)
can be related to depth (z) using an exponential function and a
porosity decay parameter with depth (k) (Athy, 1930) such thatDrilling at all Leg 194 primary sites and two alternate sites penetrated and sampled a Neogene sedimentary record that can be used to calibrate the sequence stratigraphic architecture of the Marion Plateau megasequences. The recovered sediments calibrate the regional seismic stratigraphy by identifying the lithologic signature and providing a chronostratigraphic framework and thus constraining the carbonate platform history and the magnitude and timing of sea level changes on the Marion Plateau.
In order to correlate core and log information accurately with the seismic data, a time-depth curve was constructed for each site (Fig. F11). These curves were obtained by integrating velocity data from sonic logs and shipboard velocity measurements and by calibrating these curves with results from check-shot surveys when they were available (Sites 1194, 1195, and 1196). For the remaining sites, tie points were defined that link prominent high-amplitude reflections to unique horizons in the cores. These horizons are usually hardgrounds or exposure surfaces at megasequence boundaries. Near the bottom of the drill holes, the basement reflection was used as a tie point to fix the critical lower part of the time-depth curve. Variations in the slopes of the time-depth curve reflect the high sonic velocities in the platform sediments (Sites 1193, 1196, and 1199), the medium velocities in the proximal slope sections (Sites 1194, 1197, and 1198) and the lowest velocities in the most distal locations (Sites 1192 and 1195) (Fig. F11). Time-depth curves, combined with the age model curves (Fig. F12), were used to assign chronostratigraphic datums to the seismic sequence boundaries at each site, which then can be compared across the transect (Fig. F13). The age models are mostly based on core catcher analyses and therefore have limited depth resolution. An overall good match between the ages of the sequence boundaries at the different sites, within limits of seismic and shipboard biostratigraphic resolution, confirms the concept of seismic stratigraphic correlations, in which seismic reflections have a chronostratigraphic significance. Age differences at boundaries between sites may be reduced once a more refined chronostratigraphy is established postcruise. The following sections discuss the major findings for each of the megasequences. The ages and depths of seismic megasequence boundaries, as well as lithologic unit boundaries at each site are plotted in Figure F13.
At all Leg 194 sites, Megasequence D, the youngest of the Marion Plateau megasequences, is composed of hemipelagic drift deposits. Megasequence D is only missing at the top of the southern platform edifice (Sites 1196 and 1999) where Megasequence C outcrops at the seafloor. The base of Megasequence D is a regional unconformity that is observed at all proximal platform sites (1193, 1197, and 1198) characterized by a hiatus lined with submarine hardgrounds, which provides an excellent link between seismic unconformities, biostratigraphic hiatuses, and nondepositional processes on the seafloor. At these sites, the age of the onset of Megasequence D sedimentation is dependent on the sites' locations relative to the depocenters of drift deposition. These ages are always younger than the ages of the base of Megasequence D in the distal areas where the sequence boundary is conformable. Sites 1192, 1194, and 1195 were drilled into the conformable succession at the base of Megasequence D, recovering the oldest sediments within this megasequence. At these sites, the base of Megasequence D was dated at 7.2 (Site 1192), 7.7 (Site 1194), and 7.2 Ma (Site 1195), respectively (Fig. F13A). Sites 1192 and 1195 are located in the most distal locations and have a well-constrained age model curve, so a basal age of ~7.2 Ma for Megasequence D is likely. Two smaller-scale seismic unconformities that represent current-controlled drift unconformities reflecting lateral shifts of depocenters and local hiatuses can be traced throughout the seismic grid. Their age range along the transect is 2.8-3.1 and 5.0-5.4 Ma, respectively (Fig. F13).
The C/D boundary at either side of the SMP coincides with the end of platform-derived shedding of neritic constituents onto the slopes (Fig. F13B). The age of ~7.2 Ma is thus assumed to reflect the end of the youngest phase of carbonate platform growth in this area and may date the drowning of the southern platform at least 1 m.y. earlier than originally predicted (Pigram, 1993).
Megasequence C is best developed near the southern carbonate edifice, where it records the platform-derived sedimentation of the youngest growth phase (Fig. F13B). Platform proximal sediments were drilled at Sites 1197 and 1198, and despite very low recovery, their age and platform-derived constituents indicate the existence of a late Miocene carbonate platform. Megasequence C at Site 1197 is characterized by relatively fine to medium grainstones (Unit II). At Site 1198, Megasequence C was deposited at the base of an escarpment and consists of coarse rudstone and floatstone (Subunits IIA and IIB) in the lower part of the section. Further from the platform, these proximal periplatform sediments interfinger with time-equivalent hemipelagic drift deposits. Based on the downlap of these Megasequence C drift deposits northwest of Site 1198 onto slope deposits of underlying Megasequence B, the southern platform edifice was thought to be entirely of late Miocene age. Drilling on the platform itself (Sites 1196 and 1199) revealed, however, that only the uppermost 100-180 m potentially consist of a late Miocene phase (lithologic Subunits IA-IC). Biostratigraphic dating below this depth gave a middle Miocene age. A low-amplitude, low-frequency reflection underneath the SMP top is a candidate for the boundary between these two carbonate platform-growth phases (Fig. F13B). This interval between the seafloor and ~110-130 mbsf dips gently to the southeast and coincides approximately with Megasequence C sediments on the downcurrent slope at Site 1197. The age models for the B/C boundary indicate ages of 10.5 (Site 1192), 11.0 (Site 1195), 11.5 (Site 1198), <11.8 Ma (Site 1194) and <11.3 (Site 1197). Giving more weight to the open-plateau section with the best seismic coherency and age control, an age of ~11.0 Ma can be postulated for the B/C boundary. This boundary correlates with the karstic top of the northern platform, placing an upper age limit on middle Miocene platform growth of NMP.
Within the NMP, at Site 1193, Megasequence B consists, at the base, of inclined slope deposits that can be seen on seismic data as inclined reflections underneath the platform (Fig. F13A). At Sites 1193 and 1194, these early Miocene slope deposits were composed mostly of fine silt-sized carbonate debris, which became mixed with a pelagic fraction in a periplatform environment (lithologic Units IV and V at both sites). The top of these periplatform sediments immediately underneath the platform section at Site 1193 could be confined to ~16 Ma. No age-diagnostic markers were found in the platform section above. Considering that the top of Megasequence B (platform top) was dated through seismic correlation along the B/C boundary to ~11.0 Ma, the northern platform growth may span ~5 m.y. in the middle Miocene. The top of this platform at Site 1193 shows signs of meteoric diagenesis. Topographically below the northern platform, Site 1194 penetrated an upper-slope section in a margin proximal position adjacent to NMP. The top of Megasequence B at Site 1194 is represented by a hardground surface, which caps an interval interpreted as a neritic outer ramp deposited in 30-50 m paleowater depth (lithologic Subunit IIIA). The base of this interval, marked by another hardground surface, can be clearly mapped on the seismic data. The interval overlies neritic upper-slope and hemipelagic sediments, indicating a shallowing-upward trend and thus a sea level lowering in the latest middle Miocene.
Megasequence B on the open plateau is mostly composed of a mixture of distal periplatform and pelagic components. In the most distal Site 1195, the top of Megasequence B coincides with a 20- to 30-m-thick interval rich in glauconite overlying distal periplatform sediments. The absence of neritic components indicates a reduction of neritic carbonate production at the end of the middle Miocene.
At Site 1198, situated adjacent to the escarpment in a proximal position to the southern platform, the top 70 m of Megasequence B (lithologic Subunit IIC) thickens toward the platform escarpment, documenting shedding of neritic material at the end of the middle Miocene. These periplatform sediments overlie hemipelagic deposits (lithologic Unit III). At Sites 1196 and 1199 on the platform, the >500-m-thick platform sequence below the presumed B/C boundary cannot be further subdivided because of the transparent seismic facies. It is likely that after its initiation onto a substrate of latest Oligocene age, this southern platform complex was a product of several growth phases in the early and middle Miocene. Interestingly, at both sites on the slope (Site 1197 and 1198), no early Miocene platform shedding is indicated, neither by debris in the periplatform sediments, nor geometrically by a thickening toward the platform. This observation might be related to the strong influence of currents from north to south that shape the geometry of the platform. The paleo platform shape can be estimated by tracking the weak reflection assumed to be the top of Megasequence B (equivalent to the top of the NMP) at a subsurface depth of ~100-130 mbsf. This reflection can be partly traced on the seismic data and displays a topography with a rim at the northwest, upcurrent escarpment and a gentle dip toward the southeastern, downcurrent slope at Site 1197. At Site 1197 the B/C boundary caps the coarsest interval at that site (lithologic Unit III) and forms a thick prograding slope unit that forms a downcore transition from steeper-dipping slope to almost flat hemipelagic deposits (lithologic Units III-V).
In the northern survey area, Megasequence A consists of a siliciclastic substrate that directly overlies acoustic basement. Megasequence A has a highly variable thickness, infilling small-scale basement irregularities. Only Sites 1193 and 1195 were characterized as having siliciclastic sediments directly overlaying basement. Thus, Megasequence A is not easily traced seismically and lithologically through all Leg 194 sites. At Sites 1195 and 1198, basement was overlain by a thin veneer of coarse carbonate deposits of undefined age. These sediments at Site 1198 can be seen on the seismic data as a thin onlapping unit directly overlying the basement reflection. Acoustic basement itself is highly variable, as shown by its different seismic signature across the study area. At Leg 194 sites, it consists of altered basalt flows and volcaniclastic breccias/conglomerates.
A primary objective of Leg 194 was to establish the magnitude of the major late middle Miocene sea level fall based on the expected stratigraphic relationships between the northern and the southern platforms. The SMP was thought to be nucleated on the distal slope sediments of the NMP during the maximum sea level lowstand. However, drilling results at Sites 1196 and 1199 showed that the southern platform is composed of ~120-160 m of late Miocene platform sediments overlaying a middle to early Miocene platform. This stratigraphic architecture makes estimates of sea level variations difficult using the original proposed methodology. Nevertheless, sequence stratigraphic and facies relationships from Sites 1193 and 1194 allow an estimate of the magnitude of the sea level fall to be made. The top of the platform at Site 1193 (35 mbsf) was deposited during the middle Miocene (~12-16 Ma) at an estimated paleowater depth of 10-50 m (Table T3) and represents the middle Miocene highstand. Lithologic Subunit IIIA at Site 1194 (117-154 mbsf), deposited during the latest middle Miocene (~12-14 Ma), represents an in situ shallow-water (30-50 m) interruption of a generally deeper water sedimentation. The base of Subunit IIIA today lies 145 m below the top of the northern platform at Site 1193, and seismic correlation between the two sites suggests that Subunit IIIA was deposited after the platform top was exposed (top panel of Fig. F14A). Consequently, these sequences provide a means to estimate the amount of sea level drop needed to move the shallow-water depositional environment from the top of the platform at Site 1193 to the base of Subunit IIIA at Site 1194. Sediment and water loading that have occurred since the end of the middle Miocene compacted earlier deposited sediments, and perhaps resulted in differential isostatic adjustments between the two sites, and thus affected the sequence geometry. These effects need to be removed in order to estimate the magnitude of the sea level drop.
First, the sediment
rebound is considered relative to a basement with infinite flexural strength
(Fig. F14A). As
sediment is removed, the thickness of the underlying sediments must be adjusted
to reflect the progressive decompaction of the sediment. The measure of
compaction is sediment porosity, which was measured at both sites. Porosity ()
can be related to depth (z) using an exponential function and a
porosity decay parameter with depth (k) (Athy, 1930) such that
(z)
= 0-kz.
(1)
A least-squares fit of
this function to Site 1194 porosity data yields 0
= 65% and k = 0.002/m (correlation coefficient = 0.91). This normal
compaction trend suggests hydrostatic fluid pressures were maintained during the
deposition and compaction process and that the equation above can be used to
simulate the compaction (and decompaction) of the sediment column. Removing
lithologic Units I, II, and IIIA (a total thickness of 154 m) induces an
expansion of 
e
= 56 m. Porosity at Site 1193 shows a general decrease with depth, except within
the platform sediments, where values are scattered, ranging from 10% to 45%,
reflecting the various degrees of cementation in the carbonate rocks. Ignoring
the porosity interval spanned by the cemented platform, the equivalent
least-squares fit for Site 1193 porosity data yields 0
= 61.2% and k = 0.001/m (correlation coefficient = 0.60).
Removing lithologic Units I and II (a total thickness of 35 m), expansion of the
sediment packages below the cemented carbonate platform amounts to e
= 7 m. Assuming infinite flexural strength of the lithosphere, the present-day
geometry between Sites 1193 and 1194 is maintained except for the decompaction
of underlying sediments. The reconstructed water depths W1 at Sites 1193 and
1194 without overlying sediment and e
are 376 and 472 m, respectively. The corrected geometric relief between the
platform and Site 1194 is thus 96 m. Given the paleowater depth estimates at
Sites 1193 and 1194 of 10-50 m and 30-50 m, respectively, the corresponding
range in the estimated eustatic fall is 56-116 m.
Second, local isostasy in response to sediment and water loading may have further affected the difference in elevation between the top of the platform at Site 1193 and the base of Subunit IIIA at Site 1194 (Fig. F14B). Isostatic correction is based on the density difference between the added water and sediment layers and an equivalent thickness of mantle at an assumed density of 3.3 g/cm3. The special case of complete local isostasy (zero flexural strength in basement) is calculated in a three-step procedure (Fig. F14B).
The present-day water depth in the absence of sediment can be calculated using
water
= Ssed(
m
- s)
/(m
- w),
(2)
where water
is the added water column thickness after the removal of the overburden
sediment, Ssed
is the thickness of the post-middle Miocene sediment load on top of the
NMP at Site 1193 (35 m) or on top of the base of Subunit IIIA at Site 1194 (154
m), and, m,
w,
and s
are the mantle, water, and average sediment densities, respectively (Table T3).
Present-day water depths, W0, at Sites 1193 and 1194 are 348 and 374 m,
respectively. Applying the above equation the calculated water
for Sites 1193 and 1194 are 25 m and 106 m, respectively, giving a reconstructed
present-day water depth in the absence of sediment, W2, of 373 and 480 m. From
these reconstructed water depths the previously calculated correction
e
is subtracted for the expansion of the underlying sediments. The corrected water
depths W3 for the late middle Miocene stratigraphic reference levels at Sites
1193 and 1194, therefore, are 367 and 424 m, respectively (top panel of Fig. F14B).
Next, the water depth at
the time of the middle Miocene highstand needs to be calculated. As the water
column is reduced, basement and overlying sediments are unloaded and so will
flexurally rebound. The following equation determines the sea level change (SL)
responsible for producing a given paleowater depth (PW) including the
isostatic adjustment of basement associated with the sea level change:
SL
= (W3 - PW)(m
- w)/m.
(3)
Considering the estimated
10-50 m paleowater depth at Site 1193 during the middle Miocene highstand, sea
level must be decreased by 219-246 (233 ± 14) m (the middle panel of Fig. F14B).
The water depth at Site 1194 at the time of the maximum highstand, W4, is given
by the reconstructed water depth W3 of 424 m minus the SL1
change of 219-246 m minus the isostatic adjustment (rebound) induced by the
change in sea level, which is
m2
= SL1w/(m
- w)
(4)
Evaluating gives a rebound range of 98-110 m. The corresponding highstand water depth, W4, at Site 1194 is thus 67-107 m.
Finally, the middle Miocene sea level fall can be calculated. At the time of the maximum lowstand, the paleowater depth at Site 1194, PW2, is estimated from paleontological assemblages to be 30-50 m. Using PW2 and the reconstructed water depth at Site 1194, W4, and adjusting for the isostatic rebound induced by the lowering of sea level, the required sea level fall is 12-53 m (32 ± 20 m; equation 3) (the lower panel of Fig. F14B).
In summary, an extremely conservative estimate of the late middle Miocene eustatic fall based on Leg 194 drilling, considering both the effects of sediment compaction and the possible range in flexural behavior of the lithosphere, is 12-116 m (64 ± 52 m). The assumption of local isostasy minimizes the estimated eustatic fall. The relatively small distance (~20 km) between Sites 1193 and 1194, the undisturbed and consistently dipping sediments, as well as the horizontal basement geometry between the sites as seen on seismic data (Figs. F7, F14A), suggest that no significant differential subsidence occurred between the sites. Therefore, given the much more reasonable scenario of finite flexural strength of the lithosphere, our best estimate of the eustatic fall is 86 ± 30 m. Platform erosion at Site 1193 and overall tectonic subsidence during the sea level lowering were not considered. Both effects are much smaller than the error margin of the above calculations and would increase the sea level fall. It is also possible that a record of the lowest sea level was not preserved, cored, or observed in Subunit IIIA at Site 1194, which would also increase the magnitude of the eustatic fall.
Two transects were drilled across the Miocene carbonate platforms of the Marion Plateau during ODP Leg 194. The northern transect includes the platform (Site 1193), proximal slope (Site 1194), and distal slope (Site 1195) sites composed of a mixed carbonate-siliciclastic depositional system adjacent to the Australian continent. The NMP drilled at Site 1193 represents the second phase of carbonate production that occurred during the Miocene on the Marion Plateau (Pigram et al., 1992; Pigram, 1993). The southern transect includes the platform (Sites 1196 and 1199), the leeward upcurrent slope (Site 1198), and the windward downcurrent slope (Site 1197) sites of an isolated carbonate bank ~65 km to the southeast of the NMP (Fig. F2). The SMP was labeled MP3 by previous investigators (Pigram et al., 1992; Pigram, 1993) because it was thought to be entirely composed of the third phase of carbonate platform sedimentation occurring on the Marion Plateau in the late Miocene. A primary discovery made during ODP Leg 194 is that this southern platform is a compound edifice that has undergone three main phases of carbonate accumulation.
The following sections reconstruct the evolution of these two platforms on the basis of lithologies recovered during ODP Leg 194 (Figs. F9, F15).
Sedimentation on the Marion Plateau basement initiated on a topographically irregular surface consisting of volcanic and volcaniclastic rocks (Fig. F15A). The oldest sediments at Site 1195 include inner platform carbonates of possible Eocene age that could be transported clasts (not represented in Fig. F15A). During the early Miocene (23.8-16.4 Ma) siliciclastic estuarine sediments containing large oyster shells and larger benthic foraminifers accumulated at Site 1193. A subsequent sea level rise resulted in the deposition of a thick succession of mixed carbonate/siliciclastic material bearing glauconite and phosphate, which was also recovered at Site 1195. Between these two locations, a basement high at Site 1194 apparently formed an island in this time interval. Around 18 Ma (late early Miocene), silt-sized skeletal packstone/grainstone with quartz and terrigenous clays were deposited throughout the northern transect. They likely represent the mid- to distal slope sediments of a platform that was located to the north or north-northwest of the study area. This sedimentary package formed a gentle, eastward-facing ramp. Shallow-water facies, including skeletal rudstone/floatstone with bryozoans and larger benthic foraminifers first appeared in the northern transect in the early middle Miocene (~16 Ma) (Fig. F15B). Silt-sized skeletal wackestone/grainstone with clay and planktonic foraminifers accumulated downslope from these facies at Sites 1192, 1194, and 1195. During the middle Miocene (16.4-11.2 Ma), a carbonate platform consisting of shallow-water grainstone/floatstone with bryozoan fragments and larger benthic foraminifers grew on the previous ramp morphology (Site 1193; Subunit IIIA; Fig. F15B). Sediment shedding was directed eastward, as shown by seismic data and the presence of upper-slope skeletal packstones at Site 1194 (Subunit IIIB). Distal slope sediments were retrieved from Site 1195. However, these sediments may not originate from the NMP because the corresponding lithologic horizon at Site 1192 (Subunit III) consists of hemipelagic packstones with glauconite, clay, and no neritic components. During this time interval, the NMP had a flat-top morphology with a steep upper-slope between Site 1193 and 1194 (Fig. F15B).
An important sea level fall, possibly related to a major ice-building phase in Antarctica, occurred in the late middle Miocene. The Marion Plateau carbonate platforms were exposed and karstified (Fig. F15C). A 30-m-thick succession of skeletal packstone/floatstone dominated by bryozoans (Subunit IIIA; Site 1194) was deposited seaward of NMP between ~13 and ~11 Ma (Fig. F15C), likely corresponding to a lowstand carbonate ramp. Paleowater depth estimates of these sediments were used to calculate the amplitude of this major sea level change (see "Sea Level Magnitude Recorded by Sediment Sequences on the Marion Plateau").
During the ensuing late Miocene sea level rise, the NMP was flooded (Fig. F15D), but platform growth did not reinitiate, as was observed on the SMP. A significant ferromanganese hardground was formed during this sea level rise on the top of the previously deposited lowstand platform sediments at Site 1194. From the late Miocene to the Pleistocene, a hemipelagic sediment drift consisting of greenish gray planktonic foraminiferal mudstone/wackestone with clay onlapped the middle Miocene slope lowstand platform and drowned the NMP (Fig. F15E).
In this area, the irregular basement topography likely includes numerous basalt lava flows that formed an escarpment in the vicinity of Site 1198. The earliest deposits encountered are phosphate sands mixed with siliciclastic material of latest Oligocene age (24.6-24.2 Ma) recovered from Site 1196 (Fig. F15A). It is not clear whether these sediments were produced in situ, transported laterally, or reworked from older rocks. The presence of oysters in these beds indicates a water depth <30 m. During the early Miocene (23.8-16.4 Ma) a carbonate platform developed in a topographic low at Site 1196 (Fig. F15B) (SMP; Units IV and III), building a thick series of aggrading rhodalgal carbonates. Contrary to earlier hypotheses (Pigram et al., 1992; Pigram, 1993), this platform did not nucleate on the slope sediments from the NMP, but clearly predates them and thus likely corresponds to the earliest phase of carbonate production seismically imaged along the eastern margin of the Marion Plateau (Pigram, 1993). Facies homogeneity suggests that accommodation space remained more or less constant during this time interval. Numerous exposure surfaces within Subunit IIIC at Site 1196 (variegated dolostone) indicate that water depth was less than the amplitude of the early Miocene sea level fluctuations. Off-platform shedding was restricted, but included a thin package of coarse skeletal grainstone with benthic foraminifers at Site 1197 and stable deep-euphotic accumulations of rhodalgal-foraminiferal floatstone at Site 1198. The oldest phase of platform growth ceased near the early/middle Miocene boundary, corresponding to the first occurrence of shallow-water carbonate production on the northern transect.
The next phase of carbonate production initiated in the early middle Miocene (~16.4 Ma) (Site 1196; Unit II) (Fig. F15B). Red algal-dominated carbonate buildups formed during the renewed transgression, exporting detritus to the slopes of both Sites 1197 and 1198. The occurrence of reefs and possibly shoals or islands on the platform top is inferred from the variable depth in the sedimentary section at which these facies were recovered at Sites 1196 and 1199 (Figs. F9B, F15B). Shedding of silt-sized skeletal carbonate slope deposits was asymmetric, with greater accumulation at Site 1197, probably related to the prevailing current directions. In the middle Miocene (~15.2-13.3 Ma), shallow-water carbonates accumulated in a ~150-m thick succession of fine grainstone rich in coralline algae, miliolid foraminifers, and seagrass remains. This facies filled the previously created irregular topography on the platform top in a shallow-water setting implying rapid and still unexplained creation of accommodating space. Vertically accreting reef-style growth was locally maintained, particularly on the northeast edge of the platform, where a buildup apparently existed as indicated by multichannel seismic data. At this stage, the platform was asymmetric with a flat-topped, rimmed western margin and an eastern margin showing a distally steepened ramp geometry (Figs. F13B, F15B). Platform shedding predominantly occurred in a downcurrent, downramp direction toward Site 1197 at the southeast of the section. Sediment shedding may also have taken place on the northwestern slope. The final pulse of the middle Miocene growth phase in this area was characterized by an increase in accommodation space expressed by the facies change from shallow (<20 m) seagrass-rich beds to rhodalgal floatstone (Subunit ID; Sites 1196 and 1199) (Fig. F15B) that suggest 60- to 100-m water depths. Reef growth speculatively continued on the northwestern side of the platform, where it acted as a marginal barrier system. Sediment shedding at this time was almost exclusively toward Site 1197 in the southeast.
Lithologic data from Site 1199 suggest that sea level retreat at the onset of the late Miocene was rapid because no evidence of shallowing could be observed in the depositional environment (60-100 m) of the rhodalgal floatstone forming Subunit ID, which occurs directly below an exposure horizon (Fig. F15C). Sediments produced during this regression may have also eroded and been exported downslope.
During the late Miocene (11.2-5.3 Ma), the platform was again flooded and carbonate production resumed in this area (Fig. F15D). The northwestern edge of the edifice was again occupied by a potential reef buildup. Sediments shed toward the west were possibly swept away by the strong currents flowing parallel to the platform flanks. In the platform center, this sedimentary pulse began with a phase of reef construction (Subunit IC) followed by a shallowing-upward succession consisting of rhodalgal floatstone (Subunits IB and IA; Fig. F15D). The latter accumulated in water depths that did not exceed 100 m. The uppermost sediments of Subunit IA include intertidal (beachrock) cements indicating extremely reduced accommodation space near the end of this depositional episode. Sediments were predominantly exported toward the southeast. The geometry of the southeastern margin evolved from a distally steepened to a more homoclinal ramp. The platform was exposed again during the latest Miocene as shown by the occurrence of a karst surface recovered at Site 1196 and the development of a pedogenic profile at Site 1199.
The southern platform was drowned during the latest Miocene or early Pliocene and thereafter was swept by bottom currents that contributed to the formation of a well-developed hardground surface with a laminated ferromanganese crust (Fig. F15E; also, see the volume cover photograph). The adjacent slope areas of Sites 1197 and 1198 were the locus of extensive hemipelagic drift deposition dominated by pelagic foraminifer remains, which filled the topographic lows adjacent to the platform leading to the modern bathymetrically nearly uniform seafloor (Fig. F15E).
Lithologic and biostratigraphic data obtained during Leg 194 from the Neogene carbonate platforms of the Marion Plateau reveal that platform architecture was controlled by a series of complexly related factors including sea level change, bottom-current action, and biological assemblages.
The NMP platform initiated in the late early Miocene on low-angle, southward-dipping clinoforms. What is most striking in the evolution of this platform is the rapid onset of carbonate growth in a distal periplatform setting (Fig. F15B). The initiation of neritic sedimentation over midslope deposits suggests a major sea level fall at that time. The nearly flat surface at the base of the platform further shows that the onset of carbonate production was nearly simultaneous along the area seismically imaged near Site 1193. During the subsequent growth, the platform was predominantly aggradational producing a steep-sided margin (Fig. F15B). Carbonate production ceased when sea level dropped, exposing the platform at the end of the middle Miocene. This sea level drop shifted neritic sediment production onto the former slope, producing a lowstand carbonate ramp (Fig. F15C). The high-relief margin and insufficient water depth inhibited the establishment and growth of a "healthy" platform. Consequently, the subsequent sea level rise was able to outpace neritic sediment production and a hardground developed on this lowstand edifice. Why the younger carbonate production phases observed on the southern platform did not initiate on the NMP is still unclear.
This compound carbonate edifice was initiated during the earliest Miocene and became inactive during the latest Miocene or early Pliocene. It apparently nucleated in a topographic low, questioning the paradigm that carbonate-platform nucleation requires positive antecedent topography. During its growth, this platform developed an asymmetrical architecture with an escarpment-like margin on the northwestern side and impressive high-angle prograding clinoforms on the southeastern margin. Little sediment was transported off the escarpment to the northwestern slope despite significant platform aggradation during the early and middle Miocene (Fig. F15B).
The steep-sided geometry of both the NMP and SMP originally suggested that they were constructed by tropical to subtropical faunal assemblages including corals (Figs. F7, F8, F13). In contrast, cores retrieved during Leg 194 document a cool subtropical faunal assemblage consisting primarily of red algae, bryozoans, and larger benthic foraminifers. These calcite-dominated biogenic sediments have a lower diagenetic potential than their aragonite-dominated counterparts in the tropical realm. They can therefore be reworked more easily as they undergo less cementation. In addition, the fragmentation of these sediments leads to the formation of silt- to fine sand-sized particles rather than mud which is typical for aragonite-dominated systems.
The carbonate platform architecture observed on the Marion Plateau can be best explained by the dominance of currents. Exposure to predominant wind direction results in different platform margin geometries (e.g., leeward and windward margins of the Great Bahama Bank). The Marion Plateau example demonstrates that, similarly, exposure to bottom currents also results in asymmetric platform architecture (Figs. F8, F13). Seafloor currents influence sedimentation on the periplatform apron as they inhibit sedimentation in the upcurrent position and form wide low-angle clinoforms in a downcurrent position.
The results of Leg 194 suggest that sea level in conjunction with current-dominated sedimentation is a possible cause for rapid carbonate platform growth. The fact that the observed cool subtropical faunal assemblage produces platform geometries reminiscent of tropical carbonates further indicates that physical parameters may be as or more important for producing a given platform architecture than biological ones.
Acoustic basement was penetrated at five sites during Leg 194 drilling (Sites 1193, 1194, 1195, 1197, and 1198). Prior to Leg 194 drilling, no direct information existed concerning the basement composition of the Marion Plateau.
Leg 133 drilling on the Queensland Plateau recovered altered and deformed metasedimentary and metavolcanic rocks that comprised acoustic basement (Davies, McKenzie, Palmer-Julson, et al., 1991). These metasedimentary rocks consisted of detrital quartzofeldspathic assemblages and igneous rock fragments. Low-grade quartz-muscovite-feldspar schists were also recovered. Recovered igneous rocks were determined to be altered granodioritic to tonalitic in composition and dominated by feldspars and quartz (Feary et al., 1993). Those assemblages were considered a straightforward continuation of the onshore Hodgkinson Formation, predominantly Devonian in age, consisting of thick, monotonous cleaved graywacke, siltstone, shale, and slate successions that in turn were cut by late Paleozoic-early Mesozoic dike swarms.
Basement rocks recovered during Leg 194 differ greatly from those drilled on the Queensland Plateau, although prior to drilling they were assumed to be similar in composition. Rather than metasedimentary rocks, highly altered lava flows and volcaniclastics were collected (Figs. F16, F17, F18, F19). The lack of deformation suggests that these volcanics may have been emplaced during the Late Cretaceous-Paleocene rifting of northeastern Australia from the Papuan Plateau in the north to the Lord Howe Rise in the south.
Planation of basement supposedly occurred during subaerial exposure in the Mesozoic and Paleogene. Site survey multichannel seismic reflection data suggest that the top of basement of the Marion Plateau is a topographic surface with tens to hundreds of meters of relief (Fig. F3) and often a major erosional unconformity. In particular, a north-south-trending, eastward-facing ramp exists toward the eastern edge of the plateau. The ramp has a maximum relief of 225 m and a dip of ~1°. Initiation of the oldest early Miocene reef/platform systems (and perhaps earlier systems) was focused along the eastern edge of this ramp (Pigram, 1993). Given the progression of the shallow-water early reef systems to the present water depths of the drilling sites (304-420 m), it is clear that postrift thermal subsidence remains the controlling factor on the long-term accommodation. The postrift subsidence of the region occurs even though unambiguous rift structures (e.g., high-angle normal faults, rotated fault blocks, synrift wedge packages) are not observed across the plateau.
Site 1194 was drilled on the flank of a topographic basement high (Fig. F16). Recovered volcanic rocks consist of altered amygdaloidal olivine basalts, in which the olivines are large phenocrysts within a fine matrix of pyroxenes and plagioclase laths (Fig. F16). Many of the olivines have been altered to iddingsite and opaques, probably magnetite. Occasional veins are filled with feldspars that are zoned from fine to coarse, whereas the vugs are infilled with zeolites, most likely natrolite.
Site 1197 was located in a similar setting to Site 1194 (Fig. F17). Recovered basal units of this site are polymictite volcanic breccias (tuffs) deposited in a nonmarine paleoenvironment. Glass shards are also present and vugs are filled with feldspars. At the base of the breccias is an olivine basalt (Fig. F17) showing similar characteristics as the basalts from Site 1194.
Site 1193, a volcanic flow, is highly altered even compared with Sites 1194 and 1197 (Fig. F18). Vug infilling is natrolite, characterized by fine radiating fibers (Fig. F18). Crystal outlines are well represented under normal light.
The recovered basement rocks at Site 1198, although again highly altered, are different from the basalts at the other sites, being composed principally of plagioclase (albite?) lathes. Quartz is observed to infill veins (Fig. F19).
The paleomagnetic response of basement is characterized by both relatively high intensity and consistent normal and reversed inclinations (Figs. F16, F17, F18, F19). In general, the inclination polarity is almost exactly 180° apart, implying the existence of a clean, primary thermal remanent magnetization. In the upper part of the recovered basement section, the inclination has a greater variation, implying that a magnetic overprint exists. This leads to the possibility that the inclinations, when the corresponding paleomagnetic pole is compared with the Australian apparent polar wander path, may provide age estimates for both the emplacement of the basalts and the timing of low-temperature alteration.
One of the main objectives of Leg 194 was to study fluid circulation in the Marion Plateau carbonate platforms using the sediments and pore fluids recovered from the Leg 194 drill cores. The dolomitization found in both the north and south platforms is itself indirect evidence for past fluid circulation. Dolomite formation on a large scale requires fluid flow to deliver the required magnesium to the precursor calcium carbonate sediments. But when and how fluids may have circulated and the nature of the fluids, whether they were normal seawater or hypo- or hypersaline, are open questions. It was also thought that the present-day pore waters might hold evidence of past or even continuing fluid movement. In that regard, the pore fluid sampling program on Leg 194 has yielded intriguing results. Although sampling of pore waters from within the southern platform was not possible, samples taken from sediments above and below the adjacent periplatform facies provide clear evidence that seawater continues to circulate through these sediments even though they are overlain by ~200 m of hemipelagic deposits. By inference, seawater is likely to be circulating through the southern platform at present. Similarly, the pore water samples from directly above and below the carbonate platform facies of the NMP at Site 1193 are also close to seawater in composition, suggesting seawater circulation. Site 1198 drilled through seismic Megasequences D, C, and B to basement 5 km northwest of the margin of the SMP (Fig. F20). In seismic Megasequence D, which is equivalent to lithologic Unit I, 24 samples were taken at ~10-m intervals to a depth of 196.6 mbsf, the base of lithologic Unit I. From 197 to 350 mbsf, poor recovery of the unconsolidated sediments of lithologic Subunits IIA and IIB (seismic Megasequence C and the upper ~60 m of Megasequence B) precluded pore water sampling. Sampling resumed in Megasequence B at 350 mbsf, with samples taken at ~10-m intervals to a depth of 505.0 mbsf, just above basement.
For most of the pore water constituents, nearly symmetrical, arcuate pore water profiles are found in the upper 200 mbsf. From essentially seawater values near the sediment surface, concentrations in the interval from 0 to 100 mbsf either increase or decrease depending on the ion measured. In the interval from 100 to 200 mbsf, the trends of the upper 100 mbsf reverse and concentrations return to values close to those of normal seawater. This pattern is found for alkalinity, sulfate, ammonium, strontium, potassium, and magnesium.
Two profiles, sulfate and strontium, are illustrated in Figure F20. These two constituents were chosen because they are affected by completely independent processes. The two profiles are shown plotted on top of the seismic section that crosses the platform margin and continues across the adjacent periplatform sediments now buried beneath the hemipelagic sediments of seismic Megasequence D. Note that the data points all lie on the line showing the location of Site 1198; the x-axis is concentration, not distance. Sulfate concentration initially decreases as a result of bacterial sulfate reduction. At ~100 mbsf, however, sulfate concentration increases, returning to ~29 mM at the transition from hemipelagic sediments of seismic Megasequence D to Megasequence C. A cessation of sulfate reduction in the lower part of Megasequence D is an unlikely cause for the increase in sulfate concentration because there is no decrease in the organic carbon content within this unit. At the base of seismic Megasequence C, the sulfate concentration is also ~29 mM and thereafter decreases with depth.
The strontium concentration profile found in the upper ~100 mbsf is typical of pelagic or hemipelagic, carbonate-rich sediments. Strontium is added to the pore fluids by carbonate dissolution and recrystallization. Usually, this process continues for many hundreds of meters and high strontium concentrations (500 to 1000 µM) are a standard feature of sediment pore waters below 100-200 mbsf. At Site 1198, however, the concentration of strontium in the lower portion of seismic Megasequence D reverses the trend seen in the upper ~100 mbsf and is close to the seawater value at the contact with seismic Megasequence C. The strontium concentration is also low in the uppermost sample from seismic Megasequence B and then increases with depth to the bottom of the sampled interval. As for sulfate, there is no sedimentological reason for the reversal in trend. The carbonate content of seismic Megasequence D is relatively constant and in fact increases in Megasequence C.
The strontium profile is in many ways similar to sulfate, although the chemical reaction controlling strontium concentration is independent of that which affects sulfate. The pore water profiles of both constituents in seismic Megasequence D suggest diffusion-reaction control without fluid flow. The shape of the upper portion of the pore water profiles is typical for pore waters in many parts of the ocean, but the changes in concentration seen in the lower half of lithologic Unit I are unusual. They are most easily explained by relatively constant rates of reaction in the sediments, with diffusion acting upon both the upper and lower bounds of this sediment package. The fact that the concentrations at the lower boundary of Unit I are close to seawater values implies that the fluid that fixes the concentration at the lower boundary condition is also close to seawater in composition. The near-seawater concentration of sulfate and strontium (and other constituents) at the base of seismic Megasequence D strongly suggests active circulation of seawater through the sediments of seismic Megasequence C, between 200 and 350 mbsf. Neither the mechanism nor direction of fluid flow can be determined from the available data. Based on the seismic profiles of the Marion Plateau, sequence C is not exposed on the seafloor and thus has no direct connection to seawater. The unit is in contact with the SMP; thus a hydraulic connection between the platform and the periplatform sediments of seismic Megasequence C seems likely.
Figure F21 is similar to Figure F20 but shows the concentrations of sulfate and strontium for Site 1193, drilled through the NMP. The ~200-m-thick sequence of platform facies are overlain by a thin cover of hemipelagic sediments of seismic Megasequence D and underlain by several hundred meters of hemipelagic sediments of seismic Megasequence B. Both sulfate and strontium concentration profiles of the upper 40 mbsf of Megasequence D faintly suggest the curved profiles found in Site 1198. More importantly, the pore fluids from the surface down to the base of the NMP at 230 mbsf are close to seawater in composition. The similarity was initially interpreted to be the result of a near lack of chemical reaction in highly stabilized lithologies, low-magnesium calcite and dolomite, of the platform. Although this explanation remains a possibility, a second possibility also exists. Chemical reaction may be continuing in the platform, but the resulting changes to pore fluid chemistry could be swept away by active fluid circulation. Chemical reactions certainly continue below the platform facies, in Megasequence B, where concentrations of sulfate decrease and concentrations of strontium increase with depth. The evidence for fluid circulation at Site 1193 is perhaps not as strong as found at Site 1198. Fluid circulation is, however, a viable explanation for the observed pore water profiles.
