Sample 1 8000 years old
400 000 grains
8 years represented by 1 cm thickness
50 000 grains deposited on a cm2 surface each year
Sample 2 14 000 years old
20 000 grains rm-3
25 years represented by 1 cm thickness
800 grains deposited on a cm2 surface each year
FIGURE 9.8 Examples of pollen influx (pollen flux density) calculations from measurements of pollen concentration and the rate of accumulation of the sediment matrix (Davis, 1963).
1993b; R. S. Webb et al., 1993). Consequently, pollen flux density studies have taken a back seat to pollen percentages in most modern paleoclimatic studies.
Pollen diagrams contain a large amount of information on the covariance of different pollen types through time. In order to facilitate comparison between different sites, the stratigraphic record in pollen diagrams is commonly subdivided into pollen zones; these are biostratigraphic units defined on the basis of the characteristic fossil pollen assemblage (see Fig. 9.7). Generally pollen zones contain a homogeneous assemblage of pollen and spores, but some investigators may recognize a zone that is characterized by abrupt changes. Needless to say, the definition of what constitutes a zone is a rather subjective decision and may not be agreed upon by different scientists (Tzedakis, 1994). Human nature also compels us to look for correlations with zonations previously "identified," thereby reinforcing systems which may not justify such blind faith! To avoid these problems, a number of more objective, computer-based methods have been suggested (Birks and Gordon, 1985). These are capable of identifying both major and minor zone boundaries (i.e., of defining zones and subzones) and permit objective comparisons to be made between sites. Objective computer-based zonation can also be applied to other variables in a sedimentary sequence (e.g., macrofossils, diatoms, sediment characteristics) to shed further light on the major climate-related features of the stratigraphic record (Birks, 1978; Birks and Birks, 1980). If similar local pollen assemblage zones can be identified over a large geographical area, it may be possible to define regional pollen assemblage zones, reflecting regional vegetation changes of broad paleoclimatic significance (Gordon and Birks, 1974; Birks and Berglund, 1979). The process of identifying such large-scale changes is effectively that of applying a lowpass filter to the spatial pattern of temporal change (T. Webb, personal communication). Only the major patterns of regional significance survive such scrutiny.
9.5 MAPPING VEGETATION CHANGE: ISOPOLLS AND ISOCHRONES
Studies of modern pollen rain and the distribution of different taxa in the landscape today (see Section 9.2.3) indicate that there is a fairly good spatial correspondence between them. Maps of modern pollen data can reproduce broad-scale patterns of individual taxa over large areas (Davis and Webb, 1975; Webb and McAndrews, 1976; Delcourt et al., 1984; Huntley and Birks, 1983). Studies such as these paved the way for synoptic mapping of vegetation distribution at discrete periods in the past. This approach was first suggested by Szafer (1935), who defined the term isopoll as lines of equal percentage representation of a particular pollen type in the pollen sum. Using isopolls, Szafer constructed maps showing the distribution of beech and spruce across East Germany and Poland at five intervals from late-glacial to late Holocene time. However, Szafer's time framework was speculative because, at that time, there was no means of accurately dating organic material. It was only with the extensive use of 14C dating that identical stratigraphic horizons could be identified (by interpolation between dates) at sites over large geographical areas, thus facilitating the preparation of time-sequential maps, such as those compiled by Huntley and Birks (1983), Delcourt et al. (1984), and Jacobsen et al. (1987). As an example, Fig. 9.9 shows isopolls of spruce (Picea), pine (Pinus), and oak (Quercus) over eastern North America at intervals from 18 ka B.P. to the present (T. Webb et al., 1993b). These maps indicate the vegetation response to warming, and to the retreat of the Laurentide ice sheet, with rapid migration of individual taxa in the 12-9 ka B.P. period. By 9 ka B.P., the principal features of the modern pollen distribution (0 ka) can be seen for the first time, and by 6 ka B.P. most taxa had achieved their northernmost postglacial limit. Note, however, that spruce pollen percentages show a small southward shift after 6 ka B.P., possibly reflecting a change to higher levels of available soil moisture in the northeastern U.S. (which favors spruce over pine) after the early Holocene (R. S. Webb et al., 1993).
Isopoll maps of individual taxa implicitly acknowledge that vegetation formations as we now know them have been impermanent features of the landscape (as discussed in Section 9.3). This is clearly illustrated by Huntley (1990b), who showed maps of "vegetation units," identified by associations of different pollen taxa at times in the past. Many of the vegetation units typical of western Europe today did not exist in Europe before the early Holocene and, indeed, even as recently as 1 ka B.P. there were areas of vegetation with no analog in the modern vegetation formations of Europe (Huntley, 1990a, 1990b).
Isopoll maps can be reinterpreted to show the migration of a particular genus or ecotone through time by isochrones (equal time lines). Figure 9.10a, for example, shows the location of the 15% spruce isopoll at intervals from 11,500 to 8000 yr B.P., a line thought to approximate the southern boundary of the late-glacial boreal forest in this area. Similarly, Fig.-9.10b shows the position of the conifer-
FIGURE 9.9 Isopoll maps of observed pollen data at 3 ka year intervals, from 18 ka B.P. (left-most column) to the present (at right).Three levels of shading indicate pollen percentages >1% (lightest), >5%, and >20% (darkest) (Webb et a/., 1987).
hardwood/deciduous forest ecotone derived by analogy with modern pollen, which indicates that this boundary coincides with the 20-30% isopolls for oak and pine. Both maps indicate a rapid northward migration of forests in late-glacial times. A final map, Fig. 9.10c, illustrates the position of the "prairie border," the grassland/ forest ecotone of midwestern North America based on 30% isopolls for herb pollen. Following rapid eastward shift in the early Holocene, this boundary regressed westwards after ~7000 yr B.P., indicating that the period of minimum precipitation and maximum warmth in the area had already passed.
Isopoll and isochrone maps depicting former vegetation patterns provide a qualitative perspective on past climate because vegetation formations on a broad scale are clearly determined by climate (Bryson and Wendland, 1967; Prentice et al., 1992). However, more quantitative reconstructions of past climate can be obtained by mathematically relating modern climatic conditions to modern pollen rain, and using these relationships to convert the fossil pollen record into specific paleoclimate estimates.
PRAIRIE " BORDER
FIGURE 9.10 Isochrones (in thousands of years) on the migration of various vegetation taxa or ecotones, based on the position of diagnostic isopolls considered to be characteristic of the vegetation boundary. For example, in the modern boreal forest isopolls of spruce exceed 15%. In (a) isochrones indicate the position of the 15% spruce isopoll at different times and are considered to reflect the southern margin of the boreal forest as it migrated northward. In (b) the conifer-hardwood/deciduous forest ecotone is identified by the 20% isopoll for pine and the 30% isopoll for oak, reflecting a change from dominance of oak to dominance of pine (in a northward direction). In (c) the "prairie border" is delimited by the 30% isopoll for herbaceous pollen. Shading indicates the area across which the Prairie border first expanded (to 7000 yr BP) then retreated (westward) as conditions became more moist in the region in the mid- to late Holocene (Bernabo and Webb, 1977).
PRAIRIE " BORDER
Was this article helpful?