Dating Oceanic Subduction in the Jurassic Bangong–Nujiang Oceanic Arc: A Zircon U–Pb Age and Lu–Hf Isotopes and Al-in-Hornblende Barometry Study of the Lameila Pluton in Western Tibet, China

: The subduction and close of the Mesozoic Bangong–Nujiang Ocean (BNO) led to a collision of the Lhasa and Qiangtang blocks, which formed the backbone of the Tibetan Plateau (the largest and highest plateau on Earth). However, the detailed subduction processes (in particular, the oceanic subduction processes) within the BNO are still not clear. Here, we focus on the plutonic complex of the oceanic arc in the Bangong–Nujiang suture (BNS) and report ﬁeld observations on zircon U–Pb ages, Lu–Hf isotopes, and the Al-in-hornblende barometry of quartz diorites from the Lameila pluton in western Tibet. Zircon from the quartz diorites yielded a LA-ICP-MS U–Pb age of 164 Ma. The zircon showed very positive ε Hf (t) values from 10.5 to 13.9, suggesting the Lameila pluton was likely sourced from the depleted-mantle wedge, which is in contrast with contemporary (164–161 Ma) volcanic rocks in the region that had negative ε Hf (t) values of − 7.4 to − 16.2 and a magma source from partial melting of subducted sediments. The Lameila pluton showed a temperature-corrected Al-in-hornblende pressure of 3.9 ± 0.8 kbar, corresponding to an emplacement depth of 13 ± 3 km. Therefore, the thickness of the Jurassic oceanic arc crust must have doubled since the initial growth of the oceanic arc on the BNO crust, with a crustal thickness of 6.5 km during the Middle Jurassic. In combination with previous works on volcanic rocks, this study further supports a two-subduction zone model in association with the BNO during the Middle Jurassic, namely, a north-dipping BNO–Qiangtang subduction zone and an oceanic subduction zone within the BNO. The latter oceanic subduction zone produced the depleted-mantle-derived Lameila pluton and the subducted sediment-derived volcanic rocks in the fore arc. size were used to analyze the minerals. Data were corrected online using a modiﬁed ZAF (atomic number, absorption, ﬂuorescence) correction procedure. The peak counting time was 10 s for Na, Mg, Al, Si, K, Ca, and Fe, and 20 s for Ti and Mn. The background counting time was one-half of the peak counting time on the high- and low-energy background positions. The following standards were used: sanidine (K), pyropegarnet (Fe, Al), diopsode (Ca, Mg), jadeite (Na), rhodonite (Mn), olivine (Si), and rutile (Ti).


Introduction
The Tibetan Plateau is the largest plateau on Earth, with an average elevation above 4 km [1] and crustal thickness of 60-80 km [2] over an area of 2.5 million km 2 . It was formed by the sequential amalgamation of continental or oceanic arc blocks ( Figure 1) over several orogenic cycles since the Paleozoic, and further enhanced by the collision and continuing convergence of the India and Asia continents since the Cenozoic [3][4][5]. Knowledge of the pre-Cenozoic tectonics, particularly the Lhasa-Qiangtang collision that followed the closure of the Bangong-Nujiang Ocean (BNO), are critical in understanding the formation processes of the Tibetan Plateau. While much progress has been made recently in understanding the processes of the oceanic-continental subduction and the subsequent  [43], magmatic rocks in the southern Qiangtang block are modified after Liu et al. [15], and other geologic units are modified after Yin and Harrison [5]. The boninitic andesites and 167 Ma supra-subduction zone (SSZ) type ophiolites [18,31] west of Rutog and the 161 Ma high-Mg andesites [44] north of Shiquanhe are the surface traces of the Jurassic oceanic arc within the BNS. The 164 Ma Lameila pluton newly recognized in this work lies 20 km to the WWN of the Shiquanhe high-Mg andesites. This pluton trends WWN and is 20 m long and 1-3 km wide (Figure 3). It was covered by Jurassic flysch and Cretaceous sedimentary rocks. The Lameila pluton and Jurassic flysch were thrust over the Cretaceous strata by the north-dipping Jiagang thrust [8].  [43], magmatic rocks in the southern Qiangtang block are modified after Liu et al. [15], and other geologic units are modified after Yin and Harrison [5]. The boninitic andesites and 167 Ma supra-subduction zone (SSZ) type ophiolites [18,31] west of Rutog and the 161 Ma high-Mg andesites [44] north of Shiquanhe are the surface traces of the Jurassic oceanic arc within the BNS. The 164 Ma Lameila pluton newly recognized in this work lies 20 km to the WWN of the Shiquanhe high-Mg andesites. This pluton trends WWN and is 20 m long and 1-3 km wide ( Figure 3). It was covered by Jurassic flysch and Cretaceous sedimentary rocks. The Lameila pluton and Jurassic flysch were thrust over the Cretaceous strata by the north-dipping Jiagang thrust [8].     [39]. The sample location for this study is shown as a yellow star, and published ages of ophiolites and their locations are shown as unfilled stars.

In-Situ LA-ICP-MS U-Pb Dating and Trace Elements of Zircon
Zircon separation and cathodoluminescence (CL) images were carried out at the Guangzhou Tuoyan Testing Technology Co., Ltd. (Guangzhou, China). Approximately 5 kg of a rock sample was

In-Situ LA-ICP-MS U-Pb Dating and Trace Elements of Zircon
Zircon separation and cathodoluminescence (CL) images were carried out at the Guangzhou Tuoyan Testing Technology Co., Ltd. (Guangzhou, China). Approximately 5 kg of a rock sample was crushed and sieved for separating zircon grains using standard magnetic and heavy liquid separation procedures. About 150 grains were mounted in epoxy and polished to expose the center of crystals.
In situ zircon U-Pb isotopic analyses were conducted using a RESOlution 193 nm ArF excimer laser (LA, Australian Scientific Instruments Pty Ltd., Canberra, Australia) coupled with a ThermoiCAP Qc Inductively Coupled Plasma-Mass Spectrometer (ICP-MS, Thermo Fisher Scientific Inc., Waltham, USA) at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan). The ablation system operated at a wavelength of 193 nm using a spot diameter of 33 µm at a 10 Hz repetition rate for 40 s, and ablation pits were about 20 µm deep. Helium was used as carrier gas to transport the ablated materials to the ICP-MS. Each analysis consisted of a 15 s background acquisition, 40 s sample data acquisition, and 45 s washout delay at the end. Every 10 sample analyses were followed by analysis of one glass standard (NIST 612) and two zircon standards 91500 [45]. Analytical results were calculated using ICPMSDataCal software (Version number, Liu et al., Wuhan, China) [46]. Concordia U-Pb diagrams and weighted mean 206 Pb/ 238 U age calculations were made using Isoplot v2.3 (Berkeley Geochronological Center, Berkeley, USA) [47].

In-Situ LA-MC-ICP-MS Lu-Hf Isotopes of Zircon
In situ zircon Lu-Hf isotope analyses were carried out using the RESOlution-S155 193 nm ArF excimer laser, which was attached to a Nu Plasma II MC-ICP-MS at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan). A stationary spot was used for the present analyses, with a beam diameter of 33 µm. Helium was used as the carrier gas to transport the ablated materials to the MC-ICP-MS. Atomic masses of 172, 173, 175-180, and 182 were simultaneously measured in static-collection mode. The 176 Lu/ 175 Lu = 0.02655 and 176 Yb/ 173 Yb = 0.7965 ratios were used to correct the isobaric interferences of 176 Lu and 176 Yb on 176 Hf, respectively [48]. Penglai zircon was selected as the standard during the analysis, whose analytical results ( 176 Hf/ 177 Hf = 0.282896 ± 38, mean square weighted deviation (MSWD) = 0.89, n = 25) during our one-day analyses were in good agreement with the published data within error (0.282906 ± 0.000010) [49]. To calculate the ε Hf (t) values, we adopted a decay constant for 176 Lu of 1.867 × 10 -11 year −1 [50] and chondritic present-day values of 176 Lu/ 177 Hf (0.0336) and 176 Hf/ 177 Hf (0.282785) [51]. Depleted-mantle Hf model ages (T DM ) were calculated using the measured 176 Lu/ 177 Hf ratios of zircon, assuming that the depleted-mantle reservoir had a 176 Hf/ 177 Hf of 0.283250 at present day, with a 176 Lu/ 177 Hf value of 0.0384 [52]. The mantle extraction model age (T DM C ) for the source rocks of the magmas was calculated by projecting initial 176 Hf/ 177 Hf ratios of the zircons to the depleted-mantle model growth line using a mean 176 Lu/ 177 Hf value (0.015) for average continental crust [53].

EPMA Amphibole Chemistry
Mineral compositions were determined at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (Wuhan), with a JEOL JXA-8100 Electron Probe Micro Analyzer (EPMA, JEOL Ltd., Tokyo, Japan) equipped with four wavelength-dispersive spectrometers (WDSs). The samples were coated with a thin conductive carbon film prior to analysis. The precautions suggested by Zhang and Yang [54] were used to minimize the difference of carbon film thickness between samples and standards and obtain an approximately uniform~20 nm coating. During the analysis, an accelerating voltage of 15 kV, a beam current of 20 nA and a 1-µm spot size were used to analyze the minerals. Data were corrected online using a modified ZAF (atomic number, absorption, fluorescence) correction procedure. The peak counting time was 10 s for Na, Mg, Al, Si, K, Ca, and Fe, and 20 s for Ti and Mn. The background counting time was one-half of the peak counting time on the high-and low-energy background positions. The following standards were used: sanidine (K), pyropegarnet (Fe, Al), diopsode (Ca, Mg), jadeite (Na), rhodonite (Mn), olivine (Si), and rutile (Ti). Precision of the EPMA analysis was calculated from counting statistics and was generally better than ±1% for contents of >10 wt %, and better than ±5% for contents of >0.5 wt %.

Zircon U-Pb Geochronology and Trace Elements
Zircon grains in sample 18TG49 from the Lameila pluton were euhedral, with sizes that were 100-200 µm long and 80-120 µm wide. They showed oscillatory zoning patterns ( Figure 5), and had low Th (68-318 ppm) and U (129-460 ppm) concentrations with high Th-U ratios (0.50-0.72). The U-Pb geochronological and trace element data are listed in Tables 1 and 2. Twenty-seven of the 30 total analyses yielded concordant 206 Pb/ 238 U and 207 Pb/ 235 U ages, with 206 Pb/ 238 U ages between 158 and 169 Ma. The weighted mean age was 164.0 ± 1.1 Ma (n = 27, MSWD = 1.3, Figure 6a,b). The oscillatory zoning patterns, Th and U concentrations, and typical magmatic zircon rare earth element (REE) patterns ( Figure 6c) imply that this age represents the time of magma emplacement.       Table 2) shown as gray lines in (c) are likely the result of tiny inclusions of mineral or fluid with slightly higher light rare earth elements (LREE) than the host zircon.

In-Situ Zircon Lu-Hf Isotopes
Zircons from the Lameila pluton had 176 Hf/ 177 Hf ratios ranging from 0.282972 to 0.283068 (Table 3), the initial 176 Hf/ 177 Hf ratios calculated using the corresponding zircon ages were 0.282966-0.283063, and the ε Hf (t) values ranged from 10.5 to 13.9 (  Table 2) shown as gray lines in (c) are likely the result of tiny inclusions of mineral or fluid with slightly higher light rare earth elements (LREE) than the host zircon.

In-Situ Zircon Lu-Hf Isotopes
Zircons from the Lameila pluton had 176 Hf/ 177 Hf ratios ranging from 0.282972 to 0.283068 (Table  3), the initial 176 Hf/ 177 Hf ratios calculated using the corresponding zircon ages were 0.282966-0.283063, and the εHf(t) values ranged from 10.5 to 13.9 ( Figure 7). These had very young TDM ages of 270-410 Ma, with an average of 340 Ma. The TDM C ages were between 320 and 540 Ma, with an average age of 440 Ma.   [15]; data for the BNO high-Mg andesite are from Liu et al. [19] and Zeng et al. [20].

Amphibole Chemistry and Stoichiometry
Results of the EPMA analyses for amphibole chemistry are shown in Table 4. Locations for EPMA analysis were near the margins in contact with quartz or K-feldspar. The amphibole compositions of the three quartz diorite samples were broadly the same; therefore, we present the major element compositions and assess the dominance of atomic exchange reactions by treating all of the amphibole crystals as one population. All the amphiboles from the Lameila pluton displayed a narrow compositional range in SiO 2 (45.33-49.84 wt %, average = 47.13 wt %) and CaO (10.30-11.88 wt %, average = 11.23 wt %). Their Al 2 O 3 contents were uniform and low, ranging between 6.32 and 8.81 wt % (average = 7.86 wt %), there are some variation for MgO (11.81-15.41 wt %, average = 12.96 wt %) and total FeO (12.52-17.25 wt %, average = 15.23 wt %) compositions (Table 4).   Spot  TG50-1 TG50-2 TG50-3 TG50-4 TG50-5 TG50-6 TG50-7 TG50-8 TG51-1 TG51-2 TG51-3 Formulae based on 23 oxygen and 13eCNK The structural formulae and stoichiometry of amphiboles were calculated based on an anhydrous basis assuming 23 oxygen and 13 cations, excluding Ca, Na, and K (13eCNK). Fe 3+ and Fe 2+ contents were calculated assuming a charge of 46, as suggested by Leake et al. [56]. All of the Lameila amphiboles were calcic (Na B = 0.13-0.32 atoms per formula unit (apfu), Ca B + Na B = 1.70-2.19 apfu) and are classified as magnesiohornblende (Mg# (Mg/(Mg + Fe 2+ )) = 0.65-0.93, Si = 6.66-7.07 apfu, Figure 8). Leake et al. [56] suggested that igneous amphiboles have a maximum of Si = 7.3 apfu, and those with Si > 7.3 apfu were crystallized under sub-solidus conditions in the presence of a fluid or had been altered by hydrothermal fluids; therefore, the low Si signature of the Lameila hornblende suggests a magmatic origin, which is in accordance with the petrographic observations.

Key Control Factors on the Variation of Amphibole Compositions
The amphibole compositions are sensitive to physical-chemical conditions in magma systems, such as changing melt composition, pressure, temperature, volatile content, and co-crystallizing mineral phases [57]. Therefore, it is important to determine the dominating parameters that are responsible for the compositional variations of minerals. Variations in Al for the Lameila hornblendes are largely accommodated by the tetrahedral site (Al IV = 0.93-1.34 apfu). The amount of Al in the C site (Al VI = 0.05-0.29 apfu) is minor. Atom exchanges, such as the Al-Tschermak exchange, Ti-Tschermak exchange, the edenite exchange, and the plagioclase substitution are proposed to be significant mechanisms controlling compositional variations in amphiboles.
Elemental correlation diagrams of the Lameila amphiboles ( Figure 9) imply that the contributions from both the pressure-sensitive Al-Tschermak exchange and the temperature-sensitive Ti-Tschermak exchange were negligible, as evidenced by the lack of covariations of Al VI and Ti C with Al IV (Figure 9a,b). Instead, the slopes of the regression lines in the (Na + Ca) A -Al IV and the Ca B -Al IV diagrams (Figure 9b,c) suggest that about 55% of the Al IV variations were accommodated by the temperature-sensitive edenite exchange and plagioclase substitution. The remaining proportions for the Al IV and total Al variation may have been controlled by some other factors, such as fractional crystallization of the host magma, or changes in the redox state during the crystallization of the Lameila amphiboles (Figure 9d-f). The effects of the fractional crystallization processes and the changes of the redox state on Al variation may have somewhat overlapped with those of the edenite exchange and plagioclase substitution; however, the extent to which the former two factors contributed to the Al variations of the Lameila amphiboles cannot be quantitatively evaluated in the current work. Nevertheless, the existence of the temperature-sensitive edenite exchange and the plagioclase substitution in the Lameila samples demands temperature correction for the Al-in-hornblende barometer.

Al-in-Hornblende Barometer and Depth of the Lameila Pluton Emplacement
Empirical calibration and experimental studies show that the total Al composition in calcic amphiboles increases linearly with crystallization pressure. This phenomenon forms the foundation of Al-based barometers for calcic amphiboles [57][58][59][60][61]. An important prerequisite for the applications

Al-in-Hornblende Barometer and Depth of the Lameila Pluton Emplacement
Empirical calibration and experimental studies show that the total Al composition in calcic amphiboles increases linearly with crystallization pressure. This phenomenon forms the foundation of Al-based barometers for calcic amphiboles [57][58][59][60][61]. An important prerequisite for the applications of Al-in-hornblende barometers is an appropriate buffering assemblage (amphibole + plagioclase + K-feldspar/quartz, with medium to high oxygen fugacity). The lithology and petrography of the Lameila pluton (Figure 4) show that the above buffering assemblage was satisfied. In addition, the magma in which the Lameila amphiboles crystallized had a high oxygen fugacity. This inference is supported by the elevated Ce/Ce* and Eu/Eu* in the Lameila zircons ( Figure 10) and the occurrence of primary magnetite grains within and along the margins of amphibole crystals (Figure 4c). We first calculated the emplacement pressure using Al-in-hornblende barometers without temperature correction. For this purpose, we employed several widely used barometers, including calibrations of PHZ86 [61], PH87 [60], PJR89 [59], PS92 [62], and PEL98 [58]. Results of the calculated pressures are shown in Table 4.
We then used the temperature-corrected calibration of Anderson and Smith [57] to evaluate the emplacement pressure. The temperatures for amphibole crystallization were estimated by employing several kinds of methods ( Figure 11). The first (TO84) was the Ti-in-amphibole calibration of Otten [63], which was first reported by Helz [64]. The second (TEL98) was the amphibole thermometer of Ernst and Liu [58]. Results of these two thermometers were further compared with temperatures (TTiin-zircon) derived from the Ti-in-zircon thermometer [65]. Ranges of TO84 (619-687 °C, average = 642 °C) and TEL98 (569-701 °C, average = 641 °C) largely overlapped with each other, but the results of TEL98 were more scattered than TO84. The TTi-in-zircon can record the temperature at which a zircon grain crystallizes in the host magma. Stages for zircon crystallization in an intermediate/felsic magma could be either early or late, and thus TTi-in-zircon of different zircon grains could record whole-stage temperatures of a magma system. Amphiboles crystallized early in dioritic magma during emplacement, and overlap between the ranges of TTi-in-zircon and TO84 was the best estimation for the temperatures for amphibole crystallization. The trend of TTi-in-zircon with increasingly younger ages was likely the result of heating by magma replenishment after the crystallization of zircon grains with older ages (Figure 11c). This batch of magma replenishment was recorded in the dark-colored amphibole-rich xenoliths of diorites that were in sharp contact with the host Lameila quartz diorites. The replenished magma probably delivered heat but not materials to the quartz diorites, which is evidenced by the linearly decreasing Mg with Al (Figure 9e). On this basis, the TO84 results are used in this study for correction of the Al-in-hornblende barometer (PAS95_TO84) proposed by Anderson and Smith [57]. The corrected PAS95_TO84 ranged from 3.6 to 4.3 kbar except for four outliers (2.2-2.7 kbar, Figure 11d). The outliers with elevated temperatures were likely the result of heating as recorded by the TTi-in-zircon trend (Figure 11c). The average value of the majority was 3.89 ± 0.28 kbar (n = 18, MSWD We first calculated the emplacement pressure using Al-in-hornblende barometers without temperature correction. For this purpose, we employed several widely used barometers, including calibrations of P HZ86 [61], P H87 [60], P JR89 [59], P S92 [62], and P EL98 [58]. Results of the calculated pressures are shown in Table 4. We then used the temperature-corrected calibration of Anderson and Smith [57] to evaluate the emplacement pressure. The temperatures for amphibole crystallization were estimated by employing several kinds of methods ( Figure 11). The first (T O84 ) was the Ti-in-amphibole calibration of Otten [63], which was first reported by Helz [64]. The second (T EL98 ) was the amphibole thermometer of Ernst and Liu [58]. Results of these two thermometers were further compared with temperatures (T Ti-in-zircon ) derived from the Ti-in-zircon thermometer [65]. Ranges of T O84 (619-687 • C, average = 642 • C) and T EL98 (569-701 • C, average = 641 • C) largely overlapped with each other, but the results of T EL98 were more scattered than T O84 . The T Ti-in-zircon can record the temperature at which a zircon grain crystallizes in the host magma. Stages for zircon crystallization in an intermediate/felsic magma could be either early or late, and thus T Ti-in-zircon of different zircon grains could record whole-stage temperatures of a magma system. Amphiboles crystallized early in dioritic magma during emplacement, and overlap between the ranges of T Ti-in-zircon and T O84 was the best estimation for the temperatures for amphibole crystallization. The trend of T Ti-in-zircon with increasingly younger ages was likely the result of heating by magma replenishment after the crystallization of zircon grains with older ages (Figure 11c). This batch of magma replenishment was recorded in the dark-colored amphibole-rich xenoliths of diorites that were in sharp contact with the host Lameila quartz diorites. The replenished magma probably delivered heat but not materials to the quartz diorites, which is evidenced by the linearly decreasing Mg with Al (Figure 9e). On this basis, the T O84 results are used in this study for correction of the Al-in-hornblende barometer (P AS95_TO84 ) proposed by Anderson and Smith [57]. The corrected P AS95_TO84 ranged from 3.6 to 4.3 kbar except for four outliers (2.2-2.7 kbar, Figure 11d). The outliers with elevated temperatures were likely the result of heating as recorded by the T Ti-in-zircon trend (Figure 11c). The average value of the majority was 3.89 ± 0.28 kbar (n = 18, MSWD = 0.11), which was interpreted as the pressure of the Lameila pluton emplacement. Uncertainties for the pressure estimation were the followings: (1) analytical uncertainties are better than 1%; (2) uncertainties with the Al-in-hornblende calibration are within 0.6 kbar; (3) the standard deviation for PAS95_TO84 results is 0.19 kbar. Taken all together, the uncertainties for the pressure would be better than 0.83 kbar. As a consequence, the pressure for the Lameila pluton is 3.89 ± 0.83 kbar. The average continental crust density is 2.700-2.875 g/cm 3 , and the oceanic crust is 2.700-2.950 g/cm 3 [66]. The density of the oceanic arc crust in the BNO is not clear, but should be between the above two crusts. If we employ a conservative density of 2.950 g/cm 3 , the depth of the Lameila pluton emplacement would be 13.2 ± 2.8 km.

Mechanism for the Crustal Thickening of the Jurassic Oceanic Arcin the BNO
Previous studies on high-Mg volcanic rocks within the BNS suggest a Jurassic oceanic arc that formed during the intra-oceanic subduction of the BNO. However, the configuration of this oceanic arc, including the crustal thickness, is not clear. As mentioned in Sections 5.1 and 5.2, this oceanic arc in the BNO has grown significantly. The Al-in-hornblende barometer shows that the 164 Ma Lameila pluton in western Tibet intrudes in the upper plate of the Jurassic oceanic subduction system within the BNO at a depth of about 13 km. This emplacement depth is the minimum crustal thickness of the Jurassic oceanic arc within the BNO, since the thickness of the crust underneath the emplacement level of the Lameila pluton is unknown. This depth of emplacement implies that the Jurassic oceanic arc crust has thickened significantly since the initial development of this oceanic arc on the oceanic crust of the BNO. According to studies on both modern and paleo-oceanic arcs around the world, arc crust thicknesses depend on the thickness of pre-arc basement, tectonic extension or shortening, and arc maturity (magmatic addition and evolution) [67]. The pre-arc basement crust consists of mainly mafic oceanic crust and minor sedimentary cover. Studies on the ophiolites show that the Uncertainties for the pressure estimation were the followings: (1) analytical uncertainties are better than 1%; (2) uncertainties with the Al-in-hornblende calibration are within 0.6 kbar; (3) the standard deviation for P AS95_TO84 results is 0.19 kbar. Taken all together, the uncertainties for the pressure would be better than 0.83 kbar. As a consequence, the pressure for the Lameila pluton is 3.89 ± 0.83 kbar. The average continental crust density is 2.700-2.875 g/cm 3 , and the oceanic crust is 2.700-2.950 g/cm 3 [66]. The density of the oceanic arc crust in the BNO is not clear, but should be between the above two crusts. If we employ a conservative density of 2.950 g/cm 3 , the depth of the Lameila pluton emplacement would be 13.2 ± 2.8 km.

Mechanism for the Crustal Thickening of the Jurassic Oceanic Arcin the BNO
Previous studies on high-Mg volcanic rocks within the BNS suggest a Jurassic oceanic arc that formed during the intra-oceanic subduction of the BNO. However, the configuration of this oceanic arc, including the crustal thickness, is not clear. As mentioned in Sections 5.1 and 5.2, this oceanic arc in the BNO has grown significantly. The Al-in-hornblende barometer shows that the 164 Ma Lameila pluton in western Tibet intrudes in the upper plate of the Jurassic oceanic subduction system within the BNO at a depth of about 13 km. This emplacement depth is the minimum crustal thickness of the Jurassic oceanic arc within the BNO, since the thickness of the crust underneath the emplacement level of the Lameila pluton is unknown. This depth of emplacement implies that the Jurassic oceanic arc crust has thickened significantly since the initial development of this oceanic arc on the oceanic crust of the BNO. According to studies on both modern and paleo-oceanic arcs around the world, arc crust thicknesses depend on the thickness of pre-arc basement, tectonic extension or shortening, and arc maturity (magmatic addition and evolution) [67]. The pre-arc basement crust consists of mainly mafic oceanic crust and minor sedimentary cover. Studies on the ophiolites show that the composition of the oceanic crust is dominated by regular N-MORB and E-MORB type mafic rocks, and that oceanic crust related to OIB (oceanic island basalt) or oceanic plateaus are not recognized in the Shiquanhe area [10,68]. Therefore, the thickness of pre-arc oceanic crust is most likely comparable to that of the average oceanic crust (6.5 km) [69]. Rocks with levels higher than the Lameila pluton have been removed, thus much of the Lameila pluton is exposed at the surface (Figure 4a), and the outcrops of sedimentary rocks with pre-arc ages are extremely limited in the vicinity of the Lameila pluton and the Shiquanhe high-Mg andesite ( Figure 3). Therefore, the pre-arc sedimentary covers on the floor of the BNO may be quite thin and thus their contribution to pre-arc crustal thickness is negligible. The effect of tectonic shortening on the thickening of the Jurassic oceanic arc in the BNO could not be quantitatively evaluated in this work due to the very limited preservation of structural traces.
Alternatively, magmatism may play a considerable role in the thickening of the Jurassic oceanic arc within the BNO. Studies on modern oceanic arcs, such as the Izu-Bonin-Marina [70][71][72][73], the Lesser Antilles [74][75][76], the Tonga-Kermadec [77][78][79], and the Aleutian [80][81][82] arcs show a variation of crustal thickness (9-35 km) within a single arc or between different arcs. In addition, an important finding of these studies is that the thick part of an arc is always related to high magmatic flux. This phenomenon implies that magmatic addition is a dominant mechanism responsible for oceanic arc crustal thickening. The lithological assemblage of the Jurassic oceanic arc in the BNS comprises only magmatic rocks, such as the Lameila quartz diorites, the Shiquanhe and Daruco high-Mg andesites, and the Dingqing high-Mg quartz gabbro/diabase. Pending additional studies of detailed tectonic evolution, we tentatively propose the crustal thickening of the Jurassic oceanic arc in the BNO to be a result of magmatic addition.

Oceanic Arc Development and Two Contrasting Sources for Magmatism
The east-west trending BNS in the Tibetan Plateau is a belt more than 1200 km long consisting of scattered ophiolitic fragments that are associated with thick sequences of Jurassic flysch. The BNS is extraordinarily wide, especially near Amdo in eastern Tibet and between Rutog and Shiquanhe in western Tibet (Figure 1). Besides the ophiolitic fragments and the Jurassic flysch, there are also micro-continental blocks such as the Amdo micro-block, and possibly oceanic arc blocks within the BNS. The mechanism incorporating these micro-blocks into the BNS is unclear.
In contrast to the well-exposed Amdo micro-block, surface traces of the oceanic arc rocks in the BNS are highly obscure. Previous recognition of BNS oceanic arc rocks has been mainly based on geochemical evidence from the volcanic rocks that are associated with ophiolitic fragments (SSZ type ophiolite), and suggests that an oceanic subduction occurred within the BNO [10,31,[83][84][85]. Importantly, the discoveries of boninitic rocks near Dingqing [17], Daruco [20], Shiquanhe [19], and Rutog [18] in the BNS provide further evidence of the existence of an oceanic arc within the BNO. However, the extremely localized outcrops of these volcanic rocks make it hard to trace them in the field, and thus hamper further studies on the development and configuration of this arc. The new data from the Lameila pluton of this study have three implications for the tectonic evolution of the BNO. Firstly, this study recognizes the intrusive part of the oceanic arc, and in combination with the previously reported extrusive rocks, further confirms the existence of a Jurassic oceanic arc within the BNO. Secondly, the emplacement depth of the Lameila pluton suggests that the crust of the oceanic arc within the BNO was significantly thickened during the end of the middle Jurassic. Thirdly, the zircon Lu-Hf isotopes of the Lameila pluton, together with that of the previously reported high-Mg andesites within the BNS, imply diverse sources for the oceanic arc magmatism as shown below.
Studies on the high-Mg andesites from Shiquanhe ( isotopes. It has been proposed that these high-Mg andesites are formed by the melting of subducted sediments ( Figure 12) and subsequent interaction with overlying mantle peridotites during subduction initiation [19,20]. In contrast with the enriched isotopes of the above high-Mg andesites, zircon Lu-Hf isotopes of the Lameila pluton were depleted, with ε Hf (t) ranging from 10.5 to 13.9. The depleted Lu-Hf isotopes of the Lameila pluton clearly indicate that they were formed by the partial melting of a depleted-mantle source, most likely the depleted-mantle wedge peridotites above the oceanic subduction zone in the BNO (Figure 12). Synchronous with this oceanic arc in the BNO, a Jurassic continental arc (Figure 1) in the south margin of the Qiangtang block has also developed, as suggested by many previous researchers [13,15,16,29,38,39]. These two arcs suggest the occurrence of two subduction zones during the Middle Jurassic, namely, an oceanic-continental subduction zone between the BNO and the Qiangtang block in the north, and an oceanic subduction zone within the BNO in the south ( Figure 12). The Jurassic oceanic arc presented in this study and the Amdo micro-continental block ( Figure 1) were accreted onto the southern margin of the Qiangtang block during the continued subduction of the BNO, or during the Lhasa-Qiangtang collision. The accretion of the oceanic arc and micro-block could lead to a wide Bangong-Nujiang suture in central Tibet (Figure 2d). Jurassic continental arc (Figure 1) in the south margin of the Qiangtang block has also developed, as suggested by many previous researchers [13,15,16,29,38,39]. These two arcs suggest the occurrence of two subduction zones during the Middle Jurassic, namely, an oceanic-continental subduction zone between the BNO and the Qiangtang block in the north, and an oceanic subduction zone within the BNO in the south ( Figure 12). The Jurassic oceanic arc presented in this study and the Amdo microcontinental block ( Figure 1) were accreted onto the southern margin of the Qiangtang block during the continued subduction of the BNO, or during the Lhasa-Qiangtang collision. The accretion of the oceanic arc and micro-block could lead to a wide Bangong-Nujiang suture in central Tibet ( Figure  2d).

Conclusions
This work reports a study of the plutonic part of the Jurassic oceanic arc within the Bangong-Nujiang suture in western Tibet. The Lameila pluton consists mainly of quartz diorites with a zircon U-Pb age of 164 Ma. In addition, the zircon εHf(t) values of 10.5-13.9 indicate that the Lameila pluton was most likely sourced from the depleted-mantle wedge. Temperature-corrected Al-in-hornblende pressure was 3.9 ± 0.8 kbar, and the corresponding depth of emplacement was 13 ± 3 km. The thickness of the Jurassic oceanic arc crust must have doubled since the initial growth of the oceanic arc on the Bangong-Nujiang Ocean crust, with a thickness of 6.5 km during the Middle Jurassic. Acknowledgments: We thank Zheng-Han Li for his assistance in the field, and Kui-Dong Zhao and Shui-Yuan Yang for their assistance with U-Pb-Lu-Hf isotopes and mineral chemistry measurements. We thank the editors for editorial handling and suggestions. We thank three anonymous reviewers for useful comments and suggestions.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Conclusions
This work reports a study of the plutonic part of the Jurassic oceanic arc within the Bangong-Nujiang suture in western Tibet. The Lameila pluton consists mainly of quartz diorites with a zircon U-Pb age of 164 Ma. In addition, the zircon ε Hf (t) values of 10.5-13.9 indicate that the Lameila pluton was most likely sourced from the depleted-mantle wedge. Temperature-corrected Al-in-hornblende pressure was 3.9 ± 0.8 kbar, and the corresponding depth of emplacement was 13 ± 3 km. The thickness of the Jurassic oceanic arc crust must have doubled since the initial growth of the oceanic arc on the Bangong-Nujiang Ocean crust, with a thickness of 6.5 km during the Middle Jurassic.