Modeling Density Dependence in the Presence of Observation Error

• Activity Number: 140
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Quinn Payton*+ and Paul Murtaugh and Virginia Lesser
• Companies: Oregon State University and Oregon State University and Oregon State University
• Keywords: Density Dependence ; Measurement Error ; Autoregressive Models

Abstract

The modeling of growth in ecological populations is critically important for wildlife management efforts. A population's growth rate is determined by both birth and immigration, while the population's loss rate is primarily driven by death and emigration. Density dependence refers to the relationship between the population's size and its gain and loss rates. When the magnitude and direction of these rates of change are influenced directly by the population size, they are considered density-dependent. Researchers often attempt to assess density dependence based on annual estimates of population size. The small sizes of such datasets have led many to attempt nonparametric approaches all of which have been found to have significant shortcomings and biases. However, more recently parametric alternatives are being investigated. One of the more recent and robust of these parametric methods is based on the Gompertz State-Space model, in which maximum likelihood estimation and parametric bootstrapping are used to identify and quantify density dependence. I will present problems that arise when using these methods along with some suggested remedies.