| Amazonia is the largest and most biodiverse tropical forest, harboring approximately 25% of the earth's terrestrial species. Deforestation and consequent forest fragmentation have reduced intact Amazonian rainforest to 80% of its original extent, threatening this biodiversity. The negative effect of forest fragmentation on adult tree diversity in Amazonian forest is well documented, however it remains unknown how forest fragmentation affects the diversity and composition of forest regeneration. The first chapter of my dissertation research examines the impact of fragment area on the diversity of tree and shrub seedlings regenerating in an experimentally-fragmented landscape in the Central Amazon near Manaus, Brazil. To do this, I demarcated 104 subplots, each 10 m2 in area in the centers of fragments (1 ha, 10 ha, 100 ha) and in continuous forest controls, tagged and identified all seedlings encountered in these subplots, and determined the effect of fragment size on their density, species richness, Shannon diversity, Fisher's alpha diversity, phylogenetic diversity, phylogenetic community structure and species accumulation.;Results indicate a pronounced loss of tree seedling density (3x), species richness (2x), Shannon's diversity, Fisher's alpha diversity, and phylogenetic diversity in small fragments (≤10 ha) compared to continuous forest, while larger fragments (100 ha) had intermediate density and diversity. Accumulated species richness was highest in continuous forest and significantly declined with fragment area, indicating that regional diversity is also negatively affected by fragmentation. Phylogenetic community structure was also significantly altered in fragments: the terminal phylogenetic clustering characteristic of continuous forest was lost in forest fragments. Since patterns of regeneration in forest fragments today reflect their potential for diversity persistence in the future, these results suggest future forest assemblages in fragmented Amazonian forests will be impoverished on both local and landscape scales, especially in small fragments, and indicate that protection of large blocks of continuous primary forest should remain a conservation priority. In areas lacking continuous forest, my results showing a gradient in biodiversity retention potential from large to small fragments suggest that conservation efforts should prioritize the protection of large fragments, while small fragments could also play a significant conservation role at the landscape level as partial biodiversity reservoirs, in maintenance of landscape connectivity, and for facilitating the future restoration of tree and shrub diversity in adjacent pasture.;To understand the temporal dynamics of seedling regeneration, including the impact of a catastrophic climate-change induced drought in 2005, my second chapter examines changes in tree, shrub and liana seedling density and diversity over three time periods, and it documents recruitment and mortality dynamics driving changes in seedling density and diversity. I tagged all tree, shrub and liana seedlings located in 68 subplots, each 10 m2 in area between 9/22/2005 -- 8/9/2006 (2005 census) and revisited the plots between 10/5/2007 -- 7/21/2008 (2007 census) and again between 5/13/2009 -- 6/21/2009 (2009 census). I tested whether fragment size (1 ha, 10 ha, 100 ha, continuous forest), time period (2005, 2007, 2009) or their interaction significantly affect seedling density, species richness, Shannon diversity and Fisher's alpha diversity, and I also determined the effect of these factors on seedling recruitment, mortality, turnover and population change over two time periods (2005 -- 2007, 2007 -- 2009).;Tree and shrub seedling diversity (species richness and Shannon diversity) fell significantly between 2005 and 2007 and remained significantly lower than 2005 in 2009, driven both by significantly higher species loss and lower recruitment of new species following the 2005 drought. Liana density and diversity, in contrast, increased in both time periods, allowing lianas to maintain a positive rate of population change because lianas had lower mortality immediately post-drought compared to tree and shrub seedlings. Comparing my results with those of published seedling density values in 1991 indicates an astounding decline in tree seedling density and increase in liana percentage over the intervening 14 years in fragments, which I hypothesize is driven by the high tree seedling mortality but unchanged liana mortality that I documented following the 2005 drought. This increasing proportion of lianas in relation to trees is considered one of the ten key fingerprints of global environmental change in neotropical forest, and my results are the first to indicate that differential mortality of tree compared to liana seedlings following drought may be a primary driver of this phenomena. Since lianas enhance adult tree mortality risk, my results suggest that conservation prescriptions trying to avert the negative effects of climate change on neotropical forest structure must include liana control in both fragments and continuous forest while simultaneously trying to promote tree seedling recruitment following drought by increasing connectivity in fragmented landscapes. |
