Background

Continental shelves are the main transport pathway of sediments from land to sea, playing an important role in the earth’s surface system. The sedimentary environment of low-latitude mixed siliciclastic–carbonate continental shelves is strongly influenced by sea-level change and fluvial processes (Dunbar and Dickens 2003; Schlager et al. 1994; Webster et al. 2012). Detailed research on sedimentary environments of continental shelves improves our understanding of monsoonal intensity and sea-level variability (Bourget et al. 2012; Yokoyama et al. 2000).

The Bonaparte Gulf, northwestern Australia, is a broad continental shelf with a water depth shallower than ~200 m (Bourget et al. 2013, 2014). The area of the shelf has varied with relative sea-level change during the late Quaternary (Yokoyama et al. 2001a). There are several carbonate terraces and platforms such as Londonderry, Sahul, and van Diemen Rise in the Bonaparte Gulf, which were exposed during sea-level lowstands (Fig. 1; Bourget et al. 2013, 2014; van Andel et al. 1967). Carbonate terraces and platforms contain incisions at depths exceeding 100 m, connecting the Timor Sea and Bonaparte Basin (Courgeon et al. 2016). Siliciclastic sediments, transported by monsoon-influenced fluvial activity, dominate the near-shore region (Fig. 1; De Deckker et al. 2014; Nicholas et al. 2014). Shallow sediment cores show that past relative sea level fell to ~ −120 m in the Bonaparte Gulf during the Last Glacial Maximum (LGM; ~20 ka) (De Deckker and Yokoyama 2009; Ishiwa et al. 2016; Yokoyama et al. 2000, 2001b; Nakada et al. 2016).

Fig. 1
figure 1

Bonaparte Gulf, northwestern Australia. a The location of core site KH11-1-PC01 (white square), bathymetry and named seabed features are referred to in the text. Bathymetric data is from Whiteway (2009). b Distribution of surface sediments, modified from van Andel et al. (1967). 1 recent near-shore sediments, 2 calcareous silty clays of Bonaparte Depression, 3 shelf-edge foraminiferal calcilutites and calcarenites, 4 Timor Trough silty clay and calcilutites, 5 relict transgressive quartzose calcarenites, 6 relict transgressive calcarenites, 7 bank facies. Legends refer to van Andel et al. (1967)

Full size image

Today, northwestern Australia experiences a semi-arid climate. Southeast winds prevail in the Austral summer and northwesterly winds in the Austral winter. There is a strong rainfall seasonality during the Austral summer (De Deckker et al. 2014; Gallagher et al. 2014a). This is the Australian monsoon, the intensity of which has varied with the movement of the Intertropical Convergence Zone (ITCZ) over the past 30,000 years (cf., Ding et al. 2013; Kuhnt et al. 2015; Mohtadi et al. 2011).

This mixed siliciclastic–carbonate and shallow continental shelf is influenced by monsoonal intensity and sea-level change (Gallagher et al. 2014b). Petroleum exploration and paleoclimatic reconstructions have driven interest in the long-term understanding of the sedimentary environment in this region (Anderson et al. 2011; Nicholas et al. 2014). Moreover, mixed siliciclastic–carbonate sedimentary environments in low-latitude and semi-enclosed marginal marine environments provide information on the mechanism of paleoclimatic and hydrologic change (cf., Bahr et al. 2005; Isaack et al. 2016; Soulet et al. 2011). However, our understanding of late Quaternary evolution of the Bonaparte sedimentary environment is much less well constrained.

Here, we document the evidence of environmental changes due to the exposure of these carbonate terraces and platforms before the LGM. The Bonaparte Gulf is a “far-field” (cf., Yokoyama and Esat 2011), tectonically stable site extremely suitable for reconstruction of sea level due to past continental ice-volume change (De Deckker and Yokoyama 2009; Ishiwa et al. 2016; Yokoyama et al. 2000, 2001b). We also describe the sedimentary consequences of past environmental change from a marine piston core (KH11-1-PC01) using Ca/Ti ratios, total organic carbon (TOC) and total nitrogen (TN), constrained by radiocarbon dating.

Methods

Materials

Core KH11-1-PC01 was recovered from a water depth of 140 m during the KH11-1 cruise of R/V Hakuho-Maru during January and February 2011 (Fig. 1). The top 600 cm interval of core KH11-1-PC01 (core recovery length: 951 cm) was analyzed as it is within the range of radiocarbon dating. Well-preserved macrofossils (primarily bivalves) were collected for radiocarbon dating.

Physical properties and geochemical analysis

Color reflectance was measured at 2-cm intervals after splitting the cores on ship using a Minolta CM-2002 photospectrometer. At the same time, magnetic susceptibility was measured at 2-cm intervals using a Bartington Instruments MS2C system.

TOC and TN were measured using an EA-IRMS (Elemental Analysis-Isotope Ratio Mass Spectrometry: Flash EA 1112 and Delta plus Advantage) at the Center for Advanced Marine Core Research, Kochi University, Japan (cf., Ishiwa et al. 2016; Nakamura et al. 2016; Riethdorf et al. 2015). Bulk sediments were treated with 3 M HCl to remove inorganic carbonate. Split-core scanning X-ray fluorescence (XRF) analysis was later conducted at a 1-cm interval using the TATSCAN-F2 at the Center for Advanced Marine Core Research (Sakamoto et al. 2006). The XRF core scanner analysis was affected by split-core surface condition and water content, especially for the light elements (Kido et al. 2006; Nakamura et al. 2016). Hence, we discuss the variation of relatively heavy elements, calcium (Ca), and titanium (Ti). The flux of Ca and Ti is calculated as follows,

$${\text{MAR}} = {\text{DBD}} \times {\text{LSR}}$$
$${\text{Ca}}\left( {\text{Ti}} \right){\text{flux}} = {\text{MAR}} \times {\text{Ca}}\left( {\text{Ti}} \right)\left( {\text{counts}} \right)$$

where MAR is the mass accumulation rate (g/cm2 year), DBD is the dry bulk density (g/cm2), and LSR is the linear sedimentation rate (cm/year).

Radiocarbon dating was performed on marine macrofossils (Table 1) and bulk sediment organic matter. Macrofossils were etched by 10 M HCl to remove secondary and contaminating calcium carbonate (Yokoyama et al. 2016a). Bulk sediments were pretreated twice in 3 M HCl for 12 h to digest inorganic calcium carbonate (Ishiwa et al. 2016). We followed the methods of Yokoyama et al. (2007) in the graphitization and measured the graphite by the Single Stage Accelerator Mass Spectrometry (Yokoyama et al. 2016b) and the Micro Analysis Laboratory Tandem Accelerator at the University of Tokyo.

Table 1 Age results of macrofossils in core KH11-1-PC01
Full size table

Age–depth model

Calendar ages were calculated using Oxcal (Ramsey and Lee 2013) with Marine 13 and Intcal 13 (Reimer et al. 2013) as calibration curves for macrofossils and organic matter ages. The local reservoir correction, ΔR, is undefined in the Bonaparte Gulf but expected to be minor (cf., O’Connor et al. 2010). Thus, we made no local correction, consistent with previous works (Ishiwa et al. 2016; Yokoyama et al. 2000, 2001b).

The age–depth model of KH11-1-PC01 was constrained using the BACON model (Blaauw and Christen 2011) based on macrofossil ages, since organic matter ages are affected by the transportation time of terrigenous components (cf., Ishiwa et al. 2016; Nakamura et al. 2016). This model uses the Bayesian analysis and Monte Carlo methods to constrain the smoothing age–depth model using the R statistical software package (cf., De Vleeschouwer et al. 2012; Shanahan et al. 2012).

Calculation for exposure percentage in the Bonaparte Gulf

The area of carbonate terraces and platforms in the Bonaparte Gulf exposed during lower sea level was calculated using the bathymetric dataset from Whiteway (2009). The data were interpolated to a uniform resolution of 0.5 min latitudinally and longitudinally. We calculated the area of exposure along the sea-level curve and set the exposure percentage to 0% at relative sea level = 0 m and to 100% at relative sea level = −120 m. This percentage was calculated at a 5-m interval.

Results

Lithology and physical properties

The interval from 600 to 560 cm of core KH11-1-PC01 is silty clay with mm-scale shell fragments. Fine sand is present from 560 to 520 cm and silty clay occurs from 520 to 510 cm with shell fragments, foraminifers and nannofossils (as noted in shipboard smear slides). Fine sand is present from 510 to 490 cm and silty clay with shell fragments, and nannofossils from 490 to 60 cm. The upper 60 cm of core is silt and clay with shell fragments and bioturbation. Color reflectance (b*) values gradually increase from 1 to 9 in the uppermost 300 cm with a slight shift at 200 cm, indicating an increasing yellow component. A relatively large increase in b* occurs at 30 cm (Fig. 2). Magnetic susceptibility (MS) shows a slight decrease from 7 to 4 SI 10−6 at 200 cm with a maximum of 11 SI 10−6 at 10 cm. In the upper 10 cm, the MS decreases to 0 SI 10−6 (Fig. 2).

Fig. 2
figure 2

Image of core KH11-1-PC01, X-ray image, lithology, color reflectance (b*), magnetic susceptibility, Ca, Ti, TOC, and C/N ratios

Full size image

XRF core scanning and geochemical analysis

Ca counts are constant at 600,000 from 600 to 200 cm, increasing to ~800,000 counts at 180 cm (Fig. 2). There is a sharp drop at 70 cm. From 30 cm to the core top, Ca decreases to 400,000 counts. Ti counts gradually decrease from 600 to 200 cm, sharply decreasing to 20,000 counts at ~200 cm (Fig. 2). There is a peak of ~25,000 counts from 30 cm to the core top.

TOC is ~0.7% from 600 to 200 cm, sharply increasing to 1.5% at ~180 cm with a peak of ~0.8% at ~510 cm (Fig. 2). TOC is variable in the upper 40 cm of core. C/N ratios reach a maximum of ~15 at 510 cm, then maintain a value of ~9 through the depth interval of 500–180 cm, above which values are relatively constant at ~13 (Fig. 2). In the upper 60 cm, ratios decrease upwards to the core top.

Radiocarbon dating

Tables 1 and 2 summarize the results of radiocarbon dating and Fig. 3 shows the age–depth relationship. Macrofossils dates show the range from ca. 35 to 24 ka and bulk organic matter from ca. 37 to 25 ka during the interval of 600–60 cm. The upper 60 cm is either disturbed by coring or winnowed by tidal currents (http://www.bom.gov.au/australia/tides/) and pockmark activity (Nicholas et al. 2014), and the ages in this interval are not interpreted here (Fig. 3). Below 60 cm, the age–depth model is well constrained with an offset between macrofossils and organic matter ages (cf., Ishiwa et al. 2016). The average value of this offset is ~800 14C year within the period from ca. 35 to 25 ka and ~1200 14C year with the period from ca. 25 to 24 ka. The average sedimentation rate is 0.47 m/kyr for the period from ca. 35 to 25 ka and 0.75 m/kyr for the period from ca. 25 to 24 ka.

Table 2 Age results of organic matter in core KH11-1-PC01
Full size table
Fig. 3
figure 3

The age–depth relationship of core KH11-1-PC01. a Calendar ages for macrofossils (blue) and organic matter (red). The errors are 2 sigma. b Expanded view of the age–depth model in 22–38 ka. The line is the mean value of the BACON model (Blaauw and Christen 2011) and the shade represents the minimum and maximum range of the BACON model

Full size image

Exposure of carbonate terraces and platforms in Bonaparte Gulf

Figure 4a shows the calculated percentage of exposure and the rate of sea-level change from 0 to −120 m in 5 m increments. During sea-level lowstands, the proportion of carbonate terraces and platforms exposed in Bonaparte Gulf ranged from 25% when sea level was at −60 m to in excess of 90% when sea level fell below −100 m. The rate of change from −70 to −90 m is greater than 1.5%/m. The maximum rate of change is at −80 m (Fig. 4b).

Fig. 4
figure 4

a Proportion (%) of carbonate platform area in Bonaparte Gulf exposed at lower sea levels; b rate of change in proportional area of the platforms exposed at sea level fall

Full size image

Interpretation and discussion

Sedimentary environmental change during late Quaternary

Biogenic or precipitated carbonate shows high Ca-intensities with an inverse relationship to K-, Fe-, Mn-, and Ti- intensities that are correlated with siliciclastic components (cf., Bahr et al. 2005; Kuhnt et al. 2015). Ca/Ti, Ca/K, Ca/Mn and Ca/Fe all exhibit a similar pattern of variability (see Additional file 1: Figure S1). Thus, we focus on Ca/Ti, which is less sensitive to redox changes. We suggest that calcium variation represents changes in biogenetic carbonate flux from the carbonate terraces and platform of the Bonaparte Gulf (Fig. 5; and see Additional file 1: Figure S2). The Ca flux after 25 ka is much higher than before 27 ka, indicating a carbonate flux increase caused by exposure of the carbonate platforms during lower sea level at ca. 26 ka (190 cm) (Fig. 5). In addition, we use the titanium variation to represent terrigenous sediment flux, derived from clay minerals (cf., Gingele et al. 2001; Gingele and De Deckker 2003). Magnetic susceptibility values decrease from 27 to 25 ka (Fig. 6). This pattern likely reflects the dilution by sediments dominated by biogenetic carbonate with low magnetic susceptibility.

Fig. 5
figure 5

Ca flux, Ti flux, and mass accumulation rate. Solid lines correspond to the average values during three periods, before ~27 ka, during 25–27 ka, and after ~25 ka

Full size image
Fig. 6
figure 6

Ca/Ti, magnetic susceptibility, C/N ratios, TOC, TOC flux, and sedimentation rate, and relative sea-level records. Gray shade corresponds to the timing of carbonate platform exposure due to sea-level fall. A Ca/Ti. B Magnetic susceptibility. C C/N ratios. D TOC (%). E TOC flux. F Relative sea-level records in Huon Peninsula. Black circles correspond to Yokoyama et al. (2001c). White circles to Cutler et al. (2003). Red line to Siddall et al. (2003)

Full size image

The variation in terrigenous input is interpreted to be related to changes in TOC flux and C/N ratios (cf., Ishiwa et al. 2016; Yu et al. 2010; Mackie et al. 2005). The terrigenous sediment supply increased after ca. 26 ka, as indicated by the increased TOC, C/N ratios, and sedimentation rate (Figs. 5, 6). Increased terrigenous material during this time would make mixed marine-terrigenous organic matter older, increasing the offset from macrofossil ages (Fig. 3). Additionally, mass accumulation rate calculated by the BACON model supports the exposure of carbonate terraces and platforms during this period (Fig. 5).

Mechanism for sedimentary environmental change

The carbonate terraces and platforms of the outer Bonaparte Gulf are dissected by a network of deeply incised paleochannels that would have provided sediment transport pathways during lower sea level (Courgeon et al. 2016; Fig. 7; see also Additional file 1: Figure S3). Yokoyama et al. (2000, 2001b) estimated the age of deposition in these paleochannels to be ca. 17.5 ka, suggesting that carbonate sediments were produced and transported to the depression during lower sea level by fluvial activity or ocean currents.

Fig. 7
figure 7

Bathymetry and sectional views of the Bonaparte Gulf at −60 and −90 m sea level. Note that there are bathymetric error estimates in the Bonaparte Gulf due to the hydro-isostatic effect (Yokoyama et al. 2000, 2001a, b). The effects are relatively minor

Full size image

Sedimentation patterns would have fluctuated with the relative strength of monsoon (Gallagher et al. 2014b; Kuhnt et al. 2015). The Australian Monsoonal precipitation pattern is sensitive to latitudinal ITCZ migration (Lewis et al. 2011), while speleothem records from this region indicate low monsoon variability at this time (Lewis et al. 2011; Partin et al. 2007), consistent with marine and terrestrial records (Fitzsimmons et al. 2013; Reeves et al. 2013) that show a northward ITCZ position. We suggest that the changes in monsoonal variability are not strong enough to control the sedimentary facies in the Bonaparte Gulf. A relative change in carbonate sediment flux increased at ca. 26 ka as shown by physical properties and geochemical analysis (Fig. 5). Rivers passing through the continent would have supplied siliciclastic sediments to the depression (Gingele et al. 2001; Gingele and De Deckker 2003). During this period, the supply of siliciclastic sediments did not change much due to the weak variability of monsoonal intensity. By contrast, the carbonate supply increased due to the exposure of carbonate terraces and platforms.

Sea level below −90 m resulted in sufficient exposure of carbonate terraces and platform to increase the flux of carbonate sediments (Fig. 7). During sea-level highstands, much of the shelf was submerged. During sea-level lowstands, the carbonate terraces and platforms were exposed and the Bonaparte Depression was semi-enclosed from the Timor Sea. The relative area of exposure was an important control on the sedimentary facies of the gulf.

While the hydro-isostatic effects in the Bonaparte Gulf do not significantly affect to the current interpretation (Yokoyama et al. 2000, 2001b), this factor cannot be entirely discounted due to paleotopography and paleo-water depth effects. Yokoyama et al. (2000, 2001b) estimated the amplitude of this effect (offset between the global sea level and relative sea level) is less than 20 m. Therefore, the bathymetry in Fig. 7 has an error of 10% due to hydro-isostasy effects.

These observations indicate that the sedimentary environment in the Bonaparte Gulf is primarily driven by sea-level variability due to the distinctive topography with its central depression surrounded by higher-level carbonate terraces and platforms. Exposure of carbonate terraces and platforms with a sea-level fall enhances sedimentary environmental change, characterized by carbonate sediment production and transportation. Paleobathymetry would also likely have affected local ocean circulation patterns and tidal ranges.

Sea-level change prior to the LGM has been estimated using the uplifted coral on the Huon Peninsula, indicating that sea level fell from −70 to −110 m from 30 to 24 ka (Cutler et al. 2003; Yokoyama et al. 2001c). This is consistent with sea level from the Red Sea (Siddall et al. 2003), which falls ~20 m from 28 to 26 ka (−80 to −100 m).

Comparison to other regions

Mixed carbonate-siliciclastic sedimentary systems are observed in low-latitude tropical regions. Isaack et al. (2016) present a sea-level driven model of sediment dynamics based on a multi-proxy record in the barrier-reef lagoon of Bora Bora in the South Pacific. They suggest that carbonate sediment produced in marginal reef areas is transported to the lagoons. This mechanism indicates that carbonate terraces and platforms in the Bonaparte Gulf can be the source of carbonate sediments, generating the variation of geochemical signal in the core.

The Black Sea became a semi-enclosed marginal sea, similar to the Bonaparte Gulf, during lowering sea levels (Bahr et al. 2005), and reconnected to the Mediterranean Sea at ~9000 years ago in association with rising sea level (Soulet et al. 2011). Geochemical data suggest that the hydrologic system has changed from lacustrine to marine, controlled by a water depth of sills connected to Mediterranean Sea. By contrast, during sea-level lowstands, the Bonaparte Gulf connected to the Timor Sea mainly by paleochannels at a water depth of ~200 m (Fig. 1; Yokoyama et al. 2000, 2001a). The sedimentary environment in the Bonaparte Gulf is influenced by the exposure of carbonate terraces and platforms, which played a role in pathways of carbonate sediments.

Conclusions

Geochemical and chronological records from a piston core in the Bonaparte Gulf, northwestern Australia, show a sedimentary environmental change at ca. 26 ka. Carbonate terraces and platforms and their deeply incised paleochannels play an important role in this paleoenvironmental change.

Ca/Ti ratios as an indicator of changes in mixed siliciclastic–carbonate sediments increased at ca. 26 ka, indicating the enhanced supply of carbonate sediments. TOC and C/N ratios as an indicator of terrigenous input also increased. The changes in these geochemical signals can be explained by the exposure of carbonate terraces and platforms.

Precipitation patterns shift can drive the change in sediment supply. However, during the period we investigate, the variability of Australian monsoon was not strong enough to change the sedimentary environment. By contrast, global sea-level fall to −90 m occurred at ca. 26 ka, driving the exposure of carbonate terraces and platforms and the switch from siliciclastic to carbonate-dominated sedimentation. Our research provides an understanding of a sea-level driven sedimentary environmental change in low-latitude mixed siliciclastic–carbonate environment.