Postglacial climate in the St. Lawrence lowlands, southern Québec: pollen and lake-level evidence

Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72 www.elsevier.com/locate/palaeo

Postglacial climate in the St. Lawrence lowlands, southern Que¤bec: pollen and lake-level evidence Serge D. Muller a;b;c; , Pierre J.H. Richard a;b , Joe«l Guiot d , Jacques-Louis de Beaulieu c , David Fortin a a

c

De¤partement de Ge¤ographie, Universite¤ de Montre¤al, CP 6128, Succursale Centre-Ville, Montre¤al, QC, Canada H3C 3J7 b Centre for Climate and Global Change Research (C2 GCR), McGill University, 805 Sherbrooke Street West, Montre¤al, QC, Canada H3A 2K6 Ł cologie et de Pale¤oe¤cologie (IMEP), Case 451, Faculte¤ de St-Je¤ro“me, 13 397 Marseille cedex 20, France Institut Me¤diterrane¤en d’E d Centre Europe¤en de Recherche et d’Enseignement de Ge¤osciences de l’Environnement (CEREGE), Europo“le Me¤diterrane¤en de l’Arbois, BP 80, 13 545 Aix-en-Provence cedex 04, France Received 9 October 2001; accepted 3 December 2002

Abstract Pollen and lake-level data are used to reconstruct past climate changes in the St. Lawrence lowlands, southern Que¤bec. Past lake-level changes are assessed from sedimentological, pollen and macrofossil records from a single shallow-water core from Lac Hertel, which lies in the central part of the studied area. Three low lake-level phases are recognised: prior to 8000, 7600^6600 and 4800^3400 cal. BP. The modern analogue method is applied to pollen data from seven well-dated sites from the St. Lawrence lowlands and adjacent mountain areas, constrained and unconstrained by lake-level changes. The reconstructed climate changes are congruent with the pattern of climate changes known from eastern North America: a dry and cold late-glacial episode due to the presence of pro-glacial lakes and seas; a rapid warming between 12 500 and 11 000 cal. BP possibly caused by increasing summer insolation; a dry period from 10 000 to 6500 cal. BP; a brief cooling between 9000 and 8000 cal. BP, possibly related to a summer cooling of Arctic airmasses; a temperature maximum around 8000 cal. BP; and finally, a progressive decrease in summer temperature and an increase in (winter?) precipitation over the 4500 last years. These results show that it is possible to reveal seasonal patterns in climate by combining pollen and lake-level data. ? 2002 Elsevier Science B.V. All rights reserved. Keywords: Lake levels; climate reconstruction; regional scale; modern analogue method; St. Lawrence lowlands; southern Que¤bec

1. Introduction

* Corresponding author. Present address: ISEM, Case 061, Univ. Montpellier-2, pl. Eugene Bataillon, F-34095 Montpellier cedex 05, France. E-mail address: [email protected] (S.D. Muller).

According to the assumption of a relationship between climate and plant distribution (Woodward, 1987), past climates can be reconstructed from past phytogeographies. However, our perception of this relationship depends greatly on scales (Ritchie, 1986; Webb, 1986, 1993). At the

0031-0182 / 02 / $ ^ see front matter ? 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(02)00710-1

PALAEO 3020 24-2-03

52

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

scale of continents, the present-day latitudinal zonation of vegetation matches well the north^south temperature gradient (Greller, 1989; Huntley et al., 1989). Based on that relationship, past temperatures have been reconstructed from pollen data at a continental scale (e.g., Prentice et al., 1992; Guiot et al., 1993; Webb et al., 1993b). At a regional scale, several factors such as disturbances (Ritchie, 1986), physiography or ecological

processes (Davis et al., 1986; Prentice, 1986) create non-equilibrium states between vegetation and climate, and consequently complicate pollenbased climate reconstructions. Moreover, past plant formations without modern analogues constitute another source of di⁄culties for these reconstructions. This paper presents a pollen-based climate reconstruction realised at a regional scale by ac-

Fig. 1. Study area and location of sites used in climate and lake-level reconstructions.

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

counting for these problems. Previous studies in the St. Lawrence lowlands, southwestern Que¤bec (Fig. 1), have identi¢ed the low in£uence of local factors on Holocene plant geography (Muller and Richard, 2001) and discussed the similarities between past and modern plant communities (Richard, 1989, 1993). Moreover, there are several welldated pollen diagrams (4^10 dates per site) from the study area, which allow regional climatic averages to be calculated thus reducing the in£uence of local physiographic parameters. Finally, the climate reconstruction, made using the modern analogue method, is constrained by lake-level changes, shown to be linked to climate dynamics in eastern North America (Harrison, 1989). The lake-level reconstruction of Lac Hertel, a small lake within a small catchment which lies right in the central part of the area, o¡ers a reliable and independent record of hydroclimatology (Digerfeldt, 1986; Vassiljev et al., 1998). Our study provides the ¢rst quantitative climate reconstruction in southern Que¤bec and allows to discuss it in regard to the qualitative knowledge of past climates (e.g., Richard et al., 1992; Richard, 1994; Lavoie and Richard, 2000a).

2. Study area The Montre¤al lowlands (Fig. 1) are a £at plain of marine clay deposited above tills on sandstonedolomitic bedrock (Laverdie're et al., 1972). Several intrusive hills (Monteregian hills) emerge from the plain roughly along an east^west axis. Lakes are rare in the lowlands and, except for the in¢lled Lac Romer, they are all located on Monteregian hills. The region lies at the northern limit of deciduous forests, and the natural vegetation is dominated by Acer saccharum, in association with Betula alleghaniensis on highlands, with Tilia americana in the northeastern part of the plain and with Carya cordiformis southwestwards of Sorel (Grandtner, 1966; Richard, 1987). The transitional nature of these forests is shown by the mixture of boreal and temperate elements (Maycock, 1961; Bouchard and Maycock, 1978). Most of the plain, except moraines and bedrock outcrops, is today occupied by cultivated lands.

53

3. Lake-level changes Lac Hertel (Figs. 1 and 2) occupies the central depression of Mont St. Hilaire, a Monteregian hill. Located 50 km east of Montre¤al, this intrusive massif is mainly composed of nepheline syenite in the east and of essexite (gabbro) in the west (Horva¤th and Gault, 1990). The 28.4 ha lake, situated 173 m above sea level, drains a very small watershed (about 3.47 km2 ) and is therefore likely to provide a good record of past moisture-balance changes. The present water level is controlled by the presence of a dam on the outlet (Fig. 2) and is thought to be around 3 m above the natural maximal lake level, which determines a lake surface area of about 15 ha. The natural lake :watershed ratio would have been 1:23. 3.1. Methods The lake-level changes are established from a combination of sediment composition and vegetation distribution data (Dearing and Foster, 1986; Digerfeldt, 1986), reconstructed from sedimentological, macrofossil and pollen analyses on a single marginal core. The core was taken 220 m o¡shore in a water depth of 6 m, next to an inlet (Fig. 2), with a modi¢ed Livingstone piston sampler (Livingstone, 1955; Wright, 1967). Loss-onignition (at 600‡C during 30 min; Dean, 1974) was measured on contiguous 1-cm-thick samples. Chronological control is provided by three radiocarbon dates (Table 1) and two well-dated regional pollen events : the Tsuga decline dated around 5500 cal. BP and the Ambrosia rise resulting from European deforestation between 250 and 200 cal. BP (Muller and Richard, 2001). Conventional radiocarbon dates are calibrated with the CALIB 4.0 programme (Stuiver and Reimer, 1993). Calibrated dates (cal. BP = calendar years before present) were used to construct the age^depth model (Fig. 3), by interpolating the simplest curve connecting dates within the 2c con¢dence intervals. The two dates on the basal part of the gyttja (TO-8739 and Beta-154759 ; Table 1) are older than expected given the nature of the pollen assemblages. Correlations with other sites in the region lead us to attribute an age of about 8000 cal.

PALAEO 3020 24-2-03

54

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Fig. 2. Bathymetry of Lac Hertel. (A) Map showing the location of the pro¢le Herl (black dot) and the measurements of sediment thickness (grey dots). Depth in metres. Natural lake surface is noted in grey. (B) Cross-section joining the inlet mouth to the coring site, along the transect V^W. For convenience, sediment thickness measurements are represented along a linear transect.

Table 1 Conventional and calibrated radiocarbon ages Core name

Laboratory code

Datation method

Calibrated age

Con¢dence interval Conventional 2c age

13

C/12 C ratio

Herl

TO-8736 TO-8737 TO-8738 Beta-154759 TO-8739

AMS AMS AMS conventional AMS

3700 4700 6370 9440 10440

3830^3610 4830^4530 6490^6290 9520^9250 11070^10180

n.a. n.a. n.a. 329.6x n.a.

Depth (cm)

Rejected ages are noted in italic (see text for details). AMS dates were performed on terrestrial macrofossils. n.a.: not available.

PALAEO 3020 24-2-03

3460 R 40 4160 R 50 5600 R 50 8460 B 70 9280 B 150

153^155 277^278 371^375 456^459 461^466

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

55

0

Depth (m)

1

2 3

4 5

11

10

9

8

7

6

5

4

3

2

1

0

Age (103 cal. years BP)

Fig. 3. Age^depth curve of the pro¢le Herl (Lac Hertel). Black dots represent calibrated radiocarbon dates and white dots represent well-dated regional events (Tsuga decline, European settlement) and the sediment surface. The curve is modi¢ed between 120 and 80 cm on the basis of pollen concentrations. The clay deposit is noted in grey. See text for details.

BP for the beginning of gyttja deposition at Lac Hertel. Assuming a constant pollen input between 140 and 40 cm depth, we calculated that the silty gyttja between 120 and 80 cm depth (S8, Fig. 4) accumulated between 2600 and 2050 cal. BP, i.e. three times faster than in the adjacent gyttja layers. Pollen was extracted according to the conventional protocol (Faegri and Iversen, 1989). Pollen percentages were calculated on a sum excluding hydrophilous taxa and Pteridophyta (Berglund and Ralska-Jasiewiczowa, 1986). Alnus incana was also excluded from the pollen sum, due to its local origin in some sections of the core (indicated by macroremains). A minimal number of 500 pollen grains was counted and included in the sum. Macrofossils were extracted by sifting samples under a gentle spray of water (GrosseBrauckmann, 1986). Volumes ranged between 10 and 60 cm3 , and the results were standardised to a volume of 100 cm3 . The taxa nomenclature for both pollen and macrofossils follows Birks and Birks (1980) : the su⁄x ‘-type’ groups several taxa indistinguishable by their morphology and the pre¢x ‘cf.’ indicates that the most probable taxon name is applied. Diagrams were made with the GPalWin computer programme (Goeury, 1997). The minerogenic content is assumed to reveal past variations in water depths, and consequently,

in the distance between the shoreline and the coring site (Digerfeldt, 1988). It could vary with changes in clastic input via the inlet, but given the small size of the watershed (1.1 km2 ), such changes are thought to be negligible. Macrofossil assemblages are also used to reconstruct water depth and distance to the shoreline. The abundance of aquatic macrofossils depends mainly on water depth, which determines the coverage area and the proximity of submerged plant communities. The abundance of riparian macrofossils decreases with the distance from the shore, although this relationship is in£uenced by the spatial extent of riparian communities. Due to the basin morphometry (Fig. 2), the relation between water depth and shoreline position appears to be nonlinear: as the lake ¢lled, strong shoreline variations occurred, causing only slight modi¢cations of the water depth at the coring site. Moreover, varying clastic inputs from the inlet may have disturbed macrofossil assemblages by carrying riparian and especially aquatic taxa from shallow to deep waters. 3.2. Results Stratigraphic changes in the sediment, pollen and macrofossil content are presented in Figs. 4, 5 and 6, respectively. The pollen zonation (Fig. 5) allowed us to correlate the marginal pro¢le Herl

PALAEO 3020 24-2-03

56

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Fig. 4. Stratigraphic data and loss-on-ignition of the pro¢le Herl (Lac Hertel).

with a previous diagram from a deep-water core (LaSalle, 1966) and other records in the region (Table 2) (compiled in Muller and Richard, 2001). The taxonomic diversity is summarised for each vegetation belt (aquatic, riparian and upland; Fig. 7). Nine sedimentary units (S1^S9) and seven macrofossil units (M1^M7) were identi¢ed from the sediment composition (Fig. 4) and macrofossil records (Fig. 6), respectively. Comparisons between both records are used to infer past waterlevel £uctuations at Lac Hertel (Fig. 8). Although the sometimes contradictory sedimentary and macrofossil evidence, nine periods characterised

by di¡erent lake levels were recognised. Reconstructed lake levels will be given below with regard to the maximal natural level (MNL), estimated around 3 m lower than the present-day. 3.2.1. Before 8000 cal. BP (hiatus) : low lake level, s 7.50 m lower than the MNL The basal, silty clay (unit S1, Fig. 4) implies lake level was initially high just after Lac Hertel became isolated from the receding Champlain Sea waters (LaSalle, 1966). Pollen (Fig. 5) and macrofossil evidence point to a very sparse tundra during this period. The sparseness of vegetation is indicated by the very low pollen concentrations

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

PALAEO 3020 24-2-03 Fig. 5. Simpli¢ed percentage pollen diagram of the pro¢le Herl (Lac Hertel). Pollen zones correspond to the regional vegetation history (see text). The grey zone marks the sand and Alnus leaves deposit (unit S6, Fig. 4). Dots represent less than 1%. Stratigraphic legend in Fig. 4.

57

58 S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

PALAEO 3020 24-2-03 Fig. 6. Macrofossil concentration diagram of the pro¢le Herl (Lac Hertel). Concentrations are standardised to 100 cm3 . Stratigraphic legend in Fig. 4.

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

59

and by the high amounts of silt and clay eroded from the watershed. Other sites from this region show a progressive change to a herb-rich, then a shrubby tundra, with a parallel increase in pollen concentration (Muller and Richard, 2001). Subsequently, boreal and mixed forests (regional pollen zones 2 and 3a, respectively) became established. The marginal core at Lac Hertel does not register the changes in tundra vegetation, nor are pollen zones 2 or 3a registered. This suggests there was a strong lowering of the water level causing a hiatus in sedimentation sometime during the tundra phase. This hiatus corresponds to the sharp increase in loss-on-ignition at the limit between inorganic and organic sediments (units S1 and S2, respectively; Fig. 4). The emersion at the coring site probably triggered erosion of previously deposited sediments and a subsequent redeposition of older organic material when the local zone was £ooded. Redeposition could explain the old dates obtained at the base of the pro¢le on upland macrofossils (TO-8739 and Beta-154759, Table 1). The presence of Tsuga canadensis seeds, which necessarily present a younger age (Muller and Richard, 2001), suggests that old, eroded material was being deposited at the same time as current one.

the coring site than previously. This is corroborated by the low number of Najas £exilis seeds (unit M2, Fig. 6), which production has been shown to be optimal in light shallow waters (Haas, 1996). Moreover, the regular occurrences of upland wind-dispersed macrofossils (e.g., Betula, Pinus, Tsuga; Fig. 6), which are commonly found in the centre of deep lakes (Birks, 1973), support the idea that the water depth at the coring site was relatively large. However, the constant input of organic material from surrounding forests, shown by the corresponding pollen and macrofossil records of Betula alleghaniensis and Tsuga canadensis (Figs. 5 and 6), suggests an input from the inlet and implies a more e¡ective transport of macroremains from upland plants growing along the brooklet than of riparian taxa mainly located in the quiet, shallow waters o¡ the inlet mouth. Finally, the low macrofossil diversity of riparian taxa (unit M2, Fig. 7), which could be related to the great distance between the shoreline and the coring site but also to a limited development of riparian communities, suggests a lake level 2^3 m lower than the MNL. The water surface would then have been located just at the top of the steep slope (Fig. 2), in a position unfavourable for extensive development of riparian vegetation.

3.2.2. 8000^7600 cal. BP (465^440 cm): intermediate lake level, 2^3 m lower than the MNL The low minerogenic content of sediment (unit S2, Fig. 4) suggests the shoreline was further from

3.2.3. 7600^6600 cal. BP (440^380 cm) : slight lake-level lowering In unit S3 (Fig. 4), the mineral matter mass is abundant though characterised by large £uctua-

Fig. 7. Taxonomic diversity of vegetation belts in the pro¢le Herl (Lac Hertel).

PALAEO 3020 24-2-03

60

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Fig. 8. Lake-level changes at Lac Hertel. The sediment is represented in grey.

tions. However, except for a sand layer near 400 cm depth, the mineral matter mainly consists of very thin mica particles, likely to be easily transported into deep waters. Although the increase in mineral matter may be due to changes in stream inputs, it is more probably related to a lake-level lowering. The absence of signi¢cant changes in the macrofossil assemblages (unit M2, Fig. 7) implies this lowering was rather moderate. 3.2.4. 6600^5100 cal. BP (380^320 cm): high lake level, around 0^1 m lower than the MNL The low minerogenic content (unit S4, Fig. 4) corresponds to gyttja deposition, typical of a relatively deep water. This apparently contradicts the very diversi¢ed aquatic plant remains (unit M3, Figs. 6 and 7), which indicate a low water depth. Submerged and £oating-leaved species are well represented by shallow-water species like Najas £exilis, Najas guadalupensis, Nuphar variegatum, Sparganium, Eleocharis, Potamogeton pusillus-type and Characeae (Haas, 1996; Die¡enbacher-Krall and Halteman, 2000). Moreover, the increase in diversity of riparian plant macroremains (Fig. 7), dominated by Carex sp. and Triadenum fraseri seeds (Fig. 6), seems to indicate that the coring site was close to the shoreline (Die¡enbacher-Krall and Halteman, 2000). However, considering the decrease in minerogenic content between S3 and S4 (Fig. 4), it seems

more probable that the macrofossil record corresponds to widespread riparian and aquatic formations developed on the shallow marginal terrace (Fig. 2) than to a strictly local production. Consequently, water level was probably about 0^1 m lower than the MNL between 6600 and 5100 cal. BP. 3.2.5. 5100^4800 cal. BP (320^280 cm) : progressive lake-level lowering The gyttja deposited during this period contains no macroscopic mineral particles, but loss-onignition reveals a progressive increase in minerogenic content (unit S5, Fig. 4) which is attributed to a gradual water-level lowering. The macrofossil record (unit M4, Fig. 6) is not well characterised, but is similar to that of unit M2. The end of the zone is marked by an abrupt increase in the inorganic matter mass (Fig. 4) and by an enhanced representation of upland macrofossil taxa (Betula alleghaniensis, wood fragments ; Fig. 6), which likely indicate the proximity of the shoreline. 3.2.6. 4800^3400 cal. BP (280^140 cm): low lake level, around 3^5 m lower than the MNL This period shows a straightforward increase in sedimentation rate (Fig. 3), corroborated by very low pollen concentrations (Fig. 5). The sediment is composed of ¢ne sand particles with abundant Alnus incana subsp. rugosa leaves (unit M5,

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Fig. 6). Associated with this deposit of leaves, Alnus pollen peaks (Fig. 5) point to a very local origin of A. incana. Similarly, the macrofossil and pollen or spore records of Glyceria sp. and Onoclea sensibilis (Poaceae and Monolete spores, respectively ; Fig. 5) show local presence of both taxa. Moreover, the macrofossil record shows highest diversity of riparian and upland plants, and low diversity of aquatic taxa (unit M5, Fig. 7). The sediment (unit S6, Fig. 4) is also characterised by high but £uctuating values in the inorganic matter mass, consistent with the fact that the unit is a mixture of sand and Alnus leaves. This unit is probably a litter from an A. incana riparian thicket, and indicates a very low lake level with local water depth less than 1 m. 3.2.7. 3400^2600 cal. BP (140^120 cm): high lake level, 0^1 m lower than the MNL The increase in aquatic taxa diversity (unit M6, Fig. 7), associated with the deposition of gyttja (unit S7, Fig. 4), indicates a water-level rise around 3400 cal. BP. The macrofossil record (Fig. 6) is characterised by numerous submerged, £oating-leaved and emergent aquatic taxa (Characeae, Najas £exilis, Najas guadalupensis, Nymphaea odorata subsp. odorata, Nuphar variegatum, Potamogeton sp., Sagittaria sp., Sparganium sp. and Gloeotrichia sp.), among which several thrive in shallow waters, and by the persistence of Carex seeds, which could indicate the proximity of the shoreline. However, this macrofossil record more probably results from a transport from widespread aquatic and riparian formations developed on the marginal terrace (Fig. 2). Consequently, it is considered to correspond to a high lake level, about 0^1 m lower than the MNL. 3.2.8. 2600^1800 cal. BP (120^80 cm): moderate lake-level lowering The incorporation of sand particles in the gyttja between 120 and 80 cm (unit S8, Fig. 4) probably indicates a lake-level lowering. However, this lowering has been su⁄ciently slight to leave the macrofossil assemblages unchanged (unit M6, Fig. 6). 3.2.9. 1800^0 cal. BP (80^0 cm) : high lake level The taxonomic diversity of all types of vegeta-

61

tion decreases, although aquatic taxa dominate the macrofossil record (unit M7, Fig. 7). Along with the sediment composition (unit S9, Fig. 4), this indicates a marked increase in water level around 1800 cal. BP, and a persistent high water level afterwards. This period however di¡ers from previous high lake-level stages by the abundance of Characeae, unidenti¢ed aquatic plants (leaves), Gloeotrichia sp., Chironomids and Cristatella mucedo (unit M7, Fig. 6). These taxa could indicate the complete submersion of the littoral terrace (Fig. 2), allowing the development of widespread shallow-water communities. A man-made rise of about 3 m occurred from A.D. 1775 to the present, due to construction of a series of dams on the outlet (Fig. 2).

4. Palaeoclimate reconstruction 4.1. Modern data The modern pollen data set from Sawada (2001) and from the Base de Donne¤es Polliniques et Macrofossiles du Que¤bec (BDPMQ) consists of 1674 modern pollen spectra from northeastern North America between 90‡W and 50‡W and between 35‡N and 80‡N (Fig. 9). 31 taxa (Table 3) were selected among 85 for their climatic signi¢cance, as shown by principal component analysis. Pollen percentages are calculated relative to the sum of those 31 taxa. The modern climate data set (Fig. 10) is provided by Environment Canada (1994) and the U.S. National Climatic Data Centre (1994). Monthly mean temperature (1702 stations), monthly minimal temperature (1889 stations), total precipitation (2519 stations) and sunshine (inverse of cloudiness; 120 stations) were smoothly interpolated to the 1674 pollen sites using an elevationally-sensitive interpolation. These climatic parameters are used to estimate several climatic (annual precipitation, July temperature, annual temperature and January temperature) and bioclimatic variables (growing degree days and actual evapotranspiration). The growing degree days above 0 and 5‡C are calculated after a daily interpolation of the monthly temperature by summing the daily temperature above 0 and 5‡C,

PALAEO 3020 24-2-03

62

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Table 2 Site characteristics Site name

Site code

Nature

Longitude

Latitude

Altitude

Size

Number of dates

(m)

(ha)

Atocas

Ato

lake

373.311

45.543

114

1.2

4

Bromont Hertel She¡ord St-Calixte Tania Tortue Yamaska

Brom1 Herl She Cal Tania Tor Yam

lake lake carr lake lake lake lake

372.670 373.153 372.585 373.868 374.304 373.317 372.872

45.265 45.547 45.359 45.961 45.775 45.546 45.458

135 173 282 261 305 137 265

50 28.4 4.7 1.7 2 2.5 2

8 3 10 4 6 3 7

References

Gauthier, 1981; Muller and Richard, 2001 Muller and Richard, 2001 this study Richard, 1977a, 1978 Muller and Richard, 2001 Richard, 1977b Gauthier, 1981 Muller and Richard, 2001

Coordinates are noted in decimals.

respectively, and the actual evapotranspiration is obtained from the hydrological model of BIOME1 (Prentice et al., 1992).

cal. BP) and the Ambrosia rise (250^200 cal. BP) (Muller and Richard, 2001). 4.3. Methods

4.2. Chronology of the pollen diagrams Pollen diagrams from seven sites (Fig. 1, Table 2), located in a 80 km radius around Montre¤al City, were chosen for their good chronological control (3^10 radiocarbon dates). Two regional palynological events considered as synchronous throughout the study area and used in the Lac Hertel chronology were also used here as chronological pollen markers: the Tsuga decline (V5500

The reconstruction of past climate in the St. Lawrence lowlands is based on the modern analogue method (Guiot et al., 1989; Guiot, 1990), constrained by lake-level data (Guiot et al., 1993; Cheddadi et al., 1997; Magny et al., 2001). The method consists of ¢nding a few spectra similar to a fossil spectrum in a data set of modern pollen spectra. The correlation between modern and fossil spectra is based on a chord distance measurement (Overpeck et al., 1985). The spectra, called Table 3 Taxa used in the climate reconstruction

80 70

60

50

40

Fig. 9. Distribution map of modern pollen sites.

Trees

Shrubs

Herbs

Abies Acer Betula Carpinus Carya Cupressaceae Fagus Fraxinus Juglans Picea Pinus strobus Populus Quercus Tilia Tsuga Ulmus

Alnus Corylus Myrica Salix

Artemisia Asteraceae Brassicaceae Caryophyllaceae Cyperaceae Fabaceae cf. Oxyria Poaceae Ranunculaceae Rubiaceae Saxifragaceae

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72 80

63

80 70

70

60

60

50

50

40

40

Mean monthly temperature

Minimal monthly temperature

80

80

70

70

60

60

50

50

40

40

Total precipitation

Sunshine percentage Fig. 10. Distribution map of modern climate data.

analogues, are sorted out using the values of P3E (P: annual precipitation ; E: actual evaporation): if the P3E of the analogue is higher than present but the lake level corresponding to the fossil spectra is lower, it is rejected, and conversely. In that constraint, a lake-level change 6 3 m (relative to the maximal natural level) is assumed to represent a change 6 130 mm for P3E (obtained by dividing 3 m by the natural lake:watershed ratio 23), and a lake-level change s 3 m is assumed to rep-

resent a change s 130 mm for P3E. The method is applied separately on all the pollen spectra of each site and the eight best analogues were selected for each one. The reconstructed climate values estimated separately for each pollen diagram were then averaged to time slices of 250 years to get a single regional reconstruction for the study area. Computer procedures were performed using the 3Pbase programme (Guiot and Goeury, 1996).

PALAEO 3020 24-2-03

64

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

The drastic decline in Tsuga canadensis at ca. 5500 cal. BP has been attributed to the combined impact of a climate shift (Haas and McAndrews, 2000) and a phytophagous insect outbreak (Davis, 1981; Allison et al., 1986; Bhiry and Filion, 1996). Although Davis et al. (2000) negate the idea of climate as a possible trigger to the Tsuga decline, we hypothesise that the inclusion of T. canadensis in our data set may create a bias in the climatic reconstruction. In order to evaluate the in£uence of this taxon on the climate reconstruction, the modern analogue method was applied both including and excluding it from the pollen sum. 4.4. Climate^vegetation relationships

Fig. 11. Modern climate^vegetation relationships. The wetness is approximated by E/PE and the runo¡ by P3E. E: actual evaporation; PE: potential evaporation; P: annual precipitation; GDD0, GDD5: growing degree days above 0 and 5‡C, respectively.

The ability of the modern analogue method to reconstruct climate is tested by comparing actual and estimated modern climate data (Fig. 11). Estimated values were generated for each modern spectrum by sorting out the best analogues from the entire modern spectra data set exclusive of the considered spectrum. In this validation step, the lake-level constraint is not used. The pollen spectra are more closely correlated with temperature variables (r2 around 0.90), than with precipitation or moisture balance (r2 around 0.72). Compared to previous studies (Guiot et al., 1993; Davis et al., 2000), these r2 point to a particularly close relationship between climate and vegetation in northeastern North America. As an example, a very high relationship is shown for the runo¡ (Fig. 11), which is theoretically considered to be a better predictor of lake-level changes than of vegetation changes (Guiot et al., 1993; Harrison et al., 1993). 4.5. Results Climate reconstructions obtained by excluding

Fig. 12. Climate reconstruction. (A) Unconstrained. (B) Constrained by lake levels. Full lines correspond to reconstructions including Tsuga canadensis (grey ones de¢ne the con¢dence intervals) and dashed lines to reconstructions excluding it (shown only for precipitation). Dots on the right axes of graphs represent actual modern values, calculated by averaging meteorological data from the Montre¤al lowlands area. The number of meteorological sites used in this calculation is noted in brackets. Con¢dence intervals combine the di¡erences of climate estimates between sites and the number of modern analogues found. Larger intervals between 14 000 and 10 000 cal. BP are due to fossil pollen spectra with few or without modern analogues.

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

PALAEO 3020 24-2-03

65

66

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Tsuga canadensis show no signi¢cant di¡erences for temperature and only minor departures for precipitation during the entire postglacial period (Fig. 12). Paradoxically, the steep pollen decline of Tsuga recorded at 5500 cal. BP seems to increase the reconstructed precipitation (Fig. 12). The inclusion of Tsuga within the pollen sum consequently is considered to create no bias in the climate reconstruction. Similarly, the use of lake-level changes as a constraint results in negligible climate changes, even between 5000 and 3500 cal. BP (Fig. 12), a period corresponding to the low lake level inferred from the deposit of sand and Alnus leaves (Fig. 6). Surprisingly, a test realised by accounting for the modern lake: watershed ratio (1:13) shows a decrease in annual precipitation of about 150 mm and an increase in July temperature of about 1‡C during this interval. This result shows that the lake:watershed ratio controls the sensitivity of lakes to climate changes, and points to the limit of the use of lake levels for detecting climate changes. Since the exact climatic signals associated with lake-level changes in southern Que¤bec and Acadia are not established (Almquist et al., 2001), as it is for other parts of eastern North America (Harrison, 1989), the use of lake-level reconstruction of Lac Hertel as a constraint remains subject to caution. With the exception of annual temperature, all of the reconstructions (with and without constraint, with and without Tsuga) fail to predict the modern climate of the St. Lawrence lowlands. Notably, our reconstruction appears to underestimate the seasonal contrasts of temperature (July^ January) and the growing degree day values (Fig. 12). This could be due to the fact that observed climatic values represent only 30 year averages (between 1960^1990) while pollen-based estimations concern the last 250 years.

5. Discussion 5.1. Holocene climatic moisture patterns in St. Lawrence lowlands 5.1.1. 10 000^7000 cal. BP During the early Holocene, southern Que¤bec

was characterised by very low annual precipitation (Fig. 12) and low lake levels (Lavoie and Richard, 2000a; Fig. 8). Similar conditions are shown by lake-level records from northeastern United States (Harrison and Metcalfe, 1985; Harrison, 1989; Webb et al., 1993b) and from the eastward expansion of the prairie across the Midwest (Webb et al., 1983; Winkler et al., 1986). This early Holocene widespread dry interval was believed to be the result of a greater summer insolation than at present (Kutzbach and Guetter, 1986). However, the constrained reconstruction (Fig. 12B) does not show the peak of aridity recognised soon after 8000 cal. BP in northeastern United States (Webb et al., 1983; Harrison and Metcalfe, 1985; Harrison, 1989) and in the western Great Lakes region (Davis et al., 2000). At this time, Lac Hertel exhibits an increase in water level rapidly followed by a slight lowering (Fig. 8), which could reveal subtle climatic events not recorded by vegetation. Moreover, charcoal in£ux and pollen percentages of Populus and Pinus banksiana suggest shifts, not recorded by eastern North American lake levels, from dry to wet summers in southern Que¤bec around 9000 cal. BP (Richard, 1994) and between 8000 and 7000 cal. BP (Carcaillet and Richard, 2000; Carcaillet et al., 2001). These di¡erences between vegetation and lake-level records may be related to the greater sensitivity of plants to changes in summer precipitation and of lakes to changes in winter precipitation (Carcaillet and Richard, 2000). 5.1.2. 7000^5000 cal. BP Between 7000 and 6000 cal. BP, annual precipitation increased in the St. Lawrence lowlands (Fig. 12). This feature is shown by both constrained and unconstrained reconstructions, pointing to a similar climate inference derived from pollen and lake-level data. Southern Que¤bec lakes have recorded a mid-Holocene water-level rise: Lac Hertel between 6600 and 5000 cal. BP (Fig. 8) and Lake Albion between 6900 and 6100 cal. BP (Lavoie and Richard, 2000a). The di¡erent timing could be related to the di¡erent geographical situations or to dating problems, particularly at Lake Albion (Lavoie and Richard, 2000a), but it could also re£ect a non-climatic

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

control on lake-level change, such as basin permeability changes or modi¢cations in outlet morphometry (Dearing and Foster, 1986). High lake levels (Harrison, 1989; Webb et al., 1993a) and increased precipitation reconstructed from pollen data (Davis et al., 2000) reveal similar trends in northeastern and midwestern USA, respectively. However, low lake levels are reported during the same period in several sites from Ontario (Szeicz and MacDonald, 1991; Yu and McAndrews, 1994; Bunting et al., 1996; Yu et al., 1996; Campbell et al., 1997) and Maine (Almquist et al., 2001). 5.1.3. 5000^3500 cal. BP The Lac Hertel sedimentary and macrofossil data (Figs. 4 and 5, respectively) suggest the occurrence of a dry phase in southern Que¤bec between 5000 and 3500 cal. BP, although no major climatic change was recorded by pollen during this period (Fig. 12). A dry interval was also recorded between 6100 and 4400 cal. BP in Lake Albion (Lavoie and Richard, 2000a). Similar conditions, documented between 7000 and 3000 cal. BP in Subarctic Que¤bec (Payette and Filion, 1993), in Ontario (Sreenivasa and Duthie, 1973; Szeicz and MacDonald, 1991; Yu and McAndrews, 1994; Bunting et al., 1996; Yu et al., 1996, 1997; Campbell et al., 1997), in Maine (Almquist et al., 2001) and in Wisconsin (Winkler et al., 1986; Baker et al., 1992), point to a widespread aridity. This discrepancy between pollen and lake-level records suggests that this aridity is primarily a¡ecting the winter season, rather than the growing one. Notably, we can note that the Tsuga decline dated at about 5500 cal. BP precedes the lowering of lake level. Our results suggest that the Tsuga decline was not related in southern Que¤bec to a drought episode, as it may have been in southern Ontario (Haas and McAndrews, 2000). 5.1.4. 3500^0 cal. BP The ¢nal rise of southern Que¤bec lake levels (Lavoie and Richard, 2000a; Fig. 8) is consistent with trends observed in northeastern United States (Harrison and Metcalfe, 1985; Harrison, 1989; Almquist et al., 2001) and in southern On-

67

tario (Sreenivasa and Duthie, 1973; Szeicz and MacDonald, 1991; Yu and McAndrews, 1994; Bunting et al., 1996; Yu et al., 1996, 1997; Campbell et al., 1997). This widespread increase in lake levels suggests increased precipitation throughout northeastern North America. This inference is supported by our regional climatic reconstruction which shows increasing precipitation over the 3500 last years (Fig. 12). Climatic multiproxy comparison presenting contradictory evidences from lake levels and ¢res (Carcaillet and Richard, 2000) suggests that lake levels depend mostly on winter precipitation, while ¢res depend on summer one. 5.2. Holocene temperature patterns in the St. Lawrence lowlands The annual temperature amplitude presents decreasing trend during the postglacial period, particularly well marked during the late-glacial (this trend surprisingly presents a very strong inverse relationship with annual precipitation: r2 s 3 0.97). Similar patterns were observed for sites located at similar latitudes near Lake Michigan (Davis et al., 2000). This enforces the hypothesis of a linkage between temperature seasonality and Earth orbital parameters suggested by these authors on the basis of a latitudinal transect. However, seasonality also represents a measure of continentality (Guiot et al., 1993), and thus depends on the proximity of water masses. The in£uence of Great Lakes is clearly shown by comparing an inland site to an island one, both located at similar latitudes (Davis et al., 2000). The similarity between Holocene seasonality trends in the St. Lawrence lowlands and on Great Lakes shores suggests that the St. Lawrence River may have in£uenced the past temperature pattern within the southern Que¤bec lowlands. 5.2.1. 14 000^11 500 cal. BP The late-glacial was characterised in southern Que¤bec by the occurrence of pro-glacial lakes and seas (Hillaire-Marcel, 1979; Parent and Occhietti, 1988, 1999). These cold waters induced harsh climatic conditions throughout the St. Lawrence lowland region (Ganglo¡ et al., 1971;

PALAEO 3020 24-2-03

68

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

Ganglo¡, 1981). According to our climatic reconstruction (Fig. 12), the annual temperature of the late-glacial period was around 33.7 R 0.9‡C (about 9‡C below present). An increase in temperature was recorded around 12 500 cal. BP (Fig. 12), simultaneously with regional a¡orestation and the beginning of organic sediment deposition in lakes in southern Que¤bec (Richard, 1994). Increased temperatures are reported at this date in southern Minnesota and Wisconsin (Webb and Bryson, 1972; Webb and Clark, 1977) and by sea surface temperature in the Gulf of St. Lawrence (de Vernal et al., 1993). This regional warming at 12 500 cal. BP was probably partly related to higher summer solar radiation (Kutzbach and Guetter, 1986; Ritchie and MacDonald, 1986; Cwynar and Spear, 1991). 5.2.2. 11 500^10 500 cal. BP Sea surface temperatures in the Gulf of St. Lawrence and midwestern USA temperatures attained their modern values between 11 500 and 10 500 cal. BP (Webb and Clark, 1977; de Vernal et al., 1993), or may even have been a little higher (Webb and Bryson, 1972). The St. Lawrence lowlands, however, were characterised by a temperature 1.7‡C below present at 11 000 cal. BP (3.8 R 0.4‡C ; Fig. 12). This di¡erence probably re£ects the proximity of inland ice and the impact of cold, catabatic winds, and the in£uence of cold water discharge from glacial lakes, which drained the region north of the Great Lakes (Lewis and Anderson, 1989; Anderson and Lewis, 1992). 5.2.3. 10 500^4500 cal. BP The spread of Pinus strobus and other thermophilous species between 9000 and 4500 cal. BP indicates a warmer climate (Richard, 1994). According to our reconstruction, annual temperature increased progressively between 11 500 and 8000 cal. BP (Fig. 12), as did the length of growing season (as indicated by degree days). Our results also show a slight temperature decrease from 9000 to 8250 cal. BP, mainly in July temperature (Fig. 12). This climate change is congruent with a pronounced climate cooling reported between 8900 and 8300 cal. BP from an isotopic record in an annually-laminated Minnesota lake (Hu et al.,

1999), and may con¢rm the in£uence of atmospheric circulation and especially of summer cooling in Arctic airmass suggested by these authors. After this cool episode, annual temperatures in the St. Lawrence lowlands reach values ca. 0.5‡C above the estimated modern value and 0.2‡C above the actual one (5.7 R 0.5‡C; Fig. 12). Temperatures higher than today have also been reported between 9000 and 7000 cal. BP in the western Great Lakes region (Davis et al., 2000) and in Gulf of St. Lawrence (de Vernal et al., 1993). The St. Lawrence lowlands experienced summer cooling from 5750 to 5000 cal. BP, which culminates at 5500 cal. BP (Fig. 12) simultaneously to the Tsuga decline. 5.2.4. 4500^0 cal. BP Richard (1994) notes that the increase in boreal elements (Picea, Abies) within the regional vegetation after 4500 cal. BP, could indicate a climatic cooling. However, in southern Que¤bec and northern USA, these taxa preferentially grow on peatlands. Their increase in abundance could thus also re£ect their local dynamics linked with autogenic development of peatlands, as reported at Frontenac peatland (Lavoie and Richard, 2000b). Despite this, the progressive summer cooling is observed over the 4000 last years in the Gulf of St. Lawrence (de Vernal et al., 1993) and in the midwestern USA (Webb and Bryson, 1972; Webb and Clark, 1977; Davis et al., 2000). Moreover, this pattern appears to be associated in southern Que¤bec (Fig. 12) and in the western Great Lakes region (Davis et al., 2000) with increasing winter temperature, which could explain the persistence of thermophilous species (Richard, 1994).

6. Conclusions This study provides the ¢rst quantitative climate reconstruction for southern Que¤bec. Most of the reconstructed climate changes are registered simultaneously in the Gulf of St. Lawrence, the southwestern Que¤bec lowlands and the midwestern USA. The reconstructions show : (1) a dry and cold late-glacial episode, especially harsh in southern Que¤bec due to the proximity of inland ices,

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

and to the in£uence of pro-glacial lakes and seas; (2) an abrupt warming between 12 500 and 11 000 cal. BP caused by increasing summer insolation during regional ice retreat; (3) an arid period from 10 000 to 6500 cal. BP; (4) a brief cooling event between 9000 and 8000 cal. BP, only observed from upland records and possibly related to a summer cooling of Arctic airmasses ; (5) a temperature optimum around 8000 cal. BP; and ¢nally (6) a progressive decrease in summer temperature and an increase in (winter ?) precipitation over the 4500 last years. The consistency of our reconstruction with previous results points to the reliability of the application of the modern analogue method at a regional scale. However, the south^north vegetation spread following deglaciation of southern Que¤bec was not in total equilibrium with climate: several other factors, such as seed dispersion characteristics or physiography, played a role in this process (Muller and Richard, 2001). This implies to consider carefully the minor £uctuations of reconstructed curves, especially during late-glacial and early postglacial periods. The reconstructed climate, if not representing exactly the actual past climate, corresponds at least to the minimal one necessary for the postglacial vegetation migration. Our results shed light on two particular vegetational features of the St. Lawrence lowlands. The decline of Tsuga canadensis has been attributed to a phytophagous insect outbreak (e.g., Davis, 1981; Allison et al., 1986; Bhiry and Filion, 1996) and, on the basis of a lake-level reconstruction in southern Ontario, to drought (Haas and McAndrews, 2000). The Tsuga decline appears to have occurred in southern Que¤bec independently from any precipitation change and concurrently with a pronounced summer cooling. Since the Tsuga decline does not appear to be associated everywhere with the same climate change, the role of climate was probably limited to favouring or obstructing locally the spread of the hemlock looper (Lamdina ¢scellaria, Lepidoptera). Macrofossil analyses conducted on the marginal core of Lac Hertel reveal past occurrences of several taxa not found today on Mont St. Hilaire (Maycock, 1961), including Ceratophyllum echinatum, Decodon verticillatus, Najas guadalupensis

69

and Nymphaea odorata subsp. odorata. The sporadic occurrences of N. guadalupensis during more than six millennia suggest that this species could be present albeit unidenti¢ed today on Mont St. Hilaire. Other species were encountered only once in the whole pro¢le. In regard to the present-day rarity of C. echinatum and N. guadalupensis in Que¤bec (Bouchard et al., 1983; Lavoie, 1992), their past occurrences in Lac Hertel may reveal past modi¢cations of their northern distribution limit, likely to be induced by climate change. Finally, this study highlights seasonal patterns in postglacial climate history. It corroborates the idea that pollen and charcoal data would mainly indicate summer precipitation (Carcaillet and Richard, 2000) while lake levels re£ect winter precipitation (Vassiljev et al., 1998). These relationships make sense, specially for northern zones characterised by harsh and long winters. The high amount of water liberated by spring snow melt (50^60% of the annual runo¡ in southeastern Que¤bec lowlands ; Plamondon et al., 1996) exceeds largely the stockage capacity of soils. It is then believed to contribute more to lake levels than to groundwater recharge. In comparison, summer precipitation, which is regularly dispersed during the whole season, contributes mostly to the soil stockage and consequently to the vegetation development. These considerations point to the interest of merging summer and winter precipitation proxy data by the way of pollen-based precipitation reconstructions constrained by lake levels.

Acknowledgements We thank Hans Asnong, Alayn Larouche, Isabelle Legeai and Nicole Morasse for ¢eld and laboratory assistance, and Martin Lechowicz, MarcAndre¤ Langlois and the World Biosphere Reserve of Mont St. Hilaire for the authorisation and the facilities to work. The research has been supported by the Natural Sciences and Engineering Research Council of Canada and the Fond pour la Formation de Chercheurs et l’Aide a' la Recherche of Que¤bec, grants to P.J.H.R. S.D.M. bene¢ted from a French fellowship (Allocation

PALAEO 3020 24-2-03

70

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

de recherche du Ministe're de la Recherche et de la Technologie) and was supported by the Peatland Carbon Study Program (NSERC-Strategic). Thoughtful comments by Hans Asnong, Christopher Carcaillet, Gunnar Digerfeldt, Sandy Harrison, Alayn Larouche, Martin Lavoie and Cathy Whitlock were greatly appreciated.

References Allison, T.B., Moeller, R.E., Davis, M.B., 1986. Pollen in laminated sediments provides evidence for a mid-Holocene forest pathogen outbreak. Ecology 67, 1101^1105. Almquist, H., Die¡enbacher-Krall, A.C., Flanagan-Brown, R., Sanger, D., 2001. The Holocene record of lake levels at Mansell Pond, Central Maine, USA. The Holocene 11, 189^201. Anderson, T.W., Lewis, C.F.M., 1992. Climatic in£uences of deglacial drainage in southern Canada at 10 to 8 ka suggested by pollen evidence. Ge¤ogr. Phys. Quat. 46, 255^272. Baker, R.G., Maher, L.J., Chumbley, C.A., van Zant, K.I., 1992. Patterns of Holocene environmental change in the midwestern United States. Quat. Res. 37, 379^389. Berglund, B.E., Ralska-Jasiewiczowa, M., 1986. Pollen analysis and pollen diagrams. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 455^484. Bhiry, N., Filion, L., 1996. Mid-Holocene hemlock decline in eastern North America linked with phytophagous insect activity. Quat. Res. 45, 312^320. Birks, H.H., 1973. Modern macrofossil assemblages in lake sediments in Minnesota. In: Birks, H.J.B., West, R.G. (Eds.), Quaternary Plant Ecology. Blackwell Scienti¢c Publications, Oxford, pp. 173^189. Birks, H.J.B., Birks, H.H., 1980. Quaternary Palaeoecology. E. Arnold, London. Bouchard, A., Maycock, P.F., 1978. Les fore“ts de¤cidues et mixtes de la re¤gion appalachienne du sud que¤be¤cois. Nat. Can. 105, 383^415. Bouchard, A., Barabe¤, D., Dumais, M., Hay, S., 1983. Les plantes vasculaires rares du Que¤bec. Syllogeus 48, Muse¤es nationaux du Canada, Ottawa. Bunting, M.J., Warner, B.G., Aravena, R., 1996. Late-Quaternary vegetation dynamics and hydroseral development in a Thuja occidentalis swamp in southern Ontario. Can. J. Earth Sci. 33, 1439^1456. Campbell, D.R., Duthie, H.C., Warner, B.G., 1997. Post-glacial development of a kettle-hole peatland in southern OnŁ coscience 4, 404^418. tario. E Carcaillet, C., Richard, P.J.H., 2000. Holocene changes in seasonal precipitation highlighted by ¢re incidence in eastern Canada. Clim. Dyn. 16, 549^559. Carcaillet, C., Bergeron, Y., Richard, P.J.H., Fre¤chette, B., Gauthier, S., Prairie, Y.T., 2001. Change of ¢re frequency

in the eastern Canadian boreal forests during the Holocene: does vegetation composition or climate trigger the ¢re regime? J. Ecol. 89, 930^946. Cheddadi, R., Yu, G., Guiot, J., Harrison, S.P., Prentice, I.C., 1997. The climate of Europe 6000 years ago. Clim. Dyn. 13, 1^9. Cwynar, L.C., Spear, R.W., 1991. Reversion of forest to tundra in the central Yukon. Ecology 72, 202^212. Davis, M.B., 1981. Outbreaks of forest pathogens in Quaternary history. In: 4th international Palynological Conference, vol. 3, pp. 216^227. Davis, M.B., Woods, K.D., Webb, S.L., Futyma, R.P., 1986. Dispersal versus climate: Expansion of Fagus and Tsuga into the Upper Great Lakes region. Vegetatio 67, 93^103. Davis, M.B., Douglas, C., Calcotte, R., Cole, K.L., Winkler, M.G., Flakne, R., 2000. Holocene climate in the Western Great Lakes National Parks and Lakeshores: Implication for future climate change. Conserv. Biol. 14, 968^983. Dean, W.E., Jr., 1974. Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: comparison with other methods. J. Sediment. Petrol. 44, 242^248. Dearing, J.A., Foster, I.D.L., 1986. Lake sediments and palaeohydrological studies. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 67^90. de Vernal, A., Guiot, J., Turon, J.-L., 1993. Late and Postglacial paleoenvironments of the Gulf of St. Lawrence: marine and terrestrial palynological evidence. Ge¤ogr. Phys. Quat. 47, 167^180. Die¡enbacher-Krall, A.C., Halteman, W.A., 2000. The relationship of modern plant remains to water depth in alkaline lakes in New England, USA. Paleolimnology 24, 213^229. Digerfeldt, G., 1986. Studies on past lake-level £uctuations. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 127^ 143. Digerfeldt, G., 1988. Reconstruction and regional correlation of Holocene lake-level £uctuations in Lake Bysjo«n, South Sweden. Boreas 17, 165^182. Environment Canada, 1994. Canadian Monthly Climate Data and 1961-90 Normals, version 3.0. CD-Rom. Faegri, K., Iversen, J., 1989. In: Faegri, K., Kaland, P.E., Kzrywinski, K. (Eds.), Textbook of Pollen Analysis. Wiley, Chichester. Ganglo¡, P., 1981. Signi¢cation pale¤oclimatique des formes pe¤riglaciaires reliques du Que¤bec me¤ridional. Biul. Peryglac. 28, 187^196. Ganglo¡, P., Clermont, J.P., Paquette, G., 1971. Le proble'me du tardiglaciaire au Que¤bec me¤ridional. Rev. Ge¤ogr. Montre¤al 25, 305^308. Gauthier, R., 1981. Histoire de la colonisation ve¤ge¤tale des Monte¤re¤giennes: deux sites du mont Saint-Bruno. Unpublished M.Sc. Thesis, Universite¤ de Montre¤al, Canada. Goeury, C., 1997. GPalWin: gestion, traitement et repre¤sentation des donne¤es de la pale¤oe¤cologie. In: Actes du XVe symposium de l’APLF, p. 31.

PALAEO 3020 24-2-03

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72 Grandtner, M.M., 1966. La ve¤ge¤tation forestie're du Que¤bec me¤ridional. Les Presses de l’Universite¤ Laval, Que¤bec. Greller, A.M., 1989. Correlation of warmth and temperatureness with the distributional limits of zonal forests in eastern North America. Bull. Torrey Bot. Club 116, 145^163. Grosse-Brauckmann, G., 1986. Analysis of vegetative plant macrofossils. In: Berglund, B.E. (Ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 591^618. Guiot, J., 1990. Methodology of the last climatic cycle reconstruction in France from pollen data. Palaeogeogr. Palaeoclimatol. Palaeoecol. 80, 49^69. Guiot, J., Goeury, C., 1996. 3Pbase ^ a software for statistical analysis of paleoecological and paleoclimatological data. Dendrochronologia 14, 123^135. Guiot, J., Pons, A., de Beaulieu, J.-L., Reille, M., 1989. A 140,000-year continental climate reconstruction from two European pollen records. Nature 338, 309^313. Guiot, J., Harrison, S.P., Prentice, I.C., 1993. Reconstruction of Holocene precipitation patterns in Europe using pollen and lake-level data. Quat. Res. 40, 139^149. Haas, J.N., 1996. Pollen and plant macrofossil evidence of vegetation change at Wallisellen-Langachermoos (Switzerland) during the Mesolithic-Neolithic transition 8500 to 6500 years ago. Diss. Bot. 267, 1^267. Haas, J.-N., McAndrews, J.H., 2000. The summer drought related hemlock (Tsuga canadensis) decline in eastern North America 5,700 to 5,100 years ago. In: McManus, K.A., Shields, K.S., Souto, D.R. (Eds.), Proceedings: Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America. General Technical Report NE267. United States Department of Agriculture, Forest Service, Northeastern Research Station. Durham, New Hampshire, USA, June 22^24, 1999, pp. 81^88. Harrison, S.P., 1989. Lake levels and climatic change in eastern North America. Clim. Dyn. 3, 157^167. Harrison, S.P., Metcalfe, S.E., 1985. Spatial variations in lake level since the last glacial maximum in the Americas north of the equator. Z. Gletsch.kd. Glazialgeol. 21, 1^15. Harrison, S.P., Prentice, I.C., Guiot, J., 1993. Climatic controls on Holocene lake-level changes in Europe. Clim. Dyn. 8, 189^200. Hillaire-Marcel, C., 1979. Les mers post-glaciaires du Que¤bec: quelques aspects. Unpublished Ph.D. Thesis, Universite¤ Pierre et Marie Curie, Paris VI, France. Horva¤th, L., Gault, R.A., 1990. The mineralogy of Mont Saint-Hilaire, Que¤bec. In: Wilson, W.E. (Ed.), The Mineralogical Record, Tucson. Hu, F.S., Slawinski, D., Wright, H.E., Jr., Ito, E., Johnson, R.G., Kelts, K.R., McEwan, R.F., Boedigheimer, A., 1999. Abrupt changes in North American climate during early Holocene times. Nature 400, 437^440. Huntley, B., Bartlein, P.J., Prentice, I.C., 1989. Climatic control of the distribution and abundance of beech in Europe and North America. J. Biogeogr. 16, 551^560. Kutzbach, J.E., Guetter, P.J., 1986. The in£uence of changing orbital parameters and surface boundary conditions on cli-

71

mate simulations for the past 18,000 years. J. Atmos. Sci. 43, 1726^1759. LaSalle, P., 1966. Late Quaternary vegetation and glacial history in the St. Lawrence Lowlands, Canada. Leidse Geol. Meded. 38, 91^128. Laverdie're, C., Bertrand, J., Guimont, P., 1972. La carte ge¤oŁ cologie de la Zone de morphologique; notes explicatives. E l’Ae¤roport International de Montre¤al (EZAIM), Rapport pre¤liminaire no 11, Centre de recherches e¤cologiques de Montre¤al, Montre¤al. Lavoie, G., 1992. Plantes vasculaires susceptibles d’e“tre de¤signe¤es menace¤es ou vulne¤rables au Que¤bec. Ministe're de l’Environnement, Que¤bec. Lavoie, M., Richard, P.J.H., 2000a. Postglacial water-level changes of a small lake in southern Que¤bec, Canada. The Holocene 10, 621^634. Lavoie, M., Richard, P.J.H., 2000b. The role of climate on the developmental history of Frontenac Peatland, southern Quebec. Can. J. Bot. 78, 1^18. Lewis, C.F.M., Anderson, T.W., 1989. Oscillations of levels and cool phases of the Laurentian Great Lakes caused by in£ows from glacial Agassiz and Barlow-Ojibway. J. Paleolimnol. 2, 99^146. Livingstone, D.A., 1955. A lightweight piston sampler for lake deposits. Ecology 36, 137^139. Magny, M., Guiot, J., Schoellammer, P., 2001. Quantitative reconstruction of Younger Dryas to mid-Holocene paleoclimates at Le Locle, Swiss Jura, using pollen and lake-level data. Quat. Res., in press. Maycock, P.F., 1961. Botanical studies on Mont St. Hilaire, Rouville County, Quebec. Can. J. Bot. 39, 1293^1325. Muller, S.D., Richard, P.J.H., 2001. Post-glacial vegetation migration in conterminous Montre¤al Lowlands, southern Que¤bec. J. Biogeogr. 28, 1169^1193. Overpeck, J.T., Webb, T., III, Prentice, I.C., 1985. Quantitative interpretation of fossil pollen spectra: dissimilarity coe⁄cients and the method of modern analog. Quat. Res. 23, 87^108. Parent, M., Occhietti, S., 1988. Late Wisconsinan deglaciation and Champlain Sea invasion in the St. Lawrence valley, Que¤bec. Ge¤ogr. Phys. Quat. 42, 215^246. Parent, M., Occhietti, S., 1999. Late Wisconsinan deglaciation and glacial lake development in the Appalachian uplands and piedmont of southeastern Que¤bec. Ge¤ogr. Phys. Quat. 53, 117^135. Payette, S., Filion, L., 1993. Holocene water-level £uctuations of a subarctic lake at the tree line in northern Que¤bec. Boreas 22, 7^14. Plamondon, A.P., Barry, R., Bernier, P., Pre¤vost, M., Stein, J., Trottier, F., 1996. Hydrologie forestie're et ame¤nagement du bassin hydrographique. In: Be¤rard, J.A., Co“te¤, M. (Eds.), Manuel de Foresterie, Ordre des Inge¤nieurs Forestiers du Que¤bec. Les Presses de l’Universite¤ Laval, Que¤bec, pp. 281^329. Prentice, I.C., 1986. Vegetation responses to past climatic variation. Vegetatio 67, 131^141. Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Mon-

PALAEO 3020 24-2-03

72

S.D. Muller et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 193 (2003) 51^72

serud, R.A., Solomon, A.M., 1992. A global biome model based on plant physiology and dominance, soil properties and climate. J. Biogeogr. 19, 117^134. Richard, P., 1977. Histoire post-wisconsinienne de la ve¤ge¤tation du Que¤bec me¤ridional par l’analyse pollinique. Service de la Recherche, Direction ge¤ne¤rale des Fore“ts, Ministe're des terres et fore“ts du Que¤bec. Richard, P., 1977b. Ve¤ge¤tation tardiglaciaire au Que¤bec me¤ridional et implications pale¤oclimatiques. Ge¤ogr. Phys. Quat. 31, 161^176. Richard, P., 1978. Histoire tardiglaciaire et postglaciaire de la ve¤ge¤tation au mont She¡ord, Que¤bec. Ge¤ogr. Phys. Quat. 32, 81^93. Richard, P.J.H., 1987. Le couvert ve¤ge¤tal du Que¤bec-Labrador et son histoire postglaciaire. Notes Documents 87-01, De¤partement de Ge¤ographie, Universite¤ de Montre¤al, Montre¤al. Richard, P.J.H., 1989. Les patrons de colonisation ve¤ge¤tale post-wisconsienne au Que¤bec-Labrador. In: Fulton, R.J. (Ed.), Le Quaternaire du Canada et du Groenland. Commission ge¤ologique du Canada, Ge¤ologie du Canada, vol. 1, pp. 552^559. Richard, P.J.H., 1993. Origine et dynamique postglaciaire de la fore“t mixte au Que¤bec. Rev. Palaeobot. Palynol. 79, 31^68. Richard, P.J.H., 1994. Postglacial palaeophytogeography of the eastern St. Lawrence River Watershed and the climatic signal of the pollen record. Palaeogeogr. Palaeoclimatol. Palaeoecol. 109, 137^161. Richard, P.J.H., Larouche, A.C., Lortie, G., 1992. Pale¤ophytoge¤ographie et pale¤oclimats postglaciaires dans l’ouest du Bas-Saint-Laurent, Que¤bec. Ge¤ogr. Phys. Quat. 46, 151^172. Ritchie, J.C., 1986. Climate change and vegetation response. Vegetatio 67, 65^74. Ritchie, J.C., MacDonald, G.M., 1986. The patterns of postglacial spread of white spruce. J. Biogeogr. 13, 527^540. Sawada, M., 2001. Late Quaternary Paleoclimates and Biogeography of North America. Unpublished Ph.D. Thesis, University of Ottawa, Canada. Sreenivasa, M.R., Duthie, H.C., 1973. The postglacial diatom history of Sun¢sh Lake, southwestern Ontario. Can. J. Bot. 51, 1599^1609. Stuiver, M., Reimer, P.J., 1993. Extended 14 C database and revised CALIB radiocarbon calibration program. Radiocarbon 35, 215^230. Szeicz, J.M., MacDonald, G.M., 1991. Postglacial vegetation history of oak savanna in southern Ontario. Can. J. Bot. 69, 1507^1519. U.S. National Climatic Data Centre, 1994. U.S. divisional and station climatic data and normals, vol. 1. CD-Rom. Vassiljev, J., Harrison, S.P., Guiot, J., 1998. Simulating the Holocene lake-level record of Lake Bysjo«n, Southern Sweden. Quat. Res. 49, 62^71.

Webb, R.S., Anderson, K.H., Webb, T., III, 1993a. Pollen response-surface estimates of Late-Quaternary changes in the moisture balance of the northeastern United States. Quat. Res. 40, 213^227. Webb, T., III, 1986. Is vegetation in equilibrium with climate? How to interpret late-Quaternary pollen data. Vegetatio 67, 75^91. Webb, T. III, 1993. Constructing the past from Late-Quaternary pollen data: Temporal resolution and a zoom lens space-time perspective. In: Kidwell, S.M., Behrensmeyer, A.K. (Eds.), Taphonomic Approaches to Time Resolution in Fossil Assemblages. Paleontological Society Short Courses in Paleontology, vol. 6, pp. 79^101. Webb, T., III, Bryson, R.A., 1972. Late- and postglacial climatic change in the Northern Midwest, USA: Quantitative estimates derived from fossil pollen spectra by multivariate statistical analysis. Quat. Res. 2, 70^115. Webb, T., III, Clark, D.R., 1977. Calibrating micropaleontological data in climatic terms: A critical review. Ann. N. Y. Acad. Sci. 288, 93^118. Webb, T. III, Cushing, E.J., Wright, H.E. Jr., 1983. Holocene changes in the vegetation of the Midwest. In: Wright, H.E. Jr. (Ed.), Late Quaternary Environments of the United States. Vol 2: The Holocene. University of Minnesota Press, Minneapolis, MN, pp. 142^165. Webb, T. III, Bartlein, P.J., Harrison, S.P., Anderson, K.H., 1993. Vegetation, lake levels, and climate in eastern North America for the past 18,000 years. In: Wright, H.E. Jr., Kutzbach, J.E., Webb, T. III, Ruddiman, W.F., Street-Perrott, F.A., Bartlein, P.J. (Eds.), Global Climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis, MN, pp. 415^467. Winkler, M.G., Swain, A.M., Kutzbach, J.E., 1986. Middle Holocene dry period in the northern midwestern United States: Lake levels and pollen stratigraphy. Quat. Res. 25, 235^250. Wright, H.E., Jr., 1967. A square-rod piston sampler for lake sediments. J. Sediment. Petrol. 37, 975^976. Woodward, F.I., 1987. Climate and Plant Distribution. Cambridge University Press, New York. Yu, Z.C., McAndrews, J.H., 1994. Holocene water levels at Rice Lake, Ontario, Canada: sediment, pollen and plantmacrofossil evidence. The Holocene 4, 141^152. Yu, Z.C., McAndrews, J.H., Siddiqi, D., 1996. In£uences of Holocene climate and water levels on vegetation dynamics of a lakeside wetland. Can. J. Bot. 74, 1602^1615. Yu, Z.C., McAndrews, J.H., Eicher, U., 1997. Middle Holocene dry climate caused by change in atmospheric circulation patterns: evidence from lake levels and stable isotopes. Geology 25, 251^254.

PALAEO 3020 24-2-03