ORIGINAL ARTICLE
Petrogenesis of Josoji rhyolite and intrusive rocks of NE Shimane Peninsula, SW Japan: implication on Miocene volcanism related to back-arc rifting in San’in region
2024 年 119 巻 1 号 論文ID: 230908
詳細
2024 年 119 巻 1 号 論文ID: 230908
Shimane Peninsula in the San’in region, SW Japan, has preserved a wide distribution of igneous rocks related to back-arc rifting in the Miocene. We investigated the petrography, geochemistry, and Sr-Nd isotope systematics of rhyolite lavas (Josoji rhyolite: 18-15 Ma) and basaltic to dacitic intrusions of Stage I (∼ 14 Ma) and II (∼ 13 Ma) intrusive rocks. They are classified as medium-K magma series. The Josoji rhyolite and Stage I rocks show chemical compositions of arc-type signatures, whereas the Stage II rocks show elevated Nb and Ta abundances suggesting weaker arc signatures. The geochemical characteristics indicate that the Josoji rhyolite was produced by partial melting of arc-type basalt under lower to middle crustal conditions. The arc signatures of the Stage I rocks were inherited from a remnant metasomatized lithosphere formed by the subduction of the Pacific Plate before the opening of the Japan Sea. The Stage II rocks were formed from a fertile magma that might be produced by the melting of upwelling asthenospheric mantle. We conclude that various magmatic processes took place during and just after the back-arc rifting development in the San’in region.
Current insights in elucidating the formation of the Sea of Japan propose the initiation of rifting along the eastern margin of the Eurasian plate in the late Oligocene (e.g., Shuto et al., 2015). The Southwest Japan arc underwent a rotational shift approximately from 18 to 16 million years ago (e.g., Hoshi et al., 2015). The coastal region of Shimane Peninsula in the San’in region, southwest Japan, exhibits widespread occurrences of Miocene igneous and sedimentary rocks, attributed to the expansion of the Japan Sea. The geological units formed during the evolution of this back-arc basin, transitioning through freshwater, brackish water, and marine environments (e.g., Yamauchi et al., 1980; Kano and Yoshida, 1985; Takayasu et al., 1992; Kano et al., 1994; Kogane et al., 1994; Sakai et al., 2013).
Miocene volcanic and intrusive rocks were developed along the Shimane Peninsula (Kano and Yoshida, 1985; Kano et al., 1994). They represent a significant component of the magmatic system that operated during to just after the opening of the Japan Sea, a back-arc basin in the San’in region (e.g., Iwamori, 1989; Kimura et al., 2005). Several studies were held including sedimentology, stratigraphy, and environmental investigations for the sake of understanding the formation system of the Japan Sea around the San’in region (e.g., Kano and Nakano, 1985; Morris et al., 1990; Kimura et al., 2005; Sakai et al., 2013). However, systematic geochemical examinations of extrusive and intrusive igneous rocks for understanding the magmatism in this region are somewhat inadequate.
In this aspect, this study presents petrography, major, trace and rare earth element geochemistry, and Sr-Nd isotope compositions of the Josoji rhyolite and basaltic to andesitic intrusive rocks in the northeast Shimane Peninsula (Fig. 1). These are geologically regarded as voluminous igneous activities that widely dominate the Shimane Peninsula. Hence, we consider the petrogenesis of these well-preserved examples of the Miocene volcanisms related to the opening of the Japan Sea in the San’in region. Even though the magma genesis of the intrusive rocks has been suggested by Miyake (1994) based on the bulk rock chemistry, we present new and detailed geochemical data to provide a comprehensive understanding.
The Josoji rhyolite and basaltic to dacitic intrusive rocks are exposed extensively along the Shimane peninsula and exhibit widespread distribution in its northeastern flank (Fig. 1; Kano and Yoshida, 1985; Kano et al., 1994). The Josoji rhyolite is a member of the Josoji Formation, which is composed mainly of geologically intercalated rhyolite and black mudstone. With a maximum thickness of approximately 800 m, this formation stands as a representative marine unit within the Shimane Peninsula. The Josoji rhyolite comprises of lavas and pyroclastic deposits that have accreted between 18-15 Ma (Kano et al., 1998) (Fig. 1).
The basaltic to dacitic intrusive rocks intrude into the various stratigraphic units of the Shimane Peninsula that formed between freshwater and marine environments. The lithology predominantly consists of basic to intermediate volcanic or hypabyssal rocks, accompanied by subordinate acidic formations. They are exposed as laccoliths sills, and dikes that yielded ages of 14.2-13.4 Ma (Morris et al., 1990). The intrusive rocks were formed as a response to the opening of the Japan Sea along with other volcanic rocks in the Oki Zone (see inlet map in Fig. 1) (Pouclet et al., 1995; Kimura et al., 2005; Hoshi et al., 2015; Shuto et al., 2015). Miyake (1994) stratified intrusive rocks into three distinct stages (Stages I, II, and III) based on geological characteristics and bulk rock chemistry (Fig. 1). During Stage I, numerous sills of basalt to dacite compositions intruded into unconsolidated sediments. Stage II presented the development of several small submarine volcanic edifices featuring compositions of basalt, andesite, and dacite. Finally, Stage III involved the eruption of sheet flows comprising of basalt and andesite.
Previous research about the Miocene volcanisms, especially the intrusive rocks, suggested that they are subalkaline rocks with arc signatures inherited from metasomatized continental lithosphere (Miyake, 1994; Kimura et al., 2005). For this study, intrusive rocks from Stage I (∼ 14 Ma) and Stage II (∼ 13 Ma) are examined along with the Josoji rhyolite (18-15 Ma).
Samples were collected from the Josoji rhyolite and the intrusive rocks of the Stage I and II (Fig. 1). The freshest and least-altered samples were picked for analyses of whole-rock geochemical compositions, and Sr and Nd isotopic ratios. All analyses were carried out in the Department of Earth Science, Shimane University.
Analyses of whole-rock major and selected trace elements (Ba, Ce, Cr, Nb, Ni, Pb, Rb, Sr, Th, Y, and Zr) were performed on fused glass beads using a Rigaku ZSX Primus IV X-ray fluorescence (XRF) spectrometer. The glass beads were prepared by mixing a 1:2 ratio of powdered samples and lithium tetraborate-lithium metaborate (LiBO2-Li2B4O7) flux. The analytical procedure by Kimura and Yamada (1996) was followed. Loss on ignition (LOI) measurements were found by the weight loss of powdered samples after heating at 850 °C for two hours. Other trace elements including rare earth elements (REEs) (U, Ta, La, Pr, Sr, Nd, Hf, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu) were measured using an inductively coupled plasma-mass spectrometry (ICP-MS; Aglient 7500a). A 0.05 g fragment of crushed glass beads was used for the sample preparation, which was decomposed by stepwise heating and evaporation method using a mixture of hydrofluoric acid (HF), nitric acid (HNO3), and perchloric acid (HClO4). The samples were then diluted with HNO3 before the analysis. The analytical procedure followed Kamei (2016). The Sr and Nd isotope measurements were performed using a Thermal Ionization Mass Spectrometry (TIMS) Finnigan MAT-262. The detailed procedure for separating Sr and Nd from the powdered samples is described in Kagami et al. (1987). The measured Sr and Nd isotopic ratios were normalized to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219 and were corrected referring to the standard NIST SRM987 and JNdi-1 values of 87Sr/86Sr = 0.710241 (Kagami et al., 1989) and 143Nd/144Nd = 0.512115 (Tanaka et al., 2000), respectively.
The petrographic observations of the Josoji rhyolite exhibit gray-colored, moderately altered, porphyritic to aphanitic textures. Their phenocrysts are predominantly plagioclase that are altered mostly to sericite, with a minor amount of biotite and opaque minerals. The samples from the Stage I and II intrusive rocks have no noticeable petrographic differences. They are mostly basalt to dolerite which are moderately altered. Dacites were not recognized within our study area. In basalt, the common phenocrystic minerals are plagioclase, clinopyroxene, and olivine. The plagioclase is altered to sericite in some samples, and olivine is often replaced with iddingsite. The dolerite is similar in mineral composition to basalt, but with less frequent occurrence of olivine. However, the Stage I and II intrusive rocks exhibit mineral variations indicative of compositions ranging from basic to intermediate. Certain samples show a reduction in olivine content, accompanied by an increase in plagioclase and clinopyroxene. Consequently, the lithology undergoes a transition to andesitic or porphyritic characteristics.
The whole-rock geochemical data of major and trace elements are given in Supplementary Table S1 (Supplementary Tables S1-S4 are available online from https://doi.org/10.2465/jmps.230908). Selected elements are expressed in the Harker diagrams as a function of SiO2 contents (Fig. 2). The data analyzed by Miyake (1994) are also plotted in this figure. In the K2O diagram (Fig. 2A; LeMaitre et al., 2002), almost all the samples are classified as medium-K magma series. In the trace element diagrams (Figs. 2B and 2C), two distinct compositional trends are observed for the Stage I and II intrusive rocks. Moreover, the Josoji rhyolite plots are within the same trend as those of the Stage I rocks. Such trends are also well expressed in the N-MORB-normalized multi-element spider diagram (Fig. 3) that shows somewhat more depleted patterns of the Josoji rhyolite and Stage I rocks than those of the Stage II rocks. The patterns of the Josoji rhyolite and Stage I rocks revealed arc signature characterized by the enrichment in large-ion lithophile elements (LILE) and light REEs relative to depletion of high-field-strength elements (HFSE) and heavy REEs (e.g., Tatsumi et al., 1986; Wilson, 1989). In contrast, the Stage II rocks have less pronounced arc signatures as shown by weaker Nb and Ta depletions.
The isotopic data of the Sr and Nd are given in Supplementary Table S2. The initial Sr and Nd isotopic ratios were calculated based on their respective isotopic ages, which are 15, 14, and 13 Ma for the Josoji rhyolite, Stages I, and II intrusive rocks, respectively. In the εNd (t) versus initial Sr isotopic ratio [87Sr/86Sr (i)] diagram, all the analyzed samples show narrow isotopic values in weakly depleted composition (Fig. 4).
The chemical composition of rocks undergoes changes through weathering and alteration processes. The exposed rocks in the Shimane Peninsula are frequently subjected to weathering. The samples chosen for chemical analysis were carefully selected during field and microscopic observations to minimize the influence of weathering or alteration. Despite this selection, the impact of weathering and alteration on the analysis data remains uncertain. Thus, we assessed these effects using the Chemical Index of Alteration (CIA: Nesbitt and Young, 1982) and the Alteration Index (AI: Ishikawa et al., 1976). The CIA quantifies the weathering degree of the rock, with values below 60 indicating minimal weathering (Nesbitt and Young, 1984; Pandarinath, 2022). The AI reflects the degree of sericite or chlorite alteration within the rock with unaltered rocks ranging from 20 to 60 (Hodges and Manojlovic, 1993; Large et al., 2001; Shikazono et al., 2008; Pandarinath, 2022). The representative weathering profiles of granite (Nesbitt and Markovics, 1997) and basalt (Dekayir and EL-Maataoui, 2001) illustrates an increase in these indices corresponding to the degree of weathering (Fig. 5). For the granite samples numbered 8 through 15 in the weathering profiles (Fig. 5), Nesbitt and Markovics (1997) observed notable changes in trace element compositions. Importantly, our investigation confirms that the Josoji rhyolite and intrusive rocks analyzed in this study exhibit minimal susceptibility to weathering and lack of geochemical alteration, as indicated in the CIA versus AI (Fig. 5).
N-MORB normalized multi-element spider diagram of the Josoji rhyolite reveals a typical volcanic arc signature (Fig. 3). In addition, the slightly depleted Sr-Nd isotopic composition does not rule out the possibility of igneous rocks originating in a volcanic arc environment (Fig. 4). The genesis of acidic magmas in arc systems can be attributed primarily to two crucial factors either melting of mafic rocks or fractional crystallization of mafic magma (e.g., Moyen et al., 2021).
The Josoji rhyolite is widely distributed over an area extending approximately 50 km on the Shimane Peninsula (e.g., Kano and Yoshida, 1985; Kano et al., 1994; Kano et al., 1998). In contrast, it is notable with the negligible amount of occurrence of contemporaneous basaltic to andesitic rocks (e.g., Kano and Yoshida, 1985; Kano et al., 1994; Kano et al., 1998). Ideally, when rhyolite is derived from basaltic magma through fractional crystallization, the original basaltic magma must constitute at least twice the weight of the resulting rhyolite (e.g., Nandedkar et al., 2014; Caricchi and Blundy, 2015). If the Josoji rhyolite is a product of basaltic magma differentiation, it is peculiar that such enormous parental magma intrusions are hardly distributed. These geological occurrences suggest that the Josoji rhyolite is formed by partial melting of mafic rocks rather than fractional crystallization from mafic magmas.
Beard and Lofgren (1991) conducted melting experiments of mafic rocks at 3 and 6.9 kbar, representing lower to middle crustal conditions. Under both dehydration and water-saturated conditions, they observed the generations of high-silica melts resembling the Josoji rhyolite. Discrepancies in the conditions led to variations in residual mineral assemblages, resulting in differences in melt composition. Specifically, plagioclase tended to precipitate in the residue during dehydration melting, whereas many amphiboles remained in the water-saturated melting. These effects occur in differences in the Al2O3 composition of the melts (Beard and Lofgren, 1991) (Fig. 6). The Al2O3 compositions of experimental melts suggest that the Josoji rhyolite was formed through the dehydration melting of mafic crust (Fig. 6).
A partial melting model can be constructed by considering older or coexisting mafic rocks as the starting source material of the Josoji rhyolite. When we find a potential candidate as the source material of an igneous rock in the field, the candidate must be geologically older. Although suitable lithology has not been identified in the San’in region, Kori basalt from Oki region can be used as it has been dated at 19-18 Ma (Uto et al., 1994) (Fig. 1). This Kori basalt is arc-type low-alkaline tholeiite basalt that intruded during the opening of the Japan Sea (e.g., Kimura et al., 2005; Hoshi et al., 2015).
Using the modal batch melting equation of Shaw (1970), the calculation for the partial melting of the Kori basalt was performed. The two dehydration melting conditions under 3 and 6.9 kbar were used based on Beard and Lofgren (1991). The mode of the residual phases, degrees of the melting conditions, and the values of the mineral-melt partition coefficient (D) are given in Supplementary Tables S3 and S4. The results of the calculations are presented in N-MORB-normalized multi-element spider diagram (Fig. 7). The diagram shows that the calculated patterns for the two melting conditions resemble the Josoji rhyolite samples.
Calculated Rb, Ba, and K concentrations may seem to be slightly higher than those found in the Josoji rhyolite (Fig. 7). In this investigation, although the Kori basalt is employed as a representative example of the source material, based on the calculated results, it is considered that the actual source material might had relatively lower concentrations of Large Ion Lithophile (LIL) elements, such as Rb, Ba, and K, as compared to the Kori basalt.
This result confirms that the middle to lower continental crust, relatively similar to the Kori basalt in composition, has been exposed to an upwelling heat source during the back-arc basin rifting that led to crust melting and the Josoji rhyolite magmatism.
Petrogenesis of Stage I and Stage II intrusive rocksDistinct geochemical characteristics for the Stage I and II intrusive rocks are observed in the trace element Harker diagrams (Figs. 2B and 2C) and N-MORB normalized multi-element spider diagram (Fig. 3). Miyake (1994) suggested that they originated from the melting of two different mantle sources, that is, the Stage I rocks formed from a depleted source, whereas the Stage II rocks from a less depleted source. However, their geochemical differences may also be explained by other possible magmatic processes such as, (1) different melting degrees of a single mantle source (i.e., lower melting degree for the genesis of the Stage II rocks), or (2) different degrees of input of crustal component (i.e., higher crustal contamination in the Stage II rocks). In this study, we examined these issues based on the analyzed data.
The former possibility can be ruled out for the following reasons. Igneous rocks from a single source are generally expected to have similarity in the spider diagram patterns because the residual mineral assemblage does not change significantly during the melting. However, significant differences in Nb and Ta concentrations are evident between the Stage I and II intrusive rocks, nevertheless, other trace element compositions exhibit comparatively minor distinctions (Fig. 3). This cannot be explained by differences in the melting degree. Furthermore, the SiO2 content of the least SiO2 sample in both the Stage I and II intrusive rocks is not remarkably different (Fig. 2). The SiO2 content generally decreases with increasing melting degrees in a single source melting. The similarity in SiO2 content is inconsistent with the model that the two magmas generated from the same source with different degrees of melting.
The latter possibility is also excluded from their isotopic composition. The isotopic values for the Stage I and II intrusive rocks exhibit the same trend in the εNd (t) versus 87Sr/86Sr (i) diagram (Fig. 4). The Stage II rocks should have lower εNd (t) and higher 87Sr/86Sr (i), toward the continental crust affinity, if they experienced more significant crustal contamination than the Stage I rocks (e.g., Carter et al., 1978). Consequently, the crustal contamination would not have played a significant role in causing the geochemical differences between the two groups of rocks (Fig. 3). Therefore, two different mantle sources are the most likely scenario to explain the difference in chemical composition between the Stages I and II intrusive rocks, and our new data support the petrogenetic model of Miyake (1994).
On the other hand, it is interesting to note that the Stage I and II intrusive rocks have similar isotopic compositions despite their different source materials. A relevant analogy can be drawn from the Cretaceous granitic intrusions, widely distributed across 200 × 200 km area on Kyushu Island in the southwestern Japan arc. Although almost all the granitic intrusions exhibit similar initial Sr and Nd isotopic compositions (e.g., Kagami et al., 1992; Owada et al., 1999; Muraoka et al., 2020; Yuhara et al., 2022), their origins are diverse from the mafic lower crust (e.g., Eshima et al., 2023) to the tonalitic middle crust (e.g., Kamei, 2002; Kamei et al., 2009). Hence, it is important to recognize that igneous rocks with similar initial Sr and Nd isotopic compositions do not necessarily share a common source material.
In the following sections, we delve into the differentiation process of magma within the Stages I and II intrusive rocks. The considerable SiO2 variations in both types of intrusive rocks (Fig. 2) are suggested by Miyake (1994) to be due to the crystal fractionation of plagioclase, clinopyroxene, and olivine. They focused on the nearly constant ratio of incompatible elements/La within each of the stages as a function of La contents. The distribution coefficients for the incompatible elements between crystal (plagioclase, clinopyroxene, and olivine) and melt are much smaller than one (Irving, 1978). They argued that the incompatible elements/La ratio is expected to remain nearly constant during crystal fractionation (Minster and Allègre, 1978).
In the present study, we will verify this assertion based on our data. Another plausible consideration is the potential influence of crustal assimilation during magma differentiation (e.g., Rudnick and Gao, 2003), given the widespread intrusion of these rocks into sedimentary formations at the Shimane Peninsula. The trace elements sensitive to crustal contamination (e.g., Th and Zr; Wang et al., 2013) exhibit a correlation with SiO2 contents (Figs. 2B and 2C). Notably, the Th and Zr concentrations of the least SiO2 sample in the Stage II rocks exhibit a slight elevation compared to that in the Stage I rocks (Figs. 2B and 2C). Furthermore, the trends of the Stage II rocks are clearly higher than those of the Stage I rocks. This allows for the interpretation that both stages of rocks may have undergone crustal assimilation, with Stage II rocks exhibiting a more significant influence of the contamination.
However, no consistent trend is observed for 87Sr/86Sr (i) to increase with increasing SiO2 in both stages. (Fig. 8). It is known that crustal assimilation leads to an increase in 87Sr/86Sr (i) with rising SiO2 content (e.g., Arakawa, 1990). Therefore, we interpret the considerable SiO2 variations observed in the Stage I and II intrusive rocks as a consequence of the crystal fractionation process, unaffected by the influence of crustal assimilation.
Moreover, by referring to our geological and petrographical investigations, the lack of significant xenoliths and xenocrysts is recognized in rocks from both stages. This finding reinforces the notion that crustal contamination played a non-essential role in the magmatism of Stage I and II intrusive rocks. The prevalence of plagioclase, clinopyroxene, and olivine as primary phenocrysts in petrographic examinations of basaltic rocks strongly supports their role as major fractionated phases. These outcomes align with the insights presented by Miyake (1994), supporting the argument that magma in both stages experienced a compositional change from basalt to dacite through the crystal fractionation process.
Igneous activities of the Josoji rhyolite and the Stage I and II intrusive rocksOur geochemical investigation suggests that the Josoji rhyolite (18-15 Ma) originated from the partial melting of a mafic crust displaying a volcanic arc-type composition. The source material likely shares similarities with the Kori basalts distributed in the Oki Islands. The Stage I rocks (∼ 14 Ma) exhibit volcanic arc signatures, indicative of their likely formation through the partial melting of an arc-type metasomatized mantle. These interpretations align with previous studies suggesting that some Miocene volcanic rocks in the San’in region have roots in a volcanic arc-type source (Fujibayashi et al., 1989; Miyake, 1994). In contrast, the Stage II rocks (∼ 13 Ma) are inferred to have originated from the partial melting of a fertile asthenospheric mantle.
During the Miocene in the San’in region, the subduction of the Philippine Sea Plate had not yet started (e.g., Kimura et al., 2005). Therefore, the volcanic arc-type magmatic activities observed in the Josoji rhyolite and the Stage I rocks likely resulted from the melting of a remnant of an arc-type lithosphere. This lithosphere probably formed before the Miocene, initiated by the subducting Pacific Plate slab beneath southwest Japan before the opening of the Japan Sea (Yamaoka and Wallis, 2023). Conversely, the presence of the more enriched Stage II rocks implies melting of the upwelling fertile asthenospheric mantle following the 18-14 Ma volcanic arc-type activities.
In conclusion, the magmatic events associated with the Josoji rhyolite and the Stage I and II intrusive rocks describe a diverse array of activities, each occurring independently within the context of thermal events associated with the Japan Sea opening.
We are grateful to Auer Andreas and Members of Geoscience Seminar, Shimane University, for many helpful comments and their encouragement. Thanks go to Sena Nakayama, Satomi Hirai, and Marwa Saif Al-Busaidi for helpful discussion and advice during the field survey and analytical works. Thanks are due to Yuji Ichiyama and anonymous reviewer for their critical reviews and many constructive comments which helped to significantly improve the manuscript. We would like also to express appreciation to Masao Ban for his many thoughtful comments and manuscript handling as Editor. S. Al-Busaidi acknowledges MEXT for financial support during the master research. This study was supported by JSPS KAKENHI Grant Numbers JP18H01313, JP19K21651, and JP22K03780 to A. Kamei.
Supplementary Tables S1-S4 are available online from https://doi.org/10.2465/jmps.230908.
