surveys over the plateau. Initiation of shallow marine carbonate sedimentation on the Marion Plateau began during the latest Paleogene, as the sea transgressed across the plateau basement (Megasequence A) (Davies, McKenzie, Palmer-Julson, et al., 1991). These first sediments over basement are primarily siliciclastic, with temperate-water carbonates occurring in the eastern part of the sequence.

Sedimentary facies recovered during Leg 133 and correlation to seismic profiles indicate that tropical-subtropical reef development was initiated on the Marion Plateau in the early Miocene, and by the middle Miocene, there was extensive reef growth on the plateau. These reefal sediments are part of Megasequence B and include the aggrading and prograding MP2 carbonate platform (Fig. F5). In the late middle Miocene, carbonate bank productivity rapidly diminished on the Marion Plateau, as shown by a reduced fine-grained, bank-derived component in slope sediments. This decline was primarily the result of subaerial exposure resulting from a sea level regression, which caused the demise of the MP2 platform. During the low sea level interval between 11 and 7 Ma, the MP3 platform was initiated on the eastern side of the Marion Plateau. MP2 did not reinitiate even after being reflooded during subsequent sea level increases. During the development of MP3 the western two-thirds of the Marion Plateau was exposed, forming a broad, low-relief karstic surface. Unlike MP2, MP3 was not completely drowned, but is now restricted to the area of Saumarez Reef. Partial drowning of MP3 may have resulted from sea level variations in conjunction with reduced sea surface temperatures (SSTs) (Isern et al., 1996) and greatly increased platform subsidence rates (Müller et al., 2000). Carbonate production from the Pliocene to Holocene never again achieved the areal extent of the Neogene. Instead, hemipelagic drift sediments have since dominated sedimentation on the Marion Plateau.

Measuring the amplitude and timing of eustatic sea level fluctuations is essential both for the establishment of an accurate eustatic sea level curve for the Phanerozoic and for the accurate interpretation of sediment sequences on continental margins. Defining the amplitude of the eustatic sea level curve remains one of the major challenges in sea level research (COSOD II, 1987; Sahagian and Watts, 1991; JOIDES Planning Committee, 1996). Several attempts have been made to determine the amplitude of glacioeustatic fluctuations, including passive-margin sequence stratigraphy (Vail et al., 1977; Vail and Hardenbol, 1979; Haq et al., 1987), modeling of sedimentary depositional regimes (Watts and Thorne, 1984), calibration of the oxygen-isotope curve (Major and Matthews, 1983; Miller et al., 1987; Williams, 1988), and analysis of the depositional history of carbonate sediments on atolls (Schlanger and Premoli Silva, 1986; Halley and Ludwig, 1987; Moore et al., 1987; Lincoln and Schlanger, 1987, 1991). These analyses yield a wide range of results that are often in agreement with regard to the timing of sea level events. However, there are significant differences between the different independent data sets in regard to estimates of the magnitude of sea level fluctuations.

The inability to extract tectonic subsidence effects from relative sea level signatures has hindered the quantification of eustatic sea level variations in many areas. However, a sea level shift that occurs between two sites of equal tectonic subsidence will provide an accurate record of the magnitude of eustatic change. For estimating sea level between sites on the Marion Plateau, it is necessary to demonstrate that there is no differential subsidence along the drilling transect. Two lines of evidence support this assertion. First, the Marion Plateau is not structurally compartmentalized and therefore behaves as a single structural entity (Symonds et al., 1988). Seismic lines between Leg 194 sites show that no structural elements such as faults are present that could cause the sites to have differential relative subsidence. Second, because the Marion Plateau basement surface is planated with minimal dip to the northeast, depths to basement surface contours can be considered isosubsidence lines (Fig. F3). The eight sites drilled are all near the 1-s basement contour. Thus, there is minimal basement gradient between sites and negligible differential subsidence.

Calibration of eustatic sea level variations can only be realistically estimated on slowly subsiding, structurally well-understood margins, where an accurate tectonic subsidence history can be established, and where sites of equal tectonic subsidence that have both the highstand and the lowstand history preserved, can be located. For the predicted middle Miocene Marion Plateau subsidence rates, the increase in water depth added to both sites as a result of tectonic subsidence is an order of magnitude less than that resulting from eustatic sea level change over the same interval. This additional depth is <10 m over the 1-m.y. interval of middle Miocene eustatic sea level fall, which is within the error of paleowater depth estimates in these sediments.

Next Section | Table of Contents