Hierarchical Modeling Approaches for Modeling Ecological Communities

• Activity Number: 396
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Ali Arab*+
• Companies: Georgetown University
• Keywords: hierarchical models ; ecological communities ; count data ; multivariate distributions

Abstract

The problem of modeling dependence structures and complex relationships among species in an ecological community is a challenging task and requires careful implementation of statistical methods that are both realistic and plausible (i.e., easy to implement and interpret). Most conventional methods used for modeling community data have serious shortcomings (e.g., unrealistic or restrictive assumptions on data and model). For example, assumptions of normality and independence often do not reflect the nature of the data (often count data) and the relationships among species (often better explained using multivariate distributions). Hierarchical modeling approaches discussed in this work provide a more flexible and plausible but also easy to implement alternative for modeling ecological community data. The hierarchical modeling framework seems natural for modeling complex relationships among species (e.g., predator-prey-competitor structures). The modeling approaches discussed will be illustrated using data from ecological communities (e.g., food web data, fish communities).