Geochemistry Study of Cross-castic Magma Alkalinity Evolution

The discrimination of magmatic alkalinity is a classic study that has never stopped for the past ninety years. Various methodologies have been developed since Shand’s classification using the method of alumina saturation to approach silica saturation and the methodology without involving alumina and silica such as K 2 O vs. Na 2 O and others, while the aim is to find out the evolution of alkalinity during the magmatic differentiation. The classical magmatic alkalinity evolution has been known as a castic magma alkalinity evolution, where the initial magma in the form of magma-X(a) will evolve along the stages of differentiation and remain a derivative of the initial magma {magmaX(a)}. The same philosophy is also explained in the ternary AFM diagram. Is the magmatic differentiation, followed by fractional crystallization, always an evolution of alkalinity based on caste? This question often raises current debates. This study takes the example of cogenetic volcanic and albitites. The application of the cogenetic volcanic using the selected diagram, which is ‘Three in one an overlaid diagram’. The output of the diagram presents the differentiation of magma which based on the evolution of Mg-series and Fe-series in a discontinuous branch of Bowen 1922 that can take place the castic and cross-castic, e.g. (a) from Mg-series to Mg-series {castic}, (b) from Mg-series to Feseries {cross-castic}, (c) from high-Mg tholeiitic basalt to calc-alkaline series {cross-castic}, (d) from Fe-series to Fe-series {castic}. While the evolution of magmatic alkalinity based on the continuous branch and refer to Trapezoid model generally occurring a cross-castic, e.g . (A) from sodic calc-alkaline to sodic alkaline-calcic, (B) from sodic calc-alkaline to shoshonitic alkaline-calcic, (C) from sodic calc-alkaline to potassic calc-alkaline, (D) from potassic calc-alkaline to shoshonitic alkaline-calcic, (E) sodic alkaline-calcic to sodic alkaline/peralkaline, (F) shoshonitic alkaline-calcic to potassic/ultrapotassic alkaline-calcic (cross-castic in subalkaline), (G) shoshonitic/potassic alkalinecalcic to shoshonitic/potassic alkaline/peralkaline. In this study, Fossa delle Felci volcanics (Italy) shows the evolution of magma from Mg-series to Mg-series, but the evolution of alkalinity of magma reveals the cross-caste (from sodic calc-alkaline to shoshonitic alkaline-calcic). Salak volcanics (Western Jawa) shows the evolution of magma from the Mg-series to Fe-series (cross-castic), and also the cross-castic in the evolution of alkalinity from sodic calc-alkaline to alkaline-calcic. Gothara albitites (India) clearly reveal the sodic-rich alkaline, which the magma generates from the evolution of sodic alkaline-calcic to sodic alkaline without the presence of potassic.


Introduction
The discrimination of magmatic alkalinity has been the subject of ongoing debate over the past nine decades. Various models have been developed since Shand's classification for peralkaline against calc-alkaline magma (Shand, 1927), including the recent Trapezoid model which was developed by Godang et al. (2016) for the classification of (sodic/potassic) calc-alkaline, (sodic/ shoshonitic/potassic) alkaline-calcic, alkaline, and peralkaline igneous rocks.
Determination of magmatic alkalinity can be done using the alumina saturation method (Shand, 1927 and1943;Whalen et al., 1987), silica saturation (Peacock, 1931;Rittmann, 1957Rittmann, , 1962MacDonald and Katsura, 1964;Miyashiro, 1978;Peccerillo and Taylor, 1976;Keith, 1983;Frost et al., 2001;Calanchi et al., 2002), K 2 O vs. Na 2 O without involving the alumina and silica saturation (Middlemost, 1975;Turner et al., 1996), and a combination of alumina and/or silica saturation with K 2 O and Na 2 O {Wright, 1969 andFadlin et al., 2018 (add CaO); Ishihara and Murakami, 2004;Godang et al., 2016}(Figure 1). The classic magmatic alkalinity evolution is where the initial magma in the form of magma-X(a) will evolve along the stages of differentiation and remain derivatives of the initial magma {magma-X(a)} (see in Figure 2). The same philosophy is also explained in the ternary AFM diagram as proposed by Irvine andBaragar, 1971 (in Wilson, 1989). The classic magmatic alkalinity evolution in this study is referred to as "castic magma alkalinity evolution". Is the magmatic differentiation, which is followed by fractional crystallization, always an evolution of alkalinity based on caste? This question often raises the current debates. The term of caste is adopted from Hinduism, and the castic pertains to the caste. The ternary Jensen cationic model (Jensen, 1976) is one model diagram that presents a quite different definition, that there are two types of tholeiitic namely high-Mg tholeiitic basalt (primary magma) and high-Fe tholeiitic basalt (differentiated magma). High-Mg tholeiitic basalt will evolve into calc-alkaline series ('cross-castic'), whereas high-Fe tholeiitic basalt will evolve into tholeiitic series. Aso volcanic rock (Kyushu arc, SW Japan) is one of the cases where Hunter (1998) clearly mentioned that the dominantly tholeiitic magma evolved into dominantly calc-alkaline (see also the compilation geochemistry data from Shibata et al., 2013). A similarity also occurs at the Okinawa Trough volcanic rock (active back-arc basin, East China Sea; Ishizuka et al., 1990). Another review, in which the single chain of evolution of the plagioclase during the magmatic differentiation explains the similar philosophy that the cross-castic alkalinity evolution of (calc-alkali)-plagioclases to (alkalicalcic)-plagioclases (origin of Bowen, 1922, p.190; see also in Figure 3).
The aim of this geochemistry study is to provide a clearer picture of the magmatic alkalinity evolution in the continuous branch of Bowen (1922) and the magma evolution of Mg-series and Fe-series in a discontinuous branch. The methodologies use multi-discrimination of alkalinity diagrams, Mg-series vs. Fe-series diagram, tholeiitic vs. calc-alkaline diagrams, Trapezoid model, and a new simple ternary magmatic alkalinity evolution model (CaO-K 2 O-Na 2 O).

New Model Magmatic Alkalinity Evolution
The new magmatic alkalinity evolution trend diagram presented here has been designed for non-alkaline/peralkaline magma in the form of a ternary model using components of the feldspar group (plagioclase and K-feldspar) namely in the form of variables of CaO, K 2 O, and Na 2 O, minus Al 2 O 3 (in wt.%). This ternary diagram is based on a conversion (modification) of the Trapezoid model (Godang et al., 2016), referring to the modified of continuous branch from Bowen's 1922 (Figure 4), and adopting the affirmation of the diagram 'K 2 O vs. SiO 2 ' (Harker, 1909) where there is an increase in K 2 O along the stages of magma differentiation (from basalt/gabbro to rhyolite/granite). The evolutionary trend of magmatic alkalinity from the Trapezoid Model is shown in Figure 5.  Frost and Frost, 2014).

O c e a n i c p l a g i o g r a n i t e s a n d a l b i t i t e s
tr a n si ti o n a lk a li n e  Figure 5. The cross-castic magma alkalinity evolution trend in Trapezoid model (Godang et al., 2016). (A) Evolutionary trend of sodic calc-alkaline to sodic alkaline-calcic, (B) Sodic calc-alkaline to shoshonitic alkaline-calcic, (C) Blue dash-line shows the sodic calc-alkaline to potassic calc-alkaline, (D) Potassic calc-alkaline to shoshonitic alkaline-calcic, (E) Blue dash-line shows the evolution of oceanic plagiogranites/albitites from sodic alkaline-calcic to sodic alkaline/peralkaline. (F) Red dash-line shows the evolution of S-type granites from shoshonitic alkaline-calcic to potassic/ultrapotassic alkalinecalcic (cross-castic in sub-alkaline), (G1, G2) Evolution of A-type granites from shoshonitic/potassic alkaline-calcic to shoshonitic/potassic alkaline/peralkaline.

Magmatic Source
To anticipate the possibility of magma contamination from the other magma during the process of magmatic differentiation, such as mantle magma containing rutile-melts, or magma that has interacted with the mantle plume; then the multiplotting is done as an initial step by using various diagram models. Plot of the major oxides in the Shand classification diagram shows the Salak volcanics, Fossa delle Felci volcanics, representative of calc-alkaline volcanic rocks (RCVR), and Gothara albitites that have the value of ASI < 1.1 and ascertained the origin of the magma is formed from the igneous protoliths ( Figure 6).
The crosscheck of the possibility of magma contamination with other magmas containing the rutile-melts, then the plotting is expanded by involving the ratio of Nb/Ta vs. SiO 2 (after Foley et al., 2002 andafter Asaah et al., 2014). The plotting results show the cogenetic magmatic differentiation, free of contamination from rutile-melts ( Figure 7). The ratio of Nb/Zr vs. Th/Zr for the Salak volcanic and Fossa delle Felci volcanics respectively in the range 0.0416 -0.0444 vs 0.0186 -0.0263, and 0.049 -0.0598 vs 0.0418 -0.0610 indicate that both magmatisms were generated from ACM (Continental Arc) which was not contaminated with mantle plume (after Sun et al., 2006, modified from Godang et al., 2016. The diagram using the ratio of Nb/Zr vs. SiO 2 shows the Fossa delle Felci and Salak volcanics respectively in the form of a single magma source that undergoes an ideal differentiation, and has the ratio values Nb/Zr <0.0627 strongly indicating the magma was generated from DMM (Depleted MORB Mantle). Furthermore, it is clear that there is no increase in the Nb/Zr ratio from the basaltic to dacitic composition during the magma differentiation and fractional crystallization ( Figure  8). To anticipate more possibilities for sediment recycling input, which generally carries the potassic and can influence the increase of potassic (K 2 O) in the magma source during the differentiation process, then the additional plotting was also done using the ratio Th/Ce vs. SiO 2 (after Hawkesworth et al., 1997;after He et al., 2008;Figure 6. Aluminum Saturation Index (ASI) Diagram (rectified after origin Shand, 1927). The discrimination of Alkaline ((NK)/Al > 0.85) is adapted from Whalen et al., 1987 Roex et al., 1983).  75  74  73  72  71  70  69  68  67  66  65  64  63  62  61  60  59  58  57  56  55  54  53  52  51  50  49  48  47  46  45  44  43  42 Plank and Langmuir, 1998), SS (Silicic sediments; Gasparon and Varne, 1998), PAAS (Post-Archean Australian Shale; Taylor and McLennan, 1985), NASC (North American Shale Composite; Gromet et al., 1984).  (Hartono, 1994). Jawa and Indian ocean sediments (Handley et al., 2008). Sediments (Aizawa et al., 1999). Muria and Merapi volcanics (Sendjaja et al., 2009). Italian lamproites (Kempton et al., 2018). Aeolian Islands (Francalanci et al., 2007). Symbol in Figure 9 &  (Figure 9). In contrast to Merapi volcanics, the results of plotting the isotopic composition are very close to the sediment line and the Indian Ocean sediment boundary.
A plot of the Fossa delle Felci and Salak volcanics in a spidergram displays an ideal trend of incompatible to compatible-trace elements ( Figures 11a, and b)and incompatible to compatible REE (Figures 12a and b). The enrichment in  Sun (1995) and Depleted Mantle (DM) from Salters and Stracke (2004).    Based on the result of the multi-plotting diagrams (Figures 6 to 12), it can be concluded that Fossa delle Felci volcanics (Italy) and Salak  Sun (1995) and Depleted Mantle (DM) from Salters and Stracke (2004).  (McDonough and Sun, 1995) deple ted mant le chondrite

Interpretation of results
Three in one overlaid diagram is the overlay diagram constructed from three existing diagrams using the major oxides, i.e. Mg# vs. SiO 2 (Schilling et al., 1983), FeO (t) /{FeO (t) +MgO} vs. SiO 2 (Frost and Frost, 2008), and FeO (t) /MgO vs. SiO 2 (Miyashiro, 1974). The diagram in Figure 13 shows a difference in discrimination between tholeiitic (Fe-series) and calc-alkaline (Mg-series). The interpretation of plot results is presented in Table 1.
The plot results in TAS diagram (after Le Bas, 1986) show RCVR, Fossa delle Felci, and Salak volcanics fall in sub-alkaline field (calc-alkaline series) (Figure 14), but the plot result in 'K 2 O vs. SiO 2 ' diagram (after Peccerillo and Taylor, 1976; Figure 15) shows the magmatic alkalinity of RCVR and Fossa delle Felci evolved from medium-K to high-K (cross-castic); whereas the Salak volcanics (basaltic-andesite to andesite) has a trend of cross-castic evolution. The result of plotting in the overlay diagram between Ishihara and Murakami (2004) and Godang et al. (2016), Trapezoid model (Godang et al., 2016), and New ternary magmatic alkalinity evolution trend diagram that created in this study (Figures 16, 17, and 18) are presented in Table 2.

Discussion
There are differences in the results of the plot on the overlay diagram between the versions of Miyashiro (1974) and Frost and Frost (2008) (Figure 13). The difference between these two versions of discrimination will continue to be a considerable debate in the future. The case study Figure 13. Three in one overlaid diagram. Index of differentiation diagram (Mg#; Schilling et al., 1983); Magnesian vs. Ferroan series (after Frost and Frost, 2008); Discriminates of Arc Tholeiites and Arc Calc-alkaline series (Miyashiro, 1974), Mantle-melts (after Kinzler, 1997). HMAs melts (High Magnesian Andesite; Kelemen, 1995), Lower crust (Rudnick and Gao, 2003).
The plot results on Figure 16 show RCVR and Fossa delle Felci volcanics reveal the cross-castic magma alkalinity evolution from sodic series to shoshonitic series; whereas Salak volcanic rock has a trend of cross-castic evolution. The dis-crimination line (dash lines), proposed by Ishihara and Murakami (2004), is only used as a reference which in principle explains about the cross-castic alkalinity evolution.
Plotting in the Trapezoid model ( Figure 17) provides the detailed output of magmatic alkalinity evolution, in which RCVR and Fossa delle Felci volcanics show the cross-castic from 'sodic calcalkaline' to 'shoshonitic alkaline-calcic', whereas Salak volcanic rock shows the cross-castic from 'sodic calc-alkaline' to 'sodic alkaline-calcic'. The Gothara albitites (India) clearly confirmed as sodic-rich alkaline (see also Figure 6); furthermore, Trapezoid model also displays the evolutionary trend of plagiogranites/albitites in the form of cross-castic from sodic alkaline-calcic to sodic alkaline/peralkaline without the presence of potassic. The plot results of the Gothara albitites in the new ternary CaO-K 2 O-Na 2 O diagram are interpreted as a sodic-rich series (Figure 18). The ternary diagram is a crosscheck for the Trapezoid model, which has  Fig. 6 and Fig. 17} (b) after Peccerillo & Taylor, 1976

I J O
G the same expression in discrimination of alkalinity series. This ternary diagram is just a simple diagram only used for discriminating alkalinity series. It couldn't further define such as calc-alkaline, alkaline-calcic, alkaline or peralkaline, because it does not involve the variable of alumina. The magmatic evolution of (Mg, Fe)-series in discontinuous branches do not have a direct relationship with the evolution of alkalinity in continuous branches (e.g. sodic, shoshonitic, potassic, alkaline/peralkaline), because each branch is on its own.
Finally, after going through the comprehensive studies above, it is proposed that the magmatic evolution of the Mg-Fe series such as the discon-tinuous branch of Bowen (1922) is different from the magmatic alkalinity evolution (continuous branch). The magmatic evolution (Mg-Fe series, tholeiitic vs. calc-alkaline series) can take place the castic and cross-castic (Table 3, see also Figure  13), as well as the evolution of magmatic alkalinity can also occur in castic and cross-castic. The evolutionary of magmatic alkalinity which refers to the TAS diagram (such as Rittmann, 1957;Miyashiro, 1978) or the diagram of K 2 O vs. SiO 2 (Peccerillo and Taylor, 1976) takes place in castic, the New Ternary Magmatic Alkalinity Evolution diagram can take place in castic and cross-castic, whereas the Trapezoid model generally occurs in cross-castic (Table 4, see also Figures 2 and 5).  Miyashiro (1974) Calc-alkaline to Calc-alkaline.
(4) New ternary magmatic alkalinity evolution trend (created in this study), see also in Fig. 18 * sodic series to sodic series; * shoshonitic series to shoshonitic series.