Conventional gap-phase models of temperate forests have been interpreted to suggest equilibrial long-term dynamics and maintenance of diversity through non-transitive replacement patterns in late-successional, old-growth forests. Most analyses of such dynamics use assumption-laden time-for-space substitutions. I use data from two sets of multi-decade (47 to 74 years) permanent plots in old-growth hemlock-northern hward forests in northern Michigan to assess the adequacy of gap-phase models for explaining compositional patterns in both canopy and understory. I ask, specifically, whether a) prediction of composition and diversity patterns requires specific historical knowledge beyond instantaneous data on current dynamics (i.e., does a uniformitarian principle apply?), and b) if community properties are strongly historically contingent, does this affect anticipated consequences of climate change for community pattern?
Long-term data on tree demography and canopy composition (some previously published), including initial responses to a severe blow-down, indicate that rare events play an important role in shaping community composition and maintaining diversity, even after several centuries. Without such events, alpha (plot-scale) diversity declines, likely due to competitive sorting. Some species (notably, Betula alleghaniensis) tend to be lost, and community pattern becomes more distinctly organized around edaphic patterns. Regeneration and recruitment of both shade-tolerant and shade-intolerant trees may be intimately linked to rare events and can't be fully explained by typical gap-phase dynamics.
Most of the diversity of these forests resides in the understory and new analyses of long-term data-sets show declining alpha-diversity and changing compositional patterns in understory communities that may also reflect time since major disturbance; patterns of changes are not consistent with effects of deer browse or invasive species. Initial analyses of trends in living biomass and in coarse woody-debris inputs indicate that rare events can also have consequences for biogeochemical dynamics at local scales.
Most efforts to predict vegetation changes due to anticipated climate changes have focused on regional species distributions with respect to general temperature and moisture envelopes. This and other studies suggest that changes in frequency and type of severe weather events may have significant consequences for local vegetation pattern and composition independent of effects of gross climatic trends and range shifts. Potential restructuring of community organization due to changed disturbance regimes should be taken into consideration in vegetation classification efforts.