Margaret B. Davis

Quaternary History and the Stability of Forest Communities

The concept of stability of plant communities inevitably enters discussions of forest succession. Succession following disturbance often leads to restoration of the original community, which is seen as the equilibrium community for the site. In this context, succession can be viewed as a mechanism maintaining stability.

Lags in vegetation response to greenhouse warming.

Fossil pollen in sediments documents vegetation responses to climatic changes in the past. Beech (Fagus grandifolia), with animal-dispersed seeds, moved across Lake Michigan or around its southern margin, becoming established in Wisconsin about 1000 years after populations were established in Michigan. Hemlock (Tsuga canadensis), with wind-dispersed seeds, colonized a 50,000 km2 area in northern Michigan between 6000 and 5000 years ago. These tree species extended ranges northward at average rates of 20–25 km per century. To track climatic changes in the future, caused by the greenhouse effect, however, their range limit would need to move northward 100 km per °C warming, or about 300 km per century, an order of magnitude faster than range extension in the past. Yet range extension in the future would be less efficient than in the past, because advance disjunct colonies have been extirpated by human disturbance, and because the seed source is reduced due to reductions in tree populations following logging. Many species of trees may not be able to disperse rapidly enough to track climate, and woodland herbs, which have less efficient seed dispersal mechanisms, may be in danger of extinction.

Sensitivity of cool-temperate forests and their fossil pollen record to rapid temperature change

Abstract Simulations of cool-temperate forest growth in response to climatic change using the JABOWA computer model show that a decrease of 600 growing degree-days (equivalent to a 2°C decrease in mean annual temperature) causes red spruce ( Picea rubens ) to replace sugar maple ( Acer saccharum ) as the dominant tree. These changes are delayed 100–200 yr after the climatic cooling, producing gradual forest changes in response to abrupt temperature changes, and reducing the amplitude of response to brief climatic events. Soils and disturbances affect the speed and magnitude of forest response. The delayed responses are caused by the difference in sensitivity of adult trees and younger stages. The length of the delay depends on the life history characteristics of the dominant species. Delayed responses imply that fossil pollen deposits, even if they faithfully record the abundances of trees in forests, may not be able to resolve climatic changes within 100–200 yr, or to record very brief climatic events. This explains why pollen deposits do not as yet show responses to climatic changes during the past 100 yr. Only the Little Ice Age, which lasted several centuries, caused sufficient forest change to be recorded in fossil pollen, and only at certain sites.

Climatic Changes in Southern Connecticut Recorded by Pollen Deposition at Rogers Lake

Rates of deposition of pollen grains throughout late— and postglacial time were determined from the pollen concentration in radiocarbon—dated sediment. Changes by a factor of 5 or more for all except rate pollen types from one level to the next were considered significant indication of changes in the pollen input to the lake, reflecting changes in the pollen productivity of the surrounding vegetation. Low pollen deposition rates in the oldest sediments reflect the prevalence of tundra vegetation between 14,000 and 12,000 years ago. An increase in the rate for tree pollen occurred 12,000 years ago, when boreal woodland became established. The rates continued to increase until a sudden sharp rise for white pine, hemlock, poplar, oak, and maple pollen 9,000 years ago marked the establishment of forest, similar perhaps to modern forests of the northern Great Lakes region. Pine pollen rates decline 8,000 years ago, and deciduous tree pollen became dominant. Ragweed pollen was deposited at relatively high rates...

Dispersal versus climate: Expansion of Fagus and Tsuga into the Upper Great Lakes region

Pollen records for American beech (Fagus grandifolia) and eastern hemlock (Tsuga canadensis) compiled from 50 sites in Michigan and Wisconsin, USA, show that both species entered the Upper Great Lakes region about 7000 yr B.P., reaching their western and southwestern boundaries between 2000 and 1000 yr B.P. Fagus advanced northward into lower Michigan as a continuous front, except where Lake Michigan posed a geographic barrier. Colonies were established on the far side of the lake after a 1000 year lag, implying that longdistance dispersal across a 100-km wide barrier can occur. The Fagus range may not have been in equilibrium with climate for one or two thousand years before this time, when seeds were dispersed across the lake to Wisconsin. Tsuga seeds may have been dispersed 150 km or more from Ontario to reach Upper Michigan. Scattered colonies were established 6000–7000 yr B.P. on either side of Lake Michigan, which did not pose a significant barrier to this wind-dispersed species, Tsuga spread rapidly over a large region prior to 5000 yr B.P. Subsequent expansion to the west occurred more slowly, and may reflect gradual climatic changes in northern Wisconsin during the second half of the Holocene. Tsugas range may have been limited by dispersal, rather than climate, for an unknown length of time prior to 5000 yr B.P. During this period Tsuga was expanding its range rapidly. The study shows, however, that it is difficult to devise rigorous tests to distinguish between dispersal limitations and climate as factors limiting range limits in the past.

Patch Formation and Maintenance in an Old‐Growth Hemlock‐Hardwood Forest

Cause of patch formation was investigated on a 7.2 ha study area in Sylvania Wilderness Area, a primary forest remnant in Upper Michigan comprising a mosaic of hemlock, sugar maple, and mixed-forest patches. Spatial autocorrelation analysis of the stem map indicates that, although most species pairs have a neutral association between canopy trees and understory trees of other species, hemlock and sugar maple canopy trees both have strong positive self association and negative reciprocal association with each other. No species pairs have a positive reciprocal association on regeneration with each other. MOSAIC, a Markov simulation model in which transition probabilities depend on neighborhood species composition, shows that the negative reciprocal association between hemlock and sugar maple of the intensity observed in this study, could lead to spatial separation into monodominant patches over long time periods (3000 yr). The mixed-forest patches are along spatial continua of varying steepness between sugar maple and hemlock patches. Interactions between sugar maple and hemlock overstory and understory trees, along with the pattern of invasion of hemlock, provide a reasonable explanation for the patch structure. Pedological, topographical, and disturbance history differences do not coincide with the location of patches within upland forests on the study area.

Holocene climate of New England

Stratigraphic studies of pollen and macrofossils from six sites at different elevations in the White Mountains of New Hampshire demonstrate changes in the distributions of four coniferous tree species during the Holocene. Two species presently confined to low elevations extended farther up the mountain slopes during the early Holocene: white pine grew 350 m above its present limit beginning 9000 yr B.P., while hemlock grew 300–400 m above its present limit soon after the species immigrated to the region 7000 yr. B.P. Hemlock disappeared from the highest sites about 5000 yr B.P., but both species persisted at sites 50–350 m above their present limits until the Little Ice Age began a few centuries ago. The history of the two main high-elevation conifers is more difficult to interpret. Spruce and fir first occur near their present upper limits 9000 or 10,000 yr B.P. Fir persisted in abundance at elevations similar to those where it occurs today throughout the Holocene, while spruce became infrequent at all elevations from the beginning of the Holocene until 2000 yr B.P. These facts suggest a more complex series of changes than a mere upward shift of the modern environmental gradient. Nevertheless, we conclude that the minimum climatic change which would explain the upward extensions of hemlock and white pine is a rise in temperature, perhaps as much as 2°C. The interval of maximum warmth started 9000 yr B.P. and lasted at least until 5000 yr B.P., correlative with the Prairie Period in Minnesota.

Pollen grains in lake sediments: redeposition caused by seasonal water circulation.

Annual pollen deposition per unit area measured in sediment traps is two to four times greater than deposition measured in surface sediment cores. The difference is due to repeated redeposition of pollen from the sediment surface during seasons of water circulation. This process reduces variations in the percentages of different pollen types in sediment, but causes differences in the total amount of pollen accumulated in various parts of the lake basin.

PATCHY INVASION AND THE ORIGIN OF A HEMLOCK-HARDWOODS FOREST MOSAIC

The record of forest invasion by eastern hemlock (Tsuga canadensis) during the course of Holocene migration provides useful information about invasion processes in temperate forest, a system that has been invaded by few exotic species. We used fossil pollen preserved in small forest hollows, which record forest composition on the scale of 1-3 ha, to study hemlock invasion of forests in the Sylvania Wilderness in the upper peninsula of Michigan, where there is now a mosaic of 3-30 ha stands dominated either by hemlock or by sugar maple (Acer saccharum) and basswood (Tilia americana). Fossil pollen was interpreted by comparison with 66 surface samples from small hollows in Michigan and Wisconsin, using three different statistical methods: pollen ratios, a dissim- ilarity index, and canonical variates analysis (CVA) ordination. We found that four hemlock stands along a 10-km transect across Sylvania originated as patches of white pine (Pinus strobus) forest that were invaded by hemlock -3000 yr ago (calibrated 14C dating) when hemlock expanded its range in northern Michigan. Invasion occurred at about the same time (within 800 yr) at all sites but was not associated with disturbance at any site. Over the next several thousand years hemlock coexisted with white pine, but eventually, at a time that differs from site to site, hemlock became dominant, and white pine disappeared from all but one of the four stands. These changes were apparently driven by climate changes over the last 4000 yr that caused the water table to rise (Brugam and Johnson 1997). The history of four nearby maple stands is more variable and less well understood. Unlike the hemlock stands, three of the four maple patches were not dominated by pine at the time of hemlock invasion, but instead had abundant oak (Quercus) and/or maple. Two of these stands were not invaded by hemlock, and the third, if invaded at all, was invaded for a few centuries by low densities of hemlock trees. Thus invasion by hemlock was sensitive to the species composition of the resident forest. Sugar maple and basswood increased in these stands, and by 2000 to 800 yr ago, depending on site, they resembled modern maple stands. The fourth patch was invaded by hemlock, but it was converted to a maple stand 1000-500 yr ago. A wood layer in the sediment is evidence that a catastrophic windstorm may have been responsible.

Pollen Accumulation Rates: Estimates from Late-Glacial Sediment of Rogers Lake

Absolute pollen deposition in a Connecticut lake over a 4000-year interval has been estimated from pollen frequencies in a core of late-glacial sediment dated by radiocarbon techniques. The rate of total sediment accumulation as measured after burial was statistically constant at 0.036 centimeter per year, but the rate of deposition of pollen grains onto the sediment increased from 600 to 900 grains 14,000 years ago to 9000 per square centimeter per year 10,000 years ago. A major increase in the deposition of tree pollen occurred about 11,500 years ago, at the beginning of the spruce pollen zone. Presentation of data in conventional (percentage) form masks the magnitude of this change and distorts many of the changes in accumulation rates for individual types of pollen; moreover it magnifies statistical variation in the herb zone where all pollen is scarce.