Robert Denham

Geographically assisted elicitation of expert opinion for regression models

One of the perceived strengths of Bayesian modelling is the ability to include prior information. Although objective or noninformative priors might be preferred in some situations, in many other applications the Bayesian framework offers a real opportunity to formally combine data with information available from experts. The question addressed in this paper is how to elicit this information in a form suitable for prior modelling. Particular attention is paid to geographic data for which maps might be used to assist in the elicitation. Two case studies are used to illustrate the methodology: estimation of city house prices and prediction of presence of a rare species.

Assessment and monitoring of foliage projected cover and canopy height across native vegetation in Queensland, Australia, using laser profiler data

The aim of this project was to demonstrate the potential of laser profiling for monitoring forest structure. Data were captured from flights at 30, 60, and 100 m above the canopy over three study sites in south-east Queensland at regular intervals over a 2-year period. Field measurements of foliage projected cover (FPC) and tree height were found to be highly correlated with laser derived estimates (R2 from 0.91 to 0.95). Monitoring of changes in FPC and tree height, as a result of logging or growth, was also successful. Tree heights, in particular, were measured accurately over time (residual standard error (RSE) of 0.45 m). The large RSE of the FPC model (from 5.7% to 7.3% FPC) means that subtle changes, such as seasonal variation, may be difficult to monitor. Flying height was found to be a significant explanatory variable in estimating field FPC. A transect of greater than 1000 km was also flown in a single helicopter pass to assess the technology over a range of forest types. Field measurements of FPC were collected for 21 sites along this transect. Strong relationships were observed between laser and field FPC, but these varied with forest type.

Efficient Bayesian estimation of multivariate state space models

A Bayesian Markov chain Monte Carlo methodology is developed for the estimation of multivariate linear Gaussian state space models. In particular, an efficient simulation smoothing algorithm is proposed that makes use of the univariate representation of the state space model. Substantial gains over existing algorithms in computational efficiency are achieved using the new simulation smoother for the analysis of high dimensional multivariate time series. The methodology is used to analyse a multivariate time series dataset of the Normalised Difference Vegetation Index (NDVI), which is a proxy for the level of live vegetation, for a particular grazing property located in Queensland, Australia.

The Bayesian conditional independence model for measurement error: applications in ecology

The measurement error model is a well established statistical method for regression problems in medical sciences, although rarely used in ecological studies. While the situations in which it is appropriate may be less common in ecology, there are instances in which there may be benefits in its use for prediction and estimation of parameters of interest. We have chosen to explore this topic using a conditional independence model in a Bayesian framework using a Gibbs sampler, as this gives a great deal of flexibility, allowing us to analyse a number of different models without losing generality. Using simulations and two examples, we show how the conditional independence model can be used in ecology, and when it is appropriate.

Spatially stratified sampling using auxiliary information for geostatistical mapping

This paper presents a method of spatial sampling based on stratification by Local Moran’s Ii calculated using auxiliary information. The sampling technique is compared to other design-based approaches including simple random sampling, systematic sampling on a regular grid, conditional Latin Hypercube sampling and stratified sampling based on auxiliary information, and is illustrated using two different spatial data sets. Each of the samples for the two data sets is interpolated using regression kriging to form a geostatistical map for their respective areas. The proposed technique is shown to be competitive in reproducing specific areas of interest with high accuracy.

Modelling habitat and planning surveillance using Landsat imagery: a case study using Imported Red Fire ants

This paper presents a Bayesian mixture model approach for detecting areas of habitat that are suitable for S. invicta infestation, aiding the ongoing surveillance for early detection of these exotic pests. We show that Landsat imagery is an affordable and valuable tool to assist in determining an informed surveillance strategy. In this paper, we use Landsat band 3 (visible red), Landsat band 6 (mid infrared) and a soil brightness index, in various combinations, to assess the probability that the area associated with each pixel is habitable terrain, either in a multivariate analysis, or as a univariate spatial temporal model. The multivariate analysis allows researchers to create meaningful clusters that reflect the sometimes complex combinations of conditions of conditions that form suitable habitat, rather then relying on single derived indices.

Investigating the Potential for Mapping Fallow Management Practises Using MODIS Image Data

The objective of this study was to investigate the potential for mapping fallow land management practices on local farm scale in Southern Queensland, Australia, using high temporal frequency satellite remote sensing over a period of six years. The Moderate Resolution Imaging Spectroradiometer (MODIS) was chosen as it provides a temporal resolution fine enough to detect ground cover change within cropping cycles (fallow periods). Previous studies have successfully employed MODIS data detecting cropping patterns in Kansas, North America and Northern Kazakhstan.