Spatial patterns of litterfall, nitrogen cycling, and understory vegetation in a hemlock-hardwood forest

The mosaic of old-growth hemlock and hardwood patches at Sylvania Wilderness Area, Michigan, provides an excellent opportunity to study the effects of different tree species on the underlying vegetation and nitrogen (N) cycling regimes. Understory vegetation was surveyed and sugar maple seedling growth rates were determined in 50 randomly located 4 m{dollar}sp2{dollar} subplots on a 7.2 ha mapped forest plot. Adjacent to the subplots, litterfall and seeds were collected in traps, and N cycling rates were measured by in situ incubations; light and various soil characteristics were also measured. To quantify dispersal trajectories, leaves and seeds were collected beneath seven trees within the forest but isolated from conspecifics.; Leaf litter N chemistry affects soil N mineralization rates at a fine spatial scale. Litter C:N ratios were negatively correlated with N cycling rates, especially nitrification: rates were lowest beneath hemlock trees and highest under basswood and sugar maple trees. Maximum likelihood techniques were used to parameterize a spatial model of leaf fall, which can predict N inputs to the soil. The model illustrates the importance of individual trees to soil processes and unites population biology and ecosystem ecology.; Sugar maple seedlings dominated the understory vegetation but had lower cover, density, and growth rates beneath hemlocks, where soils were acid and N-poor and where it was dark in spring, when maple seedlings accomplish most of their annual growth. Herb species composition differed under hemlocks and hardwoods. Total herb cover was negatively correlated with sugar maple seedling cover, perhaps due to competition. Seed rain was correlated with distance-weighted basal area within 20 m for all tree species on the mapped plot. Seedling density was correlated with seed number for several tree species, resulting in correlations of seedlings with conspecific canopy trees. Thus the Sylvania canopy mosaic may be maintained both by seed dispersal limitations and by leaf litter feedbacks on N cycling rates, showing that processes around individual trees may affect landscape-level patterns. Spatial patterns of trees should be exploited in studies of ecosystems or population phenomena such as seed dispersal in order to better understand forest function.