ArticlePDF Available

Chlorine-36 Surface Exposure Dating of Late Holocene Moraines and Glacial Mass Balance Modeling, Monte Sierra Nevada, South-Central Chilean Andes (38°S)

Authors:

Abstract and Figures

The development of robust chronologies of Neoglaciation from individual glaciers throughout the high-altitude Andes can provide fundamental knowledge of influences such as regional temperature and precipitation variability, and aid in predicting future changes in the Andean climate system. However, records of Late Holocene glaciation from the Central Chilean Andes are sparse, and often poorly constrained. Here, we present 36Cl surface exposure ages, dendrochronologic constraints, and glacial mass balance modeling simulations of Late Holocene glacier fluctuations in the Central-South Chilean Andes. A series of concentric moraine ridges were identified on Monte Sierra Nevada (38°S), where exposure dating of basaltic boulders was used to establish a chronology of ice recession. We infer that moraine abandonment of the most distal ridge in the valley commenced by ∼4.2 ka, and was followed by glacier margin retreat to an up-valley position. Exposure ages of the oldest Late Holocene boulders (∼2.5–0.8 ka) along the marginal extents of the moraine complex indicate fluctuations of the glacier terminus prior to ∼0.65 ka. A final expansion of the ice margin reoccupied the position of the 4.2 ka moraine, with abatement from the outermost composite moraine occurring by ∼0.70 ka, as constrained by tree-ring data from live Araucaria araucana trees. Finally, a series of nested moraines dating to ∼0.45–0.30 ka, formed from a pulsed ice recession during the latest Holocene when the lower reaches of the glacial snout was most likely debris mantled. A distributed temperature index model combined with a glacier flow model was used to quantify an envelope of possible climatic conditions of Late Holocene glaciation. The glacial modeling results suggest conditions were ∼1.5°C colder and 20% wetter during peak Neoglaciation relative to modern conditions. These records also suggest a near-coeval record of Late Holocene climate variability between the middle and high southern latitudes. Furthermore, this study presents some of the youngest 36Cl exposure ages reported for moraines in the Andes, further supporting this method as a valuable geochronologic tool for assessing Late Holocene landscape development.
Content may be subject to copyright.
Chlorine-36 Surface Exposure Dating
of Late Holocene Moraines and Glacial
Mass Balance Modeling, Monte Sierra
Nevada, South-Central Chilean
Andes (38°S)
Brittany N. Price
1
*, Nathan D. Stansell
1
, Alfonso Fernández
2
, Joseph M. Licciardi
3
,
Alia J. Lesnek
3
,
4
, Ariel Muñoz
5
,
6
,
7
, Mary K. Sorensen
1
, Edilia Jaque Castillo
8
, Tal Shutkin
9
,10
,
Isabella Ciocca
2
and Ianire Galilea
8
1
Department of Earth, Atmosphere and Environment, Northern Illinois University, Dekalb, IL, United States,
2
Department of
Geography, Mountain GeoScience Group, Universidad de Concepción, Concepción, Chile,
3
Department of Earth Sciences,
University of New Hampshire, Durham, NH, United States,
4
School of Earth and Environmental Sciences, Queens College, City
University of New York, Queens, NY, United States,
5
Institute of Geography, Ponticia Universidad Católica de Valparaíso,
Valparaíso, Chile,
6
Center for Climate and Resilience Research CR2, Santiago, Chile,
7
Center for Climate Action, Ponticia
Universidad Católica de Valparaíso, Valparaíso, Chile,
8
Department of Geography, Universidad de Concepción, Concepción,
Chile,
9
The Ohio State University, Byrd Polar and Climate Research Center and Department of Geography, Columbus, OH,
United States,
10
The Ohio State University, Department of Geography, Columbus, OH, United States
The development of robust chronologies of Neoglaciation from individual glaciers throughout the
high-altitude Andes can provide fundamental knowledge of inuences such as regional
temperature and precipitation variability, and aid in predicting future changes in the Andean
climate system. However, records of Late Holocene glaciation from the Central Chilean Andes
are sparse, and often poorly constrained. Here, we present
36
Cl surface exposure ages,
dendrochronologic constraints, and glacial mass balance modeling simulations of Late
Holocene glacier uctuations in the Central-South Chilean Andes. A series of concentric
moraine ridges were identied on Monte Sierra Nevada (38°S), where exposure dating of
basaltic boulders was used to establish a chronology of ice recession. We infer that moraine
abandonment of the most distal ridge in the valley commenced by ~4.2 ka, and was followed by
glacier margin retreat to an up-valley position. Exposure ages of the oldest Late Holocene
boulders (~2.50.8 ka) along the marginal extents of the moraine complex indicate uctuations
of the glacier terminus prior to ~0.65 ka. A nal expansion of the ice margin reoccupied the
position of the 4.2 ka moraine, with abatement from the outermost composite moraine occurring
by ~0.70 ka, as constrained by tree-ring data from live Araucaria araucana trees. Finally, a series
of nested moraines dating to ~0.450.30 ka, formed from a pulsed ice recession during the
latest Holocene when the lower reaches of the glacial snout was most likely debris mantled. A
distributed temperature index model combined with a glacier ow model was used to quantify
an envelope of possible climatic conditions of Late Holocene glaciation. The glacial modeling
results suggest conditions were ~1.5°C colder and 20% wetter during peak Neoglaciation
relative to modern conditions. These records also suggest a near-coeval record of Late
Holocene climate variability between the middle and high southern latitudes. Furthermore,
this study presents some of the youngest
36
Cl exposure ages reported for moraines in the
Edited by:
Jacob M. Bendle,
University of Northern British
Columbia, Canada
Reviewed by:
Joanna Charton,
Aix-Marseille Université, France
Lucas O. Bianchi,
Consejo Nacional de Investigaciones
Cientícas y Técnicas (CONICET),
Argentina
*Correspondence:
Brittany N. Price
bprice@niu.edu
Specialty section:
This article was submitted to
Quaternary Science, Geomorphology
and Paleoenvironment,
a section of the journal
Frontiers in Earth Science
Received: 04 January 2022
Accepted: 06 June 2022
Published: 05 July 2022
Citation:
Price BN, Stansell ND, Fernández A,
Licciardi JM, Lesnek AJ, Muñoz A,
Sorensen MK, Jaque Castillo E,
Shutkin T, Ciocca I and Galilea I (2022)
Chlorine-36 Surface Exposure Dating
of Late Holocene Moraines and Glacial
Mass Balance Modeling, Monte Sierra
Nevada, South-Central Chilean
Andes (38°S).
Front. Earth Sci. 10:848652.
doi: 10.3389/feart.2022.848652
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486521
ORIGINAL RESEARCH
published: 05 July 2022
doi: 10.3389/feart.2022.848652
Andes, further supporting this method as a valuable geochronologic tool for assessing Late
Holocene landscape development.
Keywords: cosmogenic nuclides, alpine glaciers, paleoclimate, little ice age, temperature index model, moraine
chronology, dendrochronology, glacial geomorphology
INTRODUCTION
The Central-South Chilean (CSCh, Table 1) Andes (~37.5°S)
marks a critical climatic and geographic boundary between the
low and high latitudes (Muñoz et al., 2016), thus studying glacial
uctuations in this region can provide information about past
inter-hemispheric oceanic and atmospheric dynamics (Liu and
Yang, 2003). To the south, glacial mass balances in the Northern
Patagonian Andes are largely dominated by the strength of Pacic
Ocean westerlies, El Niño Southern Oscillation (ENSO), and the
South Pacic Subtropical High (SPSH) (Aravena and Luckman,
2009;Garreaud et al., 2009;Garreaud et al., 2013). Likewise, the
Southern Annular Mode (SAM) affects the position and intensity
of the westerlies, summer temperatures, and rainfall distribution
(Gillett et al., 2006;Reynhout et al., 2019). Disentangling these
complex systems is critical for our ability to predict future
changes in the Andean environment and global climate system
in response to natural and anthropogenic forcings. Our research
aims to evaluate the climatic conditions driving glacial growth
and decay in the CSCh Andes over the Late Holocene, and assess
whether discreet leads and lags in the Andean cryosphere system
are discernable regionally over the Neoglacial period.
Glacial melt is an important source of water in CSCh, so
properly understanding the conditions and locations of glacier
survival under future warming scenarios is pivotal to mitigating
current water usages for sustainability (Huss and Fischer, 2016;
Barcaza et al., 2017;Racoviteanu et al., 2021). In this context, it is
unclear how individual glaciers might respond to localized
climatic effects vs. more regional or synoptic-scale processes
(Rupper and Roe, 2008), as future climate change scenarios
will not impact all regions of the Andes synchronously, or at
the same rate or magnitude. Furthermore, other landscape-
specic questions have yet to be addressed, such as the extent
of debris coverage on regional glaciers in the recent past, an
important factor for understanding mountain water storage.
Our ability to link Chilean glacial variability and specic
environmental forcings will be improved with a robust
chronology of past events. At present, individual records are
often broadly bracketed by both minimum and maximum ages,
resulting in chronologies that lack the resolution necessary to
fully decern the local stratigraphy. By integrating surface
exposure dating and dendrochronology (Harrison et al., 2008;
Nimick et al., 2016), we aim to better constrain the timing of Late
Holocene glacial uctuations from CSCh.
Various chronologies of glacial advances have been proposed
to reect the datasets gathered from southernmost South
America. For example, Mercer (1976) originally proposed
three distinct Neoglacial events in the Chilean Andes, while
Aniya (2013) posited ve Neoglacial events over the same
interval, with subtle differences in the timing of glacial
advances and retreat. Nevertheless, glacier uctuations in the
CSCh Andes and northern Patagonia still lack detailed regional
records to constrain large scale recessional timing during
Neoglaciation, and particularly the Late Holocene. A more
recent examination of glacial records from northern Patagonia
dates the greatest ice readvances from the Late Holocene to
between 21 ka and 0.50.2 ka, with the largest overall
advances since the mid-Holocene occurring at ~5 ka (Davies
et al., 2020).
Recent developments in cosmogenic dating now allow direct
age determination on Neoglacial moraine boulders, including the
Southern Hemisphere expression of the LIA, dened as
0.670.17 ka (García-Ruiz et al., 2014;Solomina et al., 2015;
Jomelli et al., 2016). Exposure ages this young have more
commonly been derived from
10
Be measurements on silica-
rich rocks (Licciardi et al., 2009;Schürch et al., 2016;Dong
et al., 2017;Biette et al., 2021), and until recently dating the basalt-
rich rocks in the CSCh Andes has been seldomly attempted.
These basalts often lack well-developed phenocrysts (e.g.,
feldspars) necessary for analyzing
36
Cl in mineral separates,
which can minimize complexities and uncertainties in
36
Cl
production and exposure age interpretations (Wirsig et al.,
2016). Currently, only a few studies have applied similar
methods to date moraine boulders in volcanic landscapes (e.g.,
Alcalá-Reygosa et al., 2018), largely because of
36
Cl production
systematics that yield high uncertainties in Cl-rich rocks (e.g.,
Palacios et al., 2017). Despite challenges associated with
36
Cl
surface exposure dating, this cosmogenic nuclide is well-suited
for lithologies lacking quartz, and allows for exposure dating of
moraine boulders that in some cases were emplaced only
hundreds of years ago (Jomelli et al., 2016;Jomelli et al., 2017;
Charton et al., 2021a;Verfaillie et al., 2021).
TABLE 1 | Abbreviations used in manuscript.
Term Abbreviation
Central-South Chile CSCh
El Niño Southern Oscillation ENSO
South Pacic Subtropical High SPSH
Southern Annular Mode SAM
above sea level a.s.l
accelerator mass spectroscopy AMS
Little Ice Age LIA
Southern Westerly Wind Belt SWWB
Antarctic Circumpolar Current ACC
Global Historical Climate Network GHCN
Area Altitude Balance-Ratio AABR
equilibrium line altitude ELA
distributed temperature index model DETIM
sea surface temperatures SSTs
degree day factor DDF
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486522
Price et al. Chlorine-36 Exposure Dating Chilean Andes
Glacial studies in the Chilean Andes have also been hindered
by a lack of physically-based models that test a range of climate
drivers (Pellicciotti et al., 2014). As such, our knowledge of
fundamental inuences like temperature and precipitation
changes on individual glaciers and ice caps throughout the
Andes remain minimally constrained. Moreover, different
modeling approaches produce a diverse range of outputs at
varying resolutions, making it difcult to converge on a
standardized approach for interpreting past glacial variability
(Fernández and Mark, 2016). Nevertheless, coupling local
station data with temperature index models produces rst-
order estimates of glacial sensitivity to climate change (Hock
and Holmgren, 2005), including paleo-climate scenarios (Kull
and Grosjean, 2000). Testing and deploying a physically-based
glacier mass balance model provides a tool for addressing
outstanding questions relating to the timing and causes of Late
Holocene climate change. Climatic conditions during Late
Holocene glacial advances are assessed to have been cold and
wet, corresponding to a negative SAM phase (Moreno et al.,
2018), but the magnitude of these temperature and precipitation
changes have yet to be quantied for specic regions of the Andes.
Here we present a new glacial chronology using independently
derived tree-ring data combined with
36
Cl ages on glacial
boulders from a cirque valley atop Monte Sierra Nevada, Chile
(38°3456S, 71°3447W) (Figure 1). Located at the
southernmost expression of an area (3138°S) categorized as
hosting warm- and wet-based glaciers, the glacial mass balance
here should be sensitive to both temperature and precipitation
changes (Sagredo and Lowell, 2012). This chronology is coupled
with the Hock (1999) distributed temperature index model
(DETIM), establishing a range of possible precipitation and
temperature values at the time of moraine abandonment using
proxy and modeled data.
BACKGROUND
Geologic Setting and Geomorphology
Monte Sierra Nevada stands at 2,527 m above sea level (a.s.l.),
with ice capping the highest peaks and extending partially down
several east-facing cirque valleys (Figure 2). The site is located to
the west of the Cordillera Principal, which is made up of the
granitoid North Patagonia Batholith (The Cura-Mallin
Formation) of Cretaceous and Miocene age (Glodny et al.,
2006). Pleistocene and Holocene volcanoes (e.g., Llaima,
Lonquimay, Sierra Nevada, and Tolhuaca) make up the active
volcanic arc of the Cordillera Principal that formed on the eroded,
and relatively at-lying Miocene-aged rocks (Muñoz et al., 2011).
Regionally, this southern volcanic zone is predominantly
composed of basalts and basaltic andesites (Hickey et al., 1986;
Ferguson et al., 1992;Muñoz et al., 2011). Additionally, these
active Holocene volcanos might have deposited volcanic ash on
glaciated terrain, potentially affecting glaciersresponse to climate
changes by two competing effects associated with debris thickness
and temperature: insulation from incoming energy or
enhancement of ablation (Davies and Glasser, 2012).
In our study area, there are a total of ve morainal loops along
the valley oor over a distance of ~350 m (Figure 3). Mapped
moraines exhibit variable effects of weathering and erosion
suggesting relative age differences. We number each moraine,
starting with the oldest and most distal from the current glacier
front (M1), progressing westward to the youngest, most ice-
proximal moraine (M5) (Figure 3). Visually, moraine M2 is the
largest and most intact of the moraine complex, while moraine
M1 is the sole moraine in the valley with observed soil
development, and exhibits a subtle dual-crest to the south of
the eld site. On the ice-proximal side of moraine M5 chaotic
hummocks are observed, with glacial scree deposits further up
valley (Figure 4). Additionally, several striated boulders and
roches moutonnées are observed, both within and outside of
the moraine complex.
The geomorphology of Monte Sierra Nevada yields
information on the relative timing of glacial landform
development due to the stratigraphy observed in the valley.
For example, the distal end of the valley is truncated by a
steep cliff at the contact between the relatively at-lying
Pleistocene volcanic layers and the underlying Miocene
plutonic rocks (Figure 4). The steep topography below this
FIGURE 1 | Map of South America with Chile (beige with brown outline).
Relative positions of the winds associated with the dominant precipitation
bearing South Pacic Subtropical High (SPSH), and Southern Westerly Wind
Belt (SWWB) during the Austral summer months are indicated with black
arrows. Major oceanic currents near the study site (Antarctic Circumpolar
Current, Peru-Chile Current and Cape Horn Current) are indicated with blue
arrows. Inset box shows the proximity of the identied modern climatic
transition zone of ~37.5°S(Muñoz et al., 2016) to the nearby Araucanía region
of Chile (lled in red). A yellow star identies the location of Monte Sierra
Nevada (~38°S), and the location of the Temuco weather station used for the
modeling analysis is represented by a purple circle.
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486523
Price et al. Chlorine-36 Exposure Dating Chilean Andes
contact likely precludes the establishment of moraines from
glacial advances that were more extensive than the presumed
Late Holocene moraines that end just before a steep drop in
topography to the valley below. On Monte Sierra Nevada,
moraines are generally limited to elevations between ~1,703
and 1,722 m a.s.l (Table 2) and the moraines have since been
bifurcated by Holocene stream development.
Directly south of the valley with the mapped moraine complex is
a ridge separating it from the adjacent, topographically higher valley
(Figure 3). On the ridge separating the two valleys is a well-dened
saddle and chute, both of which are currently ice free. Along this
same ridge is a trimline indicating the most recent highest extent of
ice thickness in the valley. This trimline extends down valley
intersecting moraine M1, as well as up valley past the saddle and
chutetowardsthemountains peak. The chute exhibits a cross-
cutting relationship with the trimline, having eroded any
recognizable lateral features. Notable talus deposits are identiable
along the northern and southern valley walls, particularly below the
trimline and along the chute. The modern glacial ice cap on Monte
Sierra Nevada extends into the uppermost reaches of both valleys.
Climate and the Westerlies
Central Chile has a Mediterranean-type climate with mild, wet
winters and dry, hot summers. The modern climate is largely
FIGURE 2 | Panel (A) Google Earth map of Monte Sierra Nevada and surrounding area with sample locations marked by yellow stars. Panel (B) Geologic map,
modied from Muñoz et al. (2011).QalQuaternary alluvium, PlvcpUpper Pleistocene-Holocene volcanic rocks, TcMiocen e Cura-Mallin Formation, TmMiocene
Melipeuco Plutonic Group, KgCretaceous Galletué Plutonic Group.
FIGURE 3 | Panel (A) Google Earth image of the Monte Sierra Nevada eld site, with the modern glacier outlined in blac k and the moraine complex outlined in red.
Panel (B) Ariel photograph of the eld site as delineated by red box in panel A. Map identies glacial moraines (M1-M5), locations of samples taken for
36
Cl (yellow circles
with abbreviated sample number) and tree ring (green circle) dating.
36
Cl ages are reported in thousand years before present with internal uncertainty. Probable ice ow
directions are represented by black arrows based on observed striated rock surfaces (white circles). Trimline along the southern ridge of the valley is marked (blue
dashed line), as well as the saddle and chute (black line).
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486524
Price et al. Chlorine-36 Exposure Dating Chilean Andes
modulated by Pacic and Southern Ocean processes, resulting in
a strong seasonality of precipitation south of ~35°. winter storms
dominate here, bringing ~2,000 mm of annual rainfall to the
region (Sinclair, 1996;Garreaud et al., 2009). However, Monte
Sierra Nevada (38°S) is located near what is considered to be a
transition zone between Mediterranean-subtropical and
temperate-humid climatic zones, identied at ~37.5°S
(Figure 1;Muñoz et al., 2016). Precipitation at this location is
therefore impacted by shifts in the subtropical South Pacic
Subtropical High (SPSH) and extratropical Southern Westerly
Wind Belt (SWWB) year-round (Miller, 1976;Espinoza et al.,
2020).
Glaciers in CSCh predominantly receive precipitation sourced
from the Pacic Ocean due to the moisture barrier produced by
the Andes mountains (Montecinos and Aceituno, 2003). Isotopic
analyses of precipitation suggest that Atlantic-derived moisture
rarely penetrates through to the western anks of the Andes
(Hoke et al., 2013). While regionally both the height and width of
FIGURE 4 | Photographs of the Monte Sierra Nevada eld site. Panel (A) Facing east the hummocky areas of the eld site are outlined on the glacier proximal side of
moraine M5 (purple). Panel (B) Facing west with the modern glacier in the background and contact between Pleistocene and Miocene volcanic rocks (orange dashed
line) in the foreground behind waterfall. The cliff marks the limit of preserved Neoglacial moraine bould ers, with moraine M1 (blue line) denoted to the south of the waterfall.
The combined effects of topoclimatic constrains (i.e., lower elevations, seasonally constrained snow accumulation) make the existence of large glaciers in CSCh
unlikely, as Neoglacial landforms are generally indistinguishable below this abrupt change in elevation. A grove of A. araucana trees is pictured to the south of the waterfall
in the foreground. Panel (CF) Striated boulders and roches moutonnées used for directional data to infer paleo-owlines of the glacier. Inset map in panel A shows the
location of each eld picture, with letters corresponding to each panel.
TABLE 2 | Sample information used in the calculation of the
36
Cl boulder age expressed with internal uncertainty, calculated with a rock formation age of 130 ka using a
developmental version of the CRONUS-Earth
36
Cl calculator (Balco et al., 2008; http://stoneage.ice-d.org/math/Cl36/v3/v3_Cl36_age_in.html) and LSDn scaling. The
ICE-D calibration datasets used can be found at http://calibration.ice-d.org/allcds.
Sample ID Moraine
ID
Sample
thickness
(cm)
Latitude
(°N)
Longitude
(°S)
Elevation
(m asl)
Topographic
Shielding
correction
36
Cl
concentration
36
Cl
Boulder
age (ka)
AFCL18-05 M5 1.25 38.58572 71.55426 1721 0.9618 26,455 ± 2,525 0.81 ± 0.09
AFCL18-01 M4 0.4 38.58654 71.55409 1722 0.9414 64,419 ± 2,819 2.6 ± 0.1
AFCL18-07 M4 0.75 38.58526 71.55244 1712 0.9678 6,087 ± 783 0.42 ± 0.06
AFCL18-08 M4 1 38.58487 71.55235 1711 0.9734 19,023 ± 2,354 0.46 ± 0.07
AFCL18- 09 M3 1 38.58441 71.55126 1704 0.9726 5,490 ± 631 0.40 ± 0.05
AFCL18-11B M3 1.25 38.58258 71.55158 1704 0.9689 13,408 ± 1850 0.28 ± 0.05
AFCL18-12 M2 0.75 38.58132 71.55244 1720 0.8741 58,565 ± 3,247 2.1 ± 0.1
AFCL18-13 M2 1.25 38.58108 71.55307 1717 0.743 23,502 ± 2,816 0.79 ± 0.08
AFCL18-15 M1 1.5 38.58149 71.55122 1711 0.9293 14,619 ± 2,392 0.65 ± 0.12
AFCL18-16B M1 2 38.58484 71.54964 1703 0.9677 111,396 ± 5,503 4.2 ± 0.2
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486525
Price et al. Chlorine-36 Exposure Dating Chilean Andes
the CSCh Andes are smaller in comparison to the Northern
Cordillera, orographic precipitation leads to continental
precipitation values that are up to three times larger than
corresponding oceanic amounts (Garreaud et al., 2009).
In Central Chile the SPSH typically uctuates between ~25
and 35°S and delivers moist air masses to the western slope of the
Andes (Ancapichún and Garcés-Vargas, 2015). Active year-
round, the position of the SPSH is modulated by the Bolivian
High, which migrates northward during the austral winter (May-
August) as the jet stream moves equatorward. This increase in
winter storm tracks drives midlatitude precipitation along the
central Chilean coast (Sinclair, 1996;Lamy et al., 2001;Garreaud
et al., 2009;Espinoza, et al., 2020). From February-April the SPSH
shifts south, resulting in increased precipitation off the coast of
CSCh (~37°S) (Ancapichún and Garcés-Vargas, 2015). This
hydrologic regime results in an enhanced probability of aridity
in CSCh, as streamow peaks in winter due to rainfall and again
in late spring as seasonal snow melts in the mountains.
The regional climate in the CSCh Andes is also affected by the
Antarctic Circumpolar Current (ACC). The ACC splits as it
approaches the western coast of South America between ~40
and 45°S(Boltovskoy, 1976;Villalba et al., 2012), with the Peru-
Chile Current owing north and the Cape Horn Current owing
south (Lamy et al., 2007). The Peru-Chile Current transports
cold, low-saline waters equatorward (Strub, 1998), with seasonal
changes in upwelling driven by the position and intensity of the
SPSH (Garreaud et al., 2009;Ancapichún and Garcés-Vargas,
2015). The SPSH is thus one of the dominant controls on sea
surface temperatures (SSTs) associated with the Peru-Chile
Current, with the ability to increase upwelling and offshore
Ekman transport. The position of the ACC also results in
strong winds directed at the coast, in the form of the SWWB.
Situated between the subtropical anticyclone and subpolar
cyclonic air masses, these westerlies bring strong winds and
abundant precipitation to CSCh and the adjacent northern
Patagonian Andes, intersecting the western coast of South
America between 40°S-55°S(Bertrand et al., 2014). During
Austral summer the SWWB strengthens, contracts, and shifts
to the southern extent of its range in response to southern-ocean
processes, while during the winter months the SWWB weakens
and expands both poleward and towards the equator (Bentley
et al., 2009). Comparatively, the SWWB is stronger than the
storm tracks associated with the SPSH at the study location, as it
is conned to a smaller latitudinal range and more directly affects
regional precipitation.
El Niño Southern Oscillation and Southern
Annular Mode
Located in the transition zone between Mediterranean and
temperate climates, both ENSO and SAM play important roles
in driving changes in temperature and moisture availability to
Monte Sierra Nevada. ENSO is characterized by shifting oceanic-
atmospheric conditions in the tropical Pacic, the effects of which
are superimposed on the climate system of CSCh. El Niño (La
Niña) years are associated with a weakening (strengthening) of
the Peru-Chile Current, due to a reduction in the east-west
pressure gradient (Montecinos and Aceituno, 2003), forcing a
weaker SPSH into more northerly position (Villalba, 1994). This
northward displacement of the SPSH reduces moisture
availability in CSCh. When SSTs are lower the high-pressure
system strengthens, and westerly trade winds intensify in
response (Bertrand et al., 2014). In the northern portion of the
transition zone (3540°S), ENSO is characterized as having a
stronger inuence on the SPSH (Rutllant and Fuenzalida, 1991),
whereas south of ~40°S interannual precipitation driven by the
SWWB is heavily modulated by the strength/phase of the SAM
(Aravena and Luckman, 2009;Garreaud et al., 2009;Villalba
et al., 2012).
The SAM is measured by changes in sea-surface pressure
between the mid- and high-latitudes (Villalba et al., 2012;
Ancapichún and Garcés-Vargas, 2015). Both the strength and
position of the SWWB is dictated by SSTs, and sea surface
pressure gradients associated with the ACC and the Southern
Ocean, specically the SAM (Gillett et al., 2006;Reynhout et al.,
2019). Enhancement of precipitation in Patagonia south of ~55°S
is predominantly driven by a southern displacement of the
SWWB, subsequently resulting in decreased precipitation
around 3843°S. During negative (positive) phases of the SAM
the SWWB shifts equatorward (poleward), resulting in more
humid (arid) conditions in CSCh (Garreaud et al., 2009;
Bertrand et al., 2014;Muñoz et al., 2016).
Regional Records of Late Holocene Climate
Change
Proxy paleo-climate evidence suggests that Late Holocene glacial
variability in the northern Patagonia Ice Field was driven by both
precipitation and temperature (Bertrand et al., 2012). A tree-ring
reconstruction of summer temperatures from northern Patagonia
suggests there was a cold interval from ~1.1 to 0.95 ka, followed
by warming until ~0.77 ka and another phase of cool and wet
conditions peaking between ~0.68 and 0.38 ka (Villalba, 1994).
Proxy paleoclimate records also suggest that from ~0.65 ka until
the last century was the wettest interval of the last ~5,000 years in
the southern Chilean Andes (Moreno et al., 2009). Moreno et al.
(2018) strengthened this analysis, and documented that regional
glacial advances in the southern Patagonia Ice Field were forced
from negative SAM phases (cold and wet), that were modulated
on regional scales by the position and intensity of the SWWB.
Recently, Davies et al. (2020) compiled geomorphic evidence and
recalibrated chronological data from dozens of studies assessing
uctuations of the Patagonian Ice Sheet from 35 ka to the present.
Records detailing the Late Holocene from 41°Sto52
°S noted
climatic conditions alternated between cold/wet and warm/dry
on centennial time scales (Elbert et al., 2013;Álvarez et al., 2015;
Moreno and Videla, 2016;Moreno et al., 2018;Davies et al.,
2020). Well-dened Late Holocene moraines are apparent
regionally (Glasser et al., 2008;Glasser and Jansson, 2008;
Davies and Glasser, 2012;Davies et al., 2020), but these
advances overall do not seem to surpass moraines emplaced at
~5 ka. While these Late Holocene landforms exhibit clear
evidence of later formation, such as vegetation differences and
degradation (Davies and Glasser, 2012;Davies et al., 2020), it is
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486526
Price et al. Chlorine-36 Exposure Dating Chilean Andes
possible that positions of earlier advances were reoccupied
(Sagredo et al., 2016;Davies et al., 2020).
Additional records of Neoglaciation in CSCh and to the north
are also coarsely resolved. For example, glacier advances have
been dated in the nearby Rio Grande basin (35°S) on the eastern
(Argentinian) slope of the Andes: radiocarbon dating from peat
bogs in contact with terminal moraines, which are located at
elevations around 2,500 m, yielded ages between ~0.5 and 0.2 ka
(Espizua and Pitte, 2009). On Monte Tronador (41°S), located to
the south, tree-ring dating of frontal and lateral moraines
(elevations around 1,000 m) of glaciers owing toward the
eastern side of the Andes indicated ages between ~1 and
0.4 ka, with some smaller ridges suggesting advances between
the late 19th and early 20th centuries (Masiokas et al., 2009). On
the Chilean slope,
10
Be exposure dating of boulders located atop
moraines ridges (elevations between 1,700 and 2,000 m) in the
foreground of the Cipreses glacier (34°S) yielded older ages
compared to the other sites, clustered between ~1 and 0.6 ka,
but with some samples older than 8.0 ka (Sagredo et al., 2016).
Thus, while regionally distributed records suggest ice advanced at
or around the time of the LIA, comprehensive studies that
combine multiple independent dating methods, such as tree-
ring records, with direct moraine ages in the same valley are
lacking (Fernández et al., 2022).
METHODS
Mapping and Sampling
Moraines were initially mapped and targeted for sampling using
Google Earth Pro imagery. The location and extent of these
moraines was documented in greater detail in the eld utilizing
handheld GPS methods during our eld campaign of November
2018, with additional geomorphic observations and mapping of
the valley conducted in January 2022, when directional
measurements were taken from striated rock surfaces to assess
paleo ice-ow directions. Sample locations were measured with a
single-frequency GPS receiver (Emlid reach). The receiver was
placed on a rod and positioned on top of or to the side of each
boulder, and recorded continuously for 1015 min in RINEX
format. Subsequently, these data were post-processed using the
closest GPS base station available, located in Temuco, about
100 km from Monte Sierra Nevada (Figure 1). This
differential correction delivered uncertainties in the order
of ±0.8 m.
Basaltic boulders located atop moraine ridges were sampled
for surface exposure dating, with a total of 10 processed for
36
Cl
measurement. These boulders heights ranged from ~0.42 to ~2.3
m, and included basaltic conglomerates and ne-grained
crystalline basalts (Figure 5). We preferentially selected the
largest boulders perched on or near moraine crests, which
increases the likelihood that the boulders have been stable and
experienced minimal snow cover since deposition
(Supplementary Figure S1.; Ivy-Ochs and Kober, 2008;Balco,
2011). Additionally, boulders were assessed for weathering
characteristics to distinguish moraine boulders from rock fall
(Zreda and Phillips, 2000;Briner et al., 2001). No suitable sample
material for exposure dating was found on the expression of the
proximal (M5) or near proximal (M4) moraines to the north of
the stream, as an immature drainage network had eroded large
sections of these moraines (Figure 3). Samples (~1 kg) were
chiseled from the upper few centimeters of these basaltic
boulders. Topographic shielding was measured by collecting
azimuth and zenith angles every 15°at each sample location
using a Suunto tandem compass/clinometer.
36
Cl Processing
Whole rock samples were sawed to a measured thickness of
between 0.4 and 2.0 cm before crushing and sieving to collect
grains between 250 and 125 µm. Chemical preparation of samples
was conducted at the University of New Hampshire Cosmogenic
Isotope Lab using a modied version of the protocols presented in
Stone et al. (1996) and Licciardi et al. (2008). A key component of
the modied procedures includes measuring
35
Cl/
37
Cl on a ~1 g
aliquot of rock prior to
36
Cl extraction and measurement on the
full sample. The ~1 g aliquot was spiked with a
37
Cl-enriched
solution (LLNL Spike A;
35
Cl/
37
Cl = 0.93) and ~4,000 µg of Br,
which served to increase the size of the nal precipitate. Total
sample Cl was determined through isotope dilution methods.
36
Cl
was extracted as Ag(Cl,Br) from full rock samples in two
analytical batches following standard procedures. All samples
received a carrier solution containing ~4,000 μg of Br and a
natural ratio Cl carrier (
35
Cl/
37
Cl = 3.127). This technique was
used to better control the total Cl within the analytical batch and
reduce potential AMS memory effects (Arnold et al., 2010;Finkel
et al., 2013).
36
Cl Measurements and Calculations
Measurements of
35
Cl/
37
Cl and
36
Cl/
37
Cl were conducted using
the 10 MV tandem accelerator at the Lawrence Livermore
National Laboratory Center for Accelerator Mass Spectrometry
(LLNL-CAMS). Analytical uncertainty on
35
Cl/
37
Cl
measurements ranged from 0.42 to 0.45%. Analytical
uncertainty on
36
Cl/
37
Cl AMS measurements was between 4.3
and 13%. Sample
36
Cl concentrations are corrected for lab
background using an average blank value for the two
analytical batches (4.12 × 10
4
atoms
36
Cl, n=2). Major and
trace element concentrations in the samples were determined
using x-ray uorescence and inductively coupled plasma mass
spectrometry at SGS Mineral Services in Burnaby, British
Columbia, Canada. Analytical data used to determine surface
exposure ages are provided in Table 3.
Moraine boulder exposure ages were calculated using a
developmental version of the CRONUS-Earth
36
Cl calculator
(Balco et al., 2008; http://stoneage.ice-d.org/math/Cl36/v3/v3_
Cl36_age_in.html) and LSDn scaling (Table 2;Lifton et al.,
2014). The
36
Cl calibration datasets applied by the calculator
can be found in the ICE-D database at http://calibration.ice-d.
org/allcds. Based on eld observations, no corrections were made
for rock surface erosion, vegetation cover, or snow coverage. A
bulk density value of 2.7 was used for all samples. All cosmogenic
ages here are presented as years prior to sampling (AD 2018).
We made corrections for nucleogenic
36
Cl production using a
rock formation age of 130 ka (Muñoz et al., 2011). Nucleogenic
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486527
Price et al. Chlorine-36 Exposure Dating Chilean Andes
36
Cl is produced in moderate amounts when
35
Cl absorbs
neutrons released by the decay of U and Th isotopes (Gosse
and Phillips, 2001). In rocks with high Cl, U, or Th
concentrations, nucleogenic
36
Cl can thus constitute a
substantial proportion of the total
36
Cl. In typical exposure age
calculations, the amount of nucleogenic
36
Cl in a sample is
calculated assuming steady state production, and that value is
subtracted from the total
36
Cl to obtain the cosmogenic
component (Marrero et al., 2016b). However, in young rocks
(<1 Ma formation age), production and decay of nucleogenic
36
Cl
may not be in equilibrium, and erroneously assuming steady state
could lead to uncertain and/or inaccurate surface exposure age
determinations (e.g., Sarikaya et al., 2019;Anjar et al., 2021). To
assess the sensitivity of our
36
Cl ages to the time of rock
formation, we also calculated
36
Cl ages assuming a rock
formation age of 30 ka. Calculated surface exposure ages
presented in the results and discussion of this study use a rock
formation age of 130 ka to represent the last stage of volcanism on
Monte Sierra Nevada based on lack of visual evidence of younger
vents or deposits to support a formation age of 30 ka (Muñoz
et al., 2011). Surface exposure ages representative of a younger
date of rock formation of 30 ka are presented in Supplementary
Table S1, as a possible late Pleistocene formation age has not been
dismissed by available evidence.
Dendrochronology
Dendrochronology was used to obtain an approximate date of
establishment for Araucaria araucana (Pewén) trees growing on
the most distal moraine in the valley, to provide an independent
minimum age of moraine formation and deglacial processes. To
assess the minimum age of the moraine, increment borers were
used to collect samples from mature A. araucana trees atop the
outermost moraine (M1) of the proglacial area of the study site
(Figure 3). These trees were selected with consideration to
morphological attributes such as bark, height, diameter, as well
as size and form of branches, all visual characteristics associated
with old A. araucana trees. Two cores per tree from A. araucana
(n=35) were extracted from living trees growing on the site using
the aforementioned criteria, forming the NEVchronology.
Samples were taken at 1.3 m from the soil, to assess a
minimum age of the trees. Tree cores were then sanded for
visual identication of growth rings (Stokes and Smiley, 1996), in
FIGURE 5 | Representative basaltic boulders (conglomerates and ne-grained crystalline) sampled for
36
Cl surface exposure dating.
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486528
Price et al. Chlorine-36 Exposure Dating Chilean Andes
order to be counted and measured in the Laboratory of
Dendrochonology and Environmental Studies at Ponticia
Universidad Católica de Valparaíso, Chile (https://www.pucv.
cl/uuaa/site/edic/base/port/dendrolab.html). Ring width was
measured to ±1 µm resolution using a Velmex system (https://
www.velmex.com/), after which the COFECHA software
(Holmes, 1983) was used to assess and corroborate the dating
of each tree using the ring width measurement patterns across the
suite of tree cores. Of the 35 sampled trees, 30 were able to be
properly cross-dated. From these 30 trees, 20 exhibited curvatures
close to the pith. We selected the ve oldest trees that exhibited
this curvature, suggesting a minimal age close to the total age of
the tree, and using the Duncan method (Duncan, 1989) estimated
the number of missing rings. The age of the oldest tree was used to
provide an independent and cross-dated minimum age of the
glacial landform from which the trees had colonized.
Glacial Modeling
We applied a modied version of DETIM by Hock (1999), which
is available at http://regine.github.io/meltmodel/. As inputs, we
used a climate le from the Global Historical Climatology
Network (GHCN) (https://www.ncei.noaa.gov/products/land-
based-station/global-historical-climatology-network-daily) with
daily data from AD 20042013 (Temuco station, Figure 1), as
well as the gridded NASADEM surface. The model also requires
gridded surfaces for aspect and slope values that were calculated
using the spatial analyst tools in ArcGIS 10.4.1. The model ran on
the Cygwin64 C compiler following the approach of Stansell et al.
(2022). The model ran with degree-day factors (DDFs) for ice and
snow of 7 and 3.5 mmw. e. K
1
d
1
, values typically used for
Chilean glaciers (Möller and Schneider, 2010).
The climate input le requires some pre-processing of climate
data, including the calculation of a precipitation gradient and
temperature lapse rate. The precipitation amount was set at a
value of ×1.5 the values recorded from the Temuco station, which
is similar to what is observed at other GHCN stations of similar
elevation offsets in the region. No lapse rate was applied to the
precipitation amounts along the glacier prole.
The model also requires an estimated temperature offset
(Tdiff) corresponding to the location of the oldest mapped
moraine (M1), which we independently determined by nding
the difference between present (pr) and past (pa) Equilibrium
Line Altitudes (ELAs) and applying a simple linear relationship
using a lapse rate of 6.5°C/100 m:
Tdiff ELApr ELApa×0.0065
We used the Area Altitude Balance-Ratio (AABR) method
(Osmaston, 2005) to determine ELAs. The AABR method
estimates the ELA using the hypsometry of the glacier or the
proglacial area within the limits of former ice extents, usually
demarked by moraines. We used the 12.5 pixel size DEM
available from ALOS PALSAR imagery as a source of
topographical information to build the hypsometric curve
(ASF DAAC, 2015), and a 3-m resolution true-color
composite from the Planet Constellation to interpret the
glacier margin in 2013 and to derive ELA
pr
. For the past
hypsometry, we trimmed the ALOS DEM extending the
current margin by following the pattern of the trimline to the
moraine crest closest to the cliff (M1). For the calculation, the
method iteratively searches the elevation of the zero-balance,
assuming that the mass balance gradient is linear, with a steeper
slope in the ablation zone (Kaser and Osmaston, 2002). A key
component of this approach is to select a suitable ratio between
mass balance gradients, which can be computed from mass
balance observations and assumed to hold in the past. In our
case we face the challenge that almost no mass balance data exists
for the region. The few direct measurements available from the
Mocho Glacier (Rivera et al., 2005;Schaefer et al., 2017), situated
about 115 km to the south, suggest a balance ratio of 3.2, larger
than gures commonly used, such as 1.75 or 2 (Osmaston, 2005).
A recent study found that a value of 1.56 represents glaciers at a
global scale, although with important spread (Oien et al., 2021).
Guided by these studies, we tested several ratios, opting for 2.38 as
a best approximation to represent this region. We notice,
however, that the difference in ELAs between ratios of 1.56
and 3.2 is 47 and 77 m for the present and the past,
respectively, which translates to an uncertainty of 0.2°C. Using
the median AABR value of 2.38 results in an ELA
pr
and an ELA
pa
of ~2,267 m a.s.l. and ~2029 m a.s.l. This change in ELA (ΔELA)
of -238 m translates to T
diff
of ~1.5°C.
TABLE 3 | Elemental composition measured on whole rock samples after leaching. Major elements are in wt%, and trace elements in ppm unless noted otherwise.
Major elements Trace elements
Sample ID Al
2
O
3
(wt%)
CaO
(wt%)
Fe
2
O
3
(wt%)
K
2
O
(wt%)
MgO
(wt%)
MnO
(wt%)
Na
2
O
(wt%)
P
2
O
5
(wt%)
SiO
2
(wt%)
TiO
2
(wt%)
Cl
(ppm)
B
(mg/kg)
Sm
(ppm)
Gd
(ppm)
U
(ppm)
Th
(ppm)
AFCL18-01 17.84 7.17 9.46 1.15 2.98 0.14 4.11 0.03 55.51 1.1 93.85 15 1.8 2.23 0.81 2
AFCL18-05 18.6 8.68 9.71 0.61 5.05 0.15 3.18 0.14 51.8 0.92 163.39 13 3.1 3.59 0.52 1.6
AFCL18-07 21.37 9.18 7.06 0.8 2.22 0.11 3.92 0.02 53.52 0.88 22.49 14 3.2 3.55 0.67 1.8
AFCL18-08 17.41 7.99 10.31 0.8 5.38 0.16 3.26 0.18 51.95 1 205.26 14 3.7 3.6 0.7 2.3
AFCL18-09 20.94 9.04 7.34 0.81 2.39 0.12 3.84 0.02 53.3 0.9 20.02 29 3.1 3.27 0.76 1.6
AFCL18-11B 19.79 8.83 9.09 0.84 3.32 0.14 3.65 0.15 52.04 1.03 219.01 15 3.6 4.13 0.75 1.6
AFCL18-12 19.7 8.88 9.63 0.59 4.55 0.15 3.45 0.02 51.73 0.87 128.51 15 2.8 3.2 0.38 1.1
AFCL18-13 19.66 8.84 8.66 0.88 3.6 0.14 3.49 0.08 52.68 0.88 172.85 16 3.3 3.68 0.6 1.6
AFCL18-15 17.21 5.96 9.57 1.24 2.4 0.16 4.61 0.02 56.36 1.34 122.71 19 5.1 5.34 0.9 2.4
AFCL18-16B 17.94 8.37 10.24 0.71 5.49 0.16 3.39 0.15 52.03 1.03 159.74 17 3.6 4.15 0.5 1.3
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 8486529
Price et al. Chlorine-36 Exposure Dating Chilean Andes
We then ran the DETIM with multiple climate scenarios. First,
the model was run over the interval from A.D. 2004 to 2013 with
no change in temperature and precipitation relative to today.
Next, we ran the model with 1.5°C cooling and a 20% increase in
precipitation as a rst approximation to simulate Neoglacial
conditions that are represented by the AABR method. We
assumed wetter conditions were associated with ice advances
because proxy and modeling studies suggest that this region had
increased precipitation amounts at times when it was colder
during the Late Holocene (Bertrand et al., 2014). In order to
test glacier sensitivity, we also ran the model with temperature
changes of +1°C with a doubling of precipitation amounts. A
series of other tests were run by changing only one variable at a
time using ±1°C change in temperature, ±20% change in
precipitation, and lapse rates ranging from 0.55°C/100 to
0.75°C/100. Additionally, to test the inuence of DDFs, we
adjusted the ice and snow values by ±1, following the
approach of Möller et al. (2007).
Finally, we applied the ow model of Plummer and Phillips
(2003) to simulate glacial thickness using the input of the mass
balance grid. While the mass balance model identies areas of
ablation and accumulation, the ow model predicts the shape,
size, and extent that a glacier stabilizes at for those conditions at
steady-state. The model applies the standard ow law for ice
using the shallow ice approximation, and using multiple
interpolation methods. The results of the combined energy
mass balance and ow models are shown in Figure 6. For all
model runs, we present the lowest altitude of ice limit, maximum
and average ice thickness values (Table 4).
RESULTS
Boulder
36
Cl Ages
The surface exposure ages of moraines on Monte Sierra Nevada
range from 4.2 ± 0.2 ka to 0.28 ± 0.05 ka, and represent multiple
phases of glacial growth and decay (Figure 7;Table 2). Sample
AFCL18-16B, located on the most distal moraine in the valley,
M1, is dated to 4.2 ± 0.2 ka. A second boulder sampled from M1,
AFCL18-15, yielded a younger age of 0.65 ± 0.12 ka. Two samples
dating to 2.1 ± 0.1 and 0.79 ± 0.08 ka are located on moraine M2,
near the cliff on the north end of the valley. On the center moraine
of the complex, M3, samples dated to 0.39 ± 0.05 and 0.28 ±
0.05 ka. Three samples from M4 yielded exposure ages of 2.6 ±
0.1, 0.46 ± 0.07, and 0.42 ± 0.06 ka. On M5, the innermost
moraine in the valley, a single sample dated to 0.82 ± 0.09 ka. No
apparent age biases (e.g., too young or too old) were detected
between the ne-grained crystalline vs the basaltic conglomerate
source rocks.
Dendrochronology
The cross-dating process resulted in an inter-correlation value of
0.55, a high correlation for A. araucana ring width chronologies
(Muñoz et al., 2014). The ve oldest trees comprising the NEV
chronology, which showed a complete curvature close to the pith,
were established between ve and six hundred years ago. The rst
rings in samples from this group of trees were developed around
the years AD 1407, 1462, 1511, 1559, and 1609. With the addition
of missing rings due to the exhibited curvature (Duncan, 1989),
the rst ring for each of these ve trees results in an approximate
year of AD 1318, 1378, 1482, 1535, and 1580, respectively. The
oldest tree of the NEV chronology yielded a date of AD 1318,
providing a minimum limiting age of moraine M1 at 0.70 ka. This
tree cross-dated to 0.54 with the complete set of trees measured, a
value close to the mean inter-correlation value at this site,
corroborating the age of this tree. Finally, due to the height of
the extraction (1.3 m from the ground), a number of rings were
missing with respect to the complete age of the trees, leading to an
underestimation of the age of each tree. As such, a number of
years, not determinate, should be added on to the determined age
after ring counting, necessary to obtain the growth height
observed. Following Lusk and Le Quesne (2000), the mean
height growth rates of A. araucana were 169 mm/year in sun-
grown trees, and 111 mm/year in shade-grown individuals.
Considering these estimations, a minimum period of around
10 years should be added to the age of each tree. However,
depending on competition and site requirements, this period
could be longer than 10 years.
Glacial Modeling
The combined glacier mass balance and ow models produced
simulated ice extents that mimic the modern and paleo-
conditions. Using the GHCN temperature and precipitation
values produced an ice margin at an elevation (~1905 m) that
is similar to today. Using a cooling of 1.5°C and a 20% increase in
precipitation produced a modeled ice extent that is similar to the
outer-most mapped and dated moraines on Monte Sierra Nevada
(Figure 6). As the moraine complex represents an elevation
change of only ~19 m over a distance of ~350 m,
distinguishing between moraines on the gridded volume
proves difcult. In general, the model produces the thinnest
ice near moraine positions and thicker ice elsewhere. Glacial
coverage on Monte Sierra Nevada also tends to be thicker on the
eastern slopes that are less steep than the western slopes. For
modeling purposes, we assumed that periods of colder
temperatures were associated with wetter conditions over the
last several millennia (see Discussion). Nevertheless, the
sensitivity analysis indicates that varying temperature alone by
±0.5°C shifted the elevation of the lowest ice position by ~±130 m,
while adjusting precipitation by ±20% shifted the ice margin by
~40 and +125 m (Table 4). Shifting the lapse rate by ±1.0°C/
100 m offset the lowest ice position by 230 and +330 m. Lastly,
shifting the DDF values by ±1 offset the lowest ice limit by +120
and 110 m.
DISCUSSION
36
Cl Dating of Late Holocene Moraine
Boulders
The exposure ages presented here are some of the youngest to be
developed using
36
Cl methods on glacial boulders in the Andes.
While our dataset includes only a modest number of ages, these
results demonstrate the feasibility of using
36
Cl to date young
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 84865210
Price et al. Chlorine-36 Exposure Dating Chilean Andes
FIGURE 6 | Google Earth image of Monte Sierra Nevada with ice thickness results from the combined mass balance and ow models superimposed. The yellow
circles mark the locations of dated moraine positions from this study, the modern glacial outline is represented in green. From top left to bottom middle. (A) Temperature
increase of 0.5°C. (B) Precipitation increase of 20%. (C) Lapse rate of 0.75°C/100 m. (D) Temperature decrease of 0.5°C. (E) Precipitation decrease of 20%. (F) Lapse
rate of 0.55°C/100 m. (G) DDF of +1. (H) DDF of -1.
TABLE 4 | Parameters used for modeling of the glacial mass energy balance sensitivity analysis.
Variable Lowest ice limit
(m asl)
Max. thick (m) Avg. thick (m)
baseline conditions 1.5°C, +20% precip, Lapse rate: 0.65°C/100 m 1,600 115 6
+0.5°C1,725 100 5
0.5°C1,530 125 8
+20% precip 1,560 120 8
20% precip 1,725 110 5
Lapse rate: 0.75°C/100 m 1,370 130 13
Lapse rate: 0.55°C/100 m 1,930 75 2
Degree-day factor +1 1,720 100 7
Degree-day factor 1 1,490 120 10
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 84865211
Price et al. Chlorine-36 Exposure Dating Chilean Andes
glacial deposits of similar compositions elsewhere in the Chilean
Andes. Due in part to the small number of boulders, denitive
and discreet ages cannot be condently assigned to each moraine
position. However, the chronology established for this moraine
complex provides clear documentation for multiple phases of
retreat and readvance of Monte Sierra Nevada glacier during the
Late Holocene.
Multiple sources of uncertainty must be considered when using
36
Cl to date Late Holocene moraine boulders. Processes such as
post depositional erosion and boulder rotation can contribute to
uncertainties in surface exposure dating (e.g., Balco, 2020). In
addition,the presence of nucleogenic (i.e., non-cosmogenic)
36
Cl in
rock surfaces can introduce considerable uncertainty to exposure
age calculations. In a setting such as Monte Sierra Nevada, where
the target landforms for dating and the lithologic units are likely
relatively young, low-energy neutron capture can also produce
non-negligible amounts of
36
Cl. In exposure age calculations, the
nucleogenic
36
Cl must be subtracted from the total
36
Cl to obtain
the concentration of cosmogenic
36
Cl. At this time,
36
Cl production
rates from low-energy neutron capture are the least well-
characterized production pathway (Marrero et al., 2016a),
partially because low-energy neutron ux depends on
difcult-to-constrain paleoenvironmental factors such as
snow cover and presence of water (Zreda et al., 1993;
Phillips et al., 2001;Schimmelpfennig et al., 2009;Zweck
et al., 2013;Dunai and Lifton., 2014). Improved constraints
onthetimingofvolcanicepisodes of Monte Sierra Nevada
will help to fully evaluate the nucleogenic production of
36
Cl
in the Monte Sierra Nevada moraines, as these Pleistocene
deposits have likely not reached steady-state production for
nucleogenic
36
Cl (Sarikaya et al., 2019;Anjar et al., 2021).
With these considerations in mind, we applied recent
advances in
36
Cl methods to produce robust ages for the
last several millennia.
Mid and Early Late Holocene Moraine
Chronology
The cosmogenic
36
Cl ages from Monte Sierra Nevada suggest
glaciers in the CSCh Andes uctuated throughout the latest
Middle and Late Holocene. Ice margin retreat from the oldest
and most distal moraine, M1, likely commenced by ~4.2 ± 0.2 ka
based on sample AFCL18-16B (Figure 3). A second boulder
sampled from M1 to the north of the stream yielded a much
younger age (AFCL18-15; 0.65 ± 0.12 ka), which aligns with the
establishment of A. araucana trees on this glacial landform by
0.70 ka. Additional exposure samples are needed from moraine
M1 to further assess the potential of nuclide inheritance due to
unknown timing of boulder stabilization over melting debris
covered ice, as well as to produce a statistically robust age
determination of this landform. However, the morphology of
the M1 moraine might aid in deciphering this complex glacial
setting. The subtle duel-crested nature of the southern extent of
M1 and observed pedogenesis on the moraine surface (Jenny,
1944;Mason and Jacobs, 2018) may suggest that M1 demarcates
an expansion of the ice front prior to 4.2 ± 0.2 ka, and again
before 0.65 ± 0.12 ka. Prior to 0.65 ± 0.12 ka the glacier likely
expanded to the position of moraine M1, resulting in
reoccupation of this landform, abutted the ice-proximal
expression of M1 formed in the mid-Holocene, and shaped
the dual-crest observed on the southernmost extent. The
glacier then commenced its nal retreat from the M1 position
by ~0.65 ± 0.12 ka.
The composite nature of moraine M1, with a potential
formation in the mid-Holocene followed by a latest Holocene
glacial advance reoccupying this landform, would allow for a
period of several thousands of years for soil development to occur.
While soil may develop rapidly on deglaciated terrain, this
process is largely dependent on environmental factors such as
humidity and precipitation, as well as weathering rates of the
original parent material (Mason and Jacobs, 2018). The lack of
soil development on moraines M2-M5, along with an initial
colonization age of moraine M1 by A. araucana trees of
0.70 ka, suggests that attributing a younger moraine formation
age of ~0.65 ± 0.12 ka would preclude sufcient time for
pedogenesis to support the colonization of A. araucana on
that landform.When combined, these independent datasets
suggest glaciers were at more advanced positions in the region
just prior to ~4.2 ka, and supports assigning M1 an initial mid-
Holocene formation age.
FIGURE 7 | Panel (A) Normalized kernel density estimates for all samples. Panel (B) Normalized kernel density estimates for samples that fall within the Southern
Hemisphere expression of the Little Ice Age. Sample ages in both panels are plotted with internal analytical uncertainties.
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 84865212
Price et al. Chlorine-36 Exposure Dating Chilean Andes
The trimline abutting M1 along the southern ridge of the
valley suggests these features formed concurrently, and while no
samples were dated from up-valley of the moraine complex, we
suggest the trimline formed contemporaneously with M1. The
continuation of the trimline along the southern ridge up-valley of
the chute suggests that ice from the cirque directly up-valley from
the dated moraines contributed to the formation of M1 in the
mid-Holocene. However, this up-valley source of ice was likely
minimal during the Late Holocene, with ice predominantly
owing over the southern ridge from the adjacent valley, as
interpreted from the cross-cutting stratigraphy of the chute
eroding a portion of the trimline (dashed blue, and black lines
respectively, Figure 3). The modeling results suggest that greatest
ice thicknesses in the valley were obtained when the ice margin
expanded past the mapped extent of the moraine complex, which
based on both the exposure ages and dendrochronology, has not
occurred in the Late Holocene. Thus, this expression of high ice
thickness in the valley likely formed during a period of glacial
expansion during the mid-Holocene, predating glacial retreat
from the M1 position at ~4.2 ± 0.2 ka. The glacier then retreated
to an up-valley position before retreat halted. During this interval
of glacial recession, ice from the cirque likely retreated up valley,
separating from the glacial tongue owing over the saddle on the
southern ridge.
Four samples (AFCL18-01; 2.6 ± 0.1 ka, AFCL18-05: 0.81 ±
0.09 ka, AFCL18-12; 2.1 ± 0.1 ka, and AFCL18-13; 0.79 ± 0.08 ka)
date from 2.6 ± 0.1 ka to 0.79 ± 0.08 ka, and are located at the
northern and southernmost extents of the moraine complex
(Figure 3). The exposure ages obtained from these samples
support a uctuating ice margin, and initial exposure from the
subglacial environment predating a nal glacial advance in the
valley during the latest Holocene. The mapped locations of these
samples at the outermost extents of the moraines on which they
are perched suggests reworking and transport to their present
locations by one or more readvances of the glacier. Additionally,
the older exposure ages of these marginal samples compared to
samples closer to the center of the moraine complex supports the
likelihood of multiple stages of ice marginal expansion. Ice ow
directions assessed from striated surfaces (Figure 3) indicate a
radial advance of glacial ice in the valley, which would have
obliterated previously formed moraines, and redeposited these
boulders in more lateral positions of the moraine complex. These
samples may have originally been deposited on the valley oor
during recession of the ice margin, or were potentially exposed
from subglacial conditions on the southern ridge. Additionally,
these samples may have variable amounts of
36
Cl inheritance,
which could explain some of the scatter seen in the available ages.
Finally, these boulders may have experienced destabilization and
subsequent rotation as underlying ice melted, an inherent pitfall
to interpreting surface exposure ages on ice-cored moraines.
Latest Holocene Moraine Chronology
The youngest surface exposure ages (AFCL18-07, 0.42 ± 0.12 ka;
AFCL18-08, 0.46 ± 0.07 ka; AFCL18-09, 0.40 ± 0.05 ka; AFCL18-
11B, 0.28 ± 0.05 ka; and AFCL18-15, 0.65 ± 0.12 ka), and the
development of moraines M2-M5 provide evidence for multiple
phases of ice expansion in the valley over the latest Holocene.
Two scenarios are plausible to explain the formation of moraines
M2-M5 between this nal glacial advance and the modern. The
rst would entail a glacier with minimal to no debris coverage
commencing a continuous pulsed retreat of the ice marginal
environment from 0.65 ± 0.12 ka through the modern, emplacing
moraines M2-M5. However, it is more likely that during this
interval, and perhaps the entirety of the Holocene, that the lower
portion of the glacier was debris-mantled, as debris coverage on
many regional glaciers has been increasing in modern times
(Glasser et al., 2016). If the glacial snout was indeed debris-
mantled, the response of the ice marginal environment to
climate-driven changes could be more muted (Kirkbride and
Warren, 1999;Naito et al., 2000;Davies and Glasser, 2012).
Critically thin ice margins along the frontal edge of the
glacier likely resulted in instability of slope angles on moraine
M1, and potentially along the northern and southernmost
lateral positions of the other moraines in the complex,
leading to partial or full collapse and deposition of
supraglacial debris. Several exposure ages towards the center
of moraines M3 (0.40 ± 0.05 ka) and M4 (0.42 ± 0.12 ka and
0.46 ± 0.07 ka) are in close agreement. The location of these
samples may have been better insulated from more rapid
uctuations of the ice margins at the time collapse was
initiated, as surface melt rates often vary spatially
depending on the thickness of debris coverage (Östrem,
1959;Moore, 2018). The timing between the nal
abandonment of moraine M1 at 0.65 ± 0.12 ka and the
exposure of boulders on the central portions of moraines
M3 and M4 at ~0.4 ka may represent an interval when
climatic conditions were cooler and wetter, resulting in
temporary stabilization of the ice margin at moraine M3.
Furthermore, the cluster of ages on moraines M3 and M4
between 0.46 ± 0.07 ka and 0.40 ± 0.05 ka might suggest that
latest Holocene warming and/or drying resulted in ice surface
lowering, the melt out of supra- and englacial debris, and the
continued recession of the ice margin up valley. However, as
debris can either increase insulation or enhance ablation of the
glacial ice depending on both the thickness and extent of
coverage (Davies and Glasser, 2012;Bartlett et al., 2021),
further data and modeling are needed to substantiate these
tentative interpretations.
On the ice-proximal side of moraine M5 to the north of the
stream, hummocks were surveyed (Figure 3,Figure 4). The area
over which hummocks are observed has been substantially
dissected by an immature drainage network that has also
eroded large sections of moraines M4 and M5. Development
of this chaotic terrain suggests a rapid retreat from the location of
moraine M5 (Owen and England, 1998;Figure 3). The
development of hummocks also supports the idea of a debris-
covered glacier in the lower reaches of the valley, particularly
along the glacial snout (Anderson, 2000;Bartlett et al., 2021). An
abundance of observed glacial scree and talus deposits in the
valley, particularly along the southern ridge of the valley, would
provide sufcient source material to generate debris coverage of
the snout of the paleo-glacier (Shakesby et al., 1987;Giardino and
Vitek, 1988;Moran et al., 2016;Charton et al., 2021b). As the talus
is most extensive beneath the trimline and covering the chute
Frontiers in Earth Science | www.frontiersin.org July 2022 | Volume 10 | Article 84865213
Price et al. Chlorine-36 Exposure Dating Chilean Andes