Unroofing the Ladakh Batholith: constraints from autochthonous molasse of the Indus Basin, NW Himalaya

The Indus Molasse records orogenic sedimentation associated with uplift and erosion of the southern margin of Asia in the course of ongoing India–Eurasia collision. Detailed field investigation clarifies the nature and extent of the depositional contact between this molasse and the underlying basement units. We report the first dataset on detrital zircon U–Pb ages, Hf isotopes and apatite U–Pb ages for the autochthonous molasse in the Indus Suture Zone. A latest Oligocene depositional age is proposed on the basis of the youngest detrital zircon U–Pb age peak and is consistent with published biostratigraphic data. Multiple provenance indicators suggest exclusively northerly derivation with no input from India in the lowermost parts of the section. The results provide constraints on the uplift and erosion history of the Ladakh Range following the initial India–Asia collision. Supplementary material: Method description, Sample locations, Figures S1 and S2 and Tables S1 to S3 are available at: https://doi.org/10.6084/m9.figshare.c.4858848

The uplift of the Earth's highest mountain range, the Himalaya, has been a central topic in geology because of the critical role these mountains play in the evolution of regional and global climate and oceanography (e.g. Raymo and Ruddiman 1992;Richter et al. 1992). Many methods have been suggested for inferring topographic growth, including stable-isotope-based palaeoaltimetry (e.g. Huntington and Lechler 2015), fossil leaf morphology (e.g. McElwain 2004), geomorphic features (e.g. Whipple and Tucker 1999), and low-temperature thermochronology (e.g. Reiners and Brandon 2006). Geological constraints including the exhumation of deep-seated lithological units and the deposition of intermontane sediments also serve to record the growth of high topography. In the Himalaya, continental deposits along the Indus-Yarlung-Tsangpo suture zone (IYTS) chronicle a history of collision between India and Eurasia and the attendant rise of the Himalaya-Tibetan Plateau (e.g. Searle 1983;Garzanti and Van Haver 1988;Jaeger et al. 1989;Aitchison et al. 2007;van Hinsbergen et al. 2012). Depositional ages and source-to-sink relationships constrain aspects of past ocean closure and provide information on the uplift of source terrains, yet there are challenges and complications in studying related rock units. First, these coarse-grained deposits do not always contain suitable fossils or volcanic ash layers that may provide true depositional ages. One common approach is to employ detrital zircon U-Pb or mica 40 Ar/ 39 Ar dating to constrain the maximum depositional age. However, the utility of results obtained using this methodology could be strongly influenced by the size of the dataset. The common practice of dating c. 100-200 mineral grains from a given lithological unit might not be sufficient. Second, commonly used provenance indicators, including detrital mineral ages and radiogenic isotopic compositions, geochemistry, analysis of clast compositions and sandstone petrography, do not always provide unique criteria with which to discriminate amongst potential sources, some of which may not have been adequately characterized.
This study focuses on the northern edge of the Indus Basin, where continental deposits (the Indus Molasse) lie in direct contact with the Ladakh Batholith. Because it provides important information with which to constrain uplift history, we investigate the nature of contact between the batholith and preserved basin sediments. We also provide the first U-Pb age constraints for these deposits. Using multiple lines of evidence, we further identify source regions for sediments within the basin and constrain the timing of uplift of the Ladakh Range (NW Himalaya).

Ladakh Range and Ladakh Batholith
The Ladakh Range in the western Himalaya extends ∼NW-SE for over 250 km with peak elevations of c. 6100 m (Fig. 1). It is dominated by the Ladakh Batholith, which is part of the Trans-Himalayan magmatic belt that extends for over 2500 km along the southern margin of the Eurasian continent. The range lies between the predominantly intra-oceanic arc rocks of the Kohistan arc to the west and continental margin arc rocks of the Gangdese arc to the east. Thus it bridges an important transition along-trend from intraoceanic to continental margin subduction systems (e.g. Hébert et al. 2012). The Ladakh Batholith incorporates Cretaceous to early Cenozoic plutonic rocks intruded by dykes that are as young as c. 45 Ma (Weinberg and Dunlap 2000;White et al. 2011;Heri et al. 2015). The Khardung Volcanics are remnants of a volcanic carapace for parts of the batholith and crop out along the northern flanks of the range (e.g. Dunlap et al. 1998;Lakhan et al. 2019). Their original extent is less well known.
Driven by India-Eurasia collision, the Ladakh Range is thought to have been uplifted as early as late Eocene time. Low-temperature thermochronologic data record differences between unroofing histories along the northern and southern margins (Kirstein et al. 2006(Kirstein et al. , 2009. In general, apatite fission-track ages are older (c. 35-25 Ma) towards the southern margin of the range and younger (c. 10-5 Ma) to the north, implying later exhumation or uplift for the Ladakh Range along its northern margin (Kirstein 2011).

IYTS and Indus Basin
The Indus -Yarlung -Tsangpo suture zone (IYTS) lies at the contact between two colliding continental landmasses: India and Eurasia ( Fig. 1). In NW India, the suture follows the Indus River, immediately south of the Ladakh Range (Fig. 1). The term 'Indus Basin' collectively refers to sedimentary rocks preserved along this suture. Historically, this name has generally been applied to several lithological units, many of which are in faulted contact with one another (Searle 1983;Sinclair and Jaffey 2001;Clift et al. 2002;Henderson et al. 2010Henderson et al. , 2011 and are not necessarily genetically related. Eocene-age nummulitic limestones within the basin are regarded as the last products of marine deposition within the Tethyan Ocean (Green et al. 2008) (Fig. 1). Although most units have not been directly dated, limited biostratigraphic data are available (e.g. Bajpai et al. 2004;Green et al. 2008). The majority of age constraints are provided by detrital zircon U-Pb or mica 40 Ar/ 39 Ar dates, which only give maximum depositional ages (Clift et al. 2002;Tripathy-Lang et al. 2013;Wu et al. 2007;Henderson et al. 2010Henderson et al. , 2011. Many studies have attempted to propose a stratigraphic scheme for the Indus Basin (e.g. Brookfield and Andrews-Speed 1984;Garzanti and Van Haver 1988;Searle et al. 1990;Sinclair and Jaffrey 2001;Clift et al. 2002;Henderson et al. 2010;Tripathy-Lang et al. 2013;Baxter et al. 2016).
The term 'Indus Molasse' was first applied to coarse-grained clastic deposits that crop out within and flanking the IYTS (e.g. Fuchs 1979). This early work acknowledges the existence of several different units to which the descriptive term molasse could apply and Fuchs (1979) suggested that two main types could be recognized on the basis of their structural relationships with the Ladakh Batholith. Fuchs (1979) referred to one type as an 'autochthonous molasse', which is deposited upon an exhumed Ladakh Batholith surface. This relationship has been recognized at several locations along the southern margin of the Ladakh Range (e.g. Fuchs 1979;Brookfield and Andrews-Speed 1984;Garzanti and Van Haver 1988). Here the molasse commonly involves bright purple, red or green, and beige banded rocks that were likely derived from, and deposited on, a sloping landscape eroded on the Ladakh Range (Fuchs 1979). Another type of molasse occurs in faultbounded units and is regarded as 'parautochthonous'. Several units of this nature are recognized (Hemis Conglomerate, Choksti Conglomerate and Chogdo Formation, e.g. Fuchs 1979;Clift et al. 2002;Henderson et al. 2010). The unit on which this study focuses exhibits an unambiguously depositional relationship upon an eroded surface of Ladakh Batholith rocks.

Field mapping and sedimentary logging
Field mapping was undertaken from the Basgo to Tia areas (Fig. 2), with a focus on determining the nature of the contact between the Ladakh Batholith and sediments of the Indus Molasse. We were able to trace a nonconformable contact along strike for c. 25 km. Granitic clasts ranging from pebble to boulder sizes crop out immediately above an eroded Ladakh Batholith surface with localized palaeosol development (Supplementary material). Our observations are consistent with earlier reports by Fuchs (1979), Brookfield and Andrews-Speed (1984) and Garzanti and Van Haver (1988) suggesting that any faulted contacts (Tripathy-Lang et al. 2013) are of restricted local extent. Higher stratigraphic levels in the section (up to several hundreds of metres thick) are truncated at a regionally extensive north-directed thrust fault.
Six sections were measured along the strike of the autochthonous molasse basin, from the depositional contact (nonconformity) upon the Ladakh Batholith to the footwall of the thrust fault at which the section terminates (Fig. 1d). Four semi-continuous members could be discriminated within the preserved Indus Molasse sections (Figs 3 and 4). M1 is a crudely stratified, poorly sorted, clastsupported conglomerate with clasts ranging from boulder (c. 1 m) to pebble (several centimetres) sizes (Fig. 4). Clasts are dominantly felsic and andesitic with subordinate granitic clasts. M2 is dominated by red to grey mudstone. M3 and M4 are laterally continuous conglomeratic units. M3 is a poorly sorted, clastsupported monomictic conglomerate dominated by angular to subangular andesitic clasts ranging in size from c. 3-4 cm to 10 cm. M4 is polymictic, horizontally stratified and cross-bedded and incorporates trough cross-bedded sandstone together with well-sorted pebble layers. Field counts indicate 87% felsic and andesitic volcanic clasts in M1 (total n = 384) and 80% in M4 (total n = 181) with the remainder comprising granitic lithologies. Clasts in unit M3 are entirely of volcanic origin.

U-Pb geochronology
Detrital zircon from medium-to coarse-grained sandstones of units M1, M3 and M4 were dated by laser ablation inductively coupled plasma mass spectrometry (ICP-MS) at the School of Earth and Environmental Sciences, The University of Queensland. Zircon grains were extracted by heavy-liquid and magnetic separation, before being hand-picked under a binocular microscope for mounting in epoxy resin. Zircon grains range from c. 50 to 100 µm in width. Most grains are euhedral and few are rounded, implying a short transportation distance. Analyses of a total of 1885 detrital zircon grains with −5 to 10% discordance are reported (Fig. 5). Detailed methods are provided in the Supplementary material. All units yield ages that are dominantly <90 Ma, with peaks around 23-24, 50, 57, 62 and 78 Ma (Fig. 5). The youngest peak from a single grain in M3 and 12 grains in M4 cluster around a latest Oligocene (late Chattian), with a mean age of 23.58 ± 0.11 Ma.
Detrital apatite from M1 were also dated by laser ablation ICP-MS. Grains define an isochron on the Tera-Wasserburg plot that provides a value for ( 207 Pb/ 206 Pb) o of 0.8501 ± 0.0030. This initial common lead value was used to correct all ages (Fig. 6) with the IsoplotR program (Vermeesch 2018). Detrital apatite U-Pb ages range from c. 100 to 30 Ma, consistent with major peaks observed in detrital zircon U-Pb ages.

Zircon Hf isotopic analysis
Hafnium isotopic data for 329 zircons (231 from M1, 41 from M3 and 57 from M4) were obtained using a Nu Plasma II multicollector-inductively coupled plasma mass spectrometry (MC-ICP-MS) with laser ablation (Fig. 5) at the School of Earth and Environmental Sciences, The University of Queensland. (Analytical details are provided in the data repository). ε Hf (t) values were calculated with 206 Pb/ 238 U ages and range from −10 to 17. Grains older than 45 Ma are characterized by dominantly positive ε Hf (t) values, clustering around 5-15. All ε Hf (t) values for grains within the youngest peak of c. 24-23 Ma are negative: −2.4 to −9.4 (Fig. 5c).

Discussion
Our study indicates a latest Oligocene age for initial sedimentation along the northern margin of the Indus Molasse. Detrital zircon U-Pb ages reveal a younger c. 24-23 Ma age peak for M4 and provide a maximum depositional age for this unit. This radiometric age is consistent with palaeontological (ostracod) data reported by Bajpai et al. (2004) from correlative sediments near Basgo (Fig. 2). Sediments in the studied area were entirely sourced from the north with no evidence for any southerly provenance, from the Indian Plate, for example, which contains older zircon with ages >500 Ma (e.g. Gehrels et al. 2011). Granitic clasts and sandstone with detrital zircon ages of c. 85-45 Ma are consistent with uplift and erosion of the Ladakh Batholith, which served as an important source (Brookfield and Andrews-Speed 1984;Searle 1986;Garzanti and Van Haver 1988). Abundant andesitic and felsic volcanic clasts (comprising 100% of clasts in M3 and c. 85% of clasts in M1 and M4) strongly suggest a volcanic source, which we consider most likely to be a local equivalent of the Khardung Volcanics. Such rocks must first have been eroded to expose the Ladakh Batholith above and upon which they were erupted. Documented ages for these volcanics include 51.95 ± 0.4 and 56.4 ± 0.4 Ma (whole-rock Ar/ 39 Ar plateau ages, Bhutani et al. 2009), and 60.5 ± 1.3 and 67.4 ± 1.1 Ma (zircon U-Pb sensitive high-resolution ion microprobe ages, Dunlap and Wysoczanski 2002) and are consistent with the dominant peaks identified in our zircon age data (Fig. 5). The source of the youngest age peak (c. 24-23 Ma) remains less definitive. The Karakorum Batholith potentially represents a source of younger zircon grains c. 40-20 Ma (e.g. White et al. 2012;Borneman et al. 2015) and is characterized by negative zircon ε Hf (t) values as low as −10.5 (Ravikant et al. 2009), consistent with our results. Dykes that intrude the Oligocene Saltoro Molasse have been reported from the Shyok suture to the north (Rai 1983). Oligocene-Miocene dykes have also been reported from within the Ladakh Batholith with a Rb-Sr isochron age of c. 24 Ma obtained by Ravikant (2006). However, we note that no existing studies from the Ladakh Batholith report zircon U-Pb ages younger than c. 45 Ma (Weinberg and Dunlap 2000;White et al. 2011;Heri et al. 2015).
Our work refines and builds upon earlier investigations. Garzanti and van Haver (1998) reported a Maastrichtian age for the Basgo Formation based on identifications of an ostracod fauna. However, in a more recent investigation, Bajpai et al. (2004) also reported ostracods that suggest an Oligocene age, which is consistent with our findings. Cenozoic detrital zircon ages further refine this age to the latest Oligocene. Tripathy-Lang et al. (2013) reported somewhat different detrital zircon results from sites close to the Ladakh Batholith (Figs 2 and 5d), which the authors considered to be representative of the Indus Molasse (Basgo Formation). However, our mapping suggests that the samples in question were collected from the hanging wall of a north-directed thrust fault and as such may not be related to the Basgo Formation. Moreover, the samples reported by Tripathy-Lang et al. (2013) did not contain any grains younger than 90 Ma, seemingly excluding the Ladakh Batholith as a potential source (e.g. Gehrels et al. 2011).
As sediments documented here overlie a surface eroded into the Ladakh Batholith, they place a minimum constraint on the time needed to erode the former southern margin of Eurasia. The removal of a thick crustal section (several kilometres or more) must have preceded deposition of the molasse, earlier than c. 24-23 Ma. This is consistent with thermochronological modelling from nearby locations where cooling c. 30-20 Ma is documented (Kirstein et al. 2006(Kirstein et al. , 2009. The ages we obtained from initial sediments of the Indus Molasse are consistent with those from elsewhere along the length of the IYTS (Gangrinboche facies across Tibet; Aitchison et al. 2002). In southern Tibet, Lower Miocene Gangrinboche conglomerates (including the Kailas, Qiuwu, Dazhuqu and Luobusa formations) crop out along the Yarlung Tsangpo suture zone and lie in depositional contact upon rocks of the Lhasa terrane, mostly the Gangdese Batholith (Atichison et al. 2002). This observation calls for the exhumation of the Gangdese Batholith around the same time as, or prior to, the early Miocene, consistent with previous investigations that infer Late Oligocene regional uplift of the Lhasa terrane (Harrison et al. 1992). Although the deposition of such a molasse has traditionally been regarded as a key constraint on the timing of the India-Eurasia collision, our data-which indicate that sedimentation began in the latest Oligocene-require an alternative explanation, as this event significantly post-dates even the youngest estimates for the initiation of continental collision (Klootwijk 1984;Garzanti and Van Haver 1988;Najman et al. 2001;Clift et al. 2002;Aitchison et al. 2007). The timing of the exhumation of the Ladakh Batholith and deposition of the Indus Molasse may be explained by the slab break-off event proposed c. 25 Ma (Maheo et al. 2002). In the Karakoram Batholith, igneous and metamorphic rocks associated with a high thermal gradient, enrichment of large-ion lithophile elements and depletion of highfield-strength elements have been documented and interpreted as products associated with the melting of metasomatized Asian lithospheric mantle (Maheo et al. 2002). Late Oligocene exhumation of the batholiths observed in southern Tibet (Gangdese batholith) and NW India (Ladakh Batholith) is broadly consistent with alternative mechanisms to the onset of continent-continent collision that have been proposed to explain exhumation along the IYTS. These include slab break-off (Maheo et al. 2002;Aitchison et al. 2009) or foundering of a continental root Lee et al. 2009).

Conclusions
Initial continental sediments deposited directly upon an uplifted and eroded surface that developed on the Ladakh Batholith on the southern margin of the Eurasian continent are no older than latest Oligocene, with a possible depositional age of c. 24-23 Ma. Deposits were derived entirely from the north, with the Ladakh arc and Karakoram terrane being the most likely sources. The deposition of the autochthonous molasse of the Indus Basin is analogous to the formation of the Gangrinboche facies in southern Tibet (Aitchison et al. 2002(Aitchison et al. , 2009).  (Vermeesch 2018) assuming all apatite grains have a similar common lead composition. The calculated common lead composition (0.8501 ± 0.0030) was then applied to correct all analyses to generate common-Pb-corrected ages (b). All errors are shown at the 2σ level.