Introduction

It is widely recognized that certain mantle plumes (hotspots) are located near mid-oceanic ridges in the global ocean, and those on-/off axis plumes have affected or are currently influencing the magmatic processes and crustal accretion at the mid-oceanic ridge system1,2,3. Deciphering the interaction processes between plumes and ridges using geophysical and geochemical tools can offer valuable insights for resolving the debate of “bottom-up” (plume) versus “top-down” (plate) processes as the Earth’s engine2,4,5,6,7.

Similar to the occurrences in the open ocean, intraplate processes also took place in certain back-arc basins (e.g., the South China Sea, Sea of Japan, Shikoku basin, and the West Philippine Basin-WPB) in the western Pacific convergent margin8,9,10,11,12,13. However, the impact of intraplate processes (such as mantle plumes) on the subduction-related magmatic processes and the accretion of crust in back-arc basins is not clear. The WPB in the Philippine Sea Plate (PSP) is an ideal location to investigate the interaction of a plume with a back-arc spreading center (see Fig. 1).

Fig. 1: Overview of the study region.
figure 1

a Tectonic framework of the Philippine Sea Plate (PSP) and adjacent large tectonic plates. b Sketch morphological map of the West Philippine Basin (WPB) in the Philippine Sea Plate (PSP) with the base map sourced from GeoMapApp, along with sample locations. Red solid circles represent sample locations in this study, while black solid circles represent other drilling sites of deep-sea drilling projects (DSDP)9. Magnetic isochrons with polarity chron numbers (represented by lines with different colors) are based on Hilde and Lee16. The white dashed parallel lines approximate the location of the Central Basin Fault (CBF). The detailed topography of the Akle Seamount, Toog Seamount, and Vinogradov Seamount can be found in ref. 18. In addition to the ages of samples from DSDP site 447 and ODP hole 1201D inferred from magnetic isochrones of the sampling sites, this study also presents the ages for samples from some main geological units9,11,19,28 and the present study. IBM Izu-Bonin Mariana, MT Manila Trench, PT Philippine Trench. c The variations in crustal thickness in the WPB, based on the WGM2012 gravity model and ETOP01 topographic data.

Full size image

Plate reconstructions have shown that the PSP was once located at near the Equator, and gradually moved northward since the earliest Cenozoic era. This movement was accompanied by a 90° clockwise rotation during the northward migration14,15. However, the formation and evolution of the WPB are still a topic of debate8,11,13,16. The lavas from certain important tectonic settings in the basin can offer some direct insights into the geological evolution of the WPB.

In this study, we present new ages, major and trace element compositions, and Sr-Nd-Pb-Hf isotopic compositions of lavas from significant tectonic sites such as the WPB basement, the central basin fault (CBF)-fossil spreading center, and the Benham Rise. This was done to determine the connection between intraplate and back-arc spreading processes in the WPB.

Geological setting

The Philippine Sea Plate (PSP) is situated at the junction of the Eurasian, Indo-Australian, and Pacific plates (Fig. 1). During its migration from south of the Equator to its present location, the PSP experienced multiple back-arc spreading events due to the westward subduction of the Pacific plate, resulting in the formation of several back-arc basins. These include the WPB (spreading at 65-35 Ma), the Shikoku Basin (27-15 Ma), the Parece Vela Basin (29-15 Ma), and the Mariana Trough (5 Ma to present), as well as some relic/active arcs such as the Kyushu–Palau Ridge (KPR), the West Mariana Arc, and the currently active Izu-Bonin Mariana (IBM) Arc9,10,13(see Fig. 1).

The WPB is bordered by the Philippine Trench to the west, the Ryukyu Trench to the north, the KPR to the east, and the Ayu Trough to the south13(see Fig. 1). In the middle of the basin, there exists a fossil spreading center known as the Central Basin Fault (CBF)9,17. The WPB is the deepest back-arc basin in the western Pacific. Its water depth for the WPB basin (except for those plateaus and ridges) ranges from 5000 to 6500 m8. The crustal thickness of the basin is approximately 5 km, while those of the Benham Rise reach about 15 km18. The crustal thickness of the western part of the WPB (7–15 km) exceeds that of the eastern part (5–8 km) (Fig. 1). Magnetic lineation stripes no. 26-13, corresponding to the time period of 60 to 35 Ma, have been identified in the basin16. The extending orientations of striped magnetic lineations of Chron nos. 26-19 and nos. 19-13 are NW-SE and W-E, respectively. This suggests that the basin has experienced two spreading stages: earlier NE-SW and later S-N spreading styles17 (Fig. 1). In addition, in the western and northwestern part of the WPB, there existed a series of Cretaceous to Cenozoic arc terranes and intraplate plateaus, including the Gagua Ridge (arc-related) (119-131 Ma), the Amami Plateau (arc-related) (70-117 Ma), the Urdaneta Plateau (35.9–39.8 Ma), the Oki-Daito Ridge (42.5–53.3 Ma), the Oki-Daito Rise (40.5–44.4 Ma), the Benham Rise (36.5–36.9 Ma) and many seamounts9,11,19,20 (see Fig. 1).

The Benham Rise, similar to some submarine plateaus in the global ocean, has a large basement platform18. During its earlier formation stage, the Rise may have been affected by the back-arc seafloor spreading process, as evidenced by the existence of magnetic anomaly stripes in the Rise (Fig. 1). The oceanic lithosphere formed by seafloor spreading generally retains magnetic anomaly stripes, and yet oceanic plateaus or seamounts formed by mantle plumes do not exhibit these stripes21. Previous studies have shown that magnetic anomaly stripes can be identified in the lithosphere beneath the Benham Rise and the Urdaneta Plateau in the WPB18,22. For example, magnetic anomaly stripe no. 19 (41.5 Ma) extends from the central basin of the WPB into the spur of Loro seamount in the Benham Rise, and magnetic anomaly stripe no. 20 (43.8 Ma) extends from the central basin of the WPB into the main body of the Benham Rise and the Urdaneta Plateau. The evidence above suggests that, similar to the case of the Shatsky Plateau, the formation and evolution of both the Benham Rise and the Urdaneta Plateau, produced by the Oki-Daito plume or Benham mantle plume, may have been influenced by back-arc seafloor spreading in the WPB18,22 (see Fig.1). The Benham Rise and the Urdaneta Plateau were once joined together but were subsequently separated from each other by the back-arc spreading process at the CBF11.

Results

Samples from sites 94DS (the axial site of the fossil spreading center, i.e., Central Basin Fault-CBF) and 100DS (the Vinogradov Seamount on the periphery of the Benham Rise) were collected during the China-Germany joint leg no. SO-57 cruise of R/V Sonne23(Fig. 1). Samples from DSDP (Deep Sea Drilling Program) sites 292 (located in the main body of the Benham Rise) and, for comparison, WPB basement samples (from DSDP site 447 and ODP (Ocean Drilling Program) hole 1201D) were obtained from the international ocean drilling program sample repository24,25,26. Samples were dated using the 40Ar/39Ar method and analyzed for major and trace elemental compositions, as well as Sr-Nd-Pb-Hf isotopic compositions. Sample descriptions, recovery methods, locations, water depths, analytical methods, and data are provided in the Supplementary materials.

The initial impact of the Benham plume/Oki-Daito plume at the WPB spreading center occurred approximately 51 million years ago11. The Ar-Ar age of sample 100DS (a seamount near the CBF, as shown in Fig. 1) newly obtained in this study is 23.68 (±0.04) Ma (see Supplementary Fig. S1 and Data 1), approximately indicating the end of plume activity. Thus, the duration of intraplate Benham plume magmatism is approximately 27.3 million years. In addition, the newly obtained Ar-Ar age of sample no. 94DS from CBF in this study is 27.62 (±0.56) Ma (see Supplementary Fig. S1 and Data 1), which is similar to the ages (28.1 and 27.4 Ma) reported in a previous study11. This basically reflects the cessation time of back-arc spreading processes that began at 60 Ma17.

Basaltic rock samples from the main body of the Benham Rise (site 292) and the Vinogradov seamount (site 100DS) on the periphery of the Benham Rise are composed of alkali basalt. They exhibit trace element compositions similar to oceanic island basalts (OIBs) (see Supplementary Figs. S2 and S3). In contrast, samples from the oceanic basement of the WPB (sites 447 and hole 1201D) and the Central Basin Fault (CBF) (site 94DS) belong to the subalkaline rock series. These samples from the WPB oceanic basement and the CBF have trace element compositions resembling normal mid-oceanic ridge basalts (NMORBs) and enriched-MORBs, respectively, indicating that they underwent distinct different later-stage magmatic processes. The newly obtained data in this study, combined with literature data9,20,26,27,28, show that the significant differences in Sr-Nd-Pb-Hf isotopic compositions among samples from different tectonic units of the WPB (Supplementary Data 2, Supplementary Fig. S4) imply that the region has experienced a relatively complex geological evolution.

Geochemical evidence for the contributions of plume to spreading center

Until now, three models have been proposed to account for the genetic mechanism of the WPB: back-arc spreading8,14, trapped oceanic crust fragment16, and the interaction of back-arc spreading center and mantle plume11,13. However, direct geological evidence (e.g., lava geochemical compositions) for the formation and evolution of the WPB is still scarce.

Previous studies have demonstrated that the relationship between incompatible element ratios and Sr-Nd-Pb-Hf isotopic compositions can be utilized to determine whether enriched components are introduced into the mantle sources4. In this study, all samples from the WPB exhibit clear relationships between La/SmN (a chondrite normalized value, data for chondrite are from Sun and McDonough29) and 87Sr/86Sr, 143Nd/144Nd, 176Hf/177Hf and 206Pb/204Pb (Fig. 2). Similar to a previous study9, samples from the WPB basement (site 447 and hole 1201D) exhibit low incompatible element ratios and depleted Sr-Nd-Hf-Pb compositions (Fig. 2). These samples can be considered as an end member with no influence from enriched components. Similar to samples from sites 292, 294, and 446 and the Urdaneta Plateau9,11,30, the samples from the Benham Rise (site 292) and the Vinogradov seamount (site 100DS) in this study exhibit high incompatible element ratios and enriched Sr-Nd-Hf-Pb isotopic compositions (Fig. 2), indicating a potential connection to the mantle plume. Samples from the main body of the Benham Rise and Vinogradov seamount may represent different stages of a mantle plume. The samples from the Benham Rise samples with a high extent of partial melting are related to the head material of the mantle plume, while the samples from the Vinogradov seamount with a low extent of partial melting may be produced by the tail material of the plume. This is consistent with a classical working model in which a mantle plume has a “head” (forming a voluminously large igneous province) and a thin “tail” (forming a seamount chain composed of a series of relatively isolated seamounts)2,7,31. Similar to site 29111, samples from the CBF (site 94DS) exhibit high La/SmN ratios and enriched Sr-Nd-Hf-Pb isotopic compositions (Fig. 2), indicating that their mantle source may be influenced by plume materials. Because Sr Isotopic compositions are prone to influenced by seafloor alteration, we conducted two end-member mixing calculations for the CBF lavas using La/SmN ratios (chondrite normalized values) and Nd-Hf-Pb isotopic compositions (Fig. 2). The calculated results indicate that 10–20% of the source of the CBF lavas consist of mantle plume materials, which aligns with the findings from two end-member mixing models using Sr-Nd-Hf-Pb isotopic data. Based on Fig. 2, it was determined that 20–30% of the mantle sources for the site 291 samples consists of mantle plume materials. Additionally, a previous study28 on CBF samples also indicates the presence of mantle plume material in the magmatic source.

Fig. 2: Relationship between La/SmN (chondrite normalized values) versus radiogenic isotopic ratios for the WPB lavas.
figure 2

a The plot of La/SmN versus 87Sr/86Sr. b The plot of La/SmN versus 143Nd/144Nd. c The plot of La/SmN versus 206Pb/204Pb. d The plot of La/SmN versus 176Hf/177Hf. Data sources: data for the WPB basement lavas are from literatures9,26,27; data for samples from sites DSDP-292, DSDP-294, and DSDP-446 are from literatures9; data for Urdaneta Plateau lavas are from literatures11; data for samples of site DSDP-291 are from literatures9; data for lavas from CBF and Kyushu–Palau ridge are from literatures28; data for those end-members are listed in Supplementary Data 3; data for Plume-free WPB MORB (with no influences from subduction components) were calculated using data from literatures26, Yogodzinski et al., and the present study. In addition, the average data for DSDP-292 were calculated based on Hickey-Vargas (1998) and the findings of this study, and that for 100DS were calculated using the data obtained in this study.

Full size image

Relative to the WPB basement lavas, the CBF lavas exhibit higher Nb/Y and Ba/Yb ratios, and lower Zr/Nb ratios (Fig. 3), indicating that their mantle sources may contain garnet. The data for the CBF lavas in this study and in the literature28 are plotted on the mixing line between the WPB basement lavas (without plume effect) and the Benham Rise lavas. This suggests that the enriched components of EMORB-like samples from the WPB fossil spreading center may have originated from the Benham Rise (Fig. 3). Results from two end-member mixing calculations indicate that the CBF lavas in this study and in the literature28 may contain 5–10% materials from the Benham plume, while site 291 samples may contain approximately 5% materials from the plume (Fig. 3).

Fig. 3: The discrimination of source lithology of the WPB lavas.
figure 3

a The plot of Nb/Y versus Ba/Yb. b The plot of Nb/Y versus Zr/Nb. Data sources are the same as Fig. 2. The data for those end-members are listed in Supplementary Data 3.

Full size image

Plume-spreading center interaction in the West Philippine basin

Previous studies have proposed that a typical mantle plume originating from the core/mantle boundary has a mushroom-shaped “head” and a thin “tail”2,7,31. When the plume upwells to the bottom of the lithosphere, it interacts with the lithosphere, resulting in a short-term significant “head” event and a persistent, relatively small “tail” event32,33. And the outcomes of “head” and “tail” events are large igneous provinces and age-progression seamount chains, respectively32. Alternatively, plume melting beneath a spreading center can cause greater melting (e.g., Iceland, Galapogos)34,35, while off-axis melting is linked to less ascent and lower-degree melting1,36,37.

Based on their geological and geophysical characteristics in this study and literature11(see section “Geological setting” and Fig. 1), we suggest that the Benham Rise, the Urdaneta Plateau, and Oki-Daito Rise are the results of “head” event of the Benham mantle plume and have undergone a seafloor spreading process during their early evolution stage. The positive topographies adjacent to the Benham Rise, including Narra Spur (e.g., the Toog, Alke, and Vinogradov seamounts) (Fig. 1), are the products of the “tail” event of the Benham mantle plume (the last and dying stage of plume-related magmatism18). The ages of the intraplate volcanic lavas from the WPB span a wide range, from 51 Ma11 to 23.7 Ma (this study), and the back-arc spreading period (forming the WPB) ranges from 60 Ma17 to 27.6 Ma (this study) (Fig. 1).

Thus, there was an overlapping period (51 Ma (plume initiation) minus 27.6 Ma (spreading cessation) = 23.5 Ma) during which both the intraplate plume and back-arc spreading processes existed, and they may have interacted with each other during this period. In fact, the geochemical compositions of lavas from various tectonic sites in the WPB mentioned above confirm the existence of this interaction (see Figs. 2 and 3). In this study, considering the factors mentioned above, we developed a conceptual model of tectonic evolution involving the interaction of the CBF spreading center and the Benham mantle plume, as shown in Fig. 4.

Fig. 4: A conceptual model depicting the three-stage ridge-plume interaction in the West Philippine Basin in the Philippine Sea Plate.
figure 4

Stage 1 (55-40 Ma), initial back-arc spreading of the WPB and its spreading center gradually approach to the Benham mantle plume; stage 2 (40-35 Ma), the spreading center of the WPB migrate above the Benham mantle plume; stage 3 (after 35 Ma), the spreading center of the WPB gradually away from the Benham mantle plume. AP Amami Plateau, AS Akle Seamount, BR Benham Rise, MDB Minami-Daito Basin, ODR Oki-Daito Ridge, OR Daito Ridge, RBL rheological boundary layer, TS Toog Seamount, UP Urdaneta Plateau, VS Vinogradov Seamount.

Full size image

Stage 1 (55-40 Ma): The West Philippine Basin (WPB) began to form due to the bilateral subduction of the Indo-Australian plate and the northern New Guinea plate beneath the Philippine Sea Plate, through a process of back-arc spreading (Hilde and Lee, 1984; Hall, 2002). Around 51 million years ago, the Benham mantle plume rose to the bottom of the lithospheric of the WPB, and the material from the plume moved towards the WPB back-arc spreading center along a sloping rheological boundary layer (RBL), as the cases described in previous studies36,38. During their journey to the spreading center, the plume materials can also ascend along certain lithospheric faults and erupt onto the surface, forming some OIB-like and EMORB-like lavas in specific locations within the southern Daito Basin, the Oki-Daito Ridge, and the WPB basement region9,11.

Stage 2 (40-35 Ma): Due to the persistent northward movement and clockwise rotation of the WPB, the WPB back-arc spreading center migrated over the Benham mantle plume around 40-35 Ma. The combined interaction of back-arc spreading processes and mantle plume led to the formation of the Benham Rise and the Urdaneta Plateau. Subsequently, as a result of the continuous WPB back-spreading process, the Benham Rise and the Urdaneta Plateau have been fragmented, and the WPB spreading center gradually shifted to the northeast side of the Benham plume.

Stage 3 (after 35 Ma): At approximately 35-33 million years ago, lavas from the WPB spreading center exhibited signals of the Benham plume, indicating that the plume materials had migrated along the CBF spreading center, similar to the situation in the open ocean38. At the same time, some contemporary lavas from the intersection site between the Kyushu–Palau ridge and the CBF showed plume signals28, indicating that the Benham plume material may have migrated along the spreading center and reached the Kyushu–Palau sub-arc mantle. At about 33 million years ago, the WPB spreading process ceased. Subsequently, the tail materials of the Benham plume continued to migrate along certain lithospheric faults toward the surface and formed a series of small seamounts (e.g., Vinogradov (23.7 Ma), Akle, and Toog seamounts) in the Narra spur between the CBF and the Benham Rise (Fig. 1).