Seed dispersal in gap-dynamic forests: Effects on population and community dynamics

I examine the role of seed dispersal in determining how local competition among individuals affects population and community dynamics. Variability in seed arrivals depends on spatial variation in the location and fecundity of adults and on the combined effects of two features of dispersal: (1) dispersal distance, how far from a source seeds are likely to land, and (2) seed clumping, how "noisy" or unpredictable seed arrivals are at fine spatial scales. At coarse scales, seed dispersal averages across variation in density, while at fine scales, stochasticity in dispersal generates variability. Previous analyses of seed dispersal have not treated dispersal realistically. I develop a gap-dynamic forest simulation that permits a more complete analysis of the effects of seed dispersal on population and community dynamics. I generate theoretical results in Chapters 2 and 3. I obtain empirical estimates of recruitment parameters and test the effects of seed dispersal in a natural system in Chapter 4.; In Chapters 2 and 3, I test the hypothesis that the scaling relationship between local competition and population and community dynamics depends only on the total magnitude of variability generated by seed dispersal, and not on the nature of seed dispersal (i.e. whether variability is derived principally from local dispersal or from seed clumping). In Chapter 2, I test the effects of contrasting seed dispersal dynamics on population growth rates and on equilibrium canopy areas. I find that the source of variation does affect population growth rates but not equilibrium canopy areas. Dispersal distance and seed clumping exert unique influences on populations when dynamics are governed by processes operating at low density across broad scales. In Chapter 3, I describe the effects of seed dispersal on coexistence of competing species. I find that dispersal distance and seed clumping generate qualitatively different competitive environments. Variability derived from local dispersal often results in high correlation among species in the distribution of their seeds. When comparing competition in systems with an equal magnitude of variability, coexistence is easier when dispersal distances are long and seed clumping is high because low levels of seed clumping result in high seed rain correlations among species.; In Chapter 4, I estimate fecundity, dispersal, and seedling establishment characteristics for 11 canopy and sub-canopy taxa at the Coweeta Long-Term Ecological Research site. I test the effects of stochastic seed dispersal on competition and genetic variation. I find that the predicted magnitude of variability is minimal at scales of interaction typical of forest systems ({dollar}ge{dollar}100 m{dollar}sp2{dollar}). Moderate levels of genetic variation are maintained by stochastic dispersal, but coexistence is not obtained. The effects of large differences among taxa in recruitment traits are not sufficiently moderated by variability in local densities generated by stochastic dispersal. Model predictions of competitive dominance correlate well with field abundance of "weedy" taxa at Coweeta (Acer, Betula, and Liriodendron) but not for most other taxa (e.g. Carya and Quercus). These results suggest that the influence of stochastic seed dispersal in natural forest communities is minimal relative to the influence of differences among taxa in predicted rates of recruitment.