midge-inferred holocene climate history southern · moist air masses from the pacific coast are...

The Holocene 14,2 (2004) pp. 258-271

Midge-inferred Holocene climate historyof two subalpine lakes in southern British

Columbia, Canada

Sandra M. Rosenberg,l 2* Ian R. Walker,1'2 Rolf W. Mathewes1and Douglas J. Hallett"'3('Department of Biological Sciences, Simon Fraser University, 8888 UniversityDrive, Burnaby, British Columbia, Canada, V5A ]S6; 2Department of Biology,Okanagan University College, 3333 College Way, Kelowna, British Columbia,Canada, VIV 1V7; 3Center for Environmental Sciences and Quaternary SciencesProgram, PO Box 5694, Northern Arizona University, Flagstaff AZ 86011,USA)

Received 16 July 2002; revised manuscript accepted 13 February 2003

HOLOCENERESEARCHPAPER

Abstract: To investigate postglacial environmental changes in both the coastal and interior wet belts of BritishColumbia, fossil midges were analysed from two subalpine lakes, one adjacent to the lower Fraser canyon(Frozen Lake), and the other in Mount Revelstoke National Park (Eagle Lake). The midge stratigraphy forFrozen Lake revealed an abundance of rheophilous chironomid taxa and Simuliidae larvae, reflecting the pres-ence of an inflowing stream. An abundance of Chaoborus mandibles and Microtendipes during the early Holo-cene (c. 10100-7700 'C years BP, c. 11500-8500 cal. years BP) suggests warmer temperatures. A subsequentdecline in the warm indicators and relative increases in cold stenotherms (Heterotrissocladius and Diamesa)indicate cooling until present day. This climate reconstruction is consistent with other quantitative and qualitat-ive evidence for past climatic change in southern British Columbia. At Eagle Lake the warm indicators, Dicrot-endipes and Polypedilum, are seen in the early Holocene (c. 8500-6730 1-C years BP, c. 9600-7600 cal. yearsBP), but are absent during the mid-Holocene when cooler temperatures probably prevailed. In the late Holocene(c. 3800 'dC years BP to present, c. 4200 cal. years BP to present) there is a resurgence of warm indicators,which contrasts with the evidence of continued cooling typically seen in reconstructions of southern BritishColumbia summer temperatures. The Eagle Lake record therefore appears to be anomalous. Multiproxy andmultisite investigations are needed to reconstruct Holocene climatic changes more reliably.

Key words: Chironomidae, midges, palaeoclimate, palaeolimnology, temperature reconstruction, climaticchange, British Columbia, Holocene.

Introduction

Palaeoecological techniques are powerful tools for reconstructingpast environments, and can provide long and detailed records ofpast climatic changes. They thus provide a window on naturalclimatic variations and on the role of climate in shaping terrestrialecosystems (Mathewes, 1985; Hebda, 1995; Walker and Pellatt,2001; 2004). Chironomids (Order Diptera) have become importantfor palaeoclimatic studies because distinctive assemblages ofmidges are good indicators of present and past environmental con-

*Author for correspondence. Present address: Biology Department, Langara College,100 West 49th Avenue, Vancouver, BC, Canada V5Y 2Z6 (e-mail:[email protected])

Arnold 2004

ditions (e.g., Walker, 1987; Hofmann, 1988; Walker andMathewes, 1989; Walker et al., 1995; Lotter et al., 1997). Midgeassemblages have been used to reconstruct hypolimnetic oxygencontent (Quinlan et al., 1998), acidity (Brodin and Gransberg,1993; Schnell and Willassen, 1996), salinity (Walker et al., 1995;Heinrichs et al., 1997), trophic status (Lotter et al., 1998; Littleet al., 2000; Brooks et al., 2001) and, as is the focus of this paper,temperature (Walker and Mathewes, 1989; Walker, 1991; Lotteret al., 1999; Walker et al., 1997; Olander et al., 1999).

Several studies have used fossil midges to reconstruct past cli-matic oscillations (i.e., the Killarney and Younger Dryas events)in Atlantic Canada (i.e., Walker et al., 1991 a; 199 lb; Levesqueet al., 1993a; 1993b; Wilson et al., 1993; Cwynar and Levesque,1995). In southern British Columbia, however, palaeoclimato-

10.1 191/0959683604h1703rp

© 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. at University of Victoria on May 1, 2007 http://hol.sagepub.comDownloaded from

http://hol.sagepub.com

Sandra M. Rosenberg et al.: Midge-inferred Holocene climate history from British Columbia, Canada 259

logists have until recently relied almost entirely on palaeobotan-ical studies to obtain palaeoclimate records (e.g., Hansen, 1955;Alley, 1976; Hazell, 1979; Mathewes and Heusser, 1981; Cawker,1983; Mathewes, 1985; Wainman and Mathewes, 1987; Reasonerand Hickman, 1989; Mathewes and King, 1989; Hebda, 1995;Pellatt et al., 1998; 2000; 2001; Allen et al., 1999; Heinrichset al., 1999; Hallett and Walker, 2000; Bennett et al., 2001). Rela-tively few midge studies (Walker and Mathewes, 1987; 1989;Heinrichs et al., 1997; Smith, 1997; Palmer et al., 2002; Rucket al., 1998; Smith et al., 1998) have been carried out in BritishColumbia. Since midges have short life cycles, they potentiallyrespond more quickly to climatic changes than vegetation(especially forest trees) and therefore are probably more sensitiveindicators of past climatic change than pollen, spores or plantmacrofossils (Palmer, 1998). Battarbee (2000) recently concludedthat midges are the most promising biological indicators of pastclimatic changes.

Although midges have proven to be useful indicators of late-glacial climatic changes, they have been used less often for quanti-tative Holocene climate reconstructions. The statistical errorsassociated with midge-palaeotemperatur inference models sug-gest that midges may be less well suited for the reconstruction ofsubtle, Holocene climatic changes (Lotter et al., 1999; Heiri et al.,2003; Battarbee et al., 2002). To maximize the climate signal-to-noise ratio, study sites little affected by non-climatic impactsshould be selected for midge analyses (Battarbee et al., 2002).Thus, for Holocene temperature reconstructions, we select small,shallow lakes near treeline, and avoid saline lakes, eutrophiclakes, lakes with unusually low pH and lakes which have hadmuch human impact. In this paper we use midge fossils to exam-ine postglacial environmental changes at two subalpine lakes inBritish Columbia, one in the southern Coast Mountains and thesecond in Mt Revelstoke National Park. We also critically assesstheir potential for reconstructing Holocene climate trends.

Study area

For the purposes of Holocene climate reconstruction, two smallsubalpine lakes were chosen because treeline environments areknown to be sites of high climate sensitivity (Luckman and Kear-ney, 1986; Clague and Mathewes, 1989; Walker and MacDonald,1995; Pellatt and Mathewes, 1997; Pellatt et al., 1998; 2000; Bat-tarbee, 2000). Lower Frozen Lake (49°36'N, 121°28'W) is asmall subalpine lake located in British Columbia's Coast Moun-tains, on the slopes of the lower Fraser Canyon (Figure 1). At1180 m a.s.l., the lake occupies a basin in the forested subzoneof the cool and moist Mountain Hemlock biogeoclimatic zone.The dominant trees in this zone are mountain hemlock (Tsugamertensiana (Bong.) Carr.), Pacific silver fir (Abies amabilisDougl.) and yellow cedar (Chamaecyparis nootkatensis D. Don)(Brooke et al., 1970). Frozen Lake is approximately 3 ha in areawith a maximum depth of 17 m and a bedrock sill controlling theoutlet. A small stream flows from Upper Frozen Lake through anopen meadow to the lower lake (Hallett et al., 2003). This inletdelivers a large volume of water to the lake during spring andsummer melt.Snow in the Mountain Hemlock zone begins to accumulate in

October and deep snowpacks can persist until July. The late win-ter snowpack may be up to 3 m thick (Brooke et al., 1970; Hallettet al., 2003). Climate data at nearby Hell's Gate (49°47'N,121°27'W, elevation 122 m) indicates a mean annual temperatureof 9.4°C and a mean annual precipitation of 1229 mm yr-'(Environment Canada, 1993). It should be noted that daily meantemperatures and mean annual precipitation at Frozen Lake willdiffer considerably from the Hell's Gate station due to thedifference in elevation between the two sites. Assuming an

49' 36'

49' 35* O

121' 29' 121' 28' 121° 27'

Figure 1 Map indicating location of FrozenColumbia.

elevation ontourlakes, riversrivers

- hghway* city/own

Scalem 0 500 1000

I

Lake in southern British

- elevation contourlakes, rivers

N creekstI _ ._Iroads

a city/town

Scalem O 500 t OQO

Figure 2 Map of Mount Revelstoke National Park region, indicating thelocation of Eagle Lake.

environmental lapse rate of 0.6°C/100 m (Livingstone etal.,1999), the mean annual temperature and the mean July air tem-perature at Frozen Lake are expected to be about 2.8°C and13.9°C, respectively. The actual precipitation at Frozen Lakecould be >2X that recorded at the Hell's Gate station.

Eagle Lake (51°3'N, 118°10'W) is located in Mount Revel-stoke National Park, in the Selkirk Mountains (Figure 2). The



260 The Holocene 14 (2004)

Selkirk Mountains, together with the Monashee Range to the westand Purcell Range to the east, compose the Columbia Mountains,a system of north-south trending ranges with bedrock consistingof a complex of metamorphic and igneous rocks. The ColumbiaMountains are bordered on the west by the Interior Plateau andon the east by the Rocky Mountains.The Columbia Mountains form the interior wet belt of southern

British Columbia. Moist air masses from the Pacific coast areresponsible for the mean annual precipitation of 950.2 mm yr- Irecorded at the Revelstoke climate station, 443 m a.s.l.(Environment Canada, 1993). The mean annual temperature atRevelstoke is 6.7°C. Eagle Lake, however, is located at 1845 ma.s.l., well above the Revelstoke climate station; thus, the studysite is correspondingly cooler and wetter. Assuming an environ-mental lapse rate of 0.6°C/100 m, the estimated mean annual tem-perature at Eagle Lake is -1.7°C. Mean July air temperature isestimated at 9.8°C. A temperature record is also available for mostof the past decade from a BC Hydro climate station located onMount Revelstoke (51°2'N', 1 18°8'W, 1830 m a.s.l.; BC Hydro,personal communication). The record is much shorter than thatused in calculating climate normals, but indicates a mean July airtemperature of I 1.1°C. According to Coupe et al. (1991), up to2200 mm of annual precipitation might be expected at Eagle Lakewith 50-70% as snow, leading to a maximum snowpack of nearly4 m. At this elevation, the snowpack may persist from mid-October through to mid-July. In 1999, an exceptionally wet year,extensive snow patches were still evident at Eagle Lake in mid-August (Rosenberg et al., 2003).

Eagle Lake has a maximum depth of 1.5 m, an area of 0.3 haand is located 1845 m a.s.l. (Donald and Alger, 1984). The lowerelevations of Mount Revelstoke support forests of western hem-lock (Tsuga heterophylla (Raf.) Sarg.) and western red cedar(Thuja plicata Donn), but vegetation at Eagle Lake represents ahigher-elevation community and is classified as the ESSFvc (verywet cold Engelmann Spruce-Subalpine Fir) biogeoclimaticsubzone (Coupe et al., 1991). The climax trees of this regioninclude Engelmann spruce (Picea engelmannii Parry) and subalp-ine fir (Abies lasiocarpa (Hook.) Nutt.), with a small componentof mountain hemlock (Tsuga mertensiana (Bong.) Carr). EagleLake lies near the transition from continuous forest to subalpineparkland. Many small openings are present among the trees sur-rounding the site.

Methods

Field and laboratory methodsIn 1997, a 300 cm core from the deepest part (17 m) of FrozenLake was obtained using a lightweight percussion corer(Reasoner, 1993). The core was then transported to Simon FraserUniversity where it was subsequently stored in a cold room at40C.The sediment stratigraphy consists of dark brown gyttja grading

into greyish brown clay at 238 cm depth and to grey clay nearthe core base at 240 cm. Tephra layers were found at 163-155cm, 127 cm and 60 cm (Hallett et al., 2003). For midge analysisthe core was subsampled every 5 cm from 240 to 0 cm, to give49 sediment subsamples of 2 cm3. The intervals from 320-240cm had substantially lower than 50 chironomid head capsules andwere sometimes barren (Table 1). These samples are consideredtoo small to be statistically useful (Palmer, 1998; Heiri and Lotter,2001; Larocque, 2001; Quinlan and Smol, 2001).A 110 cm core was removed from the deepest part of Eagle

Lake on 9 August 1999 in 1.5 m of water. A modified Kajak-Brinkhurst surface corer, equipped with a Glew trigger mech-anism (Glew, 1989), was used to capture the uppermost 16 cm ofthe core. The sediment below 16 cm was collected using a 5 cm

Table 1 Frozen Lake samples containing fewer than 50 chironomidhead capsules

Depth Taxa Number of(cm) head capsules

245 CricotopusJOrthocladius 0.5Corynocera nr. ambigua 1Heterotrissocladius grimshawi type 1Procladius 2Protanypus 0.5Tanytarsina (undifferentiated) 1(Ephemeroptera mandibles) (2)(Unidentifiable) (3)

250 Tanytarsina (unidifferentiated) 1270 Diamesa 1

Psectrocladius 0.5285 Diamesa 1

Pseudodiamesa 1290 Micropsectra atrofasciata type 2

diameter modified Livingstone piston corer (Wright, 1967). Thetwo core records were matched up based on the stratigraphiclengths of each of the cores. The sediment stratigraphy consistsof partially laminated gyttja with a thick tephra layer at 96-81cm, a thinner tephra at 36-35 cm, and small amounts of woodydebris throughout the core. The core was subsampled every 1 cmfrom 1 10 to 0 cm and subsequently stored in a cold room at SimonFraser University at 4°C. Subsampling (58 levels) of half cubiccentimetre sediment samples was needed to give greater than 50head capsules per interval.

Processing of each sediment sample involved head capsule iso-lation techniques outlined by Walker (1987; 2001) and Walkeretal. (199la). Treatment with 10% HCI was used to removecarbonates from the moist sediment, then 5% KOH was added toeach sample and heated to approximately 40-45°C for 5-6minutes to complete deflocculation. The sediment was sub-sequently sieved on a 95 ,um Nitex® mesh. The retained residuewas backwashed into a beaker with distilled water. Individualhead capsules were picked from a Bogorov counting tray undera dissecting microscope at X 10-25 magnification. The midgeremains were transferred to drops of water on a coverslip usingDumont #4 forceps. The coverslips were allowed to air-dry, andthen mounted on glass slides using Entellan® mounting medium.

Dipteran remains were identified at X 100-400 magnificationusing an Olympus BH-2 compound microscope. Identificationswere based on descriptions and keys by Oliver and Roussel(1983), Wiederholm (1983), Walker (1988), Uutala (1990), anextensive photograph (i.e., reference) collection and the WWWfield guide to subfossil midges (Walker, 1996-2000). If the headcapsules retained more than half of a complete mentum, they werecounted as one head capsule. If they had exactly half of a mentum,they were counted as one half. Head capsules retaining less thanhalf of the mentum were not counted.

Statistical analysisTILIA version 2.0.b.4 (Grimm, 1993) was used to collate the rawdata and generate a midge percentage diagram in TILIA-GRAPHversion 2.0.b.5 (Grimm, 1991). The program CONISS was usedto perform a stratigraphically constrained incremental sum-of-squares cluster analysis (Grimm, 1987). This allowed intervalswith major changes in midge communities to be recognizedthroughout the cores. The significance of the zones was assessedfollowing the broken-stick procedure outlined by Bennett (1996;2002) and implemented in the computer program PSIMPOLLversion 4.10. As suggested by Bennett, we designated significantboundaries as separating zones.



Sandra M. Rosenberg et a!.: Midge-inferred Holocene climate history from British Columbia, Canada 261

A midge-temperature weighted averaging inference model,based on surface sample data from 51 British Columbia lakes, wasdeveloped by Palmer et al. (2002). In addition to Chironomidae,Chaoborus were used in the model, and in other statistical analy-ses. Ceratopogonidae and Simuliidae were too rare to be included.To infer mean July air temperatures at our study sites, we used aslightly modified version of this model incorporating higher taxo-nomic resolution. Although midges probably respond moreclosely to changes in water temperature, we reconstructed airtemperature because water-temperature data were not available.Air temperatures are strongly correlated with lake surface-watertemperature and are of greater interest to palaeoclimatologiststhanwater temperatures (Livingstone and Lotter, 1998).

This modified model was developed using CALIBRATE ver-sion 0.70 (Juggins, 1997) and WA-PLS version 1.1 (Juggins andter Braak, 1996). The model revealed a significant relationshipbetween mean July air temperatures and midge assemblages, witha jack-knifed r2 of 0.73 and a jack-knifed root mean squared errorof prediction (RMSEP) of 1 .87°C. As a further assessment of thepotential usefulness of this model, we performed a partial con-strained CCA of the square-root transformed midge data, with thefirst axis constrained to represent mean July air temperature. Allother axes were unconstrained. The resulting eigenvalues for thefirst (XI) and second (X2) axes are 0.303 and 0.282, respectively,yielding X11X2 of 1.07. This high ratio for the first and secondeigenvalues suggests that temperature explains a high proportionof the variance in the species data and lends confidence to thetemperature inference model.

Using the squared chord distance as a dissimilarity coefficient,we used the computer program ANALOG (H.J.B. Birks and J.M.Line, unpublished software) to determine if adequate analoguesexisted in our surface sample data for each interval in the core(Overpeck etal., 1985; Laing et al., 1999). Those core intervalsexceeding the 95% confidence interval were assumed to have noanalogues, whereas those samples lying between the 75 and 95%confidence intervals were considered to have poor analogues.Samples within the 75% confidence interval were considered tohave good analogues in the surface sample training set (Laingetal., 1999).To further assess the temperature reconstructions, correspon-

dence analyses (CA) were performed on the fossil data. Thestrength of the correlation between the CA axis 1 sample scoresand the inferred temperature values indicates to what extent theprincipal changes in midge community structure can be explainedby the temperature inferences (Laird et al., 1998).

Results

Frozen Lake: chronology and stratigraphyCALIB 4.3 (Stuiver et al., 1998) was used to convert radiocarbon('4C) ages to calibrated years. Extensive AMS radiocarbon datingof plant macrofossils provides the chronology for the Frozen Lakecore (Table 2). A basal date of 10020 ± 50 "'C years BP(c. 11500 cal. years BP) indicates that the midge samples spanthe Holocene. Three ash layers were also found in Frozen Lake(Table 2). The lowermost tephra (163-155 cm) was identified asMt Mazama ash (6730 ± 40 "'C years BP, c. 7600 cal. years BP;Hallett etal., 1997). The second tephra (127 cm) was found bysieving (Hallett etal., 2001) and subsequently identified usingelectron microprobe analysis as being from a Glacier Peak erup-tion (5000-5080 "4C years BP, c. 5800 cal. years BP). The thirdash at 60 cm was identified as Bridge River tephra, giving anadditional date of 2435 ± 26 "'C years BP (c. 2500 cal. years BP;Clague et al., 1995).

Figure 3 depicts the percentage of total identifiable Chironomi-dae, non-chironomiddipteran remains (i.e., members of the family

Chaoboridae, family Ceratopogonidae and family Simuliidae), aswell as Ephemeroptera. For taxa where estimates of temperatureoptima have been determined, the midge taxa are arranged fromcold-tolerant to warm-adapted. Stratigraphies for the remainingchironomids are appended to the end of the diagram.

Fifty-one chironomid taxa were identified from the Frozen Lakesediments. The computer programs CONISS (Grimm, 1987) andPSIMPOLL (Bennett, 2002) indicated two significant zones in theFrozen Lake core (Table 3).

Midge assemblaqe zone 1 (FRal, 240-153 cm,c. 10100-6500 4C years BP, c. 11500-7500 cal. yearsBP)Members of the Subtribe Tanytarsina are the dominant midge taxathroughout the core including zone FRO-1 (Figure 3). Other taxacommon in zone FRO-1 include Microtendipes, Chironomus,Tribe Pentaneurini, Heterotrissocladius Sergentia, Limnophyes,Corynoneura & Thienemanniella, Rheocricotopus, Psectrocladiusand Chaoborus. Chaoborus mandibles (mainly Chaoborustrivitattus) are abundant throughout the zone and are often morenumerous than the remains of all other midges combined. Chiron-omid taxa which are found sporadically and at low abundance(<2%) include Tanytarsus chinyensis, Parochlus, Cladopelma,Cryptochironomus, Endochironomus, Pagastiella, Phaenopsectra,Stictochironomus, Hydrobaenus, Potthastia and Protanypus.Psectrocladius (Allo. or Meso.) is present only in zone FRO-1.Ephemeroptera mandibles were abundant in early zone FRO-1sediments but become less common throughout the zone (Figure3). Microtendipes decreases towards the top of FRO-1. Slightincreases are apparent in Corynoneura & Thienemanniella (to15%) and Eukiefferiella & Tvetenia abundance (to 8%) towardsthe top of this zone.

Midge assemblage zone 2 (FRa2, 153-0 cm, c. 6500 14CBP to present, c. 7500 cal. years BP to present)A decrease in the average abundance of Chironomini, andespecially Chaoborus mandibles, characterizes zone FRO-2. TheOrthocladiinae increase slightly and Microtendipes is now rare.Sergentia and Heterotrissocladiuscontinue to be present through-out this zone, whereas another cold stenotherm, Diamesa, makesits first appearance late in zone FRO-2 but remains rare. AlthoughRheocricotopus is less abundant than deeper in the core, thereare slight increases in most rheophilous taxa includingCorynoneura & Thienemanniella, Doithrix & PseudorthocladiusEukiefferiella & Tvetenia and Limnophyes. Members of the familyCeratopogonidae are uncommon although they occur more fre-quently towards the core top. Heterotrissocladius marcidus typealso increases in abundance towards the top of this zone.

Temperature inferences and analogue comparisonsMarked fluctuations are evident in the reconstructed temperaturesfor Frozen Lake (Figure 4). In FRO-1, inferred mean July airtemperature appears to gradually decrease with a maximuminferred temperature of 15.6 + 2.1°C at 235 cm and a minimuminferred temperature of 13.1 + 2.0 at 170 cm. Reconstructed tem-peratures remain cool, ranging from 11.6 ± 2.0 to 14.2 + 2.0°C,in the second zone. Although good analogues are available formost of the FRO-1 midge assemblages, only poor analogues areavailable for most of the FRO-2 midge communities (Figure 5).A high correlation (r = 0.85, p < 0.01) is evident between theCA axis 1 sample scores and the temperature inferences. Thisindicates that the temperature inferences can explain much of thevariability in the Holocene midge communities.

Eagle Lake: chronology and stratigraphyTwo Picea and one Abies needle from 106-105 cm provided anAMS date of 8330 ± 40 "'C years BP (c. 9400 cal. years BP)




Table 2 AMS radiocarbon and tephra dates for the Frozen Lake sediment core

Depth (cm) Lab. number Uncalibrated age Calibrated age* Relative area under Material dated(CAMS) (lC years BP) (cal. years BP) probability distribution

26-25 45980 1570 ± 60 1590-1580 0.009 conifer needle (Abies amabilis)1570-1330 0.9861320-1310 0.004

36-35 45981 1950 + 40 1990-1960 0.093 conifer needle (Tsuga mertensiana)1950-1820 0.907

61-60 2435 ± 26 2710-2630 0.271 Bridge River tephra2620-2590 0.0502540-2530 0.0152500-2350 0.664

90-89 45983 3560 ± 40 3970-3950 0.061 conifer needle (Tsuga mertensiana)3930-3800 0.6473800-3720 0.292

120-119 45984 4530 ± 50 5320-5030 0.974 conifer needle (Tsuga mertensiana)5010-4980 0.026

148-147 45985 6170 + 40 7220-7220 0.001 twig fragment7210-7170 0.0927160-6940 0.9036920-6910 0.004

163-155 6730 ± 40 7670-7560 0.852 Mazama tephra7540-7510 0.148

189-188 45986 8180 ± 50 9400-9390 0.006 conifer needle (Abies amabilis)9370-9360 0.0019280-9010 0.993

222-221 45987 9390 + 70 11060-11020 0.030 twig fragment11010-10960 0.03410840-10830 0.00110770-10400 0.92910340-10330 0.00110320-10310 0.00210300-10290 0.003

236-235 45988 10020 ± 50 11920-11860 0.035 conifer needle11700-11260 0.965 (Abies lasiocarpa)

*Based on Stuiver et al. (1998).

from near the base of the Eagle Lake core. Therefore Eagle Lakedoes not contain a complete Holocene sequence. The sediment-accumulation rate is estimated by linear regression and the basalage is estimated by extrapolation because no macrofossils werefound beneath the 106-105 cm interval.Two tephra layers are also identified in the Eagle Lake core

(Table 4). Dr F.F. Foit Jr at the Microbeam Facility, WashingtonState University, analysed the glass chemistry of each tephra. Thelowermost tephra, from 96-81 cm, is identified as Mt Mazamaash (6730 + 40 14C years BP; c. 7600 cal. years BP; Hallett et al.,1997). The second tephra, at 36-35 cm, is the Mt St Helens Yntephra, providing an approximate date of 3400 '4C yr BP (c. 3700cal. years BP; Luckman et al., 1986; Mullineaux, 1986).

Twenty-eight chironomid taxa were identified from the EagleLake core (Figure 6). Five distinct midge assemblage zones areidentified in this core by CONISS (Table 5). All zone boundariesare significant as assessed in PSIMPOLL (Bennett, 2002). Weignored a sixth significant zone boundary separating the sampleabove 1.5 cm from the sample immediately below, because wedid not feel it useful to include two zones consisting of onesample each.

Midge assemblage zone 1 (EAG-1, 110 cm - Mazama ash,c. 8500-6730 C years BP, c. 9600-7600 cal. years BP)The basal sediments of the Eagle Lake core are dominated byDicrotendipes, Polypedilum and Tanytarsina (Figure 6). The Chi-ronomini are more abundant in zone EAG-1 than elsewhere in thecore. Other common taxa, including Procladius, Psectrocladiusand Chaoborus trivittatus, are found throughout the zone. Someof the rarer taxa in zone EAG-1 include Sergentia, Corynoneura& Thienemanniella, Limnophyes, Acari and Ceratopogonidae(Bezzia type).

Midge assembqlae zone 2 (EAG-2, Mazama ash - 47 cm,c. 6730-3800 C years BP, c. 7600-4200 cal. years BP)There is a large increase in Tanytarsina and a pronounceddecrease in the Chironomini in EAG-2. Tanytarsina andDicrotendipes are the dominant taxa. Polypedilum is also commonin this zone but its abundance fluctuates greatly with two distinctpeaks at 75 and 65 cm. The maximum of Microtendipes and firstappearance of Endochironomus and Tribelos also occur.Cold-stenothermous taxa (e.g., Sergentia, Stictochironomus andHeterotrissocladius are rare in the Eagle Lake core, but are more




U)LI

0

.7.~~~~~~~~~~~

(C,~~~~~~~~~~~~~~~~~~~~~~~~~~~~C

~~ ~ ~ ~ ~ l

~~~~~ ~ ~~~C

~ ~ ~ A.0~~~~A A

A LA

II -." -If 4"' L

I I~~~~~~~7

4l'-.14%>

IIJ/ -

- -If/Al

E*

lI/L A/

N-I

% -

4-q,

a 0 N 0 8~~~~~~~~~~~~c NYNO ~~~~~~~~~~(W1)t4ldaG

4,c9 a

4, r'--

In&

.0

NZ

7)00

el

7)0

0

)

O 73

0 0 N0

0

0 g

(7) -00"




Table 3 Zonation of the Frozen Lake chironomidrounded to the nearest century)

stratigraphy (ages

Chironomid Depth Uncalibrated age Calibrated age*assemblage zone (cm) (radiocarbon (calibrated

years BP) years BP)

FRO-2 0-153 Present - 6500 Present - 7300FRO-1 153-240 6500-10100 7300-11500


common in zone EAG-2 than elsewhere. Among the rheophiloustaxa Limnophyes is the most abundant. Tanytarsus lugens + Cory-nocera oliveri and Monopsectrocladius also peak in zone EAG-2, coinciding with the disappearance of Psectrocladius (Allo. orMeso.).

Midge assemblage zone 3 (EAG-3, 47-29 cm, c. 3800-2400 14C years BP, c. 4200-2400 cal. years BP)A resurgence of the dominant Chironomini, most notably Dicrot-endipes and Polypedilum, occurs in conjunction with a decreaseof Tanytarsina. Microtendipes and Glyptotendipes both decline.

Midge assemblage zone 4 (EAG4, 29-3 cm, c. 2400-40014C years BP, c. 2400-500 cal. years BP)Although remains of Psectrocladius (Allo. or Meso.) type are rarein EAG-4, other Psectrocladius remains increase markedly at thebeginning of this zone. Among the less common taxa, the declinesof Microtendipes and Limnophyes and an increase in Corynoneura &Thienemanniella are most apparent.

Midge assemblage zone 5 (EAG-5, 3-0 cm, c. 400 14Cyears BP to present, c. 500 cal. years BP to present)Increases of Tanytarsina, Polypedilum and the Tribe Pentaneuriniare noted at the top of the core. Dicrotendipes and other Chirono-mini decrease in EAG-5. Although the rheophilous taxon Cory-noneura & Thienemanniella occurs in EAG-5, there are no cold-stenothermous taxa present.

Temperature inferences and analogue comparisonsThe temperature reconstruction for Eagle Lake reveals substantialfluctuations in Holocene air temperature ranging from 11.6 ± 2.0to 16.6 + 2.2°C (Figure 7). In EAG-1, inferred mean July airtemperatures range from 13.7 ± 2.0 to 16.6 ± 2.2°C. In thesecond zone, EAG-2, the inferred temperatures are mostly cooler,ranging from 11.6 ± 2.0 to 14.6 ± 2.1°C (the high temperatureof 15.4 ± 2.1°C is excluded from this zone as it corresponds witha poor analogue, Figure 8). Higher temperatures are inferred forEAG-3, 4 and 5 ranging between 14.4 ± 2.2 and 16.2 ± 2.20C.Overall, the reconstructed values suggest little temperature differ-ence between the early and late Holocene, but indicate a pro-longed cool interval in the mid-Holocene. Further analysis (Figure8) reveals that the surface sample training set contained goodanalogues for most mid-Holocene assemblages. The available ana-logues for the early- and late-Holocene assemblages are typicallypoor. A very strong correlation (r = 0.99, p < 0.01) is evidentbetween the CA axis 1 sample scores and the inferred EagleLake temperatures.

Discussion

Earlier palaeoecological studies led to the recognition of threemain Holocene (10000 "4C years BP to present, c. 11500 cal.years BP to present) climate phases in southern British Columbia.The early-Holocene phase from 10000 to 7000 "'C years BP

(c. 11500-7800 cal. years BP), often called the xerothermic per-iod, is considered to have been warmer with less precipitationthan present (Mathewes and Heusser, 1981). The second phasefrom 7000 to 3500 "'C years BP (c. 7800-3800 cal. years BP) isthought to have been a period of climatic transition, and in south-ern British Columbia has been called the mesothermic period(Hebda, 1995). This transitional period begins with warm, moistconditions and ends with cool, moist conditions (Mathewes,1985). This cooling marks the beginning of the Neoglacial inter-val, a period of renewed glacial activity, which is believed to con-tinue through the third phase, from 3500 "'C years BP to present(c. 3800 cal. years BP to present). According to Pellatt andMathewes (1997), this Neoglacial cooling allowed for the estab-lishment of modern subalpine conditions. This last phase ischaracterized by cool temperatures, maximum precipitation andincludes the Tiedemann and 'Little Ice Age' (LIA) glacialadvances (Clague and Mathewes, 1996; Ryder and Thomson,1986).

Frozen LakeThe Frozen Lake core is characterized by very diverse midgeassemblages, including many rheophilous taxa (e.g., Brillia &Euryhapsis, Corynoneura & Thienemanniella, Eukiefferiella &Tvetenia, Limnophyes and Rheocricotopus). These taxa are typicalinhabitants of streams (Oliver and Roussel, 1983; Wiederholm,1983; Walker, 1988; Ruck et al., 1998). Rheophilous taxa are alsoabundant in the sediments of Marion and Tugulnuit Lakes in Bri-tish Columbia (Walker and Mathewes, 1987; Ruck et al., 1998).This abundance of rheophilous midges reflects substantial streaminflow. This inference is further supported by the fragmentaryrecord of larval Simuliidae (black flies). Simuliid remains wereconsistently present in low numbers throughout the length of theFrozen Lake core and are considered good indicators of flowingwater (Ruck etal., 1998; Walker, 2001).

Although Tanytarsina dominated in FRO-1, the great abun-dance of Chaoborus mandibles, accompanied by Microtendipesand other temperate taxa, suggests warm early-Holocene tempera-tures (Figure 3; Walker et al., 1997), as reflected in the midge-inferred mean July air temperatures ranging from 13.1 ± 2.0 to15.6 ± 2.1°C (Figure 4). A comparison with other midge-inferredmean July air-temperature records from southern British Colum-bia indicates that the Frozen Lake temperatures are typical forearly-Holocene treeline sites (Figure 9a; Palmer et al., 2002; Pel-latt et al., 2000). The inferred mean July air temperature decreasessomewhat through zone FRO-1. This trend continues throughzone FRO-2.Mathewes and Heusser (1981) suggested that a period of higher

precipitation and cooler temperatures began before the eruptionof Mt Mazama, also recorded by Palmer et al. (2002) and Smithet al. (1998). The interval following the Mazama eruption is oftencalled the mesothermic interval, a period of decreasing tempera-ture and increased moisture (Hebda, 1995). Although this mid-Holocene cooling appears to persist longer at other subalpine sites(e.g., Cabin Lake and 3M Pond; Pellatt et al., 2000), the overalltrend at Frozen Lake is similar.

Similarly, the late Holocene is considered a time of cooling,with Neoglacial advances in the Coast, Cascade and Rocky Moun-tains. It is recorded as a cooling in the midge records of CabinLake, 3M Pond (Smith et al., 1998; Pellatt et al., 2000), Lake-of-the-Woods and North Crater Lake (Figure 9a; Palmer et al.,2002). Higher-resolution sampling would be needed to resolveLIA events such as those described by Luckman and Kearney(1986) and Clague and Mathewes (1996).

Although a decreasing inferred mean July temperature trend isevident in the midge reconstruction, and fits well with diverseother sources of palaeoclimate data, the sample specific errors forthe samples all overlap, indicating that higher-resolution details




10

20

30

40

50

E

a

70

80'90'

100

110*

120

130-

140

150

160

170

180

190

200-

210-

220

230

240

9 10 11 12 13 14 15 16 17

Inferred Mean July Air Temperature (°C)18

Figure 4 Chironomid-inferred mean July air temperatures for Frozen Lake. Bars indicate the sample specific errors in the inferred temperatures.

Table 4 AMS radiocarbon and tephra dates for the Eagle Lake sediment core

Depth (cm) Lab. number Uncalibrated age Calibrated age* Relative area under Material dated(CAMS) (radiocarbon years BP) (calibrated years BP) probability distribution

36-35 3390 ± 130 3980-3940 0.016 Mt St Helens Yn tephra3930-3360 0.984

96-81 6730 ± 40 7670-7560 0.852 Mazama tephra7540-7510 0.148

106-105 74623 8330 ± 40 9470-9440 0.091 conifer needle fragments9440-9260 0.8649170-9150 0.045


of this reconstruction are not statistically significant. This is notsurprising given the errors associated with individual temperatureinferences (Walker et al., 1997; Lotter et al., 1999), and prior evi-dence that Holocene temperature changes were of a much smallermagnitude than those recorded in longer Quaternary sections.Consequently, reliable midge-based temperature reconstructionsfor the Holocene will need to be calculated by averaging theresults from several sites. The average or 'consensus' temperaturerecord will likely provide a more reliable indication of the truepattern of Holocene climatic changes.

Eagle LakeEagle Lake sediments included approximately half the number oftaxa found at Frozen Lake. Midge-inferred mean July air tempera-tures for Eagle Lake have a larger range than those calculated forFrozen Lake (Figure 7).

In the early-Holocene (EAG-1) or 'xerothermic' period, warmsummer air temperatures are inferred, ranging from 13.7 + 2.0-16.6 + 2.2°C. A cooler climate is inferred for the mid-Holocene,or 'mesothermic' period, but the inferred return to warm con-ditions at the core top directly contradicts the cooler temperaturesordinarily inferred for the late Holocene.

In British Columbia, a late-Holocene warming is only weaklyevident in the temperature reconstruction of Mathewes andHeusser (1981), based on pollen results from Marion Lake in thesouthern Coast Mountains. The inferred warming at Marion Lakeis small. Evidence for such warming is contrary to the bulk ofpalaeoenvironmental data from southern British Columbia,including numerous pollen records, and Mathewes' (1973) ownsubjective interpretation of Marion Lake's pollen. Bennett et al.(2001) note that palaeoclimate records from British Columbia donot provide an entirely coherent picture of late-Holocene climaticchange. They suggest that a variety of factors, including the com-plex topography of British Columbia and local variations in cli-matic change, may account for discrepancies among records.These discrepancies, however, are much more evident in precipi-tation and salinity inferences than in palaeotemperature records(Heinrichs et al., 2001).

Figure 9a shows that the Eagle Lake reconstruction does notfit the typical midge-inferred temperature profile as obtained fromother subalpine lakes in southern British Columbia (Palmer et al.,2002). The late-Holocene reconstructedtemperature anomalies aremuch warmer than those inferred at other sites, and much warmerthan current temperatures at Eagle Lake. We also note that the

.$

-.Nm2.h

4----I

--

fIII



12

25-

E-C

0)

n~ ~~

50 -

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I-Irm

75 II

I-. I

-I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.

FR0 2

150

I K-175 h

.,, -

225

0 0.1 0.2 0.3 0.4 0.5

FRO-1

1570±0 ''C BP

1960A0G'C BP

oBrdge RWr2435:26 4C BP

3560*40 'C BP

4630550 "C BP

6170*40 "C BPMazama6730*40 'C BP

81 0*0 "C BP

9390+70 'C BP

10020*0 'C BP

0.6 0.7 0.8

Squared Chord Distance

Figure 5 Comparison of available analogues with Frozen Lake chironomid fossil samples (squared chord distance from the fossil sample to the bestavailable modem surface sample analogues. Confidence intervals are shown.

'--COLD

ei

cl,~~~~~~0O0'"O

NIPtiA-X

WARM-->

sS<1

k

'± ~~vro~~-~~e'\'- V tj'V~~~~" '~~~' ~~'~~'~' Zone CONISS

20EA4

30.ItSI.HelensYn 3400-

50

70- EAG

Mozomo 6730*40. 90

100EA-I8330±40.

-:-:--"': 2 240 60 2 40 20 50100015000 2 4 6

gytijo tephro cloy

Figure 6 Midge stratigraphy for Eagle Lake. Chironomid and non-chironomid (Chaoborus, Ceratopogonidae and Acari) taxon abundances are expressedas percentages of the total identifiable Chironomidae (regions in black); expanded regions illustrate a X 10 exaggeration. Species are ranked in accordancewith their temperature optima derived from the weighted averaging model. Taxa that occurred in two or fewer samples, or that never exceeded 2%abundance, are excluded.

available analogues are unsatisfactory for most of the Eagle Lakecore, with the exception of EAG-2. Although Eagle Lake is shal-low, many of the lakes included in the surface sample training setare .2 m deep; thus, Eagle Lake's current depth is insufficient

to account for the poor analogue situation. A multitude of factorsother than temperature can affect midge communities, and itseems likely that these other effects may be driving Holocenechanges at Eagle Lake. Other factors influencing midge distri-


C,

M

Totol sum of squres

--- -I

Ia1251_

. Ii -

75% C. 1. 95% C.I. radocwbondo7t-

Mfl,ft. . m

I




Table 5 Zonation of the Eagle Lake chironomidrounded to the nearest century)

stratigraphy (ages

Chironomid Depth Uncalibrated age Calibrated age*assemblage (cm) (radiocarbon (calibratedzone years BP) years BP)

EAG-5 0-3 Present-400 Present-500EAG-4 3-29 400-2400 500-2400EAG-3 29-47 2400-3800 2400-4200EAG-2 47-88.5 3800-6700 4200-7600EAG-1 88.5-110 6700-8500 7600-9600


0EAG-5 F

5

10

15 - EAG420

25

30

35 -r-

40 -EAG-3

45 .-

50 _F

.c60

65 EAG.2

, , 6 v

.Bl~ 833D44 X4C BpI _ r _ 1~~I

IV _r_

9 10 11 12 13 14 15 16 17 18 19

Inferred Mean July Air Temperature (0C)

Figure 7 Chironomid-inferred mean July air temperatures for Eagle Lake.Bars indicate the sample specific errors in the inferred temperatures.

butions and diversity may include dispersal ability and changes inwater chemistry, benthic substrates and predators (Hoffman et al.,1996). Hoffman et al. (1996) noted that the number of invertebratetaxa decreases across the transition from forest to alpine lake eco-

systems. Changes in substrate, which can affect organism abun-dance and diversity, also tend to change with elevation from more

organic substrates at lower elevations (forest/subalpine lakes) tomore inorganic substrates at higher elevations (alpine lakes).Declining lake depth can also influence water temperature andnutrient cycling. Eagle Lake water depth has been known to fluc-tuate up to 1.0 m (Donald and Alger, 1984). We could be seeinga response of midge assemblages to variables other than simplytemperature changes (Lotter et al., 1997; Walker et al., 1997).

It is especially worth noting that the midge fauna of Eagle Lakeis very unusual relative to other high-elevation British Columbialakes, including a higher proportion of presumed 'thermophilous'species than anticipated. The Eagle Lake fauna is, therefore, a

clear 'outlier' relative to other lakes we have sampled. From thispalaeoenvironmental assessment, it is impossible to discern thereason for the unusual fauna. One or more key variables that nor-

55

a 60

O) 65 -_ EAG-2

70

75

80

85 Mazuma90 8O7 O "C BP

95

100_ ~IEAG-1

105 - 8=0 Ic BP

0 0.2 0.4 a.6 0.8

Squared Chord Distance

Figure 8 Comparison of available analogues with Eagle Lake chironomidfossil samples (squared chord distance from the fossil sample to the bestavailable modern surface sample analogues. Confidence intervals areshown.

mally co-vary with mean July air temperature (e.g., winter 02),may be exerting an independent influence on the midge fauna.

Regional assessment of temperature inferencesAs discussed in relation to the Frozen Lake results, good Holo-cene reconstructions of temperature should not be based on a sin-gle core, but rather on the results obtained from as many lakes aspossible. Figure 9b offers the consensus reconstruction for sixlakes in southern British Columbia: 3M Pond, Cabin Lake, CraterLake, Lake-of-the-Woods, Eagle Lake and Frozen Lake. Duringthe lateglacial period, from approximately 11000 through 9800cal. years BC (C. 10000 14C years BP) temperatures 1 to 5°C coolerthan today are inferred. This cold period is not recorded in 3MPond, Eagle or Frozen Lakes as these lakes were not formed untilthe early Holocene.

During the early Holocene (9600-7000 cal. years BC, c. 9800-8000 "4C years BP), mean July air temperatures, as inferred frommidges, were 1 to 5°C warmer than present. This reconstructionprovides independent confirmation of the early-Holocene warmthqualitatively inferred from palaeobotanical evidence (Mathewes,1985; Hebda, 1995). Temperatures gradually decreased through-out the mid-Holocene. The coolest Holocene summer tempera-tures begin at around 1000 cal. years BC (c. 2800 "4C years BP)and persist until the present day. 3M Pond shows the greatesttemperature change, with high-inferred temperatures in the earlyHolocene and the lowest temperatures during the late Holocene.Eagle Lake does not follow the pattern typical of other lakes.During the mid-Holocene interval, Eagle Lake temperatures areinferred to be a few degrees lower than during late-Holocenetimes. Although the pattern of temperature change may havevaried somewhat regionally (e.g., on a northeast to southwestaxis), we prefer to maintain a conservative interpretation of thedata, at least, until additional data from the Columbia Mountainscan be obtained.

1t fl 1 Ir r -i fi... F I




(a)

8

6E0_- 4

CL

ez 2

o._ 0

E -20

-4

-612000

(b)

8-.

Ei

0_ 4

120

2:.°0W

0da_ :

-4

12000

10000

Age (radiocarbon years)8000 6000 4000

10000 8000 6000 40004 Age (calendar years)

10000

*10000 8000

Age (radiocarbon years)8000 6000 4000

2000 0

2000 0 2000

2000 0

6000 4000 2000 0 2000

Age (calendar years)Figure 9 (a) Chironomid-inferred palaeotemperature records for six lakes in southern British Columbia. (b) Five-point running averages of palaeotempera-ture records for southern British Columbia with and without the Eagle Lake data.

The average palaeotemperature reconstruction, with or withoutthe Eagle Lake results, generally parallels the qualitative infer-ences derived from palaeobotanical data (Figure 9b; Mathewes,1985; Barnosky et al., 1987; Vance, 1987; Hebda, 1995; Walkerand Pellatt, 2001; 2004). Our reconstruction clearly reveals therapidity of the warming at the Pleistocene/Holocene boundary,c. 9600 cal. years BC (C. 10000 14C years BP). In contrast, themid-Holocene cooling is indicated as a very gradual transitionfrom xerothermic warmth to current summer temperatures. Conse-quently, the upper and lower bounds of the 'Mesothermic' are noteasily defined. Nevertheless, the term 'Mesothermic' is useful inreference to this slow transition between the warm 'Xerothermic'(prior to c. 6000 to 7000 cal. years BC; c. 7000 to 8000 "4C yearsBP) and cool 'Neoglacial' (after c. 2500 cal. years BC; c. 4000 14Cyears BP) intervals in British Columbia.The rapid warming at the end of the Pleistocene is well

documented throughout the Northern Hemisphere. Locally, the

Cordilleran ice sheet was rapidly disintegrating, causing anabrupt decrease in surface albedo. This change in albedo wouldhave markedly reinforced the shift from glacial temperaturesto a warm Holocene climate. We also note that the maximumHolocene temperatures coincide with the summer solar in-solation maximum c. 8000 BC (c. 8800 l4C years BP; Bergerand Loutre, 1991). The cooling trend through the mid-Holocene parallels a decline in summer solar insolation(Barnosky etal., 1987; Vance, 1987; Thompson etal., 1993).We concur with Kutzbach et al. (1993) that changes inHolocene summer temperatures were driven largely by changesin summer solar insolation. Unfortunately, good proxy recordsfor winter palaeotemperatures do not exist. Since winter solarinsolation was at a minimum in the early Holocene, winter tem-peratures changes may have followed the opposite trajectory -i.e., from cold early-Holocene to mild late-Holocene wintertemperatures.



Sandra M. Rosenberg et a!.: Midge-inferred Holocene climate history from British Columbia, Canada 269

Conclusions

Both Frozen and Eagle Lakes record similar timing of climaticchange throughout the Holocene even though the direction of tem-perature fluctuations do not concur. The midge-temperature infer-ences indicate high mean July air temperatures during the earlyHolocene for both subalpine lakes. Lower mean July air tempera-tures were inferred for the mid-Holocene. Continued low-tem-perature inferences were obtained from Frozen Lake through tothe present day, but during this same late-Holocene period EagleLake temperature inferences increase contrary to most otherreconstructions from southern British Columbia. The Eagle Laketemperature record may not be providing an accurate reconstruc-tion of Holocene temperature trends. It seems likely that otherfactors apart from temperature are affecting the fauna. Alterna-tively, the climate situation surrounding Eagle Lake may beregionally complex. A comparison with multiproxy data from thesame, and other sites, is needed to substantiate the temperatureinferences.Due to the relatively small magnitude of temperature changes

throughout the Holocene, midge-based climate reconstructionsbased on single cores must be interpreted with caution. By averag-

ing the results from several sites in southern British Columbia, a

more reliable estimate of the Holocene temperature changes was

obtained. Nevertheless, changes in midge communities are notdriven only by events related to climate. Changes in water depth,chemistry or lake productivity may interfere with midge-basedtemperature reconstructions. Independent evidence from otherproxy climate indicators (e.g., in multiproxy studies) must be care-

fully considered when evaluating the results.

Acknowledgements

Funding for this project was provided by a grant from the NaturalSciences and Engineering Research Council of Canada to Ian R.Walker. Thanks also to Murray Peterson and Parks CanadaRevelstoke for permission to sample within Mount RevelstokeNational Park. Appreciation is also due to those who assisted withfield sampling: Markus Heinrichs, Marju Heinrichs and Marie-Andr6e Fallu.

References

Allen, G.B., Brown, KJ. and Hebda, RJ. 1999: Surface pollen spectrafrom southern Vancouver Island, British Columbia, Canada. CanadianJournal of Botany 77, 786-99.Alley, N.F. 1976: The palynology and palaeoclimatic significance of a

dated core of Holocene peat, Okanagan Valley, southern British Columbia,Canada. Canadian Journal of Earth Sciences 13, 1131-44.Barnosky, C.W., Anderson, P.M. and Bartlein, P.J. 1987: The north-western U.S. during deglaciation; vegetational history and paleoclimaticimplications. In Ruddiman,W.F. and Wright,, H.E. Jr, editors, NorthAmerica and adjacent oceans during the last deglaciation, v. K-3, TheGeology of North America, Boulder, CO: Geological Society of America,289-32 1.

Battarbee, R.W. 2000: Palaeolimnological approaches to climate change,with special regard to the biological record. Quaternary Science Reviews19, 107-24.Battarbee, R.W., Grytnes, J., Thompson, R., Appleby, P.G., Catalan,J., Korhola, A., Birks, HJ.B., Heegaard, E. and Lami, A. 2002:Comparing palaeolimnological and instrumental evidence of climate

change for remote mountain lakes over the last 200 years. Journal ofPaleolimnology 28, 161-79.

Bennett, J.R., Cumming, B.F., Leavitt, P.R., Chiu, M., Smol, J.P. andSzeicz, J. 2001: Diatom, pollen and chemical evidence of postglacial

climatic change at Big Lake, south-central British Columbia, Canada.Quaternary Research 55, 332-43.Bennett, K.D. 1996: Determination of the number of zones in a bio-stratigraphical sequence. New Phytologist 132, 155-70.

2002: Documentation for psimpoll 4.10 and pscombl.03. C pro-grams for plotting pollen diagrams and analysing pollen data. Uppsala:Quaternary Geology, Uppsala Universitet, 25 pp.Berger, A. and Loutre, M.F. 1991: Insolation values for the climate ofthe last 10 million years. Quaternary Science Reviews 10, 297-317.Brodin, Y.W. and Gransberg, M. 1993: Responses of insects, especiallyChironomidae (Diptera), and mites to 130 years of acidification in a Scott-ish lake. Hydrobiologia 250, 201-12.Brooke, R.C., Peterson, E.B. and Krajina, VJ. 1970: The subalpinemountain hemlock zone: subalpine vegetation in southwestern BritishColumbia, its climatic characteristics, soils, ecosystems and environmentalrelationships. In Krajina, V.J. and Brooke, R.C., editors, Ecology ofwestern North America, Vancouver: Department of Botany, University ofBritish Columbia, 147-349.Brooks, SJ., Bennion, H. and Birks, HJ.B. 2001: Tracing lake trophichistory with chironomid - total phosphorus inference model. FreshwaterBiology 46, 513-33.Cawker, K.B. 1983: Fire history and grassland vegetation change: threepollen diagrams from southern British Columbia. Canadian Journal ofBotany 61, 1126-39.Clague, JJ. and Mathewes, R.W. 1989: Early Holocene thermalmaximum in western North America new evidence from Castle Peak,British Columbia. Geology 17, 277-80.

1996: Neoglaciation, glacier-dammed lakes, and vegetation changein northwestern British Columbia, Canada. Arctic and Alpine Research28, 10-24.Clague, JJ., Evans, S.G., Rampton, V.N. and Woodsworth, GJ. 1995:Improved age estimates for the White River and Bridge River tephras,western Canada. Canadian Journal of Earth Sciences 32, 1172-79.Coupe, R., Stewart, A.C. and Wikeem, B.M. 1991: Engelmann spruce-subalpine fir zone. In Meidinger, D. and Pojar, J., editors, Ecosystems ofBritish Columbia, Victoria: Research Branch, Ministry of Forests,Province of British Columbia, 223-36.Cwynar, L.C. and Levesque, AJ. 1995: Chironomid evidence for late-glacial climatic reversals in Maine. Quaternary Research 43, 405-13.Donald, D.B. and Alger, DJ. 1984: Limnological studies in Glacier andMount Revelstoke National Parks. Edmonton: Parks Canada, CanadianWildlife Service.Environment Canada 1993: Canadian climate normals, 1961-1990(Normales climatiques au Canada, 1961-1990). Volume 1: BritishColumbia. Downsview, Ontario: Atmospheric Environment Service.Glew, J.R. 1989: A new trigger mechanism for sediment samplers.Journal of Paleolimnology 2, 241-43.Grimm, E.C. 1987: CONISS: a FORTRAN 77 program for strati-graphically constrained cluster analysis by the method of incremental sumof squares. Computers and Geosciences 13, 13-35.

1991: TILIA-GRAPH 2.0 version b.5 (computer software). Spring-field: Illinois State Museum, Research and Collections Center.

1993: TILIA 2.0 version b.4 (computer software). Springfield:Illinois State Museum, Research and Collections Center.Haliett, DJ. and Walker, R.C. 2000: Paleoecology and its applicationto fire and vegetation management in Kootenay National Park, BritishColumbia. Journal of Paleolimnology 24, 401-14.Hallett, DJ., Hills, L.V. and Clague, JJ. 1997: New accelerator massspectrometry radiocarbon ages for the Mazama tephra layer fromKootenay National Park, British Columbia, Canada. Canadian Journal ofEarth Sciences 34, 1202-209.Haliett, DJ., Lepofsky, D.S., Mathewes, R.W. and Lertzman, K.P.2003: 11 000 years of fire history and climate in the mountain hemlockrainforests of southwestern British Columbia based on sedimentarycharcoal. Canadian Journal of Forest Research 33, 292-312.Haliett, DJ., Mathewes, R.W. and Foit, F.F. 2001: Mid-HoloceneGlacier Peak and Mount St Helens We tephra layers detected in lake sedi-ments from southern British Columbia using high-resolution techniques.Quaternary Research 55, 284-92.Hansen, H.P. 1955: Postglacial forests in south central and central BritishColumbia. American Journal of Science 253, 64058.Hazell, S. 1979: Late Quaternary vegetation and climate of Dunbar Valley,British Columbia. MSc thesis, University of Toronto, Ontario, 101 pp.




Hebda, RJ. 1995: British Columbia vegetation and climate history withfocus on 6 ka BP. Geographie physique et Quaternaire 49, 55-79.Heinrichs, M.L., Walker, I.R. and Mathewes, R.W. 2001. Chironomid-based paleosalinity records in southern British Columbia, Canada: acomparison of transfer functions. Journal of Paleolimnology 26, 147-59.Heinrichs, M.L., Walker, I.R., Mathewes, R.W. and Hebda, R.J.1999: Holocene chironomid-inferred salinity and paleovegetationreconstruction from Kilpoola Lake, British Columbia. Ge'ographiephysique et Quaternaire 53, 211-21.Heinrichs, M.L., Wilson, S.E., Walker, I.R., Smol, J.P., Mathewes,R.W. and Hall, KJ. 1997: Midge and diatom-based palaeosalinity recon-structions for Mahoney Lake, Okanagan Valley, British Columbia, Can-ada. International Journal of Salt Lake Research 6, 249-67.Heiri, 0. and Lotter, A.F. 2001: Effect of low count sums on quantitativeenvironmental reconstructions: an example using subfossil chironomids.Journal of Paleolimnology 26, 343-50.Heiri, O., Wick, L., van Leeuwen, J.F.N., van der Knaap, W.O. andLotter, A.F. 2003: Holocene tree immigration and the chironomid faunaof a small Swiss subalpine lake (Hinterburgsee, 1515 m asl). Palaeo-geography, Palaeoclimatology, Palaeoecology 189, 35-53.Hofmann, W. 1988: The significance of chironomid analysis (Insecta:Diptera) for paleolimnological research. Palaeogeography, Palaeo-climatology, Palaeoecology 62, 501-509.Hoffman, R.L., Liss, W.J., Larson, G.L., Deimling, E.K. andLomnicky, G.A. 1996: Distribution of nearshore macroinvertebrates inlakes of the Northern Cascade Mountains, Washington, USA. Archiv furHydrobiologie 136, 363-89.Juggins, S. 1997: CALIBRATE version 0.82 (computer software).Juggins, S. and ter Braak, CJ.F. 1996: WAPLS version 1.1(computer software).Kutzbach, J.E., Guetter, P.J., Behling, PJ. and Selin, R. 1993: Simu-lated climatic changes: results of the COHMAP climate-model experi-ments. In Wright, H.E. Jr, Kutzbach, J.E., Webb, T. III, Ruddiman, W.F.,Street-Perrott, F.A. and Bartlein, P.J., editors, Global climates since thelast glacial maximum, Minneapolis: University of Minnesota Press, 24-93.Laing, T.E., Ruhland, K.M. and Smol, J.P. 1999: Past environmentaland climatic changes related to tree-line shifts inferred from fossil diatomsfrom a lake near the Lena River Delta, Siberia. The Holocene 9, 547-57.Laird, K.R., Fritz, S.C. and Cumming, B.F. 1998: A diatom-basedreconstruction of drought intensity, duration, and frequency from MoonLake, North Dakota: a sub-decadal record of the last 2300 years. Journalof Paleolimnology 19, 161-79.Larocque, I. 2001: How many chironomid head capsules are enough?A statistical approach to determine sample size for palaeoclimaticreconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology172,133-42.Levesque, AJ., Mayle, F.E., Walker, I.R. and Cwynar, L.C. 1993a: Apreviously unrecognized late-glacial cold event in eastern North AmericaNature 361, 623-26.

1993b: The Amphi-Atlantic Oscillation: a proposed late-glacialclimatic event. Quaternary Science Reviews 12, 629-43.Little, J.L., Hall, R.I., Quinlan, R. and Smol, J.P. 2000: Past trophicstatus and hypolimnetic anoxia during eutrophication and remediation ofGravenhurst Bay, Ontario: comparison of diatoms, chironomids, and his-torical records. Canadian Journal of Fisheries and Aquatic Sciences 57,333-41.Livingstone, D.M. and Lotter, A.F. 1998: The relationship between airand water temperatures in lakes of the Swiss Plateau: a case study withpalaeolimnological implications. Journal of Paleolimnology 19, 181-98.Livingstone, D.M., Lotter, A.F. and Walker, I.R. 1999: The decreasein summer surface water temperature with altitude in Swiss Alpine lakes:a comparison with air temperature lapse rates. Arctic, Antarctic and AlpineResearch 31, 341-52.Lotter, A.F., Birks, HJ.B., Hofmann, W. and Marchetto, A. 1997:Modern diatom, cladocera, chironomid, and chrysophyte cyst assemblagesas quantitative indicators for ihe reconstruction of past environmental con-ditions in the Alps. I. Climate. Journal of Paleolimnology 18, 395-420.

1998. Modern diatom, cladocera, chironomid, and chrysophyte cystassemblages as quantitative indicators for the reconstruction of pastenvironmental conditions in the Alps. II. Nutrients. Journal of Paleo-limnology 19, 443-63.Lotter, A.F., Walker, I.R., Brooks, SJ. and Hofmann, W. 1999: Anintercontinental comparison of chironomid palaeotemperature inference

models: Europe vs. North America Quaternary Science Reviews 18,717-35.Luckman, B.H. and Kearney, M.S. 1986: Reconstruction of Holocenechanges in alpine vegetation and climate in the Maligne Range, JasperNational Park, Alberta. Quaternary Research 26, 244-61.Luckman, B.H., Kearnery, MS., King, R.H. and Beaudoin, A.B. 1986.Revised '4C age for St Helens Y tephra at Tonquin Pass, British Columbia.Canadian Journal of Earth Sciences 23, 734-36.Mathewes, R.W. 1973: A palynological study of postglacial vegetationchanges in the University Research Forest, southwestern British Columbia.Canadian Journal of Botany 51, 2085-103.

1985: Paleobotanical evidence for climatic change in southernBritish Columbia during late-glacial and Holocene time. Syllogeus 55,397-422.Mathewes, R.W. and Heusser, L.E. 1981: A 12000 year palynologicalrecord of temperature and precipitation trends in southwestern BritishColumbia. Canadian Journal of Botany 59, 707-10.Mathewes, R.W. and King, M. 1989: Holocene vegetation, climate, andlake-level changes in the Interior Douglas-fir biogeoclimatic zone, BritishColumbia. Canadian Journal of Earth Sciences 26, 1811-25.Mullineaux, D.R. 1986. Summary of pre-1980 tephra-fall depositserupted from Mount St Helens, Washington State, USA. Bulletin ofVolcanology 48, 17-26.Olander, H., Birks, HJ.B., Korhola, A. and Blom, T. 1999: Anexpanded calibration model for inferring lakewater and air temperaturesfrom fossil chironomid assemblages in northern Fennoscandia. The Holo-cene 9, 279-94.Oliver, D.R. and Roussel, M.E. 1983: The insects and arachnids of Can-ada Part 11: the genera of larval midges of Canada; Diptera: Chironomi-dae. Ottawa: Agriculture Canada Publication, 1746.Overpeck, J.T., Webb, T. Ill, and Prentice, I.C. 1985: Quantitativeinterpretation of fossil pollen spectra: dissimilarity coefficients and themethod of modern analogs. Quaternary Research 23, 87-108.Palmer, S.L. 1998: Subfossil chironomids (Insecta: Diptera) and climaticchange at high elevation lakes in the Engelmann spruce-subalpine fir zonein southwestern British Columbia. MSc thesis, Department of Zoology,University of British Columbia, Vancouver, 105 pp.Palmer, S.L., Walker, I.R., Heinrichs, M.L., Hebda, RJ. and Scudder,G.G.E. 2002. Postglacial midge community change and Holocene palaeo-temperature reconstructions for treeline, southern British Columbia(Canada). Journal of Paleolimnology 28, 469-90.Pellatt, M.G. and Mathewes, R.W. 1997: Holocene tree line and climatechange on the Queen Charlotte Islands, Canada. Quaternary Research 48,88-99.Pellatt, M.G., Hebda, RJ. and Mathewes, R.W. 2001: High-resolutionHolocene vegetation history and climate from Hole 1034B, ODP leg 169S,Saanich Inlet, Canada. Marine Geology 174, 211-26.Pellatt, M.G., Smith, MJ., Mathewes, R.W. and Walker, I.R. 1998:Palaeoecology of postglacial treeline shifts in the northern CascadeMountains, Canada. Palaeogeography, Palaeoclimatology, Palaeoecology141, 123-38.Pellatt, M.G., Smith, MJ., Mathewes, R.W., Walker, I.R. and Palmer,S.L. 2000: Holocene treeline and climate change in the subalpine zonenear Stoyoma Mountain, Cascade Mountains, southwestern BritishColumbia, Canada. Arctic, Antarctic and Alpine Research 32, 73-83.Quinlan, R. and Smol, J.P. 2001: Setting minimum head capsuleabundance and taxa deletion criteria in chironomid-based inferencemodels. Journal of Paleolimnology 26, 327-42.Quinlan, R., Smol, J.P. and Hall, R.I. 1998: Quantitative inferences ofpast hypolimnetic anoxia in south-central Ontario lakes using fossilmidges (Diptera: Chironomidae). Canadian Journal of Fisheries andAquatic Sciences 55, 587-96.Reasoner, M.A. 1993: Equipment and procedure improvements for alightweight, inexpensive, percussion core sampling system. Journal ofPaleolimnology 8, 273-81.Reasoner, M.A. and Hickman, M. 1989: Late Quaternary environmentalchange in the Lake O'Hara region, Yoho National Park, British Columbia.Palaeogeography, Palaeoclimatology, Palaeoecology 72, 291 -316.Rosenberg, S.M., Walker, I.R. and Mathewes, R.W. 2003: Postglacialspread of hemlock (Tsuga) and vegetation history in Mount RevelstokeNational Park, British Columbia, Canada. Canadian Journal ofBotany 81,139-51.Ruck, A., Walker, I.R. and Hebda, RJ. 1998: A palaeolimnological




study of TugulnuitLake, British Columbia, Canada, with special emphasison river influence as recorded by chironomids in the lake's sediment. Jour-nal of Paleolimnology 19, 63-75.Ryder, J.M. and Thomson, B. 1986: Neoglaciation in the southern CoastMountains of British Columbia: chronology prior to the late Neoglacialmaximum. Canadian Journal of Earth Sciences 23, 273-87.Schneli, 0.A. and Willassen, E. 1996: The chironomid (Diptera)communities in two sediment cores from Store Hovvatn, S. Norway, an

acidified lake. Annales de Limnologie 32, 45-61.Smith, MJ. 1997: Postglacial changes in chironomid communities andinferred climate near treeline at Mount Stoyoma, Cascade Mountains,southwestern British Columbia. MSc thesis, Simon Fraser University,Burnaby, 71 pp.

Smith, MJ., Pellatt, M.G., Walker, I.R. and Mathewes, R.W. 1998:Postglacial changes in chironomid communities and inferred climate near

treeline at Mount Stoyoma, Cascade Mountains, southwestern BritishColumbia. Journal of Paleolimnology 20, 277-93.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 cal. BP. Radio-carbon 40, 1041-83.Thompson, R.S., Whitlock, C., Bartlein, P.J., Harrison, S.P. andSpaulding, W.G. 1993: Climatic changes in the western United Statessince 18,000 yr B.P. In Wright, H.E. Jr, Kutzbach, J.E., Webb, T. III,

Ruddiman, W.F., Street-Perrott, F.A. and Bartlein, P.J., editors, Globalclimates since the last glacial maximum, Minneapolis: University ofMinnesota Press, 468-513.Uutala, AJ. 1990: Chaoborus (Diptera: Chaoboridae) mandibles -

paleolimnological indicators of the historical status of fish populations inacid-sensitive lakes. Journal of Paleolimnology 4, 139-51.Vance, R.E. 1987: Meteorological records of historic droughts as climaticanalogues for the Holocene. In McKinnon, N.A. and Stuart, G.S.L.,editors, Man and the Mid Holocene Climatic Optimum - Proceedings ofthe Seventeenth Annual Conference of the Archaeological Association ofthe University of Calgary. Calgary: The University of Calgary Archaeol-ogical Association, 17-32.Wainman, N. and Mathewes, R.W. 1987: Forest history of the last 12 000years based on plant macrofossil analysis of sediment from Marion Lake,southwestern British Columbia. Canadian Journal ofBotany 65, 2179-87.Walker, I.R. 1987: Chironomidae (Diptera) in paleoecology. QuaternaryScience Reviews 6, 29-40.

1988: Late-Quaternary paleoecology of Chironomidae (Diptera:Insecta) from lake sediments in British Columbia. PhD thesis, SimonFraser University, Bumaby, 204 pp.

1991: Modem assemblages of arctic and alpine Chironomidae asanalogues for late-glacial communities. Hydrobiologia 214, 223-27.

1996-2000: WWW field guide to subfossil midges. http://www.ouc.ca/eesc/iwalker/wwwguide/ (last accessed 28 October 2003).

2001: Midges: Chironomidae and related Diptera. In Smol, J.P.,Birks, H.J.B. and Last, W.M., editors, Tracking environmental changeusing lake sediments, volume 4, zoological indicators, Dordrecht: Kluwer,43-66.Walker, I.R. and MacDonald, G.M. 1995: Distributions of Chironomidae(Insecta: Diptera) and other freshwater midges with respect to treeline,Northwest Territories, Canada. Arctic and Alpine Research 27, 258-63.Walker, I.R. and Mathewes, R.W. 1987: Chironomidae (Diptera) andpostglacial climate at Marion Lake, British Columbia, Canada. QuaternaryResearch 27, 89-102.

1989: Chironomidae (Diptera) remains in surficial lake sedimentsfrom the Canadian Cordillera: analysis of the fauna across an altitudinalgradient. Journal of Paleolimnology 2, 61-80.Walker, I.R. and Pellatt, M.G. 2001: Climate change and impacts insouthern British Columbia - a palaeoenvironmental perspective. Un-published report to the Climate Change Action Fund, Natural ResourcesCanada, Ottawa, 66 pp.

2004: Climate change in coastal British Columbia - a palaeo-environmental perspective. Canadian Water Resources Journal, in press.Walker, I.R., Mott, RJ. and Smol, J.P. 1991a: Allerod-Younger Dryaslake temperatures from midge fossils in Atlantic Canada. Science 253,1010-12.Walker, I.R., Levesque, AJ., Cwynar, L.C. and Lotter, A.F. 1997: Anexpanded surface-water palaeotemperature inference model for use withfossil midges from eastern Canada. Journal ofPaleolimnology 18, 165-78.Walker, I.R., Smol, J.P., Engstrom, D.R. and Birks, HJ.B. 1991b: Anassessment of Chironomidae as quantitative indicators of past climaticchange. Canadian Journal of Fisheries and Aquatic Sciences 48, 975-87.Walker, I.R., Wilson, S.E. and Smol, J.P. 1995: Chironomidae (Diptera):quantitative palaeosalinity indicators for lakes of western Canada.Canadian Journal of Fisheries and Aquatic Sciences 52, 950-60.Wiederholm, T., editor 1983: Chironomidae of the Holarctic region.Keys and diagnoses. Part 1 - larvae. Entomologica Scandinavica Sup-plement No. 19.Wilson, S.E., Walker, I.R., Mott, RJ. and Smol, J.P. 1993: Climaticand limnological changes associated with the Younger Dryas in AtlanticCanada Climate Dynamics 8, 177-87.Wright, H.E. Jr 1967: A square-rod piston sampler for lake sediments.Journal of Sedimentary Petrology 37, 975-76.



midge-inferred holocene climate history southern · moist air masses from the pacific coast are...

Documents