(U-Th)/He Thermochronology of the Indus Group, Ladakh, Northwest India: Is Neogene Cooling a Continental-Scale Thermal Event in the India-Asia Collision Zone?

The India-Asia continental collision zone archives a sedimentary record of the tectonic, geodynamic, and erosional processes that control the thermal history of the Himalayan orogenic interior since the onset of collision in early Paleogene time. In this paper, we present new (U-Th)/He thermochronometric cooling age data from 18 detrital zircons (ZHe) and 19 detrital apatites (AHe) of the early Eocene – early Miocene (ca. 50 – 23 Ma) continental facies of the Indus Group along the India-Asia collision zone in Ladakh, northwest (NW) India. This along-strike regional-scale low-temperature thermochronometric data set from the Indus basin is the ﬁ rst report of ZHe and AHe cooling ages from western and eastern Ladakh. Thermal modeling of our ZHe and AHe cooling ages indicates a postdepositional Neogene cooling signal in the Indus Group. Cooling initiated at ca. 21 – 19Ma, was operational along the ~ 300km strike of the collision zone in NW India by ca. 11Ma, and continued until ca. 3 Ma. The Miocene cooling signal, also present along the India-Asia collision zone in south Tibet, is a continental-scale cooling event likely linked to increased erosional e ﬃ ciency by the Indus and Yarlung Rivers across an elevated region resulting from the subduction dynamics of the underthrusting Indian plate.


Introduction
The India-Asia collision zone records the largest intercontinental collision of Earth's Cenozoic history and continues to provide new geological and geophysical data that improve our understanding of evolution of the Himalayan orogenic system. Drained by the Indus and Yarlung Rivers and containing the Indus-Yarlung suture, the >2000 km long India-Asia collision zone is a natural laboratory for studying early orogenic and subsequent tectonothermal processes that ensued in the orogen interior (e.g., [1][2][3]). Within the India-Asia collision zone in Ladakh, NW India, the Indus Group (also called the Indus Molasse) represents the continental sequence of a larger suite of sedimentary rocks parallel to the Indus suture called the Indus basin sedimentary rocks (Figures 1(a)-1(c)). The early Eocene-early Miocene Indus Group was deposited in an intermontane basin along the Indus suture after the onset of India-Asia collision and is currently at elevations of >4-5 km along with rest of the Indus basin sedimentary rocks [4][5][6][7][8][9][10][11]. Knowing the timing of burial, exhumation, and uplift of the Indus Group is critical in determining the tectonothermal, geodynamic, and erosional processes that governed regional exhumation and uplift of the India-Asia collision zone in NW India after the onset of collision.
Previous low-temperature thermochronometric studies in the Indus Group as well as the overall Indus basin are restricted mostly to central Ladakh. Low-temperature thermochronometry in the central part of the Indus basin, using zircon fission-track (ZFT), apatite fission-track (AFT), zircon (U-Th/He) (ZHe), and apatite (U-Th)/He (AHe) data, indicates burial to >200°C and unveils a postdepositional Miocene-Pliocene cooling signal from ca. 22 to 4 Ma [6,12]. AFT ages from two samples in central Ladakh and one sample in eastern Ladakh also record postdepositional Miocene cooling in the Indus Group [6,13]. However, Tripathy-Lang et al. [10] obtained ca. 52-28 Ma predepositional unreset ZHe ages from the late Oligocene Basgo Formation of the Indus Group in central Ladakh. This finding raises a possibility that the cooling ages from the Indus basin may vary locally alongstrike. Along the India-Asia collision zone in NW India (Figure 1), cooling ages from the Ladakh batholith vary along-strike: the western parts of the Ladakh batholith reveal ca. 46-41 Ma ZFT cooling ages [14], but ZHe, AFT, and AHe ages from the central part of the Ladakh batholith show discontinuous Eocene-Miocene cooling pulses from ca. 35 to 12 Ma [15,16].
Low-temperature thermochronometric analyses using ZHe, AFT, and AHe age data reveal a Miocene cooling signal at multiple locations along the India-Asia collision zone in south Tibet between Kailas and Zedang (Figure 1(b)).
Therefore, to determine whether the Miocene cooling signal is also present along the strike of the collision zone in NW India, we analyzed and thermally modeled a total of 37 new single-grain detrital ZHe and AHe ages from four samples of the Indus Group in western, central, and eastern Ladakh (Figures 2-4). In addition, we compile 244 published AHe, AFT, and ZHe mean cooling ages (Figures 5(a) Figure 1: (a) Regional geologic map of the India-Asia collision zone in Ladakh, NW India, modified after Buchs and Epard [65]. Primary tectonostratigraphic units and major faults are shown here. Samples are indicated with a number and a DZ-prefix at their corresponding locations. (b) Location of the study area (red hollow box) with respect to the overall length of the India-Asia collision zone along the Indus and Yarlung sutures of south Asia. KF: Karakoram fault. (c) Schematic simplified cross-section AA' through the India-Asia collision zone in NW India showing the Indus Group sample locations. GCT: Great Counter thrust. 2 Lithosphere and compare these to regional geologic observations to determine if Miocene cooling was a continental-scale thermal event along >1600 km of the India-Asia collision zone between Lehdo, western Ladakh, and Zedang, southeast Tibet ( Figure 1(b)).

Geologic Background
A critical element of the India-Asia collision zone in Ladakh, NW India, is the Indus Group that is a part of the Indus basin sedimentary rocks. These sedimentary rocks unconformably overlie the Cretaceous-Paleogene Ladakh batholith to the north and truncate against a complex assemblage of Indus suture units to the south (Figure 1(a)). The Indus suture units consist of the Cretaceous Dras-Nidar Complexes, containing ophiolitic mélange, volcanic and volcano-sedimentary rocks [28][29][30], and the Triassic-Cretaceous Lamayuru Complex, containing the continental slope facies of the northern Indian passive margin [31]. South of the Indus suture lies the Tethyan Himalaya, which is composed of a metasedimentary Neoproterozoic-early Cenozoic continental shelf succession of the northern Indian passive margin [32]. A klippe preserves the Mesozoic Spongtang oceanic arc within the Tethyan Himalaya ( Figure 1; [33]). In the southeast, the subducted-extruded margin of the Indian continental plate is preserved in the Tso Morari Complex, which is an ultrahigh-pressure metamorphic unit [34]. A series of strikeparallel north-verging thrusts, linked to the Great Counter thrust (GCT) system, deformed the collision zone between ca. 23 and 20 Ma causing inversion of the Indus basin ( Figure 1; [6,9]). The Indus basin sedimentary rocks consist of two rocks groups: the marine Tar Group, deposited in a forearc setting between Late Cretaceous and early Eocene time (Figure 2( Figure 2: Regional Indus Group stratigraphy in Ladakh, NW India (modified after Sinclair and Jaffey [50]), showing sampled formations, and ZHe-AHe thermochronology results reported in this study. Biostratigraphic ages of the marine strata and the Basgo Formation are from Green et al. [35] and Bajpai et al. [4]. U-Pb detrital zircon maximum depositional ages, calculated using YC2σ(3+) method, are from Bhattacharya et al. [5]. n is the number of zircon or apatite grains dated from a specific formation. Fm: formation. [7,35]; and the continental Indus Group, deposited after the onset of India-Asia collision in an intermontane basin between early Eocene and early Miocene time (ca. 50-23 Ma; Figure 2(b); [4][5][6][7][8]11]). Stratigraphically, the Indus Group consists of six regional-scale siliciclastic formations-the oldest Nurla Formation, the Choksti, Hemis, Nimu, and Basgo Formations, and the youngest Temesgam Formation ( Figure 2). Deposition of continental sedimentary rocks also occurred locally in the Ladakh region in Pliocene time [7,36].
Discrepancies exist concerning the timing of sedimentary rock exhumation among the few previous low-temperature detrital thermochronometric studies from the Indus basin in NW India that are limited mostly to central Ladakh. The cooling age data from the same rock group or formations using the same low-temperature thermochronometric techniques are significantly different. Tripathy-Lang et al. [10] determined ca. 52-28 Ma unreset predepositional ZHe ages from the late Oligocene Basgo Formation in central Ladakh and suggest that the ages reflect cooling of Indian plate    7) Laskowski et al. [18], (8) Dèzes et al. [61], and (9) Shen et al. [24].   Table S5 for sources of ages corresponding to GPS locations). 6 Lithosphere

Sample Collection and Methods
The primary technique used is (U-Th)/He detrital ZHe and AHe thermochronology. Sampling the Indus Group for detrital ZHe and AHe dating is challenging. Multiple sections preserve the basal, middle, and top parts of the Indus Group in Ladakh and a single section across the composite Indus Group stratigraphy is not accessible. We collected 4 detrital samples (DZ-prefix), one from each of the four key regional-scale Indus Group units: sample DZT2MH78 from the Nurla Formation (base) in eastern Ladakh, sample DZAR2HMS77 from the Hemis Formation (middle) in central Ladakh, and samples DZTSKG02 and DZASKR16 from the Basgo (middle) and Temesgam (top) Formations, respectively, in western Ladakh (Figure 2). DZT2MH78 and DZASKR16 are fine-grained sandstone samples, while DZAR2HMS77 and DZTSKG02 are very coarse-grained sandstone samples that were interbedded with conglomerates.
3.1. (U-Th)/He Thermochronology. Zircons and apatites were separated using standard techniques of crushing, grinding, magnetic, and heavy liquid separation. Four to six dateable quality sharp-faced euhedral zircon and apatite crystals with few or no inclusions were hand-picked from each sample and analyzed at the Arizona Radiogenic Helium Dating Laboratory following the procedures of Reiners et al. [40]. A total of 18 zircons and 19 apatites were dated. Selected zircon and apatite grains were packed in Nb tubes and subjected to standard procedures of He extraction analysis using a quadrupole mass spectrometer. Th and U contents were measured using ICP-MS following the methods in Reiners et al. [40]. Subsequently, raw ages were calculated by incorporating known analytical concentrations of U, Th, and He in the combined decay-diffusion equation and corrected by applying the alpha-ejection protocols [41,42]. If resulting ZHe or AHe cooling ages are older than the depositional age of the formation, then these cooling ages are thermally unreset after deposition and reflect predepositional cooling in the source regions. In contrast, ZHe and AHe cooling ages that are younger than the depositional age of the formation suggest burial temperatures in the basin exceeded 140-200°C and 40-90°C, respectively, and indicate that these ages are therefore thermally reset after deposition. However, the (U-Th)/He system is sensitive to radiation damage, crystal anisotropy, grain size, annealing mechanism, and uncertainties in burial conditions, thereby reducing the precision associated with inferred thermal histories from ZHe or AHe cooling ages [43,44]. We analyzed the variation of ZHe and AHe grain ages with effective uranium and grain size to identify if radiation damage and/or crystal size have a control on the observed distribution of cooling ages.

Thermal
Modeling. Using the thermal modeling program HeFTy v.1.9.1 [45], the ZHe and AHe ages from each sample were inverse modeled to determine the time-temperature (t-T) history of the corresponding sample. For a given input data set that include uncorrected cooling ages, U-Th-Sm concentrations, grain size, and a given diffusion model (e.g., [44,46]), the inverse algorithm extracted a family of t-T paths that a sample could have possibly experienced under the user assigned t-T constraints. Based on the calculated statistical fit between the measured and predicted cooling ages, the t-T paths generated for a particular sample can be grouped into two sets-the acceptable-fit paths that have a Kolmogorov-Smirnov probability ≥0.05 and the good-fit paths, which have a Kolmogorov-Smirnov probability ≥0.5. The t-T path with the highest goodness of fit is the best-fit path. The best-fit path is sensitive to the number of paths tested but reasonably approximates a meaningful thermal history experienced by a particular sample when a sufficiently high number of paths are tested.
Applying the inverse algorithm of the thermal modeling program HeFTy, we modeled the t-T histories of the four samples using the ZHe and AHe cooling age data and published regional geologic t-T constraints. We used depositional age constraints as discussed in "Constraining Depositional Ages" and considered a surface depositional temperature of 0-25°C for all samples. No evidence of cooling exists in the Indus Group before 23 Ma, when overthrusting by the GCT inverted the Indus basin [6]. Recent ZFT age analyses from the Indus Group suggest that basin temperatures exceeded the lower partial annealing zone temperatures of 185-200°C but stayed below higher partial annealing zone temperatures of 280-300°C [5]. An illite crystallinity estimate by Clift et al. [39] also indicates a basin temperature of~239°C from the Indus Group. Thus, based on regional geologic constraints, we allowed each model to explore the t-T space younger than 23 Ma and colder than 240°C. Considering a 240°C temperature limit also permits the model to extract t-T histories from temperatures greater than the closure temperature range of the warmest thermochronometer in the study, i.e., ZHe system, thereby relaxing the t-T space being investigated. We used mean ZHe and AHe ages in modeling the t-T histories of the samples using the diffusion models of Guenthner et al. [44] and Farley [46]. Using the same t-T constraints, we also modeled the single-grain ZHe and AHe ages from each sample, as well as all the single-grain ZHe ages individually, to test if the resultant models exhibit thermal histories that are significantly different from those obtained by using mean ZHe and AHe ages. We ran all models for >800,000 paths and until at least 100 good-fit paths were obtained and considered the best-fit path that resulted at the end of each model run to reasonably approximate a robust t-T history under the applied geologic constraints.

Constraining Depositional Ages.
Limited fossil data from the Indus Group formations make it difficult to precisely estimate their true depositional ages, which are critical initial constraints in thermal modeling. We use the depositional ages of the studied Indus Group formations from published biostratigraphic ages [7,35], detrital zircon U-Pb maximum depositional ages [5,7,11] and ages of tectonic events that paused deposition [6]. The depositional age of the Nurla Formation, from which sample DZT2MH78 is collected, is ca. 50-41 Ma. The lower age limit of 50 Ma is constrained by the ca. 50 Ma U-Pb detrital zircon maximum depositional age of the Nurla Formation as well as the fossils of same 7 Lithosphere age from the marine sequence immediately underlying the Nurla Formation [5,35]. The upper age limit of 41 Ma is imposed from the ca. 46-41 Ma U-Pb detrital zircon maximum depositional age of the overlying Choksti Formation [7,11]. The 41 Ma upper age limit for the Nurla Formation also coincides with an unconformity confirming termination of Nurla deposition [47]. The depositional age of the Hemis Formation, from which sample DZAR2HMS77 is collected, is constrained to ca. 37-28 Ma. The lower age limit of 37 Ma comes from the detrital zircon U-Pb maximum depositional age of the Hemis Formation [5]. The upper age limit of 28 Ma comes from the late Oligocene biostratigraphic age of the Basgo Formation [4], which is younger than the Hemis Formation. It is possible that Hemis deposition terminated earlier; however, the lack of fossils in the Hemis Formation or in the overlying Nimu Formation (ca. 32 Ma detrital zircon U-Pb maximum depositional age [7,12] makes it challenging to precisely constrain the upper depositional age limit of the Hemis Formation. The depositional age of the Basgo Formation, which contains late Oligocene ostracod fossils [4] and has a youngest single-grain detrital zircon U-Pb maximum depositional age of ca. 27 Ma [5], is constrained to ca. 28-26 Ma. Stratigraphically, the Basgo Formation is~10-200 m thick [48] and is conformably overlain by the Temesgam Formation that has a detrital zircon U-Pb maximum depositional age of ca. 26 Ma [5]. The depositional age of the Temesgam Formation is therefore constrained to ca. 26-23 Ma, with the upper age limit of ca. 23 Ma from the estimated age of counterthrusting by the GCT that ended regional Indus Group deposition [6].
Within individual samples, correlations between grain size and ZHe or AHe ages are either very weak or nonexistent. However, considering the 18 ZHe ages together, a moderately positive correlation between grain size and age accounts for the ZHe age dispersion in the samples ( Figure S3). Likewise, the 19 AHe ages together show a weak positive correlation with grain size, which possibly explains the intrasample AHe age dispersion ( Figure S3). No meaningful correlations exists between effective uranium and grain ages in individual samples or collectively for either thermochronometric system, suggesting radiation damage is not the primary control on observed intrasample ZHe and AHe age variability. Tables S1 and S2 and Figure S3 contain all analytical data, and cooling age versus grain size and effective uranium plots.

Thermal Modeling
Modeling individual ZHe and AHe ages for a sample, i.e., 8-10 single-grain ages together in HeFTy resulted in no acceptable or good-fit paths. That HeFTy fails to yield meaningful thermal histories by simultaneously solving all input data from multiple grain ages of the same sample is known (e.g., [17]). The t-T modeling results using mean ZHe and AHe ages are in Figures 3(a)-3(d), while the t-T modeling results from the 18 individual ZHe ages are in Figure S4. Because these two sets of t-T modeling results do not differ significantly, we continue the discussion with reference to the t-T models using mean ages from here on.
Thermal modeling of ZHe and AHe ages confirms thermal resetting of both thermochronometric systems and exhibits best-fit t-T paths that suggest initiation of postdepositional cooling between ca. 21 and 11 Ma in all samples (Figures 3(a)-3(d)). Cooling begins in the Hemis and Basgo Formations at ca. 21-19 Ma, while cooling begins later in the Nurla and Temesgam Formations at ca. 15-11 Ma. As shown by several acceptable and good-fit t-T paths in each model, cooling may have initiated earlier than the time indicated by the best-fit t-T path in the corresponding model. In this study, we consider the timing of beginning of cooling indicated by the best-fit t-T path as the minimum time by which cooling began in each model. The best-fit t-T paths in Nurla and Hemis Formations indicate cooling beginning from peak basin temperatures of~225°C (Figures 3(a) and  3(b)). The best-fit t-T paths in the shallower Basgo and Temesgam Formations indicate peak basin temperatures of 150 and 180°C, respectively (Figures 3(c) and 3(d)). The relatively low peak basin temperatures in the Basgo and Temesgam Formations are likely a consequence of modeling. Because the younger Temesgam Formation was at a higher elevation than the older Basgo Formation, the Temesgam Formation is expected to show cooler maximum basin temperatures. Nevertheless, modeling of the individual ZHe ages from the Basgo and Temesgam Formations indicate peak basin temperatures exceeding 180-200°C ( Figure S4).
The acceptable and good-fit t-T paths appear to pass through either higher temperatures (>180-200°C) early or through lower temperatures (<180-200°C) later resulting in 8 Lithosphere two distinct clusters of t-T solutions in each model. From a strict modeling point of view, these two distinct clusters of solutions indicate that either can satisfactorily explain the two input data points, i.e., a mean ZHe age and a mean AHe age, and constrain the timing of initiation of cooling at any time between ca. 23 and 5 Ma. The older higher temperature cluster predicts the mean ZHe age, while the younger lower temperature cluster predicts the mean AHe age. The two clusters are essentially part of a single cooling event that initiated from temperatures >180-200°C, which can be independently verified from regional paleo-temperature estimates using illite crystallinity and ZFT analyses (e.g., [12,39]). The clusters are likely a consequence of averaging single-grain ZHe or AHe ages, and this problem may be avoided if more ZHe and AHe individual grain ages from a single sample are available to analyze the distribution of ages and predict each age using the method of Fox et al. [49]. However, acquiring good-quality dateable zircon or apatite grains was challenging, and given the limited ZHe and/or AHe ages in our individual samples, we use available data from regional thermal proxies and thermochronometric studies to validate the results of our t-T modeling.

Discussion
The new ZHe and AHe cooling ages together with the thermal modeling reveal a postdepositional Miocene-Pliocene cooling phase initiated in the Indus Group by ca. 21-19 Ma and by 11 Ma was in operation along the~300 km strike of the India-Asia collision zone from western to eastern Ladakh in NW India. The t-T models also confirm that maximum burial temperatures in the Indus Group were close to or exceeded the upper limit of the ZHe partial retention temperatures, i.e., 180-200°C. This temperature estimate is consistent with the maximum burial temperature of~239°C determined by Clift et al. [39] using illite crystallinity. Our study expands the limited regional low-temperature thermochronometric data set and shows that the postdepositional Miocene cooling signal in the Indus basin is present from western to eastern Ladakh. This implies that the basin did not undergo differential burial and exhumation histories along the length of the India-Asia collision zone in NW India. The new ca. 20-3 Ma ZHe and AHe ages from western to eastern Ladakh overlap with the published ca. 19-8 Ma ZHe, 14-12 Ma AFT, and 7-3 Ma AHe ages from the Indus Group that infer a postdepositional Miocene-Pliocene cooling signal in central Ladakh [6,12]. Our thermal models and the ca. 10-3 Ma AHe ages also indicate that the cooling phase continued into Pliocene time at least until ca. 3 Ma. It is difficult to explain why the ZHe cooling ages from two different locations of the Basgo Formation in this study and in the work by Bhattacharya et al. [12] differ from those determined by Tripathy-Lang et al. [10] from the same formation. Nevertheless, our ZHe cooling ages from the Basgo Formation are compatible with the overall trend of the ZHe cooling ages determined from the Indus Group. Across Ladakh, the late Oligocene-early Miocene Temesgam Formation is the youngest Indus Group stratigraphy on a regional scale and has postdepositional ZHe and AHe cool-ing ages like the deeper and older stratigraphic units. Localized Pliocene deposition [7], restricted to the Zanskar Gorge area of central Ladakh, is <1 km in stratigraphic thickness and cannot explain the postdepositional cooling ages in the Temesgam Formation. Therefore, sedimentation alone cannot account for postdepositional burial temperatures >180-200°C in the~4.5 km thick Indus Group. We suggest that the Indus basin sedimentary rocks, which contains the Indus Group, were buried by the GCT hanging wall rocks between ca. 23 and 20 Ma [6,12,50]. Although thermochronometric data do not directly constrain surface uplift, the timing of Miocene cooling in the Indus Group coincides with key geodynamic and erosional processes operating in the India-Asia collision zone at that time that are linked to regional surface uplift and denudation (Figure 4). The Indus basin was affected by a combination of contemporaneous geodynamic (e.g., Indian slab dynamics), tectonic (e.g., slip along the GCT), and erosional processes (e.g., Indus River incision) in Miocene time in NW India. We suggest that, following Greater India slab rollback and breakoff, the return of northward underthrusting of the Indian plate in early Miocene time provided the mechanism for regional surface uplift of the overall India-Asia collision zone [24,27,[51][52][53], including the GCT-overthrusted Indus basin sedimentary rocks. This surface uplift was responsible for vertical elevation gains across the basin. The present-day elevations of >4-5 km of the Indus basin are therefore a direct consequence of the northward underthrusting of the Indian plate that uplifted the collision zone and crustal shortening along the GCT, although the relative contributions to elevation gains by these two processes in NW India are not yet determined. The Indus River flowed through the elevated region in Neogene time eroding the overthrusted Indus basin rocks and producing the observed Miocene cooling signal. Considering cooling from a mean maximum burial temperature of 200°C and assuming a geothermal gradient of 20-30°C/km that is consistent with the regional Miocene geothermal gradients estimated from neighboring lithotectonic units [54][55][56], we infer a removal of~7-10 km of rock from the Indus Group since the onset of Miocene cooling. Chemical weathering and alteration indices and sedimentation rates determined from Indus fan sediments show that erosion peaked in the Himalaya at ca. 15-10 Ma transporting rocks from the orogenic front to the Indus fan [57]. Our rock erosional estimate suggests that, during Miocene time, the orogenic hinterland contributed a significant amount of rock material to the Indus fan as well. The caveat here is that the Indus River might not have been the sole driver of regional erosion. The Asian monsoon, which intensified during Miocene time in the orogenic hinterland [57], may have played a role in erosion and subsequent cooling as well. However, more studies are required in the region to investigate the interplay of multiple drivers of erosion.
Compared to NW India, low-temperature thermochronometric studies in the India-Asia collision zone of south Tibet are abundant and unveil the regional thermal history. ZHe cooling ages from the late Oligocene-early Miocene Kailas Formation in south Tibet, which is contemporaneous with the upper formations of the Indus Group in NW India, 9 Lithosphere reveal basin exhumation between ca. 21 and 15 Ma [17]. AFT and AHe cooling ages indicate that the Cenozoic Liuqu Formation experienced cooling after ca. 12-10 Ma [22]. Thermal modeling of ZHe and AHe ages, and AFT age data show that the Cretaceous-early Eocene Xigaze forearc cooled in Miocene time from ca. 21-7 Ma [20,23]. ZHe, AFT, and AHe thermochronology of the adjacent southern Gangdese batholith and the northern Tethyan Himalaya in south Tibet also depict cooling between ca. 20 and 6 Ma [18,19,21,26,58]. The Miocene cooling signal along the India-Asia collision zone in south Tibet is attributed to regional dynamic surface uplift that accompanied northward underthrusting of the Indian plate after Greater Indian slab rollback and breakoff [24,27], Asian monsoon (e.g., [17]), movement on the GCT and/or Yarlung River incision (e.g., [17,19,20,22,23,26,58]).
Comparison of our thermochronometric data with a compilation of 244 published mean ZHe, AFT, and AHe cooling ages from NW India and south Tibet reveals that 75% of the ages are Miocene in age (Figures 5(a)-5(c)). The postdepositional ages of the sedimentary units along the India-Asia collision zone-the Indus, Kailas, Liuqu, and Xigaze basins-and the overall India-Asia collision zone exhibit similar cooling age distribution trends with a peak Miocene cooling age of ca. 17 Ma (Figures 5(b) and 5(c)). Although cooling ages may vary with grain size, radiation damage, presence of inclusions, and overall structural position, the ages in this compilation reflect erosional cooling through closure temperatures of ZHe (140-200°C, [44]) and AHe (40-90°C, Farley et al., 2000) corresponding tõ 2-10 km depths, assuming 20-30°C/km geotherms. The combined data set of low-temperature thermochronometric ages suggests a continental-scale Miocene erosional cooling event along >1600 km of the India-Asia collision zone between Lehdo and Zedang (Figure 1(b)). Cooling may have initiated in Paleogene time in the nonsedimentary units (e.g., Ladakh-Gangdese batholiths, Tethyan Himalaya); however, Neogene time is when all tectonic units of the India-Asia collision zone were exhuming (Figures 5(b) and 5(c)). Given the timing of key geodynamic, tectonic, and erosional processes that coincide with the timing of Miocene cooling in NW India ( Figure 4) and south Tibet, we hypothesize that a combination of these processes played a role in the observed Miocene cooling signal, i.e., the regional Miocene geology supports the thermochronology. In other words, the set of processes controlling cooling in Paleogene time (e.g., [55,59,60]) was different from those controlling cooling in Neogene time. The onset of Miocene cooling on a continental scale is coeval with the regional dynamic surface uplift and a return of northward underthrusting of the Indian plate following Greater India slab breakoff, and both these processes are well documented in south Tibet [24,27]. Using ZFT-AFT thermochronology and topography modeling, Shen et al. [24] interpret that the Miocene cooling signal along the Yarlung suture in the India-Asia collision zone of south Tibet is a consequence of dynamic topography, by which a southerly migrating wave of topographic uplifts and subsidence reflect the surface response to sublithospheric mantle flow perturbations. If the dynamic topography hypothesis is applicable to NW India, this suggests that Miocene cooling occurred regionally across the Indus and Yarlung sutures along the India-Asia collision zone in response to dynamic topographic uplift associated with the lithospheric slab dynamics of the subducting Indian plate. The onset of Miocene cooling in NW India is also coeval with the~22-19 Ma timing of normal faulting along the regional South Tibetan Detachment south of the study area (Figure 1(a); [61]) or immediately precedes regional orogen-parallel E-W extension, which initiated at 16 Ma along the India-Asia collision zone in NW India and south Tibet [18,62,63]. However, low-temperature thermochronometric data from NW India are limited, and additional data are needed to test the dynamic topography hypothesis, or any potential cause-effect relationship between the onset of Miocene cooling and orogen-parallel extension or normal faulting along the South Tibetan Detachment. Nevertheless, the role of coeval lithospheric slab dynamics in NW India in relation to the regional surface uplift and observed Miocene cooling signal cannot be ruled out, although its exact nature remains poorly understood. Therefore, we suggest that the Miocene cooling along the India-Asia collision zone was exacerbated by downcutting erosion by the Indus and Yarlung Rivers through an elevated region produced by the northward underthrusting of the Indian slab after Greater India slab breakoff. Basins along the India-Asia collision zone were buried either through sedimentation and/or by the GCT system (e.g., [17,19,23]; this study) depending on local structural relationships. Interestingly, this pronounced Neogene cooling event along the India-Asia collision zone is contemporaneous with cooling to the south in Greater Himalayan rocks (e.g., [64]), and whether these two signals might possibly be related to large-scale evolution of the orogenic wedge remains a topic for future research.

Conclusions
Our new ZHe-AHe ages and thermal models from the Indus Group indicate a regional-scale Miocene-Pliocene cooling phase along the~300 km strike of the India-Asia collision zone in NW India. Cooling initiated in the Indus Group in Miocene time at ca. 21-19 Ma following burial of the Indus basin by the GCT to >180-200°C and continued at least until ca. 3 Ma. The Miocene cooling signal in NW India is part of a continental-scale Miocene cooling signal observed along >1600 km of the India-Asia collision zone from Lehdo, NW India, to Zedang, south Tibet. Subduction dynamics, specifically a return to northward underthrusting of the Indian plate, coupled with erosion by the Indus and Yarlung Rivers in NW India and south Tibet, likely drove Neogene cooling across the India-Asia collision zone.

Data Availability
(U-Th)/He thermochronometric data presented in this paper are available in the supplementary file. 10 Lithosphere  Figure S3: plot of (U-Th)/He ZHe-AHe cooling ages vs effective uranium and grain size. Figure S5: time-temperature modeling of individual (U-Th)/He zircon ages using HeFTy.