Presentation Title
Responses of Small Mammal Communities to Local and Landscape Variables in a Forest/Agriculture Mosaic
Abstract
How do local and landscape processes affect community assembly? The relative contribution of these two spatial scales has been investigated in several taxa, but most authors analyze only the most common species in the community and fail to account for bias introduced by detection error. Hierarchical abundance models address both of these problems. They correct for detection error by leveraging information from across the community to estimate species-specific abundances and detection probabilities. Using small mammal communities in forests, old fields, and active fields in Vermont as a model system, I will construct a hierarchical abundance model to 1) estimate site-level mammal abundances while correcting for detection error, 2) use estimated abundances generated by the model to determine regional and site-level species richness and diversity, and 3) determine the environmental covariates that are most highly correlated with these diversity measures. I predict that covariates associated with the local environment (e.g. canopy cover) will explain more variation in community structure than covariates associated with the landscape (e.g. distance to nearest road). Preliminary tests of the hierarchical abundance model with simulated data demonstrate that the model produces accurate estimates of small mammal abundance, diversity, and covariate significance at mid to high detection probabilities. Model estimates are less accurate at low detection probabilities, but are still an improvement over uncorrected data. These early findings suggest that the modeling framework will successfully reduce the bias in detection error and better estimate the relative contribution of local and landscape processes to small mammal communities.
Primary Faculty Mentor Name
Nicholas Gotelli
Status
Graduate
Student College
College of Arts and Sciences
Program/Major
Biology
Primary Research Category
Biological Sciences
Responses of Small Mammal Communities to Local and Landscape Variables in a Forest/Agriculture Mosaic
How do local and landscape processes affect community assembly? The relative contribution of these two spatial scales has been investigated in several taxa, but most authors analyze only the most common species in the community and fail to account for bias introduced by detection error. Hierarchical abundance models address both of these problems. They correct for detection error by leveraging information from across the community to estimate species-specific abundances and detection probabilities. Using small mammal communities in forests, old fields, and active fields in Vermont as a model system, I will construct a hierarchical abundance model to 1) estimate site-level mammal abundances while correcting for detection error, 2) use estimated abundances generated by the model to determine regional and site-level species richness and diversity, and 3) determine the environmental covariates that are most highly correlated with these diversity measures. I predict that covariates associated with the local environment (e.g. canopy cover) will explain more variation in community structure than covariates associated with the landscape (e.g. distance to nearest road). Preliminary tests of the hierarchical abundance model with simulated data demonstrate that the model produces accurate estimates of small mammal abundance, diversity, and covariate significance at mid to high detection probabilities. Model estimates are less accurate at low detection probabilities, but are still an improvement over uncorrected data. These early findings suggest that the modeling framework will successfully reduce the bias in detection error and better estimate the relative contribution of local and landscape processes to small mammal communities.