Dynamics In Late-Successional Hemlock-Hardwood Forests Over Three Decades

Ecology, Jan, 2000 by Kerry D. Woods

Natural Sciences, Bennington College, Bennington, Vermont 05201 USA

Abstract. Permanent plots in old-growth hemlock-northern hardwood forests of Michigan's upper peninsula have been remeasured over periods of 16-32 yr. A gradient from hemlock (Tsuga canadensis) to sugar maple (Acer saccharum) dominance is associated with increasing soil pH and calcium. Secondary species include yellow birch (Betula alleghaniensis) and basswood (Tilia americana). From 1962 to 1994 hemlock increased in basal area and dominance in most plots. Sugar maple showed little overall change, while basswood and especially yellow birch showed sporadic but often large declines in basal area. Birch populations declined due to lack of recruitment, and sugar maple and basswood may be subject to similar decline; only hemlock showed a fairly stable size structure. Mortality rates were lowest for hemlock (0.3%/yr) and highest for yellow birch (1.6%/yr), corresponding to canopy residence times of 357 and 61 yr, respectively. Stem maps allowed assessment of neighborhood influences on growth and mortality. Growth and mortality rates were negatively correlated for all species. Growth rate was influenced by tree size and site conditions for all species, but hemlock and sugar maple growth rates were also affected by size- and distance-weighted indices of neighbor influence. Old-growth stands several centuries old continue to undergo compositional change related to both stand history and current population interactions. Yellow birch and basswood are probably maintained by significant disturbances and will decline under a disturbance regime of small gaps. Hemlock may be the ultimate competitive dominant in most sites but may require well over a millennium without major disturbance to displace sugar maple.

Key words: Acer saccharum; Betula alleghaniensis; canopy dynamics; competition; hemlock-- northern hardwood forest; long-term studies; old-growth forest; permanent plots; succession; tree demography; Tsuga canadensis.

INTRODUCTION

Much of our understanding of late-successional forests is based on untested anecdotes and conceptual models and lacks strong empirical support. Because tree generation times exceed a century, direct measurement of community dynamics is daunting, while indirect approaches are hampered by assumptions about environmental history (chronosequence studies and inferences from current stand patterns) or limits on resolution, precision, or generalizability of data (paleoecology and historical reconstruction) (Davis 1989, Pickett 1989). New generations of simulation models (e.g., Pacala et al. 1993, 1996) may overcome some of these limitations, but are difficult to apply retrospectively. These limitations threaten the effectiveness of conservation management focused on old-growth forests. Ultimately, direct, long-term observation is critical in testing and extending hypotheses about properties of late-successional communities (Bakker et al. 1996), but few studies of late-successional forests have been maintained over more than a decade (but see Filip et al. 1960, Leak 1970, Hough 1965, Whitney 1984, Peet 1984).

I used data from a 32-yr study of permanent plots in old-growth hemlock-northern hardwood stands in northern Michigan to interpret community and population dynamics and to assess some frequent assumptions and expectations of succession theory. Succession models often suggest that with the passage of time following stand initiation, stand composition will become more stable, community and population properties will become more fully controlled by processes and interactions internal to the stand, and stand structure and composition will become less directly attributable to particular historical events. Do late-successional stands, in fact, increase in stability? If so, what is the relevant time frame? If more than two or three generations of canopy dominants are required, infrequent major disturbances may make stable "old-growth" stands relatively rare. Do structure and composition of late-successional stands become substantially independent of the direct effects of massive disturbance (as indicated by lack of even-aged cohorts and light-demanding species), becoming increasingly shaped by processes internal to the stand?

Changes in basal area and size structure, mortality and establishment rates, and the effects of local competition on growth and mortality over three decades all suggest that successional change continues in these old-growth forests. Current composition reflects probable disturbances several centuries past, but patterns in growth and mortality are also shaped by current competitive interactions. Trends suggest an ultimate, hemlock-dominated, and potentially stable community, but only after several more centuries without major disturbance.

METHODS

Site description and field methods

The Huron Mountains, on the central south shore of Lake Superior in Michigan, USA (Fig. 1), display local relief of [tilde]270 m above mean lake surface of 182 m. Bedrock is Pre-Cambrian granite, gneiss, and sandstones (Dorr and Eschmann 1970). Soils are generally thin and acidic, derived from rocky glacial till, with small areas of deeper till and alluvium. Research sites are located within a 2500-ha reserve in mesic forests dominated by hemlock (Tsuga canadensis) and sugar maple (Acer saccharum), with yellow birch (Betula alleghaniensis) and basswood (Tilia americana). There has been no logging within the reserve.