A Downgoing Indian Lithosphere Control on Along-Strike Variability of Porphyry Mineralization in the Gangdese Belt of Southern Tibet

The E-trending Gangdese porphyry copper belt in southern Tibet is a classic example of porphyry mineraliza- tion in a continental collision zone. New zircon U-Pb geochronological, zircon Hf-O, and bulk-rock Sr-Nd isotope data for the Miocene mineralizing intrusions from the Qulong, Zhunuo, Jiru, Chongjiang, and Lakange porphyry copper deposits and Eocene igneous rocks from the western Gangdese belt, together with literature data, show that both Paleocene-Eocene igneous rocks and Miocene granitoids exhibit coupled along-arc iso- topic variations, characterized by bulk-rock ε Nd (t) and zircon ε Hf (t) values increasing from ~84° to ~92°E and then decreasing toward ~95°E. These are interpreted to reflect increasing contributions of subducted Indian continental materials from ~92° to ~84°E and from ~92° to ~95°E, respectively. The Miocene mineralizing intrusions were derived from subduction-modified Tibetan lower crust repre- sented isotopically by the Paleocene-Eocene intrusions, with contributions from Indian plate-released fluids and mafic melts derived from mantle metasomatized by subducted Indian continental materials. Involvement of isotopically ancient Indian continental materials increased from east (Qulong) to west (Zhunuo), which is interpreted to reflect an increasingly shallower angle of the downgoing Indian slab from east to west, con- sistent with geophysical imaging. Exploration of Gangdese Miocene porphyry copper deposits should focus on the Paleocene-Eocene arc where the subarc mantle was mainly enriched by fluids from the subducted Neo-Tethyan oceanic slab. Neodymium-Hf isotope data for mineralizing igneous rocks from porphyry copper deposits globally show no obvious correlations with Cu endowment. Although Nd-Hf isotopes are useful for imaging lithospheric architecture through time, caution must be taken when using Nd-Hf isotopes to evaluate the potential endowment of porphyry copper deposits, because other factors such as tectonic setting, crustal thickening, magma differentiation, fluid exsolution, and ore-forming processes all play roles in determining Cu endowments and grades.


Introduction
Porphyry-type mineralization above subduction zones generally shows some variability of mineralization styles and ages and geochemistry of mineralizing intrusions along and across strike of magmatic arcs (Kay et al., 1999;Hollings et al., 2005;Sillitoe and Perelló, 2005;Jones et al., 2015;Angerer et al., 2018). However, it is not clear whether such variability occurs in continental collision zones. Many porphyry copper deposits have been recognized in the continental collision zones of the Tethys orogen, including the Gangdese belt in southern Tibet, the Chagai belt in Pakistan, and the Arasbaran-Kerman belt in Iran (Hou and Zhang, 2015). Variability of collisionrelated porphyry copper deposits along or across these orogens has been poorly understood. In this study, we focus on porphyry mineralization in the Gangdese belt. It was recently noted that the Gangdese porphyries exhibit distinct acrossstrike metallogenic variability characterized by porphyry Cu mineralization in the southern Lhasa subterrane and porphyry Mo mineralization in the central Lhasa subterrane (Fig. 1B;Sun et al., 2017b). However, it is unclear whether there is along-strike variability of porphyry copper mineralization in the Gangdese belt, despite many studies providing geochronological and geochemical data for ore-related intrusions and associated mineralization (Hou et al., 2004(Hou et al., , 2015aLi et al., 2011;Leng et al., 2013;Lu et al., 2015;Wang et al., 2015;Yang et al., 2015Yang et al., , 2016aSun et al., 2017aSun et al., , 2018. Here, we present new zircon U-Pb and Hf-O and bulkrock Sr-Nd isotope data for the Miocene ore-related intrusions from five Gangdese porphyry copper deposits (Qulong, Lakange, Chongjiang, Jiru, Zhunuo) and Eocene felsic and gabbroic rocks from the western Gangdese belt. Combined with literature data, our new results enable recognition of the contrasting isotopic characteristics of porphyry copper mineralization along the strike of the Gangdese belt, which has important implications for exploration of collision-related porphyry copper deposits.
The Lakange porphyry Cu deposit is located ~20 km east of Lhasa (Fig. 1). Its Cu tonnage and grade are unknown at this preliminary stage of exploration. Early Jurassic Yeba Formation andesite, tuff, volcaniclastic rocks, and interbedded sandstones and slates crop out in the center and north of the district. Limestones and sandstones of the Late Jurassic Duodigou Formation crop out in the south of the district and are in fault contact with the Yeba Formation. Miocene granodiorite porphyry (zircon U-Pb age = 13.6 ± 0.4 Ma) was emplaced into the Yeba Formation (Leng et al., 2016). An Re-Os molybdenite model age constrains copper mineralization to 13.4 ± 0.2 Ma (Leng et al., 2015). Potassic alteration of the granodiorite porphyry produced secondary K-feldspar, biotite, and quartz. Phyllic alteration (muscovite + illite + quartz + pyrite) typically occurred along the contact zones between granodiorite porphyry and wall rocks of the Yeba Formation. Chalcopyrite is the dominant hypogene copper sulfide, occurring as disseminations within potassic-and phyllic-altered rocks, and in quartz-chalcopyrite ± pyrite ± molybdenite and chalcopyrite ± pyrite veins.
The Jiru porphyry Cu deposit is located ~20 km west of Nanmulin ( Fig. 1). It has a resource of 42 Mt at an average grade of 0.4% Cu (Zheng et al., 2014). An early Eocene granitoid with zircon U-Pb age of 48.6 ± 0.6 Ma (Zheng et al., 2014) was intruded by a late Eocene granite (zircon U-Pb age of 36.3 ± 0.9 Wang et al., 2014) and Miocene porphyry with zircon U-Pb ages of 16.2 ± 0.4 to 15.5 ± 0.3 Ma Zheng et al., 2014;Yang et al., 2016b). Five molybdenite samples obtained from quartz-molybdenite veins that crosscut the Miocene monzogranite porphyry yielded an Re-Os weighted mean age of 15.3 ± 0.1 Ma; by contrast, another three molybdenite samples collected from quartz-molybdenite veins that crosscut the Eocene monzogranite yielded an Re-Os weighted mean age of 48.3 ± 0.4 Ma (Zheng et al., 2014). The Eocene and Miocene Re-Os ages are similar to the zircon U-Pb ages of the crosscutting intrusions, indicating that two episodes of copper mineralization may have occurred at Jiru. The hydrothermal alteration affected the Miocene monzogranite porphyry and was characterized by early potassic alteration with mineral assemblages of hydrothermal K-feldspar, biotite, and quartz and late phyllic alteration with mineral assemblages of muscovite, illite, and quartz. Chalcopyrite is the main Cu-bearing mineral, occurring as disseminations within potassic-and phyllic-altered rocks and in veins composed of quartz, chalcopyrite, and pyrite. In combination with previous literature (Zheng et al., 2014;Yang et al., 2016b), our observations of the hydrothermal alteration and mineralization assemblages do not support the occurrence of two porphyry copper systems at Jiru; we infer the deposit to be Miocene in age.

Analytical Methods
Samples of Miocene intrusions, including the Chongjiang premineralization monzogranite, Jiru intermineralization monzogranite porphyry, and Zhunuo intermineralization monzogranite porphyry, were collected for zircon U-Pb dating and Hf-O isotope analyses. Additional samples of Miocene intrusions from Qulong (premineralization monzogranite and intermineralization monzogranite porphyry), Lakange (intermineralization granodiorite porphyry), and Chongjiang (premineralization monzogranite and intermineralization monzogranite porphyry) were collected for bulk-rock Sr-Nd isotope analyses. Eocene igneous rocks from Zhunuo and Jiru were collected for zircon Hf isotope analyses. Eocene igneous rocks from the western Gangdese belt (Dejilin dacite, Xiani granite and gabbro, Zhongba monzogranite) were collected for zircon U-Pb dating, zircon Hf isotope, and bulk-rock Sr-Nd isotope analyses. The locations and characteristics of the newly studied samples are listed in Table 2, which includes 35 Sr-Nd samples, 10 zircon Hf samples, five zircon U-Pb geochronology, and three zircon O samples.

Zircon U-Pb dating
U-Pb dating of zircons from the Jiru monzogranite porphyry (JR11-1) was performed using a sensitive high-resolution ion  (Sláma et al., 2008) was used to monitor performance. During dating, several measurements of the Plešovice zircon standard gave weighted mean 206 Pb/ 238 U age of 335 ± 2 Ma (MSWD = 3, n = 15). Details about the analytical procedures are provided in Compston et al. (1984) and Lu et al. (2012). U-Pb dating of zircons from the Chongjiang monzogranite (CJ001) was performed using a SHRIMP II at the Beijing SHRIMP Center, Chinese Academy of Geological Sciences. U-Th-Pb ratios were determined relative to the Temora standard zircon (417 Black et al., 2003), and the absolute abundances were calibrated to the standard zircon M257. During the analytical period, several measurements of the Temora zircon standard gave weighted mean 206 Pb/ 238 U ages of 416.6 ± 1.5 Ma (mean square of weighted deviates [MSWD] = 2.6, n = 54). Measured compositions were corrected for common Pb using the 207 Pb method.
U-Pb dating of zircons from the Xiani granite (XL-21) and gabbro (XL-07/14), and Zhongba monzogranite (ZB-28) was performed on the laser ablation-inductively coupled plasma-mass spectrometer (LA-ICP-MS) at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan. This instrument couples a quadrupole ICP-MS (Agilient 7500a) and a UP-193 solid-state laser (193-nm; New Wave Research) with an automatic positioning system. For this study, laser spot size was set to ~32 µm, laser energy density was set to 8.5 J cm -2 , and repetition rate was set to 10 Hz. Zircon 91500 (Wiedenbeck et al., 1995) was used as external standard for U-Pb dating and was analyzed twice every five analyses. They yielded a 206 Pb/ 238 U age of ~1064 Ma, which is similar to the recommended age of 1065 Ma (Wiedenbeck et al., 1995). Isotopic ratios and element concentrations of zircons were calculated using ICPMSDataCal (Liu et al., 2008). Common Pb correction and ages of the samples were calculated using LA-ICP-MS common lead correction, according to the method of Andersen (2002). Details of the analytical procedures are provided in Liu et al. (2010).

Zircon O isotope analyses
In situ zircon oxygen isotopes of mineralizing intrusions from Chongjiang (CJ001), Jiru (JR11-1), and Zhunuo (ZN005-189) were analyzed using the SHRIMP-IIe/MC at the Beijing SHRIMP Center. The analytical spot size is approximately 23 μm in diameter. Values of oxygen isotope ratios were standardized to Vienna-standard mean ocean water (V-SMOW, 18 O/ 16 O = 0.0020052) and reported in standard per mil notation. The instrumental mass fractionation factor was corrected using the Penglai and Qinghu zircon standards with δ 18 O values of 5.31 ± 0.10‰ (2SD) and 5.4 ± 0.2‰ (2SD), respectively (Li et al., 2010. Detailed instrument conditions, analytical procedures and data reduction procedures are described in Ickert et al. (2008).

Bulk-rock Sr-Nd isotope analyses
The Sr and Nd isotope compositions of the Miocene and Eocene igneous rocks were determined by thermal ionization mass spectrometry (TIMS) in the State Key Laboratory Geological Process and Mineral Resources. The detailed analytical procedures are provided in Liu et al. (2004) and Sun et al. (2018). The measured Sr and Nd isotope ratios were normalized against 86 Sr/ 88 Sr = 0.1194 and 146 Nd/ 144 Nd = 0.7219.

Zircon Hf and O isotopes
Results for in situ zircon Hf-O isotope analyses for three Miocene mineralizing intrusions are listed in Appendix Table A2. The Chongjiang monzogranite (CJ001) has zircon εHf(t) values of 1.7 to 5.7 and δ 18 O values of 5.8 to 6.7‰. The Jiru monzogranite porphyry (JR11) has zircon εHf(t) values of -4.5 to 4.5 and δ 18 O values of 6.3 to 7.0‰. The Zhunuo monzo-

Zircon Hf isotopes
Results for in situ zircon Hf isotope analyses for the Eocene igneous rocks are listed in Appendix Table A3

Bulk-rock Sr-Nd isotopes
Results for 35 bulk-rock Sr-Nd isotope analyses for the Eocene and Miocene igneous rocks are listed in Appendix Table A4. All the Miocene samples have evolved Sr-Nd isotope compositions, with initial 87 Sr/ 86 Sr and εNd(t) values ranging from 0.70489 to 0.70565 and -4.2 to 0.0 for the Qulong monzogranite and granodiorite, 0.70517 to 0.70532 and -0.2 to 0.2 for the Lakange granodiorite, and 0.70489 to 0.70565 and -4.2 to -2.8 for the Chongjiang monzogranite.

Isotopic variations along strike of the Gangdese belt
The Gangdese batholiths in the southern Lhasa subterrane were emplaced from the Middle Triassic to the Miocene Zhu et al., 2011Zhu et al., , 2015Hou et al., 2015b, and references therein). Our syntheses of all published and new Nd-Hf isotopes indicate that there are distinct spatial isotopic variations along the Gangdese belt from precollisional to collisional magmatism (Fig. 3).
The Paleocene-Eocene (~65-40 Ma) and Miocene (~21-13 Ma) intrusions were emplaced during the India-Asia con-tinental collision over a large area from Gaer to Nyingchi (Fig. 1B). It was previously shown that the Paleocene-Eocene rocks in the western Gangdese belt (west of ∼89°E) generally have higher initial 87 Sr/ 86 Sr ratios and lower εNd(t) values than those in the eastern Gangdese belt (east of ∼89°E; Wang et al., 2015). Our new isotopic data such as the ca. 50 Ma Xiani gabbros and granites at ∼86°E are consistent with previous interpretations (Fig. 3E, F). More importantly, both the mafic and felsic Paleocene-Eocene igneous rocks show similar spatial isotopic variations with their εNd(t) and εHf(t) values increasing from ~84° to ~92°E and then decreasing toward ~95°E (Fig. 3E, F). These spatial trends are also exhibited by the Miocene (~21-13 Ma) mineralizing and barren granitoids (Fig. 3G, H), but not the precollisional Triassic-Cretaceous intrusions (Fig. 3A-D). Given that the precollisional arc rocks at ~88° to 95°E have depleted Nd-Hf isotopes (Fig. 3A-D), the more unradiogenic Nd-Hf isotopes of the Paleocene to Miocene rocks most likely reflect the involvement of the subducted Indian continental materials with low ε Nd(t) and εHf(t) values (Chu et al., 2011). This interpretation is supported by geophysical results showing a N-dipping subducted Indian continental lithosphere and transport of Indian crust north of the Yarlung Zangbo suture (Replumaz et al., 2010;Gao et al., 2016). In addition, the contributions of subducted Indian components to collisional magmas along the Gangdese belt probably decreased from ~84° to ~92°E and then increased from ~92° to ~95°E, in keeping with the spatial isotopic variations of the Paleocene to Miocene igneous rocks (Fig. 3E-H).

Genesis of Miocene porphyry copper-related intrusions
The origin of Miocene ore-related intrusions in the Gangdese belt remains highly controversial, but it is generally accepted that either subduction-modified juvenile Tibetan lower crust or metasomatized mantle-derived mafic magmas played major roles in their formation (Hou et al., 2004(Hou et al., , 2015aYang et al., 2015Yang et al., , 2016aSun et al., 2018;. The Paleocene-Eocene intrusions crop out within or nearby all known Miocene porphyry copper deposits (Fig.  4A) and have more juvenile Nd and Hf isotope compositions than the Miocene mineralizing granitoids at individual deposits when all data are recalculated to 15 Ma (Fig. 4B, C). It suggests the Miocene rocks involved a less radiogenic component, which is most likely the subducted Indian continental materials with unradiogenic Nd and Hf isotopes (Guo et al., 2013;Yang et al., 2016a;Sun et al., 2018). The Miocene mineralizing intrusions have large variations in zircon oxygen isotope ratios, which range from mantle-like (5.3 ± 0.6‰, 2σ; Valley et al., 1998) to supracrustal δ 18 O values (up to 8‰;Fig. 5), consistent with involvement of mantle and supracrustal components in their genesis (Kemp et al., 2006).

Implications for exploration
The along-strike isotopic variation along the Gangdese belt is consistent with new geophysical evidence suggesting a seg-  (Wen, 2007;Yang et al., 2009;Zheng et al., 2014;Huang et al., 2015;Wu et al., 2016;Sun et al., 2018)  mented subducting Indian slab (Figs. 3E, F, 8;Li and Song, 2018;Liu et al., 2020). The segment between ~86° and ~92°E hosts all known Gangdese Miocene porphyry copper deposits, which are found above a steeper Indian slab segment that was less metasomatized by Indian continental materials than other segments (Figs. 3E,F,8). By contrast, the segments to the west of ~86°E and east of ~92° to ~95°E were strongly metasomatized by Indian continental components above flatter slabs-these segments do not host any Miocene porphyry deposits. The lack of porphyry deposits above flatter Indian slab is at odds with previous observations that young giant porphyry deposits coincide with domains of flat slab subduction in circum-Pacific arcs . We speculate that this discrepancy is due to the continental nature of the subducting Indian slab, which likely contains less chlorine, sulfur, and water than subducting oceanic slab.
The Nd-Hf isotopes are effective in imaging large-scale lithospheric architecture through time (Hou et al., 2015b). Previous studies have proposed that isotopic values of ore-related intrusions are correlated with metal endowment of porphyry copper deposits in collisional zones Asadi, 2018;Deng et al., 2018). However, our larger Nd-Hf isotope data show that there are no obvious positive correlations between copper metal tonnage and bulk-rock ε Nd(t) and zircon εHf(t) values on a global scale (Fig. 9). We advise caution when using Nd-Hf isotopes for predicting metal endowment of porphyry deposits, because other factors such as tectonic setting, crustal thickening, magma differentiation, Fig. 5. Zircon δ 18 O versus εHf(t) for the Miocene ore-related intrusions from Qulong (QL), Chongjiang (CJ), Jiru (JR), and Zhunuo (ZN) porphyry Cu deposits. Note that the uncertainties for zircon Hf or O isotopes from previous literature were not reported and are thus not shown (Hu, Y.B., et al., 2015(Hu, Y.B., et al., , 2017Wang et al., 2015).  JR (Wang et al., 2015) CJ (Hu et al., 2017) Lhasa basement / Indian crust δ O = 5.3 ± 0.6 % (2 18 σ) (Hu, Y.B., et al., 2015) Fig. 6. Bulk-rock initial 87 Sr/ 86 Sr versus εNd(t) for the Miocene fertile and barren intrusions from the Gangdese belt. All data are recalculated to 15 Ma and are listed in Appendix Table A4.    Lh a sa te rr a n e W E H i m a l a y a t e r r a n e  Fig. 9. εHf(t) and εNd(t) values versus Cu metal tonnage (Mt) for porphyry copper deposits. Data for zircon Hf and bulk-rock Nd isotopes and metal tonnage for deposits are listed in Appendix A5. The line and symbol (circle, square, triangle) for each deposit denote the range and average of εHf(t) and εNd(t) values (note that some symbols are not in the middle of lines because the lines represent the range of value rather than 2σ). Oyu Tolgoi fluid exsolution, and ore-forming processes all play roles in determining Cu endowment and grades (Tosdal and Richards, 2001;Chiaradia, 2014;Zhang and Audétat, 2017).

Conclusions
Gangdese Paleocene-Eocene igneous rocks and Miocene granitoids have systematic along-arc isotopic variations, characterized by bulk-rock εNd(t) and zircon εHf(t) values increasing from ~84° to ~92°E and decreasing from ~92° to ~95°E. The amounts of subducted Indian continental materials involved in mantle metasomatism decreased from ~84° to ~92°E and then increased from ~92° to ~95°E during the Paleocene-Eocene due to variations in the subduction angle of the segmented Indian slab. The Miocene porphyry Cu deposits are located above a steeper Indian slab segment, and their genesis involved Indian plate-released fluids, metasomatized mantle-derived mafic magmas, and Tibetan lower crust. Our results highlight that caution must be taken when using Nd-Hf isotopes for predicting metal endowment of porphyry deposits. Geosciences, Beijing. He received a B.Sc. degree in 2002 and a Ph.D. degree in 2008 from the Liaoning Technical University. Xiang conducted postdoctoral research at the China University of Geosciences, Beijing (July 2008 to July 2010), and cooperative research at the University of Western Australia (September 2015 to September 2016) and the Australian National University (January to March 2020). His research interests focus on magmatic-hydrothermal mineralization, particularly porphyry copper mineralization.