The geologic record of mountain glaciations is a robust indicator of terrestrial paleoclimate change. During the last glaciation, mountain ranges across the western US hosted glaciers while the Cordilleran and Laurentide
ice sheets flowed to the west and east of the continental divide,
respectively. Records detailing the chronologies and paleoclimate
significance of these ice advances have been developed for many sites across
North America. However, relatively few glacial records have been developed
for mountain glaciers in the northern Rocky Mountains near former ice sheet
margins. Here, we report cosmogenic beryllium-10 surface exposure ages and
numerical glacier modeling results, which show that mountain glaciers in the
northern Rockies abandoned terminal moraines after the end of the global
Last Glacial Maximum around 17–18 ka and could have been sustained by
Mountain glaciers are widely recognized as robust indicators of modern climate change (Oerlemans, 2005; Mark and Fernández, 2017). Investigations of past glacier fluctuations preserved in the geologic record can reveal valuable information regarding past climate oscillations and variability (e.g., Gilbert, 1890; Blackwelder, 1931; McCoy et al., 1985; Marcott et al., 2019). In the Rocky Mountain region of the western US, records of mountain glaciation have been used extensively to reconstruct the regional pattern of Pleistocene glaciation in space and time (e.g., Porter et al., 1983; Leonard, 1989; Licciardi et al., 2004; Laabs et al., 2009; Quirk et al., 2020), but few studies have focused on northern ranges along the former southern margins of the Laurentide and Cordilleran ice sheets. While surficial geologic records of Pleistocene mountain glaciation in the northern Rocky Mountains of western Montana have been available for decades (Alden, 1932; Carrara, 1987), these records have seldom been used to infer climate conditions (e.g., Murray and Locke, 1989). Many ranges were occupied by coalesced valley glaciers and ice caps with high-altitude ice divides, which are especially difficult to reconstruct based solely on mapped glacial deposits and landforms. Additionally, in much of northwestern Montana, mountain glaciers likely coalesced with the southern edges of the Laurentide and Cordilleran ice sheets, which also complicates reconstructions of paleo-glaciers and limits the usefulness of traditional methods for inferring past climate from glacier equilibrium-line altitudes or mass balance gradients.
However, discrete Pleistocene mountain glaciers occupied some ranges of western Montana, as evidenced by a well-preserved record of deposits and landforms delimiting their maximum extent during the last glaciation. Such records are found in the northern Absaroka Range in southwestern Montana and the eastern Lewis Range in northwestern Montana (Fig. 1), where glaciers incised deep valleys and in some areas constructed broad terminal-moraine complexes along mountain fronts. These records present an opportunity to reconstruct mountain glacier extents and develop cosmogenic chronologies of the last glaciation. These spatiotemporally constrained paleo-glaciers can then, in turn, be used to infer paleoclimate conditions in the northern Rocky Mountains during the last glaciation for which relatively few records exist compared to other regions of western North America.
Pleistocene ice extents in the northern US Rocky Mountains (after Porter et al., 1983; Pierce, 2003) with the locations of our two field sites, Cut Bank Creek (CB) in the Lewis Range and Pine Creek, South Fork Deep Creek, and Cascade Creek in the northern Absaroka Range (AB) indicated by green stars. Locations of previously established age control are indicated by yellow circles including the Greater Yellowstone glacial system (GYGS), Crazy Mountains (CM), Wind River (WR), Sawtooth (ST), Wasatch (WM), Uinta (UM), Front Range (FR), Sawatch (SW), and San Juan (SJ) ranges. General outlines of the Cordilleran and Laurentide ice sheets as well the northern Rocky Mountain ice cap (NRMIC) are also shown. Map of western North America (inset) with state outlines. Green stars indicate our study areas, and the red box shows the approximate coverage of the main illustration.
Here we present new surficial mapping of latero-terminal moraines of the
last Pleistocene glaciation in the Cut Bank and Lake Creek valleys in the
eastern Lewis Range and cosmogenic
The Lewis Range (48.5
The Absaroka Range (45
Reconstructions of Pleistocene glaciers in the northern Rocky Mountains of western Montana are limited (Pierce, 2003), and relatively little work has been done inferring past climate in the region from paleo-glacier characteristics. Most previous work has focused either on the Greater Yellowstone area of southern Montana or on the Glacier National Park area of northern Montana – also the foci of the current study. In these and other areas of western Montana past researchers have identified deposits and landforms from the penultimate and most recent glaciations, generally termed Bull Lake and Pinedale glaciations, following the terminology developed by Blackwelder (1915) for the Wind River Range of Wyoming (Fig. 1). Based on chronologies of glacial deposits throughout the middle and southern Rocky Mountains, these last two Pleistocene glaciations are thought to correspond broadly to intervals of global ice volume increase during Marine Isotope Stage (MIS) 2 and 6, respectively (Licciardi and Pierce, 2008, 2018; Quirk et al., 2018; Dahms et al., 2018; Schweinsberg et al., 2020; Laabs et al., 2020). Chronological work utilizing cosmogenic nuclide surface exposure dating in the Yellowstone–Grand Teton National Park area of southwestern Montana and adjacent northwestern Wyoming (Licciardi et al., 2001; Licciardi and Pierce, 2008, 2018; Pierce et al., 2018) has allowed subdivision of Pinedale-age deposits, as discussed below.
Deposits of Pleistocene mountain glaciers in the eastern Lewis Range of
western Montana were mapped and described as early as 1906 by Calhoun and
then later by Alden (1932), Carrara (1989), and Fullerton et al. (2004).
Calhoun (1906) described the broad hummocky terminal and recessional
moraines deposited on the plains to the east of Cut Bank Creek headwaters
investigated in this study as well as several recessional moraine ridges
deposited upvalley. Fullerton et al. (2004) inferred multiple Pinedale
tills, two ages of Bull Lake till, and a possible pre-Bull Lake till in
moraine deposits at Cut Bank Creek and elsewhere along the eastern front of
the Lewis Range. No numerical ages are available for these deposits,
although a radiocarbon age on a wood fragment, underlying two latest
Pleistocene tephra layers in lake sediment at Marias Pass, provides a
minimum age of 12 194
Pleistocene glacial deposits north of Yellowstone National Park and near the
northern Absaroka Range were first described and mapped by Weed (1893) and
then later by Pierce (1973, 1979, and references therein). Licciardi and
Pierce (2018) identified three distinct phases of glaciation in the Greater
Yellowstone region during the last glacial including the early (22–18 ka),
middle (18–16 ka), and late (16–13 ka) Pinedale. While the early Pinedale
phase in the Yellowstone area occurred mainly during the Last Glacial
Maximum, defined as the period of greatest global ice volume during the most
recent glacial stage (26.5–19.0 ka; Clark et al., 2009), the middle and late
Pinedale phases clearly postdated the global Last Glacial Maximum (LGM), although they appear to
have predated the Younger Dryas interval. Terminal and recessional moraines
at the southwestern front of the northern Absaroka Range and in Paradise
Valley to the south have cosmogenic
While many investigations in western Montana have focused on reconstructing
the extent and chronology of the Pinedale glaciation, few have attempted
to describe Pinedale climate conditions. Less attention has been paid by
previous researchers to the use of paleo-glaciological methods to reconstruct late
Pleistocene climate in western Montana than to reconstruction of the extent
and chronology of past glaciation. Locke (1990) examined modern and
reconstructed late Pleistocene glacier equilibrium lines throughout western
Montana, concluding that an assumed late Pleistocene temperature depression
of 10
Modern methods used to reconstruct paleo-glaciers, particularly distributed energy–mass balance or degree-day mass balance models, have been successfully applied to sites in the middle (Laabs et al., 2006; Refsnider et al., 2008; Birkel et al., 2012; Quirk et al., 2018, 2020) and southern Rocky Mountains (Ward et al., 2009; Brugger, 2010; Brugger et al., 2019a, b; Dühnforth and Anderson, 2011; Leonard et al., 2014, 2017a). In this study we apply a modified version of the Plummer and Phillips (2003) distributed energy–mass balance model to reconstructed glaciers in the Absaroka and Lewis ranges to help elucidate climate conditions in the northern Rockies during the last glaciation.
Although terminal moraines of east-flowing glaciers in the Lewis Range are known from previous studies, they were remapped here to aid with reconstructing maximum ice extent in the Cut Bank Creek and Lake Creek valleys (Fig. 2a). Moraines in both valleys were examined in aerial imagery available in Google Earth and using 1 : 24 000-scale topographic maps. The portion of the terminal moraine north of Cut Bank Creek was mapped in the field. Moraines were identified as broad (0.5–1 km wide), looping plateaus with hummocky topography (Fig. 3) on the piedmont east of the Lewis Range and featured abundant erratic boulders on their crests.
Surficial mapping of glacial deposits within our area of interest in the Absaroka Range had been previously completed by Pierce (1979, and references therein). Mapping in the Pine Creek area was subsequently updated by Licciardi and Pierce (2008). In the field, we checked and confirmed, without modification, the moraine mapping from these previous studies.
Following moraine mapping and field verification, we selected moraines and
erratic boulders atop moraine crests for in situ cosmogenic
On moraine crests, we searched for large (
At Pine Creek in the northern Absaroka Range, where cosmogenic
All samples were prepared at SUNY Geneseo for in situ cosmogenic
We calculated cosmogenic
Moraine ages and associated uncertainties are reported as the arithmetic
mean of individual boulder exposure ages and the standard error of the mean,
respectively (as in Putnam et al., 2010; Quirk et al., 2020). We do not
account for snow cover shielding on calculated exposure ages, noting that
snow shielding corrections (
The coupled energy–mass balance and ice-flow models used in this study were originally developed by Plummer and Phillips (2003) and have been successfully used to estimate paleoclimate conditions for extinct glaciers in a variety of geologic settings (Quirk et al., 2020; Rowan et al., 2014; Leonard et al., 2014; Harrison et al., 2014; Laabs et al., 2006). Additionally, several studies have verified the model's ability to successfully predict snow accumulation (Laabs et al., 2006; Leonard et al., 2014) and melt (Quirk et al., 2020), as well as small glacier extents (Plummer, 2002) for modern conditions in the western US.
Our modeling approach is to match simulated glacier extents produced under prescribed climate perturbations relative to modern (e.g., temperature depression and precipitation change) to field evidence such as terminal and lateral moraines. In this study, we match modeled glacier shapes and thicknesses to the well-defined Pinedale maximum ice extents at Cut Bank, Pine Creek, and South Fork Deep Creek. To test the validity of the ice-flow parameters used for the Cut Bank Creek glacier detailed below, we reconstructed the undated Lake Creek glacier immediately to the south of Cut Bank at its maximum mapped extent using the same parameters. We reconstructed the Cut Bank glacier using a model spatial resolution of 180 m, while we used a resolution of 30 m for the Pine Creek and South Fork Deep Creek glaciers, which were modeled in the same domain (herein the northern Absaroka domain). We did not include Cascade Creek as a target for glacier reconstructions because mapping of the glacier's exact terminal position remains unresolved.
The energy–mass balance model calculates snow accumulation and ablation at
every cell within the model domain for the time interval of interest,
typically 1 to several years. Annual mass balance depends mostly on
precipitation and temperature, which are the principal inputs to the model.
In this study, we use an approach similar to the one used by Leonard et al. (2017a) whereby we describe the monthly spatial distribution of temperature
and precipitation at every cell across the model domain with linear
regressions of elevation and PRISM (Parameter-elevation Regression on
Independent Slopes Model; Daly et al., 2008) monthly mean climatological models. It is important to
note that in our simulations we change monthly temperature and precipitation
distributions for the entire year, while glacier mass balance is primarily
sensitive to ablation season temperatures and accumulation season
precipitation as snowfall. Secondary climate parameters include estimates of
average monthly relative humidity, cloudiness, and wind speed and are taken
directly or derived from a combination of RAWS and NOAA COOP
historical weather station data. We calculated average monthly cloudiness
for the Cut Bank and Lake Creek Canyon domains by determining the fraction
of days per month with precipitation (i.e., 0.5 cloudiness equates to 15 d of
precipitation per 30 d total). For the Pine and South Fork Deep Creek
domain, cloudiness was estimated using the ERA-Interim third-generation
(1979–2015) reanalysis (
The primary output from the energy–mass balance model is a mass balance grid
for model domain. The mass balance grid is input to the ice-flow model along
with a digital elevation model of the drainage basins. The ice-flow model
designed by Plummer and Phillips (2003) used here is similar to the
finite-element ice sheet model described by Fastook and Chapman (1989) and
follows the commonly used shallow-ice approximation. Snow and ice mass is
gained in the accumulation zone and flows along the ice-surface gradient via
deformation and sliding into the ablation zone. We run glacier simulations
to steady state where the simulated terminus stabilizes at a mapped moraine
position. We define the steady-state condition in our model runs as when the
integrated surface balance errors are less than 5 % and typically
The Cut Bank glacier required a greater value of the deformation flow coefficient compared to the steeper valley glaciers in the northern Absaroka Range. Although it is likely that the Cut Bank Creek glacier was sliding at its base, we did not account for the contribution of sliding to flow because it was likely far less than the contribution to flow by deformation as indicated by the great ice thicknesses and low surface slopes. As described previously, we also simulated the Lake Creek glacier immediately south of Cut Bank using identical ice-flow parameters to test the validity of the chosen values. Through experimentation, we tuned the ice-flow parameters to produce simulated steady-state glaciers that matched the mapped paleo-glacier thickness and shape in both valleys and thus parameterized the effects the piedmont lobe and glacier shape had on the Cut Bank glacier. (Fig. S1 in the Supplement).
The suite of moraines deposited at the mountain front in Cut Bank Creek
valley features three broad, looping plateaus with hummocky topography
separated by incised meltwater channels and outwash (Fig. 4). The suite
includes a multi-crested terminal moraine deposited beyond the
mountain front and a recessional moraine deposited near the mouth of Cut
Bank Canyon (Fig. 4). The ice-distal sector of the terminal moraine has
the highest internal relief (up to 30 m) along the portion of the moraine
south of Cut Bank Creek, with numerous closed depressions, some of which are
filled with shallow lakes. The distal slope of the moraine grades to a
broad, gently sloping outwash plain known locally as Starr School Flat,
featuring low-relief (
Surficial geologic map of the Cut Bank Creek moraine complex detailing the Pinedale ice-distal, ice-proximal, and recessional moraine extents.
The moraines delimit the size and shape of the piedmont lobes formed by
glaciers in the two valleys. At Cut Bank Creek, the maximum Pinedale
glacier, as denoted by the ice-distal moraine, extended almost 30 km from
the headwall and occupied an area of
Here we present 29 cosmogenic
Cosmogenic
Cosmogenic
Bold font indicates outlier samples.
We identified three outliers among moraine exposure ages, including samples
CB-01 and CB-12 from the Cut Bank terminal and CB-23 from the Cut Bank
recessional moraine (Table 1). Sample CB-01 is more than 8 kyr older than all
other boulder exposure ages from the terminal moraine and is therefore
interpreted to reflect inherited
Abandonment ages for the two moraines in the northern Absaroka Range at
Cascade Canyon and South Fork Deep Creek are limited by the means of two
boulder exposure ages each at 16.9
The set of bedrock exposure ages from the ice-recessional path in Pine Creek
valley includes one (PC11-03, 34.0
Model simulations were completed for the Cut Bank and northern Absaroka
model domains including four simulations matching the Cut Bank terminal
moraine (CB
Ice thickness maps generated from coupled energy–mass balance and ice-flow modeling for
Multiplicative precipitation factors and temperature depressions, both with respect to modern, that produced modeled ice extents matching field mapped extents for (1) Cut Bank terminal moraine, (2) northern Absaroka Range Pinedale maxima at Pine Creek and South Fork Deep Creek, and (3) Cut Bank recessional moraine.
As previously mentioned, to match the modeled glacier shape to field
evidence at Cut Bank, we found it necessary to effectively set the
contribution of ice velocity due to sliding to zero. To test how realistic
these model conditions were for reconstructing other glaciers, we
reconstructed the Pinedale glacier that occupied Lake Creek Canyon, the
drainage immediately to the south of Cut Bank. We matched the modeled
glacier to the mapped Pinedale maximum in Lake Creek Canyon with
The
Moraines in the northern Absaroka Range have exposure ages that fall within the middle Pinedale interval of 18–16 ka, as identified in the Greater Yellowstone region by Licciardi and Pierce (2018) and after the end of the global LGM (26.5–19.0 ka; Clark et al., 2009). During this time, the Yellowstone glacier system thickened across the Yellowstone Plateau, coalesced with ice masses in some neighboring mountains (such as the Beartooth, High Absaroka, and Gallatin ranges), and formed large outlet lobes, including the northern outlet that terminated just south of the glaciated portion of the northern Absaroka Range (Licciardi and Pierce, 2008, 2018). This large glacier system persisted after the southwestern margin of the Laurentide Ice Sheet in northern Montana began retreating (Dalton et al., 2020) and middle latitudes in the Northern Hemisphere began warming (Shakun et al., 2015). Licciardi and Pierce (2018) suggest that enhanced westerly airflow into the region during the middle Pinedale interval combined with orographic effects of the thickened ice cap augmented precipitation in the northern Yellowstone region. The strengthened westerly airflow across the region likely impacted valley glaciers in the northern Absaroka Range, providing sufficient moisture for glaciers to persist at their maximum lengths despite rising summer insolation at middle latitudes (Laskar et al., 2004) and atmospheric carbon dioxide concentrations (Lüthi et al., 2008). Additionally, middle latitudes in North America may have remained cold for several millennia after the Laurentide Ice Sheet began retreating, as suggested by the persistence of other Rocky Mountain glaciers at near-maximum extents until 17 ka (Laabs et al., 2020) and model-based estimates of the regional temperatures at 17 ka (Liu et al., 2009; He, 2011).
The terminal and recessional moraines in the Lewis Range have exposure ages that also fall within the middle Pinedale interval of 18–16 ka and thus may also have been responding to similar climatic controls as in the Absaroka Range to the south. Alternatively, the post-LGM age of these moraines could be related to the Lewis Range's proximity to the southwestern margin of the Laurentide Ice Sheet. When the Shelby Lobe and other southwestern outlets of the Laurentide Ice Sheet were at their maximum extent, general circulation modeling studies suggest that a large area of high atmospheric pressure developed across the western dome of ice sheet, resulting in anticyclonic easterly airflow along the southern margins (Thompson et al., 1993; Bartlein et al., 1998). This circulation pattern likely resulted in a cold and dry climate in the Lewis Range while the southwestern outlets occupied their terminal moraines. Recent reconstructions of this sector of the Laurentide Ice Sheet suggest that the Shelby Lobe retreated to the northeast by ca. 17 ka (Dalton et al., 2020), which may have been accompanied by a weakening of easterly anticyclonic circulation at the latitude of the Lewis Range and strengthening westerly airflow that delivered moisture-laden air and enhanced precipitation in the mountains. Enhanced precipitation may have resulted in glacier advance to their maximum lengths after the Laurentide Ice Sheet began to retreat. This effect has been suggested by previous studies, including earlier interpretations of the moraine chronologies in the northern Yellowstone region (Licciardi et al., 2001) and age limits on moraines elsewhere in northern interior mountains (Licciardi et al., 2004; Thackray et al., 2004). Licciardi and Pierce (2018) note that the range of terminal-moraine exposure ages in the Yellowstone region includes some that overlap with the early Pinedale interval of 22–18 ka, which includes the latter part of the global Last Glacial Maximum when some southwestern outlets of the Laurentide Ice Sheet were at their maximum size. While the effect of the southwestern Laurentide on regional airflow may not have impacted the Yellowstone region, it may have impacted the Lewis Range as indicated by the exposure ages of the terminal and recessional moraines in Cut Bank valley. Additional age limits on moraines in the Lewis Range, other mountains in northwestern Montana, and the Shelby Lobe will aid in understanding the relative timing of mountain and continental glaciation.
Considering the glacial chronologies presented here in a larger spatial context, the exposure ages of terminal and recessional moraines show some consistency with mountain glacier moraines from elsewhere in the western United States. Elsewhere in the Rocky Mountains, moraines with age limits of ca. 18–17 ka are found in the Sawtooth Range in Idaho (Thackray et al., 2004), the Wasatch and Uinta Mountains in northern Utah, and numerous glacial valleys in the southern Rocky Mountains in Colorado (Leonard et al., 2017a; Brugger et al., 2019a, b; Schweinsberg et al., 2020). Where sequences of moraines are exposure-dated in the Rocky Mountains, the outermost moraines of the last glaciation generally have ages that fall within the early Pinedale interval of 22–18 ka and inner moraines (representing near-maximum glacier lengths) that fall within the middle Pinedale interval (Quirk et al., 2020; Laabs et al., 2020). This pattern is observed throughout the Rocky Mountains and suggests that the mountain glacier moraine chronology in western Montana differs from the rest of the region such that the outermost moraines do not represent the early Pinedale interval and only represent the middle Pinedale interval, suggesting similar or more extensive ice during the middle compared to the early Pinedale. This may reflect the importance of regional climatic effects on mountain glaciation, especially the strengthening of westerly airflow and attendant moisture delivery, as described above.
Glacier modeling results yielded a series of
Mumma et al. (2012) presented a paleoclimate record developed from Lower Red
Rock Lake in southwestern Montana, alongside a synthesis of other lacustrine
records from the region, spanning approximately from the entire LGM time
interval (i.e., 26–19 ka) through the early Holocene. The Lower Red Rock Lake chronology is constrained by several
Reconstructions of the valley glacier that occupied Big Timber Canyon in the
Crazy Mountains of western Montana by Murray and Locke (1989) provide
additional limits for regional climate during the last glacial culmination.
Their glacier model experiments, specifically the low mass balance gradients
derived from them, indicate that the climate in the northeastern Crazy Mountains
was typical of a cold, dry continental interior, with around 75 % of
modern precipitation, when the glacier reached its maximum size, although
the specific timing of the glacier maximum here is unknown. Additional work
by Locke (1990, 1995) on paleo-glacier reconstructions suggests that
last-glaciation equilibrium line altitudes (ELAs) were
While a unique temperature–precipitation combination for the culmination of
the Pinedale maximum in western Montana is difficult to infer from glacier
modeling results presented here, the consistent timing of the glacial
culmination at 18–17 ka – after the Laurentide Ice Sheet began retreating
and global LGM – suggests that a regional increase in precipitation during
the middle Pinedale interval supported glacier maxima. This is consistent
with inferred climate for the last glaciation in the Greater Yellowstone
region described by Licciardi and Pierce (2018) and earlier studies
inferring that glaciers in northern mountains in the conterminous western
United States reached the maximum size after the Laurentide Ice Sheet began
retreating (Thackray, 2008), as well as the regional airflow pattern implied
by the paleo-glacier reconstructions of Locke (1990, 1995) and pollen records
reported by Mumma et al. (2012). If strengthened westerly airflow at 18–17 ka resulted in accumulation season precipitation similar to modern amounts
as suggested by regional climate proxies and model output, then a regional
temperature depression can be inferred from glacier modeling results
presented here. Model simulations of glaciers in the Pine Creek and South
Fork Deep Creek valleys suggest a temperature depression of 8.5
Ice-margin retreat rates following the abandonment of Pinedale maximum extents in the northern Rockies are constrained by the cosmogenic exposure age chronology of glacially scoured and striated bedrock from Pine Creek Canyon in the northern Absaroka Range (Fig. 2). First, we emphasize the uncertainty associated with this deglacial chronology from the exclusion of three assumed old and one young outlier from the data set. Furthermore, the sample transect only captures a northern tributary of the Pine Creek glacier (see sample transect in Fig. 2) and may thus not be representative of the larger main-valley glacier system. However, few glaciated valleys in the northern Rockies have age controls sufficient to estimate retreat rates; therefore, the data presented here, while limited, are valuable for inferring rates of deglaciation. Keeping these considerations in mind we can use the data to describe the pattern of deglaciation in the northern Absaroka.
We adapt the age–depth algorithm of Breitenbach et al. (2012) to model
glacier age–distance and elevation relationships at Pine Creek (Fig. 8).
The algorithm uses a Monte Carlo scheme to estimate age–distance and
elevation confidence intervals. We use 10 000 age model realizations with
age uncertainty perturbations equal to the exposure age analytical
uncertainties. Additionally, all realizations are required to satisfy
superposition and morphostratigraphic relationships. The results indicate that
median (95 % confidence interval) horizontal ice-margin retreat rates
range from 1.6 km ka
Age–distance
Several studies of glaciated valleys in the western US have sufficient age controls to estimate retreat rates during the last glaciation along a north–south transect of the Rocky Mountains including (Fig. 9) the Pine Creek valley reported here, the Teton Range in Wyoming (Licciardi and Pierce, 2008), the Wasatch Range and Uinta Mountains in northern Utah (Laabs et al., 2011; Munroe and Laabs, 2017; Quirk et al., 2018, 2020), the Front Range (Ward et al., 2009; Dühnforth and Anderson, 2011), Sawatch Range (Briner, 2009; Young et al., 2011; Leonard et al., 2017b; Schweinsberg et al., 2020; Tulenko et al., 2020), and San Juan Range (Guido et al., 2007) in Colorado. Here, we consider vertical retreat rates for all sites to minimize the strong effects valley slope and glacier hypsometry have on apparent rates of retreat.
Vertical glacier retreat rates exhibit no clear relationship with respect to latitude along a north–south transect from Pine Creek in southern Montana to the San Juan Range in Colorado. Retreat rates for sites in the middle of the transect (Wasatch, Uinta, Front Range) are somewhat lower than rates calculated from the remaining sites and could reflect a response to increased moisture at these latitudes during Heinrich Stadial 1 (e.g., Munroe and Laabs, 2013; Heinrich, 1988). While the timing of initial abandonment of ice-distal positions is variable at sites across the Rockies, which range from the end of the global LGM to ca. 16 ka, the broad pattern and timing of subsequent deglaciation after ca. 16 ka are similar across the Rocky Mountains (Fig. 9).
The timing of terminal-moraine abandonment is variable across the Rocky
Mountains and spans a period of around 8 kyr, beginning during the LGM and
continuing, such as in the Lewis and Absaroka ranges, into the middle
Pinedale (18–16 ka). The large range in glacier retreat from ice-distal
positions suggests diverse controlling mechanisms of initial deglaciation
across the region. However, the coherence of ice-retreat rates in the
Absaroka Range with locations across the Rockies from ca. 16 ka through the
late glacial (i.e., 19–11.7 ka; Reitner et al., 2016) suggests common factors
driving deglaciation across the region. For example, glacier retreat in the
Rocky Mountains after ca. 16 ka coincides with sustained increases in
atmospheric CO
We present cosmogenic exposure ages for moraines in the Absaroka and Lewis
ranges of Montana that indicate glacial stadials during the middle Pinedale
interval (18–16 ka) and thus after the end of the global LGM. We propose
that regionally strengthened westerly airflow and orographic effects
associated with the thickening Yellowstone ice cap nourished valley glaciers
in the Absaroka Range with precipitation and allowed glaciers to persist at
their maximum lengths despite rising summer insolation at middle latitudes
(Laskar et al., 2004) and rising atmospheric carbon dioxide concentrations
(Lüthi et al., 2008). Similarly in the Lewis Range, glaciers maintained
their maximum extents following the retreat of the Shelby Lobe of the
Laurentide Ice Sheet by ca. 17 ka (Dalton et al., 2020), which we propose
could have been accompanied by a weakening of anticyclonic circulation and
strengthening of westerly airflow that effectively increased precipitation
in the Lewis Range. If we assume that precipitation during the middle
Pinedale was similar to or slightly drier than modern, following a cold and
likely much drier than modern early Pinedale–LGM, our model simulations of
glaciers in the Absaroka Range suggest a temperature depression around
8.5–9.0
Cosmogenic
The supplement related to this article is available online at:
BJQ, EH, BJCL, and EL conceived the project with input from JL and MAP. All listed co-authors completed field mapping and sampling. BJQ, EH, and BJCL completed prep work for
The contact author has declared that neither they nor their co-authors have any competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank the reviewers for thoughtful and helpful comments, which greatly improved the paper. We also thank Doug Steen and Alec Spears for help sampling the South Fork Deep Creek and Cascade Creek moraines as well as with preliminary glacier model results.
This paper was edited by Claudio Latorre and reviewed by two anonymous referees.