Timescales of Magma Mixing Beneath the Iheya Ridge, Okinawa Trough: Implications for the Stability of Sub-Seafloor Magmatic Systems

Abstract

Submarine volcanic eruptions can be destructive for marine environments and resources. Magma mixing is considered to be an important trigger for volcanic eruptions. Determining the magma residence time from mixing to eruption is conducive to assessing the stability of magmatic systems, especially beneath the seafloor where in situ volcano monitoring is inaccessible. Here, we estimated the timescale of magma mixing beneath the Iheya Ridge, Okinawa Trough, which is characterized by pervasive magma mixing. We focused on andesitic and rhyolitic magma generated by basalt–rhyolite mixing and rhyolite–rhyolite mixing, respectively. By taking advantage of the Mg diffusion chronometry, we showed that the andesitic magma resided in the magma chamber for very short time (~0.1–0.3 years), whereas the residence time of the rhyolitic magma was much longer (~80–120 years). The different times might be in part related to the different rheology of the mixed magmas. The short residence time of the andesitic magma suggested efficient magma mixing that allowed the andesites to be erupted, which may explain the appearance of scarce andesites in basalt–rhyolite dominant settings. However, the rapid mixing and eruption of magma is a disadvantage for the development and preservation of seafloor hydrothermal resources. Therefore, we suggest that the stability of sub-seafloor magma systems must be evaluated during the assessment of seafloor sulfide resources and mining prospects.

1. Introduction

Sub-seafloor magma systems are huge energy reservoirs, which are not only the most important heat source for seafloor hydrothermal activities but can also supply volatiles and/or metals to hydrothermal systems [1,2,3]. However, submarine volcanic eruptions can be extremely destructive. For example, they can trigger serious natural disasters such as earthquakes and tsunamis [4,5]. In addition, huge amounts of heat and toxic gases emitted from the seafloor could have catastrophic impacts on marine environments and ecosystems [3,6]. Furthermore, they can also damage previously formed hydrothermal systems, which is a potential threat that should not be ignored in the assessment of seafloor sulfide resources and mining feasibility. Therefore, it is essential to evaluate the stability of sub-seafloor magma systems and assess the residence times of magmas prior to eruption.

While it is difficult to monitor magmatic activity beneath the seafloor in situ, the instabilities of a deep magma reservoir may be partly reflected by complex magmatic processes such as magma mixing, which can be recorded by the mineralogy of erupted lavas [7,8,9,10,11,12]. The timescales of magmatic processes can be estimated by diffusive re-equilibration within crystals [13,14,15,16,17,18,19,20,21].

The Okinawa Trough (OT) is a young back-arc basin that is characterized by complex volcanic systems, widespread seafloor hydrothermal systems, and extremely high heat flow values [22,23,24]. Widespread magma mixing is suggested to have occurred in the shallow magma chambers of the OT [25,26]. However, the timescales of magma mixing have not been well constrained. In this study, we focused on the magma system beneath the Iheya Ridge, where basalt–rhyolite mixing and rhyolite–rhyolite mixing have been recorded by the chemical zonation of minerals [26], and estimated the residence times of magmas from mixing to eruption using diffusion chronometry.

2. Geological Setting and Samples

The OT is located in the eastern margin of the East China Sea Shelf (Figure 1). It is a newly developed back-arc basin, and its initial extension can be traced back to the middle-to-late Miocene in response to the subduction of the Philippine Sea plate [27]. The OT is generally divided into three segments from the north to the south, namely the NOT, MOT, and SOT. Several central grabens, which represent the active back-arc spreading centers, have developed in the MOT and SOT. The SOT and MOT are also characterized by widespread hydrothermal activities and complex volcanic systems [23,28].

The Iheya Ridge is a volcanic ridge located in the central axis of the Iheya Graben, which also represents the center of an anomalous volcanic zone (known as the VAMP area) (Figure 1). The volcanic rocks in the Iheya Ridge consist of basalts, andesites, and rhyolites. Multi-level magma storage and pervasive magma mixing are the key features of the magma plumbing system beneath this area [26]. An active hydrothermal field (CLAM) has been reported on the Iheya Ridge.

The samples used in this study included andesite (T5-2) and rhyolite (C11) collected from the Iheya Ridge. The detailed whole-rock geochemistry and mineralogical compositions of the Iheya Ridge have been reported previously [26]. Andesite T5-2 contains three groups of minerals that are in equilibrium with (1) rhyolitic melt, including low-Ca plagioclase, low-Mg orthopyroxene, and clinopyroxene with reverse zonations, (2) basaltic melt, including high-Ca plagioclase and high-Mg clinopyroxene with normal zonations, and (3) andesitic melt, including unzoned clinopyroxene and weakly zoned orthopyroxene (Figure S1). This indicates that the andesite used in this study was a mixture of basaltic and rhyolitic melts. Rhyolite C11 contains two groups of minerals that are in equilibrium with (1) host melt, including reversely zoned plagioclase and orthopyroxene, and (2) a less-evolved rhyolitic melt, including normally zoned plagioclase, unzoned orthopyroxene, and clinopyroxene (Figure S2). This suggests that the rhyolite used in this study was formed from a hybridization of rhyolitic melts with different evolution degrees.

3. Methods

We estimated the magma residence time from mixing to eruption beneath the Iheya Ridge using Mg diffusion in olivine, orthopyroxene, and plagioclase crystals. We chose olivine and reversely zoned orthopyroxene for andesite T5-2 and reversely zoned orthopyroxene and plagioclase and normally zoned plagioclase for rhyolite C11. These minerals all have euhedral crystal shapes, except for olivine (Figure 2 and Figure 3). The olivine was corroded due to it mixing with rhyolitic melt (Figure 2a). No recrystallization of the olivine occurred because the hybrid andesitic melt destabilized the olivine. Instead, some tiny pigeonites were found to surround the olivine due to the reaction between olivine and rhyolitic melt (Figure 2a) [26].

Because the Mg concentrations in olivine and orthopyroxene are proportional to the greyscale of the back-scattered electron (BSE) images of these minerals, the Mg profiles of the olivine and orthopyroxene in this study were calibrated from the greyscale values of high-resolution BSE images [29]. This method can obtain high-resolution Mg profiles. The BSE images were captured using an electron probe (EMPA) at the Ocean University of China under a beam current of 20 nA and at an accelerating voltage of 15 kV. The greyscale values of the BSE images were extracted using the ImageJ software. The greyscale values were then converted into Mg concentrations based on several (~4–6 points) EMPA analyses along the profiles. The Mg concentrations in the plagioclase were analyzed using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. The detailed analytical methods for EMPA and LA-ICP-MS are given in [26]. The analytical results are given in Data S1.

The Mg profiles of the crystals and their evolution with time can be modelled using the equation:

$$\partial C_i/\partial \mathrm{t}=D_i\times {\partial }^2{C}_i/\partial {\mathrm{x}}^2$$

(1)

where C_i and D_i represent the concentration and diffusion coefficient of the element i, respectively. The diffusion coefficients and P–T conditions used in the modelling are listed in

Table 1. Mg diffusion coefficients in olivine, orthopyroxene, and plagioclase.

	DMg (μm2/s)	Orientation	P (kbar) 1	T (°C) 1	fO2 (△NNO) 1	XMg/XAn	Reference
Ol_T5-2	4.08 × 10−7	[010]	2	970	−0.3	0.8	[31]
Opx_T5-2	6.22 × 10−8	[001]	2	970	−0.3	\	[32]
Opx_C11	3.71 × 10−9	[001]	2	870	−0.3	\	[32]
Pl_C11	1.11 × 10−5	\	2	870	−0.3	0.35	[30]

¹ According to [26].

. The crystal orientation was estimated according to the crystal shapes. Note that Mg diffusion in plagioclase is isotropic [30]. We assumed that the plagioclases had a homogeneous initial MgO concentration distribution. For the orthopyroxenes and olivines, we assumed that they had a homogeneous core with a higher/lower Mg# value surrounded by a homogeneous mantle with lower/higher Mg# value at the beginning of diffusion. The outermost rim compositions were used as the boundary conditions.

4. Results and Discussions

4.1. Timescales of Magma Mixing Beneath the Iheya Ridge

The diffusion modelling results are shown in Figure 2 and Figure 3. In the andesite T5-2, the Mg diffusion in the olivine yielded a time of 0.1–0.3 years (Figure 2a). While the olivine crystal was slightly corroded, the Fe-rich rim had constant Mg# values, suggesting that the corrosion had little effect on the Mg diffusion profile across the Fe-rich and Mg-rich boundary. This time scale was quite consistent with that obtained from the Mg diffusion in orthopyroxene (0.1–0.3 years), which had an euhedral crystal shape (Figure 2b). In addition, the textures of the olivines in the T5-2 andesite (Figure 2a) were similar to those in the experiment of [33], where the olivines contacted with rhyodacitic and rhyolitic melt for up to 622 h at 800–885 °C and 50–150 MPa. This further suggested that the olivines in the andesite T5-2 contacted with the mixed magma for a very short time. In the rhyolite C11, the orthopyroxene crystals yielded a much longer magma residence time (80–120 years) (Figure 3a,b), which was consistent with the results calculated from the normally zoned plagioclase and the reversely zoned plagioclase (Figure 3c,d), although the latter were not well constrained due to the limited data.

It is noteworthy that the timescale calculation here was not a precise dating, as the diffusion coefficients were not constant but were dependent on the crystal compositions, crystal orientation, temperature, and so on [15,17,30,31,32]. Other uncertainties came from the choice of the initial and boundary conditions [17]. For example, the initial Mg concentrations of the plagioclases in the rhyolite C11 may not have been homogeneous. However, similar timescales yielded by multiple crystals indicated that our results were a plausible estimation.

Despite this uncertainty, our results suggested different magma residence times (from <1 year to ~100 years) beneath the Iheya Graben. The controlling factors on the timescales of magma mixing might be diverse, e.g., tectonic settings and the physical and chemical conditions of the magmas. Because the andesite T5-2 and rhyolite C11 used in this study were from the same location, their different magma residence times might be in part related to the different rheology and temperature of the mixed magmas. For example, rhyolite C11 is a mixture of rhyolitic melts [26], which have higher viscosities due to their higher SiO₂ concentrations and relatively lower temperatures. A high viscosity and a lower temperature would retard the magma flow and reduce the efficiency of magma mixing [34]. A long magma residence time (~600 year) was also reported in the southernmost part of the SOT, where the rhyolitic magma was intruded by small amount of basaltic magma (present as a mafic magma enclave) [25], suggesting inefficient magma mixing and a relatively high viscosity. In comparison, the andesite T5-2 in this study was generated by the mixing of a basaltic magma with a rhyolitic magma at a ratio of 4:6 [26]. The injection of a high proportion (40%) of high-temperature and low-SiO₂ basaltic magma would form a mixed magma with relatively low viscosity.

4.2. Rapid Magma Mixing and Eruption of Andesitic Magma in Bimodal Volcanic Series

The volcanic rocks in the OT are characterized by bimodal distributions, that is, basalts and rhyolites are dominant in the OT with scarce andesites [35,36,37]. The andesites have been suggested to have mainly been sourced from the mixing of basaltic magma with rhyolitic magma [26]. Therefore, the amount of andesites in the OT is probably dependent on the efficiency of magma mixing and eruption.

Magma mixing is considered to be an important trigger for eruptions [9,10]. An extremely short time between mixing and eruption has been reported (e.g., days to months) [21,38]. The short residence time (~0.1–0.3 years) of T5-2 magma after the magma mixing implied that the injection of basaltic magma into the rhyolitic magma may have led to the instability of the magma reservoir and the subsequent eruption. The efficient magma mixing suggested that the Iheya Ridge is a place that is favorable for the eruption of andesitic magma, which was consistent with the observation that the andesites recovered in the OT (excluding the modern arc fronts) so far were mostly from the Iheya Ridge and adjacent regions [26,36].

4.3. Implications for the Development and Preservation of Seafloor Hydrothermal Activities in the OT

The formation of seafloor hydrothermal activities involves the penetration of seawater into the crust, fluid–rock interactions heated by shallow magma, the rise of hydrothermal fluids, and the precipitation of hydrothermal products (e.g., sulfides), which may take from years to decades to occur [39]. Therefore, a long-term stable sub-seafloor magma system is essential for the development of a seafloor hydrothermal system with potential mineral resources. For example, the formation of a CLAM hydrothermal field on the Iheya Ridge might be attributed to a rhyolitic magma reservoir similar to that of C11 rhyolite. Similarly, long-term magma storage (~600 years) has also been reported near the Tangyin hydrothermal field [25]. However, the coexistence of a short magma storage suggests that the CLAM hydrothermal field may not have good prospects for resource exploitation because the rapid mixing and eruption could ruin the previously formed hydrothermal deposits and also present a great threat for mining activities. Therefore, evaluating the stability of sub-seafloor magmatic systems is a critical step in the assessment of seafloor hydrothermal resources.

5. Conclusions

The timescale of magma mixing beneath the Iheya Ridge was estimated by using Mg diffusion chronometry. The andesitic magma generated by the mixing between the basaltic and rhyolitic magmas experienced a very-short-time (~0.1–0.3 years) residence prior to eruption. In comparison, the rhyolitic magma generated by rhyolite–rhyolite mixing resided in the magma chamber for a much longer time (~80–120 years), which might have been in part related to the higher viscosity of the mixed magma. The rapid basalt–rhyolite mixing and eruption was favorable to the appearance of andesites in a bimodal volcanic series, but this is a disadvantage for the development and preservation of seafloor hydrothermal resources. This study thus emphasized the importance of evaluating the stability of sub-seafloor magma systems during the assessment of seafloor sulfide resources and mining prospects.