Effects of sediment mixing on 10Be concentrations in the Zielbach catchment, central-eastern Italian Alps

Quaternary Geochronology 19 (2014) 148e162

Contents lists available at SciVerse ScienceDirect

Quaternary Geochronology journal homepage: www.elsevier.com/locate/quageo

Research paper

Effects of sediment mixing on 10Be concentrations in the Zielbach catchment, central-eastern Italian Alps S. Savi a, *, K. Norton a, b, V. Picotti c, F. Brardinoni d, N. Akçar a, P.W. Kubik e, R. Delunel a, F. Schlunegger a a

University of Bern, Institute of Geology, Baltzersstrasse 1þ3, CH-3012 Bern, Switzerland Victoria University of Wellington, School of Geography, Environment and Earth Science, PO Box 600, 6140 Wellington, New Zealand University of Bologna, Department of Earth Science, Geology and Environment, Via Zamboni 67, 40127 Bologna, Italy d University of Milano-Bicocca, Department of Geological Sciences and Geotechnologies, Piazza della Scienza 4, 20126 Milano, Italy e ETH Zürich, Labor f. Ionenstrahlphysik (LIP), Schafmattstrasse 20, 8093 Zürich, Switzerland b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 February 2012 Received in revised form 18 January 2013 Accepted 21 January 2013 Available online 4 February 2013

Basin-wide erosion rates can be determined through the analysis of in situ-produced cosmogenic nuclides. In transient landscapes, and particularly in mountain catchments, erosion and transport processes are often highly variable and consequently the calculated erosion rates can be biased. This can be due to sediment pulses and poor mixing of sediment in the stream channels. The mixing of alluvial sediment is one of the principle conditions that need to be verified in order to have reliable results. In this paper we perform a field-based test of the extent of sediment mixing for a w42 km2 catchment in the Alps using concentrations of river-born 10Be. We use this technique to assess the mechanisms and the spatiotemporal scales for the mixing of sediment derived from hillslopes and tributary channels. The results show that sediment provenance and transport, and mixing processes have a substantial impact on the 10 Be concentrations downstream of the confluence between streams and tributary channels. We also illustrate that the extent of mixing significantly depends on: the sizes of the catchments involved, the magnitude of the sediment delivery processes, the downstream distance of a sample site after a confluence, and the time since the event occurred. In particular, continuous soil creep and shallow landsliding supply high 10Be concentration material from the hillslope, congruently increasing the 10Be concentrations in the alluvial sediment. Contrariwise, a high frequency of mass-wasting processes or the occurrence of sporadic but large-magnitude events results in the supply of low-concentration sediment that lowers the cosmogenic nuclide concentration in the channels. The predominance of mass-wasting processes in a catchment can cause a strong bias in detrital cosmogenic nuclide concentrations, and therefore calculated erosion rates may be significantly over- or underestimated. Accordingly, it is important to sample as close as possible to the return-period of large-size sediment input events. This will lead to an erosion rate representative of the “mass-wasting signal” in case of generally high-frequency events, or the “background signal” when the event is sporadic. Our results suggest that a careful consideration of the extent of mixing of alluvial sediment is of primary importance for the correct estimation of 10Be-based erosion rates in mountain catchments, and likewise, that erosion rates have to be interpreted cautiously when the mixing conditions are unknown or mixing has not been achieved. Ó 2013 Elsevier B.V. All rights reserved.

Keywords: Spatial mixing Temporal mixing Mass-wasting processes Basin-wide erosion rate Cosmogenic nuclides Südtirol

1. Introduction In the past decade, the use of cosmogenic nuclides for estimating basin-wide erosion rates has experienced a large increase thanks to the wide applicability of this technique and the continuously increasing precision of calculated erosion rates. Basinwide erosion rates are generally determined through the analysis * Corresponding author. Tel.: þ41 (0) 31 6318772; fax: þ41 (0) 31 6314843. E-mail address: [email protected] (S. Savi). 1871-1014/$ e see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.quageo.2013.01.006

of in situ-produced cosmogenic nuclides (Brown et al., 1995) based on the idea that samples taken at the outlet of a catchment are representative of the entire upstream basin (Granger et al., 1996; von Blanckenburg, 2005; Dunai, 2010). Quartz grains on hillslopes are characterized by a specific nuclide concentration, which is proportional to the time they spent in the near surface and inversely proportional to the erosion rate (Lal, 1991; Gosse and Phillips, 2001). After detachment, the grains will be transported to the channels, mixed by water, and carried to the basin outlet, where they are considered to represent the natural variation of the entire

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

catchment area (Brown et al., 1995; Granger et al., 1996). However, this is only true when a certain number of assumptions are verified (see review by von Blanckenburg, 2005). The fundamental assumption is that denudation rates are uniform through space and time; i.e. the catchment is in cosmogenic steady-state. If this is the case, then the in-going nuclide flux through production is equal to the out-going flux through decay and erosion, and the system is in isotopic equilibrium (Bierman and Steig, 1996). A related assumption, which becomes more important in regions with highly variable process rates, is that the sub-catchments deliver quartz in proportion to their erosion rates. In a geomorphic sense, this requires that the sediment flux downstream of a confluence is equal to the sum of the upstream fluxes, as has been exemplified for the eastern side of the Andes (Wittmann et al., 2009). These two assumptions can be met when each of the surfaces and the tributaries deliver sediment in proportion with their supply area and longterm erosion rates (Granger et al., 1996; von Blanckenburg, 2005). When these assumptions are not met, the catchment is not in geomorphic equilibrium and the use of the cosmogenic technique in the calculation of basin-wide erosion rates becomes problematic (von Blanckenburg, 2005). In transient landscapes, catchments can be subjected to sporadic mass-wasting processes so that sediment delivery to streams becomes highly irregular (Bierman and Steig, 1996; Belmont et al., 2007; Palumbo et al., 2011; Vassallo et al., 2011). When this occurs, the nuclide concentration can only be interpreted as an erosion rate after a detailed geomorphic analysis of the basin, since its value will be biased toward a specific sub-catchment depending on the mobilized volume of sediment (Granger et al., 1996; Binnie et al., 2006; Yanites et al., 2009; Van den Berg et al., 2012). In this case, a sample collected at the outlet of the basin cannot be considered representative of the entire catchment since it would be dominated by material from only a fraction of the area. This might result in an over- or under-estimation of the erosion rate (Bierman and Steig, 1996; Granger et al., 1996; Niemi et al., 2005; Binnie et al., 2006; Yanites et al., 2009). According to these authors, there are two key-variables to take into account when working in transient landscapes: the size of the catchment and the extent to which the sediments are mixed. Large catchments (e.g. >100 km2) mitigate the effects of landslides and sporadic events, which can be distributed through the system, minimizing the repercussion on nuclide concentrations in the alluvial sediment (Clapp et al., 2002; Matmon et al., 2003; von Blanckenburg, 2005; Wittmann et al., 2007; Yanites et al., 2009). Accordingly, when the catchment is small (e.g. <50e70 km2) the mass-wasting effects cannot be neglected and the system has to be analyzed with more detail. It then becomes important that sediment is completely mixed in order for a sample to be representative of the system (Bierman and Steig, 1996; Binnie et al., 2006). Despite widespread assumptions of well-mixed samples, there are only few studies that analyze perturbations in mountain channels in detail (e.g., Kober et al., 2012). Moreover, the number of studies decreases further if we consider the size of the catchments, since most of these focus on basins with generally wellmixed sediment (Clapp et al., 2002; Matmon et al., 2003; Wittmann et al., 2009), and only one study focuses on sediment mixing in small-size catchments (Binnie et al., 2006). These authors showed that complete mixing typically occurs over w10 m distance downstream from a junction in the steady state San Bernardino Mountains, USA. However, it is unclear whether such a simple relationship will also be valid in a more complex environment. Here, we bring a contribution toward the understanding of sediment mixing effects on in-situ cosmogenic nuclide concentrations in alluvial sediment, within the framework of a transient

149

landscape. We proceed by following the methodology proposed by Binnie et al. (2006), and present the results from a sediment mixing model for a w33.5 km2 catchment in the central-eastern Italian Alps. In contrast to the study of Binnie et al. (2006), we explore the extent of mixing in a geomorphic transient catchment where masswasting processes (i.e., debris flows) are frequent. We also sample soil in order to have a reference background value and to better understand the dynamics between production and transfer of sediment. Moreover, we test the extent of sediment mixing downstream from a junction between two tributaries, sampling alluvial sediment at multiple locations along the basin’s trunk stream. The particular geomorphic setting of the catchment additionally allows evaluating the effects of different processes (shallow landslides and soil creep vs. debris flows) on 10Be concentrations in alluvial sediment. In this paper, we specifically address the limits to the applicability of the cosmogenic nuclide method in Alpine catchments and discuss the spatial and temporal repercussions of sediment supply through mass-wasting events on cosmogenic nuclide concentrations in streams. 2. Setting and study site The Zielbach catchment (33.5 km2) is located in the Südtirol region of the central-eastern Italian Alps (Jarman et al., 2011) (Fig. 1). Elevation ranges between 505 m a.s.l. (at the confluence with the Adige River) and 3337 m a.s.l. (Mt. Rosso). The basin is underlain by the Upper Austroalpine metamorphic basement (Schmid et al., 2004), where the main lithologies are ortho- and para-gneisses, mica schists, quartzites, amphibolites and marbles. From a geomorphic point of view, it can be divided in two sectors: (i) the w25.5 km2 main catchment where the Zielbach channel is characterized by steep bedrock reaches forming several metershigh cascades, tens of meters-long step-pool segments, and multiple tributaries along its descent; and (ii) the w2.3 km2 eastern tributary catchment, where a network of debris-flow channels is perched on a deep-seated sackung (here termed the debris-flow catchment) (Fig. 2). In the main catchment, the geometry of the channel network is strongly controlled by the litho-tectonic fabric and channels follow major faults and foliations. The Zielbach trunk stream originates at an altitude of ca. 20 800 m a.s.l. and reaches the Adige River at an elevation of ca. 505 m a.s.l., after building a large fan (Fig. 1). Several knick-zones are recognizable in the landscape along the main longitudinal profile. The lowest-knick zone is the most prominent and hosts one of the highest waterfalls of the region, with a vertical drop of 97 m. These knick-zones are generally associated with lithologic and structural features or morphological steps related to glacial sculpting and represent high “steps” in the channel long profile (Fig. 3). Overall, the morphology of the main catchment is characterized by steep headwaters and U-shaped tributary valleys where deposits of different origins are visible in the landscape. In the headwaters, till-mantled bedrock predominates, and rock glaciers and rock fall deposits have been deposited behind transverse topographic ridges that follow tectonic structures. These ridges have retained hillslope-derived sediment in semi-closed sedimentary traps, thus inhibiting the direct transfer of sediment to the channel network. Shallow (<1 m thick) and small (<10’000 m2) superficial landslides locally affect the Quaternary till cover. Areas occupied by loose material (as rock fall deposits, rock glaciers, moraines, scree and talus slope deposits) represent potential sediment sources, but they are generally disconnected from the Zielbach channel network (Figs. 1 and 2). Above 1500e1700 m a.s.l., soils are developed on glacial till, while below this altitude, agricultural activity is intense and grassland and orchards predominate. Soils have a variable thickness, ranging between 30 and 80 cm

150

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Fig. 1. Location and Quaternary deposits of the Zielbach catchment. The map shows the subcatchments (dotted line) and the tributaries analyzed. Sampling locations: filled diamonds represent the Zielbach main stream deposits, empty diamonds represent tributary samples, the two debris-flow samples are marked with filled circles, and the soils are marked with filled squares. Light color indicates loose material, dark color refers to residual till cover (above 1500e1700 m) or soil mantled areas (below these altitudes). Areas with no color represent bedrock. The location of S01 in this paper corresponds to the sample site of Norton et al., (2011).

(minimum), with less than 10 cm of organic horizon and welldeveloped major stratification. The tree line is located at about 2’000 m a.s.l. and occupies the lower part of the catchment only. The eastern debris flow catchment is characterized by steep bedrock headwaters (w50 on average) that have supplied material through rock fall events, and by a large amount of unconsolidated sediment stored in scree and in colluvial deposits within the channels (several meters thick), available to be transferred through debris flows to the fan and the Adige River (Fig. 2C). The high steepness of this slope creates a direct connection between the

sediment sources in the headwaters and the channels, facilitating the initiation and the transfer of debris flow events. Indeed, large debris flows and torrential floods have occurred frequently in the past years. Recently, the largest debris-flow events in this catchment occurred in 2008 (70’000 m3) and August 2009 (22’800 m3). The 2008 event destroyed 5 bridges and damaged many apple cultivations on the fan (Bolzano Province, SIRIO-IHR, ED30 database). During the same event, some of the sediment was also supplied by a small tributary of the main catchment (Fig.1, tributary 13), as fresh unconsolidated material at the tributary bottom is

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

151

Fig. 2. Photographs of the Zielbach catchment. A) Panoramic view on the eastern side of the basin with visible the tops Höhe Waisse, Lodner and Tschigat; altitudes in parenthesis (taken from Tributary 16). B) Landscape of tributary 16. C) Slope of the debris-flow catchment; the picture is taken from an altitude of ca. 1700 m a.s.l. The two debris-flow channels are partially hidden by trees and morphology. In the small picture (upper right corner of picture C) is shown the deposit within the western debris-flow channel (man as scale e circle).

visible on aerial orthophotos from that year. Debris flows between the beginning of the survey in 1998 and 2008, e.g. the events in 2004, were much smaller and did not exceed a transferred volume of w3,000 m3 (Bolzano Province, SIRIO-IHR, ED30 database). The topography of the entire catchment has been strongly affected by glaciations and glacial erosion. Despite the almost complete absence of active glaciers nowadays (few glaciers remain

in the uppermost part of the catchment), several indicators of past glacial activity can be found in the landscape. U-shaped valley morphologies, till deposits that almost entirely cover the uppermost area of the basin and moraines suggest that the landscape has been strongly glaciated in the past. In addition, several rock-glaciers are present in the major tributary catchments. Penck and Brückner (1909) show that glaciers covered the entire catchment during the

152

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Fig. 3. Sampling location in relationship with stream and catchment topographic profiles. Symbols follow Fig. 1. The ridgeline elevations are shown as black (east) and gray dashed (west) lines.

‘Würm’ period. Field observations, cosmogenic nuclides, and 14C ages also suggest that following the Last Glacial Maximum (LGM), deglaciation in this area of the Alps proceeded between 18,000 and 15,000 yrs B P, with no evidence for glacier re-advance in more recent cold periods (e.g., Younger Dryas; Bargossi et al., 2010; Ivi Ochs et al., 2006). This is supported by Castiglioni (1928) who showed that glacial advances during the Little Ice Age (LIA) were modest in this part of the Alps, when small glaciers formed above 2’700 m a.s.l only. Accordingly, while lower elevations in this catchment have likely been ice-free since the LGM deglaciation, the upper sub-catchments have been subjected to repeated glacial perturbations lasting until the LIA.

the two contributing areas. Binnie et al. (2006) sampled at a fixed distance of w10 m from the junction, while here we have sampled two points at different downstream distances from the junction, ranging between w100 m and 3 km (depending on the best achievable sample location). We did not collect further samples from all smaller tributaries between the tributary junctions and the trunk-stream samples (exceptions were the sites S06 and S08, where S07 enters in between). In this way, we build two sediment mixing models for each junction, to calculate the differences in mixing ratios as the sediment is mixed downstream. This sampling strategy allows us to analyze the spatial and temporal impact of sediment mixing on in-situ cosmogenic nuclide concentrations.

3. Methods

3.2. Measurement of cosmogenic 10Be and erosion rates calculation

3.1. Sampling strategy

The 10Be concentrations were measured in quartz, which was purified following a modified version (Akçar, 2006) of the one described in Kohl and Nishiizumi (1992). The samples were first sieved and only the fraction between 0.25 and 0.5 mm was used in the subsequent steps. The non-magnetic fraction was isolated with a Frantz magnetic separator and then treated with a series of acids using dilute HCl, HF and HNO3. Hydrochloric acid allows the dissolution of carbonates and iron oxides while HF attacks all the other silicate minerals. To accelerate the procedure, the mixture of acids and samples were heated for few hours in an ultrasonic bath at w60  C, and Aqua Regia was used for cleaning the samples of the last acid residue (Akçar, 2006). Subsequently, extraction of the cosmogenic 10Be was performed at the University of Bern (Akçar et al., 2012). Samples were spiked with 0.15e0.18 mg of Be carrier before dissolution (Table 3). The 9Be/10Be ratios were corrected for the measured long-term average blank ratio (3.13  0.12  1015) and measured using accelerator mass spectrometer at the ETH Zurich. The results were normalized to the S2007N standard with a nominal ratio of 28.1  1012 (modified from Kubik and Christl, 2010). Erosion rate calculations were performed following Balco et al. (2008) using CRONUS EARTH calculator (ver. 2.2; http:// hess.ess.washington.edu/) and are scaled for altitude and latitude using the scaling factors of Dunai (2001); with a SLHL production rate of 4.43  0.52 atoms/g/a (Tables 2 and 3). For all the samples the mean elevation of the upstream catchment was used, with

We collected a total of 21 samples for 10Be analysis: 16 were taken from alluvial sediments within the channel network and were used to analyze the effects of sediment mixing, and 4 were collected from soil, yielding information about the background erosion rate on hillslopes (Fig. 1). Alluvial sediment was collected from the surface, while soil samples were taken in the actively mixed top-soil between 10 and 50 cm depth. To investigate effects related to the mixing of sediment, samples were taken from both the Zielbach main channel and its major tributaries. For each tributary, the samples were collected at the outlet of the tributarycatchment, while in the main channel the samples were taken below the junction with the tributaries and above and below the knick-zones. Two samples were collected from the outlet of the debris-flow catchment as it hosts two separate channels. All the samples were collected in sand bars in the corresponding active channel; the location of the samples and their relationship to the stream long-profile are shown in Fig. 3. The degree of sediment mixing was determined following the scheme proposed by Binnie et al. (2006), considering three samples for each junction: one each from the tributary and main stream and one below the junction in the main river (Table 1; Fig. 4). The two debris-flow samples S09 and S10 (Figs. 3 and 4) were considered together, taking the average of their concentrations and summing

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

153

Table 1 Topographic features of the sub-catchments. Altitudes, areas and distances are calculated from a Digital Elevation Model with 2.5 m of resolution. Sample name

Sample site

Sample altitudea [m a.s.l.]

Mean altitudeb [m a.s.l.]

Areac [km2]

Downstream distanced [m]

Distance from the junctione [m]

S01 S02 S03 S06 S07 S08 S09 S10 S11 S13 S14 S15 S16 S18 S19 S20 S500 S501 S502 S504

Main channel e Fan Main channel e Fan Main channel Main channel Tributary Main channel Debris flow channel Debris flow channel Main channel Tributary Main channel Main channel Tributary Main channel Main channel Main channel Soil sample Soil sample Soil sample Soil sample

605 514 1006 1546 1527 1276 1135 1163 1859 1556 1562 2206 2224 2181 1952 2340 2200 2426 2426 2040

2363 2339 2529 2624 2267 2564 1936 2149 2674 2234 2654 2755 2773 2746 2714 2777 e e e e

31.29 31.70 25.53 22.51 0.37 24.54 1.54 0.87 17.91 0.69 21.26 13.57 6.78 14.17 15.37 5.79 e e e e

1250 230 3470 5420 4990 5040 4860 3900 3860 3900 6130 5250 5290 8020 8280 7610 e e e e

1850 2890 1290 125 e 825 e e e e e 150 e 570 e e e e e e

a b c d e

Altitude of the sample point. Mean altitude of the upstream basin relative to sample point. Area of the upstream basin relative to sample point. Downstream distance of the sample from the Zielbach headwaters. Downstream distance of the sample from the closest uppermost junction (only for the samples used in the models).

exception for the soils were the site elevation was used. Soil denudation rates were also calculated using the surface production rate. As pointed out by Granger and Riebe (2007), the nuclide concentration at any depth in a mixed eroding soil is equal to the

surface concentration at cosmogenic steady state. We use a 10Be half-life of 1.39  0.1 My (Chmeleff et al., 2010; Korschinek et al., 2010) and a sample density of 2.7 g/cm2 throughout (2.2 g/cm2 was used for the soils). For every sample, the shielding factor was

Fig. 4. Scheme of the tributaries (black outlines) analyzed in the mixing models (Figs. 5 and 6) at two distances downstream of each junction.

154

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Average snow density of 0.2 g/cm3. The total shielding factor Stot was calculated as

Table 2 Input parameters used for the CRONUS EARTH calculator. The shielding corrections are calculated for topographic and snow shielding as explained in the text (paragraph 3.2). Density used for alluvial sediment is 2.7 g/cm3, for soils is 2.2 g/cm3. Sample name

Lab IDa

Latitude [DD]

Longitude [DD]

Mean altitude [m a.s.l.]

Shielding correction

S01 S02 S03 S06 S07 S08 S09 S10 S11 S13 S14 S15 S16 S18 S19 S20 S500 S501 S502 S504

SPI01 SPI02 SPI03 SPI06 SPI07 SPI08 SPI09 SPI10 SPI11 SPI13 SPI14 SPC15 SPC16 SPC18 SPC19 SPC20 S-500 S-501 S-502 S-504

46.6780 46.6706 46.6932 46.7036 46.7013 46.6951 46.6970 46.6968 46.7103 46.7043 46.7046 46.7182 46.7225 46.7070 46.7191 46.7141 46.7304 46.7198 46.7121 46.7076

11.0660 11.0734 11.0509 11.0391 11.0378 11.0461 11.0513 11.0531 11.0300 11.0368 11.0363 11.0073 11.0226 11.0531 11.0228 11.0310 11.0194 11.0236 11.0268 11.0313

2363 2340 2530 2624 2268 2564 1937 2150 2675 2235 2655 2756 2773 2746 2714 2778 e e e e

0.85 0.86 0.85 0.85 0.84 0.85 0.84 0.84 0.85 0.83 0.85 0.85 0.86 0.85 0.85 0.84 0.87 0.87 0.87 0.87

a

Stot ¼ 1  ðð1  Ss Þ þ ð1  St ÞÞ 3.3. Mixing model

To evaluate the extent of sediment mixing in the main channel, we compared the ingoing sediment flux from the tributaries and the outgoing sediment flux below the junction with the 10Be concentration measured in the channels. Sediment mixing can be calculated in two ways; either by the nuclide balance or sediment fluxes above and below the junction (see Binnie et al., 2006). Based on the nuclide balance, if all sources are accounted for, the downstream concentration must be between the concentrations of the two upstream segments. This can be expressed as:

½10 Bemc ¼ J½10 BeS1 þ K½10 BeS2

(3)

where J and K are less than 1, and J þ K ¼ 1; the subscripts “MC” represents the main channel while “S1” and “S2” represent the two sub-catchments. If 100% of the 10Be signal comes from basin S1, then J ¼ 1 (if 60% is from S1, J ¼ 0.6, etc.). In this way, a mixing relationship based on nuclide concentrations can be determined. The sediment fluxes from each sub-basin can also be calculated from the erosion rate and upstream area:

ID used during the laboratory procedures.

calculated as the sum of topographic shielding (St) and snow shielding (Ss) effects. The topographic shielding was determined from a digital elevation model (DEM) with 2.5 m resolution using a modified version of the Codilean (2006) shielding macro (B. Guralnik, ETH, pers. comm.), while the snow shielding was calculated using an altitude dependent formula determined empirically (Norton, 2008) from snow depth-duration data (Auer, 2003). Since the snow cover duration is highly dependent on altitude and since there is a strong relationship between snow depth and elevation (Norton, 2008), we could use the formula:

Ss ¼ 1*108 z2  8*105 z þ 1:0242

(2)

Qx ¼ εx *Ax

(4)

where Qx is the sediment flux at the point “x”, εx is the erosion rate derived from the 10Be concentrations (Section 3.2) and Ax is the area of the analyzed sub-basin (determined here from the 2.5 m DEM using the hydrological tools in ArcGIS). According to the theory of steady state, if no storage occurs, then QOUT (¼QMC, sample in the main channel below the junction) has to be equal to QIN (coming from the two tributaries) so that we can write:

(1)

where Ss is the snow shielding, z is the average altitude of the subbasin (extrapolated from the DEM) and the numbers are constants.

QMC ¼ QS1 þ QS2 or; QMC ¼ ðεS1 *AS1 Þ þ ðεS2 *AS2 Þ:

(5)

Table 3 Corrected 10Be concentrations and denudation rates calculated with CRONUS EARTH calculator (ver. 2.2; http://hess.ess.washington.edu/; Balco et al., 2008). The denudation rates refer to the scaling factor of Dunai (2001) with spallogenic PSLHL ¼ 4.43  0.52 atoms/g/yr with the ETH standard S2007N. For all the samples a thickness of “0” has been used. 10

Be results:

Sample name

Sample weight [g]

Carrer weight [mg]

10 Be conc. [atoms/g] 103

Error þ/ [atoms/g] 103

Production rate [muons] [atoms/g/a]

Production rate [spallation] [atoms/g/a]

Denudation rate [mm/ka]

Internal uncertainty [mm/ka]

External uncertainty [mm/ka]

S01 S02 S03 S06 S07 S08 S09 S10 S11 S13 S14 S15 S16 S18 S19 S20 S500 S501 S502 S504

51.3437 46.0990 51.2614 39.1442 58.3232 57.4298 45.3242 59.0581 52.5342 51.7709 53.8055 45.2908 45.2037 46.0250 50.5993 45.1733 50.5537 50.5399 41.5100 50.4539

0.1762 0.1752 0.1752 0.1755 0.1586 0.1759 0.1591 0.1760 0.1755 0.1758 0.1756 0.1580 0.1586 0.1578 0.1590 0.1586 0.1589 0.1588 0.1591 0.1588

11.44 13.55 46.83 64.10 45.30 45.84 4.73 7.50 41.94 103.9 42.21 35.19 37.58 32.30 48.12 25.94 187.9 254.3 215.6 129.2

1.19 1.29 2.66 2.86 2.34 2.84 0.80 0.86 2.13 5.75 2.36 1.98 2.18 2.77 2.34 1.82 7.54 8.67 8.19 8.18

0.39 0.39 0.41 0.43 0.38 0.42 0.34 0.37 0.43 0.38 0.43 0.44 0.44 0.44 0.44 0.45 0.44 0.44 0.44 0.43

25.16 24.79 28.05 29.80 23.20 28.67 18.31 21.29 30.87 22.48 30.45 32.50 33.07 32.29 31.66 32.71 33.08 33.82 33.63 31.14

1460 1210 395 307 342 412 2650 1920 484 146 474 606 577 657 433 829 144 109 128 196

160 120 23 14 18 26 490 230 25 8 27 35 34 58 22 60 6 4 5 13

210 170 46 35 39 49 530 290 56 17 56 72 69 89 50 100 16 12 14 24

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

It is important to highlight that QMC is the mathematical sum of the fluxes of material supplied by the tributaries and quantified through the calculated erosion rates (derived from the corrected 10 Be concentrations) and the upstream sizes of catchments’ areas. The relative contribution of a sub-catchment to the total sediment flux is then (Table 4):

QSx % ¼ QSx =QMC *100:

(6)

If the two tributaries do not contribute sediment in proportion to their erosion rates (or are poorly mixed), the nuclide concentration measured below the junction does not represent the weighted mean of the two tributaries. The signal in the main channel will then be biased toward the catchment, which has supplied more material. Following the model proposed by Binnie et al. (2006) we used the 10Be concentrations from the two tributaries as end members for a possible mixing scenario and compared these values with the 10Be concentration below the junction. As described above, this “cosmogenic signature” in the main channel is the product of the mixing between the two tributaries and indicates which one has a “stronger” signal (i.e. which sub-catchment is supplying more sediment). By comparing the signal derived from the 10Be concentrations with the calculated sediment fluxes, we estimate the state of mixing of the sediment in the channel, and the relative contribution of each tributary at two downstream sampling points. 4. Results 4.1. Cosmogenic

10

Be and erosion rates

The results of the in-situ 10Be nuclide concentrations allow the identification of three groups with large differences in 10Be concentrations (Table 3, Fig. 5). Soil and till samples have the highest concentrations, which implies that they have experienced the lowest erosion rate. Samples with the lowest 10Be concentrations, thus the highest erosion rates, are derived from the debris-flow catchment while samples from the Zielbach and its tributaries yield intermediate values. There are a few exceptions in this last group: sample S13 (tributary 13) yields a 10Be concentration that

155

plots within the same order of magnitude of the soil samples, while S01 and S02 have concentration values closer to those recorded for the debris-flow channels (Fig. 5). There is an overall increasing trend in 10Be concentrations in the main channel in the downstream direction down to the junction with the debris-flow catchment, beyond which an abrupt decrease occurs. The 10Be concentrations of the soil samples decrease uniformly with elevation in the downstream direction primarily due to lower production rates at lower altitudes (see Fig. 1 for location of sites, and Fig. 5 for the concentration pattern). The concentrations in Fig. 5 are not scaled to altitude. The scaled 10Be concentrations are generally between 54 and 71% lower if the concentrations are scaled to the SLHL production rate of 4.43  0.52 atoms/g/a, but the observed trends do not change. 4.2. Mixing model The trends described above highlight the tributary basins’ contribution to the total amount of sediment transported by the Zielbach. The clear imprint of the tributaries on the main channel nuclide concentration reveals the importance of considering the degree of sediment mixing. The sediment mixing diagrams (Figs. 6 and 7) are composed of two parts. First, simple nuclide mixing calculations, based on the 10Be concentrations of the upstream segments, give the relative contributions from each tributary necessary to produce the measured downstream nuclide concentration (e.g., equation (3)). For example, the concentration measured at point S15 is mainly influenced by the tributary 16 since the overlapping area corresponds to a range between 44, and 100% of the signal being derived from S16 (Fig. 6-A1). Likewise, the cosmogenic nuclide signal registered at the point S06 (Fig. 6-B1) is mainly influenced by S14 (55e75%) and S08 is dominated by S07 (70e100%). Farther downstream, S01 is clearly biased toward the debris-flow concentrations (77e97%) (Fig. 6-C1 and 7-A1). Second, we compare the mixing ratios determined from the nuclide balances with the calculated sediment fluxes of the two tributaries (e.g., equations (5) and (6)). If the sediment is well mixed, the intersection between the cosmogenic nuclide signals will be identical with the calculated sediment fluxes (Binnie et al.,

Table 4 Parameters used in the mixing models. The samples are reported in the table following a downstream location and they are grouped as in the model. The two tributaries are highlighted in gray and precede the samples below the junction. For example, the model for tributary 16 considers S20 and S16 as input tributaries and S15 and S18 as samples after the junction with the main Zielbach channel. The same is valid for S14 and S13 as input tributaries and S06, S08 as samples after the junction, and so on. * Samples S09 and S10 are considered together in the model. Sample name

Area [km2]

10

Be conc. [atoms/g] 103

Error þ/ [atoms/g] 103

Denudation rate [mm/ka]

External uncertainty [mm/ka]

Qa [m3/yr]

Q%

S20 S16 S15 S18 S19 S11 S14 S13 S06 S08 S06 S07 S08 S03 S09 S10 S03 S01 S02

5.79 6.78 13.57 14.17 15.37 17.91 21.26 0.69 22.51 24.54 22.51 0.37 24.54 25.53 1.54 0.87 25.53 31.29 31.70

25.94 37.58 35.19 32.30 48.12 41.94 42.21 103.90 64.10 45.84 64.10 45.30 45.84 46.83 4.73 7.50 46.83 11.44 13.55

1.82 2.18 1.98 2.77 2.34 2.13 2.36 5.75 2.86 2.84 2.86 2.34 2.84 2.66 0.80 0.86 2.66 1.19 1.29

829 577 606 657 433 484 474 146 307 412 307 342 412 395 2650 1920 395 1460 1210

100 69 72 89 50 56 56 17 35 49 35 39 49 46 530 290 46 210 170

4800 3910 8230 9300 6650 8660 10,080 100 6910 10,120 6910 130 10,120 10,080 4070 1660 10,080 45,750 38,510

55 45 e e e e 99 1 e e 98 2 e e 26 10 64 e e

a

Q represents the sediment flux calculated as explain in the paragraph 3.3, while Q% is used in Figs. 6 and 7.

156

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Fig. 5. 10Be concentrations versus downstream distance. Concentrations of the Zielbach samples show a slightly increasing trend in the downstream direction. The soil sample concentrations decrease in the same direction. Debris flow samples have the lowest concentrations. Filled diamonds indicate the main Zielbach river samples; empty diamonds represent the tributaries samples. Error bars show 1-sigma; concentrations are not scaled for altitude (see text for more discussion).

2006). As shown in Fig. 6 (A1-B1-C1), for the immediate downstream sample, only S16 is well mixed. The sediment flux range overlaps with the 10Be signals between 45 and 55% of S16, expressing roughly an equal contribution from the two tributaries. The sediment is not mixed in the other two cases (Fig. 6-B1 and 6C1) where there is no overlap between the cosmogenic nuclide signal and the calculated sediment flux. Likewise, the material from the debris-flow catchment is poorly mixed with the Zielbach sediments (Fig. 7-A1), as the 10Be signal shows a clear imprint of the debris-flow concentrations (77e97%) while the calculated sediment flux suggests a lower contribution from these tributaries (26e46%). For the tributaries 16 and 13, the situation improves considerably if we consider a sample farther downstream (Fig. 6-A2 and 6-B2). The cosmogenic nuclide signals are entirely matched by the sediment flux calculations, suggesting that the samples are well mixed, with a 50:50 contribution from S16 and S20 for S18 at 570 m downstream of the junction. By 1250 m downstream at S08, S14 and S13 are well mixed, and the sediment is almost entirely derived from S14 (89e100%). Material supplied from S07 remains poorly mixed as does the debris-flow material (Fig. 6-C2 and 7-A2), where the tributary’s nuclide signal remains higher than that of the main Zielbach channel. Norton et al. (2011) also measured the 10Be concentration of the Zielbach main channel on a sample collected at the outlet of the basin in 2006. Their reported concentration, 23.8  3.3 *103 atoms g1, is almost twice as high as those measured here for the same site (S01 and S02). The mixing model of the debris-flow derivedmaterial yields a completely different result when using the value of Norton et al. (2011) (Fig. 7B), with an overlapping between cosmogenic nuclide signal and sediment flux between 41 and 46%.

The pre-debris flow 10Be concentration shows a mixing between 41 and 71% with the 41e46% of sediment derived from the debris-flow tributaries and the 54e59% supplied by the Zielbach in S03, values that are concurrent with the calculated sediment flux. 5. Discussion 5.1.

10

Be concentration in a complex Alpine setting

Taken at face value, these results show the increasing effect of sediment supply from tributaries on the main channel 10Be concentrations. In the main catchment, stream-borne 10Be concentrations average w42 * 103 atoms g1, with a distinct trend toward increasing concentrations in the downstream direction. Geomorphic features, such as knick-zones, do not seem to affect 10Be concentrations, since samples taken upstream and downstream of the knick-zones show similar values, suggesting that sediment has been bypassed over these features. Likewise, samples above a junction do not show any variation in 10Be concentration. Field observations show that the Zielbach main catchment has been mainly affected by local and shallow landslides which only involve the mobilization of unconsolidated regolith. Larger rock falls occur in the upper portions of the basin, but this material tends to be trapped on or behind talus deposits, which are disconnected from the system (e.g., Figs. 1 and 2). While we have no direct measurement of the residence time of this material, we anticipate that it is removed at a rate similar to or slower than that of the soils (w100 mm/kyr). Accordingly, if our inference is correct, this suggests that storage in these traps spans time intervals on the order of 103e104 yrs. This finding, however, awaits the results of further detailed studies.

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

157

Fig. 6. Mixing models for the tributaries in the main Zielbach catchment. The filled sectors are representative of the cosmogenic nuclide signals: the overlapping area between light filled area (proportioned signal coming from the tributaries) and dark filled area (10Be concentration registered below the junction) represents the nuclide concentration mixing of the two tributaries in the main channel. The diagonal lines illustrate the calculated theoretical sediment flux from the tributaries (representing the theoretical true mixing). In between vertical lines (space shown with an arrow) is indicated the 10Be-derived mixing signal. When the 10Be signal overlaps the modeled signal, the sediment is well-mixed and the catchment is in isotopic equilibrium. The numbers between breaks indicate sample distance downstream of the junction for tributary S16 (A1 and A2), tributary S13 (B1 and B2), and tributary S07 (C1 and C2). In each case, the left hand plot is near the junction and the right hand plot is farther from the junction.

Tributaries 16 and 13 add higher concentrations to the system, suggesting predominantly hillslope soil and sediment sources, which tend to have lower erosion rates (Norton et al., 2010). This inference is supported by the high concentrations measured locally in till and soil samples, which show an order-of-magnitude higher concentrations than stream sediment. These circumstances promote the supply of material with higher 10Be concentrations and explain why these two tributaries locally (and temporally) shift the Zielbach 10Be concentration toward higher values. Likewise, we use this mechanism to explain the higher-than-expected concentration of sample S19, as a small, shallow landslide occurred upstream of the sampling location just few days before collection, potentially adding abundant high concentration material. Similarly, we

interpret the particularly high concentration measured in sample S13 to be the result of the 2008 debris-flow event, which resulted in the supply of high-concentration material derived from the lateral hillslopes (Fig. 8). The high-concentration measured in the sample is due to the fact that this latter tributary normally does not deliver material via debris flows, as revealed by the survey of the region. Since the sampling campaign was performed during 2009 and 2010, the collected S13 sample might show the overprinting of the 2008 event, which mainly involved the entrainment and transport of regolith and rock fall-derived clasts deposited in the channel, as revealed by orthophotos and field observations. The sample location, topographically higher than the main Zielbach channel and with no coupling with the lateral earth flow visible in Fig. 8, implies

158

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Fig. 7. Mixing models for the debris-flow catchment in the Zielbach basin. A1 and A2 represent the post-debris flow conditions (after the 2008 event) while B represents the model calculated with the Norton et al. (2011) value (pre-debris flow event). Symbols and explanation follow Fig. 5.

that the material has been directly derived from the higher portions of the tributary. Accordingly, this supply could have introduced abundant material with high 10Be concentrations into the system, since the low activity of tributary 13 implies longer time for exposure of sediment prior to erosion. The case of the tributary 07 is slightly different; since its value is coherent with Zielbach average concentrations. The differences between the effects of tributary 13 and tributary 07 (which have similar areas) on the main Zielbach concentration suggests that catchment size has relatively little importance while the mechanism of sediment release and transfer within a basin has a much higher impact on alluvial sediment concentration. The two samples from the debris-flow catchment also have a clear overprint of the 2008 debris-flow event, but in the opposite direction (input of material with low concentrations). The remarkably lower concentration measured in S09 and S10 could be due to either the high frequency of debris-flow events in this catchment, or to the extraordinarily large size of the 2008 debris flow that scoured the channel bed for several meters in depth. In the first case, high-frequency events would continually strip off the surface, keeping the concentrations low. This scenario does not necessarily contradict the assumption of steady state erosion, as it can be thought of as a chipping process in which a given depth is removed per event (e.g. Granger et al., 1996; Small et al., 1999; Kober et al., 2007). In this case, the erosion rates calculated from the 10Be concentrations are representative of the true long-term rates from the debris-flow catchments. In the second case, where the depth of scour is much greater than during an average event, the 10Be concentrations would be dominated by deep, low concentration material, and they would not be representative of the long-term mean. If we can assume that the material

has been well mixed, then the true long-term erosion rate would be lower than those calculated after the 2008 event.

5.2. Sediment mixing considerations The results of the models suggest a large variability in the degree of sediment mixing. With respect to the main Zielbach channel, we can say that the sediment concentrations from the tributaries are eventually well mixed as sediment is transported downstream. The model suggests that if taken at a proper distance from the junction the sample can be reliable and proportionally representative of the tributaries’ contributions. Sediment from tributaries 16 and 13 is poorly mixed 125e150 m downstream of the junction, but well mixed after additional w500e1000 m. It is clear that in this transient Alpine catchment, sediment-mixing distances are longer than in the largely steady-state San Bernardino Mountains (California-USA) as reported by Binnie et al. (2006). However, because the rates and timing of erosion events in this setting are so variable, it is difficult to define a fixed length that can be used independently for all tributaries and for all basins. Our results illustrate that the supply mechanisms need to be carefully considered when sampling alluvial sediment, because this can highly affect the sediment 10Be concentrations. It is, therefore, difficult to identify a generalized strategy to collect samples in mountain catchments, since the erosional and transport mechanisms have to be detailed individually for each catchment. However, a valid sampling strategy should take into account that a sample collected close to a junction has generally a higher probability to be biased toward one of the two source areas, while this risk decreases when collecting further downstream.

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

159

Fig. 8. Orthophotos of catchment areas taken in 2006, 2008 and 2011 (downloaded from the geobrowser of the Bolzano province e http://www.provincia.bz.it/informatica/temi/ maps-webgis.asp). In 2008, the debris flow effects are visible for both the tributary 13 (upper images-set) and the fan (lower images-set). The comparison between the 2006 and 2008 photos show the different areas covered by sediment deposited during the 2008 event (tributary 13 and debris-flow tributary, with the deposit of the 2008 debris flow spreads on the entire length of the active channel in the fan). In the photos are also represented the major knick zones and the sampling locations. It is evident how the 2008 event affected the channels, with deposits still visible in 2011. In the lower images-set, the location of S01 also corresponds to the sampling location of Norton et al. (2011).

160

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

With respect to sediment mixing, tributary 07 provides insight into the variability in sediment transport in the Zielbach. From Fig. 6-B2, it is clear that the sediment derived from tributaries S14 and S13 are well mixed at point S08. Since S08 also includes the source area for S07, we can infer that either sediment from the tributary 07 is well mixed in S08, or that S07 has the same concentration as the trunk stream (and is therefore invisible to our analysis). Alternatively, it is likewise possible that tributary 07 has no effect on the 10Be concentration in the Zielbach. This can occur if the sediment supplied from the tributary 07 is deposited before reaching the main channel or flushed in pulses out of the system. In either case, S07 does not change the main channel concentration, and S08 and S03 are representative for the mixing of the tributaries above. The results expressed in Fig. 6C, therefore, show that nearly 100% of the sediment at points S08 and S03 is delivered by the trunk stream, which is well mixed with respect to the upper tributaries (Fig. 6-B2). The mixing models from the debris-flow catchment are harder to interpret (Fig. 7). As expected for a debris flow-dominated basin, the samples collected below the junction suggest that the sediment is not well mixed (Fig. 7-A1). Even at almost 3 km downstream of the junction (S02), the sediment is not well mixed and is clearly influenced by the debris flow (Fig. 7-A2). In 2006, Norton et al. (2011) sampled near the site S01, before the large debris-flow event that affected the catchment in 2008. The lower 10Be values of our samples highlight how debris flow activity can affect the 10Be concentration of a basin, perturbing the nuclide concentrations from their spatial and temporal averages (Bierman and Steig, 1996; Granger et al., 1996; Niemi et al., 2005; Binnie et al., 2006; Yanites et al., 2009; Vassallo et al., 2011; Kober et al., 2012). The concentration measured by Norton et al. (2011) is also below the average for the Zielbach channel (23.8*103 atoms g1 versus 42.4*103 atoms3 g1) suggesting that the low concentrations in the trunk stream are persistent, and the debris flows from catchments S09 and S10 are in fact highfrequency events (but with general smaller volumes or eroding capabilities). In support of this interpretation, the model in Fig. 6B shows that the 10Be signal is partially congruent with the calculated sediment flux suggesting almost complete mixing. This either implies that this volume of the most recent debris flow event (w3,000 m3 in 2004) was too small to have substantially perturbed the 10Be concentration of the trunk stream, or that the material derived from that event was well mixed with the Zielbach material within 2 years. The 2008 event, with a much larger volume of mobilized sediment, clearly caused a large perturbation in the 10Be concentration, leading to the question of how far in time from a large event a sample yields reliable nuclide concentrations for erosion rates calculations. 5.3. Spatial and temporal implications for sediment mixing In agreement with previous studies (Bierman and Steig, 1996; Granger et al., 1996; Clapp et al., 2002; Matmon et al., 2003; Binnie et al., 2006; Yanites et al., 2009), the models calculated in this paper show that the sampling location has to be far enough from the junction to avoid the imprint of a particular source area. Likewise, the size of the catchment can also influence the sediment mixing conditions. However, our study suggests that the complete mixing can also be achieved even for relatively small catchments, since the results of the Zielbach main catchment (size of w25.5 km2) show that the sediment is overall well-mixed and that the cosmogenic nuclide signal is congruent with the relative sediment contributions of the different tributaries. Our results highlight, moreover, the importance of knowing the sediment transfer mechanisms and the magnitude of the relevant sediment input/disturbance. In the Zielbach main catchment, the transport of sediment occurs mainly

through soil creep and rare, shallow landslides, allowing the supply of high concentration material into the system; this is confirmed by the higher concentrations of 10Be derived from the tributaries that are affected by these hillslopes processes. Contrariwise, in the debris-flow catchment, the high-frequency of debris-flow events allows the supply of low concentration sediment that results in a lower cosmogenic nuclide signal (Fig. 9). When low-frequency high-magnitude events occur (e.g., 2008 event), the amount of material supplied to the main channel causes a strong overprint on the mixing condition of the entire basin, moving the signal toward the debris-flow catchment concentration. The larger the event, the longer the response time necessary for the basin to recover to predisturbance mixing conditions (Schaller et al., 2004). Accordingly, the return period for events of large magnitude provides another variable that has to be taken into account for the sampling and erosion rate interpretation. We thus support the statement by Yanites et al. (2009) that a high frequency of homogeneous processes does not affect the cosmogenic nuclide in the system, provided that the sediment is spread homogenously through space and time. However, an episodic large event can completely throw the system out of balance over the timescale of the catchment processes, in agreement with the recent findings of Kober et al. (2012). Therefore, it appears obvious that sediment mixing has to be achieved both in space and time. Following this result, a good sampling campaign should take into consideration not only the spatial distance of the sample below a junction, but also the previous history of the catchment in terms of sediment supply and mass wasting disturbance. It is therefore advisable to collect 10Be samples with a sufficient, locally specified, time span after large debris flow events, as seems to be the case for the Norton et al. (2011) sample. 5.4. Denudation rates Catchment-wide denudation rates calculated for the Zielbach basin vary from 146 to 2650 mm/kyr (Table 3). These results are in the same range as those from other parts of the European Alps (Wittmann et al., 2007; Delunel et al., 2010; Norton et al., 2008, 2010; 2011; Kober et al., 2012; Van den Berg et al., 2012). These previous studies focused on basins encompassing the entire range of scales from the whole Alpine chain (Wittmann et al., 2007; Norton et al., 2011) to small catchments (Delunel et al., 2010; Kober et al., 2012; Norton et al., 2008, 2010). Our study is similar to the latter, but shows similar denudation rate variability as the large Alpine-wide studies (e.g., Wittmann et al., 2007). Here, we relate the large variability in denudation rates to the mixing of sediment from various sources. In particular, the mixing diagrams highlight the strong influence that sediment supply mechanisms have on 10 Be concentrations and, subsequently, on erosion rates. The results also allow us to identify the samples with biased erosion rate values (due to the 2008 event), which are related here to the poor mixing of debris-flow derived material. Removing these samples (S06, S13, S09, S10, S01 and S02) greatly reduces the variability of the data. Using only the well-mixed samples, we can observe that in the main Zielbach catchment, the denudation rates range between 342 and 829 mm/kyr and show a downstream decreasing trend (i.e., with decreasing elevations; Fig. 9). The downstream trend of decreasing erosion rates could have two potential causes: increasing glacial impact with increasing altitude, or increasing erosion and incorporation of soils at lower altitudes. The well mixed denudation rates yield integration times of w700e1800 years. Therefore, all but the uppermost reaches of the catchment will have reached cosmogenic steady state since deglaciation. It is unlikely then that the nearly consistent trend is due to non-steady state conditions due to glacial cover and erosion. The

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

161

Fig. 9. Relative basin-wide denudation rates, calculated through 10Be measured values. The Zielbach alluvial sediment (diamonds; open diamonds represent tributary’s samples) shows a decreasing trend moving downstream with local decreases due to the till contribution from hillslopes (squares). Below the junction with the debris-flow catchment the Zielbach experiences an abrupt increase of erosion rate due to the sediment-supply from the high-frequency debris-flow processes affecting this tributary (circles). The light rhomb represents the value of Norton et al. (2011).

soils erode at 109e196 mm/kyr with no clear trend with elevation. Since the aerial coverage of soils increases generally downstream (Fig. 1), erosion (even if spatially homogeneous) and incorporation of this material into the stream will result in decreasing calculated denudation rates in the downstream direction. In the Zielbach, this seems to be the more reasonable explanation. 6. Conclusion In this paper we show how sediment mixing models provide insight into sediment transfer mechanisms and the overprint related to the supply of material by of debris-flow processes on the cosmogenic nuclides concentrations of an Alpine catchment. In a transient landscape, the use of the cosmogenic nuclide technique for the calculation of basin-wide erosion rates can be difficult due to the highly irregular supply of sediment. When sediment is transferred homogeneously through the basin (both in space and time), the cosmogenic nuclide technique can be used reliably, independently from catchment size. The application of the 10Be technique is, however, more complicated when the catchment has been dominated by mass-wasting processes, since the cosmogenic signature is more sensitive to sporadic events. In this case, the calculated erosion rate might significantly over- or under-estimate the real denudation rate. It is important that the sediment is well mixed in the channel (i.e. the cosmogenic signature in the main channel equally represents the upstream tributaries) and that the processes affecting the basin are homogeneous both in frequency and magnitude. The results presented here suggest that the returnperiod of high-magnitude events needs to be considered when Alpine basins are sampled for erosion rate studies. Generally, sediment will be better mixed with increasing downstream

distances from confluence with a debris-flow dominated channel. Likewise, the extent of mixing will increase if samples are collected after sufficient, yet to be specified, time of a debris flow event. In this context, larger magnitude mass wasting events require the consideration of longer distances and times for sampling. Accordingly, every sampling campaign as to be carefully planned, depending on the site-specific processes operating within individual catchments for different spatial and temporal scales, and particularly, if landscapes are in transient state. In this sense, more detailed studies are needed to better understand the spatial and temporal effects of sporadic events on cosmogenic nuclide concentrations, where mass-wasting processes are the dominating mechanisms of sediment transfer in the catchment. Acknowledgments We greatly appreciate the financial support of ESF TopoEurope (CRP SedyMont e IP1) and the Swiss National Science Foundation (project No. 20T021-120464). Editorial handling by: S. Ivy-Ochs References Akçar, N., 2006. Paleoglacial records from the Black Sea area of Turkey field and dating evidence. Ph.D. dissertation, Bern e University, Bern, Switzerland. Akçar, N., Deline, P., Ivy-Ochs, S., Alfimov, V., Hajadas, I., Kubik, P.W., Christl, M., Schlucter, C., 2012. The AD 1717 rock avalanche deposits in the upper Ferret Valley (Italy): a dating approach with cosmogenic 10Be. Journal of Quaternary Science. ISSN: 0267-8179 ISSN: 0267-8179, 1e11. http://dx.doi.org/10.1002/ jqs.1558. Auer M. 2003. Regionalisierung von Schneeparametern e eine Methode zur darstellung von Schneeparametern im Relief, Unpublished Master’s thesis, University of Bern, 97 pp.

162

S. Savi et al. / Quaternary Geochronology 19 (2014) 148e162

Balco, G., Stone, J.O., Lifton, N.A., Dunai, T.J., 2008. A complete and easily accessible means of calculating surface exposure ages or erosion rates from 10Be and 26Al measurements. Quaternary Geochronology 3, 174e195. Bargossi, G.M., Bove, G., Cucato, M., Gregnanin, A., Morelli, C., Moretti, A., Poli, S., Zanchetta, S., Zanchi, A., 2010. Note illustrative della Carta Geologica d’Italia. Foglio 013 e Merano. Servizio Geologico d’Italia, ISPRA. Belmont, P., Pazzaglia, F.J., Gosse, J.C., 2007. Cosmogenic 10Be as a tracer for hillslope and channel sediment dynamics in the Clearwater River, western Washington State. Earth and Planetary Science Letters 264, 123e135. Bierman, P., Steig, E.J., 1996. Estimating rates of denudation using cosmogenic isotope abundances in sediment. Earth Surface Processes and Landforms 21, 125e139. Binnie, S., Phillips, W.M., Summerfield, M.A., Fifield, L.K., 2006. Sediment mixing and basin-wide cosmogenic nuclide analysis in rapidly-eroding mountainous environments. Quaternary Geochronology 1, 4e14. Bolzano Province, SIRIO-IHR, ED30 e Ripartizione 30 Opere Idrauliche. Brown, E.T., Stallard, R.F., Larsen, M.C., Bourles, D.L., Raisbeck, G.M., Yiou, F., 1995. Denudation rates determined from the accumulation of in-situ-produced 10Be in the Luquillo experimental forest, Puerto Rico. Earth and Planetary Science Letters 129, 193e202. Castiglioni, B., 1928. Ghiacciai delle Venoste Orientali. Bollettino Comitato Glaciologico 8, 91e165. Chmeleff, J., von Blanckenburg, F., Kossert, K., Jackob, D., 2010. Determination of the 10 Be half-life by multicollector ICP-MS and liquid scintillation counting. Nuclear Instruments and Methods in Physics Research B 268, 192e199. Clapp, E.M., Bierman, P.R., Caffee, M., 2002. Using 10Be and 26Al to determine sediment generation rates and identify sediment source areas in an arid region drainage basin. Geomorphology 45, 89e104. Codilean, A.T., 2006. Calculation of the cosmogenic nuclide production topographic shielding scaling factor for large areas using DEMs. Earth Surface Processes and Landforms 31, 785e794. Delunel, R., van der Beek, P.A., Carcaillet, J., Bourlès, D.L., Valla, P.G., 2010. Frostcracking control on catchment denudation rates: Insights from in situ produced 10 Be concentrations in stream sediments (EcrinsePelvoux massif, French Western Alps). Earth and Planetary Science Letters 293, 72e83. Dunai, T.J., 2001. Scaling factors for production rates of in situ produced cosmogenic nuclides: a critical reevaluation. Earth and Planetary Science Letters 176, 157e169. Dunai, T.J., 2010. Cosmogenic Nuclides. Cambridge University Press, UK. Gosse, J.C., Phillips, F.M., 2001. Terrestrial in situ cosmogenic nuclides: theory and application. Quaternary Science Reviews 20, 1475e1560. Granger, D.E., Riebe, C.S., 2007. Cosmogenic nuclides in Weathering and Erosion. In: Drever, J.I. (Ed.), Surface and Ground Water, Weathering, and Soils. Treatise on Geochemistry, vol. 5. Elsevier-Pergamon, Oxford. Holland, H.D., Turekian, K.K. (eds.). Granger, D.E., Kirchner, J.W., Finkel, R., 1996. Spatially averaged long-term erosion rates measured from in situ-produced cosmogenic nuclides in alluvial sediment. Journal of Geology 104 (3), 249e257. Ivi-Ochs, S., Kerschner, H., Reuther, A., Maisch, M., et al., 2006. The Timing of Glacier Advances in the Northern European Alps Based on Surface Exposure Dating with Cosmogenic 10Be, 26Al, 36Cl, and 21Ne. Geological Society of America. special Paper 415. Jarman, D., Agliardi, F., Crosta, G.B., 2011. Megafans and outsize fans from catastrophic slope failures in Alpine glacial troughs: the Malser Haide and the Val Venosta cluster, Italy. In: Jaboyedoff, M. (Ed.), Slope Tectonics. Special Publications, 351. Geological Society, London, pp. 253e277. http://dx.doi.org/10.1144/SP351.14. Kober, F., Ivy-Ochs, S., Schlunegger, F., Baur, H., Kubik, P.W., Wieler, R., 2007. Denudation rates and a topography-driven rainfall threshold in northern Chile: multiple cosmogenic nuclide data and sediment yield budgets. Geomorphology 83, 97e120. Kober, F., Hippe, K., Salcher, B., Ivy-Ochs, S., Kubik, P.W., Wacker, L., Hählen, N., 2012. Debris-flow-dependent variation of cosmogenically derived catchment-wide denudation rates. Geology. http://dx.doi.org/10.1130/G33406.1. Kohl, C.P., Nishiizumi, K., 1992. Chemical isolation of quartz for measurement of in-situ produced cosmogenic nuclides. Geochimica Cosmochimica Acta 56, 3583e3587.

Korschinek, G., Bergmaier, A., Faestermann, T., Gertmann, U.C., Knie, K., Rugel, G., Wallner, A., Dillmann, I., Dollinger, G., Lierse von Gosstomski, Ch, Kossert, K., Maiti, M., Poutivtsev, M., Remmert, A., 2010. A new value for the 10Be half-life by heavy-ion elastic recoil detection and liquid scintillation counting. Nuclear Instruments and Methods in Physics Research B 268, 187e191. Kubik, P.W., Christl, M., 2010. 10Be and 26Al measurements at the Zurich 6 MV tandem AMS facility. Nuclear Instruments and Methods in Physics Research B 268, 880e883. Lal, D., 1991. Cosmic ray labeling of erosion surfaces: in situ nuclide production rates and erosion models. Earth and Planetary Science Letters 104, 424e439. Matmon, A., Bierman, P.R., Larsen, J., Southworth, S., Pavich, M., Finkel, R., Caffee, M., 2003. Erosion of an ancient mountain range, the Great Smoky Mountains, North Carolina and Tennessee. American Journal of Science 303 (9), 817e855. Niemi, N.A., Oskin, M., Burbank, D.W., Heimsath, A.M., Gabet, E.J., 2005. Effects of bedrock landslides on cosmogenically determined erosion rates. Earth and Planetary Science Letters 237, 480e498. Norton, K. P., 2008. Response of the landscape in the Swiss Alps to the late glacial to Holocene climate transision. PhD thesis, Leibniz University of Hannover, Germany, 148pp. Norton, K.P., von Blanckenburg, F., Schlunegger, F., Schwab, M., Kubik, P.W., 2008. Cosmogenic nuclide-based investigation of spatial erosion and hillslope channel coupling in the transient foreland of the Swiss Alps. Geomorphology 95, 474e486. Norton, K.P., von Blanckenburg, F., Kubik, P.W., 2010. Cosmogenic nuclide-derived rates of diffusive and episodic erosion in the glacially sculpted upper Rhone Valley, Swiss Alps. Earth Surface Processes and Landforms. http://dx.doi.org/ 10.1002/esp.1961. Norton, K.P., von Blanckenburg, F., DiBiase, R., Schlunegger, F., Kubik, P.W., 2011. Cosmogenic 10Be-derived denudation rates of the Eastern and Southern European Alps. International Journal of Earth Science (Geol Rundsch) 100, 1163e 1179. http://dx.doi.org/10.1007/s00531-010-0626-y. Palumbo, L., Hetzel, R., Tao, M., Li, X., 2011. Catchment-wide denudation rates at the margin of NE Tibet from in situ-produced cosmogenic 10Be. Terra Nova 23, 42e48. Penck, A., Brückner, E., 1909. Die Alpen in Eiszeitalter, vol. III. Tauschnitz ed., Lipsia. Schaller, M., Von Blanckenburg, F., Veldkamp, A., Van den Berg, M.W., Hovius, N., Kubik, P.W., 2004. Paleo-erosion rates from cosmogenic 10Be in a 1.3 Ma terrace sequence: River Meuse, the Netherlands. Journal of Geology 112, 127e144. Schmid, S.M., Fügenschuh, B., Kissling, E., Schuster, R., 2004. Tectonic map and overall architecture of the Alpine orogen. Eclogae Geologicae Helvetiae 97, 93e 117. http://dx.doi.org/10.1007/s00015-004-1113-x. Small, E.E., Anderson, R.S., Hancock, G.S., 1999. Estimates of the rate of regolith production using 10Be and 26Al from an alpine hillslope. Geomorphology 27, 131e150. Van den Berg, F., Schlunegger, F., Akçar, N., Kubik, P., 2012. 10Be-derived assessment of accelerated erosion in a glacially conditioned inner gorge, Entlebuch, Central Alps of Switzerland. Earth Surface Processes and Landforms. http://dx.doi.org/ 10.1002/esp.3237. Vassallo, R., Ritz, J.F., Carretier, S., 2011. Control of geomorphic processes on 10Be concentrations in individual clasts: Complexity of the exposure history in GobiAltay range (Mongolia). Geomorphology 135, 35e47. von Blanckenburg, F., 2005. The control mechanisms of erosion and weathering at basin scale from cosmogenic nuclides in river sediment. Earth and Planetary Science Letters 237, 462e479. Wittmann, H., von Blanckenburg, F., Kruesmann, T., Norton, K.P., Kubik, P.W., 2007. Relation between rock uplift and denudation from cosmogenic nuclides in river sediment in the Central Alps of Switzerland. Journal of Geophysical Research 112, F04010. http://dx.doi.org/10.1029/2006JF000729. Wittmann, H., von Blanckenburg, F., Guyot, J.L., Maurice, L., Kubik, P.W., 2009. From source to sink: preserving the cosmogenic 10Be-derived denudation rate signal of the Bolivian Andes in sediment of the Beni and Mamoré foreland basins. Earth and Planetary Science Letters 288, 463e474. Yanites, B.J., Tucker, G.E., Anderson, R.S., 2009. Numerical and analytical models of cosmogenic radionuclide dynamics in landslide-dominated drainage basins. Journal of Geophysical Research 114, F01007. http://dx.doi.org/10.1029/ 2008JF001088.