The Neoglacial History of Robson Glacier, British Columbia.

Abstract

As glaciers in the Canadian Rockies recede, glacier forefields continue to yield subfossil wood from sites overridden by these glaciers during the Holocene. Robson Glacier in British Columbia formerly extended below tree line, and recession over the last century has progressively revealed a number of buried forest sites that are providing one of the more complete records of glacier history in the Canadian Rockies during the latter half of the Holocene. The glacier was advancing ca. 5.5 km upvalley of the Little Ice Age terminus ca. 5.26 cal ka BP, at sites ca. 2 km upvalley ca. 4.02 cal ka BP and ca. 3.55 cal ka BP, and 0.5-1 km upvalley between 1140 and 1350 A.D. There is also limited evidence based on detrital wood of an additional period of glacier advance ca. 3.24 cal ka BP. This record is more similar to glacier histories further west in British Columbia than elsewhere in the Rockies and provides the first evidence for a post-Hypsithermal glacier advance at ca. 5.26 cal ka BP in the Rockies. The utilization of the wiggle-matching approach using multiple 14C dates from sample locations determined by dendrochronological analyses enabled the recognition of 14C outliers and an increase in the precision and accuracy of the dating of glacier advances.

Full text

ARTICLE
Neoglacial history of Robson Glacier, British Columbia
B.H. Luckman, M.H. Masiokas, and K. Nicolussi
Abstract: As glaciers in the Canadian Rockies recede, glacier forefields continue to yield subfossil wood from sites overridden by
these glaciers during the Holocene. Robson Glacier in British Columbia formerly extended below tree line, and recession over the
last century has progressively revealed a number of buried forest sites that are providing one of the more complete records of
glacier history in the Canadian Rockies during the latter half of the Holocene. The glacier was advancing ca. 5.5 km upvalley of
the Little Ice Age terminus ca. 5.26 cal ka BP, at sites ca. 2 km upvalley ca. 4.02 cal ka BP and ca. 3.55 cal ka BP, and 0.5–1 km
upvalley between 1140 and 1350 A.D. There is also limited evidence based on detrital wood of an additional period of glacier
advance ca. 3.24 cal ka BP. This record is more similar to glacier histories further west in British Columbia than elsewhere in the
Rockies and provides the first evidence for a post-Hypsithermal glacier advance at ca. 5.26 cal ka BP in the Rockies. The
utilization of the wiggle-matching approach using multiple
14
C dates from sample locations determined by dendrochrono-
logical analyses enabled the recognition of
14
C outliers and an increase in the precision and accuracy of the dating of glacier
advances.
Résumé : Au fil du retrait des glaciers dans les Rocheuses canadiennes, les avant-fronts glaciaires continuent de produire du bois
subfossile de sites recouverts par ces glaciers durant l’Holocène. Le glacier Robson en Colombie-Britannique s’étalait auparavant
jusque sous la limite des arbres et son retrait au cours du dernier siècle a graduellement révélé différents sites boisés enfouis qui
fournissent un des registres les plus complets de l’histoire glaciaire dans les Rocheuses canadiennes durant la deuxième moitié
de l’Holocène. Le glacier s’était avancé jusqu’a
`environ 5,5 km en amont du front du Petit Âge glaciaire vers 5,26 cal ka BP, a
`
environ 2 km vers 4,02 cal ka BP et 3,55 cal ka BP, et a
`0,5–1 km entre 1140 et 1350 A.D. Il existe également certains indices
reposant sur du bois détritique d’une autre période d’avancée du glacier vers 3,24 cal ka BP. Ce registre s’apparente plus a
`
l’évolution de glaciers plus a
`l’ouest en Colombie-Britannique qu’a
`celle d’autres glaciers dans les Rocheuses et fournit les
premières preuves d’une progression glaciaire post-hypsithermale vers 5,26 cal ka BP dans ces montagnes. L’emploi d’une
approche d’appariement des oscillations en utilisant de multiples âges au
14
C d’emplacements d’échantillon établis par
analyse dendrochronologique a permis la reconnaissance de valeurs aberrantes de
14
C et une augmentation de la précision
et de l’exactitude de la datation de progressions glaciaires. [Traduit par la Rédaction]
Introduction
Most alpine glaciers in the Northern Hemisphere reached
their maximum Holocene extent during the “Little Ice Age” of
the last few centuries (Grove 2004). Subsequent recession dur-
ing the last ca. 100 years at several sites has revealed subfossil
wood and (or) other organic materials that were overridden by
earlier glacier events. Increasingly, studies of this wood and
buried, glacially overridden, forests are being used to define
earlier periods of glacier advance and link them to global-scale
climatic controls (e.g., Le Roy et al. 2015;Solomina et al. 2016).
Although increasing numbers of such forefield records are be-
coming available, including many from British Columbia (e.g.,
Koch et al. 2007;Menounos et al. 2009;Mood and Smith 2015;
St-Hillaire and Smith 2017), few such records are available for
the Canadian Rockies (Osborn et al. 2001;Wood and Smith
2004;Luckman 2006). These sites are relatively rare and, indi-
vidually, only provide a partial record of local glacier histories:
regional history must be compiled by the assembly, analysis,
and correlation of records from many sites.
Robson Glacier in British Columbia formerly extended below
tree line, and recession over the last century has progressively
revealed detrital wood and a number of buried forest sites that are
providing one of the more complete records of glacier history in
the Canadian Rockies during the latter half of the Holocene. This
paper updates and reviews new evidence from Robson Glacier and
places it within the context of regional glacier history. It also
outlines some of the problems of reconstructing glacier history
from such evidence.
Site description and overview of previous work
Robson Glacier is a ca. 6 km long valley glacier (ca. 13.9 km
2
in
2006) that drains the eastern and northern flanks of Mount Rob-
son, British Columbia (Fig. 1). It has a well-developed series of
lateral and terminal moraines that extend down to 1660 m and
abut the Continental Divide. The oldest terminal moraine has a
closed forest cover and is about 2 km downvalley of the present
snout. Sheet 32A of the Inter-Provincial Boundary Commission
(Cautley and Wheeler 1924) shows the glacier front within
Received 18 October 2016. Accepted 8 August 2017.
Paper handled by Associate Editor Timothy Fisher.
B.H. Luckman. Department of Geography, The University of Western Ontario, London, ON N6A 5C2, Canada.
M.H. Masiokas. Department of Geography, The University of Western Ontario, London, ON N6A 5C2, Canada; Instituto Argentino de Nivología,
Glaciología y Ciencias Ambientales (IANIGLA), CCT-CONICET, Mendoza, Argentina.
K. Nicolussi. Institute of Geography, University of Innsbruck, 6020 Innsbruck, Austria.
Corresponding author: B.H. Luckman (email: Luckman@uwo.ca).
Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from RightsLink.
A correction was made to the e-First version of this paper on 24 October 2017 prior to the final issue publication. The current online and print versions are
identical and both contain the correction.
1153
Can. J. Earth Sci. 54: 1153–1164 (2017) dx.doi.org/10.1139/cjes-2016-0187 Published at www.nrcresearchpress.com/cjes on 17 August 2017.
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

ca. 100 m of the terminal moraine in 1922.
1
Heusser (1956) used
dendrochronology (ring counts) to date the terminal to ca. 1783
based on the oldest of seven trees he cored on its surface, applying
a 12 year ecesis estimate. He also used ring counts to date several
additional moraines upvalley as 1801, 1864, 1891, 1907 plus several
smaller features in the 1920s–1940s (see Heusser 1956;Luckman
2000). A proglacial lake developed in front of the glacier after
ca. 1960 and has expanded upvalley as the glacier front has re-
ceded. Four buried forest sites have been exposed during glacier
recession, which allowed extensive sampling of wood remains.
Each of these will be described briefly followed by the description
of dendrochronological analyses and radiocarbon dating of these
materials and a discussion of the chronological implications of
these findings.
Heusser buried forest site
Heusser (1956) identified a buried forest site, overlain by ca. 1–2 m
of till, that was exposed in abandoned stream channels cut by
proglacial streams between ca. 1925 and 1950. He obtained an
early Libby
14
C date of 450 ± 150 years BP from an in situ stump at
1
The 1924 map is based on 1922 terrestrial photogrammetry.
Fig. 1. Location of Robson Glacier. The image is from September 2006. The image source is Image ©2011 Province of British Columbia, ©2011
Google, ©2011 Cnes/SPOT image. LIA, Little Ice Age terminal moraine (1783); H, Heusser site; R, Ridge site; GT, Glacier Toe site; E, Upper
Extinguisher Tower site; ET, Lower Extinguisher Tower site. [Colour online.]
1154 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

this site. Investigations at this locality by Luckman between 1980
and 1992 recovered over 45 detrital and in situ logs
2
and stumps
up to 30 cm in diameter (Luckman 1995). Tree-ring series from
these logs were successfully calendar crossdated, initially using a
master chronology from a site at Bennington Glacier (ca. 75 km to
the southeast) and later with the chronology from the Columbia
Icefield (Luckman 1995;Luckman and Wilson 2005). The overrid-
den trees at this site grew between 867 and 1350 A.D., and death
dates from in situ stumps indicated that the glacier advanced over
this site between ca. 1142 and 1350 A.D. at a net average rate of
ca. 2–2.5 m/year (Luckman 1995). This site provided the first evi-
dence for a calendar-dated early Little Ice Age (LIA) advance in the
Canadian Rockies.
Upper Extinguisher Tower site
Luckman et al. (1993) reported on a large, possibly in situ, Pinus
albicaulis stump and rootstock, initially discovered by a climber,
lying on the ground surface about 30 m from the lateral margin of
the glacier adjacent to Extinguisher Tower, a 300 m high bedrock
pinnacle on the north side of the glacier. The snag was located
20–30 m from the contemporary ice front on a relatively gently
sloping thinly till-covered bedrock surface, about 3.5 km upvalley
of the 1990 snout and some 300–350 m higher. Two small detrital
Abies logs were also found lying on the surface a short distance
downglacier from the stump. Four
14
C dates from these samples
yield ages between ca. 3130 and 3360
14
C years BP (E1–E4; Table 1;
Luckman et al. 1993;Luckman 1995).
Glacier Toe area
In the late 1980s a low bedrock ridge emerged from beneath the
northern flank of the glacier snout as it receded upvalley. Progla-
cial drainage from the north flank of the glacier cut a channel
between this ridge and the northern valley side, creating a small
delta where it entered the proglacial lake. In 1992 three logs were
recovered from the west shore of the lake, a short distance up-
stream of the Heusser site. Tree-ring series from two of these logs
(logs R9210 and R9212; see later in text) crossdated and provided a
strong “floating” (undated) chronology but could not be matched
with the existing calendar-dated tree-ring chronology from the
Heusser site. The source of these logs was a buried forest site close
to the 1993 glacier front. Luckman (1995) briefly described this site
and reported three radiocarbon dates between 3500 and 3710
14
C
years BP (G1–G3; Table 1) that confirm that the material found in the
lake (G1) was of similar age to in situ material at the glacier front (G2,
G3). The balance of this paper reports subsequent, mainly den-
drochronological, investigations at this and two other sites
that provide a more detailed history of glacier variations at this
glacier.
New field investigations and site descriptions
Glacier Toe site (Figs. 2A–2D)
The Glacier Toe site is approximately 2–300 m upvalley of the
lake and lies in a saddle between a broad low bedrock ridge,
ca. 40 m higher than the lake, and the valley side. The site was
exposed in the 1990s by erosion of a proglacial stream originating
from the northern flank of the glacier. In 1993 the braided progla-
cial river channel and coarse gravel bars in front of the glacier
contained many wood fragments plus sheared and shattered de-
trital logs up to 2 m long (Fig. 2C), including some protruding
vertically through the gravels. A small river bank section in one of
the gravel bars, ca. 50 m downstream of the ice front, revealed
20–30 cm of compacted organic litter with needles, cones, abun-
dant roots, and stunted trees overlying ca. 20–30 cm of till (Fig. 2D;
Luckman 1995). A radiocarbon date of 3710 ± 70
14
C years BP (G2;
Table 1) was obtained from one of these roots. Several large
2
In situ logs or stumps are preserved in growth position: detrital logs have been transported from their (usually unknown) growth location.
Table 1. Radiocarbon dates from the Robson sites.
Ref.
a
14
C date
laboratory No.
b
Log ID Setting
Ring
position
c
14
C date
(years BP)
cal
14
C date,
median
(cal ka BP)
d
cal
14
C date,
2␴(cal ka BP) Source
Upper Extinguisher site
E1 Beta-28439 GDO In situ stump Outer 3300±70 3.53 3.70–3.38 Luckman et al. (1993)
E2 Beta-33012 R8902 In situ stump Not known 3230±70 3.46 3.64–3.27 Luckman et al. (1993)
E3 Beta-35010 R8902x Small detrital log Not known 3130±70 3.34 3.56–3.16 Luckman et al. (1993)
E4 Beta-38309 R8905 Small detrital log Not known 3360±60 3.60 3.82–3.45 Luckman et al. (1993)
Glacier Toe site (chronology A, 794–1112 A)
G1 Beta-62064 R9210 Detrital wood (lake) 984–1029 A 3500±60 3.77 3.96–3.62 Luckman (1995)
G2 Beta-65381 R9311 Root in paleosol 123 rings 3710±70 4.06 4.29–3.85 Luckman (1995)
G3 Beta-65382 R9314 Large snag in gravels Not known 3650±60 3.97 4.15–3.83 Luckman (1995)
G4 Beta-200485 R9212 Detrital wood (lake) 945–1002 A 3730±70 4.09 4.35–3.88 Menounos et al. (2009)
G5 Beta-200484 R0026A Detrital wood (delta) 51–71 of 405 3320±80 3.56 3.82–3.38 Menounos et al. (2009)
G6 Beta-378910 R0026A2 Detrital wood (delta) 151–192 of 405 3260±21 3.48 3.56–3.41 This paper
G7 Beta-378911 R921014 Detrital wood (lake) 827–847 A 3870±17 4.31 4.41–4.28 This paper
Ridge site (chronology C, 1731–2017 C)
R1 Beta-84817 R95 Detrital wood 31 rings 3160±70 3.38 3.56–3.21 Luckman et al. (1996)
R2 Beta-200486 R9532 Detrital wood 1811–1901 C 3550±60 3.84 4.06–3.65 Menounos et al. (2009)
R3 MAMS-31064 R9532b Detrital wood 2009–2017 C 3300±23 3.52 3.58–3.46 This paper
Lower Extinguisher site (chronology B, 113–632 B)
ET1 Beta-187090 R0301 Log in till section 510–610 B 4780±60 5.51 5.61–5.32 Luckman (2007)
ET2 Beta-200483 R0403 Log in till section 275–325 B 4770±60 5.51 5.60–5.33 This paper
ET3 Beta-378912 R0402 Log in till section 356–396 B 4803±17 5.51 5.59–5.48 This paper
a
Reference number of sample.
b
The MAMS date is from the Klaus-Tschira-Archäometrie-Zentrum lab in Mannheim, Germany.
c
Ring position of the dated sample (when known) is given based on its position in the respective “floating” tree-ring chronology. Note that the youngest ring (closest
to the pith) provides the oldest calendar date.
d
The median calendar date and 2␴range are based on calibration results established by using OxCal 4.3 (Ramsey 2009) and the IntCal13 calibration curve (Reimer
et al. 2013) and rounded to the nearest decade.
Luckman et al. 1155
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

trunks, some partially impregnated with till, were also associated
with the paleosol and buried within the gravels (Fig. 2B). A second
radiocarbon date of 3650 ± 50
14
C years BP (G3; Table 1) was ob-
tained from ca. 31 rings of a large, detrital, 2 m high, whitebark
pine rootstock. A similar date of 3500 ± 60
14
C years BP (G1; Table 1)
from one of the logs recovered from the lake confirmed that the
wood recovered from the lake is of similar age and has been
washed in by the proglacial stream. Glacier recession after 1995
exposed the upvalley, downslope side of the ridge, and the progla-
cial drainage initially built a small sand and gravel delta at the ice
front graded to the col (possibly burying other material). With
further recession and lowering of the ice front, the proglacial
stream abandoned the col route (the “Creek” site; Fig. 2A) and
flowed downslope, entering the ice and flowing to the lake via a
subglacial conduit (Fig. 2A, left). Between 1993 and 2000, 54 cross
sections were collected from detrital logs lying in the former
stream channel, on the delta surface downstream or along
the shorelines of the proglacial lake. These logs were assumed to
have been derived by fluvial and possibly glacier erosion of the
buried forest site associated with the paleosol at this locality.
Ridge site (Fig. 2E)
In 1995 a second site with buried wood and other organic re-
mains was exposed along the Ridge crest approximately 1–200 m
southwest of the Glacier Toe site and 10–15 m higher. Many
sheared and shattered stumps and logs up to 25 cm in diameter
were lying on the surface or exposed in a series of annual mo-
raines formed between ca. 1992 and 2000 as the glacier margin
receded back down the southern flank of the ridge. Clumps of
moss and organic material were also found on the surface, and
although some tree stumps were possibly in growth position, no
actual paleo surface was exposed. A radiocarbon date of 3160 ±
70
14
C years BP (R1; Table 1) was obtained from 31 rings in a small
detrital wood fragment from this site. This date was at least
400 years younger than dates from the adjacent Glacier Toe site,
raising the possibility that trees at these two sites were killed by
different glacier events. The initial radiocarbon date indicated
that wood from the Ridge site was of similar age to the Upper
Extinguisher Tower site upvalley. Approximately 32 cross sections
from log fragments up to 2 m long were recovered from this site
between 1995 and 2000.
Fig. 2. Glacier Toe site. (A) Aerial oblique view, looking downvalley August 2004, showing the “Ridge” and “Creek” sites, the former stream delta
(upper right), and proglacial lake beyond. In 2004 the proglacial stream along the east margin of the glacier flowed into the ice along the foot of the
ridge (bottom left). (B) Sheared and splintered log in a gravel bar near the ice front, June 1993. (C) Sample R9501 in the creek, June 1995. This log had
315 rings dating between 794 and 1108 in chronology A. (D) Paleosol site exposed in the stream cut, June 1993. Till and the overlying paleosol were
exposed in the stream bank immediately upstream of the large boulder (far left). Small trees, roots, and organic debris were recovered from this section,
and a cross section of 123 rings from a root in this paleosol yielded a 3710 ± 70
14
C years BP date (G2; Table 1). (E) Splintered log and wood lying on the
surface and partially buried in till of the 1995 annual moraine immediately outside the glacier front on the Ridge crest, June 1995. [Colour online.]
1156 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

Lower Extinguisher Tower site (Fig. 3)
In 2003 British Columbia Parks personnel reported a newly
exposed site near Extinguisher Tower (Figs. 1,3). Two large logs
ca. 30–35 cm in diameter protruded from a stream cut on the east
side of the glacier. This site is some 400–500 m downvalley of the
Upper Extinguisher Tower site and separated from it by a bedrock
cliff (Fig. 3A). A radiocarbon date of 4780 ± 60
14
C years BP (ET1;
Table 1) was obtained from the outer hundred rings of one of these
logs (R0301), indicating that it is significantly older than both the
wood from the Upper Extinguisher site and the Glacier Toe and
Ridge sites over 3 km downvalley (Luckman 2007). The large
stream-bank exposure revealed that these large logs outcropped
about 1–2 m above a partially buried paleosol surface, developed
on and buried by till, which could be traced for over 50 m down-
stream. Several other logs were recovered along this section to-
gether with an erect in situ stump rooted in a rust-coloured
paleosol (Figs. 3F,3G). This colouration in the Canadian Rockies is
usually associated with the weathering of volcanic tephra (King
1984;Beaudoin and King 1994), although limited excavation did
not reveal a discrete tephra layer at this site. The most likely
source is the Mazama tephra dated at ca. 6900
14
C years BP (7.63 ±
0.15 ka; Zdanowicz et al. 1999).
Detrital logs and a second in situ stump were found in a small
dry tributary streambed some 60–100 m upvalley of the main
section (Figs. 3B,3C). Sixteen cross sections were recovered from
this new site, eight from each locality.
Methods
Dendrochronological analyses
Dendrochronological studies on the subfossil samples were
used to develop “floating chronologies” from this material at the
three sites by measuring and crossdating ring-width series from
these logs. Cross sections were prepared, sanded, and polished
using standard procedures (Stokes and Smiley 1968). Measure-
ments were carried out at The University of Western Ontario Tree-
ring Laboratory using the TRIM or Velmex systems with 0.001 or
0.01 mm precision or on a Velmex system at the Biogeography
Laboratory, Department of Earth Sciences, Brock University. Where
practical, two or more radii were measured for each sample and
cross correlation was carried out between radii to verify measure-
ment accuracy. Master chronologies were developed for each log.
Crossdating trials were carried out using the program COFECHA
(Holmes 1983;Grissino-Mayer 2001) as well as the program TSAP-
Win (Rinn 2011). For the COFECHA analyses, each tree-ring series
was first high-pass filtered, prewhitened, and log-transformed to
enhance the year-to-year variability and facilitate the identifica-
tion of possible dating errors. The TSAP analyses high-pass filter
the data and use the sign test, Gleichläufigkeit, and tvalues to
crossdate the samples.
The measured series averaged 160 years (range 42–405 years) in
length at the Glacier Toe and Ridge sites and 211 years (range
65–456 years) at the Lower Extinguisher site.
Fig. 3. Extinguisher Tower sites. (A) View of the east lateral margin of the glacier in 2003 showing the Upper (E) and Lower (ET) Extinguisher
Tower sites (photo by Chris Zimmermann, BC Parks). (B) Log in the abandoned tributary stream channel. (C) Aerial view of the lower site (ET).
The stars show critical locations: (a) in situ stump; (b) location of (B); (c) location of (D). (D) Large trunks in the lower section. R0301–R0401 has
been cut, and R0402 is being sampled. Note the orange surface material adjacent to the logs. (E) Detail of the lower section: (a) location of
R0301 and R0401; (b) location of R0403 (being cut); (c) location of R0407 and R0408. The paleosol can be traced downslope from R0301–R0401
passing below R0403 to R0408. (F) In situ stump R0408 and the cut log R0407 protruding horizontally higher up the section. Neither of these
could be crossdated. (G) Orange-stained paleosol adjacent to and below R0408. [Colour online.]
Luckman et al. 1157
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

At each site, the longest, clearest, and most complete series were
first correlated (in COFECHA) against all other series using 50 year
moving segments lagged with 50% overlap to identify provisional
sets of well-correlated series that could be grouped into several float-
ing master chronologies. Subsequently, the remaining “undated” log
chronologies were tested against these “master” series, and follow-
ing numerous trials, final master series were developed for both the
Lower Extinguisher and Glacier Toe – Ridge sites, combining series of
variable length and with different periods of overlap.
Radiocarbon dating
Initial exploratory radiocarbon dates (G1–G3, R1, and ET1; Table 1)
were obtained to determine the approximate age of the logs at
each site. These dates were from relatively large samples, prior to
any dendrochronological analyses, and were not anchored within
the floating tree-ring chronologies. Subsequently, a second set of
dates (G4, G5, R2, and ET2; Table 1) were obtained from crossdated
ring sequences to clarify temporal relationships between the dated
samples. Three additional dates (G6, G7, and ET3) were later obtained
in an attempt to resolve dating inconsistencies between the radio-
carbon and tree-ring dating from this material. These three later
dates were included in an accelerator mass spectrometry (AMS) cali-
bration run at Beta Analytic (Miami, Florida). They are the mean of
two (G6) or three (G7, ET3) separate AMS determinations and are the
most precisely constrained dates. An additional high-precision date
(R3) was obtained from the Klaus-Tschira-Archäometrie-Zentrum
laboratory in Germany. Further attempts to reconcile dating scenar-
ios were carried out by “wiggle matching” of selected
14
C dates using
the program OxCal version 4.3 (Ramsey et al. 2001).
Results
Glacier Toe and Ridge sites
At the Glacier Toe – lake site, 15 of 55 logs crossdated with other
series compared with 15 of 32 logs from the Ridge site (Tables S1,
S2
3
). This modest success reflects that many samples contained
growth anomalies such as reaction wood, radial asymmetry, or
very narrow ring series. Most samples were Engelmann spruce
(Picea engelmannii) or alpine fir (Abies lasiocarpa),
4
although a few
whitebark pine (Pinus albicaulis) may have been present. Mean ring-
widths were 0.48 mm (range 0.13–1.19 mm), and the series showed
relatively low mean sensitivity (0.215; see Fritts 1976) and high
first-order autocorrelation (mean 0.799) that indicates relatively
low interannual variability within ring series.
Rarely did contiguous ring series correlate over their entire
length: poorer correlations occurred most frequently at the begin-
ning or end of the series where tight rings or growth distortions
are usually greatest. Trees were considered to be crossdated and
assigned “floating years” if they crossdated strongly with the rele-
vant master chronology for at least 75 years or correlated strongly
with other radii or trees that did correlate with the master. Many
radial samples were only partial —i.e., the pith and (or), more
frequently, the outermost ring was missing or indistinguishable.
The outermost rings are often very narrow or poorly preserved.
However, outer ring dates are assumed to approximate death
dates when the outer surface of the sample clearly followed a ring
boundary over part of the circumference of the sample.
Ultimately, two separate chronologies were developed from
these series. Chronology A is 319 years in length, assigned dating
of 794–1112 A
5
and includes 20 logs crossdating with R9210 and
3
Supplementary data are available with the article through the journal Web site at http://nrcresearchpress.com/doi/suppl/10.1139/cjes-2016-0187.
4
Ring series of these two species are strongly correlated in this region.
5
This is a floating chronology where the relative age of tree rings is correct, but the absolute, calendar age, is undefined. The “dating” applied to the
chronology is in arbitrary “A” or “C” years.
Fig. 4. Chronology A samples. The age distribution and record of the 20 logs in chronology A at the Glacier Toe and Ridge sites. For details,
see Table S1.
3
Ages are given in chronology A years. The letters following the sample number indicate the collection site: C, creek bed or delta;
L, lake shore; R, ridge site. Logs with approximate death date (outer surface follows ring boundaries) are identified by the short vertical lines.
1158 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

R9212 (Fig. 4). Chronology C is 287 years long, assigned dating of
1731–2017 C
5
and includes nine logs that crossdate with sample
R9532 (Fig. 5).
6
The logs in chronology A are predominantly from
the Glacier Toe site (13/20), whereas the majority of samples in
chronology C are from the Ridge site (8/10; see Tables S1, S2
3
). The
longest series from log R2327 (405 years), found on the delta, did
not correlate with either master chronology and is treated as a
distinct series (see later in text).
Chronology A is composed of tree-ring series from 20 logs,
13 from the Glacier Toe site and seven from the Ridge site. The
youngest outermost dates of samples from both sites are the same
(1112 A, samples R9308 and R9533; Fig. 4; Table S1
3
). This chronol-
ogy indicates that mature trees had been growing for over
300 years at or close to the Glacier Toe site prior to the glacier
advance. Many of the logs were sheared and (or) had been
snapped, probably during the glacier advance, and several had root-
stocks attached (e.g., Fig. 2C). Although the eight youngest death
dates are between ca. 1108 and 1112 A, the full range of outer dates
indicates that these trees could have been killed or died over a 50–
150 year period.
7
Chronology C is composed of tree-ring series from 10 logs,
mainly (8/10) from the Ridge site. None of these logs crossdate with
chronology A, and the two master series do not crossdate. The end
dates of the seven youngest ring series range between 2022 and
2002 C (Fig. 5; Table S1
3
).
Seven
14
C dates have been obtained from the Glacier Toe site
and three from the Ridge site. Six of the dates cluster between
3500 and 3870
14
C years BP (4.31–3.77 cal ka BP
8
), with four diverg-
ing dates between 3320 and 3160
14
C years BP (G5, G6, R1, and R3,
range 3.56–3.21 cal ka BP). Unfortunately, most of the older (i.e.,
first acquired) radiocarbon dates have large error terms, and in-
spection of the calibration curves indicates a relatively low slope
and often multiple and wide calendar age estimates for a given
14
C
determination. Wiggle-matching analyses were applied to im-
prove the precision and accuracy of the dating of the tree-ring
chronologies. However, inconsistencies between some radiocar-
bon ages and the relative position of the dated tree rings in the
floating chronologies presented problems in determining approx-
imate calendar age equivalents for tree-ring chronology A.
9
Attempts to “wiggle match” the three radiocarbon dates from
chronology A (G1, G4, and G7) were unsuccessful, and the dating
for chronology A was finally modelled using only the G4 and G7
radiocarbon dates (Table 2). This results in a median dating of
4333–4015 cal years BP or ca. 4.33–4.02 cal ka BP (extreme range
4447–3969 cal years BP; Table 2;Fig. 6) for chronology A. Wiggle-
matched estimates for radiocarbon dates R2 and R3 (from log
R9532) were used to date chronology C. Median dates for this
chronology are from 3836 to 3545 cal years BP (full 2␴range
3874–3472 cal years BP; Table 2) and clearly do not overlap with
chronology A (Fig. 6).
Radiocarbon date G2 (4.06 cal ka BP) was taken from the trunk
of a large, 2 m high, detrital whitebark pine stump and rootstock
in gravels at the Glacier Toe site, and G3 (3.97 cal ka BP) is from a
complete cross section of a rootstock from the paleosol. Both
dates are close to and consistent with the end (youngest) date for
chronology A (3976 cal years BP) and confirm that the glacier
overran this site ca. 4.0 cal ka BP.
Log R2327
10
provided the longest ring-width record (405 years)
from samples at the Glacier Toe and Ridge sites. However, the
initial radiocarbon date (G5, ca. 3.56 cal ka BP) obtained from the
innermost rings of sample R0026 indicated this log was substan-
tially younger than other dated samples at this site and died
ca. 3.2–3.1 cal ka BP. A second AMS date of ca. 3.48 cal ka BP (G6)
from the same log confirmed this younger age. Wiggle matching
of dates G5 and G6 yielded modeled ages of 3.56 and 3.48 cal ka BP,
respectively (Table 2), and dates the floating 405 year chronology
of this log to 3647–3242 cal years BP (full 2␴range 3732–3178 cal
years BP; Table 2), indicating that it died ca. 750 years after the
6
Each sampled log is considered to be from a separate tree, although in some cases, e.g., R0021 and R0019 (Table S1
3
), they are possibly derived from the same tree.
7
However, not all outer ring dates are kill dates resulting from glacier activity: several sections are incomplete, and it is also probable that some samples
are from trees that were dead prior to the arrival of the glacier.
8
Most of the
14
C dates at Robson were initially reported as
14
C years BP. For clarity, the following discussion will only use the calendar equivalent ages i.e.,
cal ka BP = calendar years prior to 1950, rounded to decades per millennium, or in full (cal years BP). Both dating equivalents are given in Tables 1 and 2.
9
e.g., Radiocarbon dates G1 (3.77 cal ka BP, rings 984–1029 A) and G7 (4.31 cal ka BP, rings 835–876 A) are both from R9210. Date G4 (ca. 4.09 cal ka BP, rings
945–1002 A) from R9212 is consistent with G7 but not with G1.
10
R2327 was a large detrital log lying on the delta surface. This ring-width record is a composite floating chronology of 13 well-dated series from five pieces
(samples R0023–R0027) that is 405 years long. Both radiocarbon dates (G5 and G6) were from sample R0026.
Fig. 5. Chronology C. The age distribution and record of the 10 logs in chronology C. For details, see Table S1.
3
Ages are given in chronology C
years. The letters following the sample number indicate the collection site: C, creek bed or delta; L, lake shore; R, ridge site. Logs with
approximate death date (outer surface follows ring boundaries) are identified by the short vertical lines.
Luckman et al. 1159
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

trees in chronology A and ca. 300 years after the youngest tree in
chronology C. Although the wiggle-matched dating indicates
there is an overlap of ca. 100 years between chronology C and the
ring series in log R2327, the ring series from these trees do not
crossdate. The radiocarbon age of a small piece (31 rings) of de-
trital wood from the Ridge site (R1, 3.38 cal ka BP) is of similar age
to R2327.
Summary for the Glacier Toe and Ridge sites
Tree-ring and radiocarbon dating of logs from these two sites
indicate a mixed population of logs with three different ages. The
majority of dated samples from the Glacier Toe site crossdate in
chronology A and were killed over about a 100 year period. Eight
have outer tree-ring-determined dates between 1108 and 1112 A,
and nine of the remainder date between 1060 and 1105 A. Based on
the wiggle matching of chronology A, 19 of these trees died be-
tween ca. 3.98 and 4.10 cal ka BP. Although none of the tree-ring-
dated logs at the creek site is in situ (most logs sampled were
detrital, having been eroded by the river or washed out of the
glacier), the
14
C date of 3.97 cal ka BP (G3) derived from the
123 rings in a cross section of a root from the paleosol confirms
the presence of the glacier at this site ca. 4.00 cal ka BP. The
tree-ring results also indicate that the valley floor at this site was
forested for at least 300 years prior to that glacier advance.
The majority of the dated wood samples (8/10) recovered from
the Ridge site crossdate into the 292 year long chronology C, with
seven trees dying over a 21 year period (2022–2002 C). Wiggle
matching of radiocarbon dates (R2 and R3) indicates this chronol-
ogy extends from ca. 3.83–3.55 cal ka BP. The poorly constrained
radiocarbon date R1 (3.38 cal ka BP, 2␴range 3.56–3.21 cal ka BP;
Table 1) from a wood fragment on the Ridge partially overlaps the
C chronology. The earliest rings in the trees dating to chronol-
ogy C are ca. 3.83 cal ka BP, about 150 years younger than the logs
dating to chronology A, and they died ca. 450 years later. The logs
recovered from the Ridge site have been directly reworked by the
glacier as it pushed upslope against the ridge: some were lying on
the surface, and others were incorporated into a series of annual
moraines and flutings.
11
Based on this topographic setting, it is
assumed that original source of the wood is close to the surface,
and the wood is unlikely to have travelled any distance from its
11
Small blocks of moss and organic material were also incorporated into the till.
Table 2. Wiggle-matching results for radiocarbon dates.
Ref.
a14
C date laboratory No. Log ID
14
C date
(years BP)
cal
14
C,
median
(cal years BP)
cal
14
C date,
2␴range
(cal years BP)
Ring position
(midpoint) Agreement (%)
b
Chronology A (Glacier Toe site; 794–1112 A)
G4 Beta-200485 R9212 3730 ±70 4085 4348–3880 945–1002 A (974 A)
a Wiggle matched (G4, G7) 4154 4278–4108 102
G7 Beta-378911 R921014 3870±17 4311 4407–4280 827–867 A (847 A)
a Wiggle matched (G4, G7) 4280 4404–4234 105.2
b Chronology A (G4, G7) Start 4333 4457–4287
b Chronology A (G4, G7) End 4015 4139–3969 105.1
G1 Beta-62064 R9210 3500±60 3773 3959–3616 984–1029 A (1007 A)
a Wiggle matched (G1, G4, G7) 4081 4144–3997 0.5
Chronology A (G1, G4, G7) Start 4294 4357–4210
Chronology A (G1, G4, G7) End 3976 4039–3892 4.1
Chronology C (Ridge site; 1731–2017 C)
R2 Beta-200486 R9532 3550±60 3840 4062–3646 1810–1900 C (1855 C)
a Wiggle matched (R2, R3) 3712 3750–3639 61.2
R3 MAMS-31064 R9532B 3300±23 3521 3577–3462 2009–2017 C (2013 C)
a Wiggle matched (R2, R3) 3554 3592–3481 89.4
b Chronology C (R2, R3) Start 3836 3874–3763
b Chronology C (R2, R3) End 3545 3583–3472 65.2
R2327 (1–405 years)
G5 Beta-200484 R0026 A 3320±80 3556 3819–3378 41–71 of 405 years (57)
a Wiggle matched (G5, G6) 3590 3675–3526 115.6
G6 Beta-378910 R0026A2 3260±21 3482 3562–3414 151–192 of 405 years (172)
a Wiggle matched (G5, G6) 3475 3560–3411 103.1
b Chronology R2327 (G5, G6) Start 3647 3732–3583 1–405 years
b Chronology R2327 (G5, G6) End 3242 3327–3178 113.2
Chronology B (Lower Extinguisher site (ET); 113–632 B)
ET1 Beta-187090 R0301 4780±60 5511 5609–5324 510–610 B (560 B)
a Wiggle matched (ET1, ET2, ET3) 5331 5350–5310 43.2
ET2 Beta-200483 R0403 4770±60 5505 5602–5325 275–325 B (300 B)
a Wiggle matched (ET1, ET2, ET3) 5591 5610–5570 75.5
ET3 Beta-378912 R0402 4803±17 5505 5592–5481 356–396 B (376 B)
a Wiggle matched (ET1, ET2, ET3) 5515 5534–5494 61.5
b Chronology B Start 5778 5797–5757
b Chronology B End 5259 5278–5238 39.5
Note: The table lists the calibration results of selected radiocarbon dates and related tree-ring series and chronologies derived from grouping of dates using
OxCal 4.3 (Ramsey 2009). For other abbreviations, see Table 1.
a
G4, G5, etc., radiocarbon date identification; a, revised calibrated
14
C date (median and 2␴range) based on the combined calibration (wiggle matching) of the
14
C
dates in parentheses; b, wiggle-matched start and end dates of the related chronology (median and 2␴range of these start and end dates).
b
The suggested threshold for a satisfactory wiggle-matching agreement of a single date is 60%, the overall agreement threshold for a wiggle-matching analysis
depends on the number of dates included, i.e., it is 50% for two combined dates and 40.8% for three dates (Ramsey 1995,2009).
1160 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

original source. This suggests that most of the material exposed at
the Ridge site is from a younger forest bed with trees that began
growth ca. 3.83 cal ka BP and were overridden ca. 3.54 cal ka BP.
However, as both chronologies contain samples recovered from
both sites, there is clearly some admixture of debris from the two
sources. These results indicate Robson Glacier was advancing to a
position ca. 2 km upvalley of its LIA maximum at ca. 4.02 cal ka BP
(2␴range 4.14–3.97 cal ka BP). The glacier subsequently read-
vanced to a similar position ca. 3.55 cal ka BP (2␴range 3.58–
3.47 cal ka BP), incorporating material from forest trees almost
300 years old.
The
14
C dating on R2327 (G5 and G6; Table 1) indicates that this
log died ca. 3.24 cal ka BP (2␴range 3.33–3.18 cal ka BP), at least
300 years after other dated material recovered from the Glacier
Toe and Ridge sites. It grew at a site that had been ice free for
over 400 years. As this sample was found on the delta and was
not in situ, the most logical explanation is that this log was
derived from an exposure at an unknown locality upglacier and
has been reworked and transported to the delta by the glacial
meltwater stream.
The wood at the Upper Extinguisher site died after 3.36–
3.60 cal ka BP and postdates the trees killed at the Glacier Toe
site. The age range of these Upper Extinguisher samples over-
laps the dating of both chronology C and R2327 (Fig. 6). The
Upper Extinguisher samples were recovered close to the lateral
ice margin at a time (1989) when the main glacier tongue was
Fig. 6. Age distribution of radiocarbon dates and tree-ring chronologies from the Robson Glacier sites. Data for the single
14
C dates are as
given in Table 1. The wiggle-matched tree-ring chronologies are based on the
14
C dates indicated in parentheses. Data for the “wiggle
matched” chronologies are given in Table 2.
Table 3. Cross sections recovered from the Lower Extinguisher site.
Site
a
Sample No. Nrings
b
Inner
c
Outer
c
Death
d
Mean RW
e
Master
f14
C
g
TS R0401 290 343 632 × 0.59 M
TS R0301 355 255 609 0.52 M ET1
TS R0404 201 396 596 × 0.42 M
TS R0403 456 100 556 Sapwood 0.33 M ET2
SB R0416 270 279 548 × 0.38 M
SB R0413 216 306 522 0.27 Tentative
TS R0402 228 366 487 × 0.59 M ET3
TS R0418 303 113 415 × 0.53 M
SB R0414 233 116 348 × 0.26 M
SB R0417 174 119 294 × 0.51 Poor xd
SB R0419 193 90 282 0.57 Poor xd
SB R0408 174 In situ Asymmetric
TS R0406 155 Complex
SB R0411 135 × 0.37
TS R0407 112 nm
SB R0412 110 nm
TS R0405 79 nm
SB R0415 65 nm
Note: R0301 and R0401 are different cross sections from the same tree.
a
TS, till section; SB, dry former stream bed.
b
Nrings, total rings in the sample (including unmeasured or poorly dated inner or outer rings).
c
Inner and outer rings for the crossdated record (chronology B years).
d
×, outer complete ring present.
e
Mean RW, mean ring-width; nm, not measured.
f
M, included in chronology; Poor xd, weak crossdate.
g14
C identifies the
14
C dates from the sample.
Luckman et al. 1161
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

forming annual moraines at the Ridge site over 3 km downval-
ley. It is possible, therefore, that these trees were overwhelmed dur-
ing the same glacier advance ca. 3.55 cal ka BP. Alternatively, they
may have been killed during a later event of unknown extent, possi-
bly related to the death of R2327 ca. 3.24 cal ka BP.
New investigations at the Lower Extinguisher
Tower site
Eighteen cross sections were recovered from 17 logs
12
at this
site, eight from the till section, and nine from the dry creek bed
upstream of the section. Mean ring-width of the 13 measured
sections was 0.45 mm, and the series averaged ca. 250 years in
length (range 135–456 years). Ring series from eight of these cross
sections were crossdated into a 520 year long floating chronol-
ogy B (113–632 B),
13
including trees from both parts of the site
(Table 3). Unfortunately, neither of the in situ samples could be
crossdated.
Three radiocarbon dates, all dating ca. 5.51 cal ka BP, were ob-
tained from three logs at this site, but the tree-ring dating of the
samples differ (Table 1). Wiggle matching of the three dates pro-
vided consistent dates of 5.33, 5.51, and 5.60 cal ka BP, resulting in
a 519 year long chronology from 5260 to 5779 cal years BP (2␴
range of the last ring is 5279– 5239 cal years BP; Table 2). The outer
ring dates from the eight well-dated samples indicate these trees
were killed or died over a period of almost 200 years and that ice over-
rode this site by about 5.26 cal ka BP (2␴range 5.28–5.24 cal ka BP).
However, the downvalley extent of the glacier snout at this time
cannot be determined.
This site is at approximately 2100–2200 m, ca. 150–200 m above
present tree line. The presence of trees up to 450 years old at these
elevations ca. 5300 years ago, together with a relatively weathered
soil horizon, indicates these trees were probably remnants from a
higher, Hypsithermal, tree line. This also suggests that the glacier
advance that destroyed this forest ca. 5.26 cal ka BP was the first
Neoglacial advance to reach this point in the Robson Valley.
Regional correlation and discussion
The evidence available from Robson Glacier indicates at least
four major periods when Robson Glacier was advancing prior to
the LIA maximum, namely 1140–1350 A.D. (the Heusser site),
ca. 3.55 cal ka BP (Ridge site), ca. 4.02 cal ka BP (Glacier Toe site),
and ca. 5.26 cal ka BP (Lower Extinguisher site). There is also
possible evidence of glacier advance based on a detrital log that
died ca. 3.24 cal ka BP at the Glacier Toe site. However, the onset of
these advances is not known but probably occurred at least sev-
eral decades (based on tree kill dates) or possibly centuries before
the dates given. Moreover, in all cases, the maximum downvalley
extent of these events is not known, nor is the precise timing of
their culmination, though it would appear from the relative posi-
tions of the Heusser, Glacier Toe, Ridge, and Lower Extinguisher
sites that these advances were progressively more extensive over
time.
Evidence of a glacier advance ca. 5.25 cal ka BP has not been
recognized elsewhere in the Canadian Rockies. However, two
12
R0301 and R0401 are from the same log.
13
The dating of the “B” years is independent of both “A” and “C” years.
Fig. 7. Location of the main glacier sites referred to in the text. [Colour online.]
1162 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

detrital logs dating between 5.88 and 5.33 ka
14
were found at the
Castle Creek Glacier in the Cariboo Mountains approximately
75 km west of Mount Robson (Fig. 7;Maurer et al. 2012), and there
is evidence of glacier advances of similar age at many other sites in
the Coast Ranges of British Columbia (Menounos et al. 2009;Mood
and Smith 2015). There is also limited evidence for possible earlier
Neoglacial glacier events in the Rockies. Two detrital wood sam-
ples were recently recovered from the gravel outwash in front of
Dome Glacier that dated between 5.99 and 6.19 cal ka BP (5290 ±
30
14
C years BP, Beta-326374 and 5310 ± 30
14
C years BP, Beta-
326375; B.H. Luckman, this paper). This site had earlier yielded de-
trital whitebark pine snags dating between 6.79 and 7.46 ka years BP
(Luckman et al. 1993), and a branch of similar age (6.79–7.16 ka
years BP, 6090 ± 60
14
C years BP, Beta-264007; B.H. Luckman, this
paper) was recently washed out of the adjacent Athabasca Glacier.
These detrital logs indicate forest cover upvalley of these present
glacier snouts at several periods in the first half of the Holocene,
coincident with palynological evidence of higher tree lines at this
time (Kearney and Luckman 1983;Luckman and Kearney 1986).
However, the absence of in situ material and stratigraphic context
at Dome and Athabasca glaciers does not unequivocally indicate
that these trees were killed by glacial activity (see Luckman 1988);
therefore, evidence of earliest Holocene glacier events remains
fragmentary compared with the more extensive data being recov-
ered from sites in British Columbia. Nevertheless, it is probable
that this material is indirect evidence of several periods of glacier
advance ca. 6.2–6.0 and 7.5–6.8 ka in the Canadian Rockies.
Menounos et al. (2008) identified a regional period of glacier
advance across western North America ca. 4.2 ka based on glacial
and lacustrine evidence. In the Canadian Rockies, the evidence for
this event was based on two
14
C dates from Robson (G2 and G3;
Table 1) and three dates between 4.8 and 4.1 ka at Boundary Gla-
cier near the Columbia Icefield. Evidence for this event is also
found at Castle Creek Glacier in the Cariboo Mountains where
three detrital logs, moss, and an in situ stump dating 4.35–3.84 ka
(mostly 4.15–3.98 ka; Maurer et al. 2012) have been recovered. A
second in situ stump dating 4.96–4.45 ka was also found at the
Castle Creek glacier site. Detrital logs of similar age were found at
two adjacent glaciers (Chiqui Glacier, three logs 4.14–3.73 ka and
Chap Glacier, 4.97–4.85 ka; Maurer et al. 2012). The Glacier Toe site
at Robson now provides the strongest evidence for a regional
glacier advance ca. 4.2–4.0 ka in the Rockies. However, with the
exception of the site at Boundary Glacier, all of the other evidence
previously reported for this event is from sites located west of
the Continental Divide in British Columbia (Koch et al. 2007;
Menounos et al. 2009;Mood and Smith 2015).
The best documented Neoglacial glacier event in the Canadian
Rockies is the “Peyto Advance” (Luckman et al. 1993;Menounos
et al. 2009) dating between 3000 and 2800
14
C years BP (calibrated
ca. 2.90–3.4 ka). This evidence is best seen at Saskatchewan Gla-
cier where in situ stumps and abundant detrital wood have been
dated between 2940 ± 60 and 2760 ± 60
14
C years BP (3.32–2.78 ka;
Wood and Smith 2004). Recently, two large logs (ca. 10 and 7 m
long) at an adjacent site have yielded similar dates of 2860 ±
30 years BP (ca. 3.03 cal ka BP, Beta-326376) and 3010 ± 30 years BP
(ca. 3.21 cal ka BP, Beta-326377; B.H. Luckman, this paper). At
Robson, the material recovered from the Upper Extinguisher site
(3.25–3.69 ka) was originally tentatively correlated with the Peyto
Advance (Luckman et al. 1993). However, these samples are 200–
300 years older than the classic material found at the Saskatche-
wan and Peyto Glacier sites and may be related to the glacier
advance ca. 3.55 cal ka BP identified by the tree-ring and
radiocarbon-dated logs at the Ridge site downvalley. The 3.24 cal
ka death date for log R2327 provides the only evidence at Robson
that corresponds with the timing of the Peyto Advance. This date
is 300 years younger than the death date for trees at the Ridge site
and similar to dates from detrital logs killed by the Peyto Advance
at the Saskatchewan Glacier. However, stronger evidence is needed
to document the presence of the Peyto Advance at Robson Glacier.
Given the relatively large error terms on some, especially older,
radiocarbon dates, the limited availability of dates, and the ab-
sence of a stratigraphic context (the samples are mainly detrital),
it is challenging to disentangle the complex history of the period
between ca. 4.0 and 3.0 ka at Robson. It seems likely that the
glacial history of this period is as complicated as that of the Little
Ice Age and the last two millennia documented in Europe and
elsewhere (e.g., Barclay et al. 2013;Nicolussi and Kerschner 2014;
Le Roy et al. 2015). New stratigraphic sections and better radio-
carbon dating control will be needed to resolve this complex
history. There is no evidence at Robson for glacier events between
ca. 3.24 cal ka BP and the early Little Ice Age advance ca. 1140–
1350 A.D. at the Heusser site downvalley.
Summary and conclusions
Based on studies of subfossil wood, Robson Glacier was advanc-
ing at the Lower Extinguisher site, ca 5.5 km upvalley of the LIA
terminus ca. 5.26 cal ka BP; at the Glacier Toe and Ridge sites
(ca. 2 km upvalley) ca. 4.02 cal ka BP and ca. 3.55 cal ka BP; and at
the Heusser site (0.5–1 km upvalley) between 1140 and 1350 A.D.
prior to the LIA maximum ca. 1783 A.D. A single, 400 year old
detrital log that died ca. 3.24 cal ka BP may also provide evidence
of the regional Peyto Advance at Robson Glacier. The utilization of
the wiggle-matching approach using multiple
14
C dates from sam-
ple locations determined by dendrochronological analyses en-
abled the recognition of
14
C outliers as well as increasing the
precision and accuracy of the dating of glacier advances. The Rob-
son record includes periods of glacier advance not previously well
documented from the Canadian Rockies and is more similar to
that from sites further west in British Columbia than other sites in
the Rockies. These new results confirm multiple periods of glacier
advance in the Rockies in the late Holocene and provide the most
securely dated post Hypsithermal glacier advance identified to
date.
The Robson record also demonstrates some of the difficulties in
identifying discrete glacier events based on wood preserved in
glacier forefields (see Ryder and Thomson 1986). The dated mate-
rial at Robson comes mainly from valley floor sites lacking strati-
graphic context and may contain mixed assemblages of detrital
wood derived from deposits of different ages. At sites with abun-
dant material, preferably with some in situ trees or a paleosol,
evidence of the minimal extent and timing of glacier events can
be derived (see Le Roy et al. 2015). However, results from the
Glacier Toe, Ridge, and Lower Extinguisher sites at Robson indi-
cate death dates from trees buried during the same event may
have a range of 1–200 years. This, combined with radiocarbon
error terms, suggests that it is difficult to distinguish between (or
correlate) inferred glacier events that may be 200–300 years apart
from small amounts of detrital materials. Studies from lateral
moraine sections and historical materials in British Columbia,
Alaska, and the Alps (e.g., Barclay et al. 2013;Le Roy et al. 2015;
St-Hillaire and Smith 2017) indicate several periods with multiple
late Neoglacial glacier advances that may be only a few centuries
apart. It is therefore likely that the late Neoglacial history of the
Rockies is more complex than the limited story presented here
and that continued work in glacier forefields is needed to eluci-
date a more complete history.
14
Dates referred to as ka are reported as published.
Luckman et al. 1163
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.

Acknowledgements
Many individuals have been involved in the studies at Mount
Robson. We would like to thank Fred Dalley, Les Jozsa, Gordon
Frazer, Jessica Lusted, Scott St. George, Jim Hamilton, Chris Somr,
Martin Groleau, Emma Watson, Chris Zimmermann (BC Parks),
Carla Aruani, and David, Heather, and Helen Luckman plus Yel-
lowhead Helicopters (Valemount) for assistance in the field and
laboratory; Wayne van Velzen and his staff for permission to work
in Mount Robson Provincial Park; Dan McCarthy for access to the
Velmex system at the Biogeography Laboratory of Brock Univer-
sity; Ron Hatfield at Beta Analytic for assistance with radiocarbon
dating; Karen vanKerkoerle (The University of Western Ontario)
for Figs. 1,4–7; David Barclay and an anonymous reviewer for
comments on an earlier version of this manuscript; and the
Natural Sciences and Engineering Research Council of Canada
(NSERC) for financial support via Operating/Discovery Grants to
B.H. Luckman over the years.
References
Barclay, D.J., Yager, E.M., Graves, J., Kloezko, M., and Calkin, P.K. 2013, Late
Holocene glacial history of the Copper River Delta, coastal south-central
Alaska, and controls on valley glacier fluctuations. Quaternary Science
Reviews, 81: 74–89.
Beaudoin, A.B., and King, R.H. 1994. Holocene palaeoenvironmental record pre-
served in a paraglacial alluvial fan, Sunwapta Pass, Jasper National Park,
Alberta, Canada. Catena, 22: 227–248.
Cautley, R.W., and Wheeler, A.O. 1924. Report of the Commission Appointed to
Delimit the Boundary between the Provinces of Alberta and British Colum-
bia. Office of the Surveyor General, Ottawa.
Fritts, H.C. 1976. Tree rings and climate. Academic Press, London.
Grissino-Mayer, H.D. 2001. Evaluating Crossdating Accuracy: A Manual and Tu-
torial for the Computer Program COFECHA. Tree-Ring Research, 57: 205–221.
Grove, J.M. 2004. The Little Ice Age (2
nd
ed.). Routledge, London.
Heusser, C.J. 1956. Postglacial environments in the Canadian Rocky Mountains.
Ecological Monographs, 26: 253–302.
Holmes, R.L. 1983. Computer-assisted quality control in tree-ring dating and
measurement. Tree-Ring Bulletin, 43: 69–75.
Kearney, M.S., and Luckman, B.H. 1983. Holocene timberline fluctuations in
Jasper National Park, Alberta. Science, 221: 261–263.
King, R.H. 1984. Quaternary paleosols in the Canadian Rockies and their signif-
icance for establishing a chronologic framework. In Correlation of Quater-
nary chronologies. Edited by W.C. Mahaney. Geobooks, Norwich, England,
pp. 243–258.
Koch, J., Clague, J.J., and Osborn, G. 2007. Pre- ‘Little Ice Age’ glacier fluctuations
in Garibaldi Provincial Park, southern Coast Mountains, British Columbia.
The Holocene, 17: 1069–1078.
Le Roy, M., Nicolussi, K., Deline, P., Astrade, L., Edouard, J.-L., Miramont, C., and
Arnaud, F. 2015. Calendar-dated glacier variations in the western European
Alps during the Neoglacial: the Mer de Glace record, Mont Blanc massif.
Quaternary Science Reviews, 118: 1–22.
Luckman, B.H. 1988. 8000-year-old wood from the Athabasca Glacier. Canadian
Journal of Earth Sciences, 25: 148–151.
Luckman, B.H. 1995. Calendar-dated, early Little Ice Age glacier advance at Rob-
son Glacier, British Columbia, Canada. The Holocene, 5: 149–159.
Luckman, B.H. 2000. The Little Ice Age in the Canadian Rockies. Geomorphol-
ogy, 32: 357–384.
Luckman, B.H. 2006. The Neoglacial history of Peyto Glacier. In Peyto Glacier:
One century of Science. Edited by M. Demuth, S. Munro, and G. Young. Na-
tional Hydrological Research Institute Science Report Number 8, NHRI, Sas-
katoon, pp. 25–57.
Luckman, B.H. 2007. Holocene Glacier fluctuations in the central Canadian
Rockies. Session Holocene and latest Pleistocene alpine glacier fluctuations.
XVII INQUA Congress, Cairns, Australia 28 July - 3 August 2007. Quaternary
International, 167-8: 254.
Luckman, B.H., and Kearney, M.S. 1986. Reconstruction of Holocene change in
alpine vegetation and climate in the Maligne Range, Jasper National Park,
Alberta, Canada. Quaternary Research, 26: 244–261.
Luckman, B.H., and Wilson, R.J.S. 2005 Summer temperature in the Canadian
Rockies during the last millennium – a revised record. Climate Dynamics, 24:
131–144.
Luckman, B.H., Holdsworth, G., and Osborn, G.D. 1993. Neoglacial Glacier Fluc-
tuations in the Canadian Rockies. Quaternary Research, 39: 144–153.
Luckman, B.H., Colenutt, M.E., Kavanagh, T.A., Lusted, J.R., and Diraddo, C. 1996.
Field Investigations in the Canadian Rockies in 1995. Report to Parks Canada,
Alberta Parks and B.C. Parks Service, May 1996, iii + 41 p.
Maurer, M., Menounos, B., Luckman, B.H., Osborn, G.D., Clague, J.J., Smith, I.R.,
and Atkinson, N. 2012. Late Holocene glacier expansion in the Cariboo and
northern Rocky Mountains, British Columbia, Canada. Quaternary Science
Reviews, 51: 71–80.
Menounos, B, Clague, J.J., Osborn, G., Luckman, B.H., Lakeman, T.R., and
Minkus, R. 2008. Western Canadian Glaciers advance in concert with climate
change ca 4.2 ka. Geophysical Research Letters: L07501. doi:10.1029/
2008GL033172.
Menounos, B., Clague, J.J., Osborn, G.D., and Luckman, B.H. 2009. Holocene and
Latest Pleistocene Glacier Fluctuations in the Canadian Cordillera. Quater-
nary Science Reviews, 28: 2049–2074.
Mood, B.J., and Smith, D.J. 2015. Holocene glacier activity in the British Columbia
Coast Mountains, Canada. Quaternary Science Reviews, 128: 14–36.
Nicolussi, K., and Kerschner, H. 2014. Lateglacial and Holocene advance record of
the Gepatschferner, Kaunertal, Tyrol. In From the foreland to the Central
Alps. Field Trips to selected sites of Quaternary research in the Tyrolean and
Bavarian Alps. Edited by H. Kerschner, K. Krainer, and C. Spötl. Berlin (Geozon)
pp. 144–151.
Osborn, G.D., Robinson, B.J., and Luckman, B.H. 2001. Holocene and Latest Pleis-
tocene Fluctuations of Stutfield Glacier, Canadian Rockies. Canadian Journal
of Earth Sciences, 38: 1141–1155.
Ramsey, C.B. 1995. Radiocarbon calibration and analysis of stratigraphy: The
OxCal program. Radiocarbon, 37(2): 425–430.
Ramsey, C.B. 2009. Bayesian analysis of radiocarbon dates. Radiocarbon, 51(1):
337–360.
Ramsey, C.B., van der Plicht, J., and Weninger, B. 2001. ‘Wiggle matching’ radio-
carbon dates. Radiocarbon, 43(2A): 381–389.
Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., et al. 2013.
IntCal13 and Marine I3 radiocarbon age calibration curves 0– 50,000 years cal BP.
Radiocarbon, 55(4): 1869–1887.
Rinn, F. 2011. TSAP-Win: Time Series Analysis and Presentation for Dendro-
chronology and Related Applications – Version 4.64. User reference. Heidel-
berg, Germany (http://www.rinntech.com).
Ryder, J.T., and Thomson, B. 1986. Neoglaciation in the southern Coast Moun-
tains of British Columbia: chronology prior to the late-Neoglacial maximum.
Canadian Journal of Earth Sciences, 23: 273–287.
Solomina, O.N., Bradley, R.S., Jomelli, V., Geirsdottir, A., Kaufman, D.S., Koch, J.,
et al. 2016. Glacier fluctuations during the past 2000 years. Quaternary Sci-
ence Reviews, 149: 61–90.
St-Hillaire, V.M., and Smith, D.J. 2017. Holocene glacier history of Frank Mackie
Glacier, northern British Columbia Coast Mountains. Canadian Journal of
Earth Sciences, 54: 76–87.
Stokes, M.A., and Smiley, T.L. 1968. An Introduction to Tree-Ring Dating. Univer-
sity of Chicago Press, 73 p.
Wood, C., and Smith, D.J. 2004. Dendroglaciological evidence for a Neoglacial
advance of the Saskatchewan Glacier, Banff National Park, Canadian Rocky
Mountains. Tree-Ring Research, 60: 59–65.
Zdanowicz, C.M., Zielinski, G.A., and Germani, M.S. 1999. Mount Mazama erup-
tion: Calendrical age verified and atmospheric impact assessed. Geology, 27:
621–624.
1164 Can. J. Earth Sci. Vol. 54, 2017
Published by NRC Research Press
Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by Canadian Science Publishing on 11/08/17
For personal use only.