UC Davis

Land-use change currently constitutes the primary driver of modern biodiversity loss. One way that land conversion to agriculture can lead to biodiversity loss is through biotic homogenization where agriculture consistently favors the same species, causing communities to converge in species composition over space or time. Biotic homogenization could also occur phylogenetically, where species from the similar phylogenetic lineages are consistently favored by agriculture. However, whether phylogenetic homogenization occurs across multiple scales and in tropical regions is rarely investigated. In Chapter 1, I leveraged species’ phylogeny and traits to understand how land-use change affects tropical bird communities within Costa Rica and Colombia. We conducted bird surveys across Costa Rica and Colombia in 294 sites that varied in levels of local forest cover and annual precipitation. We then analyzed sites in Costa Rica and Colombia together to find that land conversion from forest to agriculture consistently favors the same types of species at a large scale. While most trait-based studies measure effects of global changes on species occurrences or abundances, changes in species abundance or occurrence provide little information about how species actually use the habitats where they occur. For example, a species detected in agriculture could simply be passing through (and not actually resilient to land-use change) or it could be using agriculture to forage, reproduce, or otherwise complete its lifecycle (meaning the species is resilient to land-use change). In Chapter 2, I developed and then validated a statistical model to analyze how species’ behavior differs between habitat types using observational data. To our knowledge, this is the first model to measure how animals use different environments while accounting for behavior-specific imperfect detection probability. We compared the model to an alternative model that did not account for imperfect detection probability and found that our new model produced more accurate estimates of the mean and uncertainty of the effects of environmental covariates on behaviors. Thus, this model allows researchers and managers to more accurately assess how species use the environments they are in using observational behavior data. In Chapter 3, I applied this model to over 14,000 behavioral observations across 55 bird species in Northwest Costa Rica to understand how local forest cover and forest protection level affect bird behaviors. We found that birds were more likely to perform reproductive behaviors in protected forest than private forest and agriculture, meaning that protected areas are likely important for bird reproduction and conservation. This chapter also demonstrates the importance of behavioral analyses in conservation planning, as species’ behavioral responses were not always consistent with their differences in abundance. Beyond measuring changes in occurrence and behavior, it is also critical to understand if habitats can support sufficient reproductive rates to allow for population persistence. Tropical cavity-nesting birds may be particularly threatened by agricultural land-use change, which removes the tree they rely on for nesting and roosting. In Chapter 4, I aimed to understand how land-use change affects tropical cavity-nesting bird reproduction by measuring nest site availability and conducting a nest-box addition experiment in forest and agriculture Northwest Ecuador. I found that land-use change likely limits the reproduction of cavity-nesting birds in agriculture, as nest boxes in agriculture had much higher levels of avian activity than nest boxes in forest. Nest boxes also had a relatively high rate of chicks fledged, meaning that nest boxes could be a successful conservation strategy for certain species. Overall, my dissertation explores the different effects of agricultural land-use change on tropical bird occurrence, behavior, and reproduction. I found that these effects can differ from each other and across scales. The results highlight the importance of forest conservation, especially in wet regions, and support investigating species’ responses beyond occurrence when comparing the conservation value of habitat types.