New models for old growth from long-term studies in eastern deciduous forests
Management agendas for eastern eastern temperate forests are often conditioned on poorly tested ideas about baseline or 'natural' dynamics. Clements' climax concept, reflected in E.L. Braun's classifications of 'original' forests, has been substantially modified, but underlying assumptions of gap-phase-driven tendencies in late succession towards compositional and functional equilibrium remain influential. Alternatively, structure and function may be historically contingent and non-equilibrial, reflecting long-persistent effects of rare events. It is difficult to compare the merit of these two notions for slow systems where processes play out over decades to centuries; researchers are generally dependent on indirect and assumption-laden approaches like space-for-time substitution. I use data from remeasurements of ~150 permanent plots over 50-80 yrs in old-growth hemlock-hardwood forests of northern MI, USA, to analyze demographic and ecosystem properties. I ask, specifically, whether patterns of community assembly and change, demographic properties and rates, and ecosystem properties are more consistent with ahistorical, equilibribial models informing Braun's maps and more recent thought, or with historically contingent, non-equilibrial models. I further explore how results might inform modern attempts at classification and approaches to conservation management and restoration.
Demographic and compositional trends for these stands support a view of late-successional forests as non-equilibrial and historically contingent, understandable only through an accounting of infrequent disturbances. In 'gap-phase'-dominated intervals between events, plot diversity declined as dominance of competitive species increased. Similar trait-driven dynamics appear to operate, at different spatial scales, for little-studied understory communities. Mortality rates for less competitive species (e.g., Betula alleghaniensis) increased over time and varied spatially, suggesting cohort-structured populations shaped by rare events; background mortality rates for competitive dominants (Tsuga canadensis, Fagus grandifolia) changed little. High biomass densities (250-400 Mg/ha) increased (~0.5%/yr) during most of the study period with increasing dominance of large individuals, but a single large windthrow balanced increases over the previous 65 yr resulting in no overall net change in above-ground biomass due, not to stand-scale equilibrium, but chance interactions between demographic trends and unpredictable rare events. These results raise questions for conservation management where targets include restoration of 'old-growth properties' based on associations defined (as Braun's) by current conditions and distributions. Further, it seems likely that the most immediate effects of climate change might be driven, not by tolerance-related distributional changes, but by changing disturbance frequencies.