Indirect elicitation from ecological experts: From methods and software to habitat modelling and rock-wallabies
Claudia Tebaldi and Richard Smith
This article focuses on techniques for eliciting expert judgement about complex uncertainties, and more specifically the habitat of the Australian brush-tailed rock-wallaby. Modelling wildlife habitat requirements is important for mapping the distribution of the rock-wallaby, a threatened species, and therefore informing conservation and management. The Bayesian statistical modelling framework provides a useful ‘bridge’, from purely expert-defined models, to statistical models allowing survey data and expert knowledge to be ‘viewed as complementary, rather than alternative or competing, information sources’. The article describes the use of a rigorously designed and implemented expert elicitation for multiple experts, as well as a software tool for streamlining, automating and facilitating an indirect approach to elicitation. This approach makes it possible to infer the relationship between probability of occurrence and the environmental variables and demonstrates how expert knowledge can contribute to habitat modelling.
Alan Gelfand and Sujit K. Sahu
This article discusses the use of Bayesian analysis and methods to analyse the demography of plant populations, and more specifically to estimate the demographic rates of trees and how they respond to environmental variation. It examines data from individual (tree) measurements over an eighteen-year period, including diameter, crown area, maturation status, and survival, and from seed traps, which provide indirect information on fecundity. The multiple data sets are synthesized with a process model where each individual is represented by a multivariate state-space submodel for both continuous (fecundity potential, growth rate, mortality risk, maturation probability) and discrete states (maturation status). The results from plant population demography analysis demonstrate the utility of hierarchical modelling as a mechanism for the synthesis of complex information and interactions.
Peter Challenor, Doug McNeall, and James Gattiker
This article examines the dynamics of the US economy over the last five decades using Bayesian analysis of dynamic stochastic general equilibrium (DSGE) models. It highlights an example application in what is commonly referred to as the new macroeconometrics, which combines macroeconomics with econometrics. The article describes a benchmark New Keynesian DSGE model that incorporates four types of agents: households that consume, save, and supply labour to a labour ‘packer’; a labour ‘packer’ that puts together the labour supplied by different households into an homogeneous labour unit; intermediate good producers, who produce goods using capital and aggregated labour; and a final good producer that mixes all the intermediate goods. It also considers the application of the model in policy analysis for public institutions such as central banks, along with private organizations and businesses. Finally, it discusses three avenues for further research in the estimation of DSGE models.