Past Statistical Ecology Projects
Multi-Scale Species Distribution Models for Optimal Prediction of Habitat Suitability over Large ExtentsThis project evaluated methods for merging scale-optimization techniques from spatial ecology with predictive model selection tools from statistical ecology in order to develop optimally predictive multi-scale models of distributions for cryptic and rare species. We evaluated a number of tools for this purpose, including continuous and discrete Bayesian model selection, statistical regularization via lasso regression, and assessment of model transferability using spatially-structured cross validation.
Learn more at: |
Hierarchical Models for Imperfectly-Observed RemovalsThis project developed a framework for estimating animal abundance and population growth using data from replicated removal experiments that are imperfectly observed, where removals and sampling efforts are estimated but their true values are unknown. This includes hierarchical models that allow for either constant or variable-effort removal sampling, spatial-temporal variation in catchability and animal density, and estimation of site-specific abundances and population growth rates.
Learn more at: |
Stochastic Modeling of Wildlife-Infrastructure Collision Data |
This project developed a likelihood-based framework for joint analysis of data from infrastructure-collision surveys and carcass removal experiments. We used a stochastic open population model to describe the underlying processes of infrastructure collision and carcass removal, and provided a framework for modeling collision and removal rate parameters as a function of covariates using generalized regression.
Learn more at: |