Browsing by Author "Poulsen, John Randolph"

Understanding how tropical forests respond to changes in the abiotic environment and human disturbance is critical for preserving biodiversity, mitigating climate change, and maintaining ecosystem services in the coming century. The lowland rainforests of Central Africa in particular are expected to lose 41% of present dense forest cover in the next 50 years to forest clearing, due in large part to forest loss resulting from the expansion of subsistence agriculture and logging. Deforestation also contributes a net increase in carbon dioxide to the atmosphere, exacerbating forest losses via increased tree mortality from drought, fire, and dispersal failure. Despite these grim circumstances, we know little about how Paleotropical tree communities are likely to respond to predicted changes in disturbance and climate.

To evaluate the unique response of Afrotropical forests to changes in the abiotic environment and disturbance, I employ diverse data including species inventories, seed rain, species traits, remotely sensed historic climatic data, future climate predictions, and soil nutrient data collected from 134 1-ha plots arrayed in two large-scale observational experiments spanning the central African countries of Gabon and the Republic of Congo (Brazzaville). I then combine these diverse data using novel modeling methods to 1) determine the relative roles of climate and human disturbance on tree community composition, 2) quantify the relative effects of human disturbance and the abiotic environment on tree fecundity and seed dispersal, and 3) forecast future tree community change given predicted changes in climate.

This work demonstrates that Afrotropical plant communities are more sensitive to human disturbance than to climate, with particular sensitivities to hunting and distance to village (a proxy for other human activities, including tree-cutting, gathering, etc.). These pressures have meaningful long-term effects on seed dispersal, increasing dispersal distances for animal dispersed seeds in disturbed forests. Finally, We forecast a 3 - 8% decrease in Afrotropical forest species richness by the end of the century, in contrast to the 30-50% loss of plant diversity predicted to occur with equivalent warming in the Neotropics.

This work reveals that community forecasts are not generalizable across regions, and more representative studies are needed in understudied biomes like the Afrotropics. Nascent data sets, increased availability of high quality remote sensing data, and new statistical techniques capable of synthesizing these various data will help in further resolving the fate of the world’s ecosystems. This study serves as an important counterpoint to work done in the Neotropics by providing contrasting predictions for Afrotropical forests with substantially different ecological, evolutionary, and anthropogenic histories.

Understanding how tropical forests respond to changes in the abiotic environment and human disturbance is critical for preserving biodiversity, mitigating climate change, and maintaining ecosystem services in the coming century. The lowland rainforests of Central Africa in particular are expected to lose 41% of present dense forest cover in the next 50 years to forest clearing, due in large part to forest loss resulting from the expansion of subsistence agriculture and logging. Deforestation also contributes a net increase in carbon dioxide to the atmosphere, exacerbating forest losses via increased tree mortality from drought, fire, and dispersal failure. Despite these grim circumstances, we know little about how Paleotropical tree communities are likely to respond to predicted changes in disturbance and climate.

To evaluate the unique response of Afrotropical forests to changes in the abiotic environment and disturbance, I employ diverse data including species inventories, seed rain, species traits, remotely sensed historic climatic data, future climate predictions, and soil nutrient data collected from 134 1-ha plots arrayed in two large-scale observational experiments spanning the central African countries of Gabon and the Republic of Congo (Brazzaville). I then combine these diverse data using novel modeling methods to 1) determine the relative roles of climate and human disturbance on tree community composition, 2) quantify the relative effects of human disturbance and the abiotic environment on tree fecundity and seed dispersal, and 3) forecast future tree community change given predicted changes in climate.

This work demonstrates that Afrotropical plant communities are more sensitive to human disturbance than to climate, with particular sensitivities to hunting and distance to village (a proxy for other human activities, including tree-cutting, gathering, etc.). These pressures have meaningful long-term effects on seed dispersal, increasing dispersal distances for animal dispersed seeds in disturbed forests. Finally, We forecast a 3 - 8% decrease in Afrotropical forest species richness by the end of the century, in contrast to the 30-50% loss of plant diversity predicted to occur with equivalent warming in the Neotropics.

This work reveals that community forecasts are not generalizable across regions, and more representative studies are needed in understudied biomes like the Afrotropics. Nascent data sets, increased availability of high quality remote sensing data, and new statistical techniques capable of synthesizing these various data will help in further resolving the fate of the world’s ecosystems. This study serves as an important counterpoint to work done in the Neotropics by providing contrasting predictions for Afrotropical forests with substantially different ecological, evolutionary, and anthropogenic histories.

Forest elephants (Loxodonta cyclotis), the smallest and least studied of the three extant elephant species, predominately inhabit the Guinean and Congolian tropical forests from Guinea to the Democratic Republic of Congo. Known as ecosystem engineers, forest elephants create and maintain forest habitat, shape faunal communities, transport nutrients, and disperse seeds to distant areas. Despite their essential ecological role, very little is known about forest elephant social behavior. Living in social groups provides individuals with many benefits, including information about resources, protection from predators, and access to mates. For highly social species, like elephants, understanding social behavior is crucial to implementing sustainable conservation practices and mitigating the negative impacts of human development. To date, what is known about forest elephant social behavior originates from observations in baïs – mineral rich forest clearings. As a result, our understanding is limited by the short periods of time forest elephants spend in baïs, less than 2% of their time, and the small area relative to the rest of their home range and lifespan that we are able to observe. In this dissertation, I present research from the first project to attempt to understand elephant social interactions from throughout the range of habitats that forest elephants exploit on a daily basis, including dense, closed-canopy forest. I combine genetic and satellite technologies (GPS tracking and remote sensing) using novel computational methods to address: (1) factors that influence fluctuations in forest elephant group size; (2) forest elephant group age-sex composition and the factors influencing the probability of interactions between two elephants; and (3) improvements to forest elephant monitoring via line transect surveys for dung by creating an adaptive dung decay model. I conclude that: (1) group size is variable with forest elephants displaying a fission-fusion social system – a flexible social system in which individuals or sub-groups intermittently join other groups – across habitats in response to fruit availability and human disturbance; (2) interaction between individuals is influenced by social, but not environmental, factors and forest elephants spend more time in mixed sex groups than Asian or savanna elephants; and (3) estimating dung degradation via remotely sensed imagery is a feasible, cost-efficient alternative or supplement to in-situ dung degradation studies for non-invasive population surveys. This dissertation highlights the value of untangling the complex interplay between environmental, social, and anthropogenic drivers of species group composition and social behavior to inform conservation action., the results herein will be informative for monitoring forest elephant populations and promoting human-elephant coexistence through improved management of potential conflict areas.