Late-Holocene fluctuations of the Bear River Glacier,northern Coast Ranges of British Columbia,Canada

Abstract

During Holocene advances of the Bear River Glacier, in northwestern British Columbia, ice pushed against a bedrock slope on the north side of Bear River Pass as it was being diverted eastward and westward along the pass. The result is a series of till sheets plastered against the rock slope and separated by wood mats. This sequence of tills provides an unusually detailed record of early-Neoglacial history and shows that Neoglacial expansion in the northern Coast Mountains began earlier than previously known. Moraines show the extent of ‘Little Ice Age’ expansion, and documents related to the long mining history of the area provide records of post-LIA recession. The oldest deposits in the area are a small lateral moraine outside of a prominent LIA moraine, and a till in the pass older than c. 3500 ¹⁴C yr BP. The small moraine is undated; it may be equivalent to the latest-Pleistocene Crowfoot Advance which produced similar moraines in the southern Coast Mountains and eastern Cordillera. Wood mats in the pass show that Neoglacial advances occurred at c. 3700, 3500, 3300, and 1000 ¹⁴C yr BP; each advance was slightly more extensive than the previous one. Documentary evidence indicates that the pass was still filled with ice in 1913, and diverging both eastward and westward, but glacier retreat has been rapid since then. Retreat from the pass in the mid-20th century resulted in the creation of Strohn Lake, through which meltwater from the glacier passes; a complicated lake history includes switching of discharge from eastward to westward and a series of jokuhlaups that damaged highway infrastructure to the west. A comparison of the history recorded at Bear River Pass with that recorded at the Todd Icefield 15 km to the north shows that different components of glacial history are preserved in different places, and many sites need to be examined to obtain a relatively complete regional history.

Keywords

Bear River Glacier British Columbia Coast Mountains glacier fluctuations late Holocene till sheets

Introduction

Significant paleoenvironmental information can be gained from the study of the deposits and landforms of contemporary alpine glaciers. Glacier mass balance is influenced by annual variations in precipitation and temperature, making glaciers sensitive indicators of climate. However, glacial records often are poorly preserved; fluvial activity or subsequent glacial advances may overprint or override previous deposits, reducing the resolution of the paleoclimate record. Such is the case for Holocene deposits in the Canadian Cordillera: the latest glacial advance, that of the ‘Little Ice Age’ (LIA), was generally the most extensive of the Holocene, and in most cases obliterated pre-LIA records. Hence, stratigraphic sequences with stacked, unconformity-bounded tills, such as those described by Desloges and Ryder (1990), Ryder and Thomson (1986), Osborn et al. (2001), Reyes and Clague (2004), and Jackson et al. (2008) are of considerable value. In this paper we interpret such a stratigraphic sequence, produced by late-Holocene fluctuations of the Bear River Glacier in the northern Coast Ranges of British Columbia. Here, tills and interbedded wood mats provide an unusually complete early-Neoglacial history. Additionally, there is a historical record of glacier size nearly a century long, because of mining activity in the area.

Radiocarbon ages are reported herein as ¹⁴C yr BP; calendar ages are reported as ka except for ages of the last few centuries which are reported as calendar years ad.

Geography and geology of the site

The glacier in question is officially known as the Bear River Glacier (NTS mapsheet 104A/4) but is referred to locally as ‘Bear Glacier’. Prior to 1910, it was known as ‘Grosse Glacier’. It is located in the Boundary Ranges of the northern Coast Mountains, 40 km northeast of Stewart, British Columbia. Deposits of the Bear River Glacier can be accessed easily from Highway 37a, northeast of Stewart, BC (Figure 1).

Figure 1.

Location map of Bear River Glacier region, which is situated in the Boundary Ranges of the northern Coast Mountains, 40 km northeast of Stewart, British Columbia.

The landscape is typical of the Coast Ranges, with high, jagged, glacially modified peaks rising from steep-walled, deeply incised valleys (Grove, 1971). Subalpine fir (Abies lasiocarpa) and mountain hemlock (Tsuga mertensiana) rise from dense alder (Alnus sp.) thickets that cling to rounded ridges in the lower elevations. Average annual snowfall in Stewart is approximately 400–600 cm/yr with the greatest amount of snow delivered to low elevations from November to March. Average annual total precipitation is 1832 mm/yr (Environment Canada, 2010). Here, and throughout much of BC, weather is most strongly influenced by the winter position of the jet stream and the strength of the Aleutian Low (Walker and Pellatt, 2003). The resultant variable patterns of maritime air mass incursion make this region prone to widely changing climatic conditions (Barnes, 2003).

Flanked by Mounts Strohn and Disraeli, the Bear River Glacier emerges from its source on the 2135 m high plateau of the Cambria Icefield (Figure 1). The 5 km long glacier descends northward into Bear River Pass, a low divide in the east–west trending Bear River Valley, and terminates in 1.7 km long and 300 m wide Strohn Lake (416 m a.s.l., Figure 2). The actual hydrologic divide has shifted over the past 100 years, depending on the interplay of ice extent, lake extent, and human intervention; presently it lies east of Strohn Lake (Figure 1). Today glacial meltwaters from Strohn Lake drain into the Bear River, flowing first west, then south, before emptying into the Portland Canal at Stewart, BC.

Figure 2.

Map of study site showing prominent geomorphological features and the location of the sample sites in Figure 3. The general area of Strohn Lake is locally referred to as ‘Bear River Pass’ although the hydrologic divide presently is located east of the lake.

The Bear River Glacier and Pass and the Cambria Icefield source area are underlain largely by dark green and maroon-weathering volcaniclastics of the Lower Jurassic Hazelton Group (Greig et al., 1994). The Hazleton Group consists of massive- to poorly sorted, coarse- to fine-grained pyroclastic fragmental volcanics and related mass flow deposits that were laid down in a shallow marine or subaerial environment. The bedrock of the northern valley wall is andesitic to dacitic lapilli and ash tuff with scattered volcanic rock fragments. Colluvium from the Mt Disraeli nunatak may contain more mafic, massive ash and lapilli tuff beds with patchy carbonate cement. Material originating from the east side of the Cambria Icefield may contain clasts of the rusty-weathering clastic rocks of the Salmon River Formation. These are thinly bedded and laminated mud- and siltstones with lenses of calcareous sandstones and limestone. Prospecting for mineral resources in the vicinity of the pass has gone on intermittently for over 100 years, and a consequence of that activity is a relatively rich documentary record of recession of the glacier.

Known framework of Holocene environmental history in British Columbia

In British Columbia, the rapid transition from Pleistocene glacial to Holocene interglacial conditions occurred 12,500 to 9000 ¹⁴C yr BP (Walker and Pellatt, 2003). By 14.0 ka most alpine glaciers were probably no larger than they were at the end of the 20th century (Menounos et al., 2009). The Pleistocene/Holocene transition was marked by a minor glacial advance known in the Canadian Rockies and southern British Columbia as the Crowfoot Advance (Osborn and Luckman, 1988), which was dated to between 11,300 and 10,100 ¹⁴C yr BP and related to the Younger Dryas climatic reversal by Reasoner et al. (1994). A warm and dry climatic period, sometimes called the Altithermal, followed the Crowfoot Advance and characterized early-Holocene time. This interval was only briefly interrupted by the ‘8200 yr cold event’ during which glaciers in western Canada may have expanded slightly in response to hemispheric cooling associated with temporary alteration of North Atlantic thermohaline circulation (Menounos et al., 2004).

Declining summer temperatures, wetter conditions and the development of a stronger Aleutian Low contributed to mid-Holocene cooling (Walker and Pellatt, 2003) and led to a period of general glacial expansion termed Neoglaciation. In southern BC, palynologic and paleolimnologic records point to climatic deterioration beginning 5000 to 6000 ¹⁴C yr BP according to Walker and Pellatt (2003), or culminating at that time according to Heusser et al. (1985). Glacially transported and in situ tree stumps are evidence of a period of expanded glacial cover ~6000 to 5000 ¹⁴C yr BP (Osborn et al., 2007; Ryder and Thomson, 1986). This expansion was termed the ‘Garibaldi Phase’ by Ryder and Thomson (1986). Further advance of at least some glaciers in the southern Coast Mountains, and Rocky Mountains and Interior Ranges, occurred c. 4300–3700 ¹⁴C yr BP (c. 4.9–3.9 ka); this is termed the ‘4.2 ka advance’ by Menounos et al. (2008). In the northern Coast Mountains Neoglacial cooling may have been delayed, with marked decreases in temperature around 4000 to 3000 ¹⁴C yr BP (Walker and Pellatt, 2003). There is no evidence of Garibaldi Phase expansion. Prior to investigation at Bear River Pass there was no evidence of 4.2 ka expansion in this region; the oldest known Neoglacial activity was at c. 3000 ¹⁴C yr BP in the northern Coast Mountains (e.g. Jackson et al., 2008; Lewis and Smith, 2005) and c. 3300 ¹⁴C yr BP in the St Elias Mountains (Denton and Karlen, 1973).

Glaciers throughout western Canada were advancing c. 3300 to 1900 ¹⁴C yr BP to positions downstream of Garibaldi phase limits. This activity is referred to as the Tiedemann Advance by Ryder and Thomson (1986). Lake sediment proxies suggest that in southern BC at least some glaciers extended almost to LIA limits by 3000 ¹⁴C yr BP (Osborn et al., 2007). Menounos et al. (2009) note that the Tiedemann interval was characterized by numerous advances and retreats, not all of which necessarily left evidence at any one site. Three major glacial pulses in the Tiedemann interval were identified by Arsenault et al. (2007) based on interpretation of sediments and fossil pollen in a Holocene bog near the Tiedemann Glacier. Advances occurred prior to 2600, 2500–2300, and after 1900 ¹⁴C yr BP, with the middle advance being the most extensive. Similar chronologies have emerged from studies in the northern Coast Mountains (Clague and Mathews, 1992; Clague and Mathewes, 1996) and southern Coast Mountains (Allen and Smith, 2007), although there is some variation in terminology (see also Clague et al., 2009, re: terminological confusion). The Peyto Advance in the Canadian Rockies (Luckman et al., 1993) appears to be synchronous with the Tiedemann advances.

In coastal British Columbia and Alaska many glaciers were advancing, and shearing rooted stumps, in the period ad 400–700 (or c. 1.5 ka). Reyes et al. (2006) refer to this episode as the first millennium ad advance (FMA). At least in two cases the first millennium expansion was comparable in magnitude with maximum LIA advances. Allen and Smith (2007) suggest that early LIA expansion of the Bridge Glacier in the Coast Mountains was more or less a continuation of the FMA. The FMA may also be recorded at Peyto Glacier in the Canadian Rockies (Luckman, 2006).

The LIA is the latest phase of Neoglacial expansion. It commenced prior to 900 ¹⁴C yr BP but may have begun as early as 1000 ¹⁴C yr BP in the Coast Ranges (Clague and Matthews, 1992). Throughout western North America, alpine glaciers remained extensive for the next several centuries with minor fluctuations. Many glaciers reached their Holocene maxima in the 16th to mid-18th centuries or near the culmination of the LIA in the late 18th and 19th centuries (Koch et al., 2007; Luckman, 1986; Luckman and Villalba, 2000). Recession has been dominant in the past century with retreat initially slow and irregular, but becoming more rapid and continuous in the late 20th century (e.g. Ryder, 1987).

Stratigraphy of major exposures

For the purposes of this study, locations of glacial deposits in the Bear River valley have been subdivided into three major areas that are defined and described below. In the eastern part of the study area, two vegetated end moraines are found at, or just below, Bear River Pass. To the west, LIA lateral and recessional moraines border Strohn Lake. The third area described, and the major focus of the study, is a 180 m high till exposure on the northern valley wall, referred to here as the ‘north till exposure’ (Figure 3). The 1.6 km long till face was further subdivided into outcrops A (east), B (central), and C (west).

Figure 3.

Location airphoto of outcrop areas A, B and C in the north till exposure. ‘M’ designators refer to wood mats. Wood Mat M3 separates the Gray Till (GT) and the Brown Till (BT). ‘B’ designators are sampling sites for radiocarbon dating (see Table 1). The dates represent average ¹⁴C ages for a particular wood mat. Airphoto 30BCC96104 NO. 217.

Table 1.

Radiocarbon dates.

Sample (lab ID)	¹⁴C age	2σ calibrated date (ka)	Description
B1-03 (B-180666)	3680 ± 60 BP	4160–3850	detrital, Wood Mat 1
B3-03 (B-180667)	3790 ± 70 BP	4380–3820, 3780–3730	detrital, Wood Mat 1
B3B-04 (B193481)	3380 ± 60 BP	3820–3780, 3730–3470	detrital, gray-brn contact,west
B4B-04 (B-193482)	3710 + /70 BP	4250–3860	in situ, Wood Mat 1
B5-03 (B-181861)	3330 ± 60 BP	3695–3435	detrital, Wood Mat 3
B7-04 (B-193483)	3540 ± 60 BP	3980–3670	in situ, Wood Mat 2
B8-04 (B-193484)	3310 ± 70 BP	3700–3380	in situ, buried by gray till, associated with Wood Mat 3
B9-04 (B-195167)	3410 ± 60 BP	3830–3480	detrital, in gray till, associated with Wood Mat 3
B10B-04 (B-193485)	1040 ± 50 BP	1060–910	detrital, Wood Mat 4

Eastern deposits and exposures

An eastern trimline of the glacier is delineated by (a) the upper limits of deeply gullied till plastered to the lower flanks of Mt Strohn and (b) the lower limit of conifers on the mountain’s vegetated cliffs. The trimline wraps around the base of the mountain at a low angle before descending steeply toward Bear River Pass. The trimline projects to a point on the pass marked by two closely spaced end moraines, each about 5 m in height and exposed in a road cut near the pass. These moraines are thought to mark the glacier’s maximum LIA extent (Figure 4). A deep, now-dry canyon carved by Strohn Creek is located south of the moraines, against the steep valley wall. Prior to 1963, eastward-flowing Strohn Creek was the major outlet of meltwater of the Bear River Glacier.

Figure 4.

Timing of recent recession of the terminus of the Bear River Glacier. Airphoto 30BCC96104 NO. 217. The steep-sided lateral moraine referred to in text is near the left edge of the frame.

Western lateral moraines

Partly perched on a bedrock ledge, a steep-sided lateral moraine of the Bear River Glacier rises 160 m above the western edge of Strohn Lake (Figure 4). The crest of the 500 m long moraine curves slightly to the northwest and plunges at 5–10° into the Bear River valley. While the distal limb of the moraine is densely vegetated, the till on the proximal face is exposed and has been significantly gullied by slopewash. At roughly three-quarters height, a line of boulders (30 cm to several meters in size) oriented subparallel to the moraine crest separates two tills of slightly different colour. The upper till is slightly lighter and grayer in colour than the underlying brownish till. The lateral moraine is therefore taken to be a composite feature consisting of material derived from successive advances. No wood was recovered at this site. A possible in situ tree stump was observed some 3 m below the moraine crest in the south-central part of the gully but hazardous conditions thwarted an attempt to retrieve a sample.

Lying outside this lateral moraine and hidden from view of the highway, lies a remnant of a smaller, densely vegetated moraine (‘pre-LIA moraine’ of Figure 2). The rounded ridge angles to the northwest more strongly than the prominent, previously described moraine. Interpretation of this feature is based on air photo interpretation as dense vegetation and steep slopes prevented access.

A prominent, recent (post-LIA maximum) lateral recessional moraine 30 m high extends from the northwestern edge of the glacier to the edge of Strohn Lake.

North till exposure

During the most extensive Holocene advances, the northerly flowing ice stream of the Bear River Glacier would impinge against the steep cliffs of Yvonne Peak, filling the Bear River Valley. The ice was forced upward against the cliff and flow was diverted, with the glacier branching to form east and west arms. Each successive advance plastered an additional layer of till against the northern valley wall to a maximum height of 180 m above modern lake levels. This has resulted in an unusually well preserved succession of tills at this location.

Mats of organic material that separate individual till units have been used to bracket periods of till deposition. Each mat is believed to preserve material growing or scattered on the glacier forefield that was buried during the subsequent advance. Each horizon contains disrupted remnants of a paleosol, as well as abundant detrital and, in some cases, in situ wood fragments. Woody material varies in size from small, centimeter-sized twigs to thick or elongate branches, to stumps greater than 1 m in length. Wood is generally friable, dessicated and/or enclosed in bark. In some cases bark fragments, rootlets, charcoal, and conifer needles are enclosed in lenses of dark, organic-rich soil. In contrast, isolated detrital fragments outside of such horizons are largely rounded, dense fragments c. 10 to 20 cm in diameter. Porous, but in places well-cemented, sand and/or silt partially encases the wood and the roots of tilted, in situ stumps. Pockets of stratified, oxidized sand with subvertical hairline fractures and soft sediment deformation are also found near wood mats; stratification in these pockets consists of millimeter-thick laminae of alternating layers of light-coloured silt and fine sand with darker, coarse to very coarse sand.

The organic mats are exposed in deep gullies in the northern till face with slope angles reaching 30° to 40°. The lower three are located near the center of the exposure, below a 20 m high bedrock cliff band, and dip between 10° and 25° to the southwest. Wood mats 1 and 2 are exposed at Outcrop B, while Wood Mat 3 was observed at Outcrop C. Wood samples recovered from the mats provide limiting ages of the enclosing tills; detrital samples provide a maximum age for an advance while sheared, in situ stumps were overrun by the ice and thus directly date the advance. The radiocarbon ages allow reconstruction of a relatively complete Neoglacial history of the Bear River Glacier.

Outcrop A

At this site, an undated woodmat located approximately 23 m above the roadway could be traced in and out of gullies for over 100 m. Although oxidation associated with the horizon was evident, organic fragments were not as large or prevalent as in the Section B wood mats. The organic-rich layer dips approximately 35° to the south and is possibly correlative to the lower wood mat at Outcrop B (see below).

Outcrop B: Wood Mats 1 and 2 (Figure 3, M1 and M2)

Wood Mat 1, the lowermost of the three exposed horizons, is, at its highest, most easterly exposure, 26 m above the Stewart–Cassiar highway (Figure 3). Two detrital wood fragments (B1-03 and B3-03) from this mat yielded dates of 3680 ± 60 and 3790 ± 70 ¹⁴C yr BP, respectively. Sample B4B-04 was a fragment obtained from an in situ tilted tree trunk measuring nearly 30 cm in diameter and greater than 1 m in length with its roots imbedded at the level of the wood mat (Figure 5). The trunk is lying on top of and nearly parallel to the lower mat, with its axis oriented roughly northeast. The roots themselves, however, are still in place and envelop decimeter-sized, oxidized pebbles and penetrate small lenses of moderately sorted silts that are distinguishable from the till above and below. Those roots on the upslope side are bent (as if by compression), while those on the downslope side are sheared and cracked. The sample, obtained from outer rings of the rootstock, yielded an age of 3710 ± 70 ¹⁴C yr BP.

Figure 5.

Sample B4B-04, an in situ tree stump and associated trunk, located in Wood Mat 1. Ice-axe for scale.

Wood Mat 2 terminates eastward at the base of the bedrock cliff, some 20 m above the lower horizon. A date of 3540 ± 60 ¹⁴C yr BP was obtained from an in situ stump (B7-04) with delicate, well-preserved rootlets anchoring small lenses of oxidized silt at the level of the mat.

Outcrop B: Gray and brown tills

Approximately 3 m above Wood Mat 2 is a contact between two tills with slightly different colour and physical and chemical properties. The tills differ from each other with regard to colour, rheology, reaction to HCl, and grain size distribution (Figure 6). Although the contact between the tills is diffuse and not everywhere obvious, it generally extends across the length of the northern till slope, and dips to the south at an angle less than that of the till slope itself. The upper till is gray, reacts vigorously with dilute HCl, and when saturated with water tends to flow in small, steep-fronted, viscous lobes. These small lobes were observed to occur naturally and can also be created artificially. The lower till is brown, reacts less strongly with acid, and does not exhibit the lobate flow behavior when saturated.

Figure 6.

Photograph showing the contact between the gray till and the brown till at outcrop-area B (Figure 2). Inset shows a sorted sand lens of the type that is common near the top of the Brown Till. Trowel blade is 8 cm wide.

Because the two tills presumably came from the same source area, there is no obvious reason why they should have different properties. Till samples were collected for laboratory analysis in an attempt to understand the conditions of deposition that resulted in the different till properties. Grain size analysis of the two tills was carried out using conventional dry-sieving techniques followed by Mastersizer 2000 particle analysis of the fine-grained (<0.5 mm) component. Analyses and field observations indicate that the brown till is more bouldery, with a matrix that is slightly coarser-grained and more strongly bimodal than in the gray till. Sieving results show that the gray till is dominated by a finer-than-sand fraction. Analysis of the <0.5 mm component yielded similar results with 53% (by volume) of the gray till composed of silt-sized or finer grains in contrast to only 36% (by volume) for the brown till.

Mineralogically, the gray and brown tills are virtually identical. In XRD analysis of the clay-sized portion, diffraction peaks corresponding to quartz, unweathered feldspars and mineral weathering products of the sheet silicates are virtually identical in the two tills. Major-element analyses on bulk samples of the two tills (expressed as weight percentages of oxides) show that the bulk chemistry is quite similar with the exception of CaO, which, at 2.52% in the gray till, is 85% greater than the 1.36% in the brown till, corresponding to the gray till’s greater reaction with weak acid.

One possibility that could explain the slight differences between the two tills is that the tills represent two glacier advances separated by a period of lake-sediment deposition. According to this hypothesis the brown till was deposited by ice that did not encounter lake sediments as it advanced through the pass. A lake similar to the present Strohn Lake formed during subsequent recession, and consequent lake sediments were deposited over the pass. The subsequent readvance incorporated lake sediments, resulting in the slightly finer and grayer character of the upper till. This hypothesis, however, cannot account for the brown/gray discontinuity in the western LIA moraine, described earlier, because that moraine is not downstream of Strohn Lake. That colour change is even less pronounced than in the north till exposure and it is not known whether the two discontinuities have the same origin.

Outcrop B: Wood Mat 3 (Figure 3, M3)

In the 2003 field season a wood mat including moderately large log fragments could be seen associated with a line of large boulders at the base of the gray till (Figure 6); this mat is interpreted to separate the gray and brown tills. A sample of detrital wood (B5-03) from the mat yielded an age of 3330 ± 60 ¹⁴C yr BP. By the summer of 2004 the large logs below the bedrock cliff had eroded away and the mat was no longer obvious.

In the western half of Outcrop B there is no evidence of an organic mat separating the gray and brown tills. However, a large detrital wood fragment apparently embedded in the uppermost brown till was located in the vicinity of the rather gradual and indistinct (when viewed close-up) contact between the two tills. The fragment was juxtaposed with lenses of laminated sand deformed by minor folds and faults such as those found associated with the wood mats farther east. This wood (B3B-04) yielded an age of 3380 ± 60 ¹⁴C yr BP, an age statistically indistinguishable from that obtained from Wood Mat 3 farther east.

Two wood samples were obtained from sites elevationally above, but not necessarily stratigraphically above, the gray/brown contact in the vicinity of the bedrock cliff. The first is from an in situ stump rooted in bedrock fractures 8 m above Wood Mat 2 and probably a few meters above the gray–brown contact (which is difficult to precisely delineate in close-up view). This gnarled, deformed stump had visible damage to growth rings on its south side. Remnants of gray till surrounding the stump suggest enclosure in the till prior to erosion. The sample (B8-04) yielded an age of 3310 ± 60 ¹⁴C yr BP.

The second sample (B9-04) is from a detrital log that was encased in gray till approximately 12 m above Wood Mat 2. It yielded an age of 3410 ± 60 ¹⁴C yr BP.

Outcrop B: Wood Mat 4 (Figure 3, M4)

The uppermost observed wood mat is situated 8 to 9 m below an overgrown road (the original Bear River Pass Trail) that traverses the north slope 180 m above Strohn Lake. This mat is similar in appearance, composition and orientation to the lower organic-rich horizons, although far less laterally extensive. Dessicated branches roughly 25 cm long and 8 cm in diameter are enclosed in unstratified silt and fine-grained sands. The wood fragments lie parallel to the southwestward dipping organic horizon. Dark, organic-rich soil and red, oxidized till is also present in the horizon. Sample B10B from this site yielded an age of 1040 ± 50 ¹⁴C yr BP.

Outcrop C

Till at Outcrop C is poorly exposed and there is evidence for considerable modification of the slope during road construction. No organic material was recovered from this site.

Other radiocarbon ages

Jackson et al. (2008) report two radiocarbon ages from detrital wood fragments from the north till exposure; they are 3680 ± 60 and 3340 ± 60 ¹⁴C yr BP. The stratigraphic contexts of these samples are not known, but it is likely they were derived from Wood Mats 1 and 3.

Interpretation

Pre-Neoglacial history

At the Bear River Glacier, deposits associated with the latest LIA advance have erased or overridden most evidence of pre-Neoglacial advances. However, the presence of the small, rounded, well-vegetated moraine ridge lying outside the conspicuous western LIA lateral moraine indicates that an older advance extended slightly farther than the subsequent LIA maximum.

Neoglacial history

Stratigraphic evidence and ages provided by detrital and in situ wood suggest that at least four periods of Neoglacial expansion are recorded in tills deposited by the Bear River Glacier. The fourth and youngest advance represents the LIA culmination of the Neoglacial period

Early Neoglacial

The till that underlies Wood Mat 1 was emplaced prior to ~3700 ¹⁴C yr BP by an otherwise undated advance of the Bear River Glacier that deposited an unknown thickness of sediments. This advance could have been of early Neoglacial, possibly Garibaldi, age. In any event, the glacier overrode trees growing on that till c. 3700 ¹⁴C yr BP, generating Wood Mat 1, and deposited about 20 m of sediment. This second moraine was subsequently overrun by an advance ~3500 ¹⁴C yr BP which generated Wood Mat 2 (Figure 3).

In the 200 ¹⁴C years subsequent to the 3500 ¹⁴C year advance, marked recession of the Bear River Glacier is believed to have occurred. During this time, detrital wood from the vegetated slopes above the northern till exposure was avalanched onto the glacier forefield. Some detrital wood, such as sample B3-04, was deposited in the laminated sands of small morainal ponds. Strohn Lake may have formed for the first time during this period, with glacial meltwaters carrying fine-grained, gray sediments into the lake. Possibly these sediments were later incorporated into the Bear Glacier’s push-moraine during an advance about 3300 ¹⁴C yr BP. As the ice ploughed into the northern valley wall it deposited till and crushed trees (i.e. B8) growing at the base of a bedrock cliff, and generating Wood Mat 3.

Lack of a radiocarbon-dated organic record for much of the upper half of the till slope limits interpretation of the Bear River Glacier’s behaviour from 3300 to 1040 ¹⁴C yr BP. The 100 m of till extending from the gray/brown till contact to the height of the upper wood mat were deposited either by the 3300 ¹⁴C yr BP advance or during subsequent, undated advances. The latter possibility is more probable, given that glaciers throughout the Coast Range record mid-Neoglacial advances c. 2200 to 2800 ¹⁴C yr BP (see section ‘Regional correlations’).

‘Little Ice Age’

The uppermost wood mat (4) contains evidence for the earliest LIA advance of the Bear River Glacier. Shortly after 1040 ¹⁴C yr BP, the glacier advanced overtop wood fragments lying on the tills deposited during prior Neoglacial advances (Figure 7). Ice surface levels at this time must have reached greater than 170 m in height above present-day Strohn Lake. Minor decadal to centennial fluctuations likely occurred throughout the LIA, although the terminus position remained advanced throughout this time.

Figure 7.

Schematic diagram of ice-advance episodes and relative ice extent recorded at the Bear Glacier River terminus. M1–M4 are the wood mats indicated in Figure 2.

Observations of early prospectors and residents suggest that a final LIA advance culminated in the late 18th or earliest 19th centuries. Moraine exposure occurred shortly thereafter (~ad 1810) based on tree-ring dating at the study site and an assumed ecesis interval of 10 to 20 years (Barnes, 2003).

Recent (20th century) recession

Owing to its location and relative accessibility along former trade routes, and present-day proximity to the international border town of Stewart BC, there are sufficient data on the recent history of the Bear Glacier that a fairly detailed reconstruction can be made. Tree ring data suggest exposure of the Bear River Glacier’s end moraine had begun by 1810 (Barnes, 2003). Rate of retreat was limited during the 19th century, however, as indicated in the observations of RG McConnell (1913): ‘The pass is covered with a glacier about one and a half miles in length, made up of two arms flowing in opposite directions, one eastwards towards the Nass [River] and the other westwards down Bear River valley’. Measurements made by McConnell’s party suggest that the western glacier arm must have been nearly 100 m thick at this point.

In contrast, over the last century, retreat of the Bear River Glacier has been both rapid and irregular, with the most significant loss of ice cover occurring from 1949 to 1968 (Barnes, 2003). In 1956 airphotos, only minor remnants of the glacier’s east and west arms remain, and by 1978 the glacier’s snout had attained its present rounded form (Figure 4). Decreased ice cover also contributed to a drainage reversal in the middle of the last century. In the past, the glacier blocked the western valley such that meltwater was forced to drain eastward from Bear River Pass into Strohn River (Mathews, 1965). By the 1940–1950s however, continued retreat led to the formation of Strohn Lake as water was impounded between the glacier’s west arm and Bear River Pass. In a 12 yr period (1950–1962) the lake expanded from 500 m to 900 m in length. Between 1958 and 1962, five jokulhaups occurred that catastrophically released water west into the Bear River. Discharge occurred every October, when lake levels reached their maximum, with floods washing out bridges downstream and causing severe disturbances in the highway construction occurring at this time (Grove, 1971; Mathews, 1965).

In 1961, the BC Department of Highways attempted to lower the maximum lake level by excavating a trench at the east end of Strohn Lake. However, in October of that year and the following year, the lake filled to the level of the trench floor before emptying westwards catastrophically. The Bear River Glacier continued to block the valley until 1962 when the ice surface had melted below the level of the original, eastern outlet into Strohn River. Another trench was dug in the winter of 1962–1963 in a second attempt to relieve pressure on the ice dam. This time, the 17 m deep trench was excavated across the toe of the glacier to permit westerly drainage into the Bear River (Mathews, 1965). By the summer of 1963 however, subglacial drainage was sufficient to keep lake levels low and no further jokulhaups occured. Strohn Lake continued to expand as the ice retreated and by July 1967, the Bear River Glacier had retreated completely from the north wall of the pass (Province of British Columbia, 2011). Rapid surface ablation and retreat have continued to the present. Today, the glacier terminates at an elevation of 457 m a.s.l. in Strohn Lake, which has achieved maximum dimensions of 1.7 km in length and 300 m in width.

Regional correlations

Pre-Neoglacial history

In the Canadian Rocky Mountains, moraine systems of the Crowfoot advance (~11,000 to 10,000 ¹⁴C yr BP), where they are preserved, lie short distances beyond LIA moraines (Osborn and Luckman, 1988). Crowfoot moraines are regarded as coeval with the Younger Dryas climatic reversal by Reasoner et al. (1994). At the Bear River Glacier, the densely vegetated moraine on the west side (‘pre-LIA moraine’ of Figure 2) is similarly located just beyond the LIA moraine, and could be correlative with the Crowfoot Advance. Recently other glacial evidence from the BC Coast Mountains has emerged in support of a Crowfoot advance in the region. A study by Friele and Clague (2002) points to a readvance of the Squamish Valley Glacier during this time interval, and a cirque moraine situated just beyond an LIA moraine east of Lillooet Lake in the southern Coast Mountains has a minimum age of 9680 ± 40 ¹⁴C yr BP. The cirque moraine is regarded as a probable Crowfoot equivalent by Minkus (2006) and Menounos et al. (2009).

Early Neoglacial

The earliest Neoglacial advance of the Bear River Glacier pre-dates the 3300–1900 ¹⁴C yr BP Tiedemann advance described by Ryder and Thompson (1986) by nearly 500 years. Further support for pre-Tiedemann glacial activity at about this time and in this region of northwestern BC is inferred from the thickness of the clastic wedge deposited in the Bear River Delta which Hanson (1934) calculated to have formed within the last 3600 to 4000 years. Glacioclimate records at the Berendon, Frank Mackie and Salmon Glaciers located 50 km northwest of the study area do not extend sufficiently far back in time to establish whether correlative advances occurred at these sites. However, similar ages to those obtained at the Bear River Glacier site have also been reported from glaciers in adjacent regions in Alaska and elsewhere in the Canadian Cordillera (Ryder, 1987 and references therein).

It seems plausible that Neoglacial cooling commenced around 4000 ¹⁴C yr BP in northern BC, based on a number of lines of evidence, and it is this cooling trend that may have initiated the early-Neoglacial advances of the Bear River Glacier. Proxy climate records at some locations in BC and Alaska point to climatic deterioration beginning ~4000 ¹⁴C yr BP that may be related to a shift in maximum summer insolation at the Arctic Circle (60° N) beginning 4500 ¹⁴C yr BP (Nesje and Johannessen, 1992). Shifts in vegetation patterns derived from the BC pollen record also suggest a trend towards generally cooler and wetter conditions from 4700 to 2200 ¹⁴C yr BP (Hebda, 1995). Ryder (1987) interpreted the burial of caribou antlers in a permanent snowbank near the Iskut River to be the result of increased snowfall beginning 4000 ¹⁴ C yr BP.

Similarly cool and wet climatic conditions likely contributed to Ryder and Thomson’s well-documented Tiedemann Advance in the BC Coast Ranges. At the Bear River Glacier, Wood Mat 3 provides further evidence supporting the early part of the Tiedemann Advance.

Glaciers throughout the Coast Range, including ones near Bella Coola (Desloges and Ryder, 1990), in Garibaldi Park (Cashman et al., 2002), and the Tiedemann (Arsenault et al., 2007), Frank Mackie (Clague and Mathews, 1992) and Berendon glaciers (Clague and Mathewes, 1996), all record mid-Neoglacial advances 2200 to 2800 ¹⁴C yr BP . As such, similar advances subsequent to the Bear River Glacier’s 3300 ¹⁴C yr BP advance are considered probable contributors to the deposits of till on the upper half of the till slope, though they are undated at this site.

‘Little Ice Age’

Coincident with the Bear River Glacier’s LIA radiocarbon dates is the earliest LIA phase of the Frank Mackie Glacier, north of Stewart BC which expanded to impound Tide Lake 1000 ¹⁴C yr BP (Clague and Mathews, 1992). Around 1000 to 1200 ¹⁴C yr BP meltwater flowing from the neighboring Berendon Glacier may have formed a thin, discontinuous mud layer found in the stratigraphic record of Berendon Fen (Clague and Mathewes, 1996). Elsewhere in the Coast Ranges, evidence from lacustrine records and tills indicate earliest LIA advances beginning in the 11th through 13th centuries (Menounos et al., 2009).

Long-term decreases in summer radiation superimposed on decadal to millenial climatic fluctuations may have been responsible for overall climatic deterioration through the late Neoglacial, pre-LIA times (Luckman, 1996). Resultant cooling may have contributed to the onset of the LIA in the Coast Ranges. Comprehensive reviews of LIA activity and assessments of LIA synchroneity in the North American Cordillera are given by Luckman (2000, 1996) and Luckman and Villalba (2000).

Recent (20th century) recession

Glaciers in the mountains near Stewart BC, including the Bear River Glacier, had already begun to retreat from their LIA maximum positions prior to the arrival of early prospectors and surveyors. Buddington (1926) was one such traveler that recorded his observations; in 1926 he noted: ‘All the glaciers [in the Hyder District] show bare rock surfaces along their sides for 100 ft or more above their present surface’. Impressive ablation of ice surfaces throughout the BC Coast Ranges has continued over the last 100 years and has resulted in the retreat of many alpine glaciers from their tributary valleys.

Throughout the region, decreased ice cover has been accompanied by the catastrophic release of ice-dammed lakes formerly impounded by the ice. Immediately northwest of Stewart BC, retreat of the Frank Mackie Glacier was responsible for outburst floods originating from Tide Lake in the 19th and early 20th century (Clague and Mathews, 1992). Mathews (1965) describes the emptying of nearby Summit Lake under the Salmon Glacier in December of 1961. Records of meteorological conditions at that time led Mathews to conclude that dam failure was due to phenomenal ice retreat and not to any unusual weather circumstances. Elsewhere in British Columbia, Marcus (1960) describes drainage of Tulsequah Lake between 1910 and 1911. The Fyles Glacier ice dam was also breached by the waters of Ape Lake beginning in 1984 (Gilbert and Desloges, 1987).

A comprehensive study by Barnes (2003) of varved sediments in Meziadan Lake suggests that recent (1948–present) variability in rates of retreat correspond to warm and cold phases of the Pacific Decadal Oscillation.

Discussion

The unique geographic setting of the Bear River Glacier, in which a valley glacier intermittently has pushed head-on into a rock cliff, has produced an unusually detailed record of early-Neoglacial history (Figure 7). The stacked tills provide direct evidence of successive minor advances and retreats closely spaced in time. Similar histories are indirectly suggested at other sites by lake-sediment records (e.g. Osborn et al., 2007). The Bear River Glacier was already close to its subsequent LIA maximum extent by c. 3700 ¹⁴C yr BP, and thereafter spent 4000 years oscillating on a small scale about that relatively extensive position. Each readvance for which there is evidence records a slightly thicker glacier, which would translate into a slightly longer glacier were it not for the interference of the rock cliff. In the vicinity of the rock face, the ice surface at the time of the LIA maximum was 150 m higher than the surface at c. 3700 ¹⁴C yr BP.

The Bear River Glacier provides the earliest evidence known so far of Neoglacial expansion in the northern Coast Mountains. The c. 3700 ¹⁴C yr BP age of the earliest recorded advance falls within the range of ages defining the 4.2 ka advance of Menounos et al. (2008), whereas the c. 3300 ¹⁴C yr BP advance falls within the Tiedemann Advance range of Ryder and Thomson (1986). The intermediate c. 3500 BP advance falls within neither but is of similar magnitude to both. These stratigraphic relationships confirm that as resolution of Neoglacial history continues to improve with new data, the established terminology of named advances becomes less meaningful, and potentially even misleading, a point suggested by Clague et al. (2009).

Although the early Neoglacial record at Bear River Pass is relatively detailed, the later Neoglacial record is not. Meanwhile, the record in the vicinity of the Todd Icefield, 15 km north of the pass, reveals no information on pre-3000 ¹⁴C yr BP history but much detail on history after that (Jackson et al., 2008). It is clear that, depending on the vagaries of local geography, forest cover, and wood preservation, different components of the record are preserved in different places, and many sites need to be examined to obtain a relatively complete regional history.

Conclusions

Subsequent to a minor advance possibly latest Pleistocene in age, the Bear River Glacier advanced out onto the pass and into the river valley that shares its name at least four times in the last half of the Holocene. Three of these advances were closely spaced in time and show that rapid, small-scale fluctuations were superimposed on general late-Holocene expansion. Minor variations in summer temperature, as well as changes in the position of the jet stream and strength of the Aleutian Low in response to the Pacific Decadal Oscillation, may have been factors influencing the region’s small-scale glacial oscillations. Dated wood fragments show that Neoglaciation began earlier in the northern Coast Mountains than previously known, but in general the glacial chronology preserved in stacked tills at this site is synchronous with those established elsewhere in British Columbia.

Footnotes

Acknowledgements

We would like to thank Jason Hendrick and Laura Exley for field assistance, and the residents of Stewart BC, in particular C Caruso and Ian MacLeod, for their help in obtaining historical documents and for providing personal accounts of the Bear River Glacier’s activity in the last century. Helpful comments on the manuscript were provided by two anonymous referees.

Funding

This work was supported by the National Science and Engineering Research Council of Canada (grant 9026 to Osborn and 130040 to Spooner).

References

Allen

Smith

(2007) Late Holocene glacial activity of Bridge Glacier, British Columbia Coast Mountains. Canadian Journal of Earth Sciences 44: 1753–1773.

Arsenault

Clague

Mathewes

(2007) Late Holocene vegetation and climate change at Moraine Bog, Tiedemann Glacier, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences 44: 707–719.

Barnes

(2003) Investigations of interdecadal climate variability in northwestern North America, Little Ice Age to present. PhD thesis, Queen’s University, Kingston Ontario.

Buddington

(1926) Mineral investigations in southeastern Alaska. U.S. Geological Survey Bulletin 783-B: 41–62.

Cashman

Clark

Clague

. (2002) New constraints on Holocene glaciation in Garibaldi Provincial Park, British Columbia, Canada. Geological Society of America Programs with Abstracts 34: A-78.

Clague

Mathewes

(1996) Neoglaciation, glacier-dammed lakes, and vegetation change in northwestern British Columbia, Canada. Arctic and Alpine Research 28: 10–24.

Clague

Mathews

(1992) The sedimentary record and Neoglacial history of Tide Lake, northwestern British Columbia. Canadian Journal of Earth Sciences 29: 2383–2396.

Clague

Menounos

Osborn

. (2009) Nomenclature and resolution in Holocene glacial chronologies. Quaternary Science Reviews 28: 2231–2238.

Denton

Karlen

(1973) Holocene climatic variations – Their pattern and possible cause. Quaternary Research 3: 155–205.

10.

Desloges

Ryder

(1990) Neoglacial history of the Coast Mountains near Bella Coola, British Columbia. Canadian Journal of Earth Sciences 27: 281–290.

11.

Environment Canada (2010) <http://www.climate.weatheroffice.gc.ca/climate_normals/>.

12.

Friele

Clague

(2002) Younger Dryas readvance in Squamish River valley, southern Coast Mountains, British Columbia. Quaternary Science Reviews 21: 1925–1933.

13.

Gilbert

Desloges

(1987) Sediments of ice-dammed, self-draining Ape Lake, British Columbia. Canadian Journal of Earth Sciences 24: 1735–1747.

14.

Greig

Anderson

Daubeny

. (1994) Geology of the Cambria Icefield; Stewart (103P/13), Bear River (104A/4), and Parts of Meziadin Lake (104A/3). Geological Survey of Canada Open File Report 2931.

15.

Grove

(1971) Geology and mineral deposits of the Stewart area, northwestern British Columbia. British Columbia Department of Mines and Petroleum Resources Bulletin 58, 240 pp.

16.

Hanson

(1934) Bear River delta, British Columbia, and its significance regarding Pleistocene and recent glaciation. Transactions of the Royal Society of Canada 28: 179–185.

17.

Hebda

(1995) British Columbia vegetation and climate history with forcus on 6 ka BP. Géographie physique et Quaternaire 49: 55–79.

18.

Heusser

Peteet

(1985) Late-Quaternary climatic change on the American North Pacific Coast. Nature 315: 485–487.

19.

Jackson

Laxton

Smith

(2008) Dendroglaciological evidence for Holocene glacial advances in the Todd Icefield area, northern British Columbia Coast Mountains. Canadian Journal of Earth Sciences 45: 83–98.

20.

Koch

Clague

Osborn

(2007) Glacier fluctuations during the past millennium in Garibaldi Provincial Park, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences 44: 1215–1233.

21.

Lewis

Smith

( 2005) Late Holocene glacier activity at Forrest Kerr Glacier, Andrei Icefield, northern British Columbia Coast Mountains (abstract). Canadian Geophysical Union Annual Meeting, Banff, Alberta, May 2005, http://cgrg.geog.uvic.ca/abstracts/LewisLateGlacier.html

22.

Luckman

(1986) Recontruction of Little Ice Age events in the Canadian Rockies. Geographie physique et Quaternaire 40: 17–28.

23.

Luckman

(1996) Reconciling the glacial and dendrochronological records of the last millennium in the Canadian Rockies. In: Jones

Bradley

Jouzel

(eds) Climatic Variations and Forcing Mechanisms of the Last 2000 Years. NATO ASI Series I, 41: 85–108.

24.

Luckman

(2000) The Little Ice Age in the Canadian Rockies. Geomorphology 32: 357–384.

25.

Luckman

(2006) The Neoglacial history of Peyto Glacier. In: Demuth

Munro

Young

(eds) Peyto Glacier: One Century of Science. National Hydrology Research Institute Science Report 8, pp. 25–57.

26.

Luckman

Villalba

(2000) Assessing the synchroneity of glacier fluctuations in the Western Cordillera of the Americas during the last millennium. In: Markgraf

(ed.) Interhemispheric Climate Linkages. Academic Press, pp. 119–140.

27.

Luckman

Holdsworth

Osborn

(1993) Neoglacial glacier fluctuations in the Canadian Rockies. Quaternary Research 39: 144–153.

28.

McConnell

(1913) Portions of the Portland Canal and Skeena Mining Divisions, Skeena District, British Columbia. Geological Survey of Canada Memoir 32, 101 pp.

29.

Marcus

(1960) Periodic drainage of glacier-dammed Tulsequah Lake, British Columbia. Geographical Review 50: 89–106.

30.

Mathews

(1965) Two self-dumping ice-dammed lakes in British Columbia. Geographical Review 55: 46–52.

31.

Menounos

Clague

Osborn

. (2008) Western Canadian glaciers advance in concert with climate change circa 4.2 ka. Geophysical Research Letters 35: L07501.

32.

Menounos

Koch

Osborn

. (2004) Early Holocene glacier advance, southern Coast Mountains, British Columbia, Canada. Quaternary Science Reviews 23: 1543–1550.

33.

Menounos

Osborn

Clague

. (2009) Latest Pleistocene and Holocene glacier fluctuations in western Canada. Quaternary Science Reviews 28: 2049–2074.

34.

Minkus

(2006) Evidence of latest Pleistocene glacier advance in the southern Coast Mountains, British Columbia, Canada. M.Sc. thesis, University of Calgary.

35.

Nesje

Johannessen

(1992) What were the primary forcing mechanisms of high-frequency Holocene climate and glacier variations? The Holocene 2: 79–84.

36.

Osborn

Luckman

(1988) Holocene glacier fluctuations in the Canadian Cordillera (Alberta and British Columbia). Quaternary Science Reviews 7: 115–128.

37.

Osborn

Menounos

Koch

. (2007) Multi-proxy record of Holocene glacial history of the Spearhead and Fitzsimmons ranges, southern Coast Mountains, British Columbia. Quaternary Science Reviews 26: 479–493.

38.

Osborn

Robinson

Luckman

( 2001) Holocene and latest Pleistocene fluctuations of Stutfield Glacier, Canadian Rockies. Canadian Journal of Earth Sciences 38: 1141–1155.

39.

Province of British Columbia (2011) ‘Bear Glacier Provincial Park- BC Parks’ < http://www.env.gov.bc.ca/bcparks/explore/parkpgs/bear_gl/>.

40.

Reasoner

Osborn

Rutter

(1994) Age of the Crowfoot advance in the Canadian Rocky Mountains: A glacial event coeval with the Younger Dryas oscillation. Geology 22: 439–442.

41.

Reyes

Clague

(2004) Stratigraphic evidence for mulitiple Holocene advances of Lillooet Glacier, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences 41: 903–918.

42.

Reyes

Wiles

Smith

. (2006) Expansion of alpine glaciers in Pacific North America in the first millennium A.D. Geology 34: 57–60.

43.

Ryder

(1987) Neoglacial history of the Stikine-Iskut area, northern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences 24: 1294–1301.

44.

Ryder

Thompson

(1986) Neoglaciation in the southern Coast Mountians of British Columbia: Chronology prior to the late Neoglacial maximum. Canadian Journal of Earth Sciences 23: 273–287.

45.

Walker

Pellatt

(2003) Climate change in coastal British Columbia – A paleoenvironmental perspective. Canadian Water Resources Journal 28: 531–566.