Revealing Unmapped Tectonic Settings through Geochemical Fingerprinting of Th-Nb-Yb Open Access

The use of geochemistry as a spearhead or a key tool in determining geodynamic settings for understanding tectonic plate movement has been ongoing for over 50 y, beginning with Pearce and Cann [1] pioneering model that used trace elements (Nb-Zr-Ti-Sr-Y) to classify the tectonic settings of basic volcanic rocks. Since then, geochemical models have been continually developed and expanded by various researchers (e.g. Pearce and Norry [2]; Wood [3]; Shervais [4]; Pearce [5]; Meschede [6]; Cabanis and Lecolle [7]; Pearce and Peate [8]; Wang et al. [9]; Condie [10]; Condie [11]; Sun et al. [12]; Pearce [13]; Hollocher et al. [14]; Saccani [15]; Pearce et al. [16]; Shervais [17]; Godang et al. [18]), incorporating a broader range of trace elements (e.g. Th-Ta-Hf-V-Rb-Ba-La-Ce-Sm-Yb).

However, significant challenges remain. Results from different models often diverge and sometimes even contradict one another, as diagrammatic discrimination models are not universally applicable and cannot always be applied directly to all igneous rocks. A more pressing issue is that mafic or basaltic rocks formed through post-collision tectonics and exposed within the continental crust are often misidentified. These rocks are sometimes incorrectly classified as continental-arc-related products, leading to confusion in tectonic interpretations. Examples include the Somma-Vesuvius volcanic rocks in Italy, the Mahbubnagar Large Igneous Province in the Eastern Dharwar Craton in India, and the rear-arc of Muria Volcano in Central Java, Indonesia. While some propose that continental crust formation may result from both oceanic subduction and continental collision (cf. Zhu et al. [19]), these two processes are distinct in tectonic setting and origin. Consequently, an approach that selects the most sensitive diagram or a combination of models is essential for precise interpretations.

One widely used tectonic discrimination model is the Th-Nb-Yb systematics, specifically expressed through Th/Yb and Nb/Yb ratios; it was first proposed by Pearce and Peate [8] and later adopted by Pearce [13, 20] as the [13] diagram. This model effectively distinguishes tectonic settings between nonsubduction environments (e.g. midoceanic ridge basalts or MORB and within-plate basalts) and subduction-related volcanic arcs (continental and island arcs or IA). While Nakamura [21] has recently discussed this model, it remains limited in addressing igneous rock samples from outside established tectonic settings.

Over time, discoveries of igneous rocks with diverse geochemical signatures have prompted new classifications, such as alkaline arc [14], enriched tholeiitic [22], intracontinental rifting and continental extensional settings [9, 12, 23], continental arc magmatism due to oceanic slab break-off [18], and the magmatic evolution of asthenospheric mantle sources due to continental rifting. These newer tectonic settings are not fully accounted for in the original Th/Yb versus Nb/Yb model. This raises the question: can the Th-Nb-Yb systematics be further developed to accommodate these new tectonic types?

In this study, we aim to refine and expand the Th-Nb-Yb whole-rock geochemical model to uncover unmapped tectonic settings and improve magmatic alkalinity discrimination (tholeiitic, enriched tholeiitic, calc-alkaline, and alkaline). Our goal is to present a more comprehensive tectonic framework and to elucidate the interactions between magmas originating from asthenospheric and lithospheric mantle sources.

In magmatic system, Niobium (Nb) is one of the key geochemical elements, which is classified as a high-field-strength element (HFSE). It serves as a powerful indicator for distinguishing tectonic environments, mantle sources (asthenosphere and lithosphere), and the alkalinity [17, 23, 24]. Nb is highly stable, having only a single isotope (⁹³Nb) with a mole fraction of 100% [25], making it very stable in various geological processes and thus a reliable tracer of magmatic evolution. High Nb values indicate that the rock originates from enriched mantle sources, such as asthenospheric mantles and enriched lithospheric mantles (e.g. alkaline arcs and initial continental rifts), whereas low Nb indicates the magmatism derived from depleted MORB mantle (DMM) sources (e.g. normal MORB or N-MORB, IA, active continental margin/continental arcs or ACM) [14, 23, 26]. Besides that, the Nb/Ta ratio is also used to determine the presence of rutile eclogite melts in a magmatic system [27], particularly when combined with ratios such as Nb/Zr and La/Yb, which are effective in determining degrees of partial melting [18, 28].

Thorium (Th), another HFSE like Nb, is similarly classified as an incompatible element [29], with very low partition coefficients (D < 0.1) in most silicate minerals (e.g. plagioclase, olivine, pyroxenes, and feldspars), making it highly incompatible during magmatic processes [30-32]. The Nernst partition coefficient (D = C_s/C_l) describes the ratio between a trace element’s concentration in a solid phase (C_s) and its equilibrium liquid (C_l) [33]. Th and Nb preferentially remain in the melt phase during fractional crystallization [34, 35]. Th is particularly useful in identifying tectonic settings, including partial melting processes, continental crust contributions, and the presence of subducted sediment material [13, 17, 36]. High Th concentrations suggest contributions from continental crust or deeply subducted alkaline arcs [13, 18, 37], whereas low Th concentrations are associated with magmatism derived from IA or nonsubduction environments (e.g. MORB, oceanic island basalts or OIB, mantle plumes, and enriched tholeiitic continental rifts) [13, 15, 23, 38, 39].

Ytterbium (Yb) is a heavy rare earth element (HREE) that is highly immobile and more compatible than Th, Nb, or lighter rare earth elements (LREEs) such as La and Ce [30]. Yb exhibits high partition coefficients (D > 1), allowing it to readily integrate into mineral structures (e.g. orthopyroxene, clinopyroxene, and garnet) during magmatic differentiation [34]. As a result, Yb tends to remain in the solid phase and becomes enriched in nonalkaline mafic phases [34, 35]. Yb is specifically valuable for characterizing magmatic source regions, tectonic processes, and for distinguishing between continental and IA, as well as determining magma alkalinity when used in combination with Nb and Th ratios (e.g. Nb/Yb, Th/Yb) [13, 39].

Reconstructing tectonic discrimination models using the Th-Nb-Yb systematics presents significant challenges but enables broader tectonic classifications beyond the traditional division between subduction and nonsubduction environments, as first noted by Nakamura [21].

In this study, we develop a new tectonic discrimination model based on Nb/Yb and Th/Nb ratios. Th and Nb are relatively immobile elements, largely unaffected by rock weathering [22, 40, 41] and various geological processes [10, 42]. However, recent studies emphasize that the Th/Nb ratio can be significantly influenced by geological processes such as partial melting, assimilation and fractional crystallization (AFC), magmatic interaction, crustal contamination, metamorphism, and hydrothermal alteration [18, 29]. Because different tectonic environments produce asymmetric variations in Th and Nb concentrations, the resulting Th/Nb ratios can be used as critical indicators for tectonic discrimination. A case study on Slamet volcanics (Central Java, Indonesia; Harijoko [43]) illustrates this application (see online supplementary material 2).

The Th/Nb ratio is one of the key parameters in tectonic discrimination, as developed by Hawkesworth et al. [44], Pearce et al. [16], Shervais [17], and Godang et al. [18] and serves as a geochemical fingerprint to distinguish nonsubduction settings from subduction arc geochemistry in igneous rocks. For magmas originating from enriched mantle sources, characterized by Nb/Zr > 0.0627, a Th/Nb ratio of 0.67 has been applied to differentiate between asthenospheric sources (e.g. intracontinental rifts) and enriched lithospheric sources (e.g. alkaline arcs and early lithospheric rifts) [22, 38].

In this study, we replace Nb/Zr with Nb/Yb, as Nb/Zr signatures are traditionally used to distinguish between depleted and enriched mantle sources [45]. Here, the Nb/Yb ratio is further developed to differentiate between alkaline and subalkaline magmas (advanced after Pearce [24]), and it also helps identify magmatic interactions due to oceanic slab break-off or assimilation with deeper subduction of alkaline magma.

The new reconstruction model, shown in Figure 1, defines nine tectonic settings: {1} N-MORB, {2} enriched-MORB (E-MORB), {3} OIB and mantle plume, {4} intracontinental rifting, {5} IA, {6} ACM, {7} subalkaline enriched tholeiitic, {8} alkaline arc, and {9} continental extensional setting. The OIB and mantle plume settings are distinguished using an Nb/Zr ratio of 0.15 [12].

The fields of {1}, {2}, and {3} are mantle arrays, which are nonsubduction settings, purely adopted from Pearce [13]. The fields of {4} intracontinental rift settings with an additional mark of La/Yb = 20 for distinguishing enriched tholeiitic and alkaline are purely adopted from Saputro et al. [22], as well as the fields of alkaline arc {8} and continental extensional setting {9} [12, 23].

The discrimination of IA {5} and ACM {6} using Th/Yb 0.35 and 0.65 refers to the popular modeling from Barrett and MacLean [39]. The fields of {7} subalkaline enriched tholeiitic field is the section proposed in this study (the case study of the evolution of E-MORB magmatism will be presented in “Discussion-The evolution of E-MORB magmatism” section). The discrimination between subalkaline and alkaline using Nb/Yb = 6 is developed from Nb/Y = 0.6561 [24]. The boundary line of Nb/Yb = 4 for the slab failure or assimilation with alkaline magma is converted from Godang et al. [23]. The magmatic interaction line between asthenosphere and lithosphere mantle sources is adapted from Wei et al. [46].

This newly developed “Nb/Yb-Th/Nb” model integrates several existing submodels and is designed to be applicable to both mafic and felsic igneous rocks simultaneously. The AFC trend vector is converted mathematically from Pearce [13], which starts from the transitional zone toward the calc-alkaline and alkaline arcs (not starting from N-MORB) (Figure 1). It is because N-MORB originates from spinel-lherzolite [28], making it unlikely to evolve into a continental arc, which has a DMM garnet-lherzolite nature. Likewise, the trend of “the subduction vector” starts from the transitional zone toward the continental arc zone, which is characterized by an increase in Th/Yb and a slow and insignificant increase in Nb/Yb (Figure 1). This increase in Th/Yb, relative to Nb/Yb, is attributed to the involvement of continental material, which generally contains more Th than Nb [47].

The enriched lithospheric mantle source field is defined by Th/Nb > 0.67 and Nb/Yb > 6, which are alkaline that involve continental extensional zones (initial lithospheric rifts) and alkaline arcs (Figure 1). The discrimination between continental extensional and alkaline arc settings is determined using a combination of TiO₂ values relative to Al₂O₃ (proposed by Shaban et al. [38] based on the model of TiO₂ vs. Al₂O₃ from Müller and Groves [48]).

TiO₂ > (−1.1610 + 0.1935 × Al₂O₃) → within-plate continental extensional setting.

TiO₂ < (−1.1610 + 0.1935 × Al₂O₃) → alkaline arc setting.

(TiO₂ and Al₂O₃ values are expressed in weight percent (wt%)).

The alkaline arc setting is typical of deep subduction, which is characterized by the presence of feldspathoid minerals (such as leucite/pseudoleucite or nepheline), interplayed with the overriding continental plate and involves significant continental materials, reflected by higher Al₂O₃ derived from plagioclase melts from the continental crust [49]. Examples include the Roman Volcanic Province, Italy [50] and leucite tephriphonolite from Biga Peninsula, Northwest Anatolia [51].

By contrast, the within-plate continental extensional setting (= initial rift) is a new type of geodynamic setting that was recently identified by Godang et al. [23], based on their differences in geochemistry against mature continental rifts. It is formed by decompression melting of enriched subcontinental lithospheric mantle (SCLM) characterized by the presence of feldspathoid minerals, unassociated with asthenospheric mantle sources and not involving continental materials, reflected by a high TiO₂ value relative to Al₂O₃ (case study discussed in “Discussion-The continental extensional setting” section).

The newly developed Nb/Yb-Th/Nb model is constrained to the use of mafic to felsic igneous rocks across various tectonic settings (as shown in Figure 1). It is noteworthy to underline that this new diagram is not designed to discriminate continental collisional tectonics, since igneous rocks resulting from collisional tectonics generally contain geochemistry that can represent a variety of tectonic settings, such as IA, ACM, deep subduction alkaline arcs, continental rifts, MORB, OIB, and mantle plumes.

Even if this diagram is applied to igneous rocks formed directly from continental collision tectonics, it should be used only as a supporting tool rather than the primary method for tectonic discrimination. Case studies supporting this include the leucite-bearing potassic rocks from the Somma-Vesuvius volcanic complex in Italy (Guarino [52]) and the high-K mafic post-collisional rocks from the northwestern part of the Tibetan Plateau, China (Guo [53]), both discussed further in the online supplementary material 3.

Likewise, the well-known Javanese rear-arc geochemical dataset from the Muria Volcano in Central Java [54-57], which is a post-collisional product [18] but has characteristics of deep subduction alkaline arcs interplaying with within-plate features (such as OIBs or mantle plumes) as previously discussed in the “Introduction” section.

To better distinguish between syn- and post-collision tectonics and subduction arcs, it is necessary to include the Rubidium (Rb) element as an additional geochemical tracer, given it is more sensitive to collisional tectonic processes [18, 58, 59].

To ensure a robust comparison, we selected several case studies with ideal whole-rock geochemical data. Each case study clearly represents its respective tectonic setting and origin, untied from components associated with post-collisional tectonics and free from carbonatite metasomatism features (Online supplementary material), and the selected comparison diagram is a model that can be used for basaltic to felsic igneous rocks.

The typical geochemical data for N-MORB basalts are sourced from the South and West Indian Oceans [60], while the OIB data come from Mauna Kea basalt lava, Hawaii [61]. Plotting these data on the newly developed diagram reveals that N-MORB and OIB basalts fall within Th/Nb values of 0.04 to 0.11, with Nb/Yb < 2 and Nb/Yb > 6 (Figure 2).

The typical alkaline geochemical data representing intracontinental rifting tectonics are sourced from Aitutaki Island (Cook Islands, South Pacific Ocean) [37]. The basalts from Aitutaki Island are characterized by high TiO₂ (2.54%–2.75%) and high MgO (9.89%–12.92%). The plot results in Figure 2 fall within the geochemical range of 0.11 < Th/Nb < 0.67 and Nb/Yb > 6.

The typical geochemical data representing the IA are taken from basalt to dacite rocks from Candlemass and Vindication Islands (South Atlantic Ocean) [62]. These rocks have a low Th/La ratio (<0.2), indicating that their magmatism is unrelated to continental material [63]. Mineralogically, they are composed of bytownite to andesine (based on Cross, Iddings, Pirsson, Washington or C.I.P.W. mineralogy). As shown in Figure 2, the data fall within the range of Th/Nb > 0.11, Th/Yb < 0.35, low Nb/Yb (<0.3), and low-K, which are typical of volcanic rocks generated from tholeiitic magma. This is further evidenced by their low La/Yb ratios (0.51, 1.03) and lower Nb/Zr values (0.0132, 0.0096), which strongly correspond to tholeiitic characteristics [18, 23, 39].

To illustrate the transition from tholeiitic to calc-alkaline magma, we selected geochemical data from southern Pagan Island lavas (Mariana Arc, western Pacific Ocean) [64]. According to C.I.P.W. mineralogy, these lavas predominantly consist of labradorite plagioclase. Geochemically, they are characterized by Th/Nb > 0.11, Th/Yb values between 0.35 and 0.65, and low Nb/Yb (<1.00), as depicted in Figure 2.

For tectonic settings formed in the forearc ACM, we used geochemical data from the Kebasen volcanic rocks (basaltic lavas in the Banyumas Basin, Central Java, Indonesia) [65]. These rocks also primarily consist of labradorite plagioclases. Geochemically, they are defined by Th/Nb > 0.11, Th/Yb > 0.65, and low Nb/Yb (<1.00), as shown in Figure 2.

The most representative geochemical data for continental arc settings come from the Fossa delle Felci Volcano (Aeolian Arc, Salina, Italy) [66]. This volcanic sequence comprises basaltic-andesite to dacite rocks, likely derived from cogenetic magma reservoirs that underwent differentiation [67]. Mineralogically, these rocks consist predominantly of potash-labradorite to potash-andesine (based on the An-Ab-Or ternary diagram from Smith and Brown [68], using C.I.P.W. mineralogy). The plot results in Figure 2 show geochemical values of Th/Nb > 0.50, Th/Yb > 0.65, and low Nb/Yb (<4), which are indicative of volcanic rocks generated from calc-alkaline magma.

We compile the most typical magmatic examples that represent the interplay between continental lithospheric and asthenospheric mantle sources using data from the Sumbing-Slamet volcanic complex (Central Java, Indonesia) [43, 69-71]. The Sumbing and Slamet are Tertiary to Quaternary volcanoes formed by subduction between the Eurasian and Indian-Australian. Their deposits are primarily composed of lavas with a basaltic to andesitic composition and assemblages of phenocryst plagioclase, clinopyroxene, hornblende, and magnetite [43, 70]. Geochemically, these basaltic to andesitic rocks consist of medium-K to high-K. The plot results in Figure 3 show an increase in Nb/Yb values (from 1.3281 to 5.3395) and a decrease in Th/Nb values (from 2.400 to 0.2918).

Meanwhile, the ideal case study representing the interplay between asthenospheric and continental lithospheric mantle sources was selected from Guangxi Permian basalts, Western China [72]. Mineralogically, they are composed of potash-labradorite to normal-andesine ([68]; not shown). The presence of potash-labradorite and normal-andesine together in a magmatism indicates the presence of magmatic interaction. These basalts exhibit higher TiO₂ values relative to Al₂O₃ [TiO₂ > (−1.1610 + 0.1935 × Al₂O₃)] and are interpreted to be formed from within-plate tectonic settings [48]. The plot results show a decrease in Nb/Yb values (from 11.9381 to 8.1017) and an increase in Th/Nb values (from 0.1097 to 0.1697) (Figure 3), with La/Yb values ranging from 8.90 to 14.08.

The type of ideal E-MORB magmatism is selected from tholeiitic basalts of the continental flood (Franz Josef Land, Arctic Russia) [73]. It is composed of a low-K tholeiitic series and bytownite to labradorite plagioclase, which has a lower Mg# (59.10, 42.62), Nb/Yb values ranging from 2.5974 to 4.6263, Th/Nb values between 0.0783 and 0.1475 (Figure 3), and low La/Yb ratios (2.08, 3.58). The relatively low Mg# and higher Th/Nb > 0.11 indicate the presence of magmatic evolution on E-MORB [12, 74].

The typical geochemical data representing the continental extensional setting were obtained from the Adang volcanics (Western Sulawesi, Indonesia, new data from 12 drill cores) with a total of 54 samples [38]. They consist of leucite (pseudoleucite)-bearing silica-undersaturated rocks that are composed of high TiO₂ values relative to Al₂O₃ [TiO₂ > (−1.1610 + 0.1935 × Al₂O₃)], which are interpreted to have formed from within-plate tectonics. The plot results in Figure 3 show that they fall in the field of an alkaline arc or initial rift setting.

These case studies validate the effectiveness of the newly developed Nb/Yb-Th/Nb discrimination model in differentiating various tectonic settings and magmatic processes, further enhancing its applicability to diverse igneous rock compositions.

To validate our newly developed discrimination diagram, we compared it with several well-established models applicable to both basaltic and rhyolitic rocks at the same time. These models included the Th/Yb–Nb/Yb from Pearce [13], the Th–Nb from Saccani [15], and the Nb/Zr–Th/Nb diagram from Godang et al. [18]. Additionally, we also involved the partial melting of the mantle genome diagram “Nb/Zr–La/Yb” from Aldanmaz et al. [28].

The plot results on our newly developed diagram (Nb/Yb vs. Th/Nb; Figure 2) for N-MORB basalts from the South and West Indian Oceans [60] and OIB Mauna Kea, Hawaii [61] are identical to those plotted in Figures 4 and 5 from Pearce [13] and Godang et al. [18], respectively.

In Figures 2, 5, and 6, the N-MORB magmatism is observed to be influenced by asthenosphere upwelling, as indicated by a trend that leans vertically upwards. In contrast, the OIB magmatism is influenced by the oceanic environment, evident from a trend that leans vertically downward. Furthermore, the OIB Mauna Kea magmatism appears to be generated from primitive mantle (PM) sources (diagram Aldanmaz et al. [28]), which originated from asthenospheric mantle.

The plot results in Figure 2 for Aitutaki Island basalt, Cook Islands [37] fall within the field of intracontinental rifting. This plot is similar to the plot using the diagram from Godang et al. [18] (see Figure 5). Further, it exhibits a high La/Yb > 20, which is a characteristic of alkaline magmatism (refer to Saputro et al. [22]). This finding aligns with the plot in Figure 7 from Saccani [15]. Besides, the basalt has a high Nb/Zr (>0.15), suggesting that the magmatism originates from mantle plume sources (refer to Sun et al. [12]).

There is no difference in tectonic interpretation of the basalt to dacite rocks from Candlemass and Vindication Islands [62], which formed from IAs, when comparing their positions on Figure 2 with those on Figures 4, 5, and 7 (diagrams of Pearce [13]; Godang et al. [18]; and Saccani [15]). These rocks are tholeiitic in nature (as shown in Figures 2, 4, 5, 7, and 8).

The plot results for the southern Pagan Island lavas, Mariana arc [64] and Kebasen volcanic, Central Java [65] in Figure 2 are identical to those plotted in the Pearce diagram [13] (Figure 4). However, these two magmatic systems formed under different tectonic conditions (please see “Discussion-Frontal arc ACM” section).

The southern Pagan Island lavas are interpreted to have purely formed from oceanic-oceanic subduction, characterizing an IA (see also Figure 5). They have a low Th/La ratio (0.14–0.16), which indicates the absence of continental crustal material (according to Wang et al. [63]; signed by Th/La < 0.2). Due to their relatively high La/Yb and Th/Yb ratios, ranging from 3 < La/Yb < 6 and 0.35 < Th/Yb < 0.65, these lavas are identified as transition magmas from tholeiitic to calc-alkaline (see Figures 2 and 8). Furthermore, the partial melting process is triggered by the dehydration of marine sediment on the oceanic slab, as evidenced by a high Ba/Nb ratio (105.67–112.22) (Figure 9).

The Kebasen volcanic (Central Java) is indicated to have formed from the shallow subduction of the oceanic plate beneath the continental plate at the forearc of the ACM (Figure 5). This volcanic is continental tholeiitic in nature, as interpreted using the multi-cationic approach of De la Roche et al. [75] (see Fadlin et al. [65]), which is supported by their low La/Yb values (2.74–3.26; Figure 8). Additionally, the magmatism of Kebasen volcanic is interpreted to be uncontaminated by deep subduction magmatism, as evidenced by the plotted trend along the subduction vector line and indicated by a slow increase in Nb/Yb (Figure 2). Furthermore, the Kebasen magmatism is triggered by the dehydration of marine sediment on the oceanic slab, characterized by a high Ba/Nb ratio (71.00–124.28; Figure 9), alongside the involvement of subducted sediment input and continental crustal material, which is indicated by Th/La values greater than 0.2 (refer to Schaen et al. [76], and Wang et al. [63]).

It is noteworthy that the plot results in Figures 2 and 5 (Nb/Yb vs. Th/Nb and Nb/Zr vs. Th/Nb diagrams) differ from those in Figure 8 (Nb/Zr vs. La/Yb diagram), which are interpreted as calc-alkaline and tholeiitic-transition, respectively. The addition of subducted sediment input and the involvement of continental crustal material, reflected in the Th/La value exceeding 0.2, directly influence the Th/Nb ratio, resulting in higher values. However, this does not significantly affect the La/Yb ratio. This observation has direct implications for the plot results in Figure 8, which employs the La/Yb ratio (based on the diagram by Aldanmaz et al. [28]), reinforcing the conclusion that the Kebasen volcanic formed under tholeiitic conditions. This finding aligns with the interpretation derived from the multi-cationic analysis of De la Roche et al. [75] (see Fadlin et al. [65]).

The case study of Kebasen volcanic in Central Java highlights the significance of employing multiple models to achieve an accurate interpretation, as each discrimination diagram possesses distinct advantages that should not be overlooked.

The plot results of the basaltic-andesite to dacite from Fossa delle Felci Volcano, Italy [66] in Figure 2 indicate a 100% similarity in interpretation with Figure 4 (Pearce diagram [13]), which is formed from the continental arc, and the magmatism is influenced by deeper subduction characterized as alkaline in nature (as indicated by the AFC trend). The plot results using Nb/Zr versus Th/Nb, Th/Th_(N) versus Nb/Nb_(N), and Nb/Zr versus La/Yb diagrams also yield consistent outputs, classifying them as calc-alkaline continental arc rocks (Figures 5, 7, and 8 from Godang et al. [18]; Saccani [15]; Aldanmaz et al. [28]).

We specifically highlight the differences between Figure 2 (Nb/Yb vs. Th/Nb diagram) and Figure 4 (Th/Yb vs. Nb/Yb diagram from Pearce [13]) concerning alkalinity discrimination. Figure 4 displays an increase in alkalinity from calc-alkaline to shoshonitic, accompanied by a rise in Th/Yb during the magmatic differentiation from basaltic-andesite to dacite. Separately, Figure 2 demonstrates that alkalinity remains unchanged during the magmatic differentiation, and it stayed in the calc-alkaline condition, where the alkalinity feature is determined by the Nb/Yb = 6 (instead of Th/Yb). Additional comparisons involving more geochemical data can be found in Figures 10 and 11, as well as the relationship plots of Zr/Ti versus Nb/Y and K₂O versus SiO₂ (Figures 12 and 13).

Here, we propose that the Th/Yb ratio is only marginally applicable for distinguishing between tholeiitic and calc-alkaline magmas but is not suitable for discriminating alkaline compositions.

The plot results of the Sumbing-Slamet volcanics (Central Java) in Figure 3 (Nb/Yb vs. Th/Nb diagram) are similar to those in Figure 6 (Nb/Zr vs. Th/Nb diagram), which indicates that both magmatisms resulted from the interplay between continental lithospheric and asthenospheric mantle sources due to slab break-up of the oceanic plate beneath the northern part of Central Java. This interpretation is 100% conformable to the previous opinion proposed by Godang et al. [18]. The interaction is marked by a decrease in the Th/Nb value from 2.400 to 0.2918, alongside an increase in the Nb/Yb from 1.3281 to 5.3395, breaking through the limit of Nb/Yb = 4 (see the newly developed model in Figure 3).

The plot results of the Guangxi Permian basalts [72] in Figure 3 fall within the intracontinental rift field, indicating that the magma originated from asthenospheric mantle sources and was contaminated with approximately 30% crustal material. This interpretation is similar to the plot in Figure 6 and consistent with the author’s findings. It is marked by an increase in Th/Nb from 0.1097 to 0.1697 and a decrease in Nb/Yb from 11.9381 to 8.1017 (see Figure 3). Moreover, the Guangxi basalts are interpreted as enriched tholeiitic, even though the Nb/Yb value is higher than 6, as they originate from an enriched mantle source and exhibit relatively low La/Yb values (<20; see Figure 8).

Comparative case studies for determining magmatic alkalinity are also presented, including the Nyiragongo volcanics from the Western Rift of the East African Rift system (Chakrabarti et al. [77]) and volcanics from Pantelleria and Linosa Islands in Italy (White et al. [78]), both of which are derived from within-plate upwelling asthenospheric mantle plumes (refer to online supplementary material 4).

The plot results for the tholeiitic basalts from Franz Josef Land [73] in Figure 3 fall within the E-MORB and subalkaline enriched tholeiitic fields. This magmatism is interpreted as resulting from a mix of oceanic-continental rifting, originating from the asthenospheric mantle with E-MORB characteristics, which also exhibit subalkaline-enriched tholeiitic traits. This interpretation is similar to the plot in the Nb/Zr versus Th/Nb diagram (Figure 6) and matches the plot results in the Saccani diagram [15] (Figure 7).

We specifically plotted datasets of enriched tholeiitic basalts from the Northern Atlantic Ocean of the 15°20'N fracture zone [79]. The result of plots shows that it falls in the E-MORB field. This interpretation fits with what the author found, which was that the magmatism in the 15°20'N fracture zone comes from E-MORB genesis (see Figures 3, 6, and 7).

The difference between the two enriched tholeiitic basalts above is that the tholeiitic from the 15°20'N fracture zone did not go through magmatic evolution (marked by Th/Nb < 0.11). In contrast, the tholeiitic Franz Josef Land has undergone magmatic evolution, shown by Th/Nb > 0.11 (see Figures 3 and 6).

The plots of the Adang volcanics in Figure 3, combined with the equation TiO₂ > (−1.1610 + 0.1935 × Al₂O₃) [48], show that the magmatism of Adang Volcano was formed from within-plate continental extensional settings and is characterized by an alkaline nature (Figure 3). These tectonic interpretations match the plot in the Nb/Zr-Th/Nb diagram (Figure 6). Furthermore, the within-plate continental extensional setting is associated with early lithospheric rifts, where magmatism is the result of SCLM originating from an enriched DMM source (see Figure 8) that is unrelated to the deeper mantle sources (asthenospheric mantle). Our final outputs are consistent with a recent finding from Godang et al. [23].

The outcomes of our review and the development of new geochemical modeling for the Th-Nb-Yb systematics led us to the following conclusions:

(1) The geochemical modeling is reconstructed in terms of Nb/Yb versus Th/Nb. This newly created diagram successfully captures new types of tectonic settings.

(2) Several new tectonic settings have been identified in the diagram, including: (1) intracontinental rifting, which originates from the continental environment; (2) mixed oceanic-continental rifts, which produce low-alkaline-enriched tholeiitic magmas; (3) within-plate continental extensional settings (early lithopheric rifts) or alkaline arcs.

(3) Moreover, this newly created diagram can also identify magmatic interactions that formed from ACM tectonics due to oceanic slab break-off.

(4) Re-identifying magmatic alkalinity discrimination using combinations of Th/Yb, Nb/Yb, and La/Yb in diagrams of Nb/Yb-Th/Nb and Nb/Zr-La/Yb has proven effective in distinguishing between tholeiitic, calc-alkaline, enriched tholeiitic, and alkaline rocks.

The raw data supporting the conclusions of this article will be made available by the author, with due reservation.

The authors declare that there is no conflict of interest regarding the publication of this paper.

This study was supported by the National Natural Science Foundation of China (No. 92162103); the Natural Science Foundation of Hunan Province (No. 2022JJ30699, No. 2023JJ10064); and the Science and Technology Innovation Program of Hunan Province (No. 2021RC4055, No. 2022RC1182).

The authors express their gratitude to Chuan-Lin Zhang (associate editor) for the efficient handling of this paper and to reviewers for providing valuable comments for the paper.

Appendix A:

Table 1: Major elements.Table 2: Trace elements.

Appendix B: Th-Nb relatively immobile elements.

Appendix C: Post-collisional settings and Tibetan potassic rocks.

Appendix D: East African rift.