dendroglaciological evidence for holocene glacial advances in the … · dendroglaciological...

Dendroglaciological evidence for Holocene glacialadvances in the Todd Icefield area, northernBritish Columbia Coast Mountains

Scott I. Jackson, Sarah C. Laxton, and Dan J. Smith

Abstract: Accelerated glacial recession and downwasting in Pacific North America is exposing land surfaces and fea-tures buried by glacial advances that, in many locations, predate the recent Little Ice Age (LIA). Dendrochronologicanalyses of increment core samples from living trees (Abies lasiocarpa, Tsuga mertensiana) and samples of subfossilwood collected in the Todd Icefield area, Boundary Ranges, British Columbia Coast Mountains, provide the basis for adendroglaciological and radiocarbon-based reconstruction of late Holocene glacier activity. Five intervals of glacier ex-pansion were recorded by trees killed or buried by advancing glaciers: (1) an advance prior to �3000 14C years BP;(2) an advance at �3000 14C years BP that coincides with the regional Tiedemann advance; (3) an unattributed advanceat 2300 14C years BP; (4) a two-phase advance at �1700 and �1450 14C years BP that corresponds with the regionalFirst Millennium advance; (5) an advance with three phases of expansion that began prior to �750 14C BP and is con-sistent with the regional early LIA interval and a two-phase interval of late LIA expansion culminating after �240 and100 years BP. This chronology of late Holocene glaciation matches that emerging from similar investigations in thecoastal cordillera of Pacific North America and provides additional support for the regional significance of both theTiedemann and the First Millennium advances.

Résumé : Une récession et une fonte glaciaire accélérées dans l’océan Pacifique en Amérique du Nord exposent des sur-faces de terrain et des traits caractéristiques enfouis sous des avancées glaciaires lesquelles, dans bien des cas, datentd’avant le récent Petit Âge glaciaire. Des analyses dendrochronologiques d’échantillons proportionnels de carottes prisd’arbres vivants (Abies lasiocarpa, Tsuga mertensiana) et des échantillons de bois peu fossilisés recueillis dans le secteurdu champ de glace Todd, des chaînons Boundary, de la chaîne Côtière de la Colombie-Britannique, fournissent un fonde-ment pour une reconstruction dendroglaciologique et radiocarbone de l’activité glaciaire à l’Holocène tardif. Cinq interval-les d’expansion glaciaire ont été enregistrés dans des arbres abattus ou enfouis par les glaciers en progression : (1) uneavancée avant environ 3000 14C années AvP; (2) une avancée vers 3000 14C années AvP, laquelle correspond à l’avancéerégionale Tiedemann; (3) une avancée non attribuée à 2300 14C années AvP; (4) une avancée en deux phases vers 1700 etvers 1450 14C années AvP, laquelle correspond à une avancée régionale du Premier Millénaire; (5) une avancée compor-tant trois phases d’expansion qui ont débuté avant environ 750 14C années AvP; cette avancée concorde avec un intervallerégional du Petit Âge glaciaire et un intervalle d’expansion en deux phases vers la fin du Petit Âge glaciaire qui a atteintson maximum après environ 240 et 100 ans AvP. Cette chronologie de la glaciation à l’Holocène tardif concorde avec lesdonnées qui proviennent d’études semblables dans la chaîne Côtière de l’océan Pacifique en Amérique du Nord et quifournissent des bases supplémentaires pour l’importance régionale des avancées Tiedmann et du Premier Millénaire.

[Traduit par la Rédaction] Jackson et al. 98

Introduction

A growing body of evidence suggests that during the Ho-locene the climate of Pacific North America (PNAM) under-went periodic regime shifts (Bond et al. 1997; Cumming etal. 2002) and that these shifts were reflected in glacial massbalance adjustments throughout the western North Americancordillera (Denton and Karlén 1973; Dyurgerov and Meier2000). There is currently considerable interest in recon-

structing the timing and extent of alpine glacier fluctuationsand associated mass balance adjustments throughout the Ho-locene as a means to place ongoing climatic changes into alonger temporal context.

High-resolution records of past climate fluctuations arerequired to detect and assess the long-term impact and sig-nificance of any climate-induced changes on glacial systemsin PNAM. However, quantifying natural climate variabilityin this region remains problematic, due to the scarcity of

Can. J. Earth Sci. 45: 83–98 (2008) doi:10.1139/E07-071 © 2008 NRC Canada

83

Received 24 May 2007. Accepted 29 November 2007. Published on the NRC Research Press Web site at http://cjes.nrc.ca on8 February 2008.

Paper handled by Associate Editor C. Hillaire-Marcel.

S.I. Jackson, S.C. Laxton, and D.J. Smith.1 University of Victoria Tree-Ring Laboratory, Department of Geography, University ofVictoria, Victoria, BC V8W 3P5, Canada.

1Corresponding author (e-mail: [email protected]).

paleoclimatic reconstruction efforts relative to the area’s sizeand the brevity of instrumental climate records (Easterling etal. 1996). Analysis of pre-20th century climate conditions isreliant upon developing detailed proxy climate indicators,with glacier–climate response relationships potentially offer-ing detailed insights into long-term paleoclimatic changes(Larocque and Smith 2005a). When alpine glaciers retreatfrom their maximum downvalley extent, they typically leavebehind terminal moraines, which if they containgeobotanical evidence in the form of detrital wood or lichengrowing on the rocks, can provide dates of maximum glacialextent (Ryder and Thomson 1986; Luckman 1993; Osborn etal. 2001; Larocque and Smith 2003; Lewis and Smith 2004;Osborn et al. 2007). Dating moraines formed when glaciersretreat from advanced positions reached during or after apositive mass balance episode commonly benchmarks the re-sponses of glaciers to local climatic conditions.

An extensive body of work describes Holocene glacialfluctuations in PNAM (Davis 1988; Wiles and Calkin 1990;Luckman et al. 1993), with Little Ice Age (LIA) glacier fluc-tuations particularly well documented (Wiles et al. 1999;Luckman 2000; Larocque and Smith 2003; Wiles et al.2002). In the Coast Mountains of British Columbia four spa-tially synchronous periods of Holocene glacier expansionhave been identified: the Garibaldi advance from 6400 to5000 14C years BP (Ryder and Thomson 1986); theTiedemann advance from 3300 to 2200 14C years BP (Ryderand Thomson 1986); the First Millennium advance (FMA)from 1700 to 1350 14C years BP (Reyes et al. 2006); and theLIA advance from 900 to 100 years BP (Larocque andSmith 2003).

This record of Holocene glacial activity remains incom-plete because at many sites evidence of glacier fluctuationsprior to the LIA was destroyed or buried by subsequent gla-cial activity. Two methods have shown promise in circum-venting this problem in PNAM. Proglacial lake (Clague andMathews 1992; Menounos et al. 2004; Levy et al. 2004;Menounos et al. 2005) and bog or fen sediments (Clagueand Mathewes 1996) often hold records of past glacial epi-sodes and have also been used to reconstruct Holocene gla-cial–climate linkages. In addition, the timing and extent ofHolocene glacier advances has successfully been detailed byradiocarbon and dendroglaciological dating of trees killedand buried by advancing glaciers (Smith and Larocque 1996;Allen and Smith 2007; Koch et al. 2007).

Dendroglaciology is the application of dendrochronologictechniques to reconstruct glacier fluctuations (Schweingruber1988; Smith and Lewis 2007). In some applications, tree-ring series from in situ stumps and boles that were killed bya glacier advance have been cross dated to create floatingtree-ring chronologies. In some instances, the radiocarbonage assigned to perimeter wood is used to establish approxi-mate periods when advancing glaciers killed or damaged thetrees included in a floating chronology (Luckman 1988;Wood and Smith 2004). In other situations, living tree-ringrecords have been used to assign calendar dates to the gla-cier-killed trees (Luckman 1995; Wiles et al. 1996). In anycase, dendroglaciological reconstruction of glacier activityhas the potential to provide insights into past climate fluctu-ations by describing intervals in time when glaciers were ei-

ther advancing or absent from the landscape (Smith andLewis 2007).

This paper presents a dendroglaciological and radiocar-bon-dated reconstruction of Holocene glacier activity in theTodd Icefield area, northwest British Columbia Coast Moun-tains (56°09′N, 129°45′W), based on results of field investi-gations undertaken at sites located 40 km northeast of thetown of Stewart, British Columbia, within the recentlydeglaciated forefields of two valley glacier systems extend-ing from different sectors of the Todd Icefield (Fig. 1).Reconnaissance investigations at both sites in July 2003 re-vealed fresh exposures of in situ and detrital subfossil woodat several locations. Further fieldwork in the area in July2004 completed the dendroglaciological surveys in Todd andSurprise valleys.

Previous researchOnly a limited amount of research focusing on the recon-

struction of Holocene glacier behaviour has been completedwithin the Boundary Ranges region of northwestern BritishColumbia. Ryder (1987) documented the Neoglacial historyof glaciers in the Stikine–Iskut area north of Todd Icefield,concluding that increased snowfall about 4000 14C years BPresulted in the prolonged expansion of perennial snowfieldsand glaciers. Historical ice-front recession had exposed gla-cially overridden trees and soil that indicated that glaciers inthe region were growing in size and advancing downvalleyaround 600–500 14C years BP.

Clague and Mathews (1992) and Clague and Mathewes(1996) examined proglacial lake and fen sediments in the vi-cinity of Berendon and Frank Mackie glaciers west of theTodd Icefield area. They identified stratigraphic,paleoecologic, and geochronologic evidence for aTiedemann-age glacier advance around 2800 14C years BP, aFMA-age glacier advance around 1300 14C years BP, and aLIA glacier advance that began prior to 615 14C years BP(Clague et al. 2004).

Haspel et al. (2005) and Spooner et al. (2005) reported ontheir investigations of a moraine built by the expansion ofBear River Glacier south of the Todd Icefield (Fig. 1). BearRiver Glacier spills northward from the Cambria Icefield toterminate in a basin now filled by Strohn Lake. Haspel et al.(2005) and Spooner et al. (2005) describe five stacked tillswithin a lateral moraine separated by four relatively continu-ous wood mats. Wood samples from the three lower matswere interpreted to belong to separate Tiedemann-age gla-cier advances around 3700, 3450, and 3300 14C years BP.An upper wood mat yielded a date of 1040 14C years BP andwas interpreted to coincide with the earliest LIA advance ofthe Bear River Glacier.

Study sites

The Todd Icefield area is characterized by high topo-graphical relief, extensive high elevation icefields, and valleyglaciers. The Wisconsinan glaciation resulted in the forma-tion of hanging valleys, steep-sided valley walls, and ice-carved trunk valleys (Hickin et al. 2001). The bedrock geol-ogy of the area consists primarily of epiclastic volcanic

© 2008 NRC Canada

84 Can. J. Earth Sci. Vol. 45, 2008

rocks and lithic tuffs from the Lower Jurassic Unuk Forma-tion (Grove 1971).

The Todd Icefield area experiences a coastal maritime cli-mate. The annual air temperature at Stewart (55°56′N,129°59′W; 7 m above sea level (asl)) averages 6 °C, with thecoldest month averaging –3.7 °C and the warmest 15.1 °C(Environment Canada 2003, Canadian Climate Normals1971–2000). The total annual precipitation at Stewart aver-ages 1843 mm, with more than 30% of this falling as snow.

The dominant forest cover in the Todd Icefield area con-sists of mixed stands of mountain hemlock (Tsugamertensiana [Bong.] Carr.) and subalpine fir (Abieslasiocarpa [Hook.] Nutt.) (Meidinger and Pojar 1991). Theupper treeline is found at about 1500 m asl.

Todd valleyTodd Glacier is the largest of several ice tongues that spill

into the headwaters of north-facing Todd valley (unofficialname) from the Todd Icefield (56°13′N, 129°46′W; Fig. 2).Until the 1950s, five glaciers coalesced and advanced to aterminal position around 2.5 km from the present snout of

Todd Glacier at 1000 m asl. North-facing Todd Glacier isthe largest of these glaciers and is the primary source of wa-ter for north-flowing Todd Creek (Fig. 3). East of the pres-ent terminus of Todd Glacier is west-facing Sage glacier(unofficial name) with a steep bedrock slope at its terminus.East-facing Two glacier (unofficial name) is located withinthe headwaters of Two glacier valley (unofficial name) andis the source of a small creek that drains into a proglaciallake located 700 m downvalley from the 2004 terminus posi-tion of Todd Glacier. Bug glacier (unofficial name) is asteeply sloping, east-facing, cirque glacier and is the north-ernmost glacier examined within the valley.

Within the last 25 years, all of the glaciers in Todd valleyhave retreated and downwasted significantly. Recent termi-nal retreat rates estimated from aerial photographs (1974–1997) range between 9 and 76 m/year (Table 1). The rapidretreat of glaciers in Todd valley and the periodic drainageof glacially ponded lakes have exposed a fluted forefieldoverlain by a series of small recessional moraines (Fig. 4).The forefield is deeply incised by active and abandonedmeltwater stream channels. The valley floor is flanked by

© 2008 NRC Canada

Jackson et al. 85

Fig. 1. Location of Todd Icefield and glaciers. 1, Todd Glacier; 2, Surprise glacier; 3, Bear River Glacier.

deeply eroded lateral moraine complexes that extend up thevalley walls to the LIA trimline at 1200 m asl in the vicinityof Todd Glacier. Buried in situ stumps and detrital boles areexposed and visible at many locations throughout theforefield (Fig. 5).

Surprise valleySurprise glacier (unofficial name, 56°12′N, 131°50′W) is

located in a headwater tributary of Surprise Creek that flowseastward into nearby Meziadin Lake (Fig. 6). The glacieroriginates at around 1900 m asl from icefalls on the ToddIcefield that spill east into a narrow, steeply sloping anddeeply incised bedrock valley. Historical aerial photographs

and our field surveys indicate that from 1986 to 2004 theterminus of Surprise glacier retreated at a rate of around25 m/year (Table 1). In 2004, the debris-covered snout ofSurprise glacier was located at around 900 m asl, 3 kmupvalley from a bouldery terminal moraine complex heavilycolonized by slide alder (Alnus sinuata [Regel] Rydb.) andyoung subalpine fir trees.

Surprise glacier is characterized by prominent lateral mo-raines that reach up to 150 m in height above the valley floor(Fig. 7). The exposed proximal slope of the north-facing lat-eral moraine is composed of multiple till horizons separatedby contiguous layers of detrital boles and wood fragments(Fig. 8).

© 2008 NRC Canada


Fig. 2. Site map of Todd valley showing the location of glaciers and study sites.

Research methods

DatingRadiocarbon and dendroglaciological dating were em-

ployed to establish the late Holocene history of glacial activ-ity in Todd and Surprise valleys. The overall aim was tocollect subfossil tree-ring evidence that could be used to as-sign dates to the specific glacial advances that killed andburied each tree (Luckman 1995, 1998). In the first instance,attempts were made to cross date the outermost rings of thesubfossil samples to living tree-ring chronologies to provideminimum calendar dates (A.D.) for the different glacier ad-vances (Smith and Larocque 1996; Smith and Lewis 2007).Where cross dating proved unsuccessful, perimeter woodsamples containing multiple annual growth rings were re-moved and submitted for conventional radiometric analysesby Beta Analytic Inc. (Miami, Florida, USA). These radio-carbon ages are presented in the text as 14C years BP and

© 2008 NRC Canada

Jackson et al. 87

Fig. 3. Todd Glacier with the Todd Icefield shown along the skyline, July 2004.

Glacier Date (aerial photography no.)Retreat rate(m/year)

Bug 1974 (BC5604-075) 91997 (30BCC97143-36)

Sage 1974 (BC5604-075) 481997 (30BCC97143-36)

Surprise 1982 (BC82018) 251997 (BC97143-046)2004 (this survey)

Todd 1974 (BC5604-075) 761997 (30BCC97143-36)

Two 1974 (BC5604-075) 351997 (30BCC97143-36)

Table 1. Estimated historical terminus retreat rates in the ToddIcefield area.

Fig. 4. Oblique aerial photograph of the Todd Glacier forefieldin July 2003. Shown in the foreground is an abandoned meltwa-ter channel incised into fluted till sediments.

Fig. 5. A buried forest found west of Todd Glacier at 1190 masl (Site 7, Fig. 2). Located above the present-day forest cover inTodd valley, the trees at site 7 were overrun and buried duringthe late Little Ice Age.

were used to assign relative 14C ages to the floating tree-ringchronologies by identifying the midpoint radiocarbon age asthe minimum kill date (Wood and Smith 2004).

Wood sampling and treatmentWood samples were collected from subfossil stumps,

boles, and detrital wood fragments. After air drying the sam-ples were sanded to a fine polish at the University of Victo-ria Tree-Ring Laboratory. The species of each subfossilwood sample was identified using a 40× microscope and astandard reference key (Hoadley 1990). Mature mountainhemlock and subalpine fir trees found growing in a foreststand adjacent to Surprise glacier (56°12′21′ ′N,129°35′13′ ′W; 955 m asl) were sampled with an incrementcorer to establish living tree-ring chronologies.

The total ring width of the annual growth rings containedwithin each subfossil or increment core sample was mea-sured using a WinDENDRO™ tree-ring image processingsystem (Guay et al. 1992). The cores were scanned and thering widths were measured to the nearest 0.01 mm. Wherethe ring boundaries were difficult to identify, a 40× micro-scope and Velmex-type stage measurement system wereused to verify the ring boundary positions.

Cross dating of both sets of samples was accomplishedwith reference to common pointer years (narrow tree rings)(Schweingruber 1988), and the accuracy of the cross datingwas checked using the COFECHA software program (Gris-sino-Mayer 2001). Series transformations were completedusing a cubic smoothing spline and the default parameters

© 2008 NRC Canada


Fig. 6. Site map showing 1997 position of Surprise glacier and sampling location.

(50% wavelength cutoff at 32 years, examined in 50-yearsegments lagged successively by 25 years).

Standardized living tree-ring chronologies were developedusing the software program ARSTAN (Cook and Kairiukstis1990). Initially a negative exponential curve was used to re-move any age-related growth effects, and then a cubicsmoothing spline was invoked to reduce the effects of varia-tion within the series. The resulting indices were subse-quently compared using a simple Pearson’s r correlation todetermine whether both tree species had responded similarlyover the length of record.

Attempts were first made to cross date the subfossil sam-ples to the living tree-ring chronologies. For samples whereno relationship to a living chronology could be established,an effort was then made to develop internally consistentfloating chronologies with radiocarbon age control bothwithin and between the study sites.

Observations and results

Living tree-ring chronologiesSamples consisted of 58 cores measured from 42 trees,

with 35 cores used to construct the mountain hemlockchronology, and 23 cores for the subalpine fir chronology(Table 2). The mountain hemlock chronology spans the in-terval from 1599 to 2004 A.D. (405 years) and the subalpinefir series from 1649 to 2004 A.D. (355 years) (Fig. 9). Themean series correlation values of both chronologies con-tained a collective signal significant at the 99% confidenceinterval when using 50-year chronology segments. Themountain hemlock master chronology had a series correla-tion value of 0.526, and the subalpine fir master series corre-lation was 0.562 (Table 2). There is a robust correlationbetween the two chronologies, particularly as mature trees inthe 1790–2004 period when there is a significant correspon-dence in their annual radial growth trends (r = 0.523) andexpressed population signals (>0.85 for six or more trees;Wigley et al. 1984). These descriptive statistics are compa-rable to those developed at other high elevation tree-ringsites in the British Columbia Coast Mountains (Larocqueand Smith 2005b) and indicate that both master chronologieswere well suited for cross dating (Smith and Larocque1996).

Study site investigations

Todd valleySubfossil wood samples were collected at eight sites, five

in the Todd valley and three in the Two glacier valley(Fig. 2). Five sites were located in the main trunk valley(sites 1, 2, 3, 4, and 8) and three sites in the adjacent east–west-trending Two glacier valley (sites 5, 6, and 7).

Site 1Prior to 1974, Bug glacier merged with Todd Glacier at a

point �2 km north of the 2004 terminus of Todd Glacier.Close to the point where they coalesced, a bole fragment(TG04-801) was found wedged within a diamict-filled depres-sion. Containing 237 annual growth rings, the tree was killedaround 450 ± 60 14C years BP (1400–1630 A.D.; Table 3).

The northern flank of the Bug glacier forefield contains anested sequence of three distinct lateral moraines. Distal spill-age from the outermost moraine decapitated a small tree andleft a remnant in situ tree stump (TG04-802, 63 annual rings)beneath a large boulder. Radiocarbon dating of the outermost18 rings of the stump indicates that the tree was killed about160 ± 60 14C years BP (1640–1950 A.D.; Table 3).

Site 2Numerous boles, detrital wood fragments, and stumps

were found exposed 75 m above the valley floor withineroded sandy lateral moraine sediments at several locationsaround the eastern wall of Todd valley. Four samples werecollected 150–300 m downvalley from the 2004 terminus ofSage glacier (TG04-868 to TG04-871; Table 4). Perimeterwood from TG04-871 (117 rings) dates to 660 ± 60 14Cyears BP (1250–1420 A.D.; Table 3). As TG04-871 crossdates (r = 0.424) to TG04-870 (80 rings) from an adjacentlocation, a contemporaneous origin for the wood at this loca-tion is suggested (Table 4).

Four rooted boles (TG04-864 to TG04-867; Table 4) werelocated in growth position 400 m north of the 2004 terminusof Sage glacier. Specimens TG04-866 (110 rings) andTG04-867 (94 rings) cross date (r = 0.424) with TG04-871,

© 2008 NRC Canada

Jackson et al. 89

Fig. 7. View of north-facing lateral moraine at Surprise glacier.

Fig. 8. View looking vertically up the face of the lateral moraineat Surprise glacier. Shown are in situ boles dating to theTiedemann advance (A) at �3000 14C BP and First Millenniumadvance (B) at �1700 14C BP.

confirming that this grove of trees was also killed at660 ± 60 14C years BP (Table 4).

Site 3The valley bottom sediments exposed by the retreat of

Todd Glacier are composed of a deeply fluted till incised bydrainage from Two glacier valley. Site 3 is located within thenorthernmost of these abandoned channels. Exposed 6 mabove the channel base is a 10 cm thick horizon of needles,small twigs, and litter overlain by 3 m of till. An in situ bolefragment (TG04-838; 89 rings) from this horizon dates to660 ± 60 14C years BP (1206–1410 A.D.; Table 3). A seg-ment of detrital wood found partially buried in the aban-doned channel (TG04-836) successfully cross dated into afloating chronology anchored by the radiocarbon age as-signed to TG04-838 (T1-4; Table 3). Based upon these find-ings, it appears that Todd Glacier advanced into a stand ofyoung trees (89–106 years old; Table 4) at this site at660 ± 60 14C years BP.

Site 4Site 4 is located 200 m north of the 2004 terminus of

Todd Glacier in the centre of Todd valley (Fig. 2). Sequen-tial aerial photographs from 1974–1997 indicate that thisarea was ice covered until 1997, after which a meltwaterstream draining Two glacier valley incised through 5 m ofvalley bottom sediments. This channel has aggraded andchanged course, leaving behind scores of bole and detritalwood fragments scattered along the abandoned channelfloor. Disk samples were collected from a number of thesepieces of wood, and seven were incorporated into a floatingchronology anchored to the radiocarbon age assigned tosample TG04-838 (TI-3; Table 4).

The abandoned meltwater channel at site 4 is distin-guished by a 75 m long east-facing cutbank exposing three

stratigraphic units. The lowermost unit consists of a 4.5 mthick matrix-supported till without distinct bedding struc-tures, organics, grading, or fabric. Overlying this unit is alaterally extensive 30 cm thick organic horizon composed ofneedle, twig, branch, and root macrofossils. Large boles,some with root balls intact, protrude from the exposure(Fig. 10). The east–west orientation of the boles is tangentialto the flow of Todd Glacier and suggests the trees were bur-ied by an advance of Sage glacier. Overlying the woody matis a 2.4 m thick organic-free matrix-supported till capped bynorth–south trending flutes.

Ten wood samples (92 to 213 rings) were collected fromboles found within the woody mat at site 4. All of the sam-ples have distorted ring sequences due to deformation fromthe weight of the overlying glacier and could not be crossdated. A radiocarbon date for TG03-806 (101 rings) indi-cates that the trees at this site were killed and buried2300 ± 60 14C years BP (420–200 B.C.; Table 3).

Site 5A large mixed assemblage of boles, stumps, and woody

debris was located on the south-facing wall of Two glaciervalley (Figs. 2 and 11). Radial samples show that these weremature trees when they were overrun (110–234 rings; Ta-ble 4). Although radiocarbon dating of perimeter rings fromTG04-879 suggests these trees were killed 360 ± 60 14Cyears BP (Table 3), four samples cross dated into the livingsubalpine fir chronology (Table 4). This finding indicatesthat Two glacier was growing in size, advancing downvalley,and killing mature trees growing on the valley walls in1898–1899 A.D.

Site 6Site 6 encompasses a broad area of the valley floor of

Two glacier valley, close to where this tributary valley joins

© 2008 NRC Canada


Species No. trees No. in series Span (years A.D.) Total length (years) Series intercorrelation Mean sensitivity

Abies lasiocarpa 19 23 1649–2004 355 0.562 0.15Tsuga mertensiana 23 35 1599–2004 405 0.526 0.14

Table 2. Characteristics of chronologies developed from living trees at Surprise glacier.

Fig. 9. Living tree-ring mountain hemlock (Tsuga mertensiana) and subalpine fir (Abies lasiocarpa) chronologies from Surprise glacier.Bold lines show unweighted 10 year running mean.

Todd valley (Fig. 2). As Two glacier retreated back up thevalley, it released meltwater that eroded through basal tillsand left a broad sheet of distal outwash deposits mantledwith detrital boles and stumps. Samples from seven of thesedetrital wood pieces (90 to 175 rings) cross dated into thefloating chronology anchored by the radiocarbon age as-signed to TG04-838 (TI-3; Table 4). Although surface mas-tication has removed peripheral growth rings from somesamples (TG04-852, TG04-859), those retaining narrow per-imeter rings (TG04-847, TG04-849, TG04-850) indicate thatthese trees were killed by around 660 ± 60 14C years BP(1362–1401 A.D.; Table 4)

The upper section of Two glacier valley is distinguished byisolated till islands separated by abandoned meltwater chan-nels. A bullet-shaped stump (TG03-815; 165 rings) was foundwithin one of these till units and yielded an age of1540 ± 60 14C years BP (405–640 A.D.; Table 3). Samples re-covered from two detrital stumps (TG04-846, TG04-856) lo-cated downstream on the outwash plain cross dated to thisstump (TI-2; Table 4), confirming that Two glacier was advanc-ing into a stand of mature trees at this time (TI-2; Table 4).

Site 7Site 7 is located on a prominent bedrock ridge 400 m

northwest and 200 m above the terminus of Todd Glacier(Fig. 4). Several broken and fragmented large bole segmentswere found scattered across the surface of a veneer of till onthe up-glacier facing slope. A perimeter sample from a bolefragment on the bedrock ridge (TG04-874, 106 rings)yielded a date of 1690 ± 60 14C years BP (230–520 A.D.;Table 3). This bole cross dated into a floating chronologyanchored by two additional radiocarbon dates, suggestingthat Todd Glacier was advancing downvalley at the sametime as Two glacier (TI-2; Table 4).

A deeply incised gully extends from the ridge crest to thevalley floor on the distal slope of the bedrock promontory.Exposed within an upper section of the gully wall is a thicksection of till containing a laterally contiguous horizon oflarge boles with attached roots, detrital wood fragments, andbranches with attached needles (Fig. 5). A radiocarbon agefrom the perimeter rings of TG04-872 indicates that thesetrees were buried around 410 ± 60 14C years BP (1410–1650A.D.; Table 3) as Todd Glacier advanced over the crest ofthe bedrock ridge.

Site 8The partially collapsed remnants of a large lateral moraine

are plastered against the western wall of Todd valley be-tween Two glacier valley and Bug glacier. Numerous piecesof detrital wood were visible along a discontinuous wood-rich horizon �60 m above the valley floor. A sample fromthe pith area of a single bole fragment (TG04-805, 103rings) yielded an age of 730 ± 60 14C years BP (1200–1390A.D.; Table 3) and cross dated into the floating chronologyconstructed from samples collected on the opposing trunkvalley wall (TI-4; Table 4).

Surprise glacierDendroglaciological investigations at Surprise glacier

were restricted to surveys of the prominent north-facing lat-eral moraine (Fig. 7). The lateral moraine is composed ofmassive units of till of variable character but dominated by

poorly sorted sands and angular boulders up to 8 m in diam-eter. Six distinct woody horizons separated by thick till unitswere identified. The wood-dominated horizons were later-ally contiguous and emphasized by an ochre-coloured stainoriginating from an up-glacier gossan deposit.

Sampling at this site was limited to cutting disks salvagedfrom wood detritus found on the basal talus slope and fromboles or stumps found within the open face of the moraine(Fig. 8). In the case of the latter location, due to the height,severe angle, and unconsolidated nature of the lateral mo-raine, the majority of samples were collected along transectsthat extended vertically down the moraine face. Sampleswere collected either by rappelling with climbing gear fromthe moraine crest or by free climbing vertical gullies etchedinto the moraine face from the talus slope below.

The lowest woody horizon found within the lateral morainewas located 125 m below the moraine crest (F in Fig. 12).This horizon is characterized by lengthy bole segments (mini-mum 2–3 m) with attached broken root caps projecting fromthe moraine face. The boles appear to dip into the moraineface suggesting they were buried by debris spilling down thedistal face of an antecedent lateral moraine. Access to this ho-rizon was limited and only a single sample was collected(SG03-801, 175 rings), dating to 2960 ± 70 14C years BP(1395–975 B.C.; Table 3). Samples retrieved from three detri-tal boles found on the underlying talus slope are assumed tohave fallen from the moraine face and provide additional sup-port for burial around 3000 14C years BP (TI-1; Table 4).

Two laterally continuous woody horizons were located ca.50 and 70 m above the lowermost wood mat, and could betraced across the breadth of the moraine face (D and E,Fig. 12). Safety concerns restricted sampling to one bolefragment (SG04-828, 153 rings) exposed within a gully facenear the distal extent of the moraine (75 m below the mo-raine crest; E, Fig. 12) that yielded an age of 1690 ± 60 14Cyears BP (230–520 A.D.; Table 3). Four detrital samplesfrom various locations on the talus slope cross dated withSG04-828 to substantiate the burial of trees around1700 years BP (TI-2; Table 4). A branch fragment recovered20 m above SG04-828 gave an age of 1440 ± 60 14C yearsBP (SG04-29, 530–680 A.D.; Table 3). The age and positionof SG04-829 suggest it could either have been buried by asucceeding glacial advance or by the continued accumula-tion of distally spilled sediments following the earlier burialof SG04-828.

The uppermost section of the Surprise glacier lateral mo-raine contains three distinct woody horizons (A–C, Fig. 12).Three buried boles (SG04-804, SG04-807, SG04-826), re-trieved from a wood-rich mat located in the lowermost ofthese three horizons (25–30 m below the moraine crest),cross dated to form a floating chronology (TI-5; Table 4).One sample retained bark (SG04-804, 219 rings) and datesto 610 ± 50 14C years BP (1290–1420 A.D.; Table 3).

Evidence for a successive burial event comes from anoverlying wood mat that contains a number of boles andstumps positioned 12–15 m below the moraine crest; threesamples from this horizon cross dated with the living moun-tain hemlock tree-ring chronology (Table 4). Sample SG04-801, recovered 12 m below the moraine crest, yielded a per-imeter date of 1746 A.D. and SG04-805, located 10 m (B,Fig. 12) below the moraine crest, gave a perimeter date of

© 2008 NRC Canada

Jackson et al. 91

1764 A.D. As bark was not found on any of these samples,the assigned perimeter ages are considered to be minimumlimiting dates for the time of their burial.

A sample collected from a continuous wood-rich horizonlocated 3 m below the moraine crest (A, Fig. 12) cross datedwith the living subalpine fir tree-ring chronology, indicatingthat it was buried in 1843 A.D. A similar kill-date of 1848A.D. was assigned to a detrital wood fragment found on thetalus slope surface (Table 4).

The stratigraphy and layers of woody debris exposed inthe Surprise glacier lateral moraine indicate that sedimenthas been progressively accumulating at this location since

prior to �3000 14C years BP. Recurring late Holocene gla-cier advances have repeatedly killed trees growing on themoraine, in most instances by spilling debris down the distalslope of the moraine and burying the trees. Based on the gla-cial record reported from the site, the most recent LIA ad-vance at Surprise glacier was initiated prior to 1843 A.D.,when the glacier last buried trees at the site.

Synthesis and regional correlation

Investigations at Todd and Surprise glaciers have allowedthe reconstruction of late Holocene glacier activity in theTodd Icefield area (Fig. 13). Five Holocene glacial events

© 2008 NRC Canada


Sample(Beta ID) Speciesa Location

14C age years BP (no.perimeter rings dated)

2σ Calibrationdateb (A.D.)

2σ Calibrationyears BPb

Minimum treeage (years) Latitude (N)

Bear River GlacierBR03-801(181857)

saf Moraine 3680 ± 60(25)

2210–1900 B.C. 4160–3850 119 56°06′19′′

BG803-806(185808)

saf Below moraine 3340 ± 60 1750–1500 B.C. 3700–3450 64 56°06′19′′

Surprise glacierSG03-801(181858)

mh Lateral moraine 2960 ± 70(42)

1395–975 B.C. 3345–2925 175 56°11′54′′

SG04-828(197984)


230–460

480–520

1720–1480

1470–1430

153 56°11′54′′

SG04-829(197986)

mh Lateral moraine 1440 ± 60(NA)

530–680 1420–1270 56°11′54′′

SG04-804(197983)


1290–1420 660–530 219 56°11′54′′

Todd valleyTG03-806(181859)

saf Site 4 2300 ± 60(47)

420–200 B.C. 2370–2150 101 56°12′14′′

TG04-874(199708)

saf Site 7 1690 ± 60(23)c

230–460480–520

1720–14801470–1430

106 56°11′44′′

TG03-815(181560)

saf Site 6 1540 ± 60(68)

405–640 1545–1310 165 56°12′01′′

TG04-805(197989)

saf Site 8 730 ± 60(14)c

1200–13201350–1390

750–630600–540

103 56°12′37′′

TG04-838(197990)

saf Site 3 660 ± 60(17)

1260–1410 690–540 89 56°12′33′′

TG04-871(197991)

saf Site 2 660 ± 60(14)

1250–1420 700–530 117 56°12′41′′

TG04-801(197986)

mh Site 1 450 ± 60(85)

1400-15201580–1630

550–430380–320

>237 56°13′21′′

TG04-872(208442)

saf Site 7 410 ± 60 1410–1530

1550–1660

540–420

400–320

185 56°12′07′′

TG04-879(208443)

saf Site 5 360 ± 60 1420–1650 530–300 233 56°11′49′′

TG04-802(197987)

saf Site 1 160 ± 60(18)

1640–1950 310–0 63 56°13′31′′

asaf, subalpine fir; mh, mountain hemlock.bAll radiocarbon dates and 2σ ranges calculated by Beta Analytic Inc. using INTCAL98 (Stuiver and van der Plicht 1998; Stuiver et al. 1998).cPith rings dated.

Table 3. Summary of radiocarbon dated samples in the Todd Icefield area.

were identified that appear to be broadly synchronous withpreviously recognized advances.

Pre-Tiedemann advance (>3000 14C years BP)Exposed below the lowest Tiedemann-age (ca. 3000 14C

years BP) woody horizon, discovered within the Surpriseglacier lateral moraine, is a structureless unit of till over20 m thick. Although the base of this unit is mantled by ta-lus and debris eroded from the overlying moraine face, thephysical setting indicates that upwards of 50 m of additionalsection may be buried at the site.

The timing of the glacier advances that deposited thebasal tills at Surprise glacier is unknown. The basal till sedi-

ments could be associated with the regional Garibaldi phaseof glacier expansion between 6000 and 5000 14C years BP(Ryder and Thomson 1986; Calkin 1988) or with the re-ported advance of glaciers in the southern British ColumbiaCoast Mountains around 8200 14C years BP (Menounos etal. 2004). They may be associated with an earlier phase ofthe Tiedemann advance as evidenced by samples collected atBear River Glacier during the course of our investigations,dating to 3680 ± 60 and 3340 ± 60 14C years BP (2210–1900 and 1750–1500 A.D.; Table 3). Dates from multiplestacked woody mats in the moraine at Bear River Glacier re-ported by Haspel et al. (2005) and Spooner et al. (2005) sug-gest that several advance and retreat phases dating to around3700, 3450, and 3300 14C years BP may characterize theearly portion of the Tiedemann advance in the Todd Icefieldarea.

Tiedemann advance (ca. 3000 14C years BP)First documented at Tiedemann and Gilbert glaciers in the

southern British Columbia Coast Ranges (Ryder andThomson 1986), the Tiedemann advance is considered coevalwith the Peyto advance in the southern Canadian RockyMountains (Luckman et al. 1993; Osborn et al. 2001; Woodand Smith 2004). Evidence for a Tiedemann-age advance ofSurprise glacier around �3000 14C years BP supports the find-ings of Clague and Mathewes (1996) at nearby Frank MackieGlacier, as well as the reports of contemporaneous glacier ad-vances at Bear River (Haspel et al. 2005; Spooner et al. 2005)and Forest Kerr glaciers (Lewis and Smith 2005). Our find-ings confirm the local synchroneity of Tiedemann-age glacierevents and further emphasize the likely significance of thisevent throughout the Coast Mountains of British Columbia(Desloges and Ryder 1990; Reyes and Clague 2004; Allenand Smith 2007; Koch et al. 2007).

Unattributed advance (ca. 2300 14C years BP)The discovery of evidence for an advance of Sage glacier

into an established valley-bottom forest at 2300 ± 60 14Cyears BP is synchronous with a period of glacial expansionat Berendon Glacier, reported by Clague and Mathewes(1996). Corroborating evidence for glacial advances in theBritish Columbia Coast Mountains at this time was reportedby Ryder and Thomson (1986) from sites in the Mt.Waddington area, by Reyes and Clague (2004) and Allenand Smith (2007) at glaciers in the Lillooet Icefield area,and by Koch et al. (2007) in Garibaldi Provincial Park. In-creasing recognition of a regional period of glacial expan-sion between �2300 and 2000 14C years BP may distinguisha later Tiedemann phase or it may represent an unrelatedmass balance response to the cool, moist conditions that arereported to have characterized the Coast Ranges at this time(Pellat et al. 2000; Spooner et al. 2003; Lamoureux andCockburn 2005).

First Millennium advance (ca. 1690–1440 14C years BP)Evidence for FMA-age advances of glaciers in the Todd

Icefield area comes from two sites in Todd valley and fromthe lateral moraine at Surprise glacier. In Todd valley thecross-dated radiocarbon samples from sites 6 and 7(1540 ± 60 and 1690 ± 60 14C years BP) identify contempo-raneous kill dates related to the advance of both Two glacier

© 2008 NRC Canada

Jackson et al. 93

Longitude (W)Elev.(m asl) Description

129°38′48′′ 435 Log in moraine from lower woodymat 10 m above road

129°38′48′′ 435 Log eroded from moraine face andlocated in roadside gully

129°35′39′′ 800 In situ log 125 m below moraine crest


129°35′39′′ 870 In situ branch fragment 55 m belowmoraine crest


129°46′29′′ 930 Log from unit 2 in cutbank exposurein front of Todd Glacier

129°46′38′′ 1100 Log on surface of bedrock outcropwest of Todd Glacier

129°46′29′′ 1010 In situ bullet-shaped stump from tillisland on Two glacier forefield

129°46′47′′ 960 In situ bole in lateral moraine westside of Todd valley

129°46′27′′ 925 In situ bole in woody mat 6 m abovechannel west side of Todd valley

129°45′47′′ 950 In situ bole in lateral moraine sedi-ments east side of Todd valley

129°46′27′′ 975 In situ bole on bedrock outcrop belowBug glacier

129°47′11′′ 990 Glacier pushed detrital pile of stumpsand boles behind bedrock outcrop

129°46′40 ′′ 1090 In situ stump with attached bole ingully-side wood mat

129°46′41′′ 1046 Stump buried by distal spillage fromlateral moraine of Bug glacier

© 2008 NRC Canada


Location(site name)

Site No.or type Sample No.

14C ageyears BPa No rings

Correlationwith master

Perimeterageb Description

(A) Cross dated to 14C anchored floating tree-ring chronologies

Floating chronology TI-1Surprise Moraine SG03-803 193 0.645 1228 B.C. Colluvial slope below SG03-801Surprise Moraine SG04-810 144 0.511 1210 B.C. Colluvial slope below SG03-801Surprise Moraine SG04-816 143 0.470 1177 B.C. Colluvial slope below SG03-801Surprise Moraine SG03-801 2960 ± 70 175 0.424 1146 B.C. In lateral moraine 125 m below crest

Floating chronology TI-2Surprise Moraine TG03-804 109 0.348 307 A.D. Detrital in meltwater channelSurprise Moraine SG04-817 124 0.447 320 A.D. Colluvial slope below SG03-801Surprise Moraine SG04-823 200 0.466 333 A.D. Colluvial slope below SG03-801Todd 6 TG04-846 66 0.553 345 A.D. Detrital in meltwater channelSurprise Moraine SG04-828 1690 ± 60 153 0.551 345 A.D. In lateral moraine 75 m below crestSurprise Moraine SG04-834 99 0.398 352 A.D. Colluvial slope below SG03-801Todd 7 TG04-874 1690 ± 60 106 0.500 373 A.D. Detrital on till surfaceTodd 6 TG04-856 134 0.525 390 A.D. Detrital in meltwater channelTodd 6 TG03-815 1540 ± 60 165 0.611 393 A.D. In situ in till island

Floating chronology TI-3Todd 6 TG04-852 98 0.441 1203 A.D. Detrital in meltwater channelTodd 4 TG04-813 178 0.440 1276 A.D. Detrital in meltwater channelTodd 4 TG04-814 213 0.448 1313 A.D. Detrital in meltwater channelTodd 6 TG04-859 125 0.465 1320 A.D. Detrital in meltwater channelTodd 4 TG04-816 134 0.569 1321 A.D. Detrital in meltwater channelTodd 4 TG04-827 178 0.388 1321 A.D. Detrital in meltwater channelTodd 6 TG04-854 175 0.581 1321 A.D. Detrital in meltwater channelTodd 4 TG04-811 179 0.402 1322 A.D. Detrital in meltwater channelTodd 4 TG04-822 145 0.343 1337 A.D. Detrital in meltwater channelTodd 4 TG04-817 65 0.405 1337 A.D. Detrital in meltwater channelTodd 3 TG04-836 103 0.391 1339 A.D. Detrital in meltwater channelTodd 6 TG04-844 90 0.372 1359 A.D. Detrital in meltwater channelTodd 6 TG04-850 100 0.356 1362 A.D. Detrital in meltwater channelTodd 3 TG04-838 660 ± 60 89 0.358 1366 A.D. Buried between two till unitsTodd 6 TG04-847 126 0.457 1390 A.D. Detrital in meltwater channelTodd 6 TG04-849 102 0.431 1401 A.D. Detrital in meltwater channel

Floating chronology TI-4Todd 8 TG04-805 730 ± 60 103 0.495 1295 A.D. In-situ bole within west lateral moraineTodd 2 TG04-871 660 ± 60 117 0.482 1366 A.D. Buried in moraine sedimentTodd 2 TG04-867 94 0.329 1419 A.D. Rooted in lateral moraine sedimentsTodd 2 TG04-870 80 0.455 1422 A.D. Detrital sample from sediment surfaceTodd 2 TG04-866 110 0.384 1429 A.D. Rooted in lateral moraine sediments

Floating chronology TI-5Surprise Moraine SG04-807 70 0.636 1334 A.D. In lateral moraine 12 m below crestSurprise Moraine SG04-826 297 0.389 1384 A.D. In lateral moraine 30 m below crestSurprise Moraine SG04-804 610 ± 60 219 0.602 1415 A.D. In lateral moraine 25 m below crest

(B) Cross dated to living tree-ring chronologies

Surprise Moraine SG04-801 316 0.385 1746 A.D. In lateral moraine 12 m below crestSurprise Moraine SG04-805 177 0.531 1764 A.D. In lateral moraine 10 m below crestSurprise Moraine SG04-803 193 0.398 1843 A.D. In lateral moraine 3 m below crestSurprise Moraine SG04-811 197 0.396 1848 A.D. Colluvial slope below SG03-801Todd 5 TG04-884 234 0.485 1898 A.D. Glacier pushed detrital pileTodd 5 TG04-879 (360 ± 60)c 233 0.433 1898 A.D. Glacier pushed detrital pileTodd 5 TG04-880 110 0.358 1898 A.D. Glacier pushed detrital pileTodd 5 TG04-882 161 0.477 1899 A.D. Glacier pushed detrital pile

aAll radiocarbon dates calculated by Beta Analytic Inc. using INTCAL98 (Stuiver and van der Plicht 1998; Stuiver et al. 1998).bCalendar age is an estimate assigned averaging the end members of the 2σ calibration date.cRadiocarbon date conflicts with cross-dated age.

Table 4. Tree-ring chronology cross dates.

and Todd Glacier. A corresponding expansion of Surpriseglacier during the same period (1690 ± 60 14C years BP) sug-gests that glaciers in the Todd Icefield area were respondingto a regional climate forcing that led to a prolonged periodof positive mass balance. Geochemical and biological prox-ies in marine and lacustrine sediments reveal pronounced re-gional cooling and changes in atmosphere–ocean circulationpatterns suggesting that First Millennium glacier advancesmay best be explained by decreased ablation, rather than byincreased accumulation (Hu et al. 2001; Moberg et al. 2005).

At Surprise glacier, two distinct FMA phases are high-lighted suggesting that the regional FMA may embrace morethan one episode of glacial expansion. An advance subse-quent to that at 1690 ± 60 14C years BP is recorded by thediscovery of a laterally contiguous woody mat located abovethe older moraine material dating to 1440 ± 60 14C years BP.

Little Ice Age (ca. 730 14C years BP to present)LIA glacier advances appear to have been initiated by 883

to 1156 A.D. in the Todd Icefield area (Haspel et al. 2005;Spooner et al. 2005). Radiocarbon-controlled dates on buriedwoody debris from three sites in Todd valley and cross dat-ing of additional wood detritus indicate that by 730–660 14C

years BP Two glacier and Sage glacier had coalesced withTodd Glacier to fill the main trunk valley. At Surprise gla-cier, buried boles at high elevation within the lateral moraineconfirm that by 610 14C years BP the glacier had filled thevalley and was burying �300-year-old trees growing on thedistal slopes of the pre-existing southern lateral moraine. Al-though the duration of this period of early LIA glacier ex-pansion is uncertain at locations within Todd valley, atSurprise glacier moraine sediments deposited during this ad-

© 2008 NRC Canada

Jackson et al. 95

Fig. 10. View northwest towards the exposed cutbank at site 4,where meltwater incision exposed a forest overrun by an advanceof Sage glacier around 2300 14C BP.

Fig. 11. View of detrital boles and stumps overrun in 1899 A.D.by an advance of Two glacier at site 5.

Fig. 12. View of north-facing lateral moraine at Surprise glacier.Total vertical exposure height is ca. 145 m. Radiocarbon-datedstratigraphic horizons corresponding to periods of glacial ad-vance are delineated. Where living tree-ring chronologies wereused to date the horizon, the chronology used is indicated in pa-rentheses. (A) 1848 A.D. (LC-SG); (B) 1764 A.D. (LC-SG);(C) 610 14C years BP; (D) 1440 14C years BP; (E) 1690 14Cyears BP; (F) 2960 14C years BP.

Fig. 13. Compilation of radiocarbon and cross-dated ages as-signed to subfossil boles and stumps in Todd and Surprise val-leys. TI designates floating tree-ring chronologies and LCdesignates living tree-ring chronologies (see Fig. 6; Table 4).Shaded intervals are times of glacier advance recorded by killdates assigned to perimeter age of cross-dated samples. Thequestion mark identifies a period of expansion described by sin-gle radiocarbon date.

vance were colonized with trees by 1430 A.D. (SG04-801;Table 4).

The discovery of glacially killed trees and transportedboles dating to the late LIA in Todd valley indicate that theearly LIA advances were followed by a period of ice reces-sion and downwasting. Considering that trees >200 years oldwere killed by this later advance, several centuries may haveelapsed before the initiation of late LIA glacier expansion inTodd valley. At Surprise glacier a corresponding interval ofdownwasting after the early LIA expansion is indicated bythe discovery of 316- to 177-year-old trees buried by thesucceeding late LIA advance.

There is limited evidence for the character of late LIAglacier activity in Todd valley. Radiocarbon dates on wooddetritus and buried boles show that Bug glacier, Two glacier,and Todd Glacier were all advancing into valley bottom for-ests 500 to 400 years ago. This advance likely correspondsto moraines constructed by glaciers farther south in the cen-tral Coast Mountains in the early- to mid-1700s A.D.(Desloges and Ryder 1990; Smith and Desloges 2000;Larocque and Smith 2003). This observation is supported bytrees killed in 1746 and 1764 A.D. at Surprise glacier whendebris spilled down the distal moraine slope.

Todd and Surprise glaciers appear to have reached theirmaximum downvalley Holocene extent during the late LIAinterval. At Bug glacier, the distal spillage of lateral morainesediments dating to 160 ± 60 14C years BP provides an indi-cation of when this period of maximum ice extent occurred.At Surprise glacier this period occurred sometime after1843–1848 A.D., when trees were killed on the morainecrest and 3 m of additional sediment was deposited. In Toddvalley, Two glacier continued to expand and kill mature treesas late as 1899 A.D.

Summary

Dendroglaciological investigations completed within twovalleys fed by glaciers originating from the Todd Icefieldprovide complementary chronologies of late Holocene gla-cier activity. Five intervals of glacier expansion were re-corded by trees killed or buried by advancing glaciers: (1) anadvance prior to �3000 14C years BP; (2) an advance at3000 14C years BP that corresponds with the regionalTiedemann advance; (3) an unattributed advance 2300 14Cyears BP that may distinguish a later Tiedemann phase or anunrelated regional mass balance response that remainspoorly appreciated; (4) a two-phase advance at �1700 and�1450 14C years BP that corresponds with the regional FirstMillennium advance; and (5) several phases of Little IceAge glacier expansion that are consistent with an early LIAphase of glacier expansion that began prior to �750 14Cyears BP and late LIA periods of glacier expansion culmi-nating after �1760 and 1900 A.D. This chronology of lateHolocene glaciation matches that emerging from similar in-vestigations in coastal Pacific North America and providesadditional support for the regional significance of both theTiedemann and the First Millennium advances.

Acknowledgements

Funding for this project was provided by grants from theNatural Sciences and Engineering Research Council of Can-ada (NSERC) and the Canadian Foundation for Climate andAtmospheric Science (CFCAS) to Smith. Field research activ-ities by Jackson and Laxton were partially supported bygrants from the Northern Scientific Training Program. Theauthors thank Sandy Allen, Dave Lewis, and Kelly Penrosefor their assistance in the field, Ole Heggen for draftingFigs. 2 and 6, and to the pilots of Prism Helicopters (Stewart,B.C.) for their always friendly logistical support. Sandy Allengraciously allowed us to reproduce Fig. 4. The authors wouldalso like to thank the two reviewers (J. Koch and B.Luckman) for comments that helped to improve the presenta-tion of our findings.

References

Allen, S.M., and Smith, D.J. 2007. Late Holocene glacial activityof Bridge Glacier, British Columbia Coast Mountains. CanadianJournal of Earth Sciences, 44: 1753-1773.

Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P.,deMenocal, P., Priore, P., Cullen, H., Hajdas, I., and Bonani, G.1997. A pervasive millennial-scale cycle in North Atlantic Holo-cene and glacial climates. Science (Washington, D.C.), 278:1257–1266.

Environment Canada. 2003. Canadian Climate Normals 1971–2000. Environment Canada, National Climate Data and Informa-tion Archive. http://climate.weatheroffice.ec.gc.ca/climate_nor-mals/index_e.html (last accessed May 2007).

Calkin, P.E., Wiles, G.C., and Barclay, D.J. 2001. Holocene coastalglaciation of Alaska. Quaternary Science Reviews, 20: 449–461.

Clague, J.J., and Mathews, W.H. 1992. The sedimentary record andNeoglacial history of Tide Lake, northwestern British Columbia.Canadian Journal of Earth Sciences, 29: 2383–2396.

Clague, J.J., and Mathewes, R.W. 1996. Neoglaciation, glacierdammed lakes, and vegetation change in northwestern BritishColumbia. Arctic and Alpine Research, 28: 10–24.

Clague, J.J., Wohlfarth, B., Ayotte, J., Eriksson, M., Hutchinson, I.,Mathewes, R., Walker, I.R., and Walker, L. 2004. Late Holoceneenvironmental change at treeline in the northern Coast Moun-tains, British Columbia, Canada. Quaternary Science Reviews,23: 2413–2431.

Cook, E., and Kairiukstis, L. (Editors). 1990. Methods ofdendrochronology in the environmental sciences. InternationalInstitute of Applied Systems Analysis. Kluwer, Dordrecht, TheNetherlands.

Cumming, B.F., Laird, K.R., Bennett, J.R., Smol, J.P., andSalomon, A.K. 2002. Persistent millennial-scale shifts in mois-ture regimes in western Canada during the past six millennia.Proceedings of the National Academy of Sciences, 99: 16 117 –16 121.

Davis, P.T. 1988. Holocene glacier fluctuations in the AmericanCordillera. Quaternary Science Reviews, 7: 129–157.

Denton, G.H., and Karlén, W. 1973. Holocene climatic variations— their pattern and possible cause. Quaternary Research, 3:155–205.

Desloges, J.R., and Ryder, J.M. 1990. Neoglacial history of theCoast Mountains near Bella Coola, British Columbia. CanadianJournal of Earth Sciences, 27: 281–290.

© 2008 NRC Canada


Dyurgerov, M.B., and Meier, M.F. 2000. Twentieth century climatechange: evidence from small glaciers. Proceedings of the Na-tional Academy of Sciences, 97: 1406–1411.

Easterling, D.R., Peterson, T.C., and Karl, T.R. 1996. On the devel-opment and use of homogenized climate datasets. Journal ofClimate, 9: 1429–1434.

Grissino-Mayer, H. 2001. Evaluating cross-dating accuracy: man-ual and tutorial for the computer program COFECHA. Tree-Ring Research, 57: 205–221.

Grove, E.W. 1971. Geology and mineral deposits of the Stewartarea, northwestern British Columbia. British Columbia Depart-ment of Mines and Petroleum Resources, Victoria, British Co-lumbia, Bulletin 58.

Guay, R., Gagnon, R., and Morin, H. 1992. A new automaticmethod and interactive tree-ring measurement system based on aline scan camera. Forestry Chronicle, 68: 138–141.

Haspel, R., Osborn, G.D., and Spooner, I. 2005. Neoglacial depos-its of Bear River Glacier, northern Coast Ranges, British Colum-bia. Annual meeting of the Western Division of the CanadianAssociation of Geographers, Lethbridge, Alta., March 2003,Program with Abstracts, pp. 11–13.

Hickin, A.S., Brooks, E.D., and Bobrowsky, P.T. 2001. NorthCoast LRMP aggregate potential map: accompanying notes.British Columbia Ministry of Energy, Mines and Petroleum Re-sources, Victoria, British Columbia, Open File 2001-19, mapscale 1 : 2 000 000.

Hoadley, R.B. 1990. Identifying wood: accurate results with simpletools. Taunton Press, Newton, Conn.

Hu, F.S., Ito, E., Brown, T.A., Curry, B.B., and Engstrom, D.R.2001. Pronounced climatic variations in Alaska during the lasttwo millennia. Proceedings of the National Academy of Sci-ences, 98: 10 552 – 10 556.

Koch, J., Clague, J.J., and Osborn, G.D. 2007. Glacier fluctuationsduring the past millennium in Garibaldi Provincial Park, south-ern Coast Mountains, British Columbia. Canadian Journal ofEarth Sciences, 44: 1215–1233.

Koch, J., Osborn, G.D., and Clague, J.J. 2007. Pre-‘Little Ice Age’glacier fluctuations in Garibaldi Provincial Park, Coast Moun-tains, British Columbia, Canada. The Holocene, 17: 1069–1078.

Lamoureux, S.F., and Cockburn, J.M.H. 2005. Timing and climaticcontrols over Neoglacial expansion in the northern Coast Moun-tains, British Columbia, Canada. The Holocene, 15: 619–624.

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: 1413–1436.

Larocque, S.J., and Smith, D.J. 2005a. Little Ice Age proxy glaciermass balance records reconstructed from tree rings in the Mt.Waddington area, British Columbia Coast Mountains, Canada.The Holocene, 15: 748–757.

Larocque, S.J., and Smith, D.J. 2005b. A dendroclimatological re-construction of climate since AD 1700 in the Mt. Waddingtonarea, British Columbia Coast Mountains, Canada.Dendrochronologia, 22: 93–106.

Lewis, D.H., and Smith, D.J. 2004. Little Ice Age glacial activityin Strathcona Provincial Park, Vancouver Island, British Colum-bia. Canadian Journal of Earth Sciences, 41: 285–297.

Lewis, D.H., and Smith, D.J. 2005. Dendroglaciological evidenceof late Holocene glacier activity at Forrest Kerr Glacier, AndreiIcefield, northern British Columbia Coast Mountains. Annualmeeting of the Canadian Association of Geographers. The Uni-versity of Western Ontario, London, Ont.

Levy, L.B., Kaufman, D.S., and Werner, A. 2004. Holocene glacierfluctuations, Waskey Lake, northeastern Ahklun Mountains,southwestern Alaska. The Holocene, 14: 185–193.

Luckman, B.H. 1988. Dating the moraines and recession ofAthabasca and Dome glaciers, Alberta, Canada. Arctic and Al-pine Research, 20: 40–54.

Luckman, B.H. 1993. Glacier fluctuation and tree-ring records forthe last millennium in the Canadian Rockies. Quaternary Sci-ence Reviews, 12: 441–450.

Luckman, B.H. 1995. Calendar-dated, early “Little Ice Age” gla-cier advance at the Robson Glacier, British Columbia, Canada.The Holocene, 5: 149–159.

Luckman, B.H. 1998. Dendroglaciologie dans les Rocheuses duCanada. Géographie physique et Quaternaire, 51: 1–13.

Luckman, B.H. 2000. The Little Ice Age in the Canadian Rockies.Geomorphology, 32: 357–384.

Luckman, B.H., Holdsworth, G., and Osborn, G.D. 1993.Neoglacial glacier fluctuations in the Canadian Rockies. Quater-nary Research, 39: 144–155.

Meidinger, D., and Pojar, J. 1991. Ecosystems of British Columbia.Research Branch, British Columbia Ministry of Forests, Victo-ria, British Columbia.

Menounos, B., Clague, J.J., Gilbert, R., and Slaymaker, H.O. 2005.Environmental reconstruction from a varve network in thesouthern Coast Mountains, British Columbia, Canada. The Ho-locene, 15:1163–1171.

Menounos, B., Koch, J., Osborn, G.D., Clague, J.J., andMazzucchio, D. 2004. Early Holocene glacier advance, southernCoast Mountains, British Columbia, Canada. Quaternary Sci-ence Reviews, 23: 1543–1550.

Moberg, A., Sonechkin, D.M., Holmgren, K., Datsenko, N.M., andKarlén, W. 2005. Highly variable Northern Hemisphere temper-atures reconstructed from low- and high-resolution proxy data.Nature (London), 433: 613–617.

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: 1141–1155.

Osborn, G.D., Menounos, B., Koch, J., Clague, J.J., and Vallis, V.2007. Multi-proxy record of Holocene glacial history of theSpearhead and Fitzsimmons ranges, southern Coast Mountains,British Columbia. Quaternary Science Reviews, 26: 479–493.

Pellat, M., Smith, M., Mathewes, R., Walker, I., and Palmer, S.2000. Holocene treeline and climate change in the subalpinezone near Stoyoma Mountain, Cascade Mountains, southwesternBritish Columbia, Canada. Arctic, Antarctic, and Alpine Re-search, 32: 73–83.

Reyes, A.V., and Clague, J.J. 2004. Stratigraphic evidence for mul-tiple Holocene advances of Lillooet Glacier, southern CoastMountains, British Columbia. Canadian Journal of Earth Sci-ences, 41: 903–918.

Reyes, A.V., Wiles, G.C., Smith, D.J., Barclay, D.J., Allen, S.,Jackson, S., Larocque, S., Laxton, S., Lewis, D.H., Calkin, P.E.,and Clague, J.J. 2006. Expansion of alpine glaciers in PacificNorth America in the first millennium AD. Geology, 34: 57–60.

Ryder, J.M. 1987. Neoglacial history of the Stikine–Iskut area,northern Coast Mountains, British Columbia. Canadian Journalof Earth Sciences, 24: 1294–1301.

Ryder, J.M., and Thomson, B. 1986. Neoglaciation in the southernCoast Mountains of British Columbia: chronology prior to thelate Neoglacial maximum. Canadian Journal of Earth Sciences,23: 273–287.

Schweingruber, F.H. 1988. Tree-rings: basics and applications ofdendrochronology. Kluwer Academic Publishers, Dordrecht,The Netherlands.

Smith, D.J., and Desloges, J.R. 2000. Little Ice Age history ofTzeetsaytsul Glacier, Tweedsmuir Provincial Park, British Co-lumbia. Géographie physique et Quaternaire, 54: 135–141.

© 2008 NRC Canada

Jackson et al. 97

Smith, D.J., and Larocque, C.P. 1996. Dendroglaciological datingof a Little Ice Age advance at Moving Glacier, Vancouver Is-land, British Columbia. Géographie physique et Quaternaire, 50:47–55.

Smith, D.J., and Lewis, D.H. 2007. Dendroglaciology. In Encyclo-pedia of Quaternary science. Edited by S.A. Elias. Elsevier Sci-entific, Vol. 2, pp. 986–994.

Spooner, I., Haspel, R., and Osborn, G.D. 2005. Holocene historyof Bear River Glacier, northern Coast Ranges, British Columbia.In Water, ice, land, and life: the Quaternary interface. CanadianQuaternary Association, 2005 Conference, June 5–8, 2005, Uni-versity of Manitoba, Winnipeg, Man. Abstract Vol., p. A87.

Spooner, I.S., Barnes, S., Baltzer, K.B., Raeside, R., Osborn, G.D.,and Mazzuchi, D. 2003. The impact of air mass circulation dy-namics on Late Holocene paleoclimate in northwestern NorthAmerica. Quaternary International, 108: 77–83.

Stuiver, M., and van der Plicht, J. 1998. Editorial Comment. Radio-carbon, 40: xii-xiii.

Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen,K.A., Kromer, B., McCormac, G., van der Plicht, J., and Spurk,

M. 1998. INTCAL98 radiocarbon age calibration, 24 000 – 0 calBP. Radiocarbon, 40: 1041–1083.

Wigley, T.M.L., Briffa, K.R., and Jones, P.D. 1984. On the averageof correlated time series, with applications in dendroclimatologyand hydrometeorology. Journal of Climate and Applied Meteo-rology, 23: 201–213.

Wiles, G.C., and Calkin, P. 1990. Neoglaciation in the southernKenai Mountains, Alaska. Annals of Glaciology, 14: 319–322.

Wiles, G.C., Calkin, P., and Jacoby, G. 1996. Tree-ring analysisand Quaternary geology: principles and recent applications.Geomorphology, 16: 259–272.

Wiles, G.C., Barclay, D.J., and Calkin, P.E. 1999. Tree-ring-dated“Little Ice Age” histories of maritime glaciers from westernPrince William Sound, Alaska. The Holocene, 9: 163–173.

Wiles, G.C., Jacoby, G., Davi, N., and McAllister, R. 2002. LateHolocene glacier fluctuations in the Wrangell Mountains,Alaska. Geological Society of America, Bulletin, 114: 896–908.

Wood, C.; and Smith, D.J. 2004. Dendroglaciological evidence fora Neoglacial advance of the Saskatchewan Glacier, Banff Na-tional Park, Canadian Rocky Mountains. Tree-Ring Research,60: 59–65.

© 2008 NRC Canada


dendroglaciological evidence for holocene glacial advances in the … · dendroglaciological...

Documents