Authors: Juan Pablo Gomez*, University of Florida, Jose Miguel Ponciano, Department of Biology, University of Florida, Jason K Blackburn, SEER Lab, Department of Geography, Emerging Pathogens Institute, University of Florida
Topics: Quantitative Methods, Migration
Keywords: Discrete Self Decomposable Models, Immigration, Residency, Abundance Time Series
Session Type: Paper
Start / End Time: 5:20 PM / 7:00 PM
Room: Napoleon B2, Sheraton 3rd Floor
Presentation File: No File Uploaded
Time series data of animal abundances at different spatial and temporal scales inform ecologists about the drivers of fundamental population dynamics processes. Arrivals and departures to particular locales, births and deaths, invasion and extinction, have all been modeled as simple stochastic processes whose trajectory is shaped by biotic and abiotic forces. We generalize a little-known discrete-time stochastic process to model time series of counts of individuals in a location and generalize it as a spatio-temporal model with sampling error and covariate effects. The proposed approach is applicable to any scenario where the objective is to understand how the processes of arrival and departures to and from a location influence abundance variability across time. Through simulations, we demonstrate that the model has good statistical properties and performs well under a wide variety of scenarios. Although the model is developed as not spatially explicit, we show that the use of spatially autocorrelated variables can be incorporated in the model to gain insights about the spatial dynamics across the landscape. Interestingly, our model can separate residence times in local spatial units from migration events at regional scales in the population of interest. The model can be applied to many biological systems like identifying good stopover sites for migratory organisms, fitting birth-death processes. It can also be used as an alternative to resource selection functions.