Population dynamics of hayscented fern ({\it Dennstaedtia punctilobula\/}) and its impact on forest composition, structure and dynamics

Hayscented fern abundance has increased over the last 50-80 years in forest understories of the northeastern United States. To understand hayscented fern\u27s role in forest communities, particularly with observed tree regeneration failure, I quantified the distribution, population dynamics, rhizome growth, spore dispersal, and adverse effects on tree seedlings. I developed integrated models of canopy and fern dynamics to examine the fern\u27s impact on forest communities.^ Fern abundance increased with light availability, but was unrelated to soil moisture availability in the understory. Yearly changes in abundance were low for established clones. Increased or decreased shading above the fern resulted in decreased and increased fern abundance, respectively.^ Rhizome growth rates, although slow, can be predicted from frond size; rhizomes of ramets with larger fronds spread more rapidly. Ramets with large fronds also produced spores and epipetiolar buds more frequently. Spores can be dispersed as far as 25 m. Path analysis indicated direct relationships of frond density, spore production, and epipetiolar bud production to light, whereas rhizome growth related directly to frond size.^ Black cherry and red maple seedlings were the most abundant beneath fern. However, fern cover resulted in reduced numbers, growth, and survival of tree seedlings. Though growth rates were sufficient to predict tree seedling growth through the fern canopy, none did during this study. Light available to tree seedlings was directly affected by fern, decreasing as fern abundance increased.^ I developed both static and dynamic models integrating canopy and fern dynamics; these models were calibrated and incorporated into the SORTIE forest simulator. The static model predicted dramatic changes in forest structure, whereas the dynamic model predicted subtle changes in composition and structure. Although the static model predictions conformed better to empirical studies and field observations, neither model predicted observed patterns precisely. Thus, considering direct effects of fern on tree seedlings alone can not explain the patterns of failed or delayed regeneration observed in the northeastern United States.