Past and Future Forest Response to Rapid Climate Change

Margaret B. Davis

University of Minnesota, St, Paui, Minnesota

1. Introduction 167

2. Long-Distance Dispersal 168

3. Estimating Jump Distances 170

4. Interactions with Resident Vegetation—Constraints on Establishment 171

5. Interactions with Resident Vegetation—Competition for Light and

Resulting Constraints on Population Growth 172

6. Discussion 173

7. Conclusions 173

References 174

In response to large changes of climate during the Ilolocene, geographical ranges of tree species shifted northward in eastern North America, with range extensions occurring at rates of 10-100 km per century. Long-distance dispersal of seeds, an important mechanism for rapid range extension, is documented by fossil evidence for colonies established well in front of the continuous population. Average jump distance as trees moved into the Great Lakes region was 80-100 km for eastern hemlock (Tsuga canadensis), with wind-dispersed seed, and 40 km for beech (Fa-gus grandifolia), which has animal-dispersed seed. Jump dispersal distances estimated from range maps, by measuring distances between outlying colonies and the continuous population, are again larger for hemlock than for beech — 40 km versus 8 km. Interactions with resident vegetation were constraints on migration rates. Invasions of individual forest stands by hemlock were restricted to stands dominated by white pine (Pimts strobus). Stands that were dominated by hardwoods at the time of invasion were not invaded by large numbers of hemlock and are now dominated by sugar maple (Acer saccharum) and basswood (Tilia americana). Fine-scale studies of fossil records from hemlock stands by T. E. Parshall show that several centuries elapsed after the first hemlock trees were established before hemlock became dominant in the stand, replacing resident white pine. Stand simulations suggest that delays of this length could be caused by competition from resident canopy trees.

If future climate changes caused by doubled greenhouse gas concentrations occur within a century, seed dispersal is inadequate to accomplish significant changes in ranges. On this time scale, interactions with resident vegetation become important. Resident vegetation will constrain colonization of microhabitats that become more favorable as the climate changes, and resident canopy trees will inhibit population expansion of minor species that are better adapted to the new climatic regime.

1. Introduction

A rise in global temperature of about 2°C is expected as greenhouse gases reach doubled concentrations in the atmosphere (Houghton et al. 1996). How will forests respond?

The Holocene fossil record provides the only observations we have of forest responses to climate warming of this magnitude. In many regions that were not actually covered by ice, plant species expanded from local refuges as the climate warmed in the early Holocene (Thompson, 1988; Tsukada, 1988; Markgraf et al, 1995; McGlone, 1997). But in eastern North America and western Europe, a major response to Holocene warming was a latitudinal shift of species ranges as tree populations "migrated" northward to populate the newly deglaciated landscape. Analyses of climate changes using global circulation models suggest that tree migrations tracked the movements of climate-spaces to which particular species were adapted (Prentice et al, 1991). Migrations involved movements of species boundaries at rates averaging 10-100 km per century (Davis, 1981; Huntley and Birks, 1982; Huntley and Webb, 1988; Delcourt and Delcourt, 1987; Pitelka et al, 1997). If greenhouse gases double in the coming century (Houghton et al, 1996), future changes may involve shifting of species boundaries an order of magnitude more rapidly. Is such rapid migration possible?

This chapter examines the Holocene fossil record of the Great Lakes region for information on the process of tree migration. Seed dispersal and interactions with resident vegetation are considered because both factors could have constrained the rate of advance of trees. Finally, I will discuss the results in the context of global warming, in order to identify the constraints that are likely to be important in the coming century.

2. Long-Distance Dispersal

Recent discussions of tree migration have emphasized the importance of seed dispersal (Pitelka et al., 1997; Clark et ai, 1998). Dispersal appears to be limiting, because in forests seed travels only a few meters from the source tree. Diffusion models that use observed dispersal parameters are unable to simulate the rates of migration observed in the Ilolocene record (Clark, 1998). Models that are able to simulate Holocene migration rates invoke seed dispersal hundreds or even thousands of meters farther than observed in forests (Clark, 1998; Shigesada et al, 1995; S. Sugita, Ehime University, personal communication). Yet despite the importance of long-distance dispersal, we know little about it. Greene and Johnson (1995) have developed models of seed dispersal, testing them against seed trap data from very large open areas. Appreciable numbers of seeds, 1- 10% of their densities in forests, were found in traps 1000-1600 m from the forest edge. Direct observation of seeds released from towers indicated that 1 -2% were caught in up-drafts, dispersing quite differently than predicted assuming a con stant wind speed (Greene and Johnson, 1995). Johnson and Ad-kisson (1985) observed bluejays (Cyanocitta cristata) transporting beechnuts (Fagus grandifolia) 4 km to cache them near their nesting areas. These rare observations are important because they demonstrate that a small proportion of seeds is available for dispersal across distances very much greater than observed in seed traps in forests. However, direct observations do not provide information on maximum dispersal distances, nor on the frequency of long-distance dispersal events.

Additional information about seed dispersal is contained in the fossil record. Dispersal events can be inferred wherever there is evidence that populations were founded at large distances from the source population. It is, of course, technically difficult to detect a small population using fossil pollen or macrofossils, and even more difficult to demonstrate that the small population was isolated from the main population (Davis et ai, 1991). Pollen studies in Sweden, however, record the establishment of individual colonies of beech (Fagus sylvatica) in the late Holocene (Bjorkman, 1996) and macrofossils demonstrate that populations of spruce (Picea abies) grew far in advance of the expanding species front for thousands of years (Kullman, 1996). East of James Bay, Canada, small colonies of larch (Larix laricina) have become established in patches during the past 1500 years as the population has expanded. Some of these colonies have fused into a continuous distribution,

2000-year intervals. Small numbers indicate the C-l'l age (in 10' year) of a steeplO-fold increase in pollen accumulation rates (grains cm2 year-') and/or pollen percentages at an individual fossil site (maps modified from Davis, 1981, Davis et al.; 1986). Within the Great Lakes region, small numbers indicate establishment dates for populations of beech using the criteria established by Webb (1987) and Woods and Davis (1989). Stippled areas indicate present range.

2000-year intervals. Small numbers indicate the C-l'l age (in 10' year) of a steeplO-fold increase in pollen accumulation rates (grains cm2 year-') and/or pollen percentages at an individual fossil site (maps modified from Davis, 1981, Davis et al.; 1986). Within the Great Lakes region, small numbers indicate establishment dates for populations of beech using the criteria established by Webb (1987) and Woods and Davis (1989). Stippled areas indicate present range.

but along the western species limit small populations remain isolated from one another (Penalba and Payette, 1997).

More detailed information is available for American beech (Fa-gus grandifolia) and eastern hemlock (Tsuga canadensis), because their migration into the Great Lakes region of North America was studied for the purpose of contrasting migration patterns of an animal-dispersed species (beech) with those of a wind-dispersed species (hemlock) (Davis, 1987). Outlying populations were relatively easy to document, because pollen evidence for colonies founded on the far side of any of the Great Lakes provides convincing evidence of disjunction between the new colonies and the parent population. In this manner long-distance dispersal was demonstrated for both species (Webb, 1987; Davis et al., 1986; Davis, 1987; Woods and Davis, 1991).

Beech expanded northward from Florida during the early Holocene, arriving in southernmost Michigan 7000-8000 years ago (Davis, 1981; Bennett, 1987; Webb, 1988)(Fig. la). Detailed studies of beech migration into the Great Lakes region reveal that while beech was spreading northward on the eastern side of Lake Michigan, colonies were established on the western shore. Colonies were established on the western shore of the lake as early as 6000 years ago. Beech had not yet colonized the northern shore of the lake, so the disjunct populations must have been established by long-distance dispersal. Seeds were transported directly across the lake, a distance of about 100 km, or by many jumps between hypothetical islands of favorable habitat within the prairie vegetation at the southern end of the lake (Webb, 1987)(Fig. la). Webb (1986) speculates that the extinct passenger pigeon could have been the vector for dispersal. Four thousand years ago, additional colonies were established farther north along the western shore of Lake Michigan. About 2500 years ago, the northern shore of the lake was colonized and all populations expanded, coalescing into a continuous population by 2000 years B. P. Within the past millennium, as beech reached its western limit, a large disjunct population was established west of Marquette, MI, 40 km beyond the species boundary (Woods and Davis, 1989).

Hemlock moved northward up the Appalachians and along the eastern seaboard before moving westward into the upper Great Lakes region (Fig. lb). Between 6000 and 5000 years ago, there was a sudden increase of pollen from trace quantities to 10-20% of tree pollen in sediment cores from many lakes in Michigan (Fig. 2). The sudden increase occurred throughout hemlocks

Hemlock Pollen Percentages b

Hemlock Pollen Percentages b

a Swain and Winkler, 1983 b Webb, 1974 c Brubaker, 1975 d Futyma, 1982

FIGURE 2 Pollen percentages of hemlock (as percent tree pollen) in sediment from 17 lakes and bogs in northern Michigan. Radiocarbon ages (indicated on ordinate) are based on bulk sediment; hard-water errors are large at the nine lakes at the eastern end of the transect, a region of calcareous bedrock. All the eastern lakes show trace quantities of pollen starting about 7000 years ago, and a sharp increase of hemlock pollen between 6000 and 5000 years ago. The western lakes document rapid westward migration between 5500 and 4500 years ago, and much slower westward migration 3000-1500 years ago (Davis et al., unpublished data, and references cited in figure).

range in lower Michigan, and in the eastern half of upper Michigan (Fig. lb). The sudden invasion of such a wide area suggests that many previously established colonies were expanding rapidly in response to climate changes that favored hemlock. Populations coalesced and the species frontier migrated westward as a continuous front (Fig. 2)(Davis et al., 1986; Davis, 1987; Davis et al., unpublished). The source of seed for the colonies established more than 6000 years ago must have been east of Lake Huron— southern and central Ontario, which had been invaded by hemlock 8000 years ago (Kapp, 1977; Bennett, 1987; Fuller, 1998). Although the precise trajectory of dispersal is unknown, the distances required are large-over 100 km. Dispersal across such great distances is believable, however, because after hemlock had spread across Michigan and into Wisconsin, reaching its present western limit 1500 years ago, colonies became established at several locations in Minnesota. The nearest colonies are 100 km from the species front, and the farthest locations are an additional 110 km beyond them. Pollen records from two of the outlying colonies in Minnesota establish their origin as 1200 years ago (Calcote, 1986).

3. Estimating Jump Distances

The distance to which seed can be dispersed to establish new populations ahead of the migrating species front has been measured in two ways. First, distances have been tabulated between disjunct

FIGURE 3 Maps showing distances between outlying colonies and the main population at the time the colonies were established, as indicated by the fossil pollen records of beech (a) and hemlock (b) (S. Webb, 1987; Woods and Davis, 1989; Davis et al., 1986; Calcote, 1986; Davis et al., unpublished data).

colonies demonstrated in the fossil record and the species frontier as it existed at that time. The most conservative estimate was chosen in all cases. For example, islands in Lakes Michigan or Huron were presumed to have been "stepping-stones," and dispersal distances were measured over stretches of open water. 1'he results are illustrated in Figure 3, and the data are summarized in Table 1. Jump-dispersal distances (Pielou, 1979) are at least twice as great for hemlock as for beech. There are several instances of leaps of 100 km or more for hemlock. The largest leap for beech is 100 km across Lake Michigan, but other jumps are smaller, in general 10-40 km.

The second method for measuring dispersal uses detailed range maps prepared for Wisconsin and a part of upper Michigan using witness tree data collected before settlement in the early 19th century (Davis et al, 1991). In the range maps shown in Figure 4, distances were measured between outlying colonies and the continuous species limit. Again, if there were colonies between the main species limit and more distant colonies, the assumption was made that intermediate colonies acted as stepping stones. Estimates of dispersal distances measured in this way are conservative, because some of the outlying colonies may have been established well before the continuous population reached its present location. This may explain why dispersal distances measured from the range maps are consistently smaller than dispersal distances documented in the fossil record (Table 1).

The important generalization that emerges from the data is that dispersal distances for hemlock, however they are measured, are at least twice as large as for beech, and possibly four or five times greater (Table 1). This is not unexpected since hemlock seeds are dispersed by wind. However, the result is significant because dispersal distance is an important parameter that affects model predictions of migration rates — both diffusion models with a "fat tail" (Clark, 1998) and modified scattered colony models that assume that dispersal can occur out to some maximum distance (S. Sugita, personal communication). If these models are used to predict future ranges of tree species resulting from global warming, dispersal parameters will have to be determined for each species—a formidable task. Even with so much data available, we can only approximate the difference in dispersal distance between beech and hemlock, and we have no precise information on the frequency of longdistance dispersal events.

TABLE 1 Inferred Jump Dispersal Distances

FIGURE 3 Maps showing distances between outlying colonies and the main population at the time the colonies were established, as indicated by the fossil pollen records of beech (a) and hemlock (b) (S. Webb, 1987; Woods and Davis, 1989; Davis et al., 1986; Calcote, 1986; Davis et al., unpublished data).

Jump distances implied by fossil records of establishment of outlying colonies Beech {Fagus grandifolia) 40 km (n = 7)

Average distance to outlying colonies on presettlement range map Beech {Fagus grandifolia) 8 km (n = 16)

Hemlock {Tsuga canadensis) 39 km (n = 46)

FIGURE 4 Maps showing the distribution of beech and hemlock in Wisconsin at the time of the Federal Land Office Survey in the early 19th century (redrawn from Davis et al„ 1991).

FIGURE 4 Maps showing the distribution of beech and hemlock in Wisconsin at the time of the Federal Land Office Survey in the early 19th century (redrawn from Davis et al„ 1991).

4. Interactions with Resident Vegetation—Constraints on Establishment

The migration of hemlock has been studied at a fine spatial scale, using sediment from small forest hollows that provide a pollen record of the history of individual forest stands a few hectares in size (Sugita, 1994; Calcote, 1995; 1998). These fine-scale studies record the invasion of individual forest stands. Did resident vegetation influence the pattern of invasion? This question has been considered at length in the literature on invasions by exotic species, with evidence cited by several authors that resident vegetation, or its absence on disturbed sites, can influence invasion success (Crawley, 1987; Drake, 1990; Lodge, 1993). For the present discussion we are interested in the effect this phenomenon could have on overall migration rate.

Fossil pollen in a series of small forest hollows about 10 m in diameter provides a record of hemlock invasion of individual forest stands along a 10-km transect in northern Michigan. The distribution of species within the present-day forest, which has never been clearcut, is patchy—a mosaic of stands dominated by hemlock interspersed with mixed stands and large patches dominated by sugar maple (Acer saccharum). Pollen diagrams from four hemlock stands and four maple stands extend back to the time hemlock invaded the forest about 3000 years ago. Prior to hemlock invasion, all the hemlock stands had been dominated by white pine (Pinus strobus). After hemlock invaded, it coexisted with pine for a thousand or more years, until hemlock displaced white pine in three of the four stands. White pine had also been abundant in one of the stands now dominated by hardwoods. This stand was also invaded by hemlock, but following a windstorm, hemlock was eliminated and maple became dominant. In contrast, other stands now dominated by maple were already dominated by hardwoods at the time hemlock was invading pine stands. They were never invaded by large numbers of hemlock. Establishment was probably prevented by the same factors that discourage the establishment of hemlock seedlings in maple stands today. Hardwood litter provides a poor seedbed for hemlock, and the light and nutrient regimes favor the growth of maple seedlings that shade hemlock seedlings (Ferrari, 1993; Davis et al, 1994; 1998).

In this example, a portion of the landscape was occupied by resident vegetation that inhibited establishment of a migrating species. At present maple stands make up 12% of the upland landscape. Lakes and wetlands compose another 34% of the area (Pastor and Broschart, 1990), leaving only about half of the landscape available for colonization by hemlock. Was the rate of hemlock

500 H 1000


10 grains/ml

10 grains/ml

103 grains/ml

103 grains/ml

10 grains/ml

FIGURE 5 Hemlock pollen percentages and concentrations plotted against the C-14 age of sediment in five small forest hollows located within hemlock stands in western Wisconsin. Arrows indicate the oldest sediment in which fossil hemlock stomata are found at each hollow. Stomata indicate the presence of one or more hemlock trees within 20 m of the hollow (Parshall, 1999). The increases in pollen concentrations (dashed lines) and percentages (solid lines) in the last 200-300 years indicate increasing hemlock population densities within the nearest 1-3 ha of forest. Most records indicate a long establishment phase between the initial colonization of the stand and the population increase. [Modified figure reprinted, with the author's permission, from Parshall (1998)].

migration slowed because establishment was restricted? Unfortunately, the answer is ambiguous. Migration was indeed slow in this part of Michigan 3000 years ago, but we cannot separate the effects of climate from the constraint on establishment (Davis, 1987; Davis et aU 1994; 1998).

5. Interactions with Resident

Vegetation—Competition for Light and Resulting Constraints on Population Growth

Many introduced species show an "establishment phase," years or decades following introduction when the invading plants or animals are not seen in their new environment. Then suddenly the invading organisms seem to be everywhere, in a rapidly expanding population. Some believe the establishment phase represents a period when the density of colonists is so low that they are undetectable, while others believe it represents a period when genetic adaptation to a new environment is taking place (Ewel, 1986; Baker, 1965).

Direct observation of the establishment phase has not been possible because the invading organisms are so rare. A retrospective record provides more information, obtainable from small hollows in forest stands that currently include the invading species. Parshall (1998) studied invasion by eastern hemlock of five hemlock stands in western Wisconsin, using fossil pollen, conifer needles, and stomates from conifer needles. Fossil pollen in the forest hollows reflects hemlock density within 50-80 m (Sugita, 1994; Calcote, 1995), while the fossil needles, or stomates from needles, demonstrate the presence of one or more hemlock trees within 20 m (Parshall 1998). A remarkable feature of Par-shall's data (Fig. 5) is the long lag at four of the five sites between the first appearance of hemlock (shown by fossil stomates: arrows in Fig. 5) and the population expansion indicated by increased amounts of hemlock pollen. At one stand, the lag—i.e., the establishment phase—lasts 1000 years, at others several hundred years (Parshall, 1998). Thus hemlock, although present in the forest, was unable to increase for several centuries in most of these stands (Parshall, 1998.)

Population growth may have been delayed by competition from resident canopy trees. Competition for light is the mechanism suggested by gap model simulations. In a simulated sugar maple forest subjected to a sudden climate cooling of 2°C, the change in canopy dominants from sugar maple to spruce (Picea rubens) took 200 years. In the simulations, sugar maple saplings were replaced by spruce, but canopy maple continued to shade the better-adapted spruce in the understory until the canopy trees reached the end of their normal lifespan (Davis and Botkin, 1985). Disturbance can speed up replacement (Davis and Botkin, 1985; Overpeck et al., 1990), but natural disturbance rates in hardwood forests of the Great Lakes region are quite low, resulting in canopy lifetimes of 150-200 years (Frelich and Lorimer,

1991; Frelich and Graumlich, 1994; Parshall, 1995; Parshall et al., in review).

6. Discussion

If future warming were to occur slowly, with temperature increases associated with C02 doubling spread out over 500 years, migrations could occur as they did during the I Iolocene. The rates of northward range extension would have to be the maximum recorded, however, 50-100 km per century. In this unlikely scenario, seed dispersal would be an important variable limiting the rate of advance. We have shown that seed dispersal can occur over long distances—a few tens of kilometers to over a hundred kilometers—but the frequency of recorded dispersal events is not high. Long-distance dispersal that resulted in successful colonies is recorded for each species about once per millennium during the Holocene, whereas a much higher frequency will be required for future change. Dispersal by humans, however, is likely to occur, making natural dispersal mechanisms less important. But for noncommercial trees, as well as the herbs, shrubs, mosses, and fungi that make up forest ecosystems, natural dispersal will remain an important limitation to adjustment to climate change. Many future forests will doubtless lack species that we now consider important components of the ecosystem.

If climate change occurs rapidly, with 2°C warming by 2100 AD. (Houghton et al., 1996), then seed dispersal will be too slow to accomplish significant vegetation adjustment. Range extensions could occur in this time frame only for trees that already had outlying colonies to the north. Hemlock, for example, had a fringe of outlying populations beyond its range limit (Little, 1971). In Wisconsin the colonies are 40 km from the range boundary on average (Fig. 4). If outlying populations like these have survived logging along the northern range limit, they can serve as centers of infection for the surrounding landscape. Expanding these preexisting populations could allow hemlock to extend its range by an average of 40 km within the coming century. Outlying colonies of beech are only 10 km from the range limit in Wisconsin, suggesting that this species could extend its range northward only by 10 km (Fig. 4, Table 1). These range extensions are small relative to the displacements of potential ranges by hundreds of kilometers that are likely with climate changes accompanying doubled C02 (Davis and Zabinski, 1992; Sykes et al, 1996).

Under a rapid climate change scenario, factors that constrain establishment and population growth become much more important than seed dispersal. On this time scale the likely responses are changes in species abundances and distributions within regions where the tree is already growing. Fossil records and simulations show that competition from canopy trees can delay the population expansion of tree species that are better adapted than resident dominants to new climate conditions (Davis and Botkin, 1985). Another likely response to a rapidly changing climate is redistribution of tree species on the land scape, involving, for example, dispersal from drier substrates to more mesic sites. In this case possible inhibition of establishment by local vegetation could be important in delaying adjustment to changing climate.

Recent literature on forest response to future climate change emphasizes dispersal limitations (Pitelka et al, 1997; Clark et al, 1998). The review I have presented here suggests that natural dispersal is unlikely to accomplish adaptation to future climate. Competition and stand dynamics are much more important constraints on the decadal scale we need to consider if greenhouse gases continue to accumulate at present rates.

7. Conclusions

1. Range shifts in response to climate changes over the past 11,000 years were slow compared to the rapid range adjustments that will be necessary in the coming century as greenhouse gases double in concentration.

2. Past range shifts were accomplished by seed dispersal 10-100 km beyond the species range limit. The frequency of long-distance dispersal events has not been adequately measured, but data show clearly that the average dispersal distance differs between species. Migration models will have to include species-specific dispersal parameters.

3. Establishment of new populations of migrating trees was limited by resident vegetation.

4. Population expansion by newly established colonies of migrating trees was delayed for decades or centuries by competition from resident canopy trees.

5. The fossil record of forest tree response to changing climate suggests that in the coming century seed dispersal will not be adequate to accomplish range shifts rapidly enough to track future climate. On the time scale of decades, resident vegetation that inhibits establishment and disturbance regimes that control forest stand dynamics will be important limitations to the rate of forest adaptation to changing climate.


I appreciate the opportunity to participate in the celebration of the founding of the Max-Planck Institute for Biogeochemie at Jena. I congratulate the Institute and wish it success as it seeks to understand the changing global ecosystem. The research reviewed here was supported by the National Science Foundation, Grants DEB8012159, DEB8407943, BSR8615196, BSR8916503, DEB9221371, and by the Mellon Foundation. I gratefully acknowledge the generosity of T.E. Parshall and S. Sugita, who allowed me to present their research results, and I thank Holly Ewing, David Lytle, Christine Douglas, Randy Calcote, and Shinya Sugita for helpful comments on an earlier version of the manuscript.


Baker, H. G. 1965. Characteristics and modes of origin of weeds. In "The Genetics of Colonizing Species." (H. G. Baker and G. L. Stebbins, Eds.), p. 147-173. Academic Press, New York.

Bennett, K. D. 1987. Holocene history of forest trees in southern Ontario. Can. J. Bol. 65, 1792-1801.

Bjorkman, I.. 1996. Long-term population dynamics of Fagus sylvatica at the northern limits of its distribution in southern Sweden: a paleoeco-logical study. Holocene 6, 225-234.

Brubaker, L. B. 1975. Postglacial forest patterns associated with till and outwash in north central upper Michigan. Q. Res. 5, 499-527.

Calcote, R. R. 1986. Hemlock in Minnesota: 1200 years as a rare species. M.S. Thesis. University of Minnesota, Minneapolis, Minnesota.

Calcote, R. R. 1995. Pollen source area and pollen productivity: evidence from forest hollows. /. of Ecol. 83, 391-602.

Calcote, R. 1998. Identifying forest stand types using pollen from forest hollows. Holocene 8, 423-432.

Clark, J. S. 1998. Why trees migrate so fast: confronting theory with dispersal biology and the paleorecord. Am. Nat. 152, 204-224.

Clark, L S., Fastie, C., Hurtt, G., Jackson, S. T., Johnson, C., King, G. A., Lewis, M„ Lynch, J., l'acala, S., Prentbice, I. C., Schupp, E. W„ Webb, T. Ill, and Wyckoff, P. 1998. Reid's Paradox of rapid plant migration. Bioscience 48 (1), 13-24.

Crawley, M. J. 1987. What makes a community invasible? In "Colonization, Succession and Stability." (A. J. Gray, M. J. Crawley, and P. J. Edwards, Eds.), pp. 1-29. Blackwell Scientific Publishers, Oxford, United Kingdom.

Davis, M. B. 1981. Quaternary history and the stability of forest communities. In "Forest Succession: Concepts and Application ." (D. C. West, H. H. Shugart, and D. B. Botkin, Eds.), pp. 132-153. Springer-Verlag, New York.

Davis, M. B. 1987. Invasions of forest commmunities during the Holocene: beech and hemlock in the Great Lakes region. In "Colonization, Succession and Stability." A. J. Gray, M. J. Crawley, and P. J. Edwards, Eds.), pp. 373-393. Blackwell Sci., Oxford, United Kingdom.

Davis, M. B. and Botkin, D. B. 1985. Sensitivity of cool-temperate forests and their fossil pollen record to rapid temperature change. Q. Res. 23, 327-340.

Davis, M. B„ Woods, K. D„ Webb, S. L„ and 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., Schwartz, M. N., and Woods, K. 1991. Detecting a species limit from pollen in sediments./. ofBiogeogr. 18, 653-668.

Davis, M. B. and Zabinski, C. 1992. Changes in geographical range resulting from greenhouse warming: effects on biodiversity in forests. In "Global Warming and Biological Diversity." (R. L. Peters and T. E. Love-joy, Eds.), pp. 297-308. Yale LJniversity Press, New Haven, Connectient.

Davis, M. B„ Sugita, S., Calcote, R. R., Ferrari, J. B„ and Frelich, L. E. 1994. Historical development of alternate communities in a hemlock-hardwood forest in northern Michigan, USA. "Large Scale Ecology and Conservation Biology." In (P. I. Edwards, R. May, and N. R. Webb, Eds.), pp. 19-39 Blackwell Scientific Publications, Oxford, United Kingdom.

Davis, M. B„ Calcote, R. R., Sugita, S., and Takahara, H. 1998. Patchy invasion and the origin of a hemlock-hardwoods mosaic. Ecology 78, 2641-2659.

Del court, P. A. and Del court, H. R. 1987. "Long-Term Forest Dynamics of the temperate zone." Springer-Verlag, New York.

Drake, J. 1990. Communities as assembled structures: do rules govern pattern? "Free 5, 159-164.

Ewel, J. J. 1986. Invasibility: lessons from south Florida. In "Ecology of Biological Invasions of North America and Hawaii." (H. A. Mooney and J. A. Drake, Eds), pp. 214-230. Springer-Verlag, New York.

Ferrari, J. B., 1999. Fine-scale patterns of leaf litterfall and nitrogen cycling in an old-growth forest. Can. }. Forestry Res. 29, 291-302

Frelich, L. E. and Lorimer, C. G. 1991. Natural disturbance regimes in hemlock-hardwood forests of the Upper Great Lakes Region. Ecol. Monogr. 61, 145-164.

Frelich L. E. and Graumlich, L. J. 1994. Age-class distribution and spatial patterns in an old-growth hemlock-hardwood forest. Can. J. Forestry Res. 24, 1939-1947.

Fuller, J. L. 1998. Ecological impact of the mid-HoIocene hemlock decline in southern Ontariuo, Canada. Ecology 74, 2337-3351.

Futyma, R. P. 1982. Postglacial vegetation of eastern Upper Michigan. Ph.D. Thesis, University of Michigan, Ann Arbor, Mchigan.

Greene, D. F. and Johnson, E. A. 1995. Long-distance dispersal of tree seeds. Can J. Bot. 73, 1036-1045.

Huntley, B. and Birks, H. J. B. 1981. "An Atlas of Past and Present Pollen Maps for Europe: 0-13,000 Years." Cambridge University Press, Cambridge, United Kingdom.

Huntley, B. and Webb, T. Ill, 1988. "Vegetation History." Kluwer Sci. Dordrecht, Netherlands.

Houghton, J. T„ Filho, L. G. M., Callander, B. A. Harris, N., Kattenberg, A., and Maskell, K„ Eds. 1996. "Climate Change 1995: The Science of Climate Change. IPPC Report." Cambridge University Press, Cambridge, United Kingdom.

Johnson, W. C. and Adkisson, C. S. 1985. Dispersal of beech nuts by blue jays in fragmented landscapes. Am. Midland Nat. 113, 319-324.

Kapp, R. O. 1977. Late Pleistocene and postglacial plant communities of the Great Lakes Region. In "Geobotany." (R.C. Romans.), pp. 1-27 Ed. Plenum, New York.

Kullman, K. 1996. Norway spruce present in the Scandes Mountains, Sweden at 8000 B.P.: new light on Holocene tree spread. Global Ecol. Bio-geogr. Lett. 5, 94-101.

Little, E. L. Jr. 1971. "Atlas of United States Trees. Vol. 1. Conifers and Important Hardwoods." Misc. Publ. 1146. USDA Forest Service, U.S. Government Printing Office, Washington D.C.

Lodge, D. M. 1993. Biological invasions: lessons for ecology. Tree 8, 133-137.

Markgraf, V„ McGlone, M., and Hope, G. 1995. Neogene paleoenviron-mental and paleoclimatic change in southern temperate ecosystems—a southern perspective. Tree 10, 143-147.

McGlone, M. S. 1997. The response of New Zealand forest diversity to Quaternary climates. PP. 73-80. In Past and Future Rapid Environmental Changes: The spatial and Evolutionary Responses of Terrestrial Biota. (B. Huntley, W. Cramer, A. V. Morgan, H. C. Prentice, and J. R. M. Allen Eds.) Springer-Verlag, Berlin.

Overpeck, J. T„ Rind, D„ and Goldberg, R. 1990. Climate-induced changes in forest disturbance and vegetation. Nature 343, 51-53.

Parshall, T. E. 1995. Canopy mortality and stand-scale change in a northern hemlock-hardwood forest. Can. J. Forestry Res. 25, 1466-1478.

Parshall, T. E. 1998. Hemlock invasion and population increase near the species limit in Wisconsin. PhD Thesis, University of Minnesota, St. Paul, Minnesota.

Parshall, T. E., Calcote, R., Davis, M. B., Walker, K. and Douglas, C. Tree mortality in an old-growth hemlock-hardwood forest. /. of Vegeta. Sci. (in review).

Pastor, J. and Post, W. M. 1988. Response of northern forests to C02-induced climate change. Nature 334, 55-58.

Penalba, M. C. and Payette, S. 1997. Late-Holocene expansion of eastern larch (Larix lariciua {DuRoil K.Koch) in northwestern Quebec. Q. Res. 48, 1 14-121.

l'ielou, E. C. 1979. "Biogeography." Wiley-Interscience, New York. Pitelka, L. F. and Plant Migration Workshop Group. 1997. Plant migration and climate change. Am. Scientist 85, 464-423. Shigesada, N., Kawasaki, K, and Takeda, Y. 1995. Modeling stratified diffusion in biological invasions. Am. Nat. 146, 229-251. Sugita, S. 1991. Pollen representation in Quaternary sediments: theory and methods in patchy vegetation. /. Eco. 82, 879-898. Swain, A. and Winkler, M. 1983. Forest and disturbance history at Apostle

Islands National Eakeshore. Park Sei. 3, 3-5. Thompson, R. S. 1988. Western North America. In "Vegetation History." (B. Huntley and T. Webb III, Eds.), pp. 454-458. Kiuwer, Dordrecht, Netherlands.

Tsukada, M. 1988. Japan. In "Vegetation Iiistory." (B. Huntley and T.Webb

III, Eds.), pp. 459-518. Kiuwer, Dordrecht, Netherlands. Webb, T III. 1974. A vegetational history from northern Wisconsin. Evidence from modern and fossil pollen. Am. Midland Nat 92, 12-34. Webb, T, III. 1988. Eastern North America. In "Vegetation History." (B. Huntley and T. Webb III, Eds.), pp. 387-414. Kiuwer Acad. Publ., Dordrecht, Netherlands. Webb, S. L. 1986. Potential roles of passenger pigeons and other vertebrates in the rapid Holocene migrations of nut trees. Q. Res. 26, 367-375.

Webb, S. 1987. Beech range extension and vegetation history: pollen stratigraphy of two Wisconsin Lakes. Ecology 68, 1993-2005. Woods, K. D. and Davis, M. B. 1989. Paleoecology of range limits: beech in the upper peninsula of Michigan. Ecology 70, 681-696.

This Page Intentionally Left Blank

0 0

Post a comment