Pre-‘Little Ice Age’ glacier fluctuations in Garibaldi ...facultyweb.kpu.ca/~jkoch/ho_2007.pdf · activity in western Canada; and (3) discuss the likely climate forcing mechanisms

The most direct evidence of pre-‘Little Ice Age’ glacier activ-ity is found in glacier forefields and includes in situ tree stumps(Ryder and Thomson, 1986; Luckman et al., 1993; Wiles et al.,1999; Wood and Smith, 2004), detrital logs and branches (Ryderand Thomson, 1986; Luckman et al., 1993; Wiles et al., 1999),and organic soils and detrital wood exposed in composite lateralmoraines (Röthlisberger, 1986; Osborn et al., 2001; Reyes andClague, 2004). Here, we present a detailed study of Holoceneglacier fluctuations prior to the ‘Little Ice Age’ in GaribaldiProvincial Park in the southern Coast Mountains of BritishColumbia (Figure 1) based on such evidence. Our objectives areto: (1) summarize new data from Garibaldi Park; (2) compare therecord from the park with other records of Holocene glacieractivity in western Canada; and (3) discuss the likely climateforcing mechanisms. Evidence for ‘Little Ice Age’ and twenti-eth-century glacier fluctuations in the park is presented else-where and is not discussed further here (Koch et al., 2007).

Study area

Garibaldi Provincial Park is located in the southern CoastMountains about 70 km north of Vancouver, British Columbia(Figure 1). The park contains more than 150 glaciers, which are

Introduction

Recent studies point to significant and rapid fluctuations of climatethroughout the Holocene (Bond et al., 2001; Mayewski et al.,2004). Alpine glaciers have long been used to document these fluc-tuations (Denton and Karlén, 1973), because they respond rapidlyto changes in their mass balance and thus to changes in tempera-ture and precipitation. Unfortunately, most studies of past glacierfluctuations allow reconstruction of climate variability only oncentennial and decadal timescales. Furthermore, many studies ofHolocene glacier fluctuations suffer from the fact that glacieradvances in the Northern Hemisphere during the past millennium(‘Little Ice Age’) were the most extensive of the Holocene andobliterated or obscured most of the evidence of previous advances.Pre-‘Little Ice Age’ moraines are rare in western Canada and theirages are generally not well constrained (Osborn, 1985; Ryder andThomson, 1986; Osborn and Karlstrom, 1988; Osborn and Gerloff,1996; Larocque and Smith, 2003). Sediments in lakes downvalleyof glaciers can provide a more complete record of Holocene glacialactivity (Leonard, 1986a, b; Desloges and Gilbert, 1995; Leonardand Reasoner, 1999; Menounos et al., 2004), but the record com-monly is complicated by non-climatic factors.

Pre-‘Little Ice Age’ glacier fluctuations in Garibaldi Provincial Park, CoastMountains, British Columbia, CanadaJohannes Koch,1* Gerald D. Osborn2 and John J. Clague1

( 1Department of Earth Sciences, Simon Fraser University, Burnaby, British Columbia V5A 1S6,Canada; 2Department of Geology and Geophysics, University of Calgary, Calgary, Alberta T2N 1N4,Canada)

Received 21 July 2006; revised manuscript accepted 10 June 2007

Abstract: Holocene glacier fluctuations prior to the ‘Little Ice Age’ in Garibaldi Provincial Park in the BritishColumbia Coast Mountains were reconstructed from geomorphic mapping and radiocarbon ages on 37 samplesof growth-position and detrital wood from glacier forefields. Glaciers in Garibaldi Park were smaller than atpresent in the early Holocene, although some evidence exists for minor, short-lived advances at this time. Thefirst well-documented advance dates to 7700–7300 14C yr BP. Subsequent advances date to 6400–5100, 4300,4100–2900 and 1600–1100 14C yr BP. Some glaciers approached their maximum Holocene limits several timesduring the past 10 000 years. Periods of advance in Garibaldi Park are broadly synchronous with advances else-where in the Canadian Cordillera, suggesting a common climatic cause. The Garibaldi Park glacier record isalso broadly synchronous with the record of Holocene sunspot numbers, supporting previous research that sug-gests solar activity may be an important climate forcing mechanism.

Key words: Glacier advances, dendroglaciology, Holocene, solar forcing, Garibaldi Park, Coast Mountains,British Columbia.

The Holocene 17,8 (2007) pp. 1069–1078

© 2007 SAGE Publications 10.1177/0959683607082546

*Author for correspondence (e-mail: [email protected])

© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. at COLLEGE OF WOOSTER on January 18, 2008 http://hol.sagepub.comDownloaded from

http://hol.sagepub.com

among the southernmost glaciers in the Coast Mountains. Thetotal area of ice cover in the park is about 390 km².

The landscape of Garibaldi Park has been shaped by Quaternarycontinental and alpine glaciation and locally by Quaternary volcan-ism (Mathews, 1958). The park is underlain mainly by graniticrocks of the Coast Plutonic Complex and Cretaceous metasedimen-tary and volcanic rocks of the Gambier Group (Monger andJourneay, 1994). The southwestern part of the park is dominated byrocks of the Plio-Pleistocene Mount Garibaldi volcanic complex(Mathews, 1958).

The northwest-trending Coast Mountains form an orographicbarrier to moisture-laden air flowing from the Pacific Ocean, thusthere is a strong west–east environmental gradient acrossGaribaldi Park. Climate is maritime in the southwestern part of thepark but becomes increasingly continental to the north and east.Average annual precipitation at Squamish, at the head of HoweSound, is 2370 mm, but it is only 1230 mm at Whistler, 60 km tothe north. Overall, climate is humid and cool, with very wet win-ters and dry summers.

The study area is located in the Mountain Hemlock biogeocli-matic zone, which is characterized by long, cool, wet winters andshort, cool summers (Brooke et al., 1970). The growing season isshort because of heavy snowfall in winter. Arboreal vegetation inforefields investigated in this study is dominated by mountainhemlock (Tsuga mertensiana) and subalpine fir (Abies lasiocarpa),but locally includes Engelmann spruce (Picea engelmannii), white-bark pine (Pinus albicaulis) and yellow cedar (Chamaecyparisnootkatensis).

Methods

Glacier forefields were visited in 2002, 2003 and 2004 and sys-tematically searched for detrital and in situ fossil wood. Discs werecut from fossil wood found in lateral and end moraines and at thesurface. They were air dried and prepared for analysis by sandingwith progressively finer grades of sandpaper to enhance the defini-tion and contrast of annual tree-ring boundaries. Tree rings were

1070 The Holocene 17,8 (2007)

6

7

11

4

5

8

9

10

123°W

1

2

3

Squam

ish R.

Che

akam

us R

.

R.

U.S.A.

Canada

BritishColumbia Alberta

HoweSound

Lillooet Lake

Lillooet

500 kmPaciicOcean

Whistler

Squamish

Study area

Pemberton

50°N

alpine

glaciers

lakes/streams4 km

N

Figure 1 Location of Garibaldi Provincial Park and glaciers investigated in this study. 1, Wedgemount Glacier; 2, Spearhead Glacier; 3, DeckerGlacier; 4, Overlord Glacier; 5, Helm Glacier; 6, Sphinx Glacier; 7, Sentinel Glacier; 8, Warren Glacier; 9, Lava Glacier; 10, West Stave Glacier;11, Stave Glacier



Table 1 Pre ‘Little Ice Age’ radiocarbon ages in Garibaldi Park

Glacier Site Laboratory no. 14 C age Cal. yr Comment(yr BP) BP

Wedgemount Beta-170671 8650 ± 60 9540–9630 branch; 90 m from snoutSpearhead Beta-157268a 3900 ± 80 4230–4430 branch; near snout

Beta-168423a 3900 ± 60 4280–4420 branch; near snoutDecker Beta-157265a 3200 ± 70 3350–3480 in situ stump on cliff

Beta-157262a 2960 ± 50 3060–3220 log near stumpsBeta-157263a 2960 ± 40 3070–3210 in situ snag on cliffBeta-157264a 2920 ± 50 2990–3160 in situ stump on cliff

Overlord ovgl-b Beta-170665a 6170 ± 70 6980–7160 log; on surface 50 m from snoutovgl-a Beta-170660a 5980 ± 70 6730–6910 log; in till 140 m from snoutovgl-a Beta-170667a 5890 ± 70 6630–6800 log; on surface 100 m from snout

Helm hegl-a Beta-168430 8900 ± 60 9920–10 090 partially buried log 140 m from 1997 snouthegl-b Beta-186523 4080 ± 40 4520–4620 stump; melted out of ice in 2003

Sphinx spgl-e GSC-6770b 7720 ± 80 8420–8560 branch; 100 m from snoutGSC-1993c 7640 ± 80 8380–8480 log; on surface; location uncertain

spgl-a Beta-186509 5830 ± 60 6560–6700 snag; in till 300 m from snoutspgl-b Beta-208685 4280 ± 70 4810–4960 log; in till 500 m from snoutspgl-c Beta-186510 3560 ± 70 3820–3930 snag; in till 800 m from snoutspgl-d Beta-186511 1570 ± 40 1450–1520 snag; on surface below bedrock knoll 600 m from snout

Sentinel segl-c Beta-148787b 7720 ± 70 8430–8550 branch; on surface near snoutsegl-c Beta-157267b 7470 ± 80 8280–8360 branch; on surface near snoutsegl-c Beta-148786b 7380 ± 80 8160–8320 branch; on surface near snout

GSC-1477d 6170 ± 150 6890–7250 branch; in till 2 km inside ‘Little Ice Age’ limitsegl-b Beta-186508 6040 ± 60 6790–6950 snag in till 300 m from present snout

GSC-2027c 5300 ± 70 5990–6130 in situ stump 1 km inside ‘Little Ice Age’ limitWarren wagl-a Beta-168425 8050 ± 60 8950–9030 stick 250 m from 2000 snout

wagl-d Beta-148789 6370 ± 70 7250–7330 in situ stump 95 m from 2000 snoutwagl-d Beta-148790 6360 ± 80 7250–7340 in situ stump 100 m from 2000 snoutwagl-c Beta-148788 5780 ± 70 6500–6660 in situ stump 600 m from 2000 snoutwagl-b Beta-168424 5700 ± 50 6410–6540 partially buried large log

Lava lagl-a Beta-168426 6170 ± 60 7000–7160 in situ stump, Garibaldi Nevelagl-a Beta-168427 6050 ± 50 6840–6970 in situ stump, Garibaldi Nevelagl-Y Y-140 bise 5850 ± 180 6450–6860 in situ stump, Garibaldi Nevelagl-Y Y-140 bisf 5260 ± 200 5890–6280 in situ stump, re-date of Y-140 bislagl-b Beta-186521 5760 ± 60 6490–6640 in situ stump 400 m from snoutlagl-f Beta-186520 5130 ± 40 5890–5930 stump on surfacelagl-e Beta-186517 3190 ± 40 3380–3450 log in woody layer in lateral moraine

West Stave wsgl-a Beta-170668 1080 ± 60 930–1010 small log in lateral moraineStave stgl-a Beta-171096 6250 ± 70 7150–7260 detrital stump in till

Note: Dated samples are outer-perimeter wood comprising less than 25 rings.a Osborn et al. (2007).b Menounos et al. (2004).c Lowdon and Blake (1975).d Lowdon and Blake (1973).e Preston et al. (1955).f Stuiver et al. (1960).

measured on a Velmex stage with a precision of ± 0.001 mm usingthe software MeasureJ2X. Two to four radii were measured toreplicate data and reduce the possibility of missing rings or falserings (Stokes and Smiley, 1968). Floating chronologies weredeveloped for sites where more than one tree was sampled. Theyshow that many trees at a site were killed at about the same timeand support the argument that tree death was caused by an advanc-ing glacier. The total number of rings is a minimum for the lengthof time the site was not covered by ice and thus a minimum for theinterval between successive glacier advances. The floatingchronologies were visually compared to correlate marker rings andchecked and verified using the International Tree-Ring Data Bank(ITRDB) software program COFECHA. They were then cross-dated (50-yr dated segments lagged by 25 years, with a criticallevel of correlation (99%) set at 0.32) to create master ring-widthchronologies (Holmes, 1983). Floating chronologies that could not

be crossdated with living trees were radiocarbon dated. Standardradiocarbon ages on outer rings were calibrated using the programCALIB 5.0 (Stuiver et al., 2005). Calibrated ages, shown in brack-ets in this paper, are 1σ ranges.

Interpretation of radiocarbon ages from glacier forefields requiresconsideration of the provenance and location of the dated material(Osborn, 1986; Röthlisberger, 1986; Ryder and Thomson, 1986;Luckman, 1998). Stumps in growth position in Garibaldi Park showevidence of having been sheared off by glaciers. Bark is commonlypreserved as a result of burial of the stumps in sediment. Detritalwood in glacier forefields shows evidence of glacial transport,including missing bark, abrasion, embedded sediment and flatten-ing. This wood is interpreted to have been eroded from in situ treesby glaciers and transported at basal and englacial positions withinthe glaciers. Preservation of bark or relatively small branches orroots is interpreted to indicate short transport. Only samples in

Johannes Koch et al.: Pre-‘Little Ice Age’ glacier flucuations in British Columbia 1071



growth position provide unequivocal evidence for glacier advanceand extent at a given time. Detrital wood in a stratigraphic context,such as a lateral moraine or till in a glacier forefield, provides amaximum age for the advance. It also defines the minimum extentof the glacier, because the glacier must have reached at least as faras the wood and probably farther. The possibility must be consid-ered that detrital wood within till could have avalanched into theforefield or died of old age, but in such cases the wood neverthelessstill provides a maximum age for the advance. If tree-ring or radio-carbon ages on detrital wood in one glacier forefield are similar tothose in other forefields, it is unlikely that the samples were intro-duced by avalanches or died of old age.

Results

Early-Holocene activity

Detrital wood in three glacier forefields in Garibaldi Park suggeststhat glaciers were generally less extensive in the early Holocenethan in 2002–2004, but that they may have advanced on severaloccasions to near present-day glacier margins. A large, abradedand partially buried log about 140 m from the 1997 snout of HelmGlacier yielded a radiocarbon age of 8900 ± 60 14C years BP(9920–10 090 cal. yr BP; Table 1). A split and frayed fragment ofdetrital wood, likely the remains of a small conifer trunk, recov-ered from the top of a bedrock knob about 90 m from the 2002margin of Wedgemount Glacier gave an age of 8650 ± 60 14C yearsBP (9540–9630 cal. yr BP). A shredded stick found on a bar of ameltwater stream about 250 m from the 2000 snout of WarrenGlacier (site a in Figure 2) yielded an age of 8050 ± 60 14C yearsBP (8950–9030 cal. yr BP).

Advance between 7700 and 7300 14C yr BP(8500–8100 cal. yr BP)Evidence was found in two glacier forefields for a small advance atthe time of the ‘8.2 ka event’, an abrupt cooling event most likely

caused by a massive discharge of freshwater into the North Atlanticaround 8470 cal. yr BP (Alley and Ágústsdóttir, 2005). Two small,shattered and partially buried fragments of coniferous wood werefound about 100 m from the 2002 snout of Sphinx Glacier (site ein Figure 3). One of the fragments yielded a radiocarbon age of7720 ± 80 14C yr BP (8420–8560 cal. BP; Menounos et al., 2004).Another, similar age was obtained on a 2 m long stem recoveredfrom among large boulders in a meltwater channel near the snoutof Sphinx Glacier in the 1970s (7640 ± 80 14C yr BP (8380–8480cal. yr BP); Lowdon and Blake, 1975). The locations of both sam-ples indicate that, prior to the event, Sphinx Glacier was less exten-sive than at present. Three small pieces of detrital wood 50–400 mfrom the snout of Sentinel Glacier gave ages of 7720 ± 70 14C yr BP(8430–8550 cal. yr BP), 7470 ± 80 14C yr BP (8280–8360 cal. yrBP) and 7380 ± 80 14C yr BP (8160–8330 cal. yr BP; Menounos etal., 2004). Sentinel Glacier was less extensive at the time of thisinferred advance than today. Neither Sphinx nor Sentinel glaciersadvanced far beyond their present limit during the 8.2 ka event.

Advances between 6400 and 5100 14C yr BP(7300–5800 cal. yr BP)Evidence for one or more glacier advances during this period wasfound in six glacier forefields in Garibaldi Park, showing theregional significance of the event. A degraded, detrital stumprecovered from till about 800 m from the 2002 snout of StaveGlacier gave a radiocarbon age of 6250 ± 70 14C yr BP (7150–7260cal. yr BP), indicating that the glacier was more extensive at thattime than today.

Surface detrital wood 40–200 m from the 2002 snout ofOverlord Glacier gave ages of 6170 ± 70 14C yr BP (6980–7160cal. yr BP) and 5890 ± 70 14C yr BP (6630–6800 cal. yr BP). Athird sample, embedded in till 40 m farther downstream, yieldedan age of 5980 ± 70 14C yr BP (6730–6910 cal. yr BP). The loca-tions of the three samples indicate that Overlord Glacier was prob-ably less extensive than today prior to this event and advanced intomature forest around 6000 14C yr BP. One of the dated samples

1072 The Holocene 17,8 (2007)

Figure 2 Forefield of Warren Glacier in 2002, showing study sites (photograph by J. Koch). View is to the southeast



crossdated with two other pieces of detrital wood, providing afloating chronology (Table 2).

Branches and snags were found in a woody layer between twotills and in an adjacent stream about 300 m from the snout of thesouth tributary glacier of Sphinx Glacier (site a in Figure 3). Oneof the snags gave an age of 5830 ± 60 14C yr BP (6560–6700 cal.yr BP). A floating ring chronology was built from two of the sam-ples in till and one in the creek (Table 2). The evidence from thissite indicates that Sphinx Glacier was about its present size priorto an advance that overrode forest around 5800 14C yr BP.

Wood collected from till near the snout of Sentinel Glacier(Figure 1) in 1970 yielded a radiocarbon age of 6170 ± 150 14C yrBP (6890–7250 cal. yr BP; Lowdon and Blake, 1973). The exactlocation of this site is unknown, but it is likely to be the same sitethat we sampled in 2002 (Figure 4a). Our samples range from smalltrunks to large snags; one of the snags gave an age of 6040 ± 60 14Cyr BP (6790–6950 cal. yr BP). The two ages overlap at their 1σlimits, supporting the inference that they are from the same site.Two samples from the till and two samples in a meltwater streamadjacent to the till exposure were used to construct a floatingchronology (Table 2). An in situ stump partially buried in till was

collected in 1973 and dated at 5300 ± 70 14C yr BP (5990–6130 cal.yr BP; Lowdon and Blake, 1975). Although we were unable tolocate this stump, it is within 1 km of the maximum ‘Little Ice Age’margin of Sentinel Glacier (Lowdon and Blake, 1975).

The radiocarbon ages from Sentinel Glacier afford a moredetailed reconstruction of this event than the ages from Stave,Overlord and Sphinx glaciers. When Sentinel Glacier advancedinto forest 6100 14C yr BP, it was about as extensive as today. Thedated wood shows little evidence of transport and thus likely hadbeen growing near the sample site prior to the advance. By 530014C yr BP, Sentinel Glacier was at least 1 km more extensive thanat present and at least half its ‘Little Ice Age’ maximum size. Theglacier thus expanded between 6100 and 5300 14C yr BP, althoughit is not known if glacier growth was slow and continuous or theresult of two separate advances.

The Warren Glacier forefield has a similar record (sites b, c, andd in Figure 2). Two in situ stumps (Figure 4c) were found at sited about 100 m from the 2000 glacier margin. They yielded ages of6370 ± 70 14C yr BP (7250–7330 cal. yr BP) and 6360 ± 80 14C yrBP (7250–7340 cal. yr BP). An exposed and fragmented in situstump (Figure 4b) at site c about 600 m from the 2000 glacier


Figure 3 Forefield of Sphinx Glacier in 2002 and study sites (photograph by J. Koch). View is to the southeast. Dashed line delineates the glacier’s ‘Litte Ice Age’ limit

Table 2 Floating tree-ring chronologies for subalpine fir samples from glacier forefields in Garibaldi Park

Glacier Chronology 14C age Length Correlation No. radii/trees(yr BP) (yr)

Overlord site ovgl-a 5890 ± 70 138 0.562 13/4Helm sites hegl-b and -c 4080 ± 40 135 0.612 22/13Sphinx site spgl-a 5830 ± 60 184 0.494 6/3

site spgl-b 4280 ± 70 115 0.698 4/2site spgl-c 3560 ± 70 197 0.427 5/3

Sentinel sites segl-a and -b 6040 ± 60 149 0.574 8/4Lava sites lagl-b, -c, and -d 5760 ± 60 482 0.437 15/8

site lagl-f 5130 ± 40 263 0.477 7/3



margin gave an age of 5780 ± 70 14C yr BP (6500–6660 cal. yrBP). A large weathered log partially buried in till at site b yieldedan age of 5700 ± 50 14C yr BP (6410–6540 cal. yr).

We interpret these ages as follows. Warren Glacier was aboutthe same size as today when it advanced into forest at site d about6350 14C yr BP. By 5800 14C yr BP, the glacier had advancedabout 600 m downvalley and was overriding a forest at site c. Thedated log at site b, farther downvalley, shows that the glacier con-tinued to advance until at least 5700 14C yr BP. As at SentinelGlacier, it is not known if Warren Glacier advanced slowly over aperiod of 650 years or two or more times during this interval.

Garibaldi Neve, an icefield east of Mt Garibaldi, feeds six gla-ciers, including Lava Glacier (Figure 5). Two in situ stumps on aridge at the northeast corner of the neve (site a in Figure 5) yieldedradiocarbon ages of 6170 ± 60 14C years BP (7000–7160 cal. yrBP) and 6050 ± 50 14C yr BP (6840–6970 cal. yr BP). An in situstump on a nunatak in the neve was sampled in the 1950s (site Yin Figure 5). It gave ages of 5850 ± 180 14C yr BP (6450–6860 cal.yr BP; Preston et al., 1955) and 5260 ± 200 14C yr BP (5890–6280cal. yr BP; Stuiver et al., 1960); the latter is assumed to be themore reliable age (Stuiver et al., 1960). The exact location of thissample is unknown, but Figure 5 shows the two possible nunataksin the icefield. Two in situ and three detrital stumps were recov-ered along the stream draining Lava Glacier about 400 m from the2003 snout (site b, Figure 5). One of the in situ stumps yielded anage of 5760 ± 60 14C yr BP (6490–6640 cal. yr BP). Two moresites with detrital wood were found in the same area, site c c. 50 msoutheast of site b and site d c. 110 m northeast of site b. Samplesfrom sites b, c and d were successfully crossdated, providing afloating chronology of almost 500 years (Table 2). Numerousbranches, large logs, snags and stumps were found on an alluvialsurface (site f, Figure 5), 20–100 m downstream of a gorge eroded

in bedrock and till. The surface detrital wood is derived from awood layer separating two tills within the gorge (arrow near site fin Figure 5). Unfortunately, the site could not be safely accessed.Three detrital stumps at site f were crossdated (Table 2), and aradiocarbon age of 5130 ± 40 14C yr BP (5890–5930 cal. yr BP)was obtained on one of them.

The ages from Garibaldi Neve indicate a thickening of icebetween 6100 and 5900/5200 14C yr BP, either slowly or duringtwo separate events. Lava Glacier advanced into mature forestswith trees more than 450 years and 250 years old 5800 and 510014C yr BP, respectively. It likely had the same extent 5800 14C yrBP as today. By 5100 14C yr BP, when it deposited a till at site f,Lava Glacier was more than half its ‘Little Ice Age’ size.

Advance about 4300 14C yr BP (4900 cal. yrBP)Evidence for this advance was found only at Sphinx Glacier (siteb in Figure 3). Detrital branches and large logs were found in ameltwater stream in the glacier forefield and within a woody layerseparating two tills about 500 m from the snout of the southerntributary glacier. The outer 15 rings of one of the logs yielded aradiocarbon age of 4280 ± 70 14C yr BP (4810–4960 cal. BP). Afloating chronology, spanning 115 years, was constructed for twosamples collected from the till (Table 2). It is unknown if thelower till at site b correlates to the lower till at site a (Figure 3, see‘Advances between 6400 and 5100 14C yr BP’), thus indicating asingle advance, or if the upper till at site a is the lower till at siteb, indicating two separate advances. However, evidence at otherforefields indicates that glaciers reached much farther downvalleyat the end of the previous advance period. It thus is more likelythat the tills record two advances and that Sphinx Glacier recededafter the 6400–5100 BP advance period to an unknown position

1074 The Holocene 17,8 (2007)

Figure 4 (a) Wood in lateral moraine of Sentinel Glacier (site segl-b; 6040 ± 60 14C yr BP; photograph by J. Koch). (b) In situ stump at WarrenGlacier (site wagl-d; 6360 ± 80 14C yr BP; photograph by G. Osborn). (c) In situ stump at Warren Glacier (site wagl-c; 5780 ± 70 14C yr BP; pho-tograph by G. Osborn)




Figure 5 Forefield and study sites at Lava Glacier and Garibaldi Neve in 2003 (photograph by J. Koch). View is to the north. The exact locationof site Y is unknown. Wood sampled on the floodplain (site f) was eroded from a wood layer in a gully (white arrow)

Figure 6 Fossil wood deposited during glacier advances between 4600and 3000 cal. yr BP. (a) Wood melting out of Helm Glacier (site hegl-b;4080 ± 40 14C yr BP). (b) Wood layer in the lateral moraine of LavaGlacier. (site lagl-e; 3190 ± 40 14C yr BP) (photographs by J. Koch)

upvalley of site b for at least 115 years. This issue aside, anadvance occurred about 4300 14C yr BP, reaching at least site b.

Advances between 4100 and 2900 14C yr BP(4600–3000 cal. yr BP)Evidence for one or more advances during this period was foundin five glacier forefields in Garibaldi Park. The evidence showsthat glaciers became more extensive over time during this period.

Thirteen branches and stumps were found in till near the termi-nus of Helm Glacier (Figure 6a) in 2003. They appear to be asso-ciated with a weathering horizon in the till. A radiocarbon age of4080 ± 40 14C yr BP (4520–4620 cal. yr BP) was obtained on theoutermost 12 rings of the sample shown in Figure 6a, and a float-ing chronology was constructed from all of the samples (Table 2).The evidence suggests that Helm Glacier was less extensive thantoday for at least 135 years around 4100 14C yr BP. Shortly after-wards, it advanced to a more extensive position than at present.

Small wood fragments about 20 m above the present surface ofSpearhead Glacier date to 3900 ± 80 14C yr BP (4230–4430 cal. yrBP) and 3900 ± 60 14C yr BP (4280–4420 cal. yr BP; Osborn et al.,2007). They suggest that Spearhead Glacier was thicker and thusmore extensive 3900 14C yr BP than today.

Additional evidence for this event was found along the creekdraining the southern tributary of Sphinx Glacier (Figure 1) about800 m from the 2002 snout (site c in Figure 3). A layer of branchesand snags occurs within till, and one of the snags gave an age of3560 ± 70 14C yr BP (3820–3930 cal. yr BP). A floating chronol-ogy was built from three samples (Table 2). The evidence indi-cates that the glacier was at least 800 m more extensive 3600 14Cyr BP than at present.

A large log in the west lateral moraine of Lava Glacier (Figure6b) yielded a radiocarbon age of 3190 ± 40 14C yr BP (3380–3450cal. yr BP). The log is part of a wood layer that can be traced overa distance of 110 m across several gullies. The till overlying the



wood layer was deposited by Lava Glacier shortly after 3200 14Cyr BP, when it was about as extensive as at the ‘Little Ice Age’maximum.

Decker Glacier also was near its ‘Little Ice Age’ limit at thistime. In situ snags and stumps on a bedrock cliff 40–75 m abovethe present glacier surface yielded radiocarbon ages of 3200 ± 7014C yr BP (3350–3480 cal. yr BP), 2960 ± 40 14C yr BP(3070–3210 cal. yr BP) and 2920 ± 50 14C yr BP (2990–3160 cal.yr BP; Osborn et al., 2007). A detrital log below the bedrock cliffgave an age of 2960 ± 50 14C yr BP (3060–3220 cal. yr BP).Decker Glacier thickened between 3200 and 2900 14C yr BP tonear its ‘Little Ice Age’ maximum extent.

Advances between 1600 and 1100 14C yr BP(1500–1000 cal. yr BP)Evidence for one or more glacier advances in the first millenniumAD was found in the forefields of two glaciers in Garibaldi Park. Adetrital snag with relatively fine roots, located below a bedrockknob at Sphinx Glacier (site d in Figure 3), yielded an age of1570 ± 40 14C yr BP (1450–1520 cal. yr BP). The roots indicateshort transport, and it is likely that the stump was rooted on thebedrock knob. The site is only 600 m from the 2002 ice margin,but the glacier would have to thicken considerably to cover it; thusSphinx Glacier was much more extensive 1600 14C yr BP than atpresent. Supporting evidence for the same advance or a youngerone comes from the forefield of West Stave Glacier. A small log

embedded in the outermost lateral moraine close to its crestyielded an age of 1080 ± 60 14C yr BP (930–1010 cal. yr BP).

Discussion

Chronology

Glaciers in Garibaldi Park were less extensive than at presentthroughout the early Holocene. Radiocarbon ages from Wedge-mount, Helm and Warren glaciers may record minor advancesbetween 8900 and 8000 14C yr BP (Figure 7). However, prior tothe 8.2 ka event (Menounos et al., 2004), glaciers must have beenless extensive than today, because detrital wood recovered nearthe margins of Sphinx and Sentinel glaciers dates to 7700–730014C yr BP.

Warren, Stave, Sentinel, Sphinx, Overlord and Lava glaciersadvanced between 6400 and 5100 14C yr BP. The glaciers advancedover forests up to 800 m downvalley from present snouts between6400 and 5800 14C yr BP. Sentinel and Lava glaciers were no morethan 1–1.5 km from their maximum Holocene limits between 5300and 5100 14C yr BP. It is possible that this event was a single con-tinuous advance that happened over 1300 years, but more likely itwas a period like the ‘Little Ice Age’ with several advances.

Evidence for an advance around 4300 14C yr BP was found onlyat Sphinx Glacier. There, ice advanced at least 500 m fartherdownvalley than at present (Figure 7). Taken alone, the evidence

1076 The Holocene 17,8 (2007)

14C years BP10,000 9000 8000

OverlordHelm

SphinxSentinel

WarrenGaribaldi

Stave

Wedgemount Spearhead

Lava

DeckerWest Stave

7000 6000 5000 4000 3000 2000 1000 0

10,00011,000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0

(a)

(b)

(c)

? ? ?

min

max

120

100

80

60

40

20

0

Cal years BP

Figure 7 Relation between Holocene glacier activity in Garibaldi Park and sunspot numbers. (a) Generalized reconstruction of relative glacierextent in Garibaldi Park. (b) Life spans of trees killed by glacier advance. (c) Reconstructed decadal (thin line) and smoothed (bold line) sunspotnumbers (Solanki et al., 2004). Vertical dark grey bars indicate periods of glacier advance in Garibaldi Park



at Sphinx Glacier could be interpreted to indicate slow growth ofglaciers between 6000 and 4300 14C yr BP. Combined with evi-dence from other forefields, however, it is more likely that glaciersreceded between 5300 and 4300 14C yr BP to positions similar tothe present before readvancing (Figure 7).

Some glaciers achieved present-day extents prior to 4100 14C yrBP, the time that Helm Glacier advanced into forest and depositedtill at the site of its present snout. Spearhead Glacier thickened,and Sphinx Glacier advanced up to 800 m, between 3900 and3600 14C yr BP. Lava and Decker glaciers thickened and advancedinto mature forests between 3200 and 2900 14C yr BP, reaching towithin 500 m of their Holocene maximum positions at this time. Itis not known whether this was a single advance with slow growthover 1200 years or was more complex with several advances.

Sphinx Glacier retreated to near its present margins after thisadvance period, before readvancing into mature forest around 1600 14Cyr BP. Stave Glacier was near its Holocene maximum by about 110014C yr BP. This advance period was also followed by recession, beforethe start of the ‘Little Ice Age’ about AD 1000 (Koch et al., 2007).

Regional comparisonsEvidence of glacier activity in the early Holocene in westernCanada is sparse, but available data indicate that glaciers weresmaller than today at that time (Ryder and Thomson, 1986;Luckman, 1988). Advances during the early Holocene were likelyminor and short-lived. Menounos et al. (2004) summarized evi-dence for an advance in Garibaldi Park at the time of the 8.2 kacooling event, and a glacier forefield in the Rocky Mountains(Luckman, 1988) also has evidence for this event. Dated samplesin Garibaldi Park are near present-day glacier margins, indicatingthat ice extent was less than today prior to the advance and that theadvance was minor.

Glacier advances between 6000 and 5000 14C yr BP in westernCanada have been termed the ‘Garibaldi phase’ (Ryder andThomson, 1986). Evidence for this event has previously been lim-ited to a few sites in the Coast Mountains (Ryder and Thomson,1986; Laxton et al., 2003; Smith, 2003). Detrital wood in theCanadian Rocky Mountains dating to this period has been inter-preted to record treelines higher than today (Luckman et al.,1993), but it may also indicate glacier advance at this time. TheGaribaldi phase evidently was a lengthy period during which gla-ciers achieved fairly extensive positions. It is likely a much morecomplex event than evidence currently suggests.

Glacier advances between 3300 and 1900 14C yr BP are termed the‘Tiedemann Advance’ in the Coast Mountains (Ryder and Thomson,1986). The same event in the Rocky Mountains, although there datedto between 3300 and 2500 14C yr BP, is termed the ‘Peyto Advance’(Luckman et al., 1993). Evidence for advances at this time is wide-spread in western Canada (Denton and Stuiver, 1966; Rampton, 1970;Denton and Karlén, 1977; Osborn, 1986; Osborn and Karlstrom, 1989;Luckman et al., 1993; Luckman, 1995, 2006; Jackson and Smith,2005; Lewis and Smith, 2005; Smith et al., 2005). In most areas, gla-ciers achieved extents only slightly smaller than during the ‘Little IceAge’. The Tiedemann/Peyto phase was probably as complex as the‘Little Ice Age’ and included multiple advances (Luckman, 1995;Reyes and Clague, 2004; Haspel et al., 2005).

Evidence for an advance in the first millennium AD is wide-spread in western North America (Reyes et al., 2006). Glaciersprobably achieved extents at this time that were comparable withthose of the preceding and following advance periods.

Climate forcingSome researchers have noted a relation between Holocene glacierfluctuations and changes in solar activity (Denton and Karlén, 1973;Karlén and Kuylenstierna, 1996). Sunspot numbers and glacierfluctuations in Alaska and western Canada correspond on a decadal

timescale over the past millennium (Lawrence, 1950; Wiles et al.,2004; Luckman and Wilson, 2005; Koch et al., 2007). Solanki et al.(2004) inferred sunspot numbers over the past 11 000 years fromdifferences in the production of atmospheric 14C recorded in trees.Comparison of this record with the record of Holocene glacier fluc-tuations in Garibaldi Park suggests that the major periods of glacieradvance correspond with times of low sunspot numbers (Figure 7).This correspondence suggests that changes in solar activity mayhave affected glacier behaviour in Garibaldi Park and likely else-where during the Holocene. The period between 3000 and 4500 cal.yr BP, however, shows little correspondence between glacier activ-ity and solar activity and requires other explanations.

Conclusions

Holocene glacier fluctuations in Garibaldi Provincial Park weredetermined through geomorphic mapping and radiocarbon datingof detrital and in situ fossil wood in glacier forefields. Glaciers inthe park were smaller than at present in the early Holocene. Fivemajor periods of pre-‘Little Ice Age’ glacier advance date to7700–7300, 6400–5100, 4300, 4100–2900 and 1600–1100 14C yrBP. These periods are broadly synchronous with advances docu-mented elsewhere in western Canada. The evidence from GaribaldiPark suggests that Holocene climate was much more varied thanpreviously thought. Glaciers approached their Holocene maximumpositions on three occasions prior to the ‘Little Ice Age’. Glaciersin Garibaldi Park, notably Sphinx Glacier, afford the most com-plete record of Holocene glacier activity in western Canada. Theglacial record is in good agreement with the record of reconstructedsunspot numbers during the Holocene, suggesting that glacier fluc-tuations in Garibaldi Park may have been forced in part by changesin solar activity.

Acknowledgements

We thank Chris Platz and the other rangers of Garibaldi Park fortheir support, and Sami Solanki for providing data on sunspot num-bers. Will Keats-Osborn, Janet Guglich, Chanone Ryane, VanessaVallis, Clay Carlson, Scott Carlson and Matt Macleod providedassistance in the field. The Natural Sciences and EngineeringResearch Council of Canada and Geological Society of Americaprovided funding for the research. Early drafts of the manu-script benefited from reviews by B.H. Luckman, K. Nicolussi, C. Schlüchter and D.J. Smith.

ReferencesAlley, R.B. and Ágústsdóttir, A.M. 2005: The 8 ka event: cause andconsequences of a major Holocene abrupt climate change. QuaternaryScience Reviews 24, 1123–49.Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N.,

Showers, W., Hoffmann, S., Lotti-Bond, R., Hajdas, I. and Bonani,

G. 2001: Persistent solar influence on North Atlantic climate duringthe Holocene. Science 294, 2130–36.Brooke, R.C., Peterson, E.B. and Krajina, V.J. 1970: The subalpinemountain hemlock zone. Ecology of Western North America 2,148–349.Denton, G.H. and Karlén, W. 1973: Holocene climatic variations –their pattern and possible cause. Quaternary Research 3, 155–205.––— 1977: Holocene glacial and tree-line variations in the WhiteRiver Valley and Skolai Pass, Alaska and Yukon Territory.Quaternary Research 7, 63–111.Denton, G.H. and Stuiver, M. 1966: Neoglacial chronology, north-eastern St. Elias Mountains, Canada. American Journal of Science264, 577–99.




Desloges, J.R. and Gilbert, R. 1995: The sedimentary record ofMoose Lake: implications for glacial activity in the Mount Robsonarea, British Columbia. Canadian Journal of Earth Sciences 32, 65–78.Haspel, R., Osborn, J. and Spooner, I. 2005: Neoglacial deposits ofBear River Glacier, northern Coast Ranges, British Columbia.Abstracts of the Annual Meeting of the Western Division, CanadianAssociation of Geographers 2005, Lethbridge, AB.Holmes, R.L. 1983: Computer-assisted quality control in tree-ringdating and measurement. Tree-Ring Bulletin 43, 69–78.Jackson, S. and Smith, D.J. 2005: Outlaw dendroglaciology atSurprise Glacier in the northwestern British Columbia CoastMountains. Abstracts of the Annual Meeting of the Western Division,Canadian Association of Geographers 2005. Lethbridge AB.Karlén, W. and Kuylenstierna, J. 1996: On solar forcing of Holoceneclimate: evidence from Scandinavia. The Holocene 6, 359–65.Koch, J., Clague, J.J. and Osborn, G.D. 2007: Glacier fluctuationsduring the past millennium in Garibaldi Provincial Park, southernCoast Mountains, British Columbia. Canadian Journal of EarthSciences 44, 1215–33.Larocque, S.J. and Smith, D.J. 2003: Little Ice Age glacial activityin the Mt. Waddington area, British Columbia Coast Mountains,Canada. Canadian Journal of Earth Sciences 40, 1413–36.Lawrence, D.B. 1950: Glacier fluctuations for six centuries in south-eastern Alaska and its relation to solar activity. Geographical Review40, 191–222.Laxton, S., Smith, D.J., Desloges, J.R. and Day, A. 2003: Late-Holocene history of Fyles Glacier, Coast Mountains, British Columbia.Abstracts of the Joint Meeting of the Association of Canadian MapLibraries and Archives, Canadian Association of Geographers,Canadian Cartographic Association, and Canadian Regional ScienceAssociation 2003. Victoria BC.Leonard, E.M. 1986a: Varve studies at Hector Lake, Alberta Canada,and the relationship between glacial activity and sedimentation.Quaternary Research 25, 199–214.––— 1986b: Use of lacustrine sedimentary sequences as indicators ofHolocene glacial history, Banff National Park, Alberta, Canada.Quaternary Research 26, 218–31.Leonard, E.M. and Reasoner, M.A. 1999: A continuous Holoceneglacial record inferred from proglacial lake sediments in BanffNational Park, Alberta, Canada. Quaternary Research 51, 1–13.Lewis, D.H. and Smith, D.J. 2005: Dendrochglaciological evidenceof late Holocene glacier activity at Forrest Kerr Glacier, AndreiIcefield, northern British Columbia Coast Mountains. Abstracts of theAnnual Meeting of the Canadian Association of Geographers 2005.London ON.Lowdon, J.A. and Blake, W., Jr 1973: Radiocarbon dates XIII.Geological Survey of Canada Paper 73-7, 61 pp.––— 1975: Radiocarbon dates XV. Geological Survey of CanadaPaper 75-7, 32 pp.Luckman, B.H. 1988: 8000 year old wood from the AthabascaGlacier, Alberta. Canadian Journal of Earth Sciences 25, 148–51.––— 1995: Calendar-dated, early ‘Little Ice Age’ glacier advance atRobson Glacier, British Columbia, Canada. The Holocene 5, 149–59.––— 1998: Dendroglaciologie dans les Rocheuses du Canada.Géographie physique et Quaternaire 52, 139–51.––— 2006: The Neoglacial history of Peyto Glacier. In Demuth,M.N., Munro, D.S. and Young, G.J., editors, Peyto Glacier: one cen-tury of science. National Hydrology Research Institute, ScienceReport No. 8, 25–57.Luckman, B.H. and Wilson, R.J.S. 2005: Summer temperatures inthe Canadian Rockies during the last millennium; a revised record.Climate Dynamics 24, 131–44.Luckman, B.H., Holdsworth, G. and Osborn, G.D. 1993:Neoglacial glacier fluctuations in the Canadian Rockies. QuaternaryResearch 39, 144–53.Mathews, W.H. 1958: Geology of the Mount Garibaldi map-area,southwestern British Columbia, Canada. Part II: geomorphology andQuaternary volcanic rocks. Bulletin of the Geological Society ofAmerica 69, 179–98.Mayewski, P.A., Rohling, E.E., Stager, J.C., Karlén, W., Maasch,

K.A., Meeker, L.D., Meyerson, E.-A., Gasse, F., van Kreveld, S.,

Holmgren, K., Lee-Thorp, J., Rosqvist, G., Rack, F., Staubwasser,

M., Schneider, R.R. and Steig, W.J. 2004: Holocene climate vari-ability. Quaternary Research 62, 243–55.Menounos, B., Koch, J., Osborn, G., Clague, J.J. and Mazzucchi, D.

2004: Evidence for early Holocene glacier advance, Coast Mountains,British Columbia, Canada. Quaternary Science Reviews 23, 1543–50.Monger, J.W.H. and Journeay, J.M. 1994: Guide to the geology andtectonic evolution of the southern Coast Mountains. GeologicalSurvey of Canada Open File 2490, 77 pp.Osborn, G.D. 1985: Holocene tephrostratigraphy and glacial fluctua-tions in Waterton Lakes and Glacier National Parks, Alberta andMontana. Canadian Journal of Earth Sciences 22, 1093–101.––— 1986: Lateral-moraine stratigraphy and Neoglacial history ofBugaboo Glacier, British Columbia. Quaternary Research 26, 171–78.Osborn, G. and Gerloff, L. 1996: Latest Pleistocene and earlyHolocene fluctuations of glaciers in the Canadian and NorthernAmerican Rockies. Quaternary International 38/39, 7–19.Osborn, G.D. and Karlstrom, E.T. 1988: Holocene history ofBugaboo Glacier, British Columbia. Geology 16, 1015–17.––— 1989: Holocene moraine and paleosol stratigraphy, BugabooGlacier, British Columbia. Boreas 18, 311–22.Osborn, G.D., Robinson, B.J. and Luckman, B.H. 2001: Holoceneand latest Pleistocene fluctuations of Stutfield Glacier, CanadianRockies. Canadian Journal of Earth Sciences 38, 1141–55.Osborn, G.D., Menounos, B.P., Koch, J., Clague, J.J. and Vallis, V.

2007: Multi-proxy record of Holocene glacial history: Spearhead andFitzsimmons Ranges, southern Coast Mountains, British Columbia.Quaternary Science Reviews 26, 479–93.Preston, R.S., Person, E. and Deevey, E.S. 1955: Yale natural radio-carbon measurements II. Science 122, 954–60.Rampton, V.N. 1970: Neoglacial fluctuation of the Natzhat andKlutlan Glaciers, Yukon Territory, Canada. Canadian Journal ofEarth Sciences 7, 1236–63.Reyes, A.V. and Clague, J.J. 2004: Stratigraphic evidence for multi-ple Holocene advances of Lillooet Glacier, southern Coast Mountains,British Columbia. Canadian Journal of Earth Sciences 41, 903–18.Reyes, A.V., Wiles, G., Smith, D.J., Barclay, D.J., Allen, S.,

Jackson, S., Larocque, S., Laxton, S., Lewis, D., Calkin, P.E. andClague, J.J. 2006: Expansion of alpine glaciers in Pacific NorthAmerica in the first millennium A.D. Geology 34, 57–60.Röthlisberger, F. 1986: 10 000 Jahre Gletschergeschichte der Erde.Verlag Sauerländer.Ryder, J.M. and Thomson, B. 1986: Neoglaciation in the southernCoast Mountains of British Columbia: chronology prior to the lateNeoglacial maximum. Canadian Journal of Earth Sciences 23, 273–87.Smith, D. 2003: Late-Holocene glacier activity in the TchaikazanValley, Ts’yl-os Provincial Park, British Columbia. Abstracts of theAnnual Meeting of the Western Division, Canadian Association ofGeographers 2003. Prince George, BC.Smith, D.J., Jackson, S., Laxton, S. and Lewis, D.H. 2005: Late-Holocene dendroglaciology in the Stewart-Cassiar region, northernBritish Columbia Coast Mountains. Abstracts of the Annual Meetingof the Canadian Association of Geographers 2005. London, ON.Solanki, S.K., Usoskin, I.G., Kromer, B., Schüssler, M. and Beer,

J. 2004: Unusual activity of the Sun during recent decades comparedto the previous 11,000 years. Nature 431, 1084–87.Stokes, M.A. and Smiley, T.L. 1968: An introduction to tree-ring dat-ing. The University of Arizona Press.Stuiver, M., Deevey, E.S. and Gralenski, L.J. 1960: Yale naturalradiocarbon measurements V. American Journal of Science,Radiocarbon Supplement 2, 49–61.Stuiver, M., Reimer, P.J. and Reimer, R.W. 2005: Calib 5.0.Retrieved 1 May 2005 from http://radiocarbon.pa.qub.ac.uk/calib/ Wiles, G.C., Barclay, D.J. and Calkin, P.E. 1999: Tree-ring dated‘Little Ice Age’ histories of maritime glaciers from western PrinceWilliam Sound, Alaska. The Holocene 9, 163–73.Wiles, G.C., D’Arrigo, R.D., Villalba, R., Calkin, P.E. andBarclay, D.J. 2004: Century-scale solar variability and Alaskan tem-perature change over the past millennium. Geophysical ResearchLetters 31, L15203.Wood, C. and Smith, D.J. 2004: Dendroglaciological evidence for aNeoglacial advance of the Saskatchewan Glacier, Banff NationalPark, Canadian Rocky Mountains. Tree-Ring Research 60, 59–65.

1078 The Holocene 17,8 (2007)



Documents

Pre-‘Little Ice Age’ glacier fluctuations in Garibaldi ...facultyweb.kpu.ca/~jkoch/ho_2007.pdf · activity in western Canada; and (3) discuss the likely climate forcing mechanisms