New whole-rock major and trace element geochemical, zircon U-Pb geochronological, and Hf isotopic data from gabbroic rocks in New Zealand’s mid-Permian Dun Mountain ophiolite belt (DMO) provide insight into the evolution of subduction systems and early stages of intraoceanic arc development. Fe-oxide-bearing gabbros yielded high εHf(t) values (+10.3 to +13) and zircon U-Pb ages of 271.6 ± 0.6 Ma. In contrast, Fe-Ti-oxide-bearing gabbros of 268.1 ± 0.6 Ma show more enriched geochemical characteristics, including a wide range of εHf(t) values (+15.5 to +6.8). New findings strengthen the evolutionary model for the DMO and place constraints on its youngest known magmatic episode. We infer that late magmatism fingerprinted by these gabbros, including consistent negative Nb-Ta anomalies, reflects early stages of arc development and formation of island arc tholeiites on the DMO. Our model is consistent with other existing regional geochronological and geochemical data, implying that the DMO had an early stage of normal-mid-ocean ridge basalt crustal accretion followed by an influx of slab-derived components and maturity of the subducting system between ca. 271.6 and 268 Ma. These results extend our understanding of the evolution of distinct intraoceanic systems.

Ophiolites are fragments of ancient oceanic lithosphere that have been incorporated into continental margins [1, 2]. They can be formed in distinct tectonic settings, including mid-ocean ridge, back-arc, and forearc [3, 4]. However, since the recognition of lava with island arc tholeiites (IAT) and calc-alkaline geochemical signatures in the Troodos ophiolite [5], a growing number of studies have associated these fragments of ancient oceanic lithosphere to intraoceanic convergent plate margins [6]. Such ophiolites, formed during sea-floor spreading above the subducting slab, are referred to as suprasubduction zone (SSZ) ophiolites [7]. They are widely interpreted to form during subduction initiation and early growth of island arcs [2, 7, 8]. As a result, studying their geochemical and geochronological signatures is crucial for understanding plate tectonic processes and intraoceanic systems [3, 9-12].

Different geochemical signatures, such as forearc basalt (FAB), boninite, and IAT, can often be found in rocks from the ophiolitic crustal section. These signatures are widely used to identify different stages of the ophiolite and the evolution of the intraoceanic system [13, 14]. However, diverse processes can affect the geochemical characteristics of ophiolitic rocks, for example, the injection of fluids and melts from the slab [15, 16], distinct episodes of melt extraction from the mantle [17], and cumulate processes [18, 19]. Additionally, specific processes, such as colder, denser slabs descending more quickly or a thicker sedimentary cover of the slab, can also contribute to the geochemical heterogeneity of ophiolitic rocks [20-22]. This complexity can make it challenging to determine an ophiolite’s tectonic setting or evolutionary model, which is usually complicated by a limited amount of accurate geochronological data in mafic-ultramafic rocks. As a result, different interpretations of the same ophiolite are often proposed.

The middle Permian Dun Mountain ophiolite belt (DMO) [23] in New Zealand’s South Island is a prime example of the difficulties in interpreting ophiolites with diverse geochemical signatures and limited geochronological data. The mantle section from the DMO has a depleted geochemical character and spinel Cr# (0.7, 0.8) and TiO2 (wt.% < 0.2%) values that resemble those of subduction-related ophiolites [24, 25]. The lava section, on the other hand, has geochemical signatures similar to mid-ocean ridge ophiolites, dominated by normal-mid-ocean ridge basalt (N-MORB) geochemical signatures, while dikes and uppermost lavas resemble FABs or IAT [26-28]. As a result of these heterogeneous geochemical signatures, several models have been proposed to explain the tectonic setting in which the DMO formed including: at a mid-ocean ridge [29, 30], back-arc [23], in association with slow-spreading ridges in a forearc [26, 31], or fast-spreading ridges [32]. In part, these differing models reflect the limited geochronological data available to support the interpretation of various magmatic rocks geochemical signatures [32-34]. Further geochemical and geochronological research is needed to better understand the magmatic stages of the DMO, particularly within the lower crustal plutonic rocks it contains.

This study presents new petrographic, whole-rock geochemical, zircon U-Pb geochronological, Lu-Hf isotopic, and geochemical characteristics of the intrusive mafic rocks of the DMO. The youngest magmatic products reported here indicate associations with subduction processes and magma extraction during the early stages of arc development. When combined with previous research, these new findings improve our understanding of the magmatic evolution of the DMO and provide insight into various stages of the development of an associated intraoceanic system.

The DMO consists of serpentinized peridotites, cumulate sequences, gabbros, volcanic rocks, and dikes (Figure 1(a) and (b)). It is overlain by forearc basin sediments of the Maitai Group and structurally overlies the Patuki and Croisilles mélanges [23, 31, 34, 35]. Outcrops of the ophiolite belt have been offset ~480 km dextrally by the Alpine Fault since the early Miocene Epoch, separating them into northern and southern sectors [35, 36].

DMO mantle sections are characterized by harzburgite and dunite, with sparse occurrences of podiform chromitite and pyroxenite [23]. Petrogenetic and geochemical studies of the Red Hills complex (northern sector; Figure 1(b)) indicate at least three igneous events associated with distinct melting episodes and refertilization processes [25, 29, 30]. The first event was related to the production of MORB-like melts through 10%–15% partial melting of mantle sources within the garnet stability field (>55 km depth). A subsequent event is interpreted to have been associated with an additional 10% partial melting of the mantle sources, which produced boninite-like melts in the spinel-stability field (30, 50 km depth) in a forearc setting. This stage was associated with the progression of the subduction system [25]. It is assumed that a final stage of development recorded in the mantle rocks reflects the channeling of slab-enriched melts and fluids. This may be attributed to the formation of plagioclase-bearing ultramafic rocks, at shallow depths inside the plagioclase stability field (about 15 km) [25, 29].

The DMO crustal segment comprises a dismembered cumulate sequence, isotropic gabbros, and mafic volcanic rocks, such as sheeted dikes, lava flows, pillow lavas, and dikes [23, 26, 32, 37]. Cumulate sequences include pyroxenite, wehrlite, and dunite, interlayered with gabbros, including isotropic hornblende- and clinopyroxene-gabbros [26, 28, 37]. Volcanic rocks are predominantly basaltic and exhibit an upward transition to basaltic-andesites and, rarely, andesites [26, 32]. Early lavas are characterized by MORB-type signatures, with IAT becoming increasingly common up-section [26, 27, 32]. The DMO also contains a complicated network of dikes with varying lithologies and geochemical signatures [27, 32], with a predominance of FAB affinity.

Due to earlier efforts to determine the ages of magmatic zircons using the Thermal ionization mass spectrometry(TIMS) method, the DMO has long been considered to have formed during the late Early Permian Period (ca. 280 ± 5 Ma [34]). In a later study, Jugum et al. [31] reanalyzed the same samples using the LA-ICP-MS method, which is also employed in this work, facilitating age comparisons. To maintain data consistency, this study predominantly utilizes the U-Pb zircon ages reported by Jugum et al. [31] to compare and constrain various stages of magmatic evolution within the DMO.

Based on zircon U-Pb ages from both mafic and felsic rocks, the genesis and geochemical evolution of the DMO are considered to have taken place during an interval of 9 ± 3 Ma [28, 32, 33]. Samples from the southern sector (i.e., south of the Alpine Fault) tend to yield older ages than those from the northern sector (Figure 1(c)) with plagiogranite and anorthosite dikes yielding concordant Kungurian Age (middle Permian Period) age determinations (206Pb/238U vs. 207Pb/235U) of 277.6 ± 3.3 (MSWD = 1.4) and 277.4 ± 3.4 Ma (MSWD = 5.5), respectively [32]. When interpreted together with zircon geochemical data, these ages are considered representative of episodes of early N-MORB (plagiogranite) and middle MORB-IAT (anorthosite) crustal formation. In the northern sector, a plagiogranite dike cross-cutting the mantle flow foliation in peridotites of the Red Hills ultramafic complex has yielded a Kungurian Age (middle Permian Period) zircon U-Pb age determination of 274.5 ± 0.4 Ma (MSWD = 0.26) [33]. This is deemed the youngest age prior to cessation of deformation in the peridotite complex of the Red Hills massif [25].

DMO formation remains a topic of discussion among researchers. Some suggest that it is a mid-ocean ridge (MOR)-type ophiolite that was subsequently exposed to a subduction system based on analyses of whole-rock and mineral geochemistry from mantle and crustal rocks [27, 29, 30]. However, others argue that it formed in a forearc setting [24, 25, 32], citing typical Penrose “stratigraphy,” the refractory nature of the mantle rocks, characteristics of some crustal rocks, and a conformably overlying succession of forearc sediments (Maitai Group). Among the forearc interpretations, DMO is considered to represent both a slow-spreading ridge [31] and a fast-spreading and short-lived ophiolite, similar to the modern Izu-Bonin-Mariana (IBM) forearc system [32].

In the northern sector of the DMO, gabbro samples were collected both in situ and from float (Figure 1(b)). In the Roding River, a foliated Fe-oxide gabbro (RR-04; 41.37467°S, 173.29011°E; Figures 2(a) and (e)) with subophitic texture was sampled in situ. Two km SW of the RR-04 site, an isotropic Fe-Ti-oxide gabbro (RR-07; 41.36678°S, 173.30453°E) intruded by thin (<10 cm) dolerite dikes (Figure 2(b)), was collected from a ~20 m wide outcrop. Near the same locality (4 km to NW of RR-07), a pegmatitic gabbro (RR-02B; 41.35869°S, 173.25853°E) was collected from boulders in the Roding River (Figure 2(d)). Given the quantity of peridotite and gabbro boulders in the riverbed and the presence of gabbroic rock outcrops upstream, the sample is considered to be derived from the DMO. Another isotropic Fe-Ti-oxide gabbro (Figure 2(f)), 500 m W of the contact between the DMO and Maitai Group, was collected from boulders in the Miner River (HT-04; 41.39433°S, 173.23656°E). Considering the occurrence of peridotite boulders and that upstream areas of the Miner River are mapped as a component of the ophiolite crustal segment with their genesis considered to be related to the DMO (Figure 1) [38]. Within the Red Hills ultramafic massif, a gabbroic dike (sample LC-03; -41.64775, 173.00594) was collected from an outcrop in Lowther Creek (Figure 2(c) and (g)). This dike is 2 m wide, shallow-dipping, and intrudes an upper mantle peridotite with steeply dipping compositional banding of orthopyroxene and olivine-rich layers. Samples were prepared for petrography, geochemistry, and zircon isotopic analyses. Table 1 provides an overview of sample petrographic characteristics that are detailed in online Supplementary Material 1.

Whole-rock major and trace element analysis was conducted at the Australian Laboratory Services facilities in Brisbane, Australia. The samples were first prepared in a series of steps including washing, drying, crushing, splitting, and pulverizing. The concentration of major element oxides was determined using an X-ray fluorescence spectrometer with a precision of ±2%–5%. Additionally, thirty trace elements were measured using an inductively coupled plasma mass spectrometer (ICP-MS) with a precision of better than ±5% for most elements. Quality control measures were implemented, and all standards were found to be within the target range, indicating highly accurate analyses. The detection limit for major oxides is 0.01%, while it varies for trace elements (online Supplementary Material 2).

Zircon separation was performed using standard methods (gravimetric, magnetic, and heavy liquid separation) at the Hebei Geological Survey, Langfang, China. The cores of the grains were exposed by mounting the zircons in epoxy and then polishing them. They were imaged using a Hitachi SU3500 scanning electron microscope that was equipped with backscatter, secondary electron, and cathodoluminescence (CL) detectors. The images were used to identify mineral fractures, zoning, internal structures, and CL characteristics (see online Supplementary Material 1 for CL images).

Isotopic and trace element analyses of zircon grains (91Zr, 45Sc, 49Ti, 88Sr, 89Y, 93Nb, 139La, 140Ce, 146Nd, 147Sm, 153Eu, 157Gd, 172Yb, 175Lu, 178Hf, 206Pb, 207Pb, 208Pb, 232Th, and 238U) were conducted using a Thermo Fisher iCAP RQ quadrupole ICP-MS equipped with an ASI RESOlution SE 193 nm laser at The University of Queensland. U-Pb isotopes and trace elements were collected in the same analytical session. The laser spot size was adjusted based on the size of the grains. For larger zircon crystals (samples RR-04, HT-04), data were collected with a 30 µm spot size in thirty cycles of 1 second at a repetition rate of 7 Hz and a fluence of 3 J/cm3. Sample RR-2B was analyzed using a 24 µm spot size, with the same laser and ICP-MS settings as the 30 µm spot size session. U-Pb age analyses on zircons were performed using the 91,500 zircon standard [39] for background, and instrumental bias [40] and TEMORA2 grains [41] were used as secondary standards to monitor precision and accuracy. Data obtained indicate precision and accuracy better than 0.5% for U-Pb age measurements (see online Supplementary Material 1). Uncertainties associated with U-Pb isotope ratios and standard deviation of primary and secondary zircon standards were propagated into sample uncertainties following procedures suggested by Horstwood et al. [42]. For zircon trace elements, a NIST612 glass standard was used as a primary reference material. Data were processed using the Iolite software [43], and Zr was used for internal normalization of the trace elements.

Isotopic analysis of Lu-Hf in zircon (masses 171–180) was conducted using a Nu Plasma II multicollector ICP-MS (MC-ICP-MS) equipped with an ASI RESOlution SE 193 nm laser at the Center for Geoanalytical Mass Spectrometry, School of Earth and Environmental Sciences at The University of Queensland. Measurements were carried out using a larger spot size of 50 µm, in thirty-five cycles of 1 second at a repetition rate of 8 Hz and an on-sample fluence of 3 J/cm3 for the same spot where the grains were analyzed for U-Pb isotopes and trace elements. Analytical procedures followed Zhou et al. [44]. Primary and secondary zircon standards were the same as those used for the U-Pb age sessions. Accuracy and precision achieved were better than 0.1% for the primary standard (91,500) and monitoring standard (TEMORA2). Additional information on the analytical procedures, data quality, geochemical analyses, zircon trace elements, and isotopic data are available in online Supplementary Materials 1 and 2.

4.1. Whole-Rock Geochemistry

Major and trace element values are presented in Table 2. In discrimination diagrams, samples plot within gabbro or gabbro diorite fields, with a subalkaline tendency (online Supplementary Material 1). Notably, geochemistry allows samples to be separated into two distinct categories according to their elemental abundances, with Fe-oxide gabbro (RR-04) and gabbroic dike (LC-03) having more depleted character than Fe-Ti-oxide isotropic gabbros (HT-04 and RR-07; Figures 3 and 4).

Sample LC-03 exhibits high SiO2 (52.6 wt.%) and MgO (10.9 wt.%) and low TiO2 (0.23 wt.%) abundance. Fe-oxide gabbro (RR-04) also has low TiO2 (0.24 wt.%) abundance but lower values of SiO2 (49.59 wt.%) and MgO (7.72 wt.%). These samples exhibit higher MgO, CaO, and Cr (440, 870 ppm) values than the Fe-Ti-oxide gabbros. In contrast, more geochemically enriched samples (HT-04, RR-07) have higher concentrations of SiO2 (51.19, 54.24 wt.%) and TiO2 (0.86, 0.94 wt.%) and lower MgO (6.03, 7.26 wt.%) and Cr (150, 340 ppm). Notably, more depleted samples exhibit lower Ti/V (1–25) and Nb/Ta (1–3) ratios, while Fe-Ti-oxide gabbros tend to exhibit higher values for Ti/V (32–40) and Nb/Ta (11–13) ratios. In major oxide geochemical diagrams (MgO vs. TiO2 and SiO2 vs. MgO; Figures 3(a) and (b)), samples are plotted within fields associated with subduction processes (island arc and FAB).

On primitive mantle-normalized trace and chondrite-normalized REE plots (Figures 4(a) and (b)), dike (LC-03) and Fe-oxide gabbro (RR-04) exhibit greater depletion of trace elements compared with Fe-Ti-oxide gabbro samples. These samples have notably lower abundances of some high-field-strength elements (HFSEs), such as Nb, Zr, Hf, and Ti, and positive anomalies among large-ion lithophile elements (LILEs; e.g., Ba, K, Rb, Cs, and Sr). In contrast, more enriched samples (Fe-Ti-oxide gabbros) resemble N- or E-MORBs and exhibit enrichment in LILE but minimal depletion in HFSE (Figure 4(a)). Similar patterns are observed in chondrite-normalized REE plot in which gabbroic dike and Fe-oxide gabbro exhibit greater depletion in REEs (Figure 4(b)) compared with Fe-Ti-oxide gabbro samples, which display REE contents similar to E- or N-MORB. Despite having distinct signatures, all samples show an overall enrichment in fluid-mobile/LILE (Cs, Rb, Ba, Th, K, and Sr) and discernible depletion of more immobile/HFSE (Nb, Ta, and Ti) compared with N-MORB lavas (Figures 4(a) and (b)).

4.2. Zircon U-Pb Geochronology and Hf Isotopes

Zircons from samples RR-04 (Fe-oxide gabbro), HT-04 (Fe-Ti-oxide gabbro), and RR-2B (pegmatite) were analyzed to determine U-Pb ages, Lu-Hf isotopes, and trace elements abundances. Samples RR-07 and LC-03 did not yield zircons of sufficient abundance or size for analysis. Recovered zircon grains exhibit, generally, low length-to-width ratios (1:1 to 3:1) and are between 15 and 250 µm long. Most crystals have weak oscillatory and sector zoning under CL light (refer to supplementary material for CL images), typically observed in mafic igneous rocks [45]. The analyzed samples contain variable U (18, 940 ppm), Th (3, 706 ppm) contents, and Th/U ratios between 0.17 and 0.86. They show an increase in average U and Th content and Th/U ratios from RR-04 (190 and 88 ppm, 0.35) and HT-04 (200 and 90 ppm, 0.41) to pegmatitic gabbro (209 and 122 ppm, 0.54). These values are consistent with U and Th concentrations and Th/U ratios expected for magmatic zircons from gabbroic rocks [46].

Zircon U-Pb Concordia ages (206Pb/238U vs. 207Pb/206Pb) are from the middle Permian Period (Roadian Age), ranging from 271.6 ± 0.6 to 267.6 ± 0.9 Ma. Of the samples measured, the oldest age was found in the Fe-oxide gabbro (RR-04; 271.6 ± 0.6 Ma), while the Fe-Ti-oxide (HT-04) and pegmatitic gabbro samples (RR-02B) are slightly younger at 268.1 ± 0.6 and 267.6 ± 0.9 Ma (2σ), respectively (Figure 5). Zircon-weighted mean plots show individual grains ages. Zircon Hf isotopes exhibit variations between samples, with values ranging between +6.8 and +15.5 (Figure 6). The Fe-oxide gabbro samples have the highest median value of εHf(t) at +12.2, and a more consistent εHf(t) values (+10.8 and +13), compared with the Fe-Ti-oxide samples (+11.1) and pegmatitic gabbro (+11.5). It is worth noting that the Fe-Ti-oxide gabbro samples exhibit a wider range of εHf(t) values, with the most and least juvenile zircon samples analyzed having values of +15.5 and +6.8, respectively.

4.3. Zircon Geochemistry

Zircon geochemical characteristics from gabbroic samples exhibit similarities with those observed in SSZ-ophiolite and plot between the continental arc and MOR fields (Figure 7(a); U/Yb vs. Nb/Yb diagram). A contrast between samples is better observed on primitive mantle-normalized trace and REE plots (Figures 7(b) and (c)). Fe-oxide gabbro zircons (RR-04) tend to exhibit lower trace and REE elemental abundances compared with zircons from Fe-Ti-oxide (HT-04) and pegmatitic (RR-2B) gabbros, especially for Y, Ce, Nd, Sm, Eu, Gd, Yb, and Lu concentrations and Th/U ratios.

5.1. Late Magmatic Stages of the DMO

Gabbro compositions together with zircon geochronological and geochemical data constrain the timing and evolution of the youngest magmatic products described from the DMO. Analyzed gabbros exhibit low TiO2 contents and general depletion in HFSEs. Such compositions are usually unexpected in MOR ophiolites [47], which sources commonly exhibit smaller degrees of partial melting [3, 7]. Fe-Ti-oxide samples (RR-07, HT-04), Fe-oxide gabbro (RR-04), and gabbroic dike (LC-03) plot near or within the oceanic/island arc fields (Figure 8a), indicating a potential origin related to subduction processes. Furthermore, the AFM discrimination diagram (Figure 8b) exhibits Fe-Ti-oxide gabbros plotting within the arc-related noncumulate gabbros and diorites fields, while Fe-oxide gabbro falls near the arc-related ultramafic and mafic cumulates field. Thus, geochemically depleted Fe-oxide gabbro (RR-04) is likely to reflect fractional crystallization processes of melts formed in an intraoceanic system, whereas the high-Mg gabbroic dike (LC-03) probably fingerprints melt extraction from depleted mantle sources. Despite variances, the consistently low Ti, Zr, Hf, Nb, Ta, and Ti/V values across all analyzed samples indicate their classification as IAT. Furthermore, zircon U-Pb ages from these samples, ranging between 271.6 and 268 Ma, alongside positive εHf(t) values (+6.8 and +15.5), constrain an important episode of oceanic crust formation in the DMO in an island arc tectonic setting.

5.2. Comparison Between Early and Late DMO Magmatic Rocks

New findings further elaborate on differences between early and late magmatic episodes in the DMO, as previously observed [32, 48]. Zircon U-Pb ages from crustal rocks indicate magmatic evolution of the DMO occurred over a ca. 10 Myr interval ([31, 32]; this study). The oldest magmatic episode (~ 277.5 Ma) is recorded from the southern sector of the DMO (Figure 1(c)), in which geochemical signatures of zircons recovered from anorthositic and plagiogranitic dikes indicate similarities to those observed in MOR and/or IAT settings [32]. This early magmatism is usually associated with lavas with MOR signatures [31, 32], which are considered to be products of deep fertile melt extraction from mantle sources [25, 30, 32].

In contrast, younger crustal rocks are mostly observed in the northern sector and yield zircon U-Pb ages between 274.5 and 267.6 Ma ([31, 32]; this study). This younger magmatism is interpreted as a product of melts associated with an intraoceanic setting and is fingerprinted by studies in the crustal and mantle sections [24, 25, 31, 32]. Gabbros analyzed as part of this study exhibits subduction-related signatures, with ages ranging between 271.6 and 268.1 Ma. These findings contribute to a better understanding of the timing of younger magmatism and indicate the establishment of an intraoceanic system. Additionally, data presented herein contribute to the understanding of the temporal dynamics of fluid and melt circulation in the subduction system associated with the DMO.

5.3. DMO Magmatic Evolution

The combination of new geochemical and geochronological data together with existing data results in a refinement of understanding of the magmatic evolution of the DMO (Figures 9(a)–(e)). Zircon U-Pb ages ranging from approximately 277.5 to 267.6 Ma are reported from a diverse range of lithologies and geochemical compositions, distributed across a wide geographic area ([31, 32]; this study]. This age interval is associated with a complex spatial and temporal magmatic evolution that involves rocks from both the southern and northern sectors.

Early magmatic episodes in the DMO have traditionally been associated with a mid-ocean ridge setting and deep-fertile melts [30]. Our research suggests that the early magmatic products in the DMO may have been triggered by the upwelling of fertile asthenospheric melts due to decompression processes in the mantle-wedge (Figure 9(b)). These processes may have occured before revented the influx of slab-derived components into the system. This interpretation is supported by considerations that MOR-like lavas can form approximately 1–10 Myr after subduction initiation [20]. This magmatic episode is likely characterized by plagiogranites and anorthosites containing zircon with geochemical signatures similar to those MORB and/or IAT dated at around 277.5 Ma [32]. In this scenario, a progressive evolution of the DMO in a subduction-related setting is preferred (Figure 9(b)–(c)).

Following the evolution of the subduction system, FAB and IAT geochemical signatures appear in gabbros, mafic dikes, and lavas, indicating the establishment of an intraoceanic subduction system (Figure 9(d)). Mafic dikes with depleted geochemical characteristics suggest that melts were extracted from a refractory mantle source (Figure 9(d)). Other studies of petrology and geochemistry in the DMO mantle section have revealed the presence of depleted sources, including harzburgite and chromitite, which exhibit signatures comparable to those found in subduction-related settings [25, 29, 30]. Furthermore, these sources display fO2 (above MORB field) in Cr-spinels [25, 30] typical of oxidizing and fluid-rich environments [49].

The presence of IAT signatures in the uppermost lava sequence [26] and gabbroic samples (Figures 3 and 8) indicates a more evolved intraoceanic system. Zircon U-Pb ages ranging from 271.6 to 268.1 Ma, occurring approximately 6–9 Myr years after the oldest recorded zircon U-Pb age from IAT-like gabbros in the DMO, could fingerprint the timing of pervasive formation of melts derived from the slab. These ages could also correlate with mantle refertilization [25, 30], potentially triggered by an influx of slab-derived melts into the mantle wedge, resulting in the combination of new fertile melts with previously depleted melts.

The magmatic evolution of the DMO shares similarities with that of the IBM forearc system [8] in terms of the progression of its crust from FAB-boninitic rocks to IAT. However, there are notable differences in the timing and distribution of subduction-related signatures. Unlike the DMO, where subducting-related rocks are mostly confined to intrusive rocks and the uppermost lavas, rocks affected by slab-derived components are widespread and were formed shortly after subduction initiation (around 1–2 Myr) in the IBM system [8, 14]. The DMO exhibits a lava sequence largely unaffected by any significant SSZ influence, and subduction-related rocks are mostly recognizable approximately 6–9 Myr after subduction initiation. This suggests slow and protracted evolution in a forearc setting for the DMO, with significant differences from the fast- and short-lived system observed in the IBM forearc system.

Petrographic and geochemical data, as well as zircon U-Pb ages, Hf isotopic, and trace element geochemical signatures from gabbroic rocks of the DMO, provide valuable insights into the evolution of an ancient intraoceanic system. Fe-oxide cumulate gabbro with zircon U-Pb ages of ca. 271.6 Ma and Fe-Ti-oxide gabbros (ca. 268.1 Ma) with IAT-like geochemical characteristics indicates evolution toward embryonic arc stage and constrain the DMO youngest magmatic episode. Our findings, when combined with previous geochronological and geochemical results, suggest that the DMO experienced an early stage of N-MORB crust evolution (ca. 277.5 Ma) before a pervasive influx of slab-derived components and refertilization of the mantle wedge resulted in the formation of arc-like rocks at 271.5–268 Ma. Results presented in this study contribute to a better understanding of the magmatic evolution of the DMO and allow comparison with other intraoceanic systems.

We thank Dr. David Kimbrough for his comments on geochronological data at our study site, which enriched our research. We also thank Dr. Gültekin Topuz, two anonymous reviewers, and handling editor Professor Bo Wang for their valuable comments and suggestions, which greatly improved our manuscript. The authors acknowledge funding support from the Australian Research Council discovery project Diamonds in ophiolites: Recycling deep mantle into suprasubduction zones (ARC DP190100814).

The authors declare no conflicts of interest.

The original contributions and data presented in this study are included in the article and the supplementary materials.

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).

Supplementary data