Zircon U‐Pb‐He Double Dating of Modern Sands From the Inn River Catchment: Assessing Resolution and Potential in a Complex Orogenic Setting

Zircon U‐Pb‐He double dating (ZDD) provides the opportunity to derive high temperature crystallization ages and low temperature cooling ages from the very same mineral grain, making it especially attractive for zircon provenance studies. We present the combination of in situ detrital zircon U‐Pb‐He double dating and Raman spectroscopy‐based heavy mineral analysis on different tributaries of the Inn river system applying a new grain embedding technique. The Inn river catchment drains a well‐studied complex nappe pile of the Eastern Alps exposing Austroalpine and Penninic crystalline basement rocks. Igneous formations of different emplacement ages, as well as various metasedimentary units, have experienced Alpine and diverse pre‐Alpine metamorphic overprints. This makes the area ideally suited for testing the sensitivity and potential of ZDD for fingerprinting various lithologically contrasting units with contrasting thermal histories. Results demonstrate that both high‐ and low‐temperature age distributions reflect the major sources and thermotectonic pulses, respectively. More specifically, the rocks of the Tauern Window with Miocene (U‐Th)/He ages as well as the Permian metagranitoids from the Tauern and the Err‐Bernina nappe in the uppermost Inn valley are recorded in the downstream Inn sample. The main mass of zircons in the lower Inn, however, derives from the Ötztal and Silvretta crystalline basement rock with Cadomian U‐Pb and Cretaceous (U‐Th)/He ages. The abundance of heavy minerals derived from metamafic formations shows no correlation to the area of such lithologies on the catchments and the different zircon fertility resulted in the overrepresentation of zircon U‐Pb ages from the Variscan igneous suites.

. In this way, the cooling in mid-crustal levels can be dated that is especially relevant in case of orogenic belts having no or subordinate syn-collisional magmatic activity like the Alps (e.g., Mark et al., 2016).
Among the low-temperature geochronometers, fission-track dating has the longest history in application to detrital grains, and is widely used nowadays (e.g., Brandon & Vance, 1992;Carter, 2007;Dunkl et al., 2009;Hurford et al., 1984;Malusà & Fitzgerald, 2020).Apatite and zircon fission-track (AFT, ZFT) age patterns allow identifying and distinguishing the major cooling/exhumation events in the source area of the sediment (e.g., Braun et al., 2006;Malusà, 2019).The apatite and zircon (U-Th)/He thermochronometers (AHe, ZHe) have even lower closure temperatures than the respective fission-track method.This enables us to identify and date near-surface exhumation events, as it has high relevance in sediment provenance and earth surface processes studies.However, the classical (U-Th)/He-dating method extracts helium from entire grains, requiring very clear, inclusion-and fracture-free, intact crystals.Thus, representative, random sampling of detrital populations is not feasible in this way.In addition, the relatively low throughput of both apatite-and zircon-based (U-Th)/He thermochronology hampers the acquisition of large data sets with high statistical relevance.An alternative solution is offered by the in situ (U-Th)/He method that applies laser ablation and extracts the radiogenic helium from the small volume of a laser pit only (Boyce et al., 2006;Vermeesch et al., 2012).This technique allows performing the analysis on an inclusion-and fracture-free part of the crystal.Thus, in-situ helium method allows (a) obtaining more ages and (b) in this way it is possible to date not only the ideal, rare, cleanest crystals, and that results in less biased sampling and better overall statistics.Due to the smaller number of radioactive decay products extracted from the laser ablation pits, this technique yields ages with slightly higher uncertainty than the classical method that performs the degassing of entire crystals (Tripathy-Lang et al., 2013).However, they demonstrated that the (U-Th)/He ages obtained by the in situ technique are geologically meaningful and interpretable.
Currently, the most advanced detrital geochronological method is based on the combination of high-T and low-T geochronometers applied to the same grains.This double dating (DD) procedure unifies the advantages of these methods and allows highly specific discrimination of the sources (e.g., Carter & Moss, 1999; D. M. Chew & Donelick, 2012;Lu et al., 2020;Reiners et al., 2005).The best examples for the application of DD are the studies when detritus of the foreland basins associated with active or young orogens were dated (Pujols & Stockli, 2021;Thomson et al., 2017;Wang et al., 2014).Detrital grains with low-T ages close to the age of deposition (i.e., short lag time, Garver et al., 1999) can derive either from coeval volcanism or from rapidly exhuming basement blocks.This geodynamically relevant difference can be resolved by DD (e.g., Barham et al., 2019;Malusà et al., 2022).
Due to the common occurrence of zircon crystals in siliciclastic sedimentary formations and the abundance of radioactive elements in zircon, the combination of the zircon in situ U-Pb dating with the fission-track or (U-Th)/He thermochronology is the most applied DD method (ZDD method; Campbell et al., 2005;Perry et al., 2009;Evans et al., 2015;Koshnaw et al., 2021).Although detrital apatite age spectra typically show more dispersion than the zircon ages, apatite DD may supply relevant age information for the near-surface, sediment generating processes (e.g., Pickering et al., 2020).Some studies used all the three methods (U-Pb, FT [U-Th]/He "triple dating") on single apatite or zircon grains (e.g., Carrapa et al., 2009;Danišík, 2019).The sources and the magnitude of the uncertainties associated with U-Pb-He DD are discussed by Horne et al. (2019) and Léger et al. (2023).

U-Pb-He Double Dating Concepts and Procedures
The U-Pb and (U-Th)/He DD method is a relatively young geochronological tool.The laboratories use different techniques, as they presented shortly below.The first paper that describes the method was by Reiners et al. (2005).They performed U-Pb geochronology by LA-ICPMS first and then extracted the helium from the grains, dissolving them in acids and determining the concentrations of the alpha-emitting elements (U, Th, and Sm) in the entire crystal (minus the aliquot that was ablated to obtain the U-Pb age) by solution ICPMS.Boyce et al. (2006) applied the laser ablation technique for both the He-extraction and for the determination of U, Th, and Sm concentrations.First, the He is measured, and after the determination of the volume of the laser pit, the same site is re-ablated by LA-ICPMS, generating a slightly wider and deeper ablation crater.Vermeesch et al. (2012) omitted the pit volume determination that is kind of an Achilles's heel of the in situ helium thermochronology, as this kind of small-scale volumetry is not yet calibrated properly.Their calculations are based on Si-U-Th elemental ratios and the 208 Pb/ 206 Pb isotope ratio supplied by the LA-ICPMS analysis.Their approach assumes a low level of common lead in the dated crystals and that in some cases it may not be fulfilled.In order to compensate for the The helium analysis took place in high to ultrahigh vacuum gas extraction systems, practically this does not allow the introduction of porous or gas-emitting substances.Thus, commonly used epoxy resins are inappropriate as grain mounting material and better suited materials such as solvent-free "Torr Seal®" epoxy, indium metal, or transparent PFA Teflon foil are used.However, all of them have serious disadvantages.The "Torr Seal" and the indium are not transparent and additionally, polishing of indium-embedded grain mounts is difficult.The PFA has very low adhesion.Thus, grains easily jump out during preparation and the PFA sheets do not keep their geometry, sometimes leading to bending during measurements.To overcome these disadvantages, we have used a grain mounting technique that generates rigid mounts with low volatile emission and enables embedding handpicked grain populations; see its description in Section 3.1.and Text S3 in Supporting Information S1.

Testing the Double Dating Method on Known Catchments and Comparing Its Resolution to Heavy Mineralogy-Based Provenance Results
In this study, modern sediments of the Inn river of the European Alps and its tributaries are analyzed by the ZDD method (U-Pb, ZHe) using a refined analytical protocol.The catchment area of the Inn is the largest in the Eastern Alps and includes basement units with highly different lithologies, formation ages and thermal histories (Figure 1).Geological studies in the Alps have a long history and thus the sediment-supplying basement units are well described and accurately mapped (see Text S1 in Supporting Information S1).Due to the high density of geo-and thermochronological data, background information for the composition is readily compiled, and thus, age distributions of the Alpine-derived sediment samples can be predicted.The ZDD data are completed by Raman spectroscopy-based heavy-mineral analysis.Our aims are (a) to test the detrital ZDD method on a well-known region with complex orogenic history focusing on the question of the reliable identification of the age components that can be attributed to individual basement units, (b) comparing the areas of the basement units in the tributaries with the observed age pattern in order to derive some hints on heavy-mineral and especially zircon fertility, and (c) initiate to compile a reference data set for heavy mineral and ZDD age distribution of the Inn sediment as one endmember for source-to-sink studies of the North Alpine Foreland Basin fill.(Bigi et al., 1990).Samples with code numbers are used for heavy mineral analysis, while sample names with characters are used for heavy minerals analysis and zircon U-Pb-He double dating.Dashed lines indicate the catchment boundaries.

Major Structural Units of the Eastern Alps
The Alps are a multiphase collisional orogen resulting from the interaction of the European and Adriatic continents (Frisch, 1979;Schmid et al., 2004;Trümpy, 1960).The first orogeny of the Cretaceous age was followed by a second Tertiary orogenic phase (e.g., Froitzheim et al., 1996).Four major units can be distinguished in the Alps according to their original paleogeographic positions: Helveticum, Southern Alps, Penninicum, and the Austroalpine unit.The latter two are exposed in the catchment of the Inn river system (Figure 1b).
The Austroalpine unit, exposed predominantly in the Eastern Alps, occupies the highest tectonic position and is subdivided into thrust nappe systems, each comprising a pre-Alpine basement and Late Palaeozoic and Mesozoic sedimentary cover (Tollmann, 1987).The dominant Upper Austroalpine units are the Mesozoic sedimentary sequences of the Northern Calcareous Alps, Palaeozoic metasediments and metavolcanics of the Greywacke zone and the crystalline basement units with remnants of Palaeozoic and Mesozoic metasediments (Silvretta and Ötztal basement units; Figure 1b; e.g., Schuster et al., 2004).The Lower Austroalpine units have a considerably smaller spatial extent in the Inn catchment.They crop out in the Err-Bernina nappe system in the uppermost Inn catchment and form a narrow stripe along the NW margin of the Tauern Window (TW) (Schmid et al., 2004).The basement nappes experienced multiple Palaeozoic and Eoalpine (i.e., Cretaceous) metamorphic overprints.
The Penninicum (e.g., TW, Engadine Window in Figure 1b) contains parts of the European margin as well as relics of the Penninic ocean.It occupies a lower structural position and is mostly covered by the Austroalpine nappes in the Eastern Alps (Frisch, 1979).The Penninic nappes are exposed along the northern margin and in the central zone of the Eastern Alps in tectonic windows (Figure 1b).In the northern Alpine (Rheno-Danubian) flysch zone, the Penninic units are mostly siliciclastic oceanic sediments that experienced only a weak diagenetic overprint due to shallow burial in an accretionary wedge.On the contrary, the Penninicum exhumed in the Engadine Window and in the TW underwent Cenozoic greenschist-to amphibolite-facies metamorphism.In these windows, the fragments of the detached European continental crust and their complex sedimentary cover schists are exposed (e.g., Schmid et al., 2013).
A detailed description of bedrock lithologies in the Inn river catchment and their geochronological and thermochronological signatures has been compiled and is available in the Text S1 in Supporting Information S1.A compilation of published zircon U-Pb ages of the major zircon supplying bedrock units is shown in Figure 2. One should consider that the analyses were performed on zircon concentrates extracted from some selected hand specimens, and thus the cumulative curves can only be used for orientation, and they do not give a representative age distribution of the given tectonostratigraphic unit.

Modern River Sand Samples
Twenty-one modern-sand samples were collected from the Inn river and its tributaries.Quantitative heavy-mineral analysis was performed on 19 samples and zircon U-Pb-He DD on six of them (Figure 1d).The sediment samples from major tributaries of the Inn river mostly represent a single tectonic or lithological unit.Thus, in the sample ERB, sediment eroded from the Err-Bernina nappes in the headwaters of the Inn river are expected.Sample UPI derives from the Upper Inn valley, dewatering partly the ERB region, but mostly the Silvretta nappe complex and its Mesozoic cover and the Engadine Window.Sample ÖTZ contains sediment that derives exclusively from the Ötztal nappe complex.Sample TW represents the central part of the TW and the catchment of sample TWR mostly contains the rim of the TW (Figure 1c).The site of sample INN was selected where the river leaves the crystalline basement-dominated area and starts crossing the Northern Calcareous Alps.
The sampling localities are listed in Table S1 in Supporting Information S1 and shown in Figure 1d.Table S2 contains the proportion of areas of the major litho-structural units of the sampled catchments.The outlining procedure of the catchments and their bedrock composition can be found in Text S2 in Supporting Information S1. Figure S1 in Supporting Information S1 demonstrates the high lithological variety of the sampled catchments.
The samples of ca. 5 kg mass were composed of multiple aliquots, taken from different parts of the sand and sandy gravel bars.This is to avoid bias by sampling just one spot, which may have experienced exceptional hydrodynamic conditions and have unusual sediment composition.The samples were sieved and treated with low-concentrated acetic acid, and heavy minerals (HM) of the 63-125 μm fraction were separated using sodium-poly-tungstate (∼2.89 g/cm³).For the preparation of zircon concentrates, a magnetic separation step was additionally used, but hand picking was not applied to minimize the operator bias.

Our New Grain Mounting Procedure
In our novel grain mounting procedure, we place the zircon crystals together with some marker magnetite grains and ∼112-125 μm sized glass beads on a "NESCHEN gudy ® 802" double side adhesive tape that is fixed on a thick glass plate.If more than one sample is placed on the grain mount, we prepare a map that mirrors the final sample arrangement.Four to six drops of Buehler EpoThin 2 epoxy resin are placed at the center of the grain-covered area (∼2 cm 2 ) and slowly distributed on the entire surface by a needle to avoid air bubbles.The resin is covered by a round 25 mm size glass slide.After resting for 24 hr at room temperature, a curing step of 2 hr at 60°C is applied.The mount is gently removed from the double side tape using a scalpel and a few drops of alcohol, applying only very little force.A minimum of 15 μm is removed during polishing using diamond abrasives to expose the interior of the crystals.The depth is controlled by the exposure of the glass beads embedded in the crystal mounts (Pickering et al., 2020).We use polishing material up to 9 μm to generate a smooth, plain, reflecting surface containing fine polishing scratches, which are useful for better optical focusing on the surface while selecting spots for laser ablation.Using this glass-epoxy sandwich, having a ∼1.3 mm thick glass substrate and a ∼80 μm thin epoxy film, we are able to create a grain mount that is form-stable and has only a minor amount of gas emitting resin.Before introducing the mounts in the gas extraction system, we keep them in a vacuum drying cabinet at 60°C for 24 hr to minimize the volatile emission.In order to reduce the pressure of the rest gases in the helium extraction line, we apply an enhanced double gettering procedure after the laser ablation.Some more hints on the preparation of the mounts and the technical description of the additional gettering can be found in Text S3 in Supporting Information S1.

In Situ (U-Th)/He Analyses
For the in situ (U-Th)/He geochronology, we applied the procedure initially used by Boyce et al. (2006) and further applied by Tripathy-Lang et al. ( 2013) and Horne et al. (2016).Helium is extracted from a relatively  2001), (h) Siegesmund et al. (2023).*: Represents the "Zentralgneiss" bodies, but the compilation also includes single-grain ages of a dacite dike from (e)."n" indicates the number of U-Pb data of 90%-110% concordance.
small pit volume and the alpha-emitting elements are measured from a larger volume that surrounds it by a quasi semi-sphere (Léger et al., 2023).The argument for choosing this geometry is rooted in the fact that the alpha particles have ∼18 μm displacement within zircon crystals (Farley et al., 1996).Therefore, the majority of the U and Th that emitted He atoms into the He-analyzed volume is situated outside the laser ablation pit used for He determination.
The grain mounts were mapped by stitched microphotographs taken in transmitted and reflected light.The extraction of the helium took place using an ESI NWR 213 nm laser in a full-metal noble gas line at the GÖochron Laboratories, University of Göttingen.Grains were randomly selected without any morphological preferences, but only grains with a minimum of 60 μm width were considered.While placing the laser spots within the grains, at least 15 μm distance from the external contour and from inclusions and fractures was ensured.The applied spot size is 30 μm, the typical depth and volume of the ablation pits are ∼9 μm and ca.6,500 μm 3 , respectively.The extracted gas was mixed with 3 He and purified for 165 s using cold SAES getter pills and a Ti-Zr getter kept at 450°C (see the description of our gas cleaning line in Text S3 in Supporting Information S1).The chemically inert noble gases and a minor amount of other rest gases were then expanded into a Hiden, triple-filter quadrupole mass spectrometer equipped with a positive ion counting detector.Beyond the detection of the 4 He and 3 He and their ratios, the partial pressures of some rest gases were continuously monitored (H 2 , CH 4 , H 2 O, N 2 , Ar, and CO 2 ).Blanks and 4 He-gas standards of known volume were measured regularly in the same way as the samples.The volumes of the 4 He ablation pits were determined by a Keyence laser scanning microscope (VK-X200) using a 408 nm wavelength and 0.95 mW laser power.
Following degassing and pit volume measurements, the mounts were placed in an ASI Resolution S155 excimer laser ablation system coupled to a Thermo Element 2 single-collector sector-field mass spectrometer to determine the U and Th concentrations and the Pb isotope ratios.The 50 μm laser spots were centered on the helium ablation pits.The laser was fired for 17 s at a repetition rate of 6 Hz.Two laser pulses were used for pre-ablation.The carrier gas was He and Ar.Analytes of 29 Si, 202 Hg, mass204, 206 Pb, 207 Pb, 208 Pb, 232 Th, 235 U, and 238 U were measured by the ICPMS.The data reduction is based on the processing of ∼34 selected time slices starting ∼3 s at the beginning of the signal.For the concentration of the alpha-emitting elements, NIST SRM 610 glass and GJ1 zircon (Jackson et al., 2004) were used as reference materials with 29 Si as the internal standard for normalization.Similar to the helium ablation pits in zircon crystals, the reference materials were pre-ablated and ∼30 μm diameter pits were generated in order to mimic the same initial geometry and provide comparable down-hole fractionation for the ICPMS ablation.
We applied a quality threshold criterion at the evaluation of the ZHe ages and data with a relative error >20% were excluded from further consideration.

Laser Ablation U-Pb Analyses
The ICPMS spot analyses provide the U and Th concentrations needed for calculation of (U-Th)/He-ages and these analyses are also used to express the U-Pb ages (see analytical conditions in Table S3 in Supporting Information S1).The calculation of the U-Pb ages and uncertainties is based on the drift-and fractionation correction by standard-sample bracketing using GJ-1 zircon reference material (Frei & Gerdes, 2009;Jackson et al., 2004).For further control, the Plešovice zircon (Sláma et al., 2008), the 91,500 zircon (Wiedenbeck et al., 1995) and the FC-1 zircon (Paces & Miller, 1993) were analyzed as secondary standards.The age results of the reference materials were consistently within the 95% confidence interval of the published ID-TIMS values.The age results of the secondary standards were consistently within the 2σ interval of the published ID-TIMS values-see detailed analytical data, their averages and MSWD values in Table S4.Drift and fractionation corrections and data reductions were performed by our in-house software (UranOS; Dunkl et al., 2008).For 206 Pb/ 238 U ages >1.5 Ga, the 207 Pb/ 206 Pb age was used.This threshold value was determined empirically according to the error trends 10.1029/2023JF007360 7 of 21 of the U-Pb and Pb-Pb ages (e.g., Spencer et al., 2016).U-Pb ages with a discordance >10% were excluded from further consideration.Age components of single-grain data were isolated by DensityPlotter (Vermeesch, 2012).

Semi-Automated, Raman Spectroscopy-Based Heavy Mineral Analyses
The identification and quantitative evaluation of transparent and translucent HM was conducted by Raman spectroscopy (Lünsdorf et al., 2019).Carefully split, representative heavy-mineral aliquots were embedded in epoxy resin and polished.The mounts were photographed in transmitted and reflected light using a Zeiss AxioImager M2 microscope with a motorized stage and a 20× objective (NA 0.4).According to Fleet (1926), all monomineralic grains were selected using the "coordsetter" software (Lünsdorf et al., 2019), and tracked coordinates were transferred to the Raman system.Raman measurements were conducted using a Horiba XploRA Plus spectrometer equipped with a motorized x-y-z stage and using a 50× objective (NA 0.5).Table S5 in Supporting Information S1 shows the parameters used in the analyses.The transmitted light intensities (absorbance) measured in the grains were used to set the proper laser power to optimize the signal intensity and to avoid the photo-oxidation of the grains.The acquisition time of each analysis was estimated by a short test-measurement in order to gain a signal height of 5,000 counts.At every 100th analysis, a 4-Acetamidophenol standard was measured.Acquired spectra were automatically corrected for temporal drift based on Raman band positions of the standard, smoothed, scaled, and corrected for the epoxy embedding medium spectrum, compared to a modified version of the RRUFF database (Lafuente et al., 2015), and a hit index (Rodriguez et al., 2011) was assigned (Lünsdorf et al., 2019).Hit index values of 0.00-0.15are considered as "good hits" and were directly accepted.Analyses with a HI of 0.15-0.30are put into the "medium hit" group and such data demands to be evaluated manually, because these can contain correctly identified minerals but often consist of mixed or poor-quality spectra.All the other spectra were classified into the "no hit" group and not considered.More details on the experimental procedure of the applied semi-automated heavy mineral analysis can be found in Lünsdorf et al. (2019).

Heavy Mineral Data
In the 19 studied sand samples between 785 and 2,678, transparent, non-micaceous heavy mineral grains were identified by the Raman spectroscopy-based technique.The heavy mineral spectra comprise diverse mineral species and the composition of samples is variable, reflecting the complexity of the catchments (Table S6; Figure 3).Metamorphic minerals dominate, especially garnet (max.80%), amphibole (up to 68%) and epidote (up to 37%).Staurolite, chloritoid, and kyanite are minor components, with an abundance of a few percentage.The ultrastable HM (zircon, tourmaline, rutile; ZTR) are present throughout, mostly in minor proportions, but ZTR may reach ∼40% in the Zillertal (mostly rutile).Titanite is also present in all samples (max.15.7%) with highest concentrations in the ERB and Upper Inn Valley samples.Pyroxene is minor (max.4.7%) and mostly restricted to the Upper Inn valley.The proportion of apatite is highly variable; in sample TWR (TW rim, Zillertal), it is the predominant species (61%).In samples TW, 41-11 (Ötztal) and 41-13 (Trisana) garnet concentration exceeds 60% (Figure 3).The association of amphibole + epidote + titanite + pyroxene accounts for more than half of the counted grains in all of the Upper Inn valley samples, and in a couple of other samples of the tributaries, reaching ∼45% in the downstream INN sample.
Notably, the Raman spectroscopy-determined concentrations are similar to optically determined mineral concentrations in the Inn sand samples of Garzanti and Andò (2007), except some differences for the Ti-oxides.Obviously, our choice of only one small grain-size window (63-125 μm) does not violate the general heavy mineral results in these proximal high-energy Alpine river settings.

Detrital Zircon U-Pb-He Ages
In total, 768 DZ crystals were dated, and after applying the quality threshold criteria (see Section 3), 637 U-Pb ages and 631 (U-Th)/He ages can be used for interpretation.We present the results separately for the U-Pb and the (U-Th)/He data with a focus on the relations to potential source regions (Figure 4).
The U-Pb data are listed in Table S7 and visualized in Figures S2-S7 in Supporting Information S1 as Wetherill concordia plots and in Figure 4b as cumulative curves.The complementary Figure 4a shows the distribution of main lithologies on the Inn catchment, and the respective zircon U-Pb ages as compiled from the literature (see Figure 2).The obtained age distributions are different for the individual sand samples.Particularly, the ERB and ÖTZ samples form distinct end-members, with ERB containing almost exclusively Permian zircon ages, while ÖTZ is exclusively composed of pre-Variscan grains.The other age spectra, including the INN sample, can be all explained as mixtures between these two end-members.All samples apart from ÖTZ contain significant Permian age groups (Figure 4b).Typical Variscan ages (∼360-300 Ma) are most prominent in the Upper Inn valley and in the Zillertal samples draining the TW (TWR).Further significant age groups clustered around 450 Ma (mostly in TWR, UPI, and INN), 600 Ma, and 1,000 Ma (mostly in ÖTZ, UPI, and INN).Major Mesoproterozoic zircon-forming igneous events known from the European crust (e.g., Avigad et al., 2022) are also seen in the cumulative plot (Figure 4b).
The ZHe analytical data are listed in Table S7 and plotted in Figure 4d.The complementary Figure 4c shows the distribution of published and inferred ZHe ages for the major units as compiled from the literature (see details in Text S1 in Supporting Information S1).The ages are much younger than the U-Pb ages as the zircon He-thermochronometer was reset in all Alpine units during the Cretaceous Eoalpine and/or in the later Tertiary tectono-metamorphic events.The samples yield distinguishable age distributions except for samples TW and TWR, which show the youngest, and rather similar age distributions (Figure 4d).This is in accordance with the well-constrained Miocene exhumation of the TW.Samples ERB and UPI are distinct given their dominant 40−20 Ma (U-Th)/He age group (Figure 4d), reflecting the Oligocene cooling event characteristic for the region of the Upper Inn valley.Although sample UPI encompasses erosional material from the Penninicum of the Engadine Window, no Miocene age component is recorded.This either reflects the low fertility of zircons from calcschists, or the Oligocene ZHe ages detected by Price et al. (2018) along the margin of the Engadine Window is characteristic for the entire window, and its exhumation preceded the Miocene exhumation of the TW.Like the

Discussion
In the discussion we will (a) evaluate the heavy mineral data of the samples for their faithful reflection of catchment lithologies, (b) evaluate the U-Pb distribution with the same intention, (c) use the U-Pb data and the spatial distribution of source rocks to constrain zircon fertility, (d) evaluate and discuss the meaning of the (U-Th)/He data for which much less directly comparable data from the source rocks exist, and, finally (v) discuss applicability and the advances of the U-Pb-He DD approach in such lithologically and structurally complex catchments like the Inn river.The individual data sets will be discussed from the headwaters to the downstream INN sample.

Heavy Mineral Relations Between Bedrocks and River Sand Samples
In the catchment of the ERB sand sample, the exposed basement contains amphibole-bearing granitoids and some minor gabbro lenses forming the Permo-Carboniferous Err-Bernina igneous suite (Figure 1d; von Quadt et al., 1994).Thus, the apparently mafic mineral association (Ep + Amp + Px + Ttn) in the ERB sample mostly mirrors an igneous rock origin rather than metamafic.Interestingly, sample 41-20 from a small subcatchment draining mostly Variscan granitoids yields much more zircon, TiO 2 -phases and apatite, suggesting that this lithology is responsible for most of the zircon from the ERB catchment.Downstream of ERB until UPI, that is, along the Upper Inn valley, the abundant metamafic lithologies of the Silvretta nappe complex contributed to the Ep + Amp + Px + Ttn budget.However, we see a slight decrease in these mafic minerals compensated mainly by increases in garnet, apatite, and ZTR (Figures 3 and 5a).
The two Trisana (Paznauntal) samples (41-12, 41-13) are contrasting (Figure 3).Sample 41-12 drains a rather small catchment, dominated by amphibolites and greenstones (Figure 1d).Besides high amphibole and epidote concentrations, it shows the highest concentration of pumpellyite, indicative for low-grade metabasic rocks.Downstream, sample 41-13 drains a much larger area dominated by medium to high grade metasedimentary basement of the Silvretta complex, which is reflected by garnet predominance along with some staurolite and kyanite.
In the Ötztal, the headwater sample (41-11) strongly differs from the three other samples (Figure 3).Its composition is pretty similar to 41-13, but here metasedimentary rocks of the Ötztal nappe complex dominate the catchment, including garnet-rich micaschists of the Schneeberg Complex that occupy the southernmost part of the Ötztal (Figure 1b; e.g., Klug & Froitzheim, 2022).Sample 41-10 represents an amphibolite-rich end-member, which was collected from a small side valley (Figure 1d, Table S2); ca.90% of the heavy mineral spectrum is covered by three minerals: amphibole, epidote and titanite.The evaluation of the geological maps indicates ca.42% metamafic lithologies in the area of this small catchment, but the pebbles at the mouth of the creek are dominated by amphibolites.This can be explained by the typically higher fracture toughness of the amphibolite over the more foliated schists (e.g., Jahnke et al., 2022).The more downstream samples in the Ötztal (41-9, ÖTZ) mainly reflect a mixture of these two main heavy mineral-supplying lithologies of the catchment: Ep + Amp + Ttn decreases to 60% in the ÖTZ sample.This sample, however, shows relatively high zircon and apatite content, which may point to the contribution from the Ordovician felsic metagranitoids documented in the basement of the lower reaches of the Ötztal catchment (e.g., Thöny et al., 2008).
In the samples from the Zillertal, the individual heavy mineral spectra show considerable variation (Figure 3), most likely being a consequence of the zoned architecture of the TW (cf.Figures 1 and 3).In the TW sample, the garnet is dominant (>80%); it probably derives from garnet-bearing micaschists.The TWR sample was collected in a smaller catchment to obtain the bulk composition of the "Schieferhülle," the variegated metasedimentary assemblage that envelops the orthogneiss cores (Figure 1d).The ca. 30% ultrastable phases and the high apatite content mark a significant difference from the other samples.Of the headwater samples, 41-3 seems also affected by the "Schieferhülle," while 41-4 reflects more mafic lithologies.The most downstream sample 41-7 represents a mixture of all the others, underpinning the high rutile content of the metasedimentary rocks of the Zillertal catchment.

How Diagnostic Are the Observed Heavy Mineral Spectra?
The bedrock formations of the Inn catchment can be grouped into three major lithologies: metagranitoids, metasediments, and metamafic rocks.The transparent-translucent heavy mineral yield from felsic rocks appears to be low, both in the number of phases and in their abundance.In fact, mostly zircon and apatite crystals can be expected from the typical metagranitoid rocks.The Err-Bernina igneous suite is an exception to the intermediate and even the felsic rocks are amphibole-bearing, and their influence (together with the minor gabbro lenses) can be seen in the ERB sample that derived mostly from this suite.
In metapelitic-metapsammitic rock garnet, staurolite, Al-silicates, and chloritoid can be present in higher percentages as rock-forming phases, which may strongly influence the HM composition.In case of the samples from the Inn tributaries, only garnet has dominated role, while the staurolite and others are minor contributors, in agreement with the presence of the garnet-bearing micaschists in the region.The TW sample is a good example of how dominant the garnet contribution from micaschist members of relatively subordinate areal proportion in a metagranitoid-dominated area can be (garnet: 75%, apatite + zircon: 3%; see Figure 3, Table S6).In the TW catchment, ca.60% of the area is composed of metagranitoids, and the schists forming the rest of the area are mostly garnet-free, low-grade phyllites and quartzites and the potentially garnet-bearing micaschist + paragneiss assemblage occupies only a few percentage of the catchment (Table S2).
Amphibole, epidote, titanite, and pyroxene are all rock-forming minerals occurring abundantly in metamafic rocks.Thus, their impact on the HM spectra is necessarily strong.The sum percentages of metasedimentary and the metamafic phases form a mixing line and they together give the majority of HM grains in most of the samples (Figure 5a).Remarkably, some samples from the Zillertal do not plot along this mixing line due to their high zircon-tourmaline-rutile (ultrastable) content (Figure 5b).Samples from the Upper Inn valley (ERB, UPI and 41-13 to 41-20) derive mostly from the Silvretta nappes that are rich in mafic lithologies.The pyroxene + titanite proportion relative to the epidote + amphibole trend in these samples is different from the Ötztal samples (Figure 5c).Thus, mafic-derived heavy mineral signatures allow the distinction of different basement units within the Austroalpine realm (Figure 3).
We compared the proportion of the metamafic-dominated areas of the catchments and the sum of the mafic mineral proportions detected in the corresponding sand samples.From such a comparison, a trend with positive correlation is expected.In the Ötztal and in the Upper Inn valley there are catchments, where both the proportion of metamafic bedrocks and the mafic mineral proportions are high and fulfill the expectations; however, the rest of the samples does not show an obvious pattern with correlation (Figure 5d).The surprising feature of this plot is the variable, but in some cases high mafic mineral content in sand samples, derived from catchments of low or even negligible metamafic bedrock areas (<1%, see Table S2).The most plausible explanation for the dominance of mafic minerals is the low HM yield of the other lithologies forming the catchment.When the micaschists and gneisses contain low amount of garnet and other typical metasedimentary phases and/or the metagranitoids are common on the catchments, then the HM contributions from these lithologies are minor.In such cases, the HM yield from metamafic rocks is strongly overproportional and can be dominating (Tables S2 and S6).
The role of apatite is not obvious in the studied modern sand samples.Apatite is present in all samples and the abundance shows a weak correlation with rutile; it is derived presumably from all metamorphic rocks.Potentially, the low-grade rocks at the rim of the TW, beyond the ultrastables, are also enriched in apatite (sample TWR).

Detrital Zircon U-Pb Age Distributions
In the ERB age spectrum, 102 concordant U-Pb data points form a narrow population with a mean age of 286 Ma, and only two crystals yield Ordovician ages (Figure 4b).These Permian ages are in accordance with the age range determined for the Err-Bernina igneous suite (Von Quadt et al., 1994).However, these magmatic rocks cover only ca.50% of the ERB catchment, the rest is mostly ortho-and paragneisses with some flysch, which are typically zircon-bearing lithologies (Table S2).Thus, here we detect a strong over-representation in the DZ age spectra of the Permian granitoids above the pre-Variscan metamorphic basement.
The Upper Inn sample was collected downstream, where the Inn had already crossed the Silvretta basement and the Engadine Window (Figures 1d and 4a).The UPI ages show a more complex age pattern that includes, in addition to the ERB-derived Permian ages, also Variscan, Ordovician, Cadomian, and Proterozoic age components (Figures 4b and 6).Remarkably, the Permian age components in the ERB and UPI samples are significantly different according to the t-test.Thus, we assume that between the sampling point of the Err-Bernina-dominated sand and the UPI sampling site some other Permian and Variscan igneous bodies are exposed and contribute zircon grains to the Inn sand.The robustness of the U-Pb data collected from the literature allows comparison of the age components identified in the modern sand and in the bedrock formations (Figure 7a).In the UPI age spectrum, the imprint of the Mesozoic siliciclastic sediments exposed in the Engadine Window is obvious; however, the influence of the mostly >500 Ma ages from the Silvretta basement is less pronounced.
The age distribution of the ÖTZ sand is in accordance with the exclusively pre-Variscan ages obtained on the Ötztal basement (see Text S1 in Supporting Information S1).The cumulative curve of basement data compiled  (Vermeesch, 2012); the mean ages of the components and their proportions can be found in Table S8.For comparison, the upper panes show the age components identified in the zircon U-Pb data determined from the bedrock formations.The sources of the literature data are listed in the caption of Figure 2.
by Siegesmund et al. (2021) yields an excellent match to the Ötz river DZ ages (Figure 7a).Interestingly, a small young tail that appears in the basement data, forming a minor Variscan age component in Figure 6 that is missing in the sand age distribution.This is in contrast with the expectation, as the sampling of modern sands should better reveal the smaller, hidden bodies of the catchment with young ages and not the hand-specimen sampling that is performed on the well-known, mapped, major units of the metamorphic basement.Most likely, the absence of these young zircons is related to the grain selection method required for the U-Pb-He DD.We dated the central part of the crystals of at least 60 μm in diameter, while for common, in situ U-Pb geochronology smaller grains can also be used that are potentially more sensitive to rejuvenating events.
The TW and TWR samples contain ca.80% and 50% of Permian ages, respectively (Figure 4b).These age components do not differ significantly at 0.05 level, while they are different from the Permian age population in the ERB sample according to the t-tests.Additionally, the TWR sample contains a minor "younger tail" that indicates that zircon grains were derived from the Permo-Mesozoic metasediments exposed along the margin of the TW.The low proportion of these young grains and their scatter do not allow the isolation of a singular age component, but the overlap of the youngest data in the TWR sample and in the literature collection of metasediments is remarkable (Figure 7a).The pre-Permian age components indicate strong similarities; the only difference is the presence of early Variscan ages in the Permo-Triassic metasediments of the TW, whose ages are underrepresented in the TWR sample (Figure 6).In the TW sample, which represents the central part of the TW, the pre-Permian The U-Pb age distribution of the INN sample occupies an intermediate position between the end-members (Figure 4b).Although the cumulative curve seems to be rather smooth and beyond the Permian and Ordovician concentration of data it does not contain characteristic stairs, the DensityPlotter software (Vermeesch, 2012) was able to identify five age components that match to the age components determined in the Upper Inn valley and in other tributary samples (Figure 6).

What Conclusions Can Be Drawn for Zircon Fertility?
In order to compare the U-Pb ages obtained from the modern sand samples with those of the basement formations, binning into age intervals was applied.The boundary of 250 Ma was set to the Permian-Triassic boundary in order to detect the contribution from Mesozoic formations.The boundaries of 380, 500, and 700 Ma were determined by the first derivatives of the cumulative curves, by other words, where the data density is low -see the upper panel of Figure 7a.Predicted DZ age distributions were calculated for each dated catchment.The areal percentages of the zircon-bearing basement formations were considered and their U-Pb age distributions according to the available published ages (Figure 2).The above-mentioned age ranges were used for the comparison of the predicted and measured age distributions (Figure 7b).
In sample TWR, the predicted and measured age distributions are rather similar; the deviations of the age bins are less than 10%.Considering the counting uncertainties, the observed deviations are minor.However, in samples ERB and TW, the 380−250 Ma (Variscan + Permian) age range is strongly overrepresented.Similar trends-though less pronounced-can be seen in samples UPI and TWR.This result indicates high zircon fertility from the Permian and Variscan metagranitoid formations relative to the metasedimentary lithologies.This can be explained by the fine-grained zircon-poor protoliths of the metapelites, which are the other most common lithology in the region.These schist sequences also contain quartzites that are rich in DZ grains, but the overall zircon fertility of these sequences is still very low.Notably, metapelite-derived garnet dominates the heavy mineral spectrum of sample TW, but the proportion of zircon is very minor (0.5%).Thus, according to the HM and DZ analyses one can draw highly different conclusions as the garnet is overrepresented in the HM spectra and within the U-Pb ages the zircons from igneous rocks are overrepresented relatively to the zircons from metasedimentary rocks.
The INN sample represents the sediment of the "bulk Inn river" in the Lower Inn valley.Here the above-mentioned high fertility of the Variscan and Permian metagranitoids does not appear, and the predicted and measured proportions within this age range are proportionately close.Where the INN sample does show bias is in the underrepresentation of the "Ordovician" (500-380 Ma) age range and overrepresentation of the s.l."Cadomian" (700-500 Ma) age range.A similar offset can also be seen in the ÖTZ sample that represents a large part of the Inn catchment.

Detrital Zircon (U-Th)/He Age Distributions
When evaluating age distributions obtained by low-T thermochronometers, one should apply slightly different concepts and statistical methods than in the case of U-Pb results.Generally, the overdispersion is more common in low-T ages than in zircon U-Pb geochronology.It comes from the lower number of radioactive events counted-this affects fission-track dating and the in situ helium dating-and also the zonation of the actinide elements in the dated crystals introduces more bias for low-T zircon ages than for zircon U-Pb ages (e.g., Fitzgerald et al., 2006).There is another, even more significant factor that results in the non-normality of the low-T age distributions.In case of the U-Pb dating of a volcanic formation or an igneous suite, the magma generating process typically has a limited time range and its duration is much shorter than the time since their formation (only very young, Pliocene-Pleistocene igneous events are exceptions).The assumption of a symmetrical, maybe even Gaussian distribution of the single-grain U-Pb ages of an igneous suite is not impossible.The low-T ages are formed by exhumation or by thermo-tectonic processes triggering a vertical retreat of isotherms in the upper crust.Such ages always have a vertical trend and for example, in the mountains at higher elevations, older cooling ages are exposed and eroded compared to lower elevations (Malusà & Fitzgerald, 2019).That is why the low-T age distributions are typically wider and oblique, typically with positive skew.
In both samples from the Upper Inn valley, Oligocene ZHe ages are the major component (Figure 4d).We can compare the data from the intervals younger than ca.40 Ma, as this age is an inflexion point in both age spectra.These intervals yield mean ages for the ERB and UPI samples of 30.8 ± 1.2 and 30.8 ± 1.9 Ma (with 95% CI of mean), respectively, and the t-test and the Shapiro-Wilk test indicate similarity and no significant deviation from the normal distribution.These tight ages match with ZHe ages obtained on the Err-Bernina region (Price et al., 2018) and these ages are probably related to the rapid post-emplacement cooling of the region triggered by the Bergell intrusion.The older, pre-40 Ma ages are derived from the structurally higher Silvretta basement.Such ages are more abundant in the downstream UPI sample compared to ERB, where the Upper Austroalpine basement has a minor share in the catchment.
From the Ötztal basement, the available ZFT and ZHe ages indicate post-Eoalpine cooling (Elias, 1998;Lünsdorf et al., 2012), and our detrital ZHe age spectra from the river sand fit the bedrock age range (Figures 4c  and 4d).The precision of the individual Cretaceous ZHe ages is slightly better than the Oligocene ages of the Upper Inn valley due to the higher helium signal at the measurement, but the overall spread of the ÖTZ data is much wider and no obvious age components can be seen.Here, the influence of the elevation-dependence of the low-T ages is manifest.The ca. 2 km elevation range of the Ötztal catchment exposes a zircon helium partial retention zone that was developed since the Cretaceous metamorphism.
The samples from the central and the marginal zones of the TW yield similar cooling age patterns (Figure 4d).The ZHe age distributions are oblique, they cannot be represented by mean and s.d.values and their splitting into age components is also not feasible.In contrast, they can be described by a lognormal distribution: P-values are 0.16 and 0.41 for TW and TWR samples, respectively, and the characteristic mean ages are 15.7 ± 1.1 and 17.8 ± 1.0 Ma (with 95% CI of mean).The TWR ages are slightly older than the TW ages, and this difference can be explained by the tectonically higher position of the schists forming the cover envelop of the metagranitoid cores of the window.The schists experienced earlier cooling due to their higher tectonic position.
In the TW catchment the relief (=vertical range of sediment generation) is similar to the Ötztal, ca. 2 km, and vertical age dependence is also described from the TW (e.g., Fügenschuh et al., 1997;Most, 2003;Wolff et al., 2021).However, the minor spread in the ZHe ages refers to the fast cooling of the TW, triggered by the young and ongoing exhumation and not the erosion of an old, quasi "frozen" helium partial retention zone like in the Ötztal.
The ZHe age distribution in the bulk sand of the Inn river is rather flat and diffuse.The characteristic age groups that were determined in the Upper Inn and Tauern samples are hardly recognizable, and it is difficult to identify age components with proper reproducibility.The reason for the diffuse age spectra is partly due to the onlap of the relatively close Miocene and Oligocene age groups, and we have to consider that the sampled tributary areas cover only ca.half of the total Inn catchment (Figure 1c).The non-dated areas are the Silvretta basement, the western part of the Ötztal basement, the transition of eastern Ötztal to western TW (Brenner region) and the Innsbruck quartzphyllite zone, north of the TW.These areas are unevenly influenced by older (Eoalpine or even pre-Alpine) and younger Neogene thermal events (see Text S1 in Supporting Information S1); thus, they probably supply complex ZHe age spectra.

Evaluation of the Zircon Double Dating of Modern Sands of the Inn Catchment
The combined plotting of the high-T and low-T ages yields a distinguishable pattern for the dated tributaries of the Inn system (Figure 8).The ERB and TW distributions are very tight, but the centers of both the U-Pb and the He-ages are different.The UPI and TWR catchments contain more variegated lithologies/tectonic units than the ERB and TW, and this is well mirrored in their wider spread of ages.The distribution of the ÖTZ sand shows very little overlap with the other samples, as it does not contain Variscan and Permian zircons and its Cretaceous cooling preceded the Neogene-dominated cooling of the other tributaries.

Conclusions
-The heavy mineral composition of the modern sand samples of the Inn river reflect the major bedrock lithologies exposed within the catchments (metamafic rocks, high-and low-grade metasediments and metagranitoids).The HM spectra can be described by the admixture of three HM assemblages: (a) amphibole + epidote + titanite + pyroxene, that derives mostly from the metamafic rocks, (b) garnet > staurolite ± chloritoid that comes mostly from the garnet-bearing micaschists, and (c) ultrastable assemblage of zircon + tourmaline + rutile that is the yield of low-grade schists typically from the margin of the TW.Zircon and apatite were also supplied from the metagranitoid units.
-There is no obvious correlation between the areas of the characteristic lithologies on the catchments and the proportions of the HM assemblages in the corresponding modern sand samples.The dominant HM are those that appear as rock-forming phases (mostly amphibole, epidote and garnet).HM phases that typically occur as accessory minerals in the basement rocks are less abundant (e.g., zircon or tourmaline).
-The DZ U-Pb age distributions in the sands from the tributaries are highly heterogeneous and represent the characteristic age patterns of the basement formations.The Permian Err-Bernina igneous suite and the Permian metagranitoids of the TW yield distinguishable age patterns.
-Relative to the areal percentages of the metagranitoids, their zircon contributions are overrepresented in the age spectra within the sands from the Err-Bernina and TW catchments.This indicates an enhanced fertility of these metagranitod formations relative to the metasedimentary basement.In the case of sample TW (central part of TW), the discrepancy between the conclusions that can be deduced from the DZ U-Pb ages and from the heavy-mineral spectrum is particularly strong.
-In the bulk INN sample collected in the lower Inn valley, it is possible to isolate nearly all U-Pb age components that were identified in the tributaries, but the Variscan age component does not appear.
-The ZHe ages of the individual catchments are well distinguishable and generally match the low-T thermochronological ages obtained from previous studies on the basement.The Err-Bernina and Upper Inn sand samples mostly reflect the impact of the Oligocene Bergell intrusion, but the expected Neogene cooling ages from Engadine Window were not detected.Beyond potential zircon fertility differences, we cannot exclude that Engadine Window experienced an earlier Oligocene exhumation than the TW.
-The Ötztal sand contains ZHe ages from the range of post-Eoalpine cooling (Cretaceous-Paleogene), while the sand samples derived from the TW contain mostly Miocene cooling ages.In case of the two Tauern catchments, a younger age range was detected in the sample representing the schist-rich catchment that occupies a higher tectonostratigraphic position compared to the metagranitoid-rich sample representing the core of the window.
-The combined evaluation of the DD results appears to be a very robust tool for distinguishing sediments from small and medium-sized catchments.In the bulk sand of the Inn River, the major sources are still recognizable, but minor differences are smoothed down.

Data Availability Statement
The plots and conclusions of this manuscript are all based on the analytical results that are submitted as Supporting Information S1.We generated a complex table that contains all analytical details obtained by the multiple analysis (helium and U-Pb) performed on single grains.The content and structure of the tables allow reproducing the presented figures and generating any type of mineralogically and isotope geochronologically relevant statistical analysis and graphics.All Supporting Information is available from the FAIR repository of Göttingen Research Online: https://doi.org/10.25625/68FGZ2(Dunkl et al., 2023).Special thanks to Michael Tatzel, Judit Nagy, and Irina Ottenbacher for help with various sample preparation procedures and to Sarah Feil for the correction of the English.All the data used in this study can be obtained in the figures, tables, references, and supplementary files (Dunkl et al., 2023).We appreciate careful reviews by Gary J O'Sullivan and an anonymous reviewer as well as guest editor Chris Mark.Open Access funding enabled and organized by Projekt DEAL.
poorly calibrated pit volumes,Pickering et al. (2020) applied a calibration factor for unknown samples.This factor is determined by the ratio of ages measured on the quasi-homogeneous, gem-like, age reference Durango apatite(McDowell et al., 2005) by the classical, whole fragment technique and by the in situ technique.Tian et al. (2017) tested and distributed a zircon age reference material that allows matrix-matched calibration of laser ablation ZHe chronology.To avoid difficulties caused by the deep, irregular laser ablation pits that typically have steep and badly characterized walls,Evans et al. (2015) used relatively shallow pits for helium extraction.Their volumes can be measured either by confocal laser scanning microscopy or atomic force microscopy.The U, Th, Pb, and Sm contents are determined by a smaller laser ablation spot that is placed inside the previously ablated pit for He measurement.

Figure 1 .
Figure 1.(a) Outline of the of the Inn river's catchment within the Eastern Alps; DEM from Copernicus Land Monitoring Service.(b) Tectonic map of the study area with a catchment of the Inn river (after Schmid et al., 2004).NCA: Northern Calcareous Alps, GWZ: Greywacke Zone, Ö: Ötztal unit, S: Silvretta unit, EW: Engadine Window, LAA: Lower Austroalpine unitsin the Upper Inn valley dominated by the Err-Bernina igneous suite, SC: Schneeberg Complex-exposed only in the southern Ötztal massif.(c) Map inset emphasizes the catchments of the six double-dated modern sand samples.(d) Modern sand sample localities on the geological map of the Inn catchment(Bigi et al., 1990).Samples with code numbers are used for heavy mineral analysis, while sample names with characters are used for heavy minerals analysis and zircon U-Pb-He double dating.Dashed lines indicate the catchment boundaries.

Figure 3 .
Figure 3. Heavy mineral spectra obtained on the modern sand samples of Inn river and its tributaries.Brookite, anatase, and their admixtures are added to TiO 2 , and due to the low grain numbers the sillimanite and andalusite concentrations are not plotted, see the raw data in the Table S6.Mineral abbreviations follow the IUGS recommendations (Whitney & Evans, 2010); Tris.: Trisana Bach (Paznauntal).

Figure 4 .
Figure 4. (a) Simplified lithological map of the Inn catchment with the double-dated sample localities (stars, see sample codes in Figure 1c).The legend comprises groups of basement formations with similar zircon U-Pb age patterns.For sources of the data see Text S1 in Supporting Information S1. White colored areas do not supply Detrital zircon (DZ) grains (carbonate-dominated units and the slates of the Grauwacke Zone), AA: Austroalpine, TW: Tauern Window.(b) Cumulative plot showing the DZ U-Pb data obtained on the modern river sand samples.Color coding of the samples corresponds to the color of stars on the left panel (a) and Figure 1c.(c) Map of the Inn catchment with the distribution of measured and inferred zircon (U-Th)/He ages of the major structural units.The most constrained units are the Err-Bernina region, the Silvretta nappe, the SW tip of the Engadine Window and the westernmost Tauern Window (TW).The inferred ages are mostly in the Ötztal block, which are deduced from the AFT and ZFT ages.The less constrained area is the Austroalpine belt north of the TW-see sources and details in the text.(d) Cumulative plot of DZ (U-Th)/He ages obtained on the modern river sand samples.

Figure 5 .
Figure 5. Heavy mineral composition of the modern sand samples from the Inn river system.The minerals are grouped according to their assumed genetic relationship.Samples derived from the same region are represented by similar symbols, and the double-dated samples are distinguished by filled symbols.(a) Relationship between metabasic-derived heavy minerals (HM) and metapelitic-derived HM.(b) Triplot emphasizing the higher proportion of the ultrastable minerals in the Zillertal (Tauern Window) samples and their deviation from the basic-metapelitic mixing line.(c) The pyroxene + titanite versus epidote + amphibole trend in the Upper Inn valley samples is different from the Ötztal and Zillertal samples.(d) Relation of the proportion of metabasite formations on the catchments and the sum of the metabasit-derived HM.

Figure 6 .
Figure6.Age components identified in the detrital U-Pb data obtained from modern sand samples.The DensityPlotter software was used for the component search(Vermeesch, 2012); the mean ages of the components and their proportions can be found in TableS8.For comparison, the upper panes show the age components identified in the zircon U-Pb data determined from the bedrock formations.The sources of the literature data are listed in the caption of Figure2.

Figure 7 .
Figure 7. (a) Comparison of the new detrital zircon U-Pb ages and the data obtained on basement samples (see the sources in Figure 2).The horizontal bar above the plot indicates the age bins used for the comparison of measured and predicted age distributions.(b) Comparison of the measured U-Pb ages (right columns) and the predicted zircon U-Pb ages according to the area and the U-Pb ages of the basement units in the catchments (left columns).

Figure 8 .
Figure 8. Double-dating results of combined (U-Th)/He and U-Pb analysis on single detrital zircons.On the two-dimensional KDE plot, the age patterns of the individual tributaries are distinguishable.Isoline border areas represent 20% cumulative steps of the data.