Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)



Committee Chair(s)

James W. Haefner


James W. Haefner


Eugene W. Schupp


James A. MacMahon


John M. Stark


James A. Powell


Tree community composition in bottomland hardwood reforestation sites is considerably less diverse than natural bottomland hardwood forests. This study sought to understand the mechanisms behind failure to develop diverse tree communities. First, I developed a mechanistic model of seed dispersal by wind in spatially variable environments. Second, I developed a spatially explicit simulation model of forest dynamics that includes this wind-dispersal model to investigate whether diversity is limited by dispersal or interactions among species and individuals. Finally, I performed model experiments to determine if manipulations of stand structure might help improve conditions for colonization of various species, thus enhancing diversity of reforestation sites. The wind dispersal model was unbiased and accurate for predicting seed dispersal patterns of four species of wind-dispersed trees, demonstrating the utility of my algorithm for making predictions of seed arrival in a forest simulation model. The forest simulation model accurately predicted basal area growth and general patterns of species relative abundance in natural and reforested bottomland hardwoods, and predicted that reforestation sites will probably never attain diversity levels of natural forests under the current management scenario. Development of diversity was hindered by competition from the species planted and limited dispersal from forests. Hence, the only reasonably successful option to enhance diversity is probably to establish sites with mixed-species plantings at the outset. However, if stands are thinned at relatively young ages (15 yr for acorn-established stands, 25 yr for seedling-established stands), before canopy closure from planted individuals results in mortality of colonizing individuals, diversity may be enhanced if adequate numbers of colonizers are able to disperse to the site. Further research is necessary on mechanistic dispersal by animals, transition rates from seeds to seedlings, and the factors that affect such transitions in order to more accurately predict forest community development.