|
Activity Number:
|
406
- Spatio-Temporal Methods in Ecology and Epidemiology
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #324862
|
|
|
Title:
|
Joint Point Pattern Modeling for Species Co-Occurrence Using Camera Trap Data
|
|
Author(s):
|
Erin M Schliep* and Alan E. Gelfand and James S Clark and Daniel Taylor-Rodriguez
|
|
Companies:
|
University of Missouri and Duke University and Duke University and Michigan State University
|
|
Keywords:
|
linear model of coregionalization ;
log-Gaussian Cox process ;
MCMC
|
|
Abstract:
|
Camera traps, which activate and record times when an animal crosses in front of the camera, can be used to investigate daily activity patterns of species. Traditional approaches "wrap time" and view the observations from the camera traps as circular data. In circular time, two species can have very similar daily activity patterns and yet never co-occur in a proximate sense. Modeling species activity patterns in linear time preserves this notion of co-occurrence. We propose a multivariate log-Gaussian Cox process to model daily activity patterns of forest mammals. Species-specific intensities are modeled jointly to capture general activity patterns of the species while accounting for possible co-occurrence or competition between the species. Model inference is obtained in a Bayesian framework with an efficient Markov chain Monte Carlo sampling algorithm. Overlap measures using the intensity functions can detect similarities in daily activity patterns. We can infer about the probability of presence of one species in a particular time interval given presence of another species in the same or adjacent interval, addressing the question of proximate co-occurrence.
|
Authors who are presenting talks have a * after their name.
