Petrogenesis and Geological Implications of the Oligocene Mingze monzodiorites, Southern Lhasa

: The Mingze Cu-Mo deposit is located in the southern margin of the Lhasa block of the Himalayan Tibetan Plateau. Here, we report the geochronological and geochemical data from Mingze monzodiorites, which hosts the Mingze deposit. Zircon dating indicates that the Mingze monzodiorites were emplaced at ca. 31 Ma (i.e., the Oligocene). The monzodiorites have variable SiO 2 and MgO contents, strongly negative high ﬁeld-strength element (HFSE, such as Ta, Nb, Zr and Hf) anomalies on the normalized trace element diagram and show uniform ( 87 Sr / 86 Sr) i (0.7066–0.7076), ε Nd(t) ( − 2.50 to − 4.04) and ε Hf(t) ( + 1.50 to + 7.50). Their geochemical compositions are di ﬀ erent from coeval (40–30 Ma) adakite-like rocks but comparable to coeval maﬁc enclaves and gabbros. We propose that Mingze monzodiorites were derived from partial melting of the lithospheric mantle, which previously metasomatized by the subducted Indian continental plate that probably subducted into the overlying mantle. The concurrency of the genetically related maﬁc enclaves and associated intermediate to maﬁc rocks implies the heterogeneity of the Lhasa lower crust.


Introduction
The Himalayan Tibetan Plateau is interpreted as the result of the Cenozoic collision between the Indian and Asian continents and subsequent continental subduction. It is considered as the most outstanding natural laboratory for studying the evolution of a continental collision orogenesis [1]. The Lhasa block is part of the Tibetan Plateau and an archetype of a Cenozoic collisional orogen related to the India-Asia collision [1][2][3][4][5]. Thus far, it has attracted much attention from numerous scholars.
The Mingze deposit is located in the Gangdese Kelu-Chongmuda metallogenic belts, which include the Chongmuda Cu-Au deposit, Chengba Cu-Mo deposit, Nuri Cu-Mo-Au deposit and Mingze Cu deposit [6]. Most of the previous studies were concentrated on the Chongmuda, Nuri and Chengba deposit, whereas the Mingze deposit was relatively understudied. Moreover, numerous previous geochronological works have been conducted, though few integrate the geochronologic data with petrogenesis studies [7][8][9][10][11]. The Mingze deposit is regionally near to the Chengba deposit (~500 m), and people could have treated them as equivalent. In reality, their genesis and ore-associated intrusions are different as pointed out in this study.
The central Lhasa sub-block was once a microcontinent, which acted as a nucleus for the forming of the northern and southern blocks, and it has ancient basement rocks of Proterozoic and Archean ages [21]. The southern Lhasa sub-block contains widely distributed Mesozoic to Cenozoic intrusive and volcanic rocks (Figure 1c), which are composed of Upper Jurassic-Upper Cretaceous (136-93 Ma) Sangri Formation, lower Jurassic (190-174 Ma) Yeba formation, Paleogene Linzizong volcanic rocks and Cretaceous to Eocene Gangdese granitoids [20].
The Mingze area, which is located in Zedong county of Tibet, is within the southern Lhasa sub-block south of Yalung Tsangpo River. The Mingze ore block and Chengba ore block, named after their respective region, are also collectively named the Mingze deposit ( Figure 1d) [6].
Zheng et al. [7] studied the Chengba granitoids, which were treated as Mingze rocks. The Mingze deposit is a typical skarn Cu deposit. The strata in the Mingze Cu deposit are mainly Triassic metamorphic siltstone, Cretaceous metamorphic siltstone, andesitic tuff and sericite hornfels. In this study, we focused on the petrogenesis of the monzodiorites from the area containing the Mingze Cu deposit.

Samples and Methods
Seven monzodiorite samples were collected from the least altered out crops of the Mingze Cu deposit. The Mingze monzodiorites show a typical massive structure and a medium-grained granular texture (Figure 2a,b). These monzodiorites are mainly composed of plagioclase (55-60%), K-feldspar (15-20%) and amphibole (10-15%), with minor amounts of pyroxene, quartz, biotite, epidote and zircon (Figure 2c,d). Some minerals have suffered slight alteration, such as Fe oxides occurring along the margins of biotites. Plagioclase and K-feldspar crystals are euhedral or subhedral, whereas quartz crystals are subhedral grains and show rounded outlines. Amphibole grains are subhedral and their colors range from yellow to brown. All samples were analyzed for whole-rock major-and trace-element geochemistry, five of which were further analyzed for whole-rock Sr-Nd isotopes, and a sample was selected for laser-ablation inductively coupled plasma mass spectrometer (LA-ICP-MS) zircon U-Pb isotope and in situ Hf isotope analyses.
Zircon crystals separation was performed by Yuneng Rock Ore Sorting Technique Services Limited, who used conventional heavy liquid and magnetic separation techniques [22]. After separation, zircons for U-Pb and Lu-Hf isotopic detection were handpicked under a binocular microscope, and then mounted on epoxy resin discs.
The zircon U-Pb ages and cathodoluminescence (CL) images were provided by Wuhan Sample-Solution Analytical Technology Co., Ltd. The instrument for the zircon U-Pb isotopic detection used an Agilent 7700 LA-ICP-MS (Agilent, Santa Clara, CA, USA) and a 193 nm excimer laser ablation system. For this study, the ablation spot size was set to 32 µm, the ablated material was carried by He-Ar gas and the external standard used for U-Th-Pb isotopic ratios was the Harvard zircon 91500 [23]. Details of the analytical technique are provided by [24]. All the calculations and Concordia diagrams were completed using Isoplot 3.0 [25].
In situ Lu-Hf isotopic detections were conducted at the Geochemistry Lab of College of Earth sciences, Chengdu University of Technology, Chengdu, China by using a Finnigan Neptune Plus multicollector inductively coupled plasma mass spectrometer (MC-ICP-MS) and a Coherent GeoLasPro 193-nm laser-ablation system. The detections were performed on the same zircons, which had been analyzed for U-Pb ages. Details of the analytical technique are provided by [26].
The major and trace element analyses were performed by the ALS Chemex (ALS Limited, Guangzhou, China). The Samples for major element analyses were prepared with lithium borate-lithium nitrate solvent and detected by X-ray fluorescence (XRF) with detection limits of were periodically analyzed to correct instrumental drift. Geochemical reference materials of USGS BCR-2, BHVO-2, AVG-2 and RGM-2 were used as quality control [27].

U-Pb Zircon Ages and Hf Isotopes
The zircons separated from the sample MZ-12 were euhedral to subhedral crystal shapes exhibiting short prismatic forms. Most of the zircons were transparent and colorless, their average size is 60-150 μm and the length-to-width ratio is from 1:1 to 3:1, and show euhedral oscillatory zoning in cathodoluminescence (CL) images ( Figure 3a). Zircon analyses from sample MZ-12 presented high Th and U contents (Th = 408-4692 ppm, U = 553-3925 ppm, Table S1) and high ratios of Th/U (0.40-1.78). The features of CL images and the Th, U contents indicated that all crystals have a magmatic origin [28][29][30]. High precision isotopic (Sr, Nd) measurements were carried out at Nanjing FocuMS Technology Co. Ltd (Nanjing, China). The Sr-, Nd-bearing elution were dried down and re-dissolved in 1.0 mL 2 wt.% HNO3. Small aliquots of each were analyzed using an Agilent Technologies 7700x quadrupole ICP-MS (Hachioji, Tokyo, Japan) to determine the Sr, Nd abundances. Raw data of isotopic ratios were corrected for mass fractionation by normalizing to 86 Sr/ 88 Sr = 0.1194 for Sr, 146 Nd/ 144 Nd = 0.7219 for Nd with exponential law. International isotopic standards (NIST SRM 987 for Sr, JNdi-1 for Nd) were periodically analyzed to correct instrumental drift. Geochemical reference materials of USGS BCR-2, BHVO-2, AVG-2 and RGM-2 were used as quality control [27].

U-Pb Zircon Ages and Hf Isotopes
The zircons separated from the sample MZ-12 were euhedral to subhedral crystal shapes exhibiting short prismatic forms. Most of the zircons were transparent and colorless, their average size is 60-150 µm and the length-to-width ratio is from 1:1 to 3:1, and show euhedral oscillatory zoning in cathodoluminescence (CL) images ( Figure 3a). Zircon analyses from sample MZ-12 presented high Th and U contents (Th = 408-4692 ppm, U = 553-3925 ppm, Table S1) and high ratios of Th/U (0.40-1.78). The features of CL images and the Th, U contents indicated that all crystals have a magmatic origin [28][29][30].  Twenty-four spots were analyzed on these zircons, and all yielded concordant 206Pb/238U ages ranging from 28.6 Ma to 34.3 Ma (Table S1). These analyses give a weighted mean age of 31.4 ± 0.6 Ma (MSWD = 7.9, 1σ for errors), which was identical to their concordant age of 34 ± 6 Ma (MSWD =2.1) (Figure 3b,c). Thus, the weighted mean age suggested that the monzodiorites were generated in the early Oligocene.
Twenty-one Hf analyses were corrected to the magmatic crystallization age of ca. 31.4 Ma, and yielded positive εHf(t) values of +1.50 to +7. 50. The values of 176 Yb/ 177 Hf and 176 Lu/ 177 Hf ranged from 0.007383 to 0.083558 and 0.000194 to 0.002016, respectively (Table S2). Most values of 176 Lu/ 177 Hf were less than 0.002, indicating that those zircons had not experienced a significant accumulation of radiogenic Hf after formation; thus, these ratios can be used as a representation of the Hf isotopic composition at the time of zircon formation [26]. The single-stage Hf model ages (TDM) were 403.71 to 642. 61 Ma and the two-stage ages were 632.75 to 1013.62 Ma.

Whole-Rock Geochemistry
Whole-rock major and trace-element data for rocks from the Mingze monzodiorites are presented in Table S3. The Mingze monzodiorites are mainly shoshonitic and metaluminous ( Figure 4b  Twenty-four spots were analyzed on these zircons, and all yielded concordant 206Pb/238U ages ranging from 28.6 Ma to 34.3 Ma (Table S1). These analyses give a weighted mean age of 31.4 ± 0.6 Ma (MSWD = 7.9, 1σ for errors), which was identical to their concordant age of 34 ± 6 Ma (MSWD = 2.1) (Figure 3b,c). Thus, the weighted mean age suggested that the monzodiorites were generated in the early Oligocene.
Twenty-one Hf analyses were corrected to the magmatic crystallization age of ca. 31.4 Ma, and yielded positive εHf(t) values of +1.50 to +7. 50. The values of 176 Yb/ 177 Hf and 176 Lu/ 177 Hf ranged from 0.007383 to 0.083558 and 0.000194 to 0.002016, respectively (Table S2). Most values of 176 Lu/ 177 Hf were less than 0.002, indicating that those zircons had not experienced a significant accumulation of radiogenic Hf after formation; thus, these ratios can be used as a representation of the Hf isotopic composition at the time of zircon formation [26]. The single-stage Hf model ages (TDM) were 403.71 to 642. 61 Ma and the two-stage ages were 632.75 to 1013.62 Ma.

Whole-Rock Geochemistry
Whole-rock major and trace-element data for rocks from the Mingze monzodiorites are presented in Table S3   The adakitic and arc calc-alkalic magmatic rocks from [31]. Data for comparison are from [7,15,16].

Whole-Rock Sr-Nd Isotopes
Whole-rock Sr-Nd isotope data are presented in Table S3. All isotopic ratios were calculated using the zircon U-Pb age of 31.4 Ma. Samples have relatively low initial Sr isotopic ratios [( 87 Sr/ 86 Sr) i = 0.7066-0.7076] and low εNd(t) values of −4.04 to −2.49 ( Figure 6). They have high Nd model ages (T DM (Nd)) ranging from 922 Ma to 990 Ma.

Coeval Magmatism in the Mingze Region
The relatively high Th/U ratios (Table S1), as well as the well-developed oscillatory growth zoning in the zircons from the Mingze monzodiorites, indicate a magmatic origin. The weighted mean age (31.4 ± 0.6 Ma) of zircon 206 Pb/ 238 U has relatively less error than the concordant age (34 ± 6 Ma). These data are within the range reported by Zheng et al., [7], who suggested that the crystallization age of Chongmuda-Mingze granodiorites and monzogranites were ~30Ma. Mo et al. [8] reported that the crystallization age of Chongmuda monzogranites was ~28 Ma. Several Zircon U-Pb age studies have been carried out in Cenozoic granitoids from the same area. For instance, Harrison et al. [41] and Chung et al. [9] reported ages of ~30 Ma from the Chengba granodiorites, and Jiang et al. [10] provided a ~30 Ma age for the Chongmuda quartz monzogranites. The above results are consistent with this study. Accordingly, we regard ~31 Ma as the crystallization time of Mingze monzodiorites, indicating that this intrusion of Mingze was emplaced in Early Oligocene.  [16]) for the Mingze monzodiorites, southern Tibet. Data sources are as follows: Cretaceous-Paleocene Gangdese are from [33]; Chengba mafic enclaves are from [7]; Wolong mafic enclaves are from [15]; Gaoligong-Tengliang basaltic dykes are from [34]; Zhengga gabbros are from [20]; IYTS (Indus-Yarlung Tsangpo Suture) ophiolites are from [35]; Yeba volcanic rocks are from [36]; Linzizong volcanic rocks are from [37] post-collisional ultrapotassic rocks are from [38]; Quguosha Gabbros are from [16]; Himalayan leucogranites are from [39]; data for the Dagze diabases are from [40].

Coeval Magmatism in the Mingze Region
The relatively high Th/U ratios (Table S1), as well as the well-developed oscillatory growth zoning in the zircons from the Mingze monzodiorites, indicate a magmatic origin. The weighted mean age (31.4 ± 0.6 Ma) of zircon 206 Pb/ 238 U has relatively less error than the concordant age (34 ± 6 Ma). These data are within the range reported by Zheng et al., [7], who suggested that the crystallization age of Chongmuda-Mingze granodiorites and monzogranites were~30 Ma. Mo et al. [8] reported that the crystallization age of Chongmuda monzogranites was~28 Ma. Several Zircon U-Pb age studies have been carried out in Cenozoic granitoids from the same area. For instance, Harrison et al. [41] and Chung et al. [9] reported ages of~30 Ma from the Chengba granodiorites, and Jiang et al. [10] provided a~30 Ma age for the Chongmuda quartz monzogranites. The above results are consistent with this study. Accordingly, we regard~31 Ma as the crystallization time of Mingze monzodiorites, indicating that this intrusion of Mingze was emplaced in Early Oligocene.

Fractional Crystallization
The major element compositions of the Mingze monzodiorites have a large range in composition (e.g., SiO 2 , 52.70-64.98 wt.%), relatively low Mg# values (48.39-65.68) and Cr abundances (32-190 ppm) compared to typical mantle-derived magmas. The compositional ranges are compatible with the concepts of a combination of fractional crystallization and/or crystal accumulation [42]. This suggestion is supported by the oxides versus SiO 2 and MgO diagrams (Figure 7a,c,d,f and Figure 9a,b,d), in which plots generally yield linear arrays, supporting the proposed fractional crystallization process to account for the geochemical trends.
Minerals 2020, 10, x FOR PEER REVIEW 9 of 18 compared to typical mantle-derived magmas. The compositional ranges are compatible with the concepts of a combination of fractional crystallization and/or crystal accumulation [42]. This suggestion is supported by the oxides versus SiO2 and MgO diagrams (Figure 7a, c, d, f and Figure  9a, b, d), in which plots generally yield linear arrays, supporting the proposed fractional crystallization process to account for the geochemical trends.  [43]. Data sources are the same as in Figure 4.
The trend of La/Sm and Dy/Yb ratios variations are relative to the concentration of SiO2 ( Figure  7h, i) suggesting the possibility of amphibole fractionation [44]. Given that Cr is compatible in clinopyroxene, the relatively good linear relationship of Cr vs. MgO (Figure 9c), the low CaO/Al2O3 ratios of 0.2-0.5 and the good linear relationship of the CaO/Al2O3 ratios vs. MgO abundances ( Figure  9e) indicate that the clinopyroxene fractionation also played a significant role during the evolution of the magmatic system. Furthermore, plagioclase fractionation is supported by a slight negative Eu anomaly (Figure 5b, δEu = 0.68-0.81). Neither layered structure in the outcrop nor textural evidence for crystal accumulation in thin sections have been observed. Together with the negative Eu anomaly, this precludes the possibility of crystal accumulation or that the initial source rock was relatively depleted in Eu 2+ .  [43]. Data sources are the same as in Figure 4.
The trend of La/Sm and Dy/Yb ratios variations are relative to the concentration of SiO 2 (Figure 7h,i) suggesting the possibility of amphibole fractionation [44]. Given that Cr is compatible in clinopyroxene, the relatively good linear relationship of Cr vs. MgO (Figure 9c), the low CaO/Al 2 O 3 ratios of 0.2-0.5 and the good linear relationship of the CaO/Al 2 O 3 ratios vs. MgO abundances (Figure 9e) indicate that the clinopyroxene fractionation also played a significant role during the evolution of the magmatic system. Furthermore, plagioclase fractionation is supported by a slight negative Eu anomaly (Figure 5b, δEu = 0.68-0.81). Neither layered structure in the outcrop nor textural evidence for crystal accumulation in thin sections have been observed. Together with the negative Eu anomaly, this precludes the possibility of crystal accumulation or that the initial source rock was relatively depleted in Eu 2+ .
The SiO 2 and MgO contents of Mingze monzodiorites are variable, and Mg# values of Mingze monzodiorites are lower than the mantle melt but higher than the partial melting of crustal materials (Figure 7g). These features are consistent with those of the mantle mafic enclaves in adakite-like rocks, which were generated by the mixing of mantle-and crustal-derived magmas. However, relatively homogeneous Sr-Nd-Hf isotopic compositions (( 87 Sr/ 86 Sr) i , 0.7066-0.7076; εNd(t),−2.50 to −4.04;εHf(t), +1.50 to +7.50; Figure 6a,b) and the absence of linear trend between isotopic ratios and certain elements (e.g., SiO 2 and MgO) in the Mingze monzodiorites are inconsistent with the magma mixing model ( Figure 8). Thus, the magma mixing model is not favored in this study.
Crustal contamination is almost inevitable during the ascent of mantle-derived melts through the continental crust [46]. In addition, crustal assimilation during magma ascent can cause an increase in ( 87 Sr/ 86 Sr) i and a decrease in εNd(t) as the magmas are emplaced in the upper crust [47]. Therefore, we can compare major elements and radiogenic isotopes compositions to evaluate the effect of crustal contamination on basaltic rocks [16,20]. The data show that the ( 87 Sr/ 86 Sr) i and εNd(t) do not vary with respect to MgO and SiO 2 contents (Figure 8). This feature suggests limited crustal assimilation. The suggestion is further supported by the following evidence: (1) absence of a correlation between the Sr content, Lu/Yb and Zr/Hf ratios with SiO 2 contents (Figure 7j-l); (2) the relatively homogenous zircon εHf(t) (+1.50 to +7.46), which argues for little or no crustal contamination involved in the parental magma [48]; (3) the negative Zr-Hf anomalies in the N-MORB-normalized trace element diagram (Figure 5a), which suggest the parent magmas might have not undergone crustal contamination, because crustal assimilation during emplacement would result in positive Zr and Hf anomalies [49]. Therefore, we conclude that crustal contamination played an insignificant role in the formation of the Mingze monzodiorites. Crustal contamination is almost inevitable during the ascent of mantle-derived melts through the continental crust [46]. In addition, crustal assimilation during magma ascent can cause an increase in ( 87 Sr/ 86 Sr)i and a decrease in εNd(t) as the magmas are emplaced in the upper crust [47]. Therefore, we can compare major elements and radiogenic isotopes compositions to evaluate the effect of crustal contamination on basaltic rocks [16,20]. The data show that the ( 87 Sr/ 86 Sr)i and εNd(t) do not vary with respect to MgO and SiO2 contents (Figure 8). This feature suggests limited crustal assimilation. The suggestion is further supported by the following evidence: (1) absence of a correlation between the Sr content, Lu/Yb and Zr/Hf ratios with SiO2 contents (Figure 7j-l); (2) the relatively homogenous zircon εHf(t) (+1.50 to +7.46), which argues for little or no crustal contamination involved in the parental magma [48]; (3) the negative Zr-Hf anomalies in the N-MORB-normalized trace element diagram (Figure 5a), which suggest the parent magmas might have not undergone crustal contamination, because crustal assimilation during emplacement would result in positive Zr and Hf anomalies [49]. Therefore, we conclude that crustal contamination played an insignificant role in the formation of the Mingze monzodiorites.
Metasomatism in lithospheric mantle could possibly have been triggered by both subduction and thickened crustal delamination [50,51]. However, the coeval Chengba-Chongmuda (~30 Ma) and Linzi (~28 Ma) adakitic intrusions have been interpreted as a partial melting of the lower crust that contained a certain amount of juvenile mantle-derived basaltic melts [7,52]. However, Mingze monzodiorites exhibited different major element compositions from those coeval adakite-like rocks (Figures 4, 7 and 9). This probably suggests that Mingze monzodiorites have different origin compared to those adakite-like rocks and have insignificant affinity to thickened crustal delamination ( Figure  4d). We suggest, therefore, that the parent magma of Mingze monzodiorites was metasomatized by an older source material within a subduction setting. Metasomatism in lithospheric mantle could possibly have been triggered by both subduction and thickened crustal delamination [50,51]. However, the coeval Chengba-Chongmuda (~30 Ma) and Linzi (~28 Ma) adakitic intrusions have been interpreted as a partial melting of the lower crust that contained a certain amount of juvenile mantle-derived basaltic melts [7,52]. However, Mingze monzodiorites exhibited different major element compositions from those coeval adakite-like rocks (Figures 4, 7 and 9). This probably suggests that Mingze monzodiorites have different origin compared to those adakite-like rocks and have insignificant affinity to thickened crustal delamination (Figure 4d). We suggest, therefore, that the parent magma of Mingze monzodiorites was metasomatized by an older source material within a subduction setting.
In the Lu/Yb ratio and Zr/Hf ratio versus SiO2 contents diagrams, some of the Mingze monzodiorites plotted in or near the field of crust composition (Figure 7k,l) suggest their continental arc affinity. This suggestion is supported by εNd(t) versus εHf(t) plots that shift away from the mantle array ( Figure 10a). Moreover, whole rock Th and Nb elements are useful indicators for crustal input [54]. Thus, when no crustal contamination was involved, the Nb/Yb versus Th/Yb ratio diagram is useful for distinguishing subduction-related and subduction-unrelated rocks [55]. Plots of Mingze monzodiorites in the diagram show continental arc affinity (Figure 10b). In the Lu/Yb ratio and Zr/Hf ratio versus SiO 2 contents diagrams, some of the Mingze monzodiorites plotted in or near the field of crust composition (Figure 7k,l) suggest their continental arc affinity. This suggestion is supported by εNd(t) versus εHf(t) plots that shift away from the mantle array ( Figure 10a). Moreover, whole rock Th and Nb elements are useful indicators for crustal input [54]. Thus, when no crustal contamination was involved, the Nb/Yb versus Th/Yb ratio diagram is useful for distinguishing subduction-related and subduction-unrelated rocks [55]. Plots of Mingze monzodiorites in the diagram show continental arc affinity (Figure 10b).

Origin of the Ancient Crustal Signature
We propose that the parental magmas of the Mingze monzodiorites were derived from the enriched lithospheric mantle metasomatized by crustal materials in a subduction setting beneath the Lhasa terrane. However, there remains uncertainty about the origin of crustal materials. Previous work has proposed three potential metasomatic agents for forming enriched lithospheric mantle beneath the Lhasa terrane: (1) ancient crust within Central Lhasa block; (2) subducted Tethyan oceanic slab; and (3) subducted Indian continental materials ( [16] and references therein).

Origin of the Ancient Crustal Signature
We propose that the parental magmas of the Mingze monzodiorites were derived from the enriched lithospheric mantle metasomatized by crustal materials in a subduction setting beneath the Lhasa terrane. However, there remains uncertainty about the origin of crustal materials. Previous work has proposed three potential metasomatic agents for forming enriched lithospheric mantle beneath the Lhasa terrane: (1) ancient crust within Central Lhasa block; (2) subducted Tethyan oceanic slab; and (3) subducted Indian continental materials ( [16] and references therein).
It has been proposed that Early Eocene gabbros (~53 Ma) from the Longge'er pluton in the western segment of the Gangdese batholith were derived from the partial melting of an enriched mantle wedge metasomatism by Central-Lhasa ancient crust-derived magmas [45]. However, the εHf(t) values of the Mingze monzodiorites (εHf(t) = +1.5 to +7.5) are significantly different from Longge'er gabbros (εHf(t) = −1.2 to +1.0). Positive and negative zircon εHf(t) values are attributed to juvenile and ancient crustal sources, respectively [58]. The εHf(t) data indicates that fluids derived from the ancient Central-Lhasa crust were not the source of metasomatism. Moreover, previous work has demonstrated that the ancient crustal basement rocks of Proterozoic and Archean ages exist only in the central Lhasa sub-block and there is no known mechanism to bring the basement rocks of central sub-block to the mantle beneath the southern sub-block during the Cenozoic [21,59]. Thus, we posit that no Central-Lhasa ancient crust-derived materials were involved in the origin of Mingze monzodiorites.
The remaining candidate to explain the enriched mechanism of the Mingze monzodiorites magma source, therefore, is the metasomatism by subducted Indian continental materials. The εNd(t), ( 87 Sr/ 86 Sr) i, and εHf(t) values of the Mingze monzodiorites are quite similar to those of Quguosha gabbros, which were interpreted as the result of lithospheric mantle metasomatized by melts of Indian continental sediments (Figure 6a,b). As shown in the εNd(t) versus ( 87 Sr/ 86 Sr) i diagram (Figure 6b), Mingze monzodiorites exhibit a limited range of values that fall on or near a mixing line between depleted mantle and the subducted Himalayan sediments represented by the Cenozoic Himalayan leucogranites. This indicates that such rocks can be interpreted to be the result of the depleted mantle that has been, to a certain degree, metasomatized by subducted Himalayan sediments. The isotopic compositions of the Mingze monzodiorites could have been generated by the mixing of approximately 15% Himalayan sediments and 85% mantle-derived components.

Implications for Geodynamic Processes
The addition of components derived from the subducted Indian continental crust could lead to the enrichment of the upper mantle below south Tibet [63]. In addition, as mentioned above, the parent magma of the Mingze monzodiorites is compatible with an interpretation of having incorporated approximately 10-15% Himalayan-derived sediments. Moreover, Zheng et al., [7] suggested that the Indian continent subducted into the lithospheric mantle beneath the Lhasa block and accelerated the detachment of the Lhasa lithospheric root. Nábělek et al. [64] proposed that a ca. 15 km thick layer above the seismic Moho observed in the seismic cross-section is the subducted Indian lower continental crust. Taken together, the evidence suggests that the Indian continental crust had subducted into the lithospheric mantle beneath the Lhasa block before~31 Ma.
The Mingze monzodiorites are similar in composition to those coeval mafic enclaves from Wolong and Chengba-Chongmuda, and to the gabbros from Quguosha in terms of geochemistry and petrogenesis (Figures 4, 7 and 9). Their crystallization ages were all about 40-30 Ma, implying that they originated from the same magmatic system that resulted from partial melts from an enriched lithospheric mantle. This system not only formed intermediate-basic intrusions, but also formed the mafic enclaves in granitoids. Chengba-Chongmuda and Wolong adakite-like rocks, which are the host rocks of the mafic enclaves, are interpreted as the results of partial melting of thickened low crust under the Lhasa terrane. The heat source for lower crust melting was proposed to be the intrusion of mantle-derived magma [7,15]. Thus, the presence of coeval mafic enclaves and intermediate-basic composition intrusions allows for two possibilities: (1) mantle-derived magma melted a part of the lower crust and mixed with it, then formed adakite-like rocks with mafic enclaves or (2) the mantle-derived magma did not melt the lower crust, yet formed as intermediate-basic intrusions. These possibilities suggest that the mantle-derived magma was heterothermal or that a different part the lower crust had a different melt point. The suggestion implies that the presence of coeval mafic enclaves and intermediate-basic intrusions is due to either the heterogeneity of the mantle-derived magma or the heterogeneity of the lower crust. It is hard to imagine heterogeneous mantle-derived magma forming similar rocks. In addition, the relatively homogeneous Sr-Nd-Hf ( Figure 6) isotopic compositions of those rocks conflict with the heterogeneity of the mantle-derived magma. However, variable degrees of modification, by the underplated mantle-derived basaltic melts at the base of the Lhasa crust, can lead to heterogeneity [21,65,66]. Therefore, we proposed that the lower crust is highly heterogeneous in composition along the strike of the Lhasa terrane, and that the heterogeneity of the lower crust led to the concurrency of the intermediate-basic intrusions and adakite-like rocks with mafic enclaves.

2.
The Mingze monzodiorites are the result of partial melting of the lithospheric mantle beneath the Lhasa terrane metasomatized by melts of Indian continental sediments. 3.
The Indian continental crust was subducted into the lithospheric mantle beneath the Lhasa block prior to~30 Ma.

4.
The concurrency of mafic enclaves and intermediate intrusions (~40 to 30 Ma) indicates that the lower crust is heterogeneous along the strike of the Lhasa terrane.
Author Contributions: Wrote the paper, K.Z.; designed the experiments, Z.S. and F.Z.; took part in the data collection, R.L. All authors have read and agreed to the published version of the manuscript.
Funding: This study was funded by the China Geological Survey (grant no. 12120113095700, 121201010000150011-08).