PUBLICATIONS

Geophysical Research Letters

RESEARCH LETTER Rapid Last Glacial Maximum deglaciation in the Indian
10.1002/2015GL066077
Himalaya coeval with midlatitude glaciers: New insights
Key Points:
• Exposure ages of ice-polished surfaces
from 10Be-dating of ice-polished bedrock surfaces
similar to moraine boulders
• Extensive MIS-2 (LGM) glaciation at the
in the Chandra Valley, NW Himalaya
southern front of the Indian Himalaya Patricia Eugster1, Dirk Scherler2,3, Rasmus C. Thiede1, Alexandru T. Codilean2,4, and Manfred R. Strecker1
• Himalayan LGM deglaciation coeval
with midlatitude LGM deglaciation 1
Institute of Earth and Environmental Science, University of Potsdam, Potsdam, Germany, 2Earth Surface Geochemistry,
German Research Centre for Geosciences GFZ, Potsdam, Germany, 3Institute of Geological Sciences, Freie University Berlin,
Supporting Information: Berlin, Germany, 4Now at School of Earth and Environmental Sciences, University of Wollongong, Wollongong, New South
• Text S1, Figures S1–S18, Tables S1 and Wales, Australia
S2, and Captions for Data Sets S1–S3
• Data Set S1
• Data Set S2
• Data Set S3
Abstract Despite a large number of dated glacial landforms in the Himalaya, the ice extent during the
global Last Glacial Maximum (LGM) from 19 to 23 ka is only known to first order. New cosmogenic 10Be
Correspondence to: exposure ages from well-preserved glacially polished surfaces, combined with published data, and an
P. Eugster, improved production rate scaling model allow reconstruction of the LGM ice extent and subsequent
eugster@geo.uni-potsdam.de
deglaciation in the Chandra Valley of NW India. We show that a >1000 m thick valley glacier retreated
>150 km within a few thousand years after the onset of LGM deglaciation. By comparing the recession of
Citation: the Chandra Valley Glacier and other Himalayan glaciers with those of Northern and Southern Hemisphere
Eugster, P., D. Scherler, R. C. Thiede,
A. T. Codilean, and M. R. Strecker (2016), glaciers, we demonstrate that post-LGM deglaciation was similar and nearly finished prior to the Bølling/Allerød
Rapid Last Glacial Maximum deglaciation interstadial. Our study supports the view that many Himalayan glaciers advanced during the LGM, likely in
in the Indian Himalaya coeval with response to global variations in temperature.
midlatitude glaciers: New insights from
10
Be-dating of ice-polished bedrock
surfaces in the Chandra Valley, NW
Himalaya, Geophys. Res. Lett., 43, 1. Introduction
1589–1597, doi:10.1002/2015GL066077.
The behavior of glaciers is an important climate proxy for changes in humidity and temperature in high
Received 4 SEP 2015
mountain ranges, where other climate archives are generally limited. Because the response time of glaciers
Accepted 25 JAN 2016 to varying climate conditions is on the order of tens to hundreds of years, glacial chronologies based on
Accepted article online 28 JAN 2016 dated moraines are frequently used to infer paleoclimatic conditions [e.g., Putnam et al., 2010]. In the
Published online 22 FEB 2016
Himalaya, previous studies suggested strong sensitivity of glaciers to variations in precipitation and thus to
orbitally driven monsoon intensity. Additional factors may constitute strong east-west gradients in moisture
sources, with western areas being influenced by the midlatitude westerlies [e.g., Benn and Owen, 1998] and
Northern Hemisphere climate oscillations [e.g., Dortch et al., 2013; Owen and Dortch, 2014]. However, despite
a rich collection of >1800 cosmogenic exposure ages that mostly stem from moraines [Dortch et al., 2013;
Murari et al., 2014; Owen and Dortch, 2014], it has proven difficult to unambiguously identify the nature of
the climatic controls on glacier fluctuations during the late Quaternary period. For example, robust data on
advances of Himalayan glaciers during the global Last Glacial Maximum (LGM), here defined as the time per-
iod of maximum global ice volume from 19 to 23 ka during Marine Isotope Stage 2 (MIS2) [Mix et al., 2001;
Lisiecki and Raymo, 2005], are limited. This particularly applies to the monsoon-influenced sectors of
the Himalaya where early last glacial, late glacial, and early Holocene advances are commonly recognized
[e.g., Owen, 2009; Scherler et al., 2010; Murari et al., 2014]. Possible reasons for the conundrum of scarce LGM
advances include (1) asynchronous or no glacial advances due to steep climatic gradients [Owen et al., 2005]
or hypsometric effects [Pratt-Sitaula et al., 2011]; (2) glacial advance due to the impact of rock avalanches and
transient increases in debris cover [e.g., Gardner and Hewitt, 1990; Tovar et al., 2008; Jamieson et al., 2015]; (3) large
age uncertainties due to unstable till deposits [Applegate et al., 2009; Heyman et al., 2011]; and (4) lack of ade-
quate calibration sites and variations between scaling schemes [Chevalier et al., 2011; Heyman, 2014]. Although
asynchronous glacial advances due to topographic effects or steep climatic gradients may exist, it is not very
likely that these effects exclusively occur in the Himalaya; yet, so far, there exists limited evidence from other
regions. Rock avalanches could trigger asynchronous glacial advances, but these are likely short-lived, local,
©2016. American Geophysical Union.
and subordinate for large glaciers. In contrast, it is well known that erosion rates in the Himalaya are high
All Rights Reserved. [Godard et al., 2014; Olen et al., 2015], which makes erosive degradation of moraines a reasonable explanation.

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1589

 Geophysical Research Letters 10.1002/2015GL066077

10
Figure 1. Sampling locations and Be exposure ages of this study. (a) Study area. (b) Upper Chandra Valley with mapped
landforms, flow directions of glacial striations [Owen et al., 1996, 2001; this study] and modern glaciers [Pfeffer et al., 2014].
10
White stars and white bottom up triangles indicate sampling locations and corresponding Be exposure ages of this
10
study. Black triangles indicate sampling locations of previous studies [Owen et al., 1996, 2001]. (c) Comparison of Be
exposure ages and uncertainties obtained from boulders (white squares) and from glacially polished surfaces black (this
study)/grey [Owen et al., 2001]. The numbers correspond to the sampling location in Figure 1b.

Furthermore, there still exist no cosmogenic nuclide calibration sites in the Himalaya, leading to additional
methodological age uncertainties. However, recently published new calibrations, including low-latitude
and high-altitude sites elsewhere [e.g., Kelly et al., 2013], newly compiled calibration data sets [Heyman,
2014; Borchers et al., 2016], and improved insights into discrepancies between production rate scaling models
[Lifton et al., 2014], furnish an improved framework for cosmogenic nuclide exposure dating in the Himalaya.
We revisited the Chandra Valley in the Lahul region, NW Himalaya, where pioneering work by Owen et al. [1995,
1996, 1997, 2001] has established the timing of Late Pleistocene glacial advances mainly based on dated
boulders. By dating ice-polished, glacially striated bedrock surfaces and reconstructing former ice extents, we
are able to refine the existing glacial chronology subsequent to the LGM and suggest that the timing and pace
of LGM deglaciation in the Chandra Valley and other Himalayan regions is similar to midlatitude glaciers, and
primarily a response to increased global temperatures.

2. Study Area
The Chandra Valley, a tributary of the Chenab Valley, lies at >3000 m elevation and is surrounded by peaks
higher than 6000 m elevation. The bedrock in the Lahul area comprises Neoproterozoic to Permian granitic
intrusions and metasedimentary rocks of the High Himalaya and Tethyan sequences [Steck, 2003]. Among
numerous smaller glaciers, the most extensive glaciers in the upper Chandra Valley are Samundar Tapu
(86 km2) [Pfeffer et al., 2014] and Bara Shigri (130 km2) [Pfeffer et al., 2014] (Figure 1). Trimlines, U-shaped
valleys, and dated glacial features attest to a major trunk-valley glaciation, previously referred to as the
pre- or syn-LGM Chandra and Batal Glacial stages. The latter stage is manifested by pronounced trimlines
and landforms (Batal I) overlain by younger drumlins that indicate readvances (Batal II) along the Chandra
and Bhaga Valleys [Owen et al., 1995, 1997, 2001], reassigned to 15.3 ± 1.6 ka [Murari et al., 2014]. Subsequent
advances during the Kulti glacial stage are related to tributary glaciers and have been attributed to the
Early Holocene [Owen et al., 1995, 1997, 2001] but redefined to the Late Glacial [Murari et al., 2014] and are
interpreted as evidence for climatic forcing by the South Asian summer monsoon [Owen et al., 2001] or the
midlatitude westerlies [Murari et al., 2014].

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1590

 Geophysical Research Letters 10.1002/2015GL066077

Figure 2. Chandra Valley glacier extent during LGM. (a) The Chandra Valley with reconstructed ice extents during the LGM and
the subsequent deglaciation. Colors correspond to retreating positions in Figure 2c. White stars and white bottom up triangles
indicate our sampling locations. Modern glaciers from Pfeffer et al. [2014] (b) Profiles a-a′, b-b′, and c-c′ show sample locations
above the present valley floor. The reconstructed ice thicknesses are shown according to colors in Figure 2c. We included
samples from neighboring locations. (c) Retreat history of LGM glacier in the Chandra Valley from Udaipur to the Baralacha La
10 14
from >20 ka to <15 ka, reconstructed by Be data [Owen et al., 2001; this study] and C data [Rawat et al., 2015]. Mean retreat
ages are combined from different studies [Owen et al., 1996, 2001; this study] by the landform they stem from (in parentheses,
the number of ages used for the mean age) and are color coded according to the retreating position they belong to. Broken lines
in the reconstructions indicate the unclear terminus because of not taking the joining of the Bhaga Glacier and Bara Shigri
Glacier into account. Confluence Chandra Valley with main tributary valleys or glaciers indicated by an arrow. CBaV = confluence
with Bhaga Valley, CBSG = confluence with Bara Shigri Glacier, CSTG = confluence with Samundar Tapu Glacier.

3. Methods
We collected 15 bedrock samples from well-preserved glacially polished surfaces and three samples from boulders
resting on these surfaces for cosmogenic 10Be exposure dating (see supporting information for field photographs
and detailed description of sampling locations). After standard mineral separation steps [e.g., Kohl and Nishiizumi,
1992], we extracted Be by ion exchange chromatography at the German Research Centre for Geosciences GFZ
in Potsdam. 10Be/9Be ratios were measured by accelerator mass spectroscopy at the University of Cologne
[Dewald et al., 2013]. We used the CRONUS-Earth Web Calculator (http://web1.ittc.ku.edu:8888/1.0/) hosted at
the University of Kansas, which is based on the calibration data set compiled by Borchers et al. [2016], for
calculating exposure ages for both the new and previously published [Owen et al., 2001] 10Be concentrations.
This calculator uses a new scaling model by Lifton et al. [2014] (later also referred to as LSD scaling) that is
specific to 10Be and accounts for its production rate sensitivity on the incident cosmic ray energy spectrum,
instead of the cosmic ray flux-based scaling utilized previously. The cosmic ray flux-based scaling schemes
do not account for nuclide-specific differences in production rate sensitivities. The CRONUS-Earth Web

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1591

 Geophysical Research Letters 10.1002/2015GL066077

Calculator (hosted by the University of

Washington) was also used to recal-
culate all previously published ages
discussed; this information is presented
in Figures 1–3 (supporting information
provides the references of Figure 3).
The reported exposure ages are based
on the production rate scaling model
by Lifton et al. [2014] and a 10Be half-life
of 1.387 ± 0.012 Ma [Chmeleff et al.,
2010; Korschinek et al., 2010]. We used
the nuclide-specific attenuation lengths
described by Lifton et al. [2014] with
145 g/cm2 for altitudes between 0 and
2 km and 160 g/cm2 for >3 km eleva-
tions and the thickness correction
included in the CRONUS-Earth Web
Calculator [Marrero et al., 2016]. In addi-
tion, we also calculated all of our ages
with the CRONUS online calculator v2.2
using the calibration of Balco et al.
[2008] and a more extensive calibration
data set compiled by Heyman [2014].
Topographic shielding was measured
in the field with a hand-held compass
using the model by Dunne et al.
[1999]. Snow cover shielding was not
accounted for because our sampling
sites comprise steep valley flanks and
wind-exposed ridges, where thick snow
cover is unlikely to remain for long.
We reconstructed the surface profile
of the former glacier stepwise from
the terminus up-glacier with a simple
model that assumes a perfectly plastic
ice rheology and a constant driving
Figure 3. Comparison of deglaciation and climate proxies of the Northern stress [Benn and Hulton, 2010]: hi + 1 = hi
and Southern Hemispheres. Glacier retreat indicated as up-valley distance
+ (fτ D/H)i(Δx/ρg), where hi is the ice sur-
in percent from location of oldest LGM record to the present-day glacier
terminus or if vanished to the catchment boundary. Smooth line helps face elevation at node i, H is the ice
to identify the proposed long-term retreat, in which subtleties of minor thickness, τ D is the driving stress, Δx is
glacial advances and retreats may be hidden. Exposure ages on the x axis the node distance, ρ is the ice density
were recalculated using the CRONUS-Earth Web Calculator hosted at the
(900 kg/m3), g is acceleration by gravity
University of Kansas. YD = Younger Dryas. BA = Bølling/Allerød. (a) NGRIP
18
from Greenland project, Gulyia ice core, (b) Mamwluh Cave δ O record; (9.81 m/s2), and f is a dimensionless
effective moisture record (c) Northern insolation; (d) Epica ice dome; shape factor that accounts for valley-
(e) Himalayan glacier retreat; (f) Northern Hemisphere glacier retreat; side drag and is calculated from H, the
(g) Southern Hemisphere glacier retreat. Full reference list is provided in cross-sectional area of the valley A, and
the supporting information.
the ice-covered perimeter p, according
to f = A/Hp. This is a 2-D model that
neglects tributary glaciers, allowing for the estimation of the ice surface profile of the former trunk glacier,
constrained by ice-polished and dated surfaces, moraines, and trimlines. We tested different values of τ D
and obtained the best matches with our field constraints using τ D = 50 kPa for the maximum ice extent.
The longitudinal valley profile and the shape factors (0.4–0.5 in the Chandra Valley) were measured from a

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1592

 Geophysical Research Letters 10.1002/2015GL066077

90 m resolution Shuttle Radar Topography Mission digital elevation model [Jarvis et al., 2008] using MATLAB
and the TopoToolbox v2 [Schwanghart and Scherler, 2013].

4. Results
4.1. Surface Exposure Dating
Our new exposure ages range between ~14 and ~20 ka (Table 1 and Figure 1b). Older ages are obtained from
locations with elevations >700 m above the present-day valley floor, while younger ages are obtained from
lower elevations. At the Kunzum La (La = pass) (Figure 2b, Profile a-a′), glacial striations indicate that ice was
flowing eastward into the Spiti Valley [Owen et al., 1997; Saha et al., 2015], crossing the drainage divide of the
Chenab and Sutlej watersheds. Ice-polished bedrock surfaces record ice-free conditions by approximately
17.6 ± 1.2 ka (16.0 ± 1.2 ka, 19.2 ± 1.2 ka), which is consistent with striations in the upper Spiti Valley dated
at 17.4 ± 1.3 and 18.3 ± 1.3 ka. While the previously dated and recalculated surface at the Kunzum La of
18.0 ± 1.3 ka [Owen et al., 2001] is consistent with our data, the boulders dated in that study yield exposure
ages of 19.7 ± 1.2 ka, 19.0 ± 1.2 ka, and 18.6 ± 1.2 ka [Owen et al., 1996, 2001], indicating possibly minor inheri-
tance or a readvance. On a bedrock ridge at the Bara Shigri/Chandra confluence (Figure 2b, Profile b-b′) at
~4600 m elevation, glacial striations occur ~700 m above the present valley floor on extensive ice-polished
surfaces. A trimline at 4800–4900 m elevation separates rugged hillslopes from ice-polished surfaces with
a mean age of 17.6 ± 2.4 ka (19.3 ± 1.2 ka, 15.9 ± 1.2 ka). Two boulders located on this ridge yield ages of
17.8 ± 1.3 ka, consistent with the average surface age, and 32.5 ± 2.3 ka, clearly indicating inheritance.
Striated surfaces at lower elevation from the opposite valley side are located 30–150 m above the valley floor
and yield a mean age of 15.8 ± 1.5 ka (15.6 ± 1.2 ka, 14.5 ± 1.1 ka, 17.4 ± 1.3 ka). These results reflect rapid glacier
retreat, with the ice thinning by >500 m within ~2 ka. Finally, at a location ~15 km farther downvalley
(Figure 2b, Profile c-c′), glacial striations are ~500 m above the valley floor and four samples collected over
~100 m in elevation yield mean exposure ages of 19.3 ± 1.2 ka (18.3 ± 1.3 ka, 19.7 ± 1.2 ka, 19.9 ± 1.2 ka,
19.3 ± 1.2 ka). We found no more striations farther downstream.
Ages calculated with the CRONUS Earth web calculator using the scaling model by Lifton et al. [2014] are
generally older than ages calculated with the CRONUS online calculator [Balco et al., 2008] and Heyman’s
[Heyman, 2014] calibrations. The ages are on average ~25% older comparing Lal/Stone time-dependent ages
by Balco et al. [2008] and on average ~11% older comparing Lal/Stone time-dependent ages by Heyman
[2014] with ages from Lifton et al. [2014] for the Himalayan glaciers (see also Table S2 and Figures S17 and S18
in the supporting information).

4.2. Glacier Reconstruction and Deglaciation History

Based on the spatial distribution of our new and previously published exposure ages we were able to recon-
struct the ice extent in the Chandra Valley between ~20 ka and 15 ka. Our reconstruction suggests that during
the LGM, the Chandra Valley was occupied by a glacier up to 1000 m thick (see Figure 2a), which we refer to as
the Chandra Valley Glacier or CVG. Both our new ages and reconstruction confirm that glacial ice was crossing
major drainage divides such as Kunzum La (~600 m of ice thickness above the present drainage divide)
eastward into the Spiti Valley [Saha et al., 2015] and the Rothang La (~400 m ice thickness above the pass)
southward into the Beas Valley, also supported by glacial striations [Owen et al., 2001] (Figure 1b). Although
no terminal moraines of the former CVG are preserved, likely due to postglacial fluvial erosion in the narrow
and deeply incised Chenab Valley, there is evidence that the glacier reached at least the village of Rape [Owen
et al., 1997]. Our reconstruction of the ice surface suggests that the glacier extended even beyond Udaipur,
reaching a length of ~200 km. However, our reconstruction does not take into account the joining of the
Bhaga arm (>100 km) [Owen et al., 1997, 2001] into the trunk valley glacier, which may have resulted in an
even longer CVG.
Prior to 19–18 ka, the ice still occupied two drainage divides, but rapid melting had started (Figure 2).
Between 17 and 15 ka, the trunk-valley glacier had retreated to a length of ~70 km. After 15 ka, the main trunk
valley must have been mostly ice-free, as organic sediments and peat started accumulating near Chandra
Lake at ~12.9 ± 0.2 ka (radiocarbon age of 11.0 ± 0.1 ka) [Owen et al., 1997; Rawat et al., 2015], which is close
to the present terminus of the modern Samundar Tapu Glacier. At the Pleistocene-Holocene transition,
tributary valley glaciers readvanced at least once into the trunk valley at around 13 ka (recalculated from

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1593

EUGSTER ET AL.
Table 1. CRN Data of This Study
Production Rate 10
1 Sample Be Be Exposure
Sample Location Atoms (g qz) 10 9 e f,g f,h
Elevation Thickness Density Shielding Denudation Weight Carrier Be/ Be Concentration Age
a b c d 15 1
Nr. Name Type Trust °N °E (masl) (cm) (g/cm3) Spallation Muons Factor (mm/yr) (g) (mg) (× 10 ) atoms (g qz) (ka)

1 WP051 pbs 1 32.4164 77.6308 4760 1.5 2.65 45.806 0.357 0.9767 0 33.2900 0.2765 1771 ± 57.0 9.81E+5 ± 3.30E+4 16.0 ± 1.2
2 PE13_01 pbs 1 32.4237 77.6357 4819 3 2.6 45.685 0.355 0.9589 0 46.1187 0.1496 5588 ± 174.0 1.21E+6 ± 3.94E+4 19.2 ± 1.2
3 PE12_013 pbs 0 32.2808 77.5779 4485 3 2.65 37.271 0.300 0.9215 0 40.0977 0.3063 1545 ± 48.2 7.88E+5 ± 2.57E+4 15.9 ± 1.2
4 WP058 pbs 1 32.2822 77.5688 4491 4 2.8 39.933 0.321 0.9944 0 28.5550 0.2666 1705 ± 53.2 1.06E+6 ± 3.47E+4 19.3 ± 1.2
Geophysical Research Letters

5 WP059 bos 1 32.2823 77.5693 4499 3.5 2.8 40.258 0.324 0.9943 0 27.9600 0.2664 3211 ± 98.6 2.04E+6 ± 6.56E+4 32.5 ± 2.3
6 WP057 bos 1 32.2806 77.5676 4483 3.5 2.8 39.949 0.322 0.9944 0 26.7800 0.2772 1387 ± 43.1 9.58E+5 ± 3.12E+4 17.8 ± 1.3
7 WP052 pbs 1 32.2981 77.5538 3928 3.5 2.8 27.517 0.238 0.9048 0 28.1550 0.2746 981 ± 30.8 6.38E+5 ± 2.10E+4 17.4 ± 1.3
8 WP053 pbs 1 32.2991 77.5531 4028 4 2.8 30.527 0.260 0.9575 0 26.5150 0.2752 828.6 ± 26.9 5.73E+5 ± 1.94E+4 14.5 ± 1.1
9 WP054 pbs 1 32.2993 77.5534 4039 2 2.8 31.245 0.266 0.9578 0 28.3850 0.2752 993.2 ± 31.6 6.42E+5 ± 2.13E+4 15.6 ± 1.2
10 PE12_061 pbs 1 32.3086 77.4058 4095 3 2.65 31.553 0.267 0.9460 0 40.1750 0.3068 1521 ± 47.8 7.76E+5 ± 2.54E+4 18.3 ± 1.3
11 PE12_062 pbs 1 32.3066 77.4070 4011 3 2.65 30.078 0.257 0.9416 0 40.0993 0.3068 1615 ± 50.1 8.25E+5 ± 2.67E+4 19.9 ± 1.2
12 PE12_063 pbs 1 32.3051 77.4072 3955 3 2.65 29.534 0.254 0.9520 0 40.4024 0.3068 1531 ± 47.5 7.76E+5 ± 2.52E+4 19.3 ± 1.2

RAPID CHANDRA VALLEY LGM DEGLACIATION

13 PE12_064 pbs 1 32.3083 77.4061 4072 3 2.65 31.700 0.269 0.9617 0 40.3521 0.3334 1554 ± 48.3 8.57E+5 ± 2.79E+4 19.7 ± 1.2
14 PE12_056 pbs 1 32.4092 77.6498 4597 5 2.65 41.413 0.328 0.9815 0 40.0345 0.3062 1984 ± 61.3 1.01E+6 ± 3.27E+4 18.1 ± 1.3
15 PE12_057 bos 1 32.4093 77.6515 4612 3 2.65 42.467 0.337 0.9831 0 19.3261 0.3072 1044 ± 32.9 1.11E+6 ± 3.64E+4 19.0 ± 1.2
16 PE12_058 pbs 1 32.4085 77.6523 4608 3 2.65 42.079 0.334 0.9760 0 40.0975 0.3070 1782 ± 55.2 9.11E+5 ± 2.83E+4 16.2 ± 1.2
17 PE13_02 pbs 1 32.4507 77.6801 4568 4 2.6 38.324 0.305 0.9114 0 16.7413 0.1519 1483 ± 47.9 8.98E+5 ± 3.03E+4 17.4 ± 1.3
18 PE13_03 pbs 1 32.4540 77.6936 4184 4 2.6 33.755 0.281 0.9698 0 46.9540 0.1512 3857 ± 120.0 8.29E+5 ± 2.70E+4 18.3 ± 1.3
a
pbs = polished bedrock surface, bos = boulder on polished bedrock surface; all the tops were exposed at the surface.
b
Trust based on the quality of the sample and the possible earlier shielding due to, e.g., till cover. 1 = very good. 0 = good.
c
Production rate based on Lifton et al. [2014] as described in Borchers et al. [2016] and Marrero et al. [2016].
d
Production rate calculated as described in Marrero et al. [2016] and Phillips et al. [2016].
e 10 9 15
AMS ratios were normalised to 07KNSTD standard and corrected using full procedural blanks (n = 5) with Be/ Be ratios ranging between 0.9 ± 0.4 and 2.5 ± 0.6 × 10 .
f
Uncertainties are reported in the 1 sigma confidence level.
g
Propagated uncertainties include error in the blank, carrier mass (1%), and counting statistics.
h 10 10
Propagated error in the model ages include a 6% uncertainty in the production rate of Be and a 4% uncertainty in the Be decay constant.
10.1002/2015GL066077

1594
 Geophysical Research Letters 10.1002/2015GL066077

Owen et al. [2001] and Murari et al. [2014]), but there exists no evidence of any trunk-valley glacier advance at
this time or later. Well-preserved flood deposits in the Chandra River bed resulting from ice dam failure in the
Upper Chandra Valley [Coxon et al., 1996; Owen et al., 2001] and preserved Holocene strath terraces [Adams
et al., 2009] support this interpretation. Our reconstruction suggests a mean ice retreat rate of 37 ± 11 m per
year in the Chandra Valley beginning at the end of the LGM.

5. Discussion
Our new field and 10Be exposure data and the reconstructed ice extent suggest that during the LGM the
Chandra Valley and its tributaries were occupied by a ~200 km long and ~1 km thick glacier, supporting
earlier observations by Owen et al. [1995, 1997]. Combining field observations of pronounced trimlines that
separate ice-polished surfaces below from rugged bedrock ridges and hillslopes above and sample heights of
our new 10Be ages suggest that earlier advances during the last glacial cycle were either similar or not much
more extensive than during the LGM. We favor this interpretation, given the excellent preservation of
glacially polished surfaces at high elevation since ~20 ka. Furthermore, our ice reconstruction and field
evidence support significant overtopping (>500 m) into both the Spiti and Beas Valleys [Owen et al., 1995,
1997, 2001; Saha et al., 2015]. Thus, our reconstructed LGM glacier may not have corresponded to the most
extensive glaciation during the last glacial cycle. Comparing our new data and earlier work [Owen et al., 1997,
2001] suggests that our reconstructed LGM ice extent is better correlated with the Chandra glacial stage than
the Batal trimlines and, as supported by Murari et al.’s [2014] reanalysis, shifts the Batal glacial stage toward
the LGM. More dedicated work near the identified trimlines is needed to resolve this issue.
Exposure ages of boulders and ice-polished bedrock surfaces from the same locations within the Chandra
Valley are in good agreement with each other and show consistent ages between 19 and 16 ka with no sys-
tematic bias (Figure 1c). Because horizontal as well as near-vertical polished surfaces at locations 15–20 km
apart from each other yield virtually identical ages, we suggest that in our study, cosmogenic nuclide inheri-
tance is an issue for only one sample, a ~33 ka boulder situated on a much younger surface. The similar ages
between boulders and surfaces also suggest that glacial erosion has been sufficient to reset all surfaces and
that postdepositional erosion of boulder surfaces is negligible.
The reconstructed CVG maintained its maximum vertical extent prior to or at 20 ka, during a weakened Indian
Summer Monsoon (Figure 3b) [Herzschuh, 2006; Dutt et al., 2015]. After 20 ka, coeval with increasing tempera-
tures, but also increasing monsoonal strength, the CVG rapidly receded. We thus argue that the retreat of the
CVG was primarily driven by temperature; although tributary glaciers with shorter response times could still
have reacted to changes in precipitation with minor readvances during a general phase of retreat [Scherler
et al., 2011]. Such readvances of tributary glaciers are well documented by remnants of moraines and dated
boulders on lateral and frontal moraines of the Kulti glacial stage [Owen et al., 2001].
In Figure 3, we compare the pace and timing of deglaciation in the Chandra Valley with other Himalayan/
Karakoram glaciers and midlatitude glaciers from the Northern and Southern Hemispheres based on 10Be
and 14C ages. We emphasize that we focus here on long-term retreat rates, but acknowledge that the
long-term retreat of most glaciers was accompanied by smaller advances and retreats during certain time
intervals. As such short-term fluctuations are generally difficult to reconstruct, we did not include them in
Figure 3. Although there exist several Himalayan glaciers that did not retreat as fast as the Chandra Valley
glacier, and large variations in long-term retreat rates are observable in all regions, most of the glaciers
started retreating rapidly close to the end of the LGM (see Figure 3). These observations are consistent with
modeled equilibrium line altitudes (ELAs) across Central Asia, which suggest that temperature changes dur-
ing the LGM controlled the onset of deglaciation [Rupper and Koppes, 2010]. Furthermore, retreat of both
Northern and Southern Hemisphere glaciers of comparable size occurred at similar rates as in the Chandra
Valley. A recent global compilation of glacial chronologies that excludes the Himalaya points toward a close
relationship between globally increasing temperatures and rapid deglaciation [Shakun et al., 2015]. Our
new data from the Chandra Valley and recalculated ages from other Himalayan glaciers indicate that rapid
post-LGM deglaciation is probably not significantly different from deglaciation elsewhere. Glaciers located
in the western Himalaya, in westernmost Tibet, and the Karakoram apparently retreated more rapidly com-
pared to glaciers in the central Himalaya, where the influence of the monsoon is stronger [Finkel et al., 2003;
Seong et al., 2007, 2009; Owen et al., 2009].

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1595

 Geophysical Research Letters 10.1002/2015GL066077

The new scaling scheme by Lifton et al. [2014] affects estimated exposure ages significantly. In the Chandra
Valley, recalculated LSD ages deviate from the exposure ages obtained from the CRONUS-Earth online calcu-
lator (v2.2) [Balco et al., 2008] by 25% and the calibration data set compiled by Heyman [2014] by 11%. Within
the Himalaya the differences are 24% and 10%, respectively (see also supporting information Data Set S2 and
Figures S17 and S18). Differences within the scaling schemes remain at approximately 10%. In contrast to
previous scaling models, the LSD scaling model uses analytical approximations to cosmic ray fluxes in the
atmosphere and includes an updated geomagnetic and atmospheric framework [Lifton et al., 2014].
Although the lack of calibration sites within the Himalaya does not yet allow testing whether these improve-
ments also result in more precise ages, a better understanding of the discrepancies between previous scaling
models and the resulting bias [Lifton et al., 2014] suggests that some of the existing scaling models have
deficiencies. In contrast to the Himalaya, however, the maximum increase of published middle- to high-
latitude exposure ages in the Northern and Southern Hemispheres is merely 8% and some locations even
show decreasing ages of the same order when using the LSD scaling model.
In light of our new observations and age constraints from the Chandra Valley and the older recalculated
moraine ages, it is thus possible that many glacial advances, previously considered to be Late Glacial may
be coeval with the LGM confirmed by Optically Stimulated Luminescence ages, e.g., in the Everest region
[Richards et al., 2000; Owen et al., 2009]. Importantly, glaciation during the LGM and the pace of retreat in
the Himalaya appear to have been more akin to midlatitude glaciers than previously thought and thus reflect
hemisphere-scale processes rather than close regional links such as monsoonal forcing.

Acknowledgments References
This research was funded by the graduate
school GRK1364 (Shaping Earth’s Surface Adams, B., C. Dietsch, L. A. Owen, M. W. Caffee, J. Spotila, and W. C. Haneberg (2009), Exhumation and incision history of the Lahul Himalaya,
in a Variable Environment) of the German northern India, based on (U-Th)/He thermochronometry and terrestrial cosmogenic nuclide methods, Geomorphology, 107(3–4), 285–299,
Science Foundation to M.S. and D.S. doi:10.1016/j.geomorph.2008.12.017.
(DFG, Deutsche Forschungsgemeinschaft, Applegate, P. J., N. M. Urban, K. Keller, and R. B. Alley (2009), Modeling the statistical distributions of cosmogenic exposure dates from moraines,
grant STR 373/19-2). R.T. is supported by Geosci. Model Dev. Discuss., 3(1), 293–307, doi:10.5194/gmdd-2-1407-2009-supplement.
DFG grant TH 1371/5-1. We are indebted Balco, G., J. O. Stone, N. A. Lifton, and T. J. Dunai (2008), A complete and easily accessible means of calculating surface exposure ages or
10 26
to V. Jain and T. Tsering Longpo for erosion rates from Be and Al measurements, Quat. Geochronology, 3(3), 174–195, doi:10.1016/j.quageo.2007.12.001.
logistical support during our fieldwork Benn, D. I., and N. R. J. Hulton (2010), An ExcelTM spreadsheet program for reconstructing the surface profile of former mountain glaciers and
in India. We thank T. Schildgen and ice caps, Comput. Geosci., 36(5), 605–610, doi:10.1016/j.cageo.2009.09.016.
W. Düsing for help with sample pre- Benn, D. I., and L. A. Owen (1998), The role of the Indian summer monsoon and the mid-latitude westerlies in Himalayan glaciation: Review
paration and discussions. We thank and speculative discussion, J. Geol. Soc., 155, 353–363, doi:10.1144/gsjgs.155.2.0353.
N. Lifton and L. Owen for constructive Borchers, B., S. Marrero, G. Balco, M. Caffee, B. Goehring, N. Lifton, K. Nishiizumi, F. Phillips, J. Schaefer, and J. Stone (2016), Geological calibration
comments that improved the paper. of spallation production rates in the CRONUS-Earth project, Quat. Geochronology, 31, 188–198, doi:10.1016/j.quageo.2015.01.009.
Field photographs and additional Chevalier, M.-L., G. Hilley, P. Tapponnier, J. Van Der Woerd, J. Liu-Zeng, R. C. Finkel, F. J. Ryerson, H. Li, and X. Liu (2011), Constraints on
information on calculation results are the late Quaternary glaciations in Tibet from cosmogenic exposure ages of moraine surfaces, Quat. Sci. Rev., 30(5–6), 528–554,
provided in the supporting information doi:10.1016/j.quascirev.2010.11.005.
10
of this manuscript. Chmeleff, J., F. von Blanckenburg, K. Kossert, and D. Jakob (2010), Determination of the Be half-life by multicollector ICP-MS and liquid
scintillation counting, Nucl. Instrum. Methods Phys. Res., Sect. B, 268(2), 192–199, doi:10.1016/j.nimb.2009.09.012.
Coxon, P., L. A. Owen, and W. A. Mitchell (1996), A late Quaternary catastrophic flood in the Lahul Himalayas, J. Quat. Sci., 11(6), 495–510,
doi:10.1002/(SICI)1099-1417(199611/12)11:6<495::AID-JQS268>3.0.CO;2-M.
Dewald, A., et al. (2013), CologneAMS, a dedicated center for accelerator mass spectrometry in Germany, Nucl. Instrum. Methods Phys. Res.,
Sect. B, 294, 18–23, doi:10.1016/j.nimb.2012.04.030.
Dortch, J. M., L. A. Owen, and M. W. Caffee (2013), Timing and climatic drivers for glaciation across semi-arid western Himalayan–Tibetan
orogen, Quat. Sci. Rev., 78, 188–208, doi:10.1016/j.quascirev.2013.07.025.
Dunne, J., D. Elmore, and P. Muzikar (1999), Scaling factors for the rates of production of cosmogenic nuclides for geometric shielding and
attenuation at depth on sloped surfaces, Geomorphology, 27(1–2), 3–11, doi:10.1016/S0169-555X(98)00086-5.
Dutt, S., A. K. Gupta, S. C. Clemens, H. Cheng, R. K. Singh, G. Kathayat, and R. L. Edwards (2015), Abrupt changes in Indian summer monsoon
strength during 33,800 to 5500 years B.P, Geophys. Res. Lett., 42, 5526–5532, doi:10.1002/2015GL064015.
Finkel, R. C., L. A. Owen, P. L. Barnard, and M. W. Caffee (2003), Beryllium-10 dating of Mount Everest moraines indicates a strong monsoon
influence and glacial synchroneity throughout the Himalaya, Geology, 31(6), 561, doi:10.1130/0091-7613(2003)031<0561:BDOMEM>2.0.CO;2.
Gardner, J. S., and K. Hewitt (1990), A surge of Bualtar Glacier, Karakoram Range, Pakistan: A possible landslide trigger, J. Glaciol., 36(123), 159–162.
Godard, V., D. L. Bourlès, F. Spinabella, D. W. Burbank, B. Bookhagen, G. Burch Fisher, A. Moulin, and L. Léanni (2014), Dominance of tectonics
over climate in Himalayan denudation, Geology, 43(3), 243–246, doi:10.1130/G35342.1.
Herzschuh, U. (2006), Palaeo-moisture evolution in monsoonal Central Asia during the last 50,000 years, Quat. Sci. Rev., 25(1–2), 163–178,
doi:10.1016/j.quascirev.2005.02.006.
Heyman, J. (2014), Paleoglaciation of the Tibetan Plateau and surrounding mountains based on exposure ages and ELA depression estimates,
Quat. Sci. Rev., 91, 30–41, doi:10.1016/j.quascirev.2014.03.018.
Heyman, J., A. P. Stroeven, J. M. Harbor, and M. W. Caffee (2011), Too young or too old: Evaluating cosmogenic exposure dating based on an
analysis of compiled boulder exposure ages, Earth Planet. Sci. Lett., 302(1–2), 71–80, doi:10.1016/j.epsl.2010.11.040.
Jamieson, S. S. R., M. W. Ewertowski, and D. J. A. Evans (2015), Rapid advance of two mountain glaciers in response to mine-related debris
loading, J. Geophys. Res. Earth Surf., 120, 1418–1435, doi:10.1002/2015JF003504.
Jarvis, A., H. I. Reuter, A. Nelson, and E. Guevara (2008), Hole-filled seamless SRTM data V4, Int. Cent. for Trop. Agric. (CIAT). [Available at http://srtm.
csi.cgiar.org, Accessed 26 December 2008.]

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1596

 Geophysical Research Letters 10.1002/2015GL066077

Kelly, M. A., T. V. Lowell, P. J. Applegate, F. M. Phillips, J. M. Schaefer, C. A. Smith, H. Kim, K. C. Leonard, and A. M. Hudson (2013), A locally
10
calibrated, late glacial Be production rate from a low-latitude, high-altitude site in the Peruvian Andes, Quat. Geochronology,
doi:10.1016/j.quageo.2013.10.007.
Kohl, C. P., and K. Nishiizumi (1992), Chemical isolation of quartz for measurement of in-situ-produced cosmogenic nuclides, Geochim. Cosmochim.
Acta, 56(9), 3583–3587, doi:10.1016/0016-7037(92)90401-4.
10
Korschinek, G., et al. (2010), A new value for the half-life of Be by heavy-ion elastic recoil detection and liquid scintillation counting, Nucl. Instrum.
Methods Phys. Res., Sect. B, 268(2), 187–191, doi:10.1016/j.nimb.2009.09.020.
Lifton, N., T. Sato, and T. J. Dunai (2014), Scaling in situ cosmogenic nuclide production rates using analytical approximations to atmospheric
cosmic-ray fluxes, Earth Planet. Sci. Lett., 386, 149–160, doi:10.1016/j.epsl.2013.10.052.
18
Lisiecki, L. E., and M. E. Raymo (2005), A Pliocene-Pleistocene stack of 57 globally distributed benthic δ O records, Paleoceanography, 20,
PA1003, doi:10.1029/2004PA001071.
Marrero, S. M., F. M. Phillips, B. Borchers, N. Lifton, R. Aumer, and G. Balco (2016), Cosmogenic nuclide systematics and the CRONUScalc program,
Quat. Geochronol., 31, 160–187, doi:10.1016/j.quageo.2015.09.005.
Mix, A., E. Bard, and R. Schneider (2001), Environmental processes of the ice age: Land, oceans, glaciers (EPILOG), Quat. Sci. Rev., 20(4), 627–657,
doi:10.1016/S0277-3791(00)00145-1.
Murari, M. K., L. A. Owen, J. M. Dortch, M. W. Caffee, C. Dietsch, M. Fuchs, W. C. Haneberg, M. C. Sharma, and A. Townsend-Small (2014), Timing
and climatic drivers for glaciation across monsoon-influenced regions of the Himalayan–Tibetan orogen, Quat. Sci. Rev., 88, 159–182,
doi:10.1016/j.quascirev.2014.01.013.
Olen, S., B. Bookhagen, B. Hoffmann, D. Sachse, D. Adhikari, and M. R. Strecker (2015), Understanding erosion rates at the Himalayan orogenic
front: A case study from the Arun Valley, Nepal, J. Geophys. Res. Earth Surf., 120, 2080–2102, doi:10.1002/2014JF003410.
Owen, L., R. Bailey, E. Rhodes, W. Mitchell, and P. Coxon (1997), Style and timing of glaciation in the Lahul Himalaya, northern India: A framework
for reconstructing late Quaternary palaeoclimatic change in the western Himalayas, J. Quat. Sci., 12(2), 83–109.
Owen, L. A. (2009), Latest Pleistocene and Holocene glacier fluctuations in the Himalaya and Tibet, Quat. Sci. Rev., 28(21–22), 2150–2164,
doi:10.1016/j.quascirev.2008.10.020.
Owen, L. A., and J. M. Dortch (2014), Nature and timing of Quaternary glaciation in the Himalayan–Tibetan orogen, Quat. Sci. Rev., 88, 14–54,
doi:10.1016/j.quascirev.2013.11.016.
Owen, L. A., D. I. Benn, E. Derbyshire, D. J. A. Evans, W. A. Mitchell, D. Thompson, S. Richardson, M. Lloyd, and C. Holden (1995), The geomorphology
and landscape evolution of the Lahul Himalaya, Northern India, Z. Geomorphol., 39(2), 145–174.
Owen, L. A., E. Derbyshire, S. Richardson, D. I. Benn, D. J. A. Evans, and W. A. Mitchell (1996), The Quaternary glacial history of the Lahul Himalaya,
northern India, J. Quat. Sci., 11(1), 25–42.
Owen, L. A., L. Gualtieri, R. C. Finkel, M. W. Caffee, D. I. Benn, and M. C. Sharma (2001), Cosmogenic radionuclide dating of glacial landforms in
the Lahul Himalaya, northern India: Defining the timing of Late Quaternary glaciation, J. Quat. Sci., 16(6), 555–563, doi:10.1002/jqs.621.
Owen, L. A., R. C. Finkel, P. L. Barnard, M. Haizhou, K. Asahi, M. W. Caffee, and E. Derbyshire (2005), Climatic and topographic controls on the
10
style and timing of Late Quaternary glaciation throughout Tibet and the Himalaya defined by Be cosmogenic radionuclide surface
exposure dating, Quat. Sci. Rev., 24(12–13), 1391–1411, doi:10.1016/j.quascirev.2004.10.014.
Owen, L. A., R. Robinson, D. I. Benn, R. C. Finkel, N. K. Davis, C. Yi, J. Putkonen, D. Li, and A. S. Murray (2009), Quaternary glaciation of Mount Everest,
Quat. Sci. Rev., 28(15–16), 1412–1433, doi:10.1016/j.quascirev.2009.02.010.
Pfeffer, W. T., et al. (2014), The Randolph Glacier Inventory: A globally complete inventory of glaciers, J. Glaciol., 60(221), 537–552,
doi:10.3189/2014JoG13J176.
Phillips, F. M., et al. (2016), The CRONUS-Earth Project: A synthesis, Quat. Geochronol., 31, 119–154, doi:10.1016/j.quageo.2015.09.006.
Pratt-Sitaula, B., D. W. Burbank, A. M. Heimsath, N. F. Humphrey, M. Oskin, and J. Putkonen (2011), Topographic control of asynchronous
glacial advances: A case study from Annapurna, Nepal, Geophys. Res. Lett., 38, L24502, doi:10.1029/2011GL049940.
Putnam, A. E., G. H. Denton, J. M. Schaefer, D. J. A. Barrell, B. G. Andersen, R. C. Finkel, R. Schwartz, A. M. Doughty, M. R. Kaplan, and C. Schlüchter
(2010), Glacier advance in southern middle-latitudes during the Antarctic Cold Reversal, Nat. Geosci., 3(10), 700–704, doi:10.1038/ngeo962.
Rawat, S., A. K. Gupta, S. J. Sangode, P. Srivastava, and H. C. Nainwal (2015), Late Pleistocene–Holocene vegetation and Indian summer
monsoon record from the Lahaul, Northwest Himalaya, India, Quat. Sci. Rev., 114, 167–181, doi:10.1016/j.quascirev.2015.01.032.
Richards, B. W., L. A. Owen, and E. J. Rhodes (2000), Timing of Late Quaternary glaciations in the Himalayas of northern Pakistan, J. Quat. Sci.,
15(3), 283–297, doi:10.1002/(SICI)1099-1417(200003)15:3<283::AID-JQS525>3.0.CO;2-X.
Rupper, S., and M. Koppes (2010), Spatial patterns in Central Asian climate and equilibrium line altitudes, IOP Conf. Ser.: Earth Environ. Sci.,
9(1), 012009, doi:10.1088/1755-1315/9/1/012009.
Saha, S., M. C. Sharma, M. K. Murari, L. A. Owen, and M. W. Caffee (2015), Geomorphology, sedimentology, and minimum exposure ages of
streamlined subglacial landforms in the NW Himalaya, India, Boreas, doi:10.1111/bor.12153.
Scherler, D., B. Bookhagen, M. Strecker, F. von Blanckenburg, and D. Rood (2010), Timing and extent of late Quaternary glaciation in the western
10
Himalaya constrained by Be moraine dating in Garhwal, India, Quat. Sci. Rev., 29(7–8), 815–831, doi:10.1016/j.quascirev.2009.11.031.
Scherler, D., B. Bookhagen, and M. R. Strecker (2011), Hillslope-glacier coupling: The interplay of topography and glacial dynamics in High
Asia, J. Geophys. Res., 116, F02019, doi:10.1029/2010JF001751.
Schwanghart, W., and D. Scherler (2013), Short communication: TopoToolbox 2—An efficient and user-friendly tool for Earth surface sciences,
Earth Surf. Dyn. Discuss., 1, 261–275, doi:10.5194/esurfd-1-261-2013.
Seong, Y. B., L. A. Owen, M. P. Bishop, A. Bush, P. Clendon, L. Copland, R. Finkel, U. Kamp, and J. F. Shroder Jr. (2007), Quaternary glacial history
of the Central Karakoram, Quat. Sci. Rev., 26(25–28), 3384–3405, doi:10.1016/j.quascirev.2007.09.015.
Seong, Y. B., L. A. Owen, C. Yi, and R. C. Finkel (2009), Quaternary glaciation of Muztag Ata and Kongur Shan: Evidence for glacier response
to rapid climate changes throughout the Late Glacial and Holocene in westernmost Tibet, Geol. Soc. Am. Bull., 121(3–4), 348–365,
doi:10.1130/B26339.1.
Shakun, J. D., P. U. Clark, F. He, N. A. Lifton, Z. Liu, and L. Otto-Bliesner (2015), Regional and global forcing of glacier retreat during the last
deglaciation, Nat. Commun., 6, 8059, doi:10.1038/ncomms9059.
Steck, A. (2003), Geology of the NW Indian Himalaya, Eclogae Geol. Helv., 96(2), 147–196, doi:10.1007/s00015-003-1091-4.
Tovar, D. S., J. Shulmeister, and T. R. Davies (2008), Evidence for a landslide origin of New Zealand’s Waiho Loop moraine, Nat. Geosci., 1(8),
524–526, doi:10.1038/ngeo249.

EUGSTER ET AL. RAPID CHANDRA VALLEY LGM DEGLACIATION 1597