Century-scale climate changes reshaped circulation patterns over the North
Atlantic and adjacent regions during the last glacial-to-interglacial transition.
Here, we show that vegetation across eastern North America shifted dramatically
at the beginning and end of the Younger Dryas chronozone (YDC: 12,900-11,600
cal yr B.P.), when changes in ocean circulation rapidly cooled and then warmed
the North Atlantic sea-surface. On both the site-specific scale and the continental-scale,
vegetation changed only gradually during the millennia before (15,000-13,000
cal yr B.P.) and after (11,000-9000 cal yr B.P.) the YDC, but climate changes
ca 12,900 and 11,600 cal yr B.P. altered the vegetation on both spatial scales
within centuries. Plant associations changed and some taxa rapidly migrated
hundreds of kilometers (>300 km within ~100 yr). In limited regions near
the North Atlantic coast, abrupt cooling ca 12,900 cal yr B.P. resulted in a
return to earlier vegetation types. Elsewhere, however, the vegetation patterns
during the YDC were distinct from those of both earlier and later intervals.
They indicate abrupt, `non-reversing' seasonal temperature changes that were
probably related to atmospheric circulation changes during the YDC, rather than
to the direct influence of North Atlantic sea-surface temperatures. Atmospheric
circulation patterns during the YDC were unique within the last 21,000 yr because
of a unique combination of climate controls. Insolation, ice sheet extent, and
atmospheric composition were significantly different from their full-glacial
states, even when the North Atlantic returned to near full-glacial conditions.
The YDC vegetation patterns demonstrate (1) rapid ecological responsiveness
to abrupt climate change and (2) spatially varied patterns of YDC climate change.
|Figure 1. Spatial patterns of climate change and vegetation response during the YDC. Maps show the difference between YDC and modern surface temperatures during January (A) and July (B), as simulated by the GISS GCM (Rind et al., 1986; Rind, 1994). Colder-than-modern temperatures over the North Atlantic are shown in shades of blue, and warmer-than-modern temperatures over the mid-continent are shown in red. Changes in mean January and July temperatures (C), simulated by the GISS model (Rind, 1994), are used to represent the possible differences between four locations along an eastwest transect (B). Red symbols (C) represent simulated conditions before the YDC and blue symbols represent simulated conditions during the YDC. Red arrows illustrate the simulated magnitude and direction of temperature change. The modern percentages of spruce (Picea), northeastern pine (Pinus), and elm (Ulmus) pollen (C) are also plotted as large grey and black symbols, which represent two levels of abundance for each taxon. Small grey symbols represent the climatic position of all modern pollen surface samples from eastern North America. The modern percentages of each taxon is also shown with respect to a single climate variable, mean July temperatures (D), with responses to simulated conditions at locations 1 and 2 before (red) and during the YDC (blue) also shown (D). Smaller black symbols (D) highlight a selection of modern pollen samples within a specific range of moisture conditions, in order to show that moisture explains much of the variance at a given temperature. To avoid the ranges of southern and western pine species, which have different climatic tolerances (Thompson et al., 1999), the pine data shown are from the region north of 39o N latitude and east of 110o W longitude.|
|Figure 2. Late glacial pollen stratigraphies from nine sites in eastern North America (from left to right: King, 1981; Williams, 1974; Shane, 1987; Shane and Anderson, 1993; Spear and Miller, 1976; Davis, 1969; Suter, 1987; Davis et al., 1975; Mayle and Cwynar, 1995). Light grey bars mark the European Bölling/Alleröd chronozone (B/A), from ca 15,000 to 12,900 cal yr B.P. Dark grey bars indicate the Younger Dryas chronozone (YDC), from ca 12,900 to 11,600 cal yr B.P. (see Appendix A with regard to these ages). Sedge (Cyperaceae) and birch (Betula) pollen percentages are shown in grey. Oak (Quercus) pollen percentages are also shown in grey, but are shown only for Chatsworth Bog, Illinois, where no signi.cant pine (Pinus) pollen (black) was recorded (King, 1981). Triangles indicate the position of calibrated radiocarbon dates. Grey triangles denote dates that have been adjusted for old carbon.|
|Figure 3. Pine and spruce pollen percentages track abrupt range shifts. Records of pine (Pinus) (A) and spruce (Picea) (B) pollen percentages are plotted with time on a vertical axis to show changes in their geographic distributions. Arrows indicate the eastwest range shifts of pine (A) and the northsouth range shifts of spruce (B). Each site is labeled by state as in Fig. 2. The Stotzel-Leis site in Ohio (Shane, 1987) is shown in grey in inset A, and superimposed upon the stratigraphy from Pretty Lake in Indiana (Williams, 1974).|
|Figure 4. The abundance of pine pollen across eastern North America before, during, and after the YDC. Histograms represent the percent pine (Pinus) pollen at each of nine sites (Fig. 2) at 13,000, 12,000 (YDC), and 11,000 cal yr B.P. The percentages are plotted with respect to longitude to show changes in abundance across the range of pine. Each site is labeled by state, as in Fig. 2. Horizontal grey bars mark the range of sites where the presence of pine populations is indicated by pollen percentages of >10%. For comparison to modern, see Fig. 1D.|
|Figure 5. Maps of plant associations and vegetation change at 1000-cal-yr intervals between 14,000 and 10,000 cal yr B.P. Different colors illustrate different vegetation assemblages. Individual plant taxa are mapped as red, yellow, or blue with overlapping ranges represented by the combinations of primary colors: orange, purple, green, or gold. Grey represents the absence of the mapped taxa, and white represents regions with no data. Two combinations of three taxa are mapped: (A) regions with greater than 20% spruce (Picea; red), 5% sedge (Cyperaceae; yellow), and 20% pine (Pinus; blue) pollen, and (B) regions with greater than 5% ash (Fraxinus; red), 15% oak (Quercus; yellow), and 6% elm (Ulmus; blue) pollen. Maps of square chord distances (SCD) represent the dissimilarity between fossil and modern pollen assemblages (C). SCD values greater than 0.15 (black) represent pollen assemblages that have no modern equivalent. Offset maps (D) use SCDs to show the amount of change between 1000-yr intervals. High SCDs document large changes.|
|Figure 6. Lake-level elevations and pollen percentages in Wisconsin and Minnesota from 14,000 to 7000 cal yr B.P. Lake-level elevations at Lake Mendota, Wisconsin (A), are inferred from the elevations of radiocarbon-dated sediments (Winkler et al., 1986). Aquatic macrofossil assemblages and sediment characteristics from Almora Lake, Minnesota (Digerfeldt et al., 1992), constrain an envelope of possible water levels there (B). The lake-level data are compared to changes in the relative abundance of spruce (Picea), pine (Pinus), and oak (Quercus) pollen at near-by sites: (A) Devils Lake, Wisconsin (Maher, 1982) and (B) Reidel Lake, Minnesota (Almquist-Jacobson et al., 1992).|