OLD-GROWTH NORTHERN HARDWOOD FORESTS: SPATIAL AUTOCORRELATION AND PATTERNS OF UNDERSTORY VEGETATION

Abstract

The goal of this research was to examine spatial patterns of forest understory vegetation at a fine resolution for future work on underlying processes. We used a cyclic, two-dimensional sampling design, sampling plants in 2100 quadrats (0.25 m2), placing one 62 × 29.5 m grid within each of four old-growth northern hardwood (Acer saccharum–Betula alleghaniensis) stands. The specific plan used was designed to maximize spatial information and sampling efficiency. The study was done in the Sylvania Wilderness Area, Ottawa National Forest, Upper Michigan, USA. Spatial patterns of ground-layer species vary with the environment, species ecological characteristics, and their interactions. Competition with maple saplings may be a strong determinant of understory spatial patterns of these forests. Based on 95% confidence intervals, spatial analysis showed that most ground-layer species were positively autocorrelated to distances of <2.5 m in stands with high sapling density, while many of these same species were autocorrelated at up to 21 m in stands with low sapling density. Most ground-layer species also had distributions indistinguishable from random at three other resolutions (9.0 × 10.5 m to 9.0 × 31.5 m blocks) in stands with high sapling density, but aggregated distributions in stands with low sapling density. Logistic regression analysis yielded direct and indirect negative correlations between ground-layer species and maple saplings. Plant species temporal guilds were autocorrelated according to the following rank distances: spring ephemerals > evergreen ≥ early summer > late summer > dimorphic. More species were autocorrelated to greater distances on loamy soils than on sandy soils. Plant species dispersal guilds were autocorrelated according to the following rank distances: ballistic ≥ spores ≥ ant > ingested > adhesive. Thus there is a general inverse relationship between autocorrelation distance and migration rates of dispersal guilds. The spatial distribution of microtopography and decayed, coarse woody debris appears to be important for the maintenance of plant diversity and heterogeneity in old-growth stands. Forest managers can help maintain biological diversity by giving preferential management to those species that are rare and, if once locally extirpated, have poor re-colonization ability. This requires maintaining regional landscape diversity, as well as within-stand microhabitats.