Dr John B Baumgartner1, Dr Peter D Wilson2, Dr Linda J Beaumont1, Dr Manuel Esperón-Rodríguez3
1Macquarie University, North Ryde, Australia, 2Royal Botanic Gardens Sydney, Sydney, Australia, 3Western Sydney University, Richmond, Australia
Correlative species distribution models (SDMs) are presently the most common tool for predicting habitat suitability. Maxent, a machine-learning regression-type approach to fitting SDMs based on the principle of maximum entropy, is used by a large proportion of the SDM community. The ‘rmaxent’ package for R (available at Github) facilitates rapid, Java-free projection, interpretation, and interactive interrogation of Maxent models, alongside the benefits of reproducibility and extensibility provided by R. In this Impacts talk I provide a brief overview of the key features of rmaxent, demonstrating its utility to Maxent practitioners. In particular, the increased speed with which rmaxent is able to project models is particularly valuable when projecting numerous Maxent models, such as when exploring sensitivity of suitability surfaces to model settings, or when projecting models to multiple environmental scenarios, as is increasingly common when considering potential climate change. In addition, the package permits rapid construction of multivariate environmental similarity surfaces (MESS), maps highlighting factors that limit habitat suitability, and web-based (Shiny), interactive inspection of variables’ influence on suitability at specific locations.
John is a postdoc at Macquarie University in Sydney, with particular strengths in R and ecological modelling. He is interested in the uncertain impacts of climate change on habitat suitability, and the associated implications for management of biodiversity.