Combined cosmogenic ¹⁰Be, in situ ¹⁴C and ³⁶Cl concentrations constrain Holocene history and erosion depth of Grueben glacier (CH)

Abstract

Surface exposure dating with a single cosmogenic nuclide relies on the assumption of simple constant exposure. In contrast, the combination of nuclides with different half-lives or production rate depth profiles allows constraining complex exposure histories. Here we present the systematics of how the combination of ¹⁰Be, in situ ¹⁴C and ³⁶Cl can be used to understand both burial and erosion in a rock surface, in addition to yielding an exposure time. Concordant ages of all three nuclides characterize simple exposure. Apparently ‘too young’ ¹⁴C ages or ‘too old’ ³⁶Cl ages indicate burial or erosion of the rock surface, respectively. We further report a first data set from a proglacial granite bedrock ridge in a small cirque in the Central Swiss Alps. Three samples were exposed for 9.6 ± 0.3 ka during the Holocene. The use of same-sample ¹⁰Be, in situ ¹⁴C and ³⁶Cl allowed us to quantitatively show that these rock surfaces experienced simple constant exposure with ≤500 years of ice cover and ≤6 cm of erosion. In addition, the fact that samples inside and outside the Little Ice Age (LIA) ice extent yield indistinguishable ages reinforces the hypothesis that Grueben glacier reached comparable dimensions only briefly during the Holocene. Notably, at sites just inside of the LIA extent, which do exhibit fresh glacial polish, concordant ages suggest that subglacial erosion there was minimal.

1 Introduction

The concentration of cosmogenic nuclides in rocks is used by geomorphologists to gain knowledge on the age of landforms such as moraines or landslides (Gosse and Phillips 2001; Bierman et al. 2002; Ivy-Ochs and Kober 2008). In many studies, considerations based on geomorphological field evidence allow the assumption that rock surfaces were exposed in a single continuous episode. Then the analysis of one cosmogenic nuclide species suffices for dating purposes. This assumption is violated, however, if temporary burial or erosion of the rock surface cannot be excluded. In these cases, the combination of multiple cosmogenic nuclides can potentially resolve the details of the landform history (Miller et al. 2006). Indeed, the complexity itself is often the subject of interest: how long was a surface covered by ice or sediment? How deep was the bedrock underneath a glacier eroded? Whereas in the past, mainly the pair of ¹⁰Be and ²⁶Al has been utilized successfully to unravel complex exposure histories (Bierman et al. 1999; Corbett et al. 2013), in this study we combine same-sample results from ¹⁰Be, in situ ¹⁴C and ³⁶Cl.

The system of ¹⁰Be is the most frequently used and best understood amongst all cosmogenic nuclides. It has proven to be a reliable dating method on a broad temporal scale from centuries (Schimmelpfennig et al. 2014) to millions of years (Ivy-Ochs et al. 1995). In situ ¹⁴C has the potential to add substantial information about recent episodes of burial. Compared to ²⁶Al it decays more rapidly (t_1/2 = 5.7 ka) and can thus allow detection of shorter burial periods down to a few centuries (Miller et al. 2006; Goehring et al. 2011; Hippe et al. 2012). Owing to the challenges in the extraction process of in situ ¹⁴C from quartz mineral separates (Lifton et al. 2001), only a few studies employing this nuclide have been reported. Both nuclides, ¹⁰Be and in situ ¹⁴C, are mainly produced by spallation reactions from high-energy neutrons with secondary contributions from muon interactions. ³⁶Cl, in contrast, is produced in a variety of pathways depending on the chemistry of the host rock (Gosse and Phillips 2001; Alfimov and Ivy-Ochs 2009). In particular, capture of thermal neutrons by ³⁵Cl can contribute significantly to the production of ³⁶Cl. A part of the thermal neutrons are lost close to the rock surface due to diffusion (Fabryka-Martin 1988; Phillips et al. 2001). Therefore, ³⁶Cl production in rocks typically peaks at a depth of a few decimeters (depending on the concentration of natural chlorine), not at the rock surface (cf. Figure 1). When analyses of ³⁶Cl and ¹⁰Be or ¹⁴C are combined, the difference in the shape of their production rate depth profiles can be used to quantify the depth of erosion.

Fig. 1

Modelled production rate depth profile for sample Grub2, see chemical composition in Table 3. Solid lines are total production rates for ¹⁰Be (red), ¹⁴C (blue) and ³⁶Cl (green). Detailed contribution of different production pathways are shown for ³⁶Cl. The magnitude of the characteristic subsurface production peak caused by thermal and epithermal neutrons (dotted line) depends on the concentration of ³⁵Cl in the rock

The following three scenarios serve to illustrate how combined analyses of ¹⁰Be, ¹⁴C and ³⁶Cl can allow distinguishing between constant exposure, burial and erosion scenarios (Figs. 2, 3). (a) In a scenario with one continuous period t_e = 9.6 ka of exposure (Fig. 2, solid lines), cosmogenic nuclides form constantly and decay according to the different half-lives. The final concentrations of each nuclide correspond to an exposure age of 9.6 ka (Fig. 3a). Concordant ages (within uncertainties) calculated from all measured nuclides therefore indicate constant exposure. (b) If after the exposure period the samples were temporarily covered and shielded from cosmic radiation, for example by a glacier for a duration of t_b = 1.0 ka (Fig. 2, dashed lines), no new nuclides are produced during that episode whilst radioactive decay continues. Due to their long half-lives, ¹⁰Be and ³⁶Cl experience only negligible decay while the concentration of ¹⁴C drops significantly because of its comparably short half-life. Apparently too young ¹⁴C ages (compared to ¹⁰Be and ³⁶Cl) therefore characterize burial (Fig. 3b). (c) Additionally, the samples could be affected by an erosion rate of 0.1 mm/a during the period of burial (‘subglacial erosion’, dotted lines in Fig. 2). Due to the above-mentioned subsurface production peak the concentration of previously accumulated ³⁶Cl slightly increases in the first decimeters below the rock surface (Fig. 1). Erosion causes the rock surface to approach the peak. Hence the concentrations of ¹⁰Be and ¹⁴C decrease more rapidly than ³⁶Cl concentrations and apparently too old ³⁶Cl ages (compared to ¹⁰Be and ¹⁴C) indicate erosion. Conflicting apparent exposure ages with t₃₆ > t₁₀ > t₁₄ therefore characterize the burial and erosion scenario (Fig. 3c).

Fig. 2

Modelled evolution of ¹⁰Be (red), ¹⁴C (blue) and ³⁶Cl (green) nuclide concentrations in sample Grub2 for three different scenarios. a Solid constant exposure for 9.6 ka. b Dashed 9.6 ka exposure followed by 1.0 ka burial. c Dotted same as b, but total erosion of 10 cm during the period of burial, equivalent to a subglacial erosion rate of 0.1 mm/a

Fig. 3

Apparent exposure ages calculated from the final nuclide concentrations of the scenarios described in Fig. 2. The constant exposure scenario (a) yields identical exposure ages from each nuclide. The burial scenarios with or without erosion (b, c) are characterized by systematic deviations in apparent ages: compared to ¹⁰Be, ‘too low’ ¹⁴C ages indicate burial periods, ‘too high’ ³⁶Cl ages point to erosion of the rock surface

In the following we present a pilot study combining same-sample ¹⁰Be, ¹⁴C and ³⁶Cl analyses. In a simple deglaciation setting in a well-mapped study area, the aim is to gain knowledge on the dimensions of a cirque glacier in the Alps throughout the Holocene. How long was the glacier as extensive as during the LIA, or even more extensive? How deeply did the glacier erode the bedrock? The multi-nuclide approach was chosen as a tool with the theoretical capability to address these questions. Parts of the proposed method have been employed in several studies to detect either burial or erosion (e.g. Fabel et al. 2004; Goehring et al. 2011), but never before in this combination to constrain both simultaneously.

2 Setting

Gruebengletscher is a relatively small cirque glacier in a tributary valley to the Haslital close to Grimsel Pass in the central part of the Swiss Alps (Fig. 4). The surrounding area is characterized by widespread granite lithologies, in our study area mostly Central Aare granite crosscut by a band of metamorphic units (‘Altkristallin’; Abrecht 1994). These weathering resistant rocks display well-preserved glacial landscape elements such as U-shaped trough valleys and roche moutonnées. During the Last Glacial Maximum (LGM) the Grueben cirque was filled with more than 500 m of ice that was connected to the Aare Glacier, which extended out into the Swiss Alpine Foreland (Florineth and Schlüchter 1998; Bini et al. 2009; Wirsig et al. 2016). Due to its high elevation above 1700 m it can be assumed that it was occupied by glaciers throughout the Lateglacial period until the end of the Younger Dryas cold spell at 11.5 ka (Maisch et al. 1999). Glacier sizes during most of the Holocene were probably confined to within the well-documented maximum extent of the Little Ice Age (LIA) glacier re-advance (Swisstopo 1865; Maisch et al. 1999; Ivy-Ochs et al. 2009), even though a few moraine ridges are preserved outside of this perimeter (see Figs. 4, 5). Bedrock that was covered by ice during the LIA is clearly discernible based on its fresh appearance in comparison to bedrock outside this extent that is significantly more weathered and covered with lichen. Since deglaciation, slope processes such as rock falls and debris flows constituted a major impact on the Grueben valley. Kame terraces on the NW-slope above Lake Grueben indicate the approximate ice levels in the first half of the twentieth century (see Fig. 4; Kämpfer 2012; Kämpfer and Hählen 2014).

Fig. 4

Map of the study area including the four sampling sites (modified from Kämpfer (2012)) with a hill-shade model in the background [reproduced by permission of swisstopo (JA100120)]. The red square in the inset in the upper right corner indicates the location of the area within Switzerland

Fig. 5

3D-model of the study area including the four sampling sites and the footprint of Grueben glacier at ~1850 (Swisstopo 1865). Grub2-4 were covered by Grueben glacier during the LIA, Grub1 not. The model is based on swissALTI3D DEM and Landsat images used with permission of swisstopo (JA100120)

Four samples were collected from the bedrock ridge that rises almost vertically for more than 100 m to the north of Lake Grueben (Fig. 5). Due to its orientation parallel to the general valley slope, Holocene glaciers were directed down-valley by this wall of bedrock. Presumably, only at the largest Holocene glacier extents, such as the LIA, did significant amounts of ice cross over this ridge. The bedrock samples were taken along the ridge crest, one outside (Grub1) and three inside (Grub2-4) the maximum LIA extent forming a transect with decreasing elevation from 2465 to 2374 m. The sampled bedrock within the LIA extent is smooth and glacially polished. The conserved glacial polish on the rock surface indicates negligible erosion in the absence of ice. Outside of the LIA, the ridge also exhibits clear signs of glacial moulding. However, a surface roughness of approximately 1 cm (i.e. protruding quartz grains) indicates a somewhat higher degree of weathering.

3 Methods

Less than 3 cm thick pieces of bedrock were gathered using hammer and chisel. In order to minimize the influence of temporary cover by sediment or snow, samples were collected at carefully chosen spots on the top of prominent roche moutonnée features. Locations on the abrasion side were preferred and areas affected by plucking avoided. Grub4 originates not from the highest point of a roche moutonnée, but near to a small depression.

The material was crushed and sieved to grain sizes of <800 μm for Be and C and 88 μm < grain size < 500 μm for Cl aliquots. Be sample preparation from quartz mineral separates is based on the method of Kohl and Nishiizumi (1992) with some modifications (Ivy-Ochs et al. 2006). The ¹⁰Be/⁹Be ratio was measured on the 6 MV TANDEM AMS facility (Kubik and Christl 2010; Christl et al. 2013) at the Laboratory of Ion Beam Physics (LIP) at ETH Zürich relative to the 07KNSTD standard (Nishiizumi et al. 2007). The reported values are corrected for a long-term average full process blank ratio of (3.6 ± 2.6) × 10^{−15 10}Be/⁹Be. In situ ¹⁴C was extracted from ~5 g of quartz mineral separates following the procedures outlined in Hippe et al. (2009, 2013). Samples were then measured at the LIP on a MICADAS AMS system equipped with a gas ion source (Fahrni et al. 2013; Wacker et al. 2013). Data analysis follows Hippe et al. (2013) using a long-term average blank value of (3.0 ± 1.4) × 10^{4 14}C atoms based on 20 measurements in the same year. Chlorine was separated from whole rock (Aare granite) samples following the methods published by Zreda (1994) and Stone et al. (1996) with modifications by Ivy-Ochs (1996) using a ³⁵Cl-enriched spike provided by Oak Ridge National Laboratory for isotope dilution (Ivy-Ochs et al. 2004). ³⁶Cl AMS and stable Cl measurements were conducted at the LIP TANDEM system at ETH Zürich relative to the internal K382/4 N standard (Christl et al. 2013). A user blank correction of ³⁶Cl/³⁵Cl = 3.4 × 10⁻¹⁵ was applied. The chemical composition was determined by SGS in Ontario, Canada with XRF and ICP techniques.

The CRONUS-EARTH online calculator (Balco et al. 2008) was used to calculate ¹⁰Be exposure ages with a spallogenic production rate at SLHL of 3.93 ± 0.19 at/g/a (‘NENA’, ‘St’ scaling; Balco et al. 2009). For in situ ¹⁴C, ages are based on a spallogenic production rate at SLHL of 12.3 ± 0.5 at/g/a (‘St’ scaling; Young et al. 2014). Contributions from muogenic production pathways were modelled in MATLAB with the code from Balco et al. (2008) based on the methods of Heisinger et al. (2002a, b). ³⁶Cl exposure ages were calculated by a MATLAB model implementing the equations and constants presented in Alfimov and Ivy-Ochs (2009), in particular spallogenic production rates (without muons) at SLHL of 48.8 ± 3.4 at/g/a for Ca (Stone et al. 1996) and 162 ± 24 at/g/a for K (Evans et al. 1997). These values agree well within errors with recently published production rates (Borchers et al. 2016; Marrero et al. 2016).

The stated age uncertainties (1σ) include AMS standard reproducibility, counting statistics, standard mean error of samples and the propagation of input uncertainties, in particular of the production rate and the ‘St’ scaling scheme of Lal (1991) and Stone (2000). For in situ ¹⁴C, the reproducibility of the extraction system contributes an additional uncertainty of 5 %. ‘Apparent ages’ are the theoretical ages calculated from nuclide concentrations in a simple exposure scenario with no erosion. We use them here as a more accessible representation of nuclide concentrations. They do not define a true exposure age of a rock surface, as long as its burial and erosion history are unclear.

4 Results

Sample information and measured cosmogenic nuclide concentrations are listed in Table 1. All three nuclides were successfully measured in Grub2-4. Grub1 yielded results for ¹⁰Be as well as ¹⁴C, but not ³⁶Cl due to lack of sufficient sample material. Additional data on the ¹⁴C measurements are reported in Table 2. The chemical compositions of Grub2-4 are presented in Table 3. Natural chlorine is present in all samples at relatively low levels of 21.6–31.7 ppm. Finally, Table 4 summarizes calculated apparent exposure ages and uncertainties at the 1σ level. The obtained apparent exposure ages range from 7.4 ± 0.4 to 10.6 ± 0.8 ka. They are graphically summarized in Fig. 6. For each individual sample, the ages derived from different nuclides agree within the 2σ uncertainty range. The average apparent exposure ages (arithmetic mean) of Grub1, Grub2 and Grub3 are indistinguishable at 9.5–9.9 ka with ~0.7 ka uncertainty (1σ). Grub4 appears to be slightly younger with a mean apparent age of 8.1 ± 0.5 ka.

Table 1 List of samples, basic parameters and measured cosmogenic nuclide concentrations

Table 2 Additional ¹⁴C data

Table 3 Elemental composition of samples analyzed for ³⁶Cl

Table 4 Calculated apparent surface exposure ages and uncertainties (1σ)

Fig. 6

Comparison of apparent surface exposure ages and uncertainties (1σ) calculated from ¹⁰Be, ¹⁴C and ³⁶Cl concentrations. The dashed grey line and surrounding area mark the mean age and uncertainty of samples Grub1-3 of 9.6 ± 0.3 and of 8.1 ± 0.5 ka for Grub4, respectively. All nuclide ages agree with the corresponding mean age when the 2σ-uncertainty interval is considered

If the bedrock surfaces are assumed to have been covered by 100 cm of seasonal snow (0.3 g/cm³) for 6 months per year (Hippe et al. 2014), the mean ages increase to 10.8–11.2 ka for Grub1-3 and 9.2 ± 0.5 ka for Grub4 (Gosse and Phillips 2001). The discussion is based on the apparent ages without snow shielding, as the average amount of seasonal snow cover during the entire Holocene is not known. Snow cover does not affect the conclusions drawn here.

5 Discussion

The analysis of multiple cosmogenic nuclide concentrations enables us to understand exposure duration, cover duration, and erosion depth of the sampled rock surfaces. The key to the interpretation is in the relation between the apparent ages derived from the different nuclides. As illustrated above, the measured nuclide ratios characterize a certain type of exposure history (Figs. 2, 3). The production systematics and decay constants of the individual nuclides allow systematic deviations only in one direction, however: same-sample apparent ¹⁴C ages can only be younger than or equal to apparent ¹⁰Be or ³⁶Cl ages. Likewise, given the elemental composition of our samples, apparent ³⁶Cl ages can only be older than or equal to apparent ¹⁰Be or ¹⁴C ages. In this study, however, the different nuclides appear in a random order, i.e. in some cases ¹⁰Be ages are slightly older than ¹⁴C or ³⁶Cl ages, sometimes younger. Since for each sample the apparent ages of the three analyzed nuclides agree with that sample’s mean age within 2σ uncertainty (seven of eleven ages within 1σ), this strongly indicates that there is no systematic deviation from concordant nuclide ages in our data set. The spread in individual nuclide ages thus represents the expected statistical distribution given the uncertainty of individual measurements and age calculations. The concordant same-sample ages imply that the rock surfaces were constantly exposed (see Figs. 2, 3). The influence of burial and erosion was too small to manifest in offsets of cosmogenic nuclide concentrations. In this simple exposure setting, the calculated ‘apparent ages’ can therefore be considered as true exposure ages of the rock surfaces.

Simple exposure and/or known inheritance is one of the major assumptions of surface exposure dating. Here, the combination of different cosmogenic nuclides allows us to test this assumption and eventually infer that this is actually the case. Essentially, they enable us to formulate constraints: how long could a cover period be without being detected?—i.e., what is the minimum duration of cover that leads to significant differences in apparent ages of the individual nuclides? Likewise, how deep can the rock surfaces have been eroded before we expect to detect it? Numerical modelling of scenarios with varying exposure histories provides insight into the sensitivity of the method. The influence of the duration of burial on apparent ages is investigated in Fig. 7a. It shows apparent ages that result from 9.6 ka exposure followed by varying durations of burial (glacial cover) as predicted by hundreds of model runs. The longer the cover, the more the apparent ¹⁴C ages deviate from apparent ¹⁰Be and ³⁶Cl ages. The arithmetic mean of errors of individual nuclide ages is ~6 % (1σ). This propagates to ~8.5 % uncertainty on age ratios. According to Fig. 7a apparent ¹⁴C ages would deviate from ¹⁰Be and ³⁶Cl significantly (>8.5 %) if the rock surface had been covered for more than ~500 years.

Fig. 7

Sensitivity to burial and erosion. Apparent exposure ages of Grub2 calculated for two sets of scenarios. Concentration uncertainties of 6 %, resulting in 8.5 % uncertainties on nuclide ratios, are used for sensitivity statements. A 9.6 ka of exposure followed by a burial period of t_b = 0–1500 years. t_b = 1000 years is identical to scenario B in Figs. 2 and 3. If burial lasts longer than ~500 years, in situ ¹⁴C is expected to yield significantly lower apparent ages than ¹⁰Be or ³⁶Cl. B 9.6 ka of exposure followed by a burial period of 500 years. While buried, the surface is eroded by 0–15 cm by glacial erosion. The starting point at erosion depth = 0 is the same as t_b = 500 years in panel A. If more than ~6 cm are eroded, ³⁶Cl yields significantly higher apparent ages than ¹⁰Be

The sensitivity of apparent ages to the depth of erosion is portrayed in Fig. 7b. It shows apparent ages that result from 9.6 ka exposure followed by 500 years of glacial cover. The erosion depth occurring during the cover is varied from 0 to 15 cm. The deeper the erosion, the more ³⁶Cl ages appear older than ¹⁰Be ages. Following the same reasoning as above, our uncertainties allow detection of age deviations if they are greater than ~8.5 %. The minimum erosion depth we expect to detect therefore is ~6 cm (Fig. 7b). This converts to a subglacial erosion rate of 0.12 mm/a, if the site had been covered for e.g. 500 years. Taken together, the concordant ages of the three nuclides on each sample imply that our samples were covered for less than 500 years and eroded by less than 6 cm. There was effectively no erosion and only very brief ice cover at the lateral margin of the LIA Grueben glacier.

These conclusions are reinforced when the sample locations relative to the LIA ice margin are taken into consideration (Fig. 5). Despite the fact that Grub2 and Grub3 were covered by ice during the LIA advance, while Grub1 was not, the three samples yield indistinguishable mean ages. This leads to two main implications. First, during the Holocene, Grueben glacier grew to dimensions similar to the maximum LIA extent only briefly, i.e. too short for the ice cover to cause younger ¹⁴C ages in Grub2 and Grub3. Assuming that no nuclides were inherited from exposure periods before the Younger Dryas (e.g. Ivy-Ochs et al. 2007), then Grueben glacier was smaller than its maximum LIA extent for more than 9.6 ± 0.3 ka during the Holocene. This is in excellent agreement with data from other sites that suggest that glaciers in the Alps are considered to have been smaller than today for most of the Holocene (Holzhauser et al. 2005; Joerin et al. 2006; Ivy-Ochs et al. 2009; Goehring et al. 2011). Second, the total erosion of bedrock at the lateral ice margin of the LIA Grueben glacier was negligible. Otherwise we would have observed apparently younger ages inside than outside the LIA glacier footprint. Even though the fresh glacial polish indicates that the bedrock within the LIA perimeter was moulded by the glacier, it must have been only thinly abraded rather than deeply eroded. The ubiquitous glacial erosional features encountered on the ridge were therefore formed earlier in time and not destroyed since. Similar results were first obtained by Fabel et al. (2004), demonstrating that glacial scouring near the lateral ice margins in trough valley cross sections was low despite the presence of striations. Both implications outlined here support the conclusions drawn from the multi-nuclide analysis.

The age of sample Grub4 (8.1 ± 0.5 ka) appears to be younger than the other samples, which can be explained by three scenarios. First, it could be assumed that the sample spot was exposed roughly 1.5 ka later than the others, after prolonged cover by ice or till deposited at this site by the Egesen stadial glacier, and continuously exposed since. This would result in concordant ¹⁰Be, in situ ¹⁴C and ³⁶Cl ages, as is actually the case when the 2σ-uncertainty interval is considered. Second, constant exposure for 9.6 ka like Grub1-3, but covered by ~150 cm of snow for 6 months per year would result in an apparent age of 8.1 ka for Grub4. Since snow cover also enhances the neutron flux in the rock surface (Dunai et al. 2014), ³⁶Cl ages would appear ‘too old’. The comparably flat sample surface close to a local depression could serve as an argument to support these two scenarios, even though the difference in snow depth in comparison to the other samples appears rather large. However, the relative distribution of ages with t₃₆ > t₁₀ > t₁₄ hints at the third possible interpretation: a burial and erosion scenario as illustrated in Fig. 2 (dotted lines) and Fig. 3c. Because of its lower elevation, i.e. its position ‘more inside the ice’, Grub4 was likely covered by glacier advances more frequently during the Holocene than the other sample spots that are almost at the limit (Grub2,3) or even beyond (Grub1) the maximum LIA extent. Consequently, the difference in apparent exposure ages could be interpreted to record the integrated time that Grueben glacier was big enough to cover Grub4, but not Grub1-3. Yet, compared to the ¹⁰Be age from Grub4, the in situ ¹⁴C age does not appear significantly younger and the ³⁶Cl age does not appear significantly older. As argued above, this implies that glacial coverage of Grub4 was shorter than 500 years and subglacial erosion less than 6 cm. Therefore, only a part of the ~1500 years age difference could be explained by late burial and erosion. Conceivably, the nuclide concentrations measured in Grub4 could result from a combination of the three scenarios described here: delayed initial exposure, snow cover and burial + erosion caused by Grueben glacier.

We expect that Grueben glacier was big enough to cover and erode the bedrock at lower elevations and proximity to Lake Grueben more often during the Holocene. This should lead to generally younger apparent ages if our transect were extended downhill. Additionally, the mismatch between apparent ages calculated from ¹⁰Be, in situ ¹⁴C and ³⁶Cl would increase. Similar observations were initially made along a profile on proglacial bedrock at the nearby Rhône glacier with the ¹⁰Be and in situ ¹⁴C nuclide pair (Goehring et al. 2011).

6 Conclusions

Two independent lines of arguments were provided to address the major objective of this study; to gain information on the size of and erosion caused by an Alpine cirque glacier during the Holocene. First, concordant apparent ages for individual samples acquired by cosmogenic ¹⁰Be, in situ ¹⁴C and ³⁶Cl analysis imply a simple exposure history. This could be quantified by deducing that the rock surfaces were covered for less than 500 years, otherwise apparent in situ ¹⁴C ages would be younger than apparent ¹⁰Be and ³⁶Cl ages. Because ³⁶Cl yields identical apparent ages that are not older than ¹⁰Be and in situ ¹⁴C, total erosion was calculated to be less than 6 cm.

Secondly, samples Grub1-3 yield indistinguishable mean ages of 9.5–9.9 ka. Since exposure ages outside (Grub1) and within (Grub2,3) the LIA maximum glacier extent agree, Grueben glacier was rarely big enough to cover Grub2 and Grub3 and did not cause sufficient erosion at its lateral ice margin to significantly impact the nuclide inventory in the rock. This conforms well to the results of the multi-nuclide analysis.

Grub4 appears ~1.5 ka younger than the other samples: 8.1 ± 0.5 ka. The difference could be caused by a combination of (a) delayed initial exposure, (b) snow cover and/or (c) burial + erosion caused by Grueben glacier. Future studies are planned to investigate if cover duration and erosion depth increase towards the center of the glacial trough (Fabel et al. 2004; Goehring et al. 2011). This is a likely possibility because Holocene glacier advances reached these positions more frequently and covered them by thicker ice than at the lateral ice margins.

Single-nuclide studies rely on the assumption of a simple exposure history and a priori estimation of local erosion rates to calculate surface exposure ages. We showed that the combination of cosmogenic ¹⁰Be, in situ ¹⁴C and ³⁶Cl proves to be a valuable tool to avoid or to test these assumptions. In addition to the determination of exposure ages for the sampled rock surfaces, the multi-nuclide approach enabled us to place significant constraints on the duration of temporary burial (e.g. underneath a glacier) and depth of erosion caused by the glacier.