Spatial prediction of species distribution: an interface between ecological theory and statistical modelling

     The ecosystem is a delicate and complex biological community.  Any effort for conservation or for obtaining accurate species distribution data requires more information about the ecosystem as a whole.  Scientists are able to overcome this obstacle by using GIS software and satellite imagery to map out ecosystems and identify important features.

     In order to better understand a particular ecosystem, researchers begin by identifying the vegetation in the area of interest.  Any patterns of vegetation identified provide a basis for further research and analysis. These patterns do not relate to any particular species, but rather to the vegetation’s growth and density.  As a result, the data does provide information as to the possible life found in a particular ecosystem.

    A statistical model is then developed to identify species distribution.  The statistical model uses several different tests on maps and images to analyze the ecosystem.  These tests include: statistical method, error function, regression analysis and significance tests. Various maps are needed to see the interaction that exists between models so as to identify obvious and subtle patterns.  The diagram below shows a regression analysis of  E. rossil stem density in Australia.

Austin, M. P. (2002). Spatial prediction of species distribution: an interface between ecological theory and statistical modelling. Ecological modelling,157(2), 101-118.