Zircon U – Pb geochronology and geochemistry of granitic rocks in central Mongolia

The Central Asian Orogenic Belt had been formed by amalgamation of voluminous subduction–accretionary complexes during the Late Neoproterozoic to the Mesozoic period. Mongolia is situated in the center of this belt. This study presents new zircon U–Pb geochronological, whole-rock major and trace element data for granitoids within central Mongolia and discusses the tectonic setting and evolution of these granitic magmas during their formation and emplacement. The zircon U–Pb ages indicate that the magmatism can be divided into three stages: the 564–532 Ma Baidrag granitoids, the 269–248 and 238–237 Ma Khangai granitoids. The 564–532 Ma Baidrag granitoids are adakitic, have an I-type affinity, and were emplaced into metamorphic rocks. In comparison, the 269–248 Ma granitoids have high-K, calcalkaline, granodioritic compositions and are I-type granites, whereas the associated the 238–237 Ma granites have an A-type affinity. The 564–532 Ma Baidrag and 269 –248 Ma Khangai granitoids also both have volcanic arc-type affinities, whereas the 238–237 Ma granites formed in a post-collisional tectonic setting. These geochronological and geochemical results suggest that arc magmatism occurred at the 564–532 Ma which might be the oldest magmatic activity in central Mongolia. Between the Baidrag and the Khangai, there might be paleo-ocean and the oceanic plate subducted beneath the Khangai and produced voluminous granite bodies during the 269–248 Ma. After the closure of the paleo-ocean, the post collisional granitoids were formed at the 238–237 Ma based on the result of later granitoids in the Khangai area.


INTRODUCTION
Mongolia is located within the central part of the Central Asian Orogenic Belt (CAOB), which is bounded to the north by the Precambrian Siberian craton and to the south by the Tarim and North China cratons (Badarch et al., 2002;Windley et al., 2007). The central part of Mongolia is also known as the Khangai region (including the Bayankhongor area) and records multiple orogenic and magmatic events between the early Proterozoic and Mesozoic (Takashi et al., 2000;Jahn et al., 2004). However, the geology of this region remains relatively poorly characterised, with previous studies focusing on only a small number of granitic rocks. This means that systematic and

Mongolian Geoscientist
representative sampling and the associated analysis of these samples are needed to understand the tectonic processes that generated these voluminous granitoids, in addition to furthering our understanding of the tectonic evolution of the CAOB. grained, and interbedded siltstones, sandstones, and shales that contain rounded quartz grains, preserve sedimentary structures, and most likely represent a turbidite sequence (Fig. 1). The protoliths for the Zag schists are thought to be deep-water passive margin sedimentary units that overlie the Bayankhongor ophiolite, but are located beneath the sedimentary cover of the Khangai Basin (Badarch et al., 2002). Detrital zircon U-Pb ages for the Zag schists yield a depositional age of <445 ± 6 Ma (Kröner et al., 2011), whereas metamorphic white micas from the Zag metasediments yield K-Ar ages of 454 ± 9, 445 ± 9, and 395 ± 20 Ma (Jian et al., 2010). Devonian and Carboniferous sedimentary rocks are widely distributed in the Khangai area ( Fig.  1) and together constitute the Khangai-Khentii Basin (Tomurtogoo et al., 1998;Kelty et al., 2008;Purevjav and Roser, 2012). This basin is located within the central and northeastern parts of Mongolia and extends into Russia, where it forms the western end of the Khangai-Khentii-Daurian zone of the Mongol-Okhotsk Fold Belt (MOFB) (Zorin, 1999). These sedimentary rocks host voluminous early Permian and Triassic granitoids that are associated with large contact metamorphic aureoles (Orolmaa et al., 2008). Together these granitoids and sedimentary rocks form the NW-SE trending Khangai Mountains (Fig. 1). The sampling undertaken during this study was undertaken along a longitudinal traverse, yielding a total of 105 granitic samples (Fig. 1).

Field occurrences and petrology 564-532 Ma Baidrag granitoids
Fieldwork during this study identified Kfeldspar porphyritic biotite granite and redcolored K-feldspar porphyritic biotitehornblende granite units, with the latter either containing or free of mafic microgranular enclaves (MME; Fig. 2). These granitic rocks were generally emplaced into and are in direct contact with metamorphic rocks that include garnet-biotite gneiss, amphibolite, and garnet-sillimanite-biotite gneiss units (Fig. 2a, d). However, the sharp boundaries between these units are not associated with significant contact metamorphism, apart from skarn-type calcsilicates that are developed along the contact between the K-feldspar porphyritic biotite granite and limestone or weakly recrystallized marble units (Fig. 2c). The biotite granite is leucocratic and contains very large but variably sized K-feldspar (Fig. 2b). The K-feldspar porphyritic biotite-hornblende granite also contains very large K-feldspar, but this granite is more melanocratic than the biotite granite and the K-feldspar present in this unit is reddish to pinkish in color (Fig. 2e). The boundary between the Baidrag granitoids and Bayankhongor ultramafic belt is characterized by a red-colored K-feldspar porphyritic biotitehornblende granite that contains MME (Fig. 2f). Representative plutonic rocks from the Baidrag rock emplaced into metamorphic country rock, (e) Red-colored K-feldspar porphyritic biotite-hornblende granite (sample 15082102), (f) Red-colored K-feldspar porphyritic biotite-hornblende granite containing a MME (sample 15082106), (g) K-feldspar porphyritic biotite-hornblende granite (sample 15082102), (h) MME (sample 15082106C) within the K-feldspar porphyritic biotite-hornblende granite, (i) K-feldspar porphyritic biotite granite (15082105A), (j) Biotite-muscovite granite (15082107B). area ( Fig. 2g-j) include K-feldspar porphyritic biotite-hornblende granite, K-feldspar porphyritic biotite granite, and biotite-muscovite granite units, with the K-feldspar porphyritic biotite-hornblende granites being divided into two types depending on the presence or absence of MME.
K-feldspar porphyritic biotite granite (15082105A, Fig. 2i) contains biotite, Kfeldspar, plagioclase, quartz, and muscovite along with accessory apatite, zircon and opaque minerals. This granite is free of hornblende and accessory phases are both matrix-hosted and present as inclusions within biotite.

269-248 Ma Khangai granitoids
The 269-248 Ma Khangai granitoids are dominated by fresh and undeformed phaneritic biotite-hornblende granodiorite and medium-to fine-grained biotite granite units. The southwestern part of the Khangai area is dominated by phaneritic biotite-hornblende granodiorite units that commonly contain MME and are crosscut by a granitic dike (Fig. 3a, b). The west-central part of the Khangai area contains phaneritic biotite-hornblende granodiorite units with MME and K-feldspar porphyritic biotite granite units (Fig. 3c, d). The central to northern parts of this area contain phaneritic biotite-hornblende granodiorite and medium-to fine-grained biotite granite units, whereas the majority of the northern part of this area contains phaneritic biotite-hornblende granodiorite units that have been emplaced into metamorphic country rocks, including granitic, garnet, and biotite-garnet gneisses (Fig. 3e, f).

238-237 Ma Khangai granitoids
The central to northern parts of the Khangai area contain coarse-grained pegmatitic granites with crystals that are up to 2 cm in size ( Fig. 3g-l).

METHODS
All analyses were performed using instruments housed at Kyushu University, Fukuoka, Japan.

LA-ICPMS U-Pb Zircon dating
The analytical procedure used for zircon U-Pb dating has been described in detail by Adachi et al. (2012b). Zircon crystals from the samples were separated and mounted in epoxy disc with 25 mm diameter and 4 mm thickness, following separation from powdered samples by panning with a beaker and watch glass then hand picking (Kitano et al., 2014). Internal textures and mineral inclusions of individual zircon crystals were observed using a scanning electron microscope (JEOL JED2140JSM-5301S) with a cathodoluminescence (CL) detector (GatanMiniCL). Zircon U-Pb dating was undertaken by LA-ICPMS using an Agilent 7500cx quadrupole ICPMS with a New Wave Research UP-213 laser. For U-Pb geochronology, abundance of the isotopes 202 Hg, 204 Pb, 206 Pb, 207 Pb, 208 Pb, 232 Th, and 238 U were monitored during all analyses, with ablations performed using laser diameters of 25 to 40 m. The integrated isotopic ratios were corrected against standard zircon Temora (417 Ma; Black et al., 2003), and the NIST SRM-611 glass standard was used to calculate Th/U ratios. The standard zircon FC-1 (1099 Ma; Paces and Miller, 1993) was used to ensure consistency. All zircon data reductions and calculations were processed with the GLITTER software package (Griffin et al., 2008), using the analytical and calculation protocols for time-resolved analysis described by Jackson et al. (2004). Concordia diagrams were calculated using the Isoplot/Ex 4.1 software package (Ludwig, 2008).

Geochemistry
Whole-rock major (SiO 2 , TiO 2 , Al 2 O 3 , TFe 2 O 3 , MnO, MgO, CaO, Na 2 O, K 2 O, and P 2 O 5 ) and trace elements (V, Cr, Co, Ni, Cu, Zn, Ga, Rb, Sr, Y, Zr, Nb, Ba) were analyzed with a Phillips PW2400 wavelength dispersive X-ray fluorescence (WDXRF) spectrometer using fused glass beads (2:1 dilution of sample). Prior to fusion, all samples were heated at 900 0 C for 2.5 hours to remove volatiles, and to oxidize ferrous iron. The concentrations of rare earth elements (REEs) and five trace elements (Hf, Ta, Pb, Th, and U) were determined by LA-ICPMS using an Agilent 7500cx quadrupole ICPMS with a New Wave Research UP-213 laser, and these analyses were made directly on the glass bead that had been used for XRF analysis. The analytical procedures and accuracy of the LA-ICPMS approach used during this study are given in Nakano et al. (2012). The analytical uncertainties are less than 5% of the concentration of each element.

Zircon U-Pb dating
A total of 21 samples were dated, with sample locations shown in Fig. 1 and the results given in Table 1.
564-532 Ma Baidrag granitoids. Biotite granite sample 15082205A yielded 18 concordant analyses from zircon rims and a 206 Pb/ 238 U age of 532.5 ± 3.7 Ma (n = 18) with a Th/U value of 0.67 (Fig. 4a). A further three concordant analyses yielded ages of 943 ± 28, 1668 ± 53, and 1900 ± 51 Ma that reflect the ages of inherited zircon cores. Biotite-hornblende granite sample 15082102 yielded three discordant and subsequently discarded analyses and a further 11 concordant analyses that yielded a 206 Pb/ 238 U age of 532.0 ± 5.2 Ma (n = 11) with a Th/U value of 0.59 (Fig. 4b). A single further concordant analysis yielded an age of 769 ± 22 Ma, reflecting the presence of an inherited zircon core.

DISCUSSION
The new zircon U-Pb dates obtained during this study indicate that the Khangai region records three stages of magmatism at 564-532, 269-248, and 238-237 Ma. The 564-532 Ma magmatic event within the Baidrag area is thought to be related to subduction and accretion processes (Demoux et al., 2009), whereas the Permian-Triassic (269-248 and 238 -237 Ma) magmatic events are thought to be related to the formation of the Khangai-Khentii -Daurian zone within the MOFB (Zorin, 1999).

564-532 Ma magmatic event
Zircons from the Baidrag granitic rocks are euhedral, elongate, and have regular oscillatory zoning visible during CL imaging, suggesting a magmatic origin (Corfu et al., 2003), which is consistent with their high Th/U ratios (0.59-0.94). The zircon U-Pb dating of seven granitic samples from different plutons in the Baidrag area indicates that these intrusions were emplaced at 564-532 Ma, consistent with recently reported ages for 547-546 Ma granites and mafic enclaves within the Ulaan-Uul batholith in this region (Zhang et al., 2015). These samples also contain inherited entire zircons and zircon cores that yield ca. 700, ca. 800, ca. 900, ca. 1100, ca. 1200, ca. 1700, and ca. 1900 Ma ages.

Petrogenesis of granitic rocks in the Baidrag area
The biotite-muscovite granites from the Baidrag area have adakitic affinities as evidenced by their plotting within the adakite field in a Sr/Y vs. Y diagram (Fig. 6e-f), whereas the biotitehornblende granites from this area plot in the typical arc magma field of this diagram. As such, the rest of the discussion of these samples focuses on the characteristics of the adakitic and non-adakitic suites. The adakitic samples are high-K and calc-alkaline, are very depleted in the HREE (Yb = 0.53-0.94 ppm; Y = 6.6-16.4 ppm), have small to no Eu anomalies (Fig. 5), and have high Sr/Y ratios (Fig. 6e-f), all of which are similar to the characteristics of typical adakites. The HREE-and Y-depleted nature of adakites suggests they are derived from a magma source containing garnet and/or hornblende are as a residual phases (Defant and Drummond, 1990;Drummond et al., 1996;Martin, 1999). The absence of significant Eu anomalies and the high Sr concentrations of the Baidrag granites also suggest that they were derived from plagioclase free mafic source rocks in very thick crustal section (Kay and Mahlburg-Kay, 1991). Adakitic magmas can be  Whalen et al. (1987). (b) Rb vs. (Y + Nb) after Pearce et al. (1984). (c) Ternary Hf vs. Rb/30 vs. Ta*3 diagram after Harris et al. (1986). Abbreviations are as in Fig. 6 with OGT = unfractionated granite and FG = fractionated granite. generated by the partial melting of subducted oceanic slab material or the fractional crystallization of basaltic magmas (Defant and Drummond, 1990;Castillo, 2012). It is likely that the melting of oceanic slab material, including basalts and overlying metasediments, beneath the Baidrag Craton (Adachi et al., 2012a) generated the adakitic magmas that were emplaced in this area. Previous research (Jahn et al., 2004) determined that the biotite-muscovite granites have 87 Sr/ 86 Sr ratios (I Sr ) of ~0.7078 and an  Nd (T) value of -7.0, suggesting that metasediments were involved in the oceanic basalt melting-dominated petrogenesis of these granites. This is consistent with the presence of detrital zircons within these samples. The nonadakitic rocks within the Baidrag area are classified as normal volcanic arc granites ( Fig.  6e-f), suggesting that the two sets of granites in this area were derived from separate sources, with adakites derived from the melting of oceanic slab material (including basalts and metasediments) and non-adakites derived from the contemporaneous melting of juvenile crustal material.

Tectonic setting of the 564-532 Ma Baidrag granitoid magmatism
The Lake zone within the eastern margin of the Baidrag area contains eclogites that yield an 40 Ar/ 39 Ar cooling age of ca. 540 Ma (Štípská et al., 2010), indicating that these units formed same period with the granitic magmatism in this area. The geochemistry of the adakitic rocks is indicative of formation within a volcanic arc  Whalen et al. (1987). (b) Y vs. 10000*Ga/Al diagram discriminating between I-, S-, and A-type granites after Whalen et al. (1987). (c) Ternary Y vs. Nb vs. Zr/4 diagram after Eby (1992). (d) Rb vs. (Y + Nb) diagram after Pearce et al. (1984). (e) Ternary Hf vs. Rb/30 vs. Ta*3 diagram after Harris et al. (1986). Abbreviations are as in Figs 6 and 9 with A1 = intraplate rifting and A2 = post-collisional, I-igneous, S-sedimentary. environment ( Fig. 6a-d), consistent with the presence of negative Nb, Ta, and Ti anomalies in the primitive-mantle-normalized multielement patterns for these samples (Fig. 5). This indicates that the slab-derived adakitic magmas that were emplaced in this area did not provide sufficient heat to drive the subduction-related partial melting of the lower crust in this region. The fact that the adakitic and non-adakitic Itype magmatism in the Baidrag area was contemporaneous suggests that these granitoids formed in a continental arc-type environment.

269-248 and 238-237 Ma magmatic events
The zircons from the biotite-hornblende granodiorites, MME, and biotite granites in this region are euhedral, elongate, have regular oscillatory zoning visible in CL imaging, and are free of inherited cores, suggesting they have a magmatic origin (Corfu et al., 2003). The zircons from the biotite-hornblende granodiorites and associated MME yield U-Pb ages from 263 to 249 Ma, whereas the zircons from the biotite granites yield U-Pb ages from 269 to 248 Ma, suggesting that all of these units formed contemporaneously. Most of the zircons from the pegmatitic granites have oscillatory zoning visible in CL imaging and also lack inherited cores, but some are unzoned and appear dark during CL imaging, suggesting they contain high concentrations of Uranium. However, all of these characteristics are also indicative of a magmatic origin. In addition, zircons from the granitic dike contain inherited cores surrounded by oscillatory zoned overgrowths that suggest the former were generated during a thermal event that occurred before the crystallization of the latter (Corfu et al., 2003). The zircons from the pegmatitic granites yield a U-Pb age of 237 Ma that is similar to the 238 Ma age obtained for zircons from the granitic dike, suggesting that both  Whalen et al. (1987) with abbreviations as in Fig. 9. (b) Rb vs. Sr diagram showing compositional vectors reflecting the fractionation of biotite (Bt), hornblende (Hbl), and plagioclase (Pl) within granitic systems. Each vector reflects 30% crystallization with pink triangles indicating 10% fractionation intervals. Partition coefficients used are from Arth (1976) and the initial composition is the same as sample 15082801A. magmatic units formed contemporaneously. In addition, the zircons within the granitic dike contain inherited cores that yield a U-Pb age of 267 Ma, consistent with the timing of formation of the hosting granitoids. Here, we provide regional context for the 269-248 and 238-237 Ma granitoids within the Khangai area by comparing the timing of this magmatism to the timing of other magmatic events within the MOFB. A 269-248 Ma magmatic event has also been identified within the Erguna massif of northeastern China (Tang et al., 2015), whereas post-orogenic granitoids from the Khentii area yielded younger ages of 208-205 Ma (Khishigsuren et al., 2012). This suggests that the Khangai granitoids formed within the western margin of the Khangai-Khentii-Daurian zone, which was a constituent part of the MOFB (Zorin, 1999;Donskaya et al., 2012;Tang et al., 2015).

Petrogenesis of granitoids in the Khangai area 269-248 Ma granitoids.
Both the 269-248 Ma biotite-hornblende granodiorites that contain MME and the medium-to fine-grained biotite granites in this area are free of muscovite and Al -rich minerals such as garnet, aluminosilicates, and cordierite. The presence of hornblende and the absence of Al-rich minerals suggest that these samples probably have I-type affinity, consistent with their aluminum saturation index (ASI) values, all of which are >1.1 (Chappell and White 2001). Both the biotite-hornblende granodiorites and biotite granites have SiO 2 contents that range from 54.44 to 75.93 wt.% and both define a single trend on Harker diagrams that likely reflects fractional crystallization processes (Fig. 7). The concentrations of some of the LILE within these samples (e.g., Sr, Eu, and Rb) also define linear trends on Harker diagrams (Fig. 7), suggesting that the evolved biotite granites were derived from the more primitive biotite-hornblende granodiorites. The biotite granites also have high TFe 2 O 3 /MgO values that are indicative of fractionated granites, whereas the lower values of the biotite-hornblende granodiorites are indicative of unfractionated granites (Fig. 11a).
Plotting these samples on a hornblende, biotite, and plagioclase vector diagram indicates that these two suites of granitoids can be linked by the fractionation of plagioclase and hornblende (Fig. 11b). In addition, the primitive-mantlenormalized multi-element patterns of the biotite -hornblende granodiorites are characterized by positive Ba and Sr and negative Nb, Ta, and Ti anomalies. These anomalies suggest that these granodiorites formed from magmas derived from a source containing residual amphibole and/or rutile (Martin et al. 2005). The relatively flat HREE patterns for these samples also suggest derivation from a source containing more residual amphibole than garnet (Moyen, 2009). The primitive-mantle-normalized multielement patterns for the biotite granites have positive Rb, Th, and U, and negative Ba, Sr, Nb, Ta, P, and Ti anomalies. Depletions in Nb, Ta, and Ti are usually indicative of the separation of Ti-bearing phases such as ilmenite and/or rutile (Martin et al. 2005). In addition, the P 2 O 5 depletions are most likely indicative of the fractionation of apatite together with plagioclase and hornblende as discussed above, and the negative Eu anomalies within these samples reflect plagioclase fractionation. This suggests that the biotite-hornblende granodiorites and biotite granites were derived from the same magma source, and these granitoids are geochemically similar to coeval granitoids in the Erguna massif, suggesting these magmatic events might be linked (Tang et al., 2015).  (Fig. 7) and they have high TFe 2 O 3 /MgO and Ga/Al values (Fig. 10a, b), all of which suggest that these rocks have undergone significant magmatic differentiation and have A-type granite affinities, as evidenced by their classification as A-type in a Y-Nb-Zr/4 ternary diagram (Fig. 10c). These units also contain elevated concentrations of the HFSE (e.g., Nb and Y), contrasting with the 269-248 Ma biotite -hornblende granodiorites and biotite granites. They have primitive-mantle-normalized multielement patterns that are characterized by positive Rb, Th, U, and Pb, and negative Ba, Sr, P, Eu, and Ti anomalies. These depletions in Ba, Sr, and Eu are indicative of the fractionation of feldspar, whereas the P 2 O 5 depletions in these samples are indicative of the fractionation of apatite. However, A-type granites can form in both post-orogenic and anorogenic settings (Bonin, 2007), and it is often difficult to discriminate between A-type and post-orogenic granites, although there are some geological, mineralogical, and chemical differences between these two types of granite (Bonin, 1990). Anorogenic granites are commonly associated with syenite-alkali feldspar syenitenepheline syenite suites and are depleted in Ba and Sr, but are enriched in Fe. In contrast, postorogenic granites are generally enriched in Ba and Sr and commonly have high-K calk-alkaline compositions. The pegmatitic granites in the study area have relatively flat REE patterns, significant negative Eu anomalies, and high-K calk-alkaline compositions. Combining these observations with the absence of associated mafic minerals such as arfvedsonite and aegirine suggests these units represent a post-orogenic suite of magmas that formed by the partial melting of highly fractionated I-type granitic rocks, similar to the post-orogenic granitoids of the Bogd Uul pluton in the Khentii area (Khishigsuren et al., 2012).

Tectonic setting of the granitoids in the Khangai area.
The 269-248 Ma biotite-hornblende granodiorites are volcanic arc granites (Fig. 9b,  c), whereas some of the biotite granites in this area are classified as collision-related granites, most likely reflecting the effects of fractionation and crustal contamination (Pearce et al., 1984). The 269-248 Ma granitoids in the Khangai area are similar to coeval granitic rocks of the Erguna massif (Tang et al., 2015), which consist of a suite of quartz diorites, granodiorites, and monzogranites, all of which have high-K and calc-alkaline affinities. These rocks are also similar to the granitoids in the Khangai area in that they are LILE and LREE enriched, but are depleted in the HFSE and HREE. The 269-248 Ma biotite-hornblende granodiorites in the Khangai area are also similar in composition to representative subduction-related continental arc granitoids of the Sierra Nevada in California, USA (Cecil et al., 2012;Fig. 12). This suggests that the 269-248 Ma magmatism in the Khangai area occurred in an active continental margin setting related to the subduction of the Mongol-Okhotsk oceanic plate. The 238-237 Ma pegmatitic granites are classified as A2-type granites, suggesting that they formed in a post-collisional setting rather during intraplate rifting (Fig. 10c). This is consistent with the ca. 208 Ma timing of formation of post-collisional granites from the Bogd Uul pluton in the Khentii area (Khishigsuren et al., 2012). All of these data suggest that the Mongol-Okhotsk Ocean had closed by the Triassic at the latest.

Tectonic model for the formation of the Baidrag and Khangai granitoids
Combining the new field evidence, geochemical, and zircon U-Pb geochronological data presented in this study with the results of previous research enables a tectonic model for the formation of the Baidrag and Khangai granitoids to be established (Fig. 13). a) Initial metamorphism within the Baidrag craton occurred at ca. 1800 Ma (Adachi et al., 2012a). Later (ca. 560-530 Ma) northward subduction of oceanic crust beneath the craton generated adakites as a result of slab melting (Fig. 13a). This adakitic magmatism is consistent with contemporaneous (ca. 540 Ma) eclogitefacies metamorphism in the lake zone (Štípská et al., 2010). It is also likely that these slab-derived adakitic magmas did not Fig. 12. Primitive-mantle-normalized multi-element patterns showing the composition of biotite-hornblende granodiorites of the Khangai area compared to subduction-related continental arc granitoids from the Sierra Nevada of California (shaded area; Cecil et al., 2012), showing the similarity of the compositions of the granitoids from these two areas. Data are normalized to the primitive mantle composition of Sun and McDonough (1989).
provide enough heat to cause partial melting of the lower crust at this time, as evidenced by a lack of contemporaneous normal arc granites and the record of polyphase metamorphism in this area at ca. 540 Ma (Adachi et al., 2012a). b) An ocean most likely existed between the Baidrag and Siberian (?) cratons, which started to be subducted beneath the latter at ca. 270-250 Ma. This generated biotitehornblende granodiorite and biotite granite magmas that were emplaced into the Devonian to Carboniferous Khangai Basin (Orolmaa et al., 2008) during the Permian (Fig. 13b). c) The presence of ca. 240 Ma post-collisional granites in this area is indicative of magmatism related to the closure of the ocean and onset of continental collisional tectonism (Fig. 13c).

CONCLUSIONS
The new zircon U-Pb ages and whole-rock major and trace element data presented in this study allows the following conclusions to be reached: 1. Zircon U-Pb dating indicates that the magmatic activity in central Mongolia can be broadly divided into three stages that involved the formation of the 564-532 Ma Baidrag granitoids and the 269-248 and 238 -237 Ma Khangai granitoids. 2. The geochemistry of the Baidrag granitoids is indicative of an adakitic and I-type affinity, suggesting that the magmas that formed these granites were derived from a source containing two different types of source material. 3. The Khangai granitoids consist of high-K calc-alkaline granodiorite and granite units with I-type affinities that are associated with a suite of A-type granites. 4. The Baidrag granitoids suggest that arc magmatism occurred in central to southcentral Mongolia at 564-532 Ma. 5. The time gap between the Baidrag and Khangai granitoids may be associated with the development of an ocean, with the associated oceanic plate subsequently subducted beneath the Khangai area, generating voluminous granitic magmas at 269-248 Ma that subsequently formed the Khangai arc granitoids. 6. The closure of this ocean was followed by post-collisional granitoid magmatism at 238-237 Ma as evidenced by the later Khangai granitoids. These granitoids may form the western end of the scattered granitic magmatism at this time within the MOFB.
International Cooperation Agency (JICA), and I would like to express my sincere thanks to the Japanese government for this financial support.