Home WakeSpace Scholarship › Electronic Theses and Dissertations

A Dynamic Occupancy Model for Multiple Species in Serengeti National Park

Electronic Theses and Dissertations

Item Files

Item Details

A Dynamic Occupancy Model for Multiple Species in Serengeti National Park
Zhang, Xinyu
Occupancy models for different species have been used by researchers to estimatespecies’ migratory trends and to understand how the environmental and human fac-tors are affecting it. Developed from the original single species occupancy model [34],recent approaches attempt to consider multiple interacting species. However, mostof these approaches ignore the dynamicity of the migratory process. Recently devel-oped research methods about accounting for dynamicity into occupancy models aremainly divided into two directions: The first use a Bayesian hierarchical model and acentered auto-logistic regression model [23] to add in temporal dependence, and thesecond builds a probabilistic model about occupancy status within time adjacency[7]. This thesis outlines a framework to study how cross-species influences vary withrespect to environmental elements and makes some improvement based on the exist-ing methods of Rota et al. [42]. We employ the multivariate Bernoulli distribution[11] to allow the species correlations to vary over spatial and temporal scales, andwe considered dynamicity by breaking the occupancy probability into disjoint migra-tory processes. This paper employs reversible jump Markov chain Monte Carlo [19]to implement model selection. We apply our model to three pairs of herbivores anduse WAIC (Watanabe-Akaike Information Criterion) [59] to compare the estimationresults with the multiple species model without considering within or cross speciesdependence and single species occupancy models.
Dynamic Occupancy Model
Multivariate Bernoulli Distribution
Reversible Jump MCMC algorithm
Watanabe-Akaike Information Criterion
Hepler, Ph.D., Staci A (committee chair)
Erhardt, Ph.D., Robert J (committee member)
Moore, Ph.D., William F (committee member)
2020-05-29T08:36:07Z (accessioned)
2021-05-28T08:30:11Z (available)
2020 (issued)
Mathematics and Statistics (discipline)
2021-05-28 (terms)
http://hdl.handle.net/10339/96843 (uri)
en (iso)
Wake Forest University

Usage Statistics