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

Ecology, Oct, 1998 by Janice L. Fuller

INTRODUCTION

Hemlock (Tsuga canadensis) was an abundant tree species in the mixed forests that covered eastern North America in the mid-Holocene (Gaudreau and Webb 1985, Ritchie 1987). Paleoecological evidence suggests that T. canadensis declined dramatically in abundance throughout this region [approximately]5000 yr before present (Davis 1978, 1981, Webb 1982). This decline was rapid, more or less synchronous throughout eastern North America, and unique to T. canadensis (Davis 1981, Webb 1982, Allison et al. 1986). It is therefore considered to have been the result of a pathogenic outbreak (Davis 1981, Allison et al. 1986), possibly by an insect pest (Bhiry and Filion 1996). In many locations, T. canadensis populations recovered 1000-2000 yr after the decline, whereas in others it remained a sparse member of the forest flora (Davis 1978, 1981).

Several studies have focused on the cause of the mid-Holocene hemlock decline (Davis 1981, Allison et al. 1986, Filion and Quinty 1993, Bhiry and Filion 1996), but few have examined the ecological or ecosystem-level consequences. Davis (1978, 1981) notes that Betula, Fagus grandifolia, Acer saccharum, and/or Quercus increased in abundance after the decline at a number of sites in the Northeast. Hall and Smol (1993) examined communities of diatoms and chrysophytes preserved in lake sediments at a number of sites in southern Ontario to determine whether they responded to changes in the watershed vegetation cover during the hemlock decline. They found shifts in algal communities, but only in lakes with large watersheds did they observe a change in lake trophic status, with a short-lived period of eutrophication. The present study examines the impact of the hemlock decline on forest dynamics by investigating the response of other forest taxa using fine-resolution pollen data from two sites in southern Ontario, Canada.

T. canadensis is a long-lived, slow-growing, and highly shade-tolerant conifer, which, along with F. grandifolia, A. saccharum, and Betula alleghaniensis, is a major component of mature forests throughout the Great Lakes-St. Lawrence and Acadian forest regions (Rowe 1977, Ritchie 1987, Godman and Lancaster 1990). T. canadensis is generally found today in regions with cool, humid climates, and it grows on a wide variety of soils that are characterized as moist to very moist but with good drainage (Godman and Lancaster 1990). Its modern distribution is largely a function, at least locally, of recent history and human activity, as it is sensitive to disturbance and tends to be most abundant on sites least impacted by land use (Rogers 1978). T. canadensis casts deep shade and produces a thick litter layer, which often restricts the understory vegetation to T. canadensis seedlings and saplings (Rogers 1978, Benzinger 1994). Its removal from the forest canopy is, therefore, expected to result in a response from other tree and shrub species, and to influence ecosystem properties.

It is important to understand the short- and long-term impacts of pests and disease on forest dynamics for conservation and forest management purposes. The hemlock decline provides an opportunity to examine the long-term impact on forest dynamics of a species-specific, pathogenic outbreak. The chestnut decline, which occurred in North America in the early 1900s, due to the fungus Cryphonectaria parasitica, and the spread of Dutch elm disease (due to another fungus, Ceratocytis ulmi) in Europe and North America, demonstrated the potentially devastating impact of introduced pathogens (Allison et al. 1986, Whitney 1993). Currently T. canadensis populations at the southern end of their range are being attacked by the woolly adelgid (Adelges tsugae), an introduced insect whose larvae can kill trees through defoliation (McClure 1987). This pest is spreading rapidly and could decimate T. canadensis populations in North America (Orwig and Foster 1998).

The hemlock decline provides a unique opportunity to investigate long-term forest dynamics in response to a major disturbance event, such as the removal of a single, dominant tree species. The long generation time of most tree species prevents empirical studies of forest succession or tree population dynamics that cover several generations (Bennett 1983, Chen 1986). However, a paleoecological approach using fine-resolution fossil pollen analysis can provide records of forest dynamics that span thousands of years (several generations of trees) with a temporal resolution of decades (within the life-span of most tree species). In this paper I address the following questions: What is the impact on forest dynamics of the removal of an abundant tree species? At what rate does vegetation respond to such a disturbance? Are there long-term implications for forest composition of such an historical event?

Study area

The study area is southern Ontario, Canada, the forests of which are mainly classified within the Great Lakes-St. Lawrence Forest Region (Halliday 1937, Rowe 1977), and are composed of a mosaic of conifer-dominated bogs and swamps, pine plains, and upland hardwoods (Braun 1950, Rowe 1977). The two study sites [ILLUSTRATION FOR FIGURE 1 OMITTED] occur within a subdivision of this forest region, the Middle Ottawa Forest Section (Rowe 1977). This is an upland forest type, the main components of which include Acer saccharum, Fagus grandifolia, Betula alleghaniensis, Acer rubrum, Tsuga canadensis, Pinus strobus, and Pinus resinosa (Rowe 1977). Picea [TABULAR DATA FOR TABLE 1 OMITTED] glauca, Abies balsamea, Populus tremuloides, Betula papyrifera, Quercus rubra, and Tilia americana occur throughout in varying amounts. Hardwood and mixed wood swamps are common, composed of Thuja occidentalis, Juniperus communis, Picea mariana, Fraxinus nigra, Acer rubrum, and Ulmus americana (Rowe 1977). Vascular plant nomenclature follows Gleason and Cronquist (1991).