1. Introduction
Pollen-based climate reconstructions have been extensively used to document climate variability on several time and space scales. However, most of the previous work has consisted of mapping certain time intervals, such as 6 ka (e.g., COHMAP Members 1988; Gajewski et al. 2000; Sawada et al. 2004), or producing time series from one or several sites (e.g., Gajewski 1988; Sawada et al. 1999; Kerwin et al. 2004). Using key time intervals, data–model comparisons show strong regional responses to orbital forcing (e.g., COHMAP Members 1988).
Radiocarbon-dated pollen (and other) records, however, show variability on suborbital scales. Studies have shown that vegetation responds rapidly to climate change (Webb 1986; Gajewski 1987) and that there is evidence of a synchronous vegetation response to abrupt climate changes during the late glacial period and Holocene (Grimm et al. 1993; Williams et al. 2002; Viau et al. 2002, 2006; Gajewski et al. 2006) as well as during the more recent Little Ice Age (LIA) and Medieval Warm Period (MWP) of the past 1000 yr (Gajewski 1987; Viau et al. 2006). Thus, regional millennial to centennial climate changes are important, particularly because they are relevant to understanding the current global warming issues. However, regional syntheses of the time–space evolution of the climate are lacking.
Much of the recent interest in millennial-scale climate variability has centered on analysis of ice core and marine sediments; terrestrial evidence, such as pollen sequences from lake sediments, has been less often studied in this context (Gajewski et al. 2007). The large number of available pollen diagrams, each a multivariate time series, have resisted easy synthesis, in part because of the low temporal resolution of the data and ambiguities of the chronologies. Although climate change is invoked as an explanation for vegetation change in pollen diagrams, most palynological analysis has focused on the response of plants and vegetation to climate variability on orbital scales (e.g., Wright et al. 1993). However, significant changes in pollen sequences across North America and Europe occur synchronously, arguing for a response to changes in the climate regime at both a century and millennial scale (Gajewski 1987, 2000; Viau et al. 2002; Gajewski et al. 2006, 2007, and references therein). In addition, this relative lack of interest in lake sediments has been partly based on a perception that vegetation responds slowly to climate change, a conclusion that has, however, been questioned for many years (Gajewski 1988; Viau et al. 2002; Williams et al. 2002; Gajewski et al. 2007). By computing regional averages (Viau et al. 2006), we can attempt a continental-scale synthesis of the boreal zone of Canada and mitigate against some of these problems.
Here we explore the use of regional paleoclimatic averages to study Holocene changes within the modern boreal zone of Canada. Although this region has fewer available sites than in more temperate regions, we focus on higher latitudes because they are most affected by the current global warming, mainly because of strong snow/ice albedo feedbacks. Additionally, Viau et al. (2006) have shown that the development of regional composite reconstructions permits higher-temporal-resolution paleotemperature estimates through time–space averaging, which partly counteracts the low temporal resolution and sparse density of sites in some regions.
In this paper, we quantify Holocene paleoclimates across northern Canada between 50° and 70°N, encompassing the boreal and low Arctic regions. Through this determination of regional-scale patterns of climate variability we can provide a context for global warming. To accomplish this, we use a network of fossil pollen records that has the advantage of reducing the temporal and spatial uncertainties associated with site-specific uncertainties. The data were extracted from the North American Pollen Database (NAPD) and represent the most extensive network of Holocene terrestrial proxy-climate data available for North America (Grimm 2000). Temperature and precipitation reconstructions for each region are discussed on both orbital [low-frequency (104) variations of the past 12 ka] and millennial [higher-frequency (103) variations of the past 12 ka] time scales.
2. Data and methods
a. Data
We extracted 117 fossil pollen records from the NAPD (see the appendix; Grimm 2000) including all sites available from northern Canada between 50° and 70°N and between 50° and 140°W. The pollen sum used in this study consists of 89 taxa, comprising all abundant taxa and those important in boreal and tundra regions.
We then divided the region into four subregions (Fig. 1) to examine the spatiotemporal patterns of change through time. The regions were chosen as much as possible to reflect the physiographic and climate zones of Canada, but they were constrained by the need to ensure sufficient sites to permit the computation of regional averages. Modern July temperatures across the regions range from 8°C in central Canada to 10.1°C in Labrador, modern January temperatures range from −18.9°C in Labrador to −27.2°C in central Canada, and annual precipitation ranges from 529 mm in central Canada to 935 mm in Labrador. Because of the lower available site density, the central Canadian region spans 40° of longitude, compared to the roughly 20° span of the other three regions. There are only four sites from southern British Columbia available in the database, so the Mackenzie region primarily contains sites from the Yukon and upper Mackenzie valley; that is, the boreal portion of this region. This desire for objective site selection with no exclusion means that our analysis includes one site on Baffin Island and another from West Greenland from the Arctic zone.
At each site, January and July temperature and annual precipitation were estimated from the fossil pollen records using the modern analog technique (MAT; Overpeck et al. 1985). The modern pollen database used for calibration consists of 4590 sites and associated climate data, as described in Sawada et al. (2004), which includes data from most regions of North America, including the Arctic and Greenland regions. Approximately 75% of the “climate space” of North America is represented by the modern pollen site distribution (Whitmore et al. 2005; Sawada et al. 2004). The 14C age–depth relation of each pollen time series was remodeled using linear interpolation between dates based on the author’s approved dates. Calendar-year age–depth models were then derived using internationally ratified calibration curves (INTCAL98) (Stuiver et al. 1998). In this paper, we refer to dates as ka (1000 yr before a.d. 1950).
b. Paleoclimate reconstructions
We used the modern analog technique to estimate January and July temperatures and annual precipitation, following the methodology of Viau et al. (2006). Essentially, this technique successively compares each fossil pollen sample of every core to all modern pollen samples in the database. The modern pollen samples that have the smallest dissimilarity, measured by the squared chord distance (SCD), represent the best analogs (Overpeck et al. 1985). While debate remains over how many analogs should be retained or averaged (Viau and Gajewski 2007; Shuman et al. 2007; Williams and Shuman 2008; Viau et al. 2008, hereafter VIA), we simply retained the one least dissimilar modern pollen sample and used the July and January temperature and annual precipitation values at the location of the best analog as the best estimate for the reconstructions (e.g., see Viau et al. 2006). While averaging several analogs reduces the impact that an outlier point may have on the climate assigned to the fossil data point, the spatial averaging in this study also accomplishes this goal. However, a threshold is needed that specifies what is an acceptable analog, and there is at present no concrete agreement upon methods to do this, especially because this is variable in time and space (Sawada et al. 2004; Williams and Shuman 2008; VIA), as well as a function of taxonomic diversity (Waelbrock et al. 1998). In this study, we use the single best analog that is the modern sample that most resembles the fossil data point. Averaging several analogs produces temperature and precipitation estimates with more or less variability, depending on the site density of the modern calibration samples. Moreover, the averages may be based on samples spanning a wide region. For example, in the boreal forest, “good analogs” are found across a wide geographic region (Anderson et al. 1989) and averaging the climate data from these widely separated areas would result in an assigned climate different from both regions. Using both an expanded pollen set and the best analog should better separate different regional climates (Sawada et al. 2004), at least for boreal Canada. The application of the MAT in this context is discussed in Sawada et al. (2004) and Viau et al. (2006).
Once either the temperature or precipitation was estimated for all sample points of each time series, we estimated the value of each time series at 100-yr intervals using a simple linear interpolation. Because the original pollen diagrams have an average 250-yr resolution, this interpolation has minimal effect on the position of the peaks and troughs found in the original time series. We then averaged across all sites in each region to create four composite time series, one for each region, with a resolution of 100 yr.
c. Dating control and temporal resolution
The temporal resolution of these composite reconstructions is based on hundreds of dates and thousands of samples. The 117 individual pollen diagrams used in this study contain an average of 4.5 radiocarbon dates for a total of 490 radiocarbon dates for chronological control. This translates to a radiocarbon date every ∼100 yr for each regional curve, although the dates are not uniformly distributed (Viau et al. 2002). Although individual pollen records may not be sampled at a 100-yr resolution, this does not necessarily exclude their use. The pollen-sampling interval of any individual core varies greatly from site to site, with a mean value of approximately 10 cm, which translates to a median value of approximately 250 yr between samples. However, in the composite curve, any 250-yr period contains many samples and dates upon which to compute an average value (see Figs. 2 –5).
The resulting temporal resolution of our paleoreconstructions remains difficult to determine but appears to be a function of the sampling interval, age–depth model, and especially the spatiotemporal correlation. Nevertheless, the presence of spatial and temporal coherency ensures that the climate signal is preserved in the composite records. Interpolating to higher intervals than the sample depths will not add any new information to the resulting time series because peaks and troughs remain in the same place; therefore, any derived signal must be in the original data. The converse, however, is less desirable, because interpolating at lower intervals may mask out millennial-scale frequencies because of aliasing. The use of a network of sites has the advantage of reducing the temporal and spatial uncertainties associated with site-specific uncertainties.
d. Mapping space–time patterns of change
The temperature and precipitation reconstructions for each region are discussed on both orbital [low-frequency (104) variations of the past 12 ka] and millennial [higher-frequency (103) variations of the past 12 ka] time scales. We separately discuss the reconstructions of the past 2 ka, resulting from the interest in that time period as a context for understanding global warming using paleoclimate records under similar boundary conditions to those of today. To compare the range of variability among the regions on 104 time scale we simply compared the maximum and minimum values of the estimated temperatures; the variability on 103 time scales is presented as the standard deviation of the rate of change between adjacent points (i.e., the difference between each time interval).
Site density decreases in older time periods and this is associated with larger temporal variability in the reconstructions. A decrease in site density at 0 ka does not represent a problem because the fossil pollen sample represents modern-day conditions. In general, when the regional reconstructions are based on very few sites they become less reliable; the decrease in site density is quite abrupt because of the pattern of deglaciation of the area. Although we retained the complete reconstructions in the figures, as is the practice in dendroclimatic studies, for example, we do not consider the older periods when site density was too low.
To summarize the regional climate reconstructions, we use a Hovmöller diagram (contour plots with space and time as axes) approach, which is commonly used in meteorology, to illustrate the broad-scale evolution of climate patterns through time and across northern Canada. This approach has the advantage of synthesizing the results without relying on an exhaustive mapping exercise. The space–time diagrams of temperature and precipitation anomalies are used to synthesize the regional composite records of Figs. 2 –5 on two time scales of variability: the past 12 000 yr, as a way to visualize low-frequency changes (orbital scale), and a more detailed depiction of the last 2000 yr, as a way to visualize higher-frequency changes (millennial scale) across boreal Canada. Anomalies are with reference to the 0-ka reconstruction.
3. Results
a. Orbital scale
1) Labrador
July and January temperatures increased and precipitation generally increased between 12 ka and the present in the Labrador region (Fig. 2). July temperature anomalies were nearly +1°C circa 4 ka, and have slightly decreased subsequently. January temperature anomalies peaked at +1°C around 1.5 ka. Annual precipitation anomalies closely parallel January temperatures, with maximum anomalies of +60 mm occurring at 1.5 ka. Although we have less confidence for the period between 12 and ∼7 ka, the range over the entire period for July temperature anomalies is from −5° to +1°C, for January temperature anomalies from −14° to +1°C, and for annual precipitation anomalies from ∼−650 to +60 mm (Table 1). Site density decreased between 6 and 8 ka; the extreme negative values in the reconstructions, especially before 10 ka, may be disregarded.
2) Quebec
Site density in the Quebec region is low before 7 ka, and large fluctuations in reconstructed temperatures before that time are not reliable. Results show a general progression toward a warmer and moister climate from 6 to 2 ka (Fig. 3). July temperature anomalies peak at +0.35°C at 3.2 ka, and January temperature anomalies increased abruptly around 3.7 ka, with maximum values of +0.66°C around 1.8 ka. This was followed by an abrupt decrease in temperature, reaching minimum value between 0.4 and 0.5 ka (−0.96°C). Annual precipitation anomalies show maximum values of +50 mm occurring between 1.6 and 1.8 ka. The range of the temperature reconstruction is lower than that of Labrador (Table 1).
3) Central
There was a general progression toward a cooler and drier climate since 8 ka in central Canada. From 8 to ∼5.5 ka, summer temperature anomalies were relatively stable (mean of +0.62°C) after which temperature decreased toward modern values (Fig. 4). January temperature anomalies show a similar pattern; from 9 to ∼∼3 ka, January temperature anomalies fluctuated around a mean of +1.4°C and then decreased to modern values. Annual precipitation anomalies were maximum (+54 mm) at 5.5 ka. In summary, in this region the early Holocene was warm and dry and the middle Holocene relatively warm and humid. January and July temperatures decreased during the late Holocene.
4) Mackenzie
The available pollen records from this region tend to be longer, so site density remains high for nearly 12 ka. July temperatures increased between 11 and 9 ka, and remained high until after 5 ka (Fig. 5). January temperature anomalies gradually decreased since 11 ka and precipitation has decreased from maximum values at approximately 8.5 ka.
5) Space–time patterns
To summarize the above results, space–time diagrams are presented illustrating the anomalies of temperature and precipitation of the past 12 ka (Fig. 6). Warmest temperature anomalies are found in central Canada (Fig. 6) and, as expected, the western regions had maximum warmth occurring earlier than in the eastern region (Fig. 6). This is not surprising because deglaciation occurred much earlier in the west while the remnants of the Laurentide ice sheet (LIS) still influenced local to regional climates in the east until at least 8000 yr ago. During the mid-Holocene, western Canada was wetter than present, whereas eastern Canada was drier.
b. Millennial scale
1) Labrador
Millennial-scale variability during the past 12 000 yr in the Labrador region, quantified as the standard deviation of the first difference, is estimated at ±0.28°C for July temperature anomalies, ±0.69°C for January, and ±32 mm for annual precipitation anomalies (Table 1). Millennial-scale variability appears greater during the period from 12 to ∼7 ka (Fig. 2); however, the lower site density is a contributing factor.
2) Quebec
Millennial-scale variability during the past 8 ka in the Quebec region is estimated at ±0.28°C for July temperature anomalies, ±0.50°C for January, and ±18 mm for annual precipitation anomalies (Table 1). The greatest and most abrupt millennial-scale variability appeared during winter, with large transitions at 2 and 3.7 ka.
3) Central
Millennial-scale variability in central Canada region during the past 12 ka is estimated at ±0.32°C for July temperature anomalies, ±0.44°C for January, and ±14 mm for annual precipitation anomalies (Table 1). Variability of this scale was greater during the winter.
4) Mackenzie
The reconstructions for the Mackenzie region are marked by abrupt and large changes. For example, January temperatures show several abrupt coolings, and annual precipitation shows several peaks of increased precipitation. Variability is comparable to that measured in the Quebec and central Canada regions, but is less than that in Labrador (Table 1).
5) Space–time patterns
On the millennial time scale, both July and January temperature anomaly patterns show smaller oscillations superimposed on the orbital-scale trends during the Holocene (Fig. 6).
c. The last 2000 yr
1) Labrador
During the last 2000 yr, temperature anomalies in the Labrador region varied between +0.26°C at 1.1 ka, corresponding to the MWP and −0.4°C at 0.5 and 1.7 ka, corresponding to the LIA and Dark Ages cool periods (Fig. 2). January temperature and annual precipitation anomalies are in phase and show oscillations with a period of around 500 yr. January temperatures were cool and annual precipitation was lower 1000 yr ago than today, when July temperatures were relatively warm. The average range of the anomalies for the past 2000 yr is ±0.19°C for July temperatures, ±0.62°C for January temperatures, and ±28 mm for annual precipitation (Table 1).
2) Quebec
For the Quebec region, a decreasing trend in July temperatures between 2 and 0.8 ka was interrupted by a brief warming centered on 1 ka, corresponding to the Medieval Warm Period (Fig. 3). July temperatures have been increasing since 0.5 ka. January temperatures were low between 0.5 and 1.5 ka, again with relatively warm anomalies around 1.1 ka. Annual precipitation was higher prior to 1 ka. Higher-frequency annual precipitation anomalies are out of phase with those from Labrador (Figs. 2, 3). The average range of the anomalies is less than that in Labrador (Table 1).
3) Central
July temperatures in central Canada remained similar to today during the past 2 ka, except for a brief period of warmer conditions centered at 1 ka (Fig. 4). January temperatures were relatively high prior to 1 ka, and steadily decreased during the past 1000 yr. Annual precipitation anomalies vary about zero, with relatively dry periods centered on 0.5 and 1.4 ka.
4) Mackenzie
July temperature anomalies in the Mackenzie region decreased between 1.2 ka until around 0.4 ka, and then increased (Fig. 5). January temperature anomalies were generally below the present-day, except for a large and abrupt peak around 1.1 ka. Annual precipitation was generally similar to today, again except for higher values centered on 1.1 ka. Interestingly, the January temperature anomaly curve for the Mackenzie region resembles the July curve from the central region and vice versa (Figs. 4, 5).
5) Space–time patterns
Warming during the Medieval Warm Period, around 1 ka, and cooling during the Little Ice Age, centered on 0.5 ka, are seen across northern Canada. The magnitude of warming of the MWP was, as would be expected, greater in central Canada because of its continentality, whereas the LIA appears to be more strongly expressed in eastern and western Canada (Fig. 7). However, the magnitude of change between 1 and 0.5 ka is comparable across all regions (approximately −0.7°C). Annual precipitation is out of phase between Labrador and the Mackenzie region.
4. Discussion and conclusions
Despite questions that Quaternary scientists have concerning site density/distribution and chronology of terrestrial (i.e., pollen) records, it is still possible to get usable quantitative estimates of Holocene millennial- to centennial-scale climatic changes using these data. We have identified coherent patterns among the reconstructions across the study area, suggesting large-scale forcing of vegetation change at orbital to centennial time scales. The results of the analyses for the past 2000 yr (Fig. 7) suggests that even relatively brief climate variations cause responses in regional pollen production and vegetation patterns (Gajewski 1988, 2000; Williams et al. 2002; Gajewski et al. 2007), and further that that these are coherent across large areas. Because the temperature reconstructions are averages based on a number of sites from relatively large regions, the particular sequence of pollen changes at any one site differs across the region. For example, cooling could manifest itself as a decrease in spruce pollen at the tree line (at the expense of pollen from tundra plants) but an increase in spruce at the southern portion of the boreal forest (at the expense of pollen from deciduous trees). Both would reconstruct a cooling using the statistical method, and our results can be reasonably related to transitions in the vegetation as interpreted by the authors of the original pollen studies.
In Labrador, many sites record a relatively long period of tundra pollen following deglaciation (Short and Nichols 1977; Lamb 1980), and this leads to cold temperatures being reconstructed for the early Holocene (Figs. 2 and 6). A transition from herb tundra to shrub tundra in northern Labrador (Short and Nichols 1977), and a decrease in fir and white spruce at the expense of black spruce in southern Labrador (Engstrom and Hansen 1985) centered around 6.5 ka, resulted in an abrupt warming in winter temperature to be reconstructed. Maximum summer temperatures between 4 and 2.5 ka correspond to maximum spruce percentages and pollen accumulation rates in southeastern (Lamb 1980) and central (Short and Nichols 1977) Labrador. Decreasing pollen influx and sediment organic matter and increasing values of herbaceous pollen were interpreted as being due to late-Holocene cooling (Lamb 1980, 1985; Short and Nichols 1977), which is evident in the reconstruction of summer, but not winter temperatures.
In northwestern Quebec, spruce percentages increased between 6 and 2 ka, depending on region (Richard 1979; Richard et al. 1982; Gajewski et al. 1993; Leitner and Gajewski 2004), and this leads to a reconstruction of warming temperatures and increasing precipitation. If the period prior to 7 ka in Quebec is ignored because of low site density, the summer temperature curves from Labrador and Quebec closely resemble each other. However, late-Holocene winter temperatures and annual precipitation decreased in Quebec, but continued to increase in Labrador. In both Labrador and northern Quebec, the annual precipitation reconstruction closely parallels that of winter temperatures.
The location of the tree line in Quebec has not varied greatly in the past 5000 yr; climate changes have impacted the vegetation through alterations of the density and pollen production of the trees and shrubs that comprise the various zones of the broad forest–tundra that comprises the transition in this region (Gajewski et al. 1993). This impact on the vegetation is particularly noticeable in the winter, and to a lesser extent in the summer temperature reconstruction (Fig. 3), where the period of maximum temperatures began when alder pollen (characteristic of the forest–tundra) decreased and ended when spruce pollen (characteristic of lichen woodland) began to decrease (Gajewski et al. 1993, 1996; Gajewski and Garralla 1992; Leitner and Gajewski 2004). Interpretation of the tree line in Labrador is more difficult because of the presence of both altitudinal and latitudinal gradients, but the cooling temperatures reconstructed for the past 3 ka corresponds to an increase in tundra vegetation (Lamb 1985)
Site density is lower in central Canada, so more work is needed here to better quantify the Holocene climates of the region. Further, many records are short, and we have less confidence in the reconstructions prior to 9 ka. The maximum in temperature and precipitation during the mid-Holocene is more pronounced here than in eastern Canada, as expected in the central part of a continent. In this region, a movement northward of tree line has been documented between 6 and 4 ka (Moser and MacDonald 1990; MacDonald and Gajewski 1992; Gajewski and MacDonald 2004). South of the tree line, increases in the relative proportions of alder and spruce pollen at the expense of shrub birch are not as pronounced, but are still observable (e.g., MacDonald 1987a).
In western Canada (the Mackenzie region), maximum temperatures are reconstructed between approximately 10 and 6 ka. During this time, the tree line was located farther north than today in the Mackenzie delta region, and increases in several pollen taxa indicated warmer temperatures (Ritchie et al. 1983; Spear 1993). These changes have been associated with orbital warming and atmospheric circulation changes caused by the Laurentide ice sheet to the east (Ritchie et al. 1983).
Millennial-scale variations in temperature are observable across Canada during the Holocene (Fig. 6); however, the lower number of data before 9 ka makes it more difficult to determine the synchronicity during the early Holocene. Winter temperatures are warm across all of Canada between 6 and 4 ka. In general, millennial-scale changes occur slightly earlier in western Canada than in the east.
In the last 2000 yr, the pollen records document variations in temperature and precipitation of both the Little Ice Age and Medieval Warm Period. This warming and subsequent cooling are identified across Canada in both winter and summer, although perhaps beginning slightly earlier in the Mackenzie region (Fig. 7). The greatest amount of warming during the MWP was in central Canada, whereas the greatest cooling during the LIA occurred in northern Quebec. Precipitation changes are out of phase between eastern and western Canada, as had been observed in the northern United States (Gajewski 1988). This out-of-phase relation is seen in the overall postglacial curves as well as during the most recent 2000 yr (Fig. 6). In general, the pollen changes on century to millennial scales represent changes in the relative abundance of taxa rather than replacement of one pollen type by another. Thus, these are responses of the pollen production in the already established vegetation patterns rather than migration of taxa from one region to another.
Contrary to the results presented by Smol et al. (2005), and also reported in ACIA (2005, p. 370), there is no evidence of “climate stability” in Quebec/Labrador during the past 2000 yr. The assertion of stability was apparently based on several short diatom records, but these pollen-based paleoclimate estimates confirm climate and ecosystem variability at several scales, as was indicated by the analysis of the pollen records (Gajewski 2000). Our interpolated pollen records do not, however, have the resolution to determine the recent warming impacts on the vegetation.
Although the Little Ice Age has long been known to be a global event (Matthews and Briffa 2005), there are still questions about the nature of the Medieval Warm Period (Crowley and Lowery 2000). Our results indicate that across northern Canada, it was relatively warm at this time. Work is needed to develop high-resolution multiproxy-based reconstructions to better specify the nature of this time period.
Our results show that at no time during the Holocene have millennial-scale temperature variations exceeded +0.7°C in boreal Canada (Table 1). These results therefore show that presently observed temperature increases in northern Canada far exceed natural variability found in this study (Solomon et al. 2007), providing paleoclimatic support for human cause of the present-day global warming.
Acknowledgments
This study is a contribution of the Polar Climate Stability Network (PCSN) funded by the Canadian Foundation for Climate and Atmospheric Sciences (CFCAS). We gratefully acknowledge the contributors to the North American Pollen Database (NAPD) and North American Modern Pollen Database (NAMPD). We thank an anonymous reviewer for comments and suggestions.
REFERENCES
ACIA, 2005: Arctic Climate Impact Assessment. Cambridge University Press, 1042 pp.
Anderson, P., P. Bartlein, L. B. Brubaker, K. Gajewski, and J. C. Ritchie, 1989: Modern analogues of late-Quaternary pollen spectra from the western interior of North America. J. Biogeogr., 16 , 573–596.
Beaudoin, A. B., and R. H. King, 1990: Late Quaternary vegetation history of Wilcox Pass, Jasper National Park, Alberta. Paleogeogr. Paleoclimatol. Paleoecol., 80 , 129–144.
Bender, M. M., R. A. Bryson, and D. A. Baerreis, 1965: University of Wisconsin radiocarbon dates I. Radiocarbon, 7 , 399–407.
Bender, M. M., R. A. Bryson, and D. A. Baerreis, 1966: University of Wisconsin radiocarbon dates II. Radiocarbon, 8 , 522–533.
COHMAP Members, 1988: Climatic changes of the last 18,000 years: Observations and model simulations. Science, 241 , 1043–1052.
Crowley, T. J., and T. S. Lowery, 2000: How warm was the Medieval warm period? Ambio, 29 , 51–54.
Cwynar, L. C., 1982: A late-Quaternary vegetation history from Hanging Lake, northern Yukon. Ecol. Monogr., 52 , 1–24.
Cwynar, L. C., 1990: A late Quaternary vegetation history from Lily Lake, Chilkat Peninsula, southeast Alaska. Can. J. Bot., 68 , 1106–1112.
Cwynar, L. C., 1995: Paleovegetation and paleoclimatic changes in the Yukon at 6 ka BP. Géogr. Phys. Quat., 49 , 29–35.
Cwynar, L. C., and R. W. Spear, 1991: Reversion of forest to tundra in the central Yukon. Ecology, 72 , 202–212.
Davis, A. M., 1980: Modern pollen spectra from the tundra-boreal forest transition in northern Newfoundland, Canada. Boreas, 9 , 89–100.
Dyer, A. K., 1986: A palynological investigation of the Late Quaternary vegetational history of the Baie Verte Peninsula, Northcentral Newfoundland. M.S. thesis, Department of Geography, Memorial University of Newfoundland, 221 pp.
Eisner, W. R., T. E. Törnqvist, E. A. Koster, O. Bennike, and J. F. N. van Leeuwen, 1995: Paleoecological studies of a Holocene lacustrine record from the Kangerlussuaq (Søndre Strømfjord) region of West Greenland. Quat. Res., 43 , 55–66.
Engstrom, D. R., and B. C. S. Hansen, 1985: Postglacial vegetational change and soil development in southeastern Labrador as inferred from pollen and chemical stratigraphy. Can. J. Bot., 63 , 543–561.
Gajewski, K., 1987: Climatic impacts on the vegetation of eastern North America for the past 2,000 years. Vegetatio, 68 , 179–190.
Gajewski, K., 1988: Late-Holocene climates of eastern North America estimated from pollen data. Quat. Res., 29 , 255–262.
Gajewski, K., 2000: Climate of the last millennium in eastern North America and impacts on the vegetation. Pr. Geogr., 107 , 101–104.
Gajewski, K., and S. Garralla, 1992: Holocene vegetation histories from three sites in the tundra of northwestern Quebec, Canada. Arct. Alp. Res., 24 , 329–336.
Gajewski, K., and G. M. MacDonald, 2004: Palynology of North American Arctic Lakes. Long-Term Environmental Change in Arctic and Antarctic Lakes, R. Pienitz, M. S. V. Douglas, and J. P. Smol, Eds., Springer, 89–116.
Gajewski, K., S. Payette, and J. C. Ritchie, 1993: Holocene vegetation history at the boreal-forest–shrub-tundra transition in north-western Québec. J. Ecol., 81 , 433–443.
Gajewski, K., S. Garralla, and V. Milot-Roy, 1996: Postglacial vegetation at the northern limit of lichen woodland in northwestern Québec. Géogr. Phys. Quat., 50 , 341–350.
Gajewski, K., and Coauthors, 2000: The climate of North America and adjacent ocean waters ca. 6 ka. Can. J. Earth Sci., 37 , 661–681.
Gajewski, K., A. E. Viau, M. C. Sawada, D. E. Atkinson, and P. Fines, 2006: Synchronicity in climate and vegetation transitions between Europe and North America during the Holocene. Climatic Change, 78 , 341–361. doi:10.1007/s10584-006-9048-z.
Gajewski, K., A. E. Viau, and M. C. Sawada, 2007: Millennial-scale climate variability in the Holocene—The terrestrial record. Climate Variability and Change: Past, Present and Future: John E. Kutzbach Symp., Madison, WI, University of Wiscosnin—Madison, 133–154.
Grimm, E. C., cited. 2000: North American pollen database. WDC for Paleoclimatology, NCDC. [Available online at http://www.ncdc.noaa.gov/paleo/napd.html.].
Grimm, E. C., G. L. Jacobson, W. A. Watts, B. C. S. Hansen, and K. Maasch, 1993: A 50 000 year record of climate oscillations from Florida and its temporal correlation with the Heinrich events. Science, 261 , 198–200.
Hazell, S., 1979: Late Quaternary vegetation and climate of the Dunbar Valley, British Columbia. M.S. thesis, Department of Botany, University of Toronto, 101 pp.
Jordan, R. H., 1975: Pollen diagrams from Hamilton Inlet, central Labrador, and their environmental implications for the northern maritime Archaic. Arct. Anthropol., 12 , 92–116.
Kay, P. A., 1979: Multivariate statistical estimates of Holocene vegetation and climate change, forest-tundra transition zone, NWT, Canada. Quat. Res., 11 , 125–140.
Kearney, M. S., 1981: Late Quaternary vegetational and environmental history of Jasper National Park, Alberta. Ph.D. dissertation. University of Western Ontario, 318 pp.
Kerwin, M. W., J. T. Overpeck, R. S. Webb, and K. H. Anderson, 2004: Pollen-based summer temperature reconstructions for the eastern Canadian boreal forest, subarctic, and Arctic. Quat. Sci. Rev., 23 , 1901–1924.
Lamb, H. F., 1978: Post-glacial vegetation change in southeastern Labrador. M.S. thesis, Department of Ecology, University of Minnesota, Minneapolis, 101 pp.
Lamb, H. F., 1980: Late Quaternary vegetational history of southeastern Labrador. Arct. Alp. Res., 12 , 117–135.
Lamb, H. F., 1982: Late Quaternary vegetational history of the forest-tundra ecotone in North-Central Labrador. Ph.D. dissertation, University of Cambridge.
Lamb, H. F., 1984: Modern pollen spectra from Labrador and their use in reconstructing Holocene vegetational history. J. Ecol., 72 , 37–59.
Lamb, H. F., 1985: Palynological evidence for postglacial change in the position of tree limit in Labrador. Ecol. Monogr., 55 , 241–258.
Leitner, R., and K. Gajewski, 2004: Modern and Holocene stomate records of tree-line variations in northwestern Quebec. Can. J. Bot., 82 , 726–734.
MacDonald, G. M., 1983: Holocene vegetation history of the upper Natla River area, Northwest Territories, Canada. Arct. Alp. Res., 15 , 169–180.
MacDonald, G. M., 1987a: Postglacial development of the subalpine-boreal transition forest of western Canada. J. Ecol., 75 , 303–320.
MacDonald, G. M., 1987b: Postglacial vegetation history of the Mackenzie River Basin. Quat. Res., 28 , 245–262.
MacDonald, G. M., and K. Gajewski, 1992: The northern treeline of Canada. Geographical Snapshots of North America, D. G. Janelle, Ed., Guilford Press, 34–37.
MacDonald, G. M., R. P. Beukens, and W. E. Kieser, 1991: Radiocarbon dating of limnic sediments: A comparative analysis and discussion. Ecology, 72 , 1150–1155.
Mathewes, R. W., and G. E. Rouse, 1975: Palynology and paleoecology of postglacial sediments from the lower Fraser River canyon of British Columbia. Can. J. Earth Sci., 12 , 745–756.
Matthews, J. A., and K. R. Briffa, 2005: The “Little Ice Age”: Re-evaluation of an evolving concept. Geogr. Ann., 87A , 17–36.
McAndrews, J. H., and G. Samson, 1977: Analyse pollinique et implications archéologiques et géomorphologiques, lac de la Hutte Sauvage (Mushuau Nipi), Nouveau-Québec. Géogr. Phys. Quat., 31 , 177–183.
McAndrews, J. H., J. L. Riley, and A. M. Davis, 1982: Vegetation history of the Hudson Bay Lowland: A postglacial pollen diagram from the Sutton Ridge. Nat. Can., 109 , 597–608.
Morrison, A., 1970: Pollen diagrams from interior Labrador. Can. J. Bot., 48 , 1957–1975.
Moser, K. A., and G. M. MacDonald, 1990: Holocene vegetation change at treeline north of Yellowknife, Northwest Territories, Canada. Quat. Res., 34 , 227–239.
Mott, R. J., 1973: Palynological studies in central Saskatchewan: Pollen stratigraphy from lake sediment sequences. Geological Survey of Canada Paper 72-49, 18 pp.
Mott, R. J., 1976: A Holocene pollen profile from the Sept-Iles area, Québec. Nat. Can., 103 , 457–467.
Nichols, H., 1975: Palynological and paleoclimatic study of the late Quaternary displacements of the boreal forest-tundra ecotone in Keewatin and Mackenzie, N.W.T., Canada. Institute of Arctic and Alpine Research, University of Colorado, Occasional Paper 15, 87 pp.
Osborn, G., C. Clapperton, P. T. Davis, M. A. Reasoner, D. T. Rodbell, G. O. Seltzer, and G. Zielinski, 1995: Potential glacial evidence for the Younger Dryas event in the cordillera of North and South America. Quat. Sci. Rev., 14 , 823–832.
Overpeck, J. T., T. Webb III, and I. C. Prentice, 1985: Quantitative interpretation of fossil pollen spectra: Dissimilarity coefficients and the method of modern analogs. Quat. Res., 23 , 87–108.
Pellatt, M. G., and R. W. Mathewes, 1994: Paleoecology of postglacial tree line fluctuations on the Queen Charlotte Islands, Canada. Ecoscience, 1 , 71–81.
Pellatt, M. G., and R. W. Mathewes, 1997: Holocene tree line and climate change on the Queen Charlotte Islands, Canada. Quat. Res., 48 , 88–99.
Reasoner, M. A., and M. Hickman, 1989: Late Quaternary environmental change in the Lake O’Hara region, Yoho National Park, British Columbia. Paleogeogr. Paleoclimatol. Paleoecol., 72 , 291–316.
Richard, P. J. H., 1977: Histoire post-wisconsinienne de la végétation du Québec méridional par l’analyse pollinique. Service de la Recherche, Direction Générale des Forêts, Ministère des Terres et Forêts du Québec, Publications et Rapports Divers, Vol. 1, xxiv+312 pp.; Vol. 2, 142 pp.
Richard, P. J. H., 1979: Contribution à l’histoire postglaciaire de la végétation au nord-est de la Jamésie, Nouveau-Québec. Géogr. Phys. Quat., 33 , 93–112.
Richard, P. J. H., 1981: Paléophytogéographie postglaciaire en Ungava par l’analyse pollinique. Collection Paléo-Québec, No. 13, 154 pp.
Richard, P. J. H., A. C. Larouche, and M. A. Bouchard, 1982: Age de la déglaciation finale et histoire postglaciaire de la végétation dans la partie centrale du Nouveau-Québec. Géogr. Phys. Quat., 36 , 63–90.
Ritchie, J. C., 1969: Absolute pollen frequencies and carbon-14 age of a section of Holocene lake sediment from the Riding Mountain area of Manitoba. Can. J. Bot., 47 , 1345–1349.
Ritchie, J. C., 1977: The modern and late Quaternary vegetation of the Campbell-Dolomite Uplands, near Inuvik, N.W.T. Canada. Ecol. Monogr., 47 , 401–423.
Ritchie, J. C., 1982: The modern and Late-Quaternary vegetation of the Doll Creek Area, North Yukon, Canada. New Phytol., 90 , 563–603.
Ritchie, J. C., 1984: A Holocene pollen record of boreal forest history from the Travaillant Lake area, Lower Mackenzie River Basin. Can. J. Bot., 62 , 1385–1392.
Ritchie, J. C., and F. K. Hare, 1971: Late-Quaternary vegetation and climate near the arctic tree line of northwestern North America. Quat. Res., 1 , 331–342.
Ritchie, J. C., and K. A. Hadden, 1975: Pollen stratigraphy of Holocene sediments from the Grand Rapids area, Manitoba, Canada. Rev. Paleobot. Palyn., 19 , 193–202.
Ritchie, J. C., L. C. Cwynar, and R. W. Spear, 1983: Evidence from north-west Canada for an early Holocene Milankovitch thermal maximum. Nature, 305 , 126–128.
Rowe, J. S., D. Spittlehouse, E. Johnson, and M. Jasieniuk, 1975: Fire studies in the Upper Mackenzie Valley and adjacent precambrian uplands. Department of Indian and Northern Affairs, Ottawa, INA Publication QS-8045-000-EE-A1, 128 pp.
Samson, G., 1983: Prehistoire du Mushuau Nipi, Nouveau-Quebec: Etude du mode d’adaptation a l’interieur des terres hemi-arctiques. Ph.D. dissertation, University of Toronto, 810 pp.
Sawada, M., K. Gajewski, A. deVernal, and P. Richard, 1999: Comparison of marine and terrestrial Holocene climates in eastern North America. Holocene, 9 , 267–278.
Sawada, M., A. E. Viau, G. Vettoretti, W. R. Peltier, and K. Gajewski, 2004: Comparison of North American pollen-based temperature and global lake-status with CCCma AGCM output at 6 ka. Quat. Sci. Rev., 23 , 225–244.
Short, S. K., and H. Nichols, 1977: Holocene pollen diagrams from subarctic Labrador-Ungava: Vegetational history and climatic change. Arct. Alp. Res., 9 , 265–290.
Shuman, B., P. Bartlein, and T. Webb III, 2007: Response to “Comments on: ‘The magnitude of millennial- and orbital-scale climatic change in eastern North America during the Late-Quaternary’ by Shuman et al.”. Quat. Sci. Rev., 26 , 268–273.
Smol, J. P., and Coauthors, 2005: Climate-driven regime shifts in the biological communities of arctic lakes. Proc. Natl. Acad. Sci. USA, 102 , 4397–4402.
Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.
Spear, R. W., 1993: The palynological record of Late-Quaternary arctic tree-line in northwest Canada. Rev. Paleobot. Palyn., 79 , 99–111.
Steventon, R. L., and J. E. Kutzbach, 1983: University of Wisconsin radiocarbon dates XX. Radiocarbon, 25 , 152–168.
Stravers, L. K. S., 1981: Palynology and deglaciation history of the central Labrador-Ungava Peninsula. M.S. thesis, Department of Geology, University of Colorado, 171 pp.
Stuiver, M., and Coauthors, 1998: INTCAL98 radiocarbon age calibration, 24000-0 cal BP. Radiocarbon, 40 , 1041–1083.
Szeicz, J. M., G. M. MacDonald, and A. Duk-Rodkin, 1995: Late Quaternary vegetation history of the central Mackenzie Mountains, Northwest Territories, Canada. Paleogeogr. Paleoclimatol. Paleoecol., 113 , 351–371.
Terasmae, J., 1969: A discussion of deglaciation and the boreal forest in the northern Great Lakes region. Proc. Entomol. Soc. Ont., 99 , 31–43.
Vance, R. E., 1986: Pollen stratigraphy of Eaglenest Lake, northeastern Alberta. Can. J. Earth Sci., 23 , 11–20.
Viau, A. E., and K. Gajewski, 2007: Comments on “The magnitudes of millennial- and orbital-scale climatic change in eastern North America during the Late Quaternary” by Shuman et al. 2005. Quat. Sci. Rev., 26 , 264–267. doi:10.1016/j.quascirev.2006.07.010.
Viau, A. E., K. Gajewski, P. Fines, D. E. Atkinson, and M. C. Sawada, 2002: Widespread evidence of 1500 yr climate variability in North America during the past 14 000 yr. Geology, 30 , 455–458.
Viau, A. E., K. Gajewski, M. C. Sawada, and P. Fines, 2006: Millennial-scale temperature variations in North America during the Holocene. J. Geophys. Res., 111 , D09102. doi:10.1029/2005JD006031.
Viau, A. E., K. Gajewski, M. C. Sawada, and J. Bunbury, 2008: Low-and-high-frequency climate variability in Eastern Beringia during the past 25,000 years. Can. J. Earth Sci., in press.
Waelbrock, C., L. Labeyrie, J-C. Duplessy, J. Guiot, M. Labracherie, H. Leclair, and J. Duprat, 1998: Improving past sea surface temperature estimates based on planktonic fossil faunas. Paleoceanography, 13 , 272–283.
Webb III, T., 1986: Is vegetation in equilibrium with climate? How to interpret late-Quaternary pollen data. Vegetatio, 67 , 75–91.
Whitmore, J., and Coauthors, 2005: A North American modern pollen database for multi-scale paleoecological and paleoclimatic applications. Quat. Sci. Rev., 24 , 1828–1848.
Williams, J. W., and B. Shuman, 2008: Obtaining accurate and precise environmental reconstructions from the modern analog technique and North American surface pollen dataset. Quat. Sci. Rev., 27 , 669–687.
Williams, J. W., D. M. Post, L. C. Cwynar, A. F. Lotter, and A. J. Levesque, 2002: Rapid and widespread vegetation responses to past climate change in the North Atlantic region. Geology, 30 , 971–974.
Wright, H. E., J. E. Kutzbach, T. Webb III, W. F. Ruddiman, F. A. Street-Perrott, and P. J. Bartlein, 1993: Global Climates since the Last Glacial Maximum. University of Minnesota Press, 569 pp.
APPENDIX
Site Used in This Study
Location of pollen diagrams used in this study (references in the appendix).
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
July and January temperature and annual precipitation anomalies for the Labrador region (50°–70°N, 50°–65°W) for the past 12 000 and 2000 yr. Site density, chronological control, and sampling interval are also presented. Shaded area represents less confident reconstructions resulting from a decreasing number of sites.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
July and January temperature and annual precipitation anomalies for the northern Quebec region (50°–70°N, 65°–80°W) for the past 12 000 and 2000 yr. Site density, chronological control, and sampling interval are also presented. Shaded area represents less confident reconstructions resulting from a decreasing number of sites.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
July and January temperature and annual precipitation anomalies for the central Canada region (50°–70°N, 80°–120°W) for the past 12 000 and 2000 yr. Site density, chronological control, and sampling interval are also presented. Shaded area represents less confident reconstructions resulting from a decreasing number of sites.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
July and January temperature and annual precipitation anomalies for the MacKenzie region (50°–70°N, 120°–140°W) for the past 12 000 and 2000 yr. Site density, chronological control, and sampling interval are also presented. Shaded area represents less confident reconstructions resulting from a decreasing number of sites.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
Space–time patterns of (a) July temperature (°C), (b) January temperature (°C), and (c) annual precipitation (mm) anomalies during the Holocene for the regions in Fig. 1. Area above (i.e., older than) bold line represent less reliable reconstructions resulting from a decreasing number of sites. Note that for the Quebec region, values from 9 to 12 ka represent interpolations.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
Space–time patterns of (a) July temperature (°C), (b) January temperature (°C), and (c) annual precipitation (mm) anomalies during the past 2000 yr for the regions in Fig. 1.
Citation: Journal of Climate 22, 2; 10.1175/2008JCLI2342.1
Holocene climate variability for regions shown in Fig. 1. For orbital scales, the range between the minimum and maximum temperature (T) and precipitation (P) anomaly during the past 12 ka is presented. For the millennial scale and past 2 ka, numbers are the standard deviation of the first differences of T and P anomalies.
