Glaciers and subsequent mass wasting events create impressive mountain landscapes; however, the ruggedness that defines these beautiful landscapes also makes it challenging to monitor erosion in the field. The result is that spatial patterns and rates of erosion in alpine landscapes are understudied. Field locations are steep and remote and hillslope processes, including rockfalls, avalanches, and landslides, are stochastic and difficult to measure directly. This study uses talus fan sediments to deepen our understanding of individual fan deposition and catchment averaged erosion processes in the alpine setting of Garnet Canyon in the Teton Range, Wyoming, USA. We measured cosmogenic 10Be concentrations from bedrock and talus deposits to compare them to volumetric estimates of erosion rates, lichen growth, and surface weathering on talus surfaces. Amalgamated pebbles from the talus deposits contained lower 10Be concentrations than any bedrock surfaces or stream sediments. The young talus surface exposure ages are all younger than 11 ka, reflecting the importance of continued rockfall activity long after glacial retreat. Only one talus fan corresponded to known seismic events. Talus deposits contribute sediments to stream systems; 10Be concentrations were lower in amalgamated talus pebbles than in amalgamated stream sands. Lichen cover, volumetric estimates of erosion rates, and 10Be concentrations showed similar spatial trends reflecting the migration of active rockfalls to higher elevations and validating the applicability of 10Be concentrations to quantify talus surface ages. Distinct 10Be concentrations on various surfaces within Garnet Canyon indicate that future work with amalgamated samples from talus deposits can contribute to investigations about landscape evolution in alpine landscapes.

Mass movements in mountain environments shape hillslope topography and influence sediment flux in glacial and fluvial systems (Burbank et al., 1996; Hales and Roering, 2009; Straumann and Korup, 2009; Stock and Uhrhammer, 2010; Ward and Anderson, 2011). The stochastic nature of mass movements, however, makes quantifying erosion rates in remote high-altitude regions difficult and introduces uncertainties in measurement techniques (Heimsath and McGlynn, 2008). Observed rockfalls typically range from a few small blocks (<1 m) to large volumes of millions of cubic meters in various mountain environments (Dussauge et al., 2003; Strunden et al., 2015). Hillslope erosion via rockfall occurs rapidly after glacial retreat (Arsenault and Meigs, 2005; Meigs et al., 2006; Sanders and Ostermann, 2011) because glaciation deeply scours canyons to create steep valley walls (Hallet et al., 1996; Alley et al., 2003; Brocklehurst and Whipple, 2004; Foster et al., 2008). Low gradients of overdeepened valley floors reduce the efficiency of fluvial excavation; therefore, rockfall sediments accumulate to form talus fans that subsequently influence stream systems (MacGregor et al., 2000; Dühnforth et al., 2008). If we can define spatial patterns of hillslope erosion and quantify when talus sediments were deposited, we can use talus deposits to investigate climate variability, individual tectonic events, and connections between rock properties and erosion rates.

To date, talus fan volumes and glacial maximum ages are used to quantify hillslope erosion rates to understand the role of mass wasting in mountain landscape evolution (Olyphant, 1983; Sass and Wollny, 2001; Moore et al., 2009; O’Farrell et al., 2009; Tranel et al., 2015). In these estimates, the total volume of material accumulated on a valley floor is assumed to be derived from adjacent valley walls (Olyphant, 1983; Moore et al., 2009). Volume estimates require projecting the wall surface below the fan deposit unless subsurface images or intensive laser acquired wall images are obtained (Sass and Wollny, 2001; Stock et al., 2011; Strunden et al., 2015). Additional uncertainty associated with evaluating rockfall extent and rates is due to the lack of detail in the timing between events. The timing and size of rockfall events are often only recorded if they are directly observed or cause damage to anthropogenic structures. Repeat aerial imagery, laser scanning, or lichenometric dating are methods used to determine the timing of mass movements; however, environmental conditions, access to equipment, or data images may limit widespread use (Jomelli, 2013). Cosmogenic radionuclide analyses on amalgamated gravels or sands are increasingly used to constrain fluvial catchment averaged erosion rates (Anderson et al., 1996; Bierman and Steig, 1996; Granger et al., 1996, 2001; Cockburn et al., 2000; Balco et al., 2008; Portenga and Bierman, 2011). Amalgamated sediments or pebble samples allow a single sample to represent complex surfaces or events (Muzikar, 2009). Fewer studies have also used amalgamated samples to study the timing of alluvial fan deposition in arid environments and valley wall retreat in glacial valleys (Repka et al., 1997; Heimsath and McGlynn, 2008; Ward and Anderson, 2011; Ivy-Ochs et al., 2013).

In this work we used cosmogenic in situ 10Be concentrations to study talus accumulation, the influence of rockfalls on alpine landscapes, and the applicability of amalgamating pebbles to date talus surfaces. We assess ridge bedrock and talus surface 10Be concentrations and compare them to talus volume erosion rates (by Tranel et al., 2015), lichen cover, and weathering on pebble surfaces. Then, we consider how well various talus observations help us understand the evolution of Garnet Canyon, an alpine catchment in the Teton Range, after glacial retreat. We use 10Be concentrations to evaluate relative age differences between bedrock surfaces and sediment deposits, spatial patterns of erosion, and relationships between concentrations previously studied erosion rates and geologic events in the Teton Range (Wyoming, USA).

The Teton Range is located in northwestern Wyoming, south of the Yellowstone volcanic high (Fig. 1). The collection and observation sites for this study were located in and around Garnet Canyon (Fig. 2), an east-draining watershed adjacent to the central highest peak, the Grand Teton (elevation 4198 m). The bedrock in Garnet Canyon and neighboring Cascade Canyon to the north consists of Archean Webb Canyon Gneiss and Mount Owen Quartz Monzonite (Love et al., 1992; Zartman and Reed, 1998). Paleozoic strata unconformably overlie the Archean igneous and metamorphic rocks (Love et al., 1992).

Exhumation of the Teton Range began with development of thrust sheets during the Sevier-Laramide orogeny (Love, 1973; Craddock et al., 1988; Lageson, 1992) and continued with Basin and Range extension and development of the Teton normal fault 11–9 Ma (Smith et al., 1993; Roberts and Burbank, 1993; Brown, 2010). Brown (2010) also completed three-dimensional volumetric estimates of bedrock removed above the Paleozoic unconformity and calculated an exhumation rate of 0.18 mm/yr if uplift began ca. 10 Ma. Ongoing offset along the Teton fault created scarps in Pinedale-age moraines, preserving evidence of displacement in the past 13 k.y.; the most recent displacement occurred between 8 and 4.8 ka (Smith et al., 1993; Byrd, 1995; Thackray and Staley, 2014).

Intense glaciation during the Pleistocene carved a rugged landscape out of the west-dipping uplifted fault block. Glaciers incised steep U-shaped canyons, created low-relief steps in the longitudinal canyon profiles, and deposited moraines extending into Jackson Hole to the east (Pierce and Good, 1992; Foster et al., 2010; Tranel et al., 2011). Glaciers most recently extended into Jackson Hole ca. 14 ka, and retreated by ca. 11.5 ka (Licciardi and Pierce, 2008; Larsen et al., 2016). Several small glaciers remain at high elevation in the Teton Range. Licciardi and Pierce (2008) dated boulders at the head of Cascade Canyon that were 13–10 ka (Lake Solitude) and bedrock surfaces that were ca. 13 ka. Exposure ages of 12–11 ka on polished bedrock surfaces in central Garnet Canyon were obtained (Tranel et al., 2015). The most recent retreat in the nearby Wind River Range was ca. 11 ka based on moraine exposure, although these ages may be older given recent advancements in cosmogenic dating methods (Gosse et al., 1995; Licciardi and Pierce, 2008).

Although glaciers controlled the shape of the Teton landscape, rockfalls also heavily influenced the landscape by rapidly eroding valley walls and depositing extensive talus fans on the valley floors (Foster et al., 2010; Tranel et al., 2011). In Tranel et al. (2015), the contribution of rockfalls to valley wall erosion was estimated with mapped talus deposits in Garnet, Avalanche, and Glacier Gulch Canyons; field surveys of talus volumes obtained erosion rates ranging from 0.03 to 6 mm/yr and averaging 0.8 mm/yr.

Pebble and Bedrock 10Be

We collected samples of bedrock from two ridge locations and amalgamated pebbles from four talus fans in and around Garnet Canyon. We chose locations where we could access bedrock on the ridge almost directly above a talus deposit. Bedrock from the ridges was sampled from relatively flat and stable surfaces where shielding due to the surface slope would be minimal (Figs. 2 and 3). We collected samples from fans that faced east, west, north, and south around Garnet Canyon. On each of 4 fans, we collected 16–20 uniformly sized pebbles 5 cm thick or less to avoid variable accumulation of 10Be due to size or rolling pebbles (Balco, 2011; Applegate et al., 2012; Mackey and Lamb, 2013). We chose to sample sediment smaller than boulders to reduce variability of inheritance, and larger than sand to reduce water, vegetation, or animal disturbance (Schmidt et al., 2011; Ivy-Ochs et al., 2013). We collected multiple pebbles from each fan to account for variable inheritance from degradation of the fan surface and inputs from random locations on the valley walls (Matmon et al., 2006; Ward and Anderson, 2011; Applegate et al., 2010, 2012). Each pebble was gathered along a transect across the fan at the same elevation and spaced 1–2 m apart, depending on fan width, for 3 of the 4 fans (Fig. 3). Because the fourth fan (fan 18) was narrower, we collected the pebbles at equal intervals (1–2 m) from the top to the bottom of the fan (Fig. 3). The difference in collection methods could potentially influence mixing along the surface of the fan if materials rolled downslope along a similar path. However, our observations of a recent rockfall deposit suggest that pebbles and materials of all sizes are scattered across a wide area on the fan surface (Fig. 4).

We crushed and sieved samples to 0.25–1 mm at Illinois State University and sent samples to the Purdue University Rare Isotope Measurement Laboratory (PRIME Lab, West Lafayette, Indiana) for digestion, beryllium separation, and accelerator mass spectrometry (Kohl and Nishiizumi, 1992). Exposure age calculations for all samples used the CRONUS online calculator (version 2.2; Balco et al., 2008). All results listed in Table 1 used the time-varying production rate from Dunai (2001) for comparison to results in Tranel et al. (2015). The elevation, latitude, and longitude of each talus fan and bedrock area were estimated using the centroid of each polygon plotted with the Feature to Point tool in ArcGIS.

We determined the shielding factor for talus surfaces using 3D Analyst tools in ArcGIS and a 10 m U.S. Geological Survey digital elevation model ( We used the ArcGIS Skyline tool to create polygons including all of the areas visible from the sample points on the talus surface. Then we used the Create Lines tool (ArcGIS) to draw lines extending from each sample point to the outer limit of the skyline polygon (the horizon). Usually a line was created approximately every 3° (100 m), although some distances were greater or smaller. We selected the lines with a positive angle to the horizon, and calculated the topographic shielding factor (CT) for each point (Dunne et al., 1999; Codilean, 2006; Norton and Vanacker, 2009). We averaged CT values for each sample point to obtain the final shielding correction factor for each amalgamated sample.

Because snow also shields the exposed surfaces and reduces 10Be production, we incorporated a correction factor to account for snow cover based on modern snowfall estimates. Gosse and Phillips (2001) provided a calculation to quantify snow shielding based on the density of snow, the depth of snowfall each month, and attenuation length (Binnie, 2004). We applied the formulation using the average snowfall recorded by Dirks and Martner (1982) at stations in Moran and Moose, Wyoming (elevation ∼2000 m; Fig. 1). These values allowed us to calculate a minimum shielding correction because the talus locations were likely to receive higher snowfall at higher elevations than the weather station locations in Jackson Hole. We multiplied the snow correction factor (0.94) and the shielding correction factor (see Table 1) and used the combined value as the input for the shielding correction element in the CRONUS online calculator.

Pebble Weathering

We described weathering features on each pebble composing the amalgamated samples before crushing for quartz extraction and dissolution. First we studied physical weathering with Krumbein (1941) roundness and Zingg (1935) shapes to characterize if corners were broken down either through chemical weathering and dissolution of minerals or abrasion from rolling or washing down the talus slope after initial deposition. We assume that all materials start out very angular because they break apart on impact with the hillslope walls or on other rock debris as they fall. If pebbles are weathering, the corners will become more rounded over time (Birkeland, 1973).

We also used the color or percentage of lichen cover to compare relative differences in exposure on fan surfaces. Pebbles deposited earlier were exposed to chemical weathering longer, and would possibly display more surficial color change in comparison to the unweathered interior of the pebble (Birkeland, 1973; Whitehouse and McSaveney, 1983). Each pebble was therefore sawed or broken to record any gradations in color compared to the Munsell color chart ( We also observed the lichen cover on each side of the collected pebbles, assuming that pebbles showing more lichen cover would indicate longer residence and no renewed movement along the surface (Bradwell, 2009). The exposed surface was marked on each pebble during field collection. We estimated the percentage of lichen cover on each pebble face, and averaged the values for each face to produce a whole pebble percent.

To use lichen as a quantitative measurement of surface age requires extensive work to determine the lichen growth rate curve for the specific environments where the lichen are observed or calibrate the size to the age of known surfaces (Hanson, 2008; Armstrong and Bradwell, 2010). We chose to look at the overall coverage on the pebbles rather than measuring the maximum or average lichen size because the total surface area of the pebbles was too small for other quantitative estimates. Competition between lichen and elevation could also influence growth (Kodros, 1997). We assume that the percentage of lichen cover increases with exposure time (Grab et al., 2005; Bradwell, 2009); therefore, fans containing rocks with higher percentages of lichen cover are relatively older than talus fans where few samples demonstrated lichen growth.

Pebble Weathering

Lichen cover showed the most variability in the qualitative observations on the amalgamated pebble samples (Table 2). The highest elevation talus fan (fan 11) was the only fan with no lichen growth on any of the pebbles (Figs. 2 and 3). Lichen covered a maximum of 31% of the clasts on fan 153 (sample TTC-61), which faced north. The clast with the highest percentage of lichen cover (44%) was sampled from the lowest elevation and south-facing fan (fan 56, sample TTC-71).

Additional weathering observations were collected and listed in Table 2 because differences in vegetation on talus surfaces indicated varying degrees of soil development from low to high elevation. Weathering rinds were not developed in our samples, although surface colors were different from interior colors. We observed red discoloration on samples from three talus fans throughout the pebbles, possibly related to weathered iron-bearing minerals (Birkeland, 1973), but it occurred throughout the sample rather than concentric to the surface. The most common Munsell rock colors for each talus fan are listed in Table 2. The most common color differed for each fan. Roundness was similar for most of the samples. Krumbein values were uniform, and all samples had a high number of bladed pebbles with a little variation in the number of oblate or prolate pebbles (Table 2).

Pebble and Bedrock 10Be Concentrations

We expected ridge bedrock samples to have the highest 10Be concentrations and represent the oldest surfaces around Garnet Canyon because they were collected from stable surfaces. All ridge samples except one matched our assumption (Table 1). The exception was sample TTC-08 (5.15 × 10510Be atm/g), collected near the top of Middle Teton, which had a concentration closer to the glaciated floor surfaces (Fig. 5; 59.6 × 104 atm/g in TTC-11; Tranel et al., 2015). Concentrations in all other bedrock samples were >1.00 × 106 atm/g. Although sample TTC-08 was taken from a flat surface (Fig. 3), we assumed that shielding by snow cover would be minimal due to the high wind exposure near a peak at an elevation of 3903 m.

We expected the talus concentrations to be less than all the floor bedrock concentrations because we assume that the rockfalls accumulated after glaciers retreated from the canyon. Our results are consistent with this expectation. Concentrations of 10Be in the amalgamated pebbles were similar in three of the talus fans (2.00–5.00 × 104 atm/g), but much higher in sample TTC-71 (1.30 × 105 atm/g). Sample TTC-71 was collected from talus fan 56, which was the largest fan. The fan size and 10Be concentration both suggest that fan 56 is older than the other fans. Trends between concentrations and elevation were similar between talus deposits and bedrock surfaces. The lowest elevation talus fan (fan 56, sample TTC-71) had the highest 10Be concentration, and the highest elevation talus fan (fan 11, sample TTC-66) had the lowest 10Be concentration. The average 10Be concentration on the valley floor (41.4 × 104 atm/g) reported in Tranel et al. (2015) was six times greater than the average concentration in the talus deposits (6.4 × 104 atm/g). To put our results in the context of the geologic history of the Teton Range, we used concentrations of 10Be from the amalgamated talus samples to calculate exposure ages of the talus surfaces. Talus surface exposure ages ranged from 0.6 to 7.9 k.y. (Table 1).

Studies using concentrations of 10Be to estimate erosion rates or exposure ages assume that erosion rates have been constant for a sufficiently long time and that sediment deposits represent an average of multiple processes across a catchment area (Bierman and Steig, 1996; Cockburn et al., 2000; Niemi et al., 2005; Binnie et al., 2008; Balco et al., 2008; Yanites et al., 2009; Balco, 2011; Applegate et al., 2012). Because alpine postglacial landscapes are rapidly changing from recent glacial retreat and stochastic processes, the assumptions commonly applied to cosmogenic nuclide rates do not fit well. Instead, we compare 10Be concentrations between different surfaces and deposits in the landscape to consider the relative ages of geomorphic features, incorporation into stream sediments, and spatial patterns of erosion. Then we consider how the features we observed relate to the geologic history of the Teton Range.

Relative Ages of Geomorphic Features

In Figure 6 we divide the catchment features into four surface classes: bedrock ridges, bedrock hillslopes, talus deposits, and exposed bedrock on the valley floor. The areas were mapped and classified from field observations and geographic information system maps of elevation, slope, bedrock, and Quaternary deposits. Each surface class is subjected to different combinations of geomorphic processes that influence 10Be concentrations. Ridges are high-elevation surfaces between catchments with low slopes near the drainage divide or in close proximity to peaks. Higher 10Be concentrations in ridge samples reflect relatively little erosion and exposure during glaciation, making them the oldest surfaces. Bedrock floors are low-elevation surfaces demonstrating evidence of glacial scour, striations, and chatter marks along the valley bottom near ice, lakes, and between talus deposits (Fig. 6). Concentrations (from Tranel et al., 2015) reflect postglacial exposure, with little erosion since glacial retreat. Talus deposits are Quaternary colluvium (mapped by Love et al., 1992; Tranel et al., 2015). Talus concentrations are potentially a composite of high concentration material from near the ridges, steep bedrock hillslopes (either similar in age to ridges or to glacially scoured floors), and fresh rock fragments from boulders several meters thick. Talus samples contained the lowest 10Be concentration, and therefore are the youngest geomorphic features. Although older source material from bedrock hillslopes is incorporated in the deposit, it does not have a noticeable impact on the talus 10Be concentration.

Bedrock hillslopes are valley walls without talus cover, and are the most complex surface to evaluate for 10Be concentrations (Fig. 6). We did not collect samples from bedrock hillslopes, but we would expect that the 10Be concentration would be less than ridge values and greater or equal to valley floor values because many wall areas were probably covered by glacial ice at some time. Other surfaces on the bedrock hillslopes, however, were possibly exposed because observations and models suggest the valley was not completely filled with ice during the last ice age (Foster et al., 2010). In addition, random locations on the bedrock hillslopes may have very low concentrations where boulders recently fell from the slope.

Comparison of Bedrock and Talus Surfaces to Stream Sediments

In the ideal setting, stream 10Be concentrations reflect combined inputs from all surfaces in a catchment (bedrock ridges, bedrock hillslopes, talus, and floor bedrock) eroding and contributing sediment equally. Weighting the concentrations of each surface class in Figure 6 by the corresponding contributing area should equal the stream sediment concentration. The result of the weighted average of talus, bedrock, and floor concentrations (∼40 × 104 atm/g) was almost 3 times greater than the average observed concentration of 14.5 × 104 atm/g in stream sand-sized sediments from Tranel et al. (2015). The difference between the stream and weighted concentrations indicates that talus contributes disproportionate sediment amounts to the relative surface area covered in Garnet Canyon. Although the grain sizes vary between the talus and stream deposits, the low concentrations in both indicate that hillslope processes interact with the stream system to make stream concentrations much lower than bedrock surface concentrations. This is consistent with models and other field studies that predict that hillslope processes cause catchment averaged stream erosion rates to be higher than bedrock erosion rates (Niemi et al., 2005).

The talus deposits are not the only sediment source to the stream system. If the talus deposits were the only contributor, we would expect similar talus and stream concentrations. One talus concentration (TTC-71; 13.78 × 104 atm/g) was similar to the stream concentrations (11–19 × 104 atm/g in 3 stream sand samples; Tranel et al., 2015; Table 2); however, the rest were lower (<6 × 104 atm/g). The average of these stream 10Be concentrations is approximately two times greater than the average of concentrations in the talus. Although bedrock weathering in Garnet Canyon is slow, some sediment from the valley floor or bedrock wall must enter the stream channel. As discussed in more detail in the following, Garnet Canyon demonstrates spatial variability in erosion. If sediments are well mixed, catchment averaged erosion rates can still represent a long-term average (Foster and Anderson, 2016, and references within). However, small catchments tend to be poorly mixed due to variable storage and incision along the stream length (Yanites et al., 2009). Garnet Canyon appears poorly mixed because stream sediment concentrations are similar to the concentration of sample TTC-71 from fan 56. Because we are comparing pebbles to sands, however, our observations may instead reflect differences in concentrations recorded by distinct grain size classes (Lukens et al., 2016). Sand may be eroded from lower elevation sources. If we were to sample pebbles at the mouth as well, they might be similar to higher elevation talus fan sources (Riebe et al., 2015).

Spatial Patterns of Erosion

To understand how well 10Be concentrations reflect ages or rates of processes, we compare results to other qualitative and quantitative measures of age and erosion rates. Talus 10Be concentrations, talus volume erosion rates (Tranel et al., 2015), and lichen cover all support erosion activity increase or surface age decline with increasing elevation. Talus deposits exhibited lower 10Be concentrations at higher elevations (Fig. 5). Talus volume erosion rate estimates were faster at higher elevations, as expected from 10Be concentrations (Fig. 7A; data from Tranel et al., 2015). Lichen cover was reduced at higher elevations (Fig. 7B). The fan with the highest 10Be concentration (fan 56, sample TTC-71) also held the single clast with the most lichen cover. The spatial trends from our results suggest that as glaciers retreated from the canyon, new deposits covered surfaces previously covered by ice. Deposition on lower elevation fans became stagnant, but continued at higher elevations. These patterns imply that a catchment averaged erosion rate represents an average of processes in this complex alpine system, but that erosion is variable and dependent on factors affected by elevation (Riebe et al., 2015). To understand overall landscape changes we need to separate the catchment into zones defined by elevation, local climates, and dominant geomorphic processes in each area.

Implications for Inherited 10Be Uncertainty Estimates

Concentrations are sufficiently different between the surface types that inheritance does not limit a study of differential erosion across a catchment. Inheritance is more problematic when comparing individual ages of talus fans, but to address this problem we consider a number of factors that influence 10Be accumulation. In addition to sediment mixing and snow shielding, uncertainties can include inheritance in bedrock and sediment deposits, as well as degradation of geomorphic features.

We accounted for snow shielding as described herein using standard methods (Gosse and Phillips, 2001). Corrections accounted for a 5% difference in results. Temperatures influenced by elevation could affect the amount of snow cover across the catchment. For example, the length of snow cover (shielding) per year would increase with elevation because temperatures decrease with elevation. If all talus fell at the same time with close to zero inheritance, then a reduction in concentration with elevation is expected. With the historic climate variability since 11 ka (Larsen et al., 2016), it would be challenging to account for more complex shielding at individual talus locations. Another problem is the inheritance from a complex exposure history of the bedrock source. Materials accumulated on talus fans were produced from large blocks falling away from the valley wall and breaking apart in the fall. Outer surfaces of those blocks would contain 10Be accumulated when the original block was in place; therefore, some pebbles might contain inherited 10Be from the parent material.

Several methods have been proposed in the literature to account for inheritance in geomorphic deposits. Studies using amalgamated cobble samples to date alluvial terraces sampled subsurface cobbles in at least one location to create a vertical profile of 10Be concentrations and quantify the exact amount of inheritance in the sediments (Anderson et al., 1996; Repka et al., 1997). The nature of talus deposits in our study area precluded us collecting samples to create a vertical profile. Large boulders would make the profile difficult to collect and skew the profile concentrations. Given that all the talus surface ages in our study are younger than valley floor ages estimated with 10Be (youngest 11 ka; Tranel et al., 2015) and that the relationships between talus volume, lichen cover, and 10Be concentrations described here were positive, none of our results seem anomalously old due to significant inheritance.

Talus Ages in Relation to Climate and Tectonics in the Teton Range

The parallels between maximum lichen cover, talus volumes, and 10Be concentrations support our use of the 10Be concentrations to estimate talus surface ages. With ages for the talus deposits (Table 1), we can evaluate how rockfall timing relates to climate and tectonic events in the Teton Range. In Tranel et al. (2015) it was observed that the Garnet Canyon valley floor was exposed between 12 and 11 ka, and Larsen et al. (2016) observed nonglacial sediments began to dominate in Jenny Lake at about the same time. After glacial retreat, lake sediments preserved records of a warm period between 10 and 6.5 ka and considerable climate variability after 6 ka (Larsen et al., 2016). At least one earthquake with a magnitude sufficient to displace surficial deposits occurred between 8 and 4.8 ka (Byrd, 1995).

The age results in Garnet Canyon indicate that talus accumulation is not a one-time event where all deposits formed immediately after deglaciation. All talus surfaces are younger than bedrock on the valley floor. As we would expect in a setting with rockfall mass wasting, those sediments accumulate over time as more rocks fall onto the surfaces. The oldest talus fan formed within a window of time where climate was warm and a magnitude 7 earthquake occurred in the Teton Range. If talus fan 56 (sample TTC-71) formed from an earthquake-triggered rockfall, perhaps that one event was sufficient to stabilize the slope above the fan so that few rockfalls have occurred since. The rest of the talus fans, however, are much younger. Major events are not required to cause failures on steep rock slopes, and recent studies in Yosemite Valley, California, observed that warm temperatures can propagate fracture expansion to trigger rockfalls (Collins and Stock, 2016). Fluctuations in climate likely influenced daily and monthly temperatures, precipitation, and water flow in bedrock joints to propagate the already fractured bedrock around Garnet Canyon and eventually led to failures that produced talus fans over time.

A study of cross-valley profiles throughout the Teton Range described valley shapes as asymmetric, reflecting variable erosion related to local climate controlled by north or south aspects and rock strength (Marston et al., 2011). Based on their observations of channel asymmetry, we would expect to see lower 10Be concentrations and less lichen cover on the sides with steeper slopes. The 10Be concentrations, weathering, and lichen cover in our study do not capture significant differences in slope stability or effective erosion related to aspect in Garnet Canyon. Samples TTC-71 and TTC-61 were collected from fans at similar elevations. The average and maximum slopes calculated for the contributing bedrock wall areas were similar (Table 2). The slopes of bedrock wall profiles were 45° above all talus fans except fan 153, where the angle was 37° (Fig. 8). The south-facing fan surface (TTC-71, fan 56) had a higher 10Be concentration than the north-facing fan surface (TTC-61, fan 153). More frequent lichen growth was observed on pebbles on the north-facing talus fan (31% of the clasts showed evidence of lichen growth on TTC-61), but the single pebble with the greatest lichen cover (44%) was on the south-facing talus fan (TTC-71).

Talus surface features and 10Be concentrations have strong potential to help us understand geomorphic process interactions, as well as geologic and climatic events. Ages from talus 10Be concentrations capture the significantly younger age of the deposits relative to bedrock surfaces. Rockfalls distribute a significant volume of sediment with little to no inheritance despite complicated exposure histories on the source bedrock. The equivalent spatial patterns observed between 10Be concentrations, volumetric erosion rates, and lichen cover validate the use of these methods to quantify hillslope processes.

The overall low concentrations of 10Be in the talus deposits confirm that hillslope erosion is a significant and active geomorphic process contributing to the shape of the Teton landscape. Rockfall activity is closely linked to changing climate conditions following glacial retreat, and only loosely connected to earthquake events. Combined observations of 10Be concentrations in talus pebbles, lichen cover, and volumetric erosion rates support a progression of rockfall activity from low to high elevations. In addition, those high-elevation, low-concentration sediments are incorporated into fluvial sands. Only one talus deposit age correlated with a documented earthquake event. Most deposits likely occurred randomly through failures related to temperature and precipitation conditions contributing to joint fracture propagation.

Amalgamated pebble 10Be concentrations can provide valuable insight into the geomorphic mechanisms that shape alpine landscapes and contribute sediments to larger downstream systems. Pebble concentrations highlight the variability of erosion processes even within the limited area of a single small catchment. To fully understand how mountain landscapes evolve, we need to consider how local conditions within individual catchments influence the system as a whole. Future work to investigate rockfall processes should include models and further field studies of sediment mixing and transport along the talus surfaces and within stream channels, calibration of lichen growth, and more intensive sample collection within talus deposits or across bedrock hillslopes for 10Be analyses.

We thank Graham Andrews, Shanaka de Silva, and anonymous reviewers for valuable feedback on this manuscript. We also thank Audrey Happel, Amber Ritchie, Jill Tranel, and Anthony Abraham for help in the field with sample collection. Funding to travel to the field to collect samples was available through the Illinois State University College of Arts and Sciences New Faculty Initiative Grant. We also thank Tom Clifton and others at the Purdue Rare Isotope Measurement Laboratory (PRIME Lab, West Lafayette, Indiana) for help with sample preparation and analyses.