Petrogenesis of high-Mg andesites from the Proterozoic Shillong Basin, Northeast India: evidence for continuation of the Central Indian Tectonic Zone to the Assam-Meghalaya Gneissic Complex and its implications for the Columbia supercontinent reconstruction

ABSTRACT This study provides comprehensive geochemical and petrological data for the newly discovered high-magnesium andesites (HMAs) from the Shillong Group of rocks in the Proterozoic Shillong Basin of the Assam-Meghalaya Gneissic Complex (AMGC). It is the first report of HMAs from the easternmost part of the Indian shield region. The AMGC holds immense geodynamic significance because this region has recorded magmatism related to Columbia, Rodinia, and Pangea supercontinental events. The present work discusses geochemical signatures of the Shillong HMAs characterized by calc-alkaline traits with high-K contents and significant concentrations of TiO₂ (0.54–0.76 wt %) higher than boninites (TiO2 < 0.5 wt %), and Mg# (53–56). They exhibit high LILE/HFSE and LREE/HREE ratios, which are distinctive features of magmas generated in subduction zones. Tectonic discrimination diagrams clearly suggest that the Shillong HMAs were generated in a back-arc basin regime. We argue that the Shillong HMAs were derived by 5% to 10% partial melting of a depleted lherzolite mantle source within the spinel – garnet transition zone. We also argue that the Shillong HMAs may have formed as a result of the interaction between subducted sediment-derived melts and mantle. The Shillong HMAs have undergone low- to medium-grade greenschist facies metamorphism characterized by significant mineral alteration, however, they still preserve bulk original alteration-resistant trace element compositions. From a regional perspective, the synchronous nature of origin and geotectonic setting, associated geochronological ages of formation, and regional structural trends suggest that the Shillong Basin is the easternmost continuation of the Mahakoshal Mobile Belt and specifically, the Bathani volcano-sedimentary sequence of the Central Indian Tectonic Zone (CITZ); and together they form the northern fragments of the same suture zone consisting of the CITZ, the Chotanagpur Granite Gneiss Complex, and the AMGC, which demarcates the northern and southern Indian blocks of the Greater Indian Landmass.


Introduction
One of the rarest rock types to occur on the modern Earth's surface is the high-magnesium andesites (HMAs).The genesis of such volumetrically sparse magmas is an extremely important topic in geochemistry as it may have contributed greatly to continental crust formation in the Earth's early history (Kelemen 1995;Rudnick and Fountain 1995).Hence, unusual tectonic conditions must exist for the production of such rare magmas.The current conventional hypothesis for the genesis of HMA magmas states that they form as a reaction between slab-derived partial melts and the wedge mantle in convergent boundaries and suture zones (Kelemen 1995;Tang and Wang 2010;Liu et al. 2018).Hydrous melting of upper-mantle peridotites, triggered by direct addition of slab-derived aqueous fluids, is likely the root cause for the production of such magmas (Crawford et al. 1989).Apart from the mantle wedge melt, other factors like subduction-derived fluids (Ishikawa et al. 2005), and melting of subducted sediments (Defant and Drummond 1990;Shimoda et al. 1998;Tatsumi 2001;Hanyu et al. 2002;Zeng et al. 2016) also trigger the formation of such magmas.Furthermore, the geothermobarometric conditions of the HMAs may provide insight about the tectonic evolution prevalent in subduction zones and convergent margins.These rocks also contain important information about elemental transfer in such zones (Tatsumi and Hanyu 2003;Wood and Turner 2009;Wang et al. 2011;Ma et al. 2015;Liu et al. 2018).Macpherson et al. (2006) have suggested that crystal fractionation of low Mg# garnet from mantlederived basaltic magmas, with or without slab components, may produce such magmas.On the other hand, mixing between crustal-derived dacitic melts with high-Mg basaltic magmas may also lead to the generation of high-Mg andesitic magmas (Streck et al. 2007).
The rare occurrence of HMAs is also prevalent in the crustal domain of the Assam-Meghalaya Gneissic Complex (AMGC) of Northeast India, which is now widely accepted as the easternmost extension of the Central Indian Tectonic Zone (Acharyya 2003;Saikia et al. 2017).Diverse records of geological events spanning from Proterozoic to Cenozoic can be found in this region.Two prominent massifs, namely the Meghalaya and Mikir Massifs having an overall area of approximately 7,614 sq km, separated from the rest of the Indian Peninsula by the Garo-Rajmahal Trap constitute the AMGC (Evans 1964).The AMGC is a strategically significant domain and is considered as one of the oldest rift-related, intracratonic Purana basins of the Indian subcontinent.The Proterozoic Shillong Basin formed within the Shillong Plateau during Paleo-Mesoproterozoic Era followed by sequential deposition of thick piles of sedimentation and extensive magmatism and volcanism and subsequent metamorphism on a regional scale.It houses a well distinguished meta-volcanogenic succession along with associated metasediments, which are all classified as the Shillong Group of rocks.They unconformably overlie the Basement Gneissic Complex (BGC) and are exposed in the Shillong Plateau as well as in Mikir Hills of Assam.The Shillong Group of rocks was first intruded by dolerite sills and dykes, which were later partially metamorphosed to a green coloured rock locally named as Khasi Greenstone.The underlying BGC and the Shillong Group were later intruded by several discordant granitic plutons towards the Neoproterozoic Era, which are both porphyritic and non-porphyritic in nature.Named after the locality found in, the Kyrdem and Mylliem plutons are intrusive into the Shillong Group of rocks, whereas the Nongpoh and South Khasi batholiths intrude the Shillong Group as well as the BGC.Thus, the rocks of the Shillong Basin are classified as Paleo-Mesoproterozoic, Neoproterozoic to Early Paleozoic in age and the underlying BGC is considered as Paleoproterozoic in age (Mitra and Mitra 2001;Bidyananda and Deomurari 2007;Chatterjee and Ghose 2011).
Previous literature has primarily focused on the granites and dolerites of the Shillong Group, but its associated counterpart, the intermediate to felsic volcanic rocks that occur within the meta-volcano-sedimentary sequence has not been studied to such an extent.As a matter of fact, the present work happens to be the first to report the occurrence of HMAs in the Proterozoic Shillong Basin.The available published works are mostly confined to limited and sporadic field areas within the Shillong Basin without any facies classification or corelatable tectono-stratigraphy. Therefore, this paper gives a detailed debut regarding the intermediate volcanics of the Shillong Group.Here, petrological study of the HMAs involving extensive field, detailed petrographic work, geochemical analysis involving major and trace element data and mineral chemical work are done to achieve the desirable objectives.We aim to discuss the most probable mechanisms for the genesis, tectonic setting and characteristics of the magmatic source, which will help in the comprehension and development of a precise geodynamic evolution model of the Shillong Basin during the Proterozoic Era.This study represents the first inclusive and detailed investigation of the andesitic rocks in the Shillong Plateau and an appraisal of these rocks may provide insight in the geological evolution and terrain amalgamation of the Indian subcontinent in supercontinent reconstruction models.

Geological setting
The Assam-Meghalaya Gneissic Complex (AMGC) or the Shillong Plateau, which is the northeastern extension of the Indian Peninsula, occupies a unique position amidst the Precambrian shields of Indian stratigraphy.Geologically, the Proterozoic NE-SW trending Shillong Basin covers the central-eastern part of AMGC and extends northeast to its twin part, Mikir Massif, both are separated from each other by the NW-SE trending Kopili Fault (Figure 1).The AMGC is a complicated tectonic zone, due to it being caught in between two young subduction regions, which formed the east-west trending Himalayas to the north and northeast, and the north-south trending Indo-Burma Range to the east and southeast and is separated from Peninsular India by the Garo-Rajmahal gap.Several lineaments, trending NE-SW and E-W are seen in AMGC; the most prominent one being the E-W trending Dauki fault, which marks the southernmost boundary of the plateau.
The AMGC is composed of different types of rocks, starting from the oldest and most dominant Proterozoic rocks of gneissic composition, which is considered to be the basement.It includes biotite-gneiss, biotitehornblende gneiss, sillimanite-gneiss and granitic gneiss.Granite gneisses are the dominant lithological unit of the basement gneissic complex and are exposed mainly in the central and northern parts of the Shillong Plateau.After the preliminary basement formation as early as 2670 Ma (Bidyananda and Deomurari 2007;Majumdar and Dutta 2016), numerous generations of subsequent magmatism occurred periodically from 1700 Ma till 117 Ma, generating igneous rocks of varied nature, which now occur as granite plutons and basic intrusives.The basement gneissic rocks are unconformably overlained by the Shillong Group of metasedimentary rock cover (Figure 2(a)).Heterogeneous Proterozoic granite plutons, ranging from equigranular to porphyritic and medium to coarse grained, subsequently crosscut the basement gneisses and metasediments of the Shillong Group at various places.Doleritic dykes of Mesoproterozoic age, popularly known as the Khasi Greenstone, which are a collection of basic intrusives within the Shillong Group of rocks that underwent lowgrade metamorphism, are also observed in the Shillong Plateau (Mazumder 1986).The southern part of the plateau is covered by Cretaceous Sylhet basalts and Tertiary shelf sediments.
The rift-controlled Shillong Basin is located in the central and eastern parts of the AMGC, where sedimentation as well as extensive magmatism and volcanism occurred throughout the Proterozoic Eon.The Shillong Basin contains the Shillong Group of rocks, which are metamorphosed to greenschist facies and rests unconformably, as indicated by a basal conglomerate (Nandy 2001), on an array of rock types consisting of sillimanite-bearing gneisses, amphibolite, banded iron formation, granulites and granite gneisses together known as the basement gneissic complex.The boundary between the Shillong Group of rocks and the underlying basement gneisses is well defined.The Paleo-Mesoproterozoic Shillong Basin comprises metavolcanics and metasediments, which is further divided into Upper Quartzitic Formation and Lower Phyllitic Formation.The Lower Phyllitic Formation comprises mainly phyllite, quartzite, slate, conglomerate, and schists with calc-silicate rocks, while the Upper Quartzitic Formation consists mainly of quartzites intercalated with phyllite and conglomerate.The Shillong Group metasediments have been dissected by Mesoproterozoic dykes and sills named as Khasi Greenstone.The Shillong Group of rocks is further intruded by Neoproterozoic and Cambrian granitic plutons.

Field relationships
The Shillong Basin consists of a well distinguished meta-volcano-sedimentary succession exposed in the Shillong Plateau as well as Mikir Hills of Northeast India.The NE-SW trending meta-volcano-sedimentary sequence, having a regional dip towards SE, is restricted to approximately an aerial distance of 2500 sq km with the type area being located in the south western part of the basin around Sohiong and along the Umiam River section in the central part of the present-day basin configuration of the Shillong Plateau (Naik et al. 2020).It comprises of intermediate to felsic lava flows and pyroclastic rocks, along with epiclastic deposits and metasedimentary bands.The volcanogenic sequence comprises coherent andesitedacite-rhyolite lava flows and scoria beds, whereas different variants of tuff and ignimbrites constitute the pyroclastic deposits (Figure 2(b)).The metavolcanics occur as alternating, interbedded sequences within thick stratigraphically distinct horizons of metasedimentary tuffaceous phyllite and argillite of Lower Phyllitic Formation and quartzite of the Upper Quartzitic Formation.The metavolcanics can be traced to about 40 km from SW (Sohiong area) to NE (Sonidan) part of the basin in Meghalaya but are not compositionally and petrographically similar.
The meta-volcano-sedimentary sequence records a wide range of rock types, comprising of lithounits that lie in close association with one another.They are described as follows:

Volcanogenic sequence
The volcanogenic sequence consists of 1 to 10 m thick intermediate to felsic lava flows and volcanoclastic rocks, comprising basically of andesitic and rhyolitic flows and dacitic-rhyolitic ignimbrites as well as volcanoclastic conglomerates.

Lava flows
The aphyric coherent lava flows are massive and hard in nature with very fine-grained groundmass noticeable in fresh exposed outcrops (Figure 3(a-c)).At some places, they exhibit flow banding, whereas in other locations, scoriaceous vesiculated lava flows are well-preserved.The scoria beds exposed in the study area consist of effervescent glassy lava flow which has more than 40% vesicles by volume (Figure 3(d)).The rock type is aphanitic, mesocratic, mostly weathered, and restricted to south western part of the basin.Close-up photos of the rock samples are presented in Figure S1.

Pyroclastic deposits
The pyroclastic deposits are primary explosive deposits and classified based on their size, shape, type, amount of constituent materials and their primary structure preserved.Accordingly, the brecciated volcano-lithic fragments are classified into tuff, lapilli tuff, tuff breccias, volcanic bomb and ignimbrite deposits.
Tuff is the most common and widely distributed variant of felsic volcanics in the study area.The variants of tuffs are classified based on the grain size, constituent materials and percentage of clasts with respect to groundmass.Accordingly, in the studied area, these are classified as fine ash (64 mm; McPhie 1993) tuff.The tuffaceous beds are leucocratic, foliated, crenulated and interbedded with coherent lava flow, ignimbrite, epiclasts, phyllite and quartzite.It displays planar bedding at most places.Fine, coarse, lapilli and brecciated varieties of tuff are also observed in the field area.
Ignimbrites are hot pyroclastic surge deposits formed by the consolidation of materials like pumice, volcanic ash, gas, and crystal.They are commonly scattered with lithic fragments and a holohyaline groundmass, albeit the original texture of the groundmass may be obliterated due to high degrees of welding.The ignimbrites of our study area are rich in lithic and pumice clasts and display fiamme and fractured crystals within a glassy groundmass, with variation in clast size from bomb to lapilli (Figure 3(e)).They are a densely welded variety and display a foliated texture defined by aligned, flattened pumice lapilli.The glassy clasts and crystals are poorly sorted, and groundmass is rich in quartzofeldspathic minerals.
Quartzite of the Upper Quartzitic Formation is the most common lithounit of the Shillong Group of rocks.The textural elements like colour, grain size, composition, and degree of recrystallization of the quartzites vary from place to place.It occurs as thickly bedded unit intercalating with thin beds of phyllite, quartz-mica schist and metavolcanics (Figure 3(f)).The quartzites display a range of sizes from extremely fine-grained cryptocrystalline variety to medium to coarse grained and immature variety consisting of angular grains.Colour varies from buff to reddish brown to greyish white and is siliceous to arkosic in nature.Sedimentary structures like planar cross-bedding, ripple mark, herringbone cross stratification along with compositional layering as well as graded bedding are well preserved at places.
The Lower Phyllitic Formation consists of phyllites that occur as thin intercalated bands within quartzite and quartz mica schist as well as the felsic metavolcanics.The phyllite bands are leucocratic but sometimes achieve reddish brown colour due to oxidation.Thickness of bands varies from few centimetres to more than 10 m.The bands are friable, finely laminated and well foliated, and display crenulations at certain places.They are carbonaceous in nature.
Inter as well as intra-formational conglomerate beds are commonly observed within the Shillong Group of rocks.The conglomerate horizons are polymictic, matrix supported, with clast shape and size varying from subrounded elliptical to rounded in nature.

Analytical methods
Mineral chemical analyses of selective mineral phases were performed using the electron probe microanalyser (EPMA) Cameca SX-5 instrument at DST-SERB National Facility, Department of Geology (Center of Advanced Study), Institute of Science, Banaras Hindu University.A Cameca SX-5 software was used for operating the instrument with an accelerating voltage of 15 kV, beam current at 10 nA, and a LaB 6 source was used for the production of electron beam.The positions of crystals (SP5-PET, SP4-LTAP, SP3-LPET, SP2-LiF, and SP1-TAP) were verified with the help of natural silicate mineral andradite in wavelength dispersive (WDS) mode.The mineral analyses were performed with the help of the listed X-ray lines: Fe-Kα, Ca-Kα, Mg-Kα, Si-Kα, Mn-Kα, K-Kα, Na-Kα, Al-Kα, P-Kα, Ti-Kα, Cr-Kα, F-Kα, and Cl-Kα.The listed natural mineral standards were used for analytical quantification, routine calibration, and X-ray elemental mapping: wollastonite, periclase, orthoclase, rhodonite, corundum, chromite, apatite, fluorite, halite, rutile, and haematite.Analytical quantification, calibration, and data acquisition and processing were performed with the help of SX-results and SX-SAB (version 6.1) softwares.The raw data were corrected following the PAP procedure (Pouchou and Pichoir 1987).
The major oxides, trace elements and rare earth element analyses were carried out at the Wadia Institute of Himalayan Geology, Dehradun (India).Major oxides and selected trace elemental analyses were performed using Bruker S8 Tiger Sequential X-ray Spectrometer with Rh excitation source, following the processes explained by Saini et al. (1998Saini et al. ( , 2000)).The operating conditions for major oxides were: No filter, Vacuum path, 20/40 kV and for trace elements were: No filter, Vacuum path, 55/60 kV.For major and minor oxides the overall accuracy in relative standard deviation (RSD) percentage is<5% and for trace elements is<12%.The average precision is better than 2% (Purohit et al. 2006;Saini et al. 2007).The ICPMS analyses of the rare earth elements (REE) were done using the equipment PerkinElmer SCIEX quadrapole type ICPMS, ELAN DRC-e.For rare earth element analysis sample solutions were introduced into the argon plasma by means of a peristaltic pump and a cross flow nebulizer.Sample digestion and preparation of solutions were carried out following the procedures of Balaram et al. (1990).To minimize matrix effect, internal standard (AMH) and USGS (BHVO-1) sample were used as rock standards.The analysed and reported values of the standards are presented in Supp.Table 1.RSD for most of the samples is better than 10%.Minimum detection limit for major oxides is 0.01 wt% and for most of the trace elements is ~ 5 ppm.

Petrography
The Shillong HMAs, found associated with the Shillong Group of rocks, are fine grained in nature, which consist predominantly of biotite, chlorite, muscovite, plagioclase, quartz, and iron oxides as their mineral assemblage.The rock thin-sections as a whole display fine-grained, holocrystalline texture, with exceedingly rare and occasional specks of devitrified glass.The compositions of biotite, plagioclase, chlorite, and muscovite were determined by electron microprobe.Representative mineral chemical analyses of these mineral phases are given in Supp.Tables 2, 3, and 4.
The HMAs are aphyric with biotite, chlorite, and plagioclase as the most common minerals (Figure 4(a,  b)).They are homogeneously distributed throughout the fine-grained groundmass, and are mostly anhedral to subhedral in shape.No large phenocrysts were observed in thin sections.Biotite crystals are brown in colour, and strongly pleochroic under plane polarized light.They display distinctive bird's eye extinction with high-order interference colours.On the biotite classification diagram (Speer 1984), the biotites show a homogeneous composition and are classified as siderophyllite (Figure S2).On the FeO+MnO -MgO-10TiO 2 ternary diagram (Nachit et al. 2005), the analysed biotite crystals distinctly plot within the reequilibrated biotite field (Figure S3).Such signatures suggest that biotite crystals grew at the expense of earlier formed primary Fe-Mg minerals (Ramos et al. 2019).On the other hand, colourless and nonpleochroic muscovite occurs as small grains throughout the fine-grained groundmass.Plagioclase is also abundant, and present as small anhedral to subhedral grains.They display common distinctive optical properties like polysynthetic/albite twinning and their characteristic first-order grey interference colour.The analysed plagioclase crystals plot in the field of albite (Figure S4) and have compositions between An 0 to An 3 .Also, quartz microliths showing undulose extinction are present throughout the rock thin-sections.Chlorite, which occurs as greenish specks, is one of the most abundant minerals present.It occurs as small, anhedral grains and displays low-order interference colours.Iron oxides are also present as dark opaque specks.The mineral phases observed in the HMAs are mostly altered (recrystallized) products of regional metamorphism with their original forms (minerals) absent.A characteristic feature observed in thin sections is that the mineral grains are aligned in a particular direction, which indicates a weak foliation prevalent in the studied HMAs (Figure 4(c,d)).
The SiO₂ (wt %) versus Na₂O+K₂O (wt %) plot categorizes the HMA samples of the Shillong Group of rocks as andesite and basaltic andesite (Figure 5 They also show high-K signature as established by the Th versus Co plot (Figure 6(c); Hastie et al. 2007).Notably, the study samples also exhibit significant concentrations of TiO₂ (0.54-0.76 wt %) higher than boninites (TiO 2 <0.5 wt%, Le Bas and Streckeisen 1991) (Figure 6(d)).These geochemical features adhere to the standards set for high-magnesium andesites (HMAs) as given by Kelemen et al. (2003).They are analogous to the well-documented  HMAs from the western Aleutian arc (Yogodzinski et al. 1995;Kelemen et al. 2003), Mt.Shasta region in N California (Grove et al. 2002) as well as the Setouchi Volcanic Belt in SW Japan (Tatsumi and Hanyu 2003).
In the Chondrite-normalized REE diagram, the Shillong high-magnesium andesites (HMAs) are characterized by moderately enriched light rare earth elements (LREEs) ([La/Yb] N = 7.6-16.9)relative to the flat HREE trends ([Gd/Yb] N = 1.67-3.12)(Figure 7(a)).In the Primitive Mantle-normalized trace elemental plots, the HMA samples are significantly depleted in high field strength elements (HFSEs, such as Nb and Ti) (Figure 7(b)) and exhibit high and variable enrichments in large-ion lithophile elements (LILEs, such as Th, Rb, Pb, and K).Such elemental geochemical signatures may be attributed to the role of slab-derived aqueous fluids in a subduction zone tectonic setting.Additionally, the studied samples also show negative Eu anomalies (Eu/Eu* = 0.58-0.72).The negative Eu, Ba, Sr and P anomalies imply feldspar and apatite separation during the process of fractional crystallization or the presence of these residual phases in the source rocks; whereas positive Rb, Th, Pb and K, and negative Nb and Ti presumably suggest the role of slab-derived aqueous fluids and the presence of minerals such as rutile and garnet in the source rocks, as is evident in most arc magmas.The trace and rare earth element patterns of the Shillong HMAs are similar to that of the Aleutian, Mt.Shasta, and Setouchi HMAs (Figure 7(a,b)).

Post-crystallization alteration
The HMAs of the Shillong Group of rocks have undergone greenschist facies metamorphism characterized by significant mineral alteration.However, they still preserve their original alteration-resistant trace elements, which show homogenous, coherent, and subparallel concentration patterns (Figure 7(a,b)), indicating the preservation of the bulk original alteration-resistant trace element compositions.Such geochemical trend is consistent with the conclusions given forth by earlier studies (Ludden and Thompson 1978;Bienvenu et al. 1990;Xu et al. 2000), which suggest that the concentration of certain trace elements like the high field strength elements and rare earth elements in magmatic rocks are not usually affected by rock alteration.However, it is very likely that some mobile elements (like K, Ba, Rb, Sr, etc.) get mobilized during metamorphism and alteration; nonetheless our samples do not display such mobilization.Thus, we can say that the Shillong HMAs have their original bulk geochemical signatures preserved even though they have been affected by low-to mediumgrade regional metamorphism.

Petrogenesis of the Shillong HMAs
In the Harker variation plots, the Shillong HMAs display abundances of MgO, Fe 2 O 3 , CaO, Al 2 O 3 , K 2 O, and TiO 2 ,  Sun and McDonough (1989).
showing moderately decreasing or negative correlation trends with increasing SiO 2 , whereas Na 2 O, MnO, and P 2 O 5 do not show any such correlation (Figure S5).This moderate degree of geochemical correlation between MgO, Fe 2 O 3 , CaO, Al 2 O 3 , K 2 O, TiO 2 and SiO 2 shown by the Shillong HMAs can be described by the fractional crystallization of pyroxene, biotite, plagioclase, and certain Ti-bearing mineral phases during magma evolution.However, the lack of strong covariation patterns observed in the Shillong HMAs suggest that fractional crystallization did not play a dominant role in their petrogenesis.This is also corroborated by the La versus La/ Yb and La versus La/Sm plots (Figure 8(a,b)), in which the andesitic samples follow the partial melting trend rather than fractional crystallization.However, in accordance with the Harker variation diagrams, some degree of fractional crystallization of clinopyroxene, plagioclase, and biotite (Figure 8(c-f)) occurred during evolution of the basaltic andesites to the andesites of our study area.
From the Yb N versus La N /Yb N plot (Figure 9(a); Drummond and Defant 1990), it can be inferred that the Shillong HMA samples have higher Yb N values (9.57-16.78)and lower La N /Yb N ratios (8-17).Similarly, the samples show higher Y contents (28-36 ppm) and lower Sr/Y ratios (1-5) in the Y versus Sr/Y plot (Figure 9 (b)).In both the plots, the Shillong HMAs fall under the region of 'normal' arc andesite -dacite-rhyolite array, which is indicative of a subduction zone setting.Moreover, in the Nb/Yb versus Th/Yb variation diagram, used for the detection of subduction signatures in magmas, where the subduction components are positioned with higher Th/Yb values (Pearce 2008); the plot displays a diagonal MORB-OIB array with N-MORB, E-MORB and OIB as its end-members, wherein, our samples fall under the volcanic arc field (Figure 9(c)).From the above discussion, it is quite evident that the Shillong HMA were generated in a subduction zone setting.The Shillong Basin represents a subduction zone setting (Ray et al. 2013), more specifically a back-arc region (Basumatary et al. 2023).Thus, processes associated with subduction tectonics were responsible for the generation of high-Mg andesitic magmas in our study area.
Trace elemental signatures, like the enrichment of LILEs such as U (0.8-1.1 ppm) and Th (14.8-18.9ppm) indicate significant input from sediments in the Shillong HMAs.This is comprehensible from the Th versus U/Th (Zeng et al. 2016) and the Th versus Th/La (Plank and Langmuir 1998) plots, which delineates the Shillong HMA samples in the field of arc volcanic rocks with a sediment component (Figure 9(d)), as well as in the region of marine sediments and upper continental crust signatures (Figure 9(e)).We also conclude that the Shillong HMAs may have formed as a result of the interaction between subducted sediment-derived melts and mantle peridotite based on the low U/Th and high Th/La values (Shimoda et al. 1998;Tatsumi 2001;Wang et al. 2011;Zeng et al. 2016;Liu et al. 2018).Low U/Th (Figure 9(d)), low Ba/Th ratios (Figure 9(f)), and high Th concentration also indicate that subducting slab-derived fluids did not play any significant role in the formation of the Shillong HMAs (Turner et al. 1996).
The higher Th/La ratios of the Shillong HMAs than that of GLOSS (global subducting sediments; Figure 9(e)) suggest fractionation of Th/La, which is further indicative of a monazite-poor source (Plank 2005), as the partition coefficient of La is very high compared to that of Th in monazite.On the other hand, allanite is a common phase in metasediments that preferentially assimilates La over Th (Hermann and Rubatto 2009).Thus, the Shillong HMAs were possibly generated from an allanitebearing source.Additionally, the higher concentrations of Th suggest a high-temperature source region for the Shillong HMAs, as the sediment melting experiments (Johnson and Plank 1999) proved that sediment-hosted Th is partitioned into melts at temperatures exceeding 900°C.
We can comprehend the degree of partial melting as well as the nature of a mantle source from the concentrations and ratios of rare-earth elements (D'Orazio et al. 2001).The Sm/Yb ratio, for instance, is almost entirely unaffected by partial melting of spinel lherzolite.However, it gets strongly fractionated in the garnet stability field due to melting (Aldanmaz et al. 2000).The mineral phases spinel and garnet are stable at around 50-80 kilometres and>80 km depth, respectively (Wyllie 1981).The Shillong HMAs plot in the field of the spinel to garnet transition (50:50), indicating a lherzolitic mantle source with approximately 5% − 10% partial melting (Figure 10).Hence, we can concur that the studied HMAs melted at a depth of around 80 km.Therefore, the genesis of the Shillong HMAs can be explained by partial melting of a depleted lherzolite mantle source within the spinel -garnet transition zone, where metasomatism by fluids and subsequent fractional crystallization of clinopyroxene, plagioclase, and biotite occurred at relatively shallow depths.These results coincide with the nature of the mantle source inferred for the metadolerites of the Shillong Basin (Basumatary et al. 2023).

Geotectonic framework and geodynamic implications
To discuss the geotectonic framework of the Shillong HMAs, we must first evaluate the geotectonic setting for the mafic rocks of the Shillong Group.The published data of the Shillong Basin mafic rocks (Ray et al. 2013) were    Sun and McDonough (1989).
plotted in tectonic discrimination diagrams (Figure 11), alongside which, we compared the data representing the Shillong HMAs.The Shillong HMAs display high LILE/HFSE and LREE/HREE ratios, which are distinctive features of magmas generated in subduction zone setting (Hawkesworth et al. 1993;Gorton and Schandl 2000;Gogoi 2022).On the tectonic discrimination diagrams (Figure 11(a-d)), the Shillong mafic rocks as well as the andesites distinctly plot in the region of back-arc basin basalts (BABB).However, in the Zr versus Ti/Zr plot, the Shillong HMAs fall proximately below the field of BABB.From the geotectonic discrimination diagrams, we can infer that back arc-related magmatism occurred in the Shillong Plateau during Mesoproterozoic times.Thus, we can conclude that the magmas that fed the andesitic volcanism in the Shillong Basin were produced in a backarc extensional setting associated with subduction dynamics, i.e, the Shillong Basin represents a back-arc rift basin caused due to a tensional stress regime.
The Central Indian Tectonic Zone (CITZ) constitutes an ENE-WSW trending Proterozoic mobile belt along which the northern and southern Indian crustal blocks sutured to form the Greater Indian Landmass (Acharyya 2003).The Mahakoshal Mobile Belt (MMB) constitutes the northern margin of the CITZ, while the broader southern domain of the CITZ, from west to east, is occupied by the Sausar Mobile Belt, the Chotanagpur Granite Gneiss Complex (CGGC), and the Assam-Meghalaya Gneissic Complex (AMGC).The Bathani volcano-sedimentary sequence (BVSs) is a volcano-sedimentary suite of rocks that constitutes the northern margin of the CGGC and is now considered as the eastern continuation of the MMB (Saikia et al. 2017).Arc-related magmatism associated with subduction setting has been suggested for the granites of the Mahakoshal Mobile Belt (MMB) in the northern margin of the Central Indian Tectonic Zone (CITZ), which were emplaced at approximately 1.8-1.5 Ga (Roy and Prasad 2003).Bora et al. (2013) also reported a U-Pb SHRIMP zircon 206 Pb/ 238 U age of 1753 ± 9 Ma, whereas Yadav et al. (2015) attained a U -Pb ID-TIMS zircon age of 1873 ± 13 Ma from the calc-alkaline granitoids of the MMB.Similarly, subduction of the South Indian Block (Singhbhum Craton) beneath the North Indian Block (Bundelkhand Craton) would have transformed the southern region of the CGGC into an island arc, now represented by the Dalma Volcanic Belt (Gogoi 2022).This would have given rise to a back-arc extensional setting as a consequence at around 1700 -1600 Ma, now represented by the Bathani volcano-sedimentary sequence (Saikia et al. 2017;Gogoi 2022).Chatterjee and Ghose (2011) gave a monazite crystallization age of 1697 ± 17 Ma from the porphyritic granites of northern CGGC close to BVSs, while Saikia et al. (2017Saikia et al. ( , 2019) ) reported U -Pb ID-TIMS age and Rb-Sr whole rock isochron age of 1700-1600 Ma and 1664 ± 130 Ma for the granites of BVSs respectively.The coeval association between these suites of rocks is extremely interesting because contemporary ages of 1650 -1500 Ma have also been recorded from the granites of the Shillong Basin (Gogoi et al. 2019), which corroborates the same back-arc extensional tectonic setting associated with subduction dynamics.Another striking factor is the NE-SW lateral trend with a regional SE dip that has been observed between the MMB, BVSs, and Shillong Basin, which is directly tantamount to the regional trend of the CITZ (ENE-WSW).Also, Paleo-Mesoproterozoic felsic magmatism recorded in AMGC (Bidyananda and Deomurari 2007;Yin et al. 2010;Kumar et al. 2017;Gogoi et al. 2019;Doley et al. 2022) is comparable to that of the MMB of central India (Bora et al. 2013;Yadav et al. 2015) and CGGC of east India (Ray Barman et al. 1994;Acharyya 2003;Hossain et al. 2007;Saikia et al. 2017;Gogoi 2022), Eastern Ghat Mobile Belt of southeast India (Dobmeier et al. 2006), and Aravalli Belt of northwest India (Kaur et al. 2013).Thus, from the synchronous nature of origin, geotectonic setting, regional structural trend, and associated geochronological ages given by previous authors, we can conclude that the Shillong Basin is the easternmost continuation of the MMB and BVSs of the CITZ; and together they form the northern fragments of the same suture zone consisting of the CITZ, the CGGC, and the AMGC, which demarcates the Northern Indian Block and Southern Indian Block of the Indian Subcontinent (Figure 12).
On a worldwide scale, the Greater Indian Landmass holds an indispensable role in the various hypothesized supercontinent reconstruction models, especially in Columbia reconstruction (Meert 2002;Rogers and Santosh 2002;Zhao et al. 2002;Hou et al. 2008;Zhang et al. 2012;Pesonen et al. 2012;Pisarevsky et al. 2013;Bhowmik 2019;Chattopadhyay et al. 2020;Sequeira et al. 2022).These hypothesized models have referred that Columbia assembled at ca. 2.1-1.8Ga, and disintegrated at ca. 1.3-1.2Ga, with the intracontinental rifting existing at ca. 1.8-1.6Ga (Evans 2013;Wang et al. 2019;Meert and Santosh 2022).However, due to insufficient geochronological and palaeomagnetic inputs from the Indian Shield, these models have put a constraint on the validity of India's position in Columbia reconstruction.Hou et al. (2008) placed India near Laurentia and the North China Craton, whereas Zhang et al. (2012) placed it vicinal to Australia.Pisarevsky et al. (2013), on the other hand, positioned India next to the Sarmatia margin of Baltica.Thus, India's position within Columbia varies considerably in the recently proposed models.Furthermore, it has been debated whether the northern and southern Indian blocks were contiguous or not throughout Proterozoic times (Evans 2013), i.e. until the amalgamation of Columbia, India was not a single continental block as portrayed in most of the models of Columbia formation (Saikia et al. 2017;Gogoi 2022).However, the Indian proto-continent was a single entity during the Grenvillian Orogeny (1.2-1.1 Ga), at the time of its incorporation into Rodinia.Therefore, in order to comprehend the assemblage, configuration and disintegration of the greater Indian landmass, in relation with the supercontinents Columbia and Rodinia, we need to scrutinize a continuation of the mobile belts similar to CITZ in the adjoining proto-continental blocks.Identification of such analogous Paleo-Mesoproterozoic mobile belts with subduction dynamics and associated back-arc rift setting would allow us to accurately position India alongside hypothesized proto-continental blocks in supercontinent reconstruction models.

Conclusions
(1) The present work happens to be the first to report the occurrence of high-Mg andesites from the Assam-Meghalaya Gneissic Complex of Northeast India.
(2) Geochemical characteristics suggest that the Shillong HMAs were generated in a subduction zone setting, more specifically a back-arc extensional regime, by partial melting of a depleted lherzolite mantle source within the spinel -garnet transition zone.
(3) Based on the synchronous nature of origin, geotectonic setting, regional structural trend, and associated geochronological ages, we conclude that the Shillong Basin represents the easternmost continuation of the Bathani volcano-sedimentary sequence and Mahakoshal Mobile Belt, thereby marking the continuation of the Central Indian Tectonic Zone to the Assam-Meghalaya Gneissic Complex.(4) India was not a single continental block before Columbia supercontinent formation as hypothesized in several Columbia reconstruction models.Therefore, in order to comprehend the assemblage, configuration, and disintegration of the greater Indian landmass, in relation to Columbia, we need to scrutinize the continuation of mobile belts similar to the Central Indian Tectonic Zone in the adjoining proto-continental blocks.

Figure 2 .
Figure 2. (a) Regional geological map of the Assam-Meghalaya Gneissic Complex displaying the location of the study area (modified after Yin et al. 2010; Sadiq et al. 2017).The study area is featured by the black box.Abbreviations: BF = Brahmaputra Fault, DF = Dauki Fault, JFS = Jamuna Fault System, and KFZ = Kopili Fault Zone.(b) a detailed geological map of the study area showing the disposition of the high-Mg andesites in relation to other rock types and sample locations (modified after Naik et al. 2020; Bhukosh portal, Geological Survey of India).

Figure 3 .
Figure 3. Field photographs displaying (a-c) Massive outcrops of high-Mg andesitic lava flows along with associated quartzites and phyllites.(d) Scoria with vesicular structure.(e) Ignimbrites displaying lithic and pumice clasts, fiamme, and fractured crystals within a glassy groundmass.Clast size vary from bomb to lapilli.(f) Sharp contact between andesitic lava flow and quartzite shown by the yellow demarcation.

Figure 4 .
Figure 4. Photomicrographs displaying (a,b) PPL and CPL images of HMAs with anhedral to subhedral biotite, chlorite, and plagioclase grains.They are homogeneously distributed throughout the fine-grained groundmass.(c,d) a prominent alignment of the mineral grains was observed indicating a weak foliation.Mineral abbreviations: Bt = Biotite, Chl = Chlorite.

Figure 7 .
Figure 7. (a) Chondrite-normalized REE plots and (b) Primitive mantle-normalized multi-element patterns for the representative andesitic samples of the Shillong HMAs.Normalizing values are after Sun and McDonough (1989).

Figure 10 .
Figure 10.Plot of Sm/Yb versus Sm (after Zhao and Zhou 2007) for the Shillong HMAs.Numbers along lines signify the geochemical modelling degree of partial melting in different mantle sources.And the compositions of different mantle sources are from Sun and McDonough (1989).

Figure 12 .
Figure 12.A schematic map of Indian Subcontinent showing the approximately 1500 km-long ENE -WSW-trending Central Indian Tectonic Zone (CITZ), Chotanagpur Granite Gneiss Complex (CGGC) and Assam-Meghalaya Gneissic Complex (AMGC).This continuous orogenic belt sutured the North Indian Block (NIB) and South Indian Block (SIB) during the Proterozoic (Acharyya 2003).The NIB consists primarily of the Bundelkhand Craton (BuC), whereas the SIB consists of the Dharwar Craton (DC), the Bastar Craton (BC) and the Singhbhum Craton (SC).Two other mobile belts, namely the Aravalli Delhi Mobile Belt (ADMB) and the Eastern Ghats Belt (EGB), are also depicted in the diagram.The northern periphery of the CITZ is occupied by a volcano-sedimentary sequence that was restricted to the Mahakoshal Mobile Belt (MMB) and the Bathani Volcano Sedimentary sequence (BVSs) earlier.However, our present study projects the Shillong Basin (SB) as the eastern continuation of the MMB and the BVSs, thereby extending the CITZ to the AMGC.The significance of our study is that while positioning India in the Columbia supercontinent formation, traces of a mobile belt, synchronous to the CITZ, should be looked for in the adjacent continental blocks lying east and west of India.The ages reported from the SB are from: 3. Gogoi et al. (2019).The ages reported from the MMB are from: 1. Roy and Prasad (2003); 2. Bora et al. (2013); 3. Yadav et al. (2015).The ages reported from the BVSs are from: 1. Chatterjee and Ghose (2011); 2. Saikia et al. (2017, 2019); 3. Gogoi (2022).The ages reported from the CITZ are from: 1. Bhandari et al. (2011); 2. Bhowmik et al. (2011).The ages reported from the CGGC are from: 1. Pandey et al. (1986); 2. Chatterjee et al. (2008).The ages reported from the AMGC are from: 1. Chatterjee et al. (2007); 2. Yin et al. (2010).