Ecological impact of the mid-Holocene hemlock decline in southern Ontario, Canada

Ecology, Oct, 1998 by Janice L. Fuller

High Lake. - Rates of change increased after the hemlock decline at High Lake, but the values were lower than those estimated for Graham Lake. Pollen percentages and influx values of T. canadensis did not decline as markedly at High Lake. Ulmus appears to have been the main pollen taxon that increased exponentially after the decline. The increase of other forest taxa was not highly significant and/or explained well by the exponential equation. Ulmus, Betula, F. grandifolia, and A. saccharum were the main tree taxa to increase.

PCA shows the pattern of vegetation change during the hemlock decline at this site [ILLUSTRATION FOR FIGURE 8B OMITTED]. There was a shift in forest composition after the decline of hemlock; however, unlike the pattern observed at Graham Lake, once T. canadensis populations recovered, the pollen assemblage was not markedly different from that present prior to the decline. There is considerable overlap between the pre- and postdecline samples.

DISCUSSION

Nature of the hemlock decline

Paleoecological data from numerous sites in northeastern North America record the rapid decline of Tsuga canadensis in the mid-Holocene (Davis 1981, Webb 1982, Ritchie 1987). At Graham Lake there appears to have been a double decline in the abundance of T. canadensis during the mid-Holocene. Pollen influx values dropped sharply [approximately]6000 yr BP, recovered briefly, and then declined further at 5500 yr BP [ILLUSTRATION FOR FIGURE 5 OMITTED]. T. canadensis remained at low levels for [approximately]1500 yr before recovering to reach former abundances. The apparent double hemlock decline recorded at Graham Lake may reflect an initial outbreak of a pathogen, followed by recovery, and then a further wave of infestation, which wiped out the already weakened population. Bhiry and Filion (1996) found in situ hemlock macrofossils and insect remains in a paludified dunefield in Quebec. Their evidence suggests that two defoliation events occurred during the mid-Holocene hemlock decline at the site in Quebec, which may also have been the case at other sites, including Graham Lake, depending on the pathogen dynamics and resistance of the hemlock populations.

At High Lake, pollen data suggest that T. canadensis populations declined in the mid-Holocene, although perhaps not as dramatically as recorded at Graham Lake [ILLUSTRATION FOR FIGURE 5 OMITTED] or other sites in the region (McAndrews 1981, Ritchie 1987). Heterogeneous dispersal, distribution, [TABULAR DATA FOR TABLE 3 OMITTED] and impact of the pathogen that caused the decline (assuming this was the cause) may have resulted in variation with respect to mortality. There also appears to have been a double decline in T. canadensis pollen abundances at this site [ILLUSTRATION FOR FIGURE 5 OMITTED].

Vegetation response to the hemlock decline

The fine temporal resolution of the pollen influx and pollen percentage data from Graham Lake and High Lake provide detailed records of vegetation change during the mid-Holocene. It is advantageous to examine both pollen influx and pollen percentage data when investigating forest dynamics, such as those during the hemlock decline, to obtain data that are not dependent on other taxa (as percentages are) and to separate out the effects of changes in sedimentation rates (which affect pollen influx values).