Geochemistry and provenance of core sediments from the southwestern Okinawa Trough

XRD analysis of the core sediments reveals that the predominant clay minerals are illite, chlorite, and kaolinite, while the principal non-clay minerals include quartz, feldspar, calcite, and minor amounts of amphibole (Table 2). The results of major and trace element analyses for the cored sediments are presented in Tables 3–7. Major element compositions exhibit substantial variability across the different sampling stations. SiO₂ and Al₂O₃ are the most abundant oxides, jointly comprising approximately 63%–77% of the total sediment composition.

Manganese oxide (MnO), titanium dioxide (TiO₂), and phosphorus pentoxide (P₂O₅) occur in relatively low concentrations across all stations, while CaO and total iron (ΣFeO) are consistently higher than K₂O, MgO, and Na₂O. The concentrations of K₂O, MgO, and Na₂O are relatively uniform at stations ST2 and ST5, whereas more pronounced variability is observed at stations ST11, ST13, and ST15.

Among all cores, sediments from station ST13 display the highest average SiO₂ content, whereas their Al₂O₃ content is the lowest. This inverse relationship may reflect a relatively lower degree of chemical weathering. The lower Na₂O content compared to K₂O across the cores is interpreted as a result of the greater susceptibility of albite to chemical weathering relative to K-feldspar.

Fig. 2 presents the major element compositions of sediment cores from various stations, normalized to the Upper Continental Crust (UCC) values of Taylor and McLennan (1985). The results show that SiO₂ concentrations in all core samples are lower than the UCC reference. Notably, the decrease in SiO₂ content at Station ST15 is not accompanied by a corresponding increase in Al₂O₃, a pattern that may be linked to the elevated CaO content observed in these sediments.

Figure 2

UCC (Taylor & McLennan, 1985) normalized major element variations for cored sediments of this study. UCC, Upper continental crust.

Fig. 3 illustrates a positive correlation between ΣFeO and MgO with Al₂O₃, suggesting that the abundances of iron and magnesium are largely influenced by the presence of chlorite. Calcium, sodium, and potassium are primarily hosted in feldspar minerals. Variations in these elements across the cores likely reflect differences in source rock composition as well as the extent of chemical weathering. Although K₂O and Na₂O concentrations are relatively consistent across all cores, they are generally lower than UCC averages. This depletion may result from a reduced feldspar content in the source material or from feldspar breakdown during weathering.

Figure 3

Al₂O₃ versus ΣFeO and MgO plots for cored sediments of this study.

Elevated CaO concentrations at Stations ST5, ST13, and ST15, relative to UCC values, are likely due to the presence of biogenic carbonate material (e.g., shell fragments) in some sediment layers. The relatively high MnO and ΣFeO contents at Station ST15 may reflect localized hydrothermal influence. TiO₂ concentrations are relatively uniform across all stations. Still, they are higher than UCC levels, possibly due to the relative immobility of Ti during weathering, allowing for its progressive enrichment in the residual sediment fraction.

Cox et al. (1995) introduced the index of compositional variability (ICV) as a means of evaluating the variability of aluminum relative to other major oxides in sedimentary rocks. The ICV is defined as: ICV=( Fe2O3+K2O+Na2O+CaO+MgO +MnO+TiO2 )/Al2O3{\rm{ICV}} = \left( {{\rm{F}}{{\rm{e}}_2}{{\rm{O}}_3} + {{\rm{K}}_2}{\rm{O}} + {\rm{N}}{{\rm{a}}_2}{\rm{O}} + {\rm{CaO}} + {\rm{MgO}}} \right.\left. { + {\rm{MnO}} + {\rm{Ti}}{{\rm{O}}_2}} \right)/{\rm{A}}{{\rm{l}}_2}{{\rm{O}}_3}

The ICV values of sediment cores analyzed in this study (Fig. 4) show notably higher variability at Stations ST13 and ST15, suggesting significant fluctuations in clay mineral content. Higher ICV values are typically indicative of a lower abundance of clay minerals and a greater proportion of non-clay detrital input.

Figure 4

ICV values of the sediment cores in this study (+: the mean values of ICV for each station). ICV, Index of compositional variability.

Trace element concentrations of Co, Cr, Ni, V, and Sc are elevated in the ST2 core relative to other stations. These elements exhibit positive correlations with Al₂O₃, MgO, and total Fe (ΣFeO) (Figs. 5–7), suggesting that their distribution is closely associated with Fe–Mg-bearing minerals, particularly chlorite. A cross-plot of Cr/Th versus Sc/Th ratios in the sediment cores (Fig. 8) reveals a positive correlation, implying that the abundances of Cr, Sc, and Th are influenced by heavy mineral input and the presence of Fe–Mg-rich phases in the source rocks.

Figure 5

High field strength elements versus Al₂O₃ plots for cored sediments in this study.

Figure 6

High field strength elements versus MgO plots for cored sediments in this study.

Figure 7

High field strength elements versus ΣFeO plots for cored sediments of this study.

Figure 8

Cr/Th versus Sc/Th plots for the cored sediments of this study.

Data in Tables 3–7 indicate that Nb and Ta display relatively low mobility during weathering, whereas Th and U show broader compositional ranges, potentially reflecting differential geochemical behavior and partitioning during sedimentary differentiation and weathering processes.

Two samples—ST13 (44–46 cm) and ST15 (45–50 cm)—display elevated concentrations of Zr, Hf, and Y (Fig. 9), possibly indicating the presence of heavy minerals such as zircon. In contrast, cores from ST2 and ST5 exhibit higher Nb and Ta concentrations, which may reflect input from accessory phases such as rutile or ilmenite.

Figure 9

Nb versus Ta and Zr versus Hf, Y plots for cored sediments of this study.

Fig. 10 illustrates the relationship between REE concentrations and Al₂O₃ in the sediment cores. A general positive correlation suggests that REEs are likely adsorbed to clay minerals, which serve as key carriers during sediment transport and deposition.

Figure 10

REEs versus Al₂O₃ plots for cored sediments of this study. REEs, rare earth elements.

Fig. 11 shows chondrite-normalized REE patterns for samples with the highest and lowest REE concentrations in this study, alongside reference values for shale, quartzite, and the UCC. The REE concentrations of the cored sediments lie between those of shale and quartzite, with a modest enrichment in light REEs (LREEs) and a pronounced negative europium (Eu) anomaly. This REE signature is broadly consistent with that of the UCC and may reflect the removal or alteration of plagioclase during weathering or diagenesis.

Figure 11

Range of chondrite-normalized REE patterns for cored sediments in this study as compared with shale, quartzite, and UCC. Shale from Govindaraju (1989). Quartzite from Meisel et al. (1990). Upper continental crust from Taylor and McLennan (1985). REE, rare earth element; UCC, upper continental crust.

The geochemical behavior of REEs in marine sediments is closely linked to their interaction with fine-grained mineral phases, particularly clay minerals. In the present study, REE concentrations were found to exhibit a positive correlation with Al₂O₃ and other clay-associated elements, suggesting a key role for clay minerals in REE retention. This observation is consistent with recent findings by Cai et al. (2023), who demonstrated a significant positive correlation between total REY concentrations and clay mineral content in deep-sea sediments. Their study emphasized that kaolinite, in particular, acts as a major carrier of REYs, and that clay minerals—rather than only Fe–Mn oxides or biogenic apatite—play an active role in REY accumulation through surface adsorption and mineral–water interactions near the sediment-water interface.

Furthermore, Dou et al. (2010) analyzed the temporal evolution of clay mineral assemblages in the central Okinawa Trough over the last 28 ka, showing that variations in illite, smectite, and kaolinite abundances reflect changes in terrigenous input, driven primarily by the East Asian monsoon and shifts in fluvial sources. While their study did not directly analyze REEs, the compositional variability of clay minerals suggests that source-related changes could significantly influence the REE adsorption capacity and fractionation patterns, especially along vertical sediment profiles.

In addition, sediment transport time and depositional dynamics also affect the availability and binding of REEs to clay particles. Li et al. (2016) used Nd isotope systematics to constrain the timescales of lithogenic sediment transport from the Yangtze River to the Okinawa Trough, revealing that fine-grained particles (including clays) undergo delayed deposition due to extended transport distances. This prolonged residence time potentially enhances the opportunity for REEs to be adsorbed onto clay surfaces during suspension in seawater.

Taken together, these findings support the interpretation that REE enrichment in the studied sediments results from a combination of clay mineral abundance, compositional variation related to sediment provenance, and prolonged particle–seawater interaction during sediment transport. Therefore, the spatial and vertical variability of REE concentrations in the core sediments may reflect both mineralogical control and dynamic depositional conditions in the Okinawa Trough.

Owing to its geographical location, geomorphological setting, and prevailing hydrodynamic conditions, the southernmost segment of the SOT serves as an efficient sediment sink, receiving input from both Taiwan and the East China Sea continental shelf (Lee et al., 2004). Geochemically stable elements such as REEs, Th, Sc, and Co are resistant to alteration during weathering, transport, diagenesis, and metamorphism, and are therefore commonly employed as proxies for determining sediment provenance. The La–Th–Sc ternary diagram is particularly useful for discriminating between felsic and mafic source rocks (Cullers, 1994; Nyakairu & Koeberl, 2001).

As shown in Fig. 12, the sediment samples analyzed in this study plot within the compositional field representative of mixed sources—including clay, silt, sand, and gravel—similar to sediments offshore northern Taiwan (Chao & Chen, 2003). Roser and Korsch (1986) proposed a classification scheme based on major element geochemistry, delineating four provenance types: felsic igneous, quartzose sedimentary, mafic igneous, and intermediate igneous sources. In Fig. 13, the REE distribution patterns and associated trace element signatures suggest that a portion of the sediment material is indeed derived from the weathering of volcanic rocks on Guishan Island, as evidenced by enrichment in LREEs and elevated Th and La concentrations. However, a more comprehensive interpretation must also consider the multi-source nature of the sediments and the influence of post-depositional processes. The normalized REE patterns indicate not only a volcanic contribution but also the presence of terrigenous detritus, likely sourced from the Lan-Yang River and the East China Sea continental shelf, as supported by the three-end-member mixing model results. Samples showing relatively flat REE patterns with slight Eu anomalies could be indicative of sediment mixing between felsic continental sources and local mafic volcanic input. Additionally, variations in (La/Yb)_N and (Gd/Yb)_N ratios among the samples suggest differential sediment contribution from sources with contrasting degrees of weathering and sorting during transport. It is also important to recognize that REE mobility can be influenced by sedimentary redox conditions, grain size sorting, and clay mineral associations. The positive correlation between REEs and Al₂O₃ in the same samples, as discussed earlier, further implies that fine-grained clay minerals serve as major carriers of REEs and that authigenic or diagenetic processes may have modified the primary geochemical signature. Taken together, the patterns presented in Fig. 13 should be interpreted not solely as evidence of Guishan volcanic input, but as the integrated result of multi-source sediment supply, mechanical sorting during transport, chemical weathering intensity, and early diagenetic overprinting under varying bottom-water conditions in the Okinawa Trough.DF1=−1.773TiO2+0.607Al2O2+0.76tFeO−1.5MgO+0.616CaO+0.509Na2O−1.224K2O−9.09DF2=0.445TiO2+0.07Al2O3−0.25tFeO−1.142MgO+0.438CaO+0.475Na2O+1.426K2O−6.861\matrix{ {{\rm{DF}}1\; = \; - 1.773\;{\rm{Ti}}{{\rm{O}}_2} + \;0.607\;{\rm{A}}{{\rm{l}}_2}{{\rm{O}}_2} + \;0.76\;{\rm{tFeO}}\; - \;1.5\;{\rm{MgO}}\; + \;0.616\;{\rm{CaO}}\; + \;0.509\;{\rm{N}}{{\rm{a}}_2}{\rm{O}}\; - \;1.224\;{{\rm{K}}_2}{\rm{O}}\; - \;9.09} \hfill \cr {{\rm{DF}}2\; = \;0.445\;{\rm{Ti}}{{\rm{O}}_2}\; + \;0.07\;{\rm{A}}{{\rm{l}}_2}{{\rm{O}}_3}\; - \;0.25\;{\rm{tFeO}}\; - \;1.142\;{\rm{MgO}}\; + \;0.438\;{\rm{CaO}}\; + \;0.475\;{\rm{N}}{{\rm{a}}_2}{\rm{O}}\; + \;1.426\;{{\rm{K}}_2}{\rm{O}}\; - \;6.861} \hfill \cr }

Figure 12

La–Th–Sc variations for cored sediments (variation fields modified from Cullers, 1994). Shale and limestone from Govindaraju (1989). Greywacke from Condie (1993). Quartzite from Meisel et al. (1990).

Figure 13

DF1 versus DF2 plots for cored sediments in this study (variation fields modified from Roser & Korsch, 1986).

Fig. 14 illustrates that most of the cored sediments from this study are situated within the active continental margin field, supporting the interpretation that the SOT is currently undergoing early-stage back-arc extension.

Figure 14

K₂O/Na₂O versus SiO₂ plots for cored sediments in this study (variation fields modified from Roser & Korsch, 1986).

The La–Th–Sc plots (Fig. 12) indicate that the cored sediments analyzed in this study are derived from mixed provenance sources. Based on the geochemical signatures of these sediments, three end-member compositions were selected for provenance modeling: (1) cored sediments from the Lan-Yang area in northeastern Taiwan, (2) sediments from the East China Sea continental shelf, and (3) volcanic rocks from the Okinawa Trough, as reported in previous studies (Chen & Kato, 1989; Chen & Lin, 1994; Lee, 2002; Table 7). These end members were incorporated into a three-component mixing model developed by Ho and Chen (1996) to estimate the relative contributions from each source.

To simplify the modeling approach, it is assumed that the sediments reflect a single mixing event involving these three distinct end members. It is worth noting that the geochemical signature produced by a single mixing event may be similar to that generated through polycyclic sedimentary processes. The model results are presented in Table 8. According to the calculations, the dominant source of sediments is the Lan-Yang area (averaging 54.63%), followed by the East China Sea continental shelf (24.12%) and volcanic input from the Okinawa Trough (21.25%). These results support the interpretation that fluvial input from Taiwan and sediment from the East China Sea shelf represent the primary sources for sediment accumulation in the SOT.

Li et al. (2009) proposed that the Lan-Yang River in northeastern Taiwan is a major sediment contributor, transporting approximately 10 million metric tons of material annually to the adjacent coastal and offshore regions. Some of this material is likely transported northward by ocean currents into the southwestern Okinawa Trough. Huh et al. (2006) noted that sedimentation rates in the SOT generally decrease with increasing water depth and distance from the coastline, except in localized bathymetric depressions below the 1400 m isobath, where episodic turbidite deposition enhances sediment accumulation. The principal supply of detrital sediment to the region is derived from fluvial runoff originating in Taiwan’s central mountain ranges, driven by steep topographic gradients and intense monsoonal rainfall (Hsu et al., 2004). In northern Taiwan, sediment discharge is primarily funneled through the Lan-Yang River, which debouches into the sea adjacent to the SOT.

Detrital zircon U–Pb geochronology from Zhu and Zeng (2022) confirms that over the past 700 years, sediment accumulation in the SOT reflects contributions from Taiwanese rivers, the East China Sea shelf, and the Yangtze River. In the present study, the proportion of Lan-Yang-derived material in the sediment cores generally decreases with increasing distance from the northeastern Taiwan coastline (Fig. 15).

Figure 15

Variation diagram of the contribution proportion of core sediments from the Lan-Yang area at distances from the coastline of Taiwan.