1

Approximate maximum extent of ice during Pleistocene (last ~ 2 million years). Ice

cover fluctuated between glacial and ‘interglacial’ periods over 20 times during this

period. We’re currently in an ‘interglacial’. Prior to the Pleistocene, global climates

were warmer and ice extent limited (during some periods, no ice persisted anywhere on

the planet…).

We have talked about how disturbances like fire and wind, occurring at periods of

decades to centuries and at scales of a few m to km, might affect dynamics of ecosystems

and communities. Now consider response to such global, long-term changes as

glaciation and deglaciation….

2

Long temporal view first: Climate has varied at multiple scales. Over last 2 million

years (Pleistocene era), variation has been dominated by cyclic advance and retreat of

continental ice-sheets.

3

The last North American ice-sheet is referred to as the ‘Wisconsin’ glaciation. It lasted

about 100,000 years, but reached maximum extent about 20,000 years before present

(ybp) and began rapid retreat about 14,000 ybp. This map shows its shrinkage from

about 18000 ybp

A series of more detailed maps of ice in eastern North America: at maximum and at

beginning of retreat; note lakes forming at ice margin and rivers draining melt.

The weight of ice actually depressed the land surface so areas near ice margins were

Low-lying and thus became temporary lakes (as surface ‘rebounded’ they drained).

5

Ice retreats from our area just after ca.13,000 ybp, with arrival of fauna and flora typical

of tundra and treeline forests shortly thereafter. Massive proglacial lakes formed and

glacial meltwater was drained off through progressively shifting channels corresponding

to modern river valleys. These rivers were many times larger than any modern river.

6

Early post-glacial ecosystems here may have looked much like far-northern modern

Ecosystems (but with mastodons)

7

8

Modern rivers like Mississippi are much smaller than the rivers that create their valleys

9

Because of depression of land surface, ocean water intruded up the valley of the

St. Lawrence and into the Lake Champlain basin (temporarily). Lake Aggasiz was one

of the largest freshwater lakes ever; it drained rather abruptly when the St. Lawrence

River valley was uncovered by retreating ice.

10

The north-south valleys of western New England and eastern New York were filled with

a series of proglacial lakes. The Lake Champlain basin was, briefly, an arm of the north

Atlantic before the crust ‘rebounded’ from the weight of the glacier sufficiently to lift it

above sea-level. Ice did not leave the highlands of Labrador until about 5000 ybp. The

crust around Hudson Bay is still rebounding so that the bay is shrinking.

11

12

Many northeastern mountains show extensive glacial ‘carving’ even though no

mountain glaciers remain (this is Mt. Katahdin, in Maine)

13

“Glacial striations’ in rock were one of the first clues to the existence of continental

glaciation. The thousands of ‘pot-hole’ lakes of parts of northern U.S. and Canada were

formed when remnant blocks of ice, buried in debris left by the glacier (‘till’) melted,

and the surrounding materials collapsed into the hole (these collapses might have

been centuries after the glacial retreat).

14

Glaciers are flowing streams of ice; their terminus is determined by the balance

between ice formation at their source and melt rates as they move away from that source.

When glaciers were stable for some period, they acted as ‘conveyor belts for glacial

debris, leaving material in large piles at their terminus; these piles are called morraines.

Long Island is a large terminal morraine.

15

Other things changed during glacial cycles. During glacial phases, much of the western

U.S., even though not ice-covered, became considerably cooler an moister. Many of the

intermountain valleys, now dry, even deserts, held lakes. Great Salt Lake is a remnant of

one of these. Others have dried completely, leaving salt flats.

16

17

All of these processes repeated many times, although the magnitudes and periods of

glacial cycles changed. The most recent five or six were particularly severe and long-

lasting.

18

So what happened to ecological systems? For a long time, ecologists imagined that

communities simply moved with the climate in zones – that communities would have

been present in same sequence along, say, a transect northward from the Gulf of Mexico,

just in a more compressed form, like an accordion squeezed or stretched depending on

climate. (This figure comes from a classic ecology text from the 1950s.) This model sees

communities as entities that behave as an integrated whole. Look back at the discussion

of history of ‘community concepts’ from previous notes; note that this idea is closely

aligned with the Clements ‘pseudo-organismic’ view of communities

However, as a hypothesis, this model was very difficult to test. Until the 1960s, there

was no easy way to reconstruct past communities to assess this model, so little evidence

against which these visualizations could be compared.

19

However, some lake researchers had long realized that lake sediments might record long-

term records of surrounding vegetation in the form of pollen. Many trees are wind-

pollinated and, every year, large quantities of dispersed pollen fall in bodies of water (the

yellow clouds floating along the shore here are made of pine pollen).

Sediments in lakes and peat accumulated in bogs can be over 1/3 pollen by mass, but can

also preserve other types of plant material collected from surrounding vegetation. Since

these sediments are often anaerobic (or highly acidic in the case of peats), decay is very

slow. Since sediments accumulate over time, remains are in temporal sequence, from the

deepest sediments deposited immediately after glacial retreat to uppermost sediments

being deposited currently. Technologies for extracting cores of lake sediments allowed

access to a vast new data set. The most abundant fossils in lake sediments are pollen

grains. Wind-pollinated species of trees shed vast quantities of pollen, and pollen walls

are made of highly resistant sporopollenin (very similar to chitin). Pollen assemblages at

a particular depth reflect forest composition (but complexly; different species produce

pollen at different rates, some species can’t be easily distinguished from pollen, etc.)

20

Some pollen grains

21

Technologies for coring lake sediments were developed in the 1960s and later. Tubes are

driven into sediments from stabilized raft platforms. This involves a lot of smelly mud,

tangled ropes, etc.

22

Lake basins can hold 5-10 m of post-glacial sediments. Below the first few cm,

sediments tend to become condensed and take on the texture of a very firm, black jelly

made up almost entirely of organic material. This type of sediment is called ‘gyttja’

(derived from Swedish; there are many lakes in Sweden.)

23

Winter coring has some advantages and disadvantages

24

Pollen extracted from different levels (ages) in a core can be identified and counted to

develop ‘assemblages’ – species compositions – at various ages. These counts can be

compiled into pollen diagrams.

‘Pollen diagrams’ show trends in abundance for different species or groups of plants.

The vertical axis is time, with older deposits at the bottom. Each curve represents a

single type of tree (species or genus), and shows changes in its percentage representation

in assemblages at each depth/age. Rogers Lake is in Connecticut, and sediments began

accumulating here with the retreat of the ice sheet about 14,000 ybp. Note that pine and

oak pollen are most abundant – but these groups are very prolific pollen producers, so

their representation in the pollen assemblage does not reflect actual relative abundance

on the landscape. Maple pollen, on the other hand is fairly rare – but maple trees

produce little wind-blown pollen, so might be pretty abundant in actual communities.

And it’s not possible to fully discriminate the various species in each genus.

Despite these complications, it’s possible to see changes from assemblages that suggest

colder, tundra-like communities, to communities of species typical of northern/boreal

forests, to a series of forests compositions resembling forests of more temperate

climates.

25

Margaret Davis compared pollen assemblages at various depths at Rogers Lake to

modern assemblages elsewhere and judged what past vegetation around the lake might

have looked like in the past on the basis of the most similar modern pollen assemblages

(for example, pollen assemblages at Rogers Lake from 9500-12000 years ago looked like

pollen currently being deposited in lakes about 800 mi north in north-central Quebec in

northern boreal forests). So far, these findings are reasonably consistent with the idea of

vegetation ‘zones’ expanding and contracting with climate change. However, this is just

one lake.

26

27

Patterns became more complicated when paleoecologist Margaret Davis compiled data

from dozens of pollen coring sites. Here, small numbers indicate the age of first

occurrence of a species in 1000s of years. The red lines suggest the northern range limit

of each species for a given number of millennia in the past. Any values of 20,000 or

more years before present (ybp) are taken to indicate areas where a species persisted

through the full-glacial period. Larch (Larix), for example, appears to have occurred in

the lower midwest at full-glacial (about 20,000 ybp), but never on the eastern seaboard. It

spread northward to the Great Lakes region, then expanded eastward and westward into

current range (stippled area). All of the species shown here currently co-occur broadly

across northern North America and are typically considered co-dominants of the boreal

forest community ‘type’. But note that 20,000 ybp, they did not occur together anywhere;

jack pine and balsam fir occurred only along the east coast, while larch and spruce

occurred only inland. There was no ‘boreal forest’ as we would recognize it today. This

reconstruction is more in line with the Gleason ‘individualistic species’ hypothesis of

community pattern. As climate changes, species’ ranges shift independently, combining

and recombining in unique and changing ways. Some paleoecologists use the term ‘no-

modern-analogue’ communities to describe past assemblages of species that don’t exist

anywhere in the modern world. It may be that no-analogue communities were typical of

past climates.

28

Similarly, the dominant species of the northern hardwood forests (typical of older forests

on mesic sites in central and northern New England) did not occur together at full-

glacial. White pine and hemlock appear to have lived in ‘glacial refugia’ along the east

coast in areas subsequently flooded as sea level rose with the melting of the ice. At some

times in the past, then, species occurred in combinations not generally observed today

(and vice versa); these are often referred to as ‘no-analogue communities’.

29

There was substantial additional land area available for occupancy at peak glacial.

30

Animals show similar patterns. Beetle exoskeletons can also be well-preserved in

sediments. Species found together from 16,000 ybp near the glacier’s edge in Iowa are

now found in widely separate locations from Alaska to the mountaintops of New

England; the community that existed then is nowhere present today.

31

More local patterns can be examined similarly. This map of the White Mountains of

New Hampshire shows a collection of coring sites at a range of elevations

32

Pollen diagrams from this elevational transect also show striking changes (and ‘no-

analogue’ situations) along elevational gradients. For example, eastern hemlock (Tsuga

canadensis) underwent a massive population collapse everywhere in its range,

simultaneously, at about 4800 ybp. Because the ‘hemlock decline’ happened

simultaneously across its range, and other late-successional trees did not show large

changes, this is usually interpreted as the result of a new disease (or, possibly, an insect

pest), rather than as the result of climate changes.

33

But reconstruction of vegetation patterns along this elevational transect for various

intervals over the last 12,000 years show complex shifts that don’t always ‘make sense’

in terms of modern vegetation types and patterns; again, there appear to be ‘no-analogue’

types of vegetation. For example, note the ‘mixed woodland’ type of 9500 years ago and

the ‘paper birch and fir woodland’ of 8500 ybp.

34

Some diagrams from the White Mts. transect; this work was done by Ray Spear, a student

of Margaret Davis

35

Immediately following the hemlock decline, species like oak, elm, and pines showed

sharp increases in pollen abundance. Beech, maple, and birch pollen increased a couple

of hundred years later. These dynamics are typical responses to canopy disturbance in

forests today.

36

One further example of community change over longer time periods as revealed by

paleoecological techniques.

Sometimes, changes in community structure and composition appear to have been more

dramatic and, perhaps, less predictable. Note that dark blue area in SE Minnesota. This

is often referred to as the ‘Big Woods’ region. It was, prior to European settlement,

dominated by mesic forests composed primarily of maple, basswood, and elm. To the

west, south, and southeast was prairie grassland (yellow); to the northeast and north, oak

woodland/savanna (green). (Red and orange colors are northern pine and aspen an boreal

forests.) When settlers first arrived, the ‘Big Woods’ forests were seen as indicators of

particularly fertile soils, and they were some of the first areas agriculturally settled.

(These pre-settlement patterns are recognized from the notes of initial land surveys

(discussed in earlier notes). Most ecologists, for many years, took these ‘natural

communities’ to represent the ‘climax’ or mature, natural vegetation of the state,

assuming that it would have looked like this, more or less, for very long periods.

37

A photo of ‘Big Woods’ forest in Nerstrand Woods State Park near Northfield, MN, and a

map focusing on the ‘Big Woods’ area. These forests would look very similar to mesic

forests of much of the northeastern U.S. (although they lack beech and hemlock.)

38

Prairies and oak savanna/woodland vegetation burns very frequently (typically every few

years) and are fire-maintained. While mature oaks can resist fire, seedlings can’t survive

it. The ‘Big Woods’ species are completely intolerant of fire, so frequent fires can

maintain grasslands. On the other hand, mesic forests of maple and basswood are

extremely difficult to burn; there’s little fuel on the forest floor and the foliage is not very

flammable when green (unlike resinous conifer needles).

39

Eric Grimm, as a grad student at University of Minnesota, studied the history of the Big

Woods region. Here, he mapped ‘witness trees’ of various species as recorded by the

early land surveyors. Elm, basswood, and sugar maple are all dominants restricted to

mesic forests; their distributions show that the ‘big woods’ area was pretty sharply

defined. In fact, most of its boundaries correspond to topographic features that might

serve as firebreaks.

40

Oak and aspen – fire-adapted tree species – occur around the edges of the Big Woods

region where the transition is not directly to grasslands/prairies.

41

Eric Grimm collected pollen cores from lakes across the Big Woods area and in adjacent

areas to the east and west. He found that the mesic ‘Big Woods’ forest had been present

only for the last few centuries; prior to that, the area was occupied by either oak

woodland, or even prairie. The establishment of the Big Woods forests coincided with a

cool, moist period known as the Little Ice Age – but these forests persisted into drier,

warmer climates subsequently. Grimm suggested that either mesic forest OR oak

woodland OR prairie could exist in the same climate by promoting fire frequencies that

are favorable to their own maintenance. Once an area burns, fire-tolerant species are

favored and they tend to promote more frequent fires. But mesic trees of the ‘big woods’,

once established, tend to suppress fire. Thus, either type might persist indefinitely under

the same conditions unless either a) climate changes to a point where only one type can

survive, or b) some unusual event (maybe a severe drought – or the opposite)‘tips’ the

system into the other state – a ‘two-state’ system that tends to be stable in whichever state

it occupies. (Technically, a metastable system vulnerable to ‘cusp catastrophes’).

(Upper right graph shows ‘states’ of communities for a transect of sites from W to E

across the Big Woods: darkest shading indicates mesic forest (big woods); other shadings

are either oak savanna/woodland or prairie. The lower right graph illustrates Grimm’s

notion of ‘alternate stable states’ along a moisture gradient. Each type can persist along

the range of conditions illustrated…)

42

Paleoecology studies have suggested to many scientists that, as we move forward into

new climatic conditions, we should expect ecosystems to change in unpredictable ways –

that we may anticipate new ‘no-analogue’ communities.

43

Researchers have attempted to estimate future species distributions by inferring climatic

tolerances (niche parameters) from current distributions, and applying these tolerances to

predicted future climate maps. Here, modern distribution of sugar maple is shown at

upper left; one projection of likely distributions in 2100, under a typical ‘global warming’

forecast, is in the upper right graph (sugar maple would be absent from all but a few U.S.

states at the northern border)

44

The same for beech

45

And red oak; much of the northeast will be much more oak-dominate.

46

Balsam fir (and other boreal species) will be essentially eliminated from the eastern U.S.

(in these models).