Document: KEV-3-91-9

A Bayesian approach to estimating demographic matrices.

GROSS, K.* ¹ and B.A.CRAIG ²

University of Wisconsin-Madison, Madison, WI 53706 USA ¹
Purdue University, West Lafayette, IN 47907 USA ²

Abstract:
Estimating the growth rates, survival rates, and fecundities in a demographic matrix model can be challenging when individuals cannot be monitored directly. Here, we present a hierarchical Bayesian approach to estimating these parameters from unequally-spaced field estimates of stage-specific abundances. Our model hierarchy uses the demographic matrix to model the trajectories of the unobserved dynamics and an observational model that links these dynamics to field data. With this model, computational techniques such as Markov chain Monte Carlo can be used to estimate a joint posterior distribution of the matrix parameters. This joint posterior also provides posterior distributions for other matrix quantities of interest, such as the population multiplication rate and the stable stage distribution. We illustrate this approach by estimating demographic matrices for populations of pea aphids, Acyrthosiphon pisum (Harris), in alfalfa in south central Wisconsin.

Keywords:

This abstract is being presented at: 10:30 AM in session:
STATISTICAL ECOLOGY