香京julia种子在线播放

End-members of sediment grain-size obtained from EMMA analyses have been frequently associated with sediment transport and/or deposition mechanisms in various environmental settings. Previous studies suggest that in the Southern Ocean, oceanic currents, icebergs, sea ice and wind could be potential media that deliver terrestrial sediments to the seas (Pudsey, 1992; Gilbert et al., 1998; McCave et al., 2014; Wu et al., 2018; Wu et al., 2021). Sea ice, however, forms in deep water continental shelf (~500 m) in Ross Sea, with little sediment encased. Earlier studies indicate that the WRS coast could be a possible source area for wind-blown dust (Delmonte et al., 2010). Nonetheless, none of the four end-members displays a spatial distribution pattern decreasing seaward from the WRS coast, excluding wind-blown dust as a major contributor to the sediments. Therefore, we associate the grain-size end-members with sediment delivery primarily by icebergs and bottom currents.

In polar oceans, icebergs can carry sediments with a broad range of grain-size compositions. Therefore, the grain-size distribution of IRD is typically described as poorly sorted, even though coarse sediment fractions such as the >150 μm and/or >250 μm fractions from polar oceans are often considered as IRD. This is because sediment of such diameters cannot be moved and sorted by typical oceanic currents (Lamy et al., 2015; Wu et al., 2020). The grain-size distribution of EM1 ( Figure 2D ) displays poor sorting and bears a resemblance to the previously published grain-size distribution of IRD (Prins et al., 2002; Wu et al., 2018). Moreover, there is a robust correlation between EM1 and the >250 μm component obtained from wet sieving (R=0.63, p<0.01). We thus interpret EM1 as IRD.

EM2 has the best sorting and a relatively coarse mean grain-size ( Figure 2D ). This grain-size distribution is likely produced by strong current winnowing (e.g., Abuodha, 2003). However, there is a weak positive correlation between EM2 and the sorting parameter of the <150 μm fraction (r=0.31, n = 41), implying that higher EM2 content does not correspond to better sorting of the finer fraction of the sediments. Modern observations indicate that the bottom current velocity of the Ross Sea continental shelf generally remains below 15 cm/s (Jacobs et al., 1970), which is insufficient to transport particles >63 μm (McCave and Hall, 2006). Thus, the good sorting of the relatively coarse EM2 cannot be interpreted as bottom current winnowing lags. Instead, previous studies have identified well-sorted IRD in other ocean areas (Prins et al., 2002; Wang et al., 2023). It is postulated that this IRD may originate from the transport of icebergs in regions with intense hydrodynamics, such as coastal sand beaches (Prins et al., 2002; Wang et al., 2023). In such environments, the sediment itself is sorted well before being entrained in icebergs. Indeed, paleo-beaches forming during Middle Holocene Climate Optimum have been frequently found situating above the modern sea level along the western coast of Ross Sea (Colhoun et al., 1992). EM2 is thus interpreted as IRD delivered by icebergs derived from coastal beaches in the Western coast of Ross Sea.

The grain-size distributions of EM3 and EM4 exhibit significant overlap with the sortable silt interval (10-63 μm) ( Figure 2D ). This indicates that they can be transported and/or altered by bottom currents in deep-sea environments (McCave and Hall, 2006; McCave et al., 2017). The presence of a larger modal grain-size combined with relatively better sorting suggests that EM3 would likely accumulate in an environment with stronger bottom currents. Conversely, the finer grain-size and slightly poorer sorting of EM4 are indicative of deposition in an environment with weaker hydrodynamics (Jonkers et al., 2015; Wu et al., 2018).

The Ross Sea region has been identified as a significant source of icebergs in the Southern Ocean, and extensive presence of icebergs has been observed in this area (Turnbull, 2006). Indeed, the contents of EM1, EM2 and the wet sieving-based >250 μm fraction in most samples are above zero, suggesting widespread iceberg unloading throughout the sampling area. Iceberg unloading is primarily associated with dynamics of the ice shelf/glacier that produces the iceberg, but might be further influenced by a range of factors such as the sea water temperatures, ocean current pathways, debris type and amount in icebergs, distance from iceberg sources, and geomorphology of the seafloor (Teitler et al., 2010; Cook et al., 2013; Cook et al., 2014; Weber et al., 2014; Wu et al., 2021).

Antarctic ice shelves as large as the RIS, are typically devoid of basal debris as most of the debris have been unloaded near their grounding zone prior to calving (Patterson et al., 2014; Smith et al., 2019). This accounts for the extremely low content of IRD immediately in front of the RIS to the east of 178°E ( Figures 5A , 7A ). The presence of the RIS provides protection against wave and tidal action, whereas the ice shelf-free coastlines of the WRS have experienced intense wave, tide, and storm activity since the Mid-Late Holocene, after the retreat of the RIS to its current position (McKay et al., 2008), favoring the development of beaches (e.g., Colhoun et al., 1992). We noted that the maximum value of EM2 decreased significantly with increasing longitude ( Figure 8B ), suggesting that EM2 originated from the ice shelf-free western coast of Ross Sea (See Section 6.1.1). Conversely, there was no observable trend in the spatial distribution of EM1 or the wet sieving-based >250 μm fraction across longitudes ( Figures 8A, C ), in combination with the fact that EM1 and EM2 display dissimilar grain-size distribution patterns ( Figure 2D ), suggesting that the source regions of EM1 and EM2 are distinct.

On the other hand, we observed a complex relationship between IRD and water depth. When the water depth is below ~660 meters, there is significant variation in the proportion of EM1 and the weight percent of the wet sieving-based >250 μm fraction among different stations, ranging from 0 to 72% and from 0 to 85 wt.%, respectively ( Figures 8G, I ). However, when the water depth exceeds ~660 meters, this variation range abruptly narrows to between 0 and 30%, and between 0 and 22 wt.%, respectively. Notably, there is no such relationship between EM2 and water depth ( Figure 8H ).

The RIS has an average thickness approximately 600 meters (McKay et al., 2018), implying that large icebergs calved from the RIS are more likely to remain grounded for extended periods when the water depth on the continental shelf is close to this value. This may lead to extended unloading of basal and englacial detritus of the icebergs there and even disturbance of previously deposited sediments, as described by O'Brien and Leitchenkov (1997) for the scenario in Prydz Bay, East Antarctica. In contrast, smaller icebergs could unload their detritus at any water depth deeper than the thickness of their marine portion. This mechanism could explain the complex relationship between water depth and EM1 as well as the >250 μm fraction ( Figures 8G, I ). The presence of different patterns between EM1 and water depth ( Figure 8G ), and between EM2 and water depth ( Figure 8H ), thus may indicate that at least part of EM1 was released from larger icebergs, while EM2 was released from smaller ones.

Additionally, significant positive correlations were observed between all the indicators of IRD and latitude ( Figures 8D-F ), indicating a tendency for IRD deposition towards the outer shelf, coinciding with the intrusion of MCDW onto the continental shelf (Whitworth and Orsi, 2006). The warmer MCDW that enters the Ross Sea compared to other water masses can enhance the basal melting of icebergs, promoting the release of IRD. Consequently, the overall increase in IRD towards the north suggests that variations in sea water temperature, influenced by the intrusion of MCDW, play a significant role in governing delivery of IRD onto the Ross Sea continental shelf.

The analysis of sediment grain-size offers valuable insights into the dynamics of sediment transport. In the deep-sea setting, the duration of sediment settling from the water column to the seafloor is relatively brief in comparison to its residence time near the sediment-water interface, particularly in areas characterized by low sedimentation rates. Consequently, when establishing a connection between sediment grain-size and ocean currents, the primary focus lies on discerning variations in the intensity of bottom currents (McCave and Hall, 2006; McCave and Andrews, 2019).

McCave and Hall (2006) proposed using the geometric mean grain-size of the 10-63 μm fraction, which is known as the sortable silt mean (SSM), as a proxy for changes in bottom current strength, provided that it correlates well with the percentage of the sortable silt (SS%). The presence of IRD in high-latitude oceans, however, can complicate this relationship, particularly when bottom current strength is weak (Jonkers et al., 2015; Stevenard et al., 2023). In such cases, coarser components in IRD are not effectively sorted by the bottom current, resulting in an increase in SSM with the addition of coarse component content (Jonkers et al., 2015; Stevenard et al., 2023). In our samples, we found a significant correlation between SSM and SS% ( Figures 7C, D , 9A ), but an insignificant correlation between SSM and the IRD end-member EM1 ( Figure 9B ). Interestingly, there was a significant correlation between SSM and EM2 ( Figure 9C ), and the spatial distribution pattern of SSM overlapped with that of EM2 as well ( Figures 3B , 7C ), leading to suspicions that the SSM here could be used as a proxies for bottom current strength change ( Figure 2D ). In view of this, we propose a new index, EM3/(EM3+EM4) ( Figure 7B ), to evaluate the relative change of local bottom current strength. This index utilizes indicators that reflect transport by relatively strong and weak bottom currents in EMMA end-members, which are unaffected by IRD (e.g., Prins et al., 2002; Jonkers et al., 2015; Wu et al., 2018; Wang et al., 2023).

The EM3/(EM3+EM4) ratio exhibits a significant correlation with both SSM and SS% ( Figures 9D, E ). However, no significant correlation was observed between the EM3/(EM3+EM4) ratio and any IRD indicator ( Figures 9F, G ), and the spatial distribution of EM3/(EM3+EM4) ( Figure 7B ) also differs from that of the IRD indicators ( Figures 3A, B ). Moreover, a significant negative correlation was found between the EM3/(EM3+EM4) ratio and the sorting of the <100 μm fraction ( Figure 9H ). This suggests that higher values of EM3/(EM3+EM4) correspond to better sorting of the fine fractions (e.g., Passchier, 2011; Patterson et al., 2014; Wu et al., 2021), thus supporting its validity as an indicator for relative changes of the bottom current strength in our study. Higher EM3/(EM3+EM4) values indicate stronger bottom current activities, and vice versa.

Spatially, elevated values of EM3/(EM3+EM4) are predominantly observed in the WRS, particularly in the Joides basin, Joides trough, and the northernmost region of Victoria Land ( Figure 7B ). This spatial distribution indicates that these specific areas underwent notable activities of relatively strong bottom currents.

In the Southern Ocean, wind stress curl is a crucial factor that influences the oceanic current activities across the water column. Furthermore, the energy derived from the wind stress curl is regulated by sea ice cover (e.g., Toggweiler et al., 2006; McCave et al., 2014). However, we found no correlation between the EM3/(EM3+EM4) ratio and water depth in our samples ( Figure 9I ). This suggests that the energy that prompts bottom current activity does not originate from the sea surface. Otherwise, shallower sampling stations would display higher bottom current strength due to greater energy gain, resulting in a higher EM3/(EM3+EM4) ratio. Therefore, we propose that bottom current strength in our study is not related to wind stress curl, and other factor(s) must be involved to explain its spatial distribution and variability.

It is well know that the WRS serves as a significant source region for the generation of Antarctic Bottom Water (AABW), accounting approximately for 25% of its overall production (Orsi et al., 2002; Orsi and Wiederwohl, 2009). Formation of AABW in the WRS is closely linked to the rejection of brine resulting from sea ice production within the polynya (Tamura et al., 2008; Gordon et al., 2015). The influx of warm and saline MCDW from the WRS leads to raised salinity within the local water mass. Concurrently, continuous sea ice formation in the polynya further elevates the salinity of the expelled brine. As a result, highly saline brine descends and accumulates in deep water basins on the continental shelf, giving rise to Dense Shelf Water (DSW). Subsequently, the DSW moves northward through submarine canyons and, driven by tidal activity, passes the sills on the outer continental shelf. Eventually, it descends the continental slope, thereby culminating in the formation of AABW (Tamura et al., 2008; Gordon et al., 2015).

Recent observations indicate that the Joides Basin and Joides Trough in the WRS play a pivotal role in the northward advection of DSW generated within the region. Conversely, the Glomar Challenger Basin and areas east of Pennell Trough do not exhibit the presence of this water mass (Gordon et al., 2015). Interestingly, this spatial distribution pattern of DSW closely aligns with the EM3/(EM3+EM4) ratio depicted in our findings ( Figure 7B ). Hence, we deduce that the stronger activity of bottom currents in the WRS is primarily governed by the formation and transport of DSW.

Clay minerals in marine sediment serve as valuable indicators of sediment origin due to their small grain-size, enabling easy transportation and distribution through various media (Hillenbrand et al., 2003; Liu et al., 2003; Liu et al., 2007). Illite and chlorite are commonly found in high-latitudes of the Southern Ocean as a result of physical weathering and glacial erosion of basement rocks in Antarctica (Chamley, 1989; Ehrmann, 1991; Ehrmann, 1998). Illite originates from acidic rocks including granulites, gneisses, and granites. Chlorite is derived from igneous and metamorphic rocks that contain chlorite (Chamley, 1989; Hillenbrand et al., 2003). In contrast, smectite and kaolinite form by chemical weathering. Smectite forms through hydrolysis in environments with slow water movement and is often associated with volcanic regions in the Southern Ocean (Petschick et al., 1996). Kaolinite develops through extensive hydrolysis of feldspars in humid and warm environments (Chamley, 1989). The occurrence of kaolinite around Antarctica is usually related to recycling of ancient sediment (Junttila et al., 2005; Borchers et al., 2011).

To identify the source areas of the clay minerals in our samples, we utilized S-IC-K and S-I-C ternary diagrams to analyze our surface sediment data alongside a compilation of previously published data ( Figures 10A, B ). In the S-IC-K diagram, we observed that surface sediments from the WRS were closer to the IC apex, while those from the Eastern and CRS were closer to the S apex. All samples followed a line adjacent to the IC-S edge ( Figure 10A ). In the I-S-C diagram, we found that the WRS samples were nearer the I-C edge, while the ERS and CRS samples were closer to the S vertex. Additionally, all samples fell along a line perpendicular to the I-C edge roughly in the middle of the diagram ( Figure 10B ). These diagrams indicate that the studied sediments were the result of binary mixing between two end-members: an illite- and chlorite-rich end-member, and a smectite-rich end-member. Notably, neither of the end-members overlaps with the clay mineral composition of the surface sediments from the continental shelf north of Marie Byrd Land ( Figures 10A, B ), suggesting a minimal contribution from the Amundsen Sea via the westward flowing Antarctic Coastal and Slope Currents. Additionally, the contribution of the clay fraction by impinging MCDW from the north could be ruled out as well since the outer shelf samples are located between the two identified end-members in the diagrams. Thus, we conclude that both the illite- and chlorite-rich end-member and the smectite-rich end-member were primarily originated from the Ross Sea Embayment itself, which is consistent with the results of Nd and Sr isotopic analyses conducted on the clay-sized fractions of the surface sediments within the investigated region by Shao et al. (2022).

The examination of Cenozoic sediment sequences from McMurdo Sound, as documented in previous studies, reveals a correlation between warmer climate conditions and higher concentrations of illite and chlorite, while colder climate conditions are associated with higher concentrations of smectite ( Figures 10A, B ; Ehrmann et al., 2003). The illite and chlorite components are believed to originate from the Beacon Supergroup on the Southern Victoria Land coast, whereas the smectite component is thought to derive from areas where the McMurdo Volcanic Group outcrops. Our findings support an East Antarctic craton origin for the significant abundance of illite and chlorite in our samples from the WRS, particularly in close proximity to the Southern Victoria Land ( Figures 4A, C ). However, the overall declining trend of smectite from east to west ( Figure 4B ) contradicts the notion of a potential origin from the McMurdo Volcanic Group which is located in the WRS ( Figure 1B ). Instead, it is more reasonable to interpret the spatial distribution of smectite as originating from beneath the glaciers of West Antarctica. This is supported by the presence of Cenozoic mafic volcanic structures beneath the glacier systems that drain into the Ross Sea from the WAIS (Behrendt et al., 1995; Behrendt et al., 2004), as well as the elevated levels of smectite found in tills derived from the Bindschadler Ice Stream and MachAyeal Ice Stream ( Figure 1A ; Balshaw, 1981). Nevertheless, the sporadic occurrence of high smectite levels in front of Ross Island and near the northernmost coast of Victoria Land likely originates locally due to their proximity to the McMurdo Volcanic Group ( Figure 1B ). Our study thus highlights the need for caution when considering the origin of smectite in sedimentary records from the Ross Sea, given its potential multiple sources.

The consistency between kaolinite and smectite in terms of spatial distributions in the ERS ( Figures 4B, D ) may suggest that they were derived from spatially adjacent sources beneath the WAIS. This is supported by the occurrence of kaolinite in sediments from the Amundsen Sea continental shelf where clay minerals were mainly derived from beneath the WAIS as well (Hillenbrand et al., 2003), and the absence of kaolinite from sedimentary records in the WRS ( Figure 10A ). The lower kaolinite abundance in our samples compared to the sediments from Amundsen Sea indicates less entrainment of ancient chemically weathered sediments filled in the West Antarctica basin (Rooney et al., 1988) by glaciers draining into Ross Sea from the WAIS.

The congruity in spatial distributions between kaolinite and smectite within the ERS ( Figures 4B, D ) implies that they were likely originated from neighboring sources beneath the WAIS. This is reinforced by the presence of kaolinite in sediments collected from the Amundsen Sea continental shelf, where clay minerals are predominantly derived from beneath the WAIS as well (Hillenbrand et al., 2003), and the absence of kaolinite in sedimentary records within the WRS ( Figure 10A ; Ehrmann et al., 2003). The lower abundance of kaolinite in our samples, in comparison to the sediments from the Amundsen Sea, suggests a reduced incorporation of ancient chemically weathered sediments, which fill the basin of West Antarctica (Rooney et al., 1988), by glaciers draining into the Ross Sea from the WAIS.

The region surrounding the Ross Sea is predominantly covered by thick ice sheets, but exposed outcrops reveal the intricate lithologies of the bedrock. The uneven distribution of rock outcrops presents an opportunity to follow the origin of the coarse-grained IRD through an analysis of its lithological composition (Licht et al., 2005). Despite the intricate and diverse composition within the coarse fraction, Factor Analyses indicate the presence of only three primary components ( Figure 6 ), characterized by high levels of angular and rounded quartz, mafic rock grains, and deformed silt, respectively. These components are likely associated with three main sources that dominate the coarse fraction, enabling the provenance analyses of the coarse fraction to be simplified as based on them.

Factor 1, as evidenced by high loadings of both angular and rounded quartz ( Figure 6C ), is likely derived from acidic source rocks. In the Antarctic region, quartz is commonly enriched in the IRD fractions compared to other minerals due to its high mechanical hardness, rendering it resistant to breakage during glacial grinding and retains a larger particle size (Wu et al., 2020). Licht et al. (2005) observed that tills beneath the Kamb Ice Stream and Whillans Ice Stream ( Figure 1A ) contained a high abundance of both angular and rounded quartz but almost an absence of mafic volcanic rocks, despite Cenezoic volcanoes being scattered beneath these glaciers. Furthermore, quartz was the most abundant mineral species in till samples across the ERS and CRS during the Last glacial periods ( Figures 10C, D ), possibly eroded from the Granite Harbor Intrusives, as well as the granodiorite and granite of Marie Byrd Land ( Figure 1B ; Licht et al., 2005). In our samples, Factor 1 is predominantly present in the ERS and CRS regions and extends to the central Victoria Land coast. Additionally, in the Q-M-S and Q-M-F diagrams, our ERS and CRS samples mostly occupy the domains of the West Antarctic & ERS tills and the RISP (RIS Drilling Project; Licht et al., 2005) tills, supporting that the quartz-enriched Factor 1 represents a source of icebergs originating from West Antarctica.

Factor 2, characterized by a significant presence of mafic igneous rock grains, is likely indicative of materials primarily derived from a basic volcanic region. Tills in East Antarctica exhibit varying lithologies in comparison to those in West Antarctica ( Figures 10C, D ), reflecting the diverse composition of bedrock throughout the Transantarctic Mountains ( Figure 1B ). However, a substantial mafic component is consistently found in most East Antarctic tills, which is believed to originate from the Ferrar Group (Licht et al., 2005). In the Q-M-F ternary diagram, our samples from the WRS align well with the East Antarctic & WRS tills ( Figure 10D ). In the Q-M-S diagram, however, these samples disperse over a broader range than the East Antarctic & WRS tills ( Figure 10C ). Some WRS samples display higher quartz content compared to the remaining samples from the WRS, and their distribution tends to overlap with the West Antarctic & ERS tills, indicating a mixture of materials transported by icebergs from both West and East Antarctica. These particular samples are located between 74°S and 75°S in the WRS ( Figure 5B ). In contrast, certain WRS samples contain a significantly higher proportion of mafic rock grains than the tills from East Antarctica and the WRS, suggesting an additional source of mafic igneous rocks. These samples are situated between 72°S and 73°S, as well as between 76°S and 78°S ( Figure 5C ), close to areas where the McMurdo Volcanic Group is exposed ( Figure 1 ). This suggests that the McMurdo Volcanic Group may contribute significantly to the modern IRD found in the WRS.

The moderate scores (0.1-0.2) observed in Factor 2 within the ERS area likely correspond to the presence of Cenozoic mafic volcanic structures underlying the glacier systems that drain into the Ross Sea from the WAIS. However, an apparent discrepancy arises when considering that there is a high abundance of smectite in the clay fraction but a disproportionately low amount of mafic rock grains in the ERS. It has been proposed by Behrendt et al. (1995); Behrendt et al. (2004) that (ancient) glacial erosion likely planed off the mafic volcanic structures in West Antarctica before the development of the current ice streams. Rooney et al. (1988) suggest that the West Antarctic basin, which lies beneath the WAIS, is filled with glacial and marine sediments up to a kilometer thick, masking the volcanic structures. Consequently, the coarse mafic rock grains may be more commonly found in older successions beneath West Antarctica (Licht et al., 2005), making it less accessible to the present glaciers.

Factor 3 exhibits a prominent abundance of deformed silt, which comprises coarse-grained materials that form during the stabilization of a glacier at its grounding zone (Cowan et al., 2012). The prevalence of these materials within the sediment indicates the proximity of the sediment location to a glacier grounding zone (Cowan et al., 2012). Geospatially, the highest score in Factor 3 is primarily observed in the David Glacier-Drygalski Ice Tongue (DG-DIT) area, providing strong evidence for its derivation from beneath this specific glacier system.

The DG-DIT system represents the largest outlet glacier within the Victoria Land section of the EAIS (Frezzotti and Mabin, 1994). Over the past century, this system has undergone numerous calving events (Frezzotti and Mabin, 1994). Consequently, as glaciers detach from this ice tongue, they have a propensity to incorporate deformed silt as basal debris and subsequently disperse it along their drifting trajectories.

To summarize, the clay minerals and coarse fraction (IRD) found in the surface sediments of Ross Sea were sourced from both the West Antarctica and East Antarctica. Provenance analysis exhibits a complementary spatial distribution pattern for these two sediment components. Considering that approximately 50% of the mass loss of the Antarctic Ice Sheet is through iceberg calving (Depoorter et al., 2013), these findings collectively indicate that both the WAIS and the EAIS, which discharge into the Ross Sea Embayment, have maintained a high level of dynamism under the modern climatic conditions.