Graduate Certificate in Health (Specialisation)

Using Bayesian Belief Networks to model urban bird macroecology

Lawrie is an ecologist with about 30 years of diverse experience in zoological and ecological survey, research and management. The masters research project was inspired by both the interesting and frustrating elements of working as an ecological consultant, as well as a wide-raging insatiable curiosity, and a love of birds and the ecological research process.

One of the constantly recurring themes in consulting work is the problem of quickly assessing scenarios and making diagnoses on the likelihood of serious impacts on biodiversity, and recommendations for avoiding or mitigating impacts – with varying degrees of missing data or imperfect understanding, and under uncertainty. That’s both frustrating and interesting!

The aims of this research are to model the relationship between a range of urban environmental variables with urban bird diversity and distribution in Melbourne. The research will employ Bayesian Belief Networks (BBNs); and will draw together data from existing sources such as the Birds Australia ‘Atlas II’ project, a range of landscape metrics developed at the Australian Research Centre for Urban Ecology (ARCUE), together with expert knowledge from a select panel, and further ecological variables relating to bird assemblage structure, climate, topography, etc. developed during this research.

It is expected that these models will be used to identify key environmental risks to survival of native urban bird communities in an urbanising landscape, predict bird species richness changes under various development scenarios, and predict important changes to bird assemblage structure resulting from various development scenarios. The models will be constructed in such a way as to be capable of being constantly updated and enhanced, and to be suitable for further development into end-user ecological risk assessment tools (for biodiversity managers, ecologists working in multidisciplinary urban design teams, and so on).