Lenton, S. Fa, J.E. and Perez del Val J. 2000. A simple non-parametric GIS model for predicting species distribution: endemic birds in Bioko Island, West Africa. Biodiversity and conservation. 9:869-885

            The authors in this study describe the computer-based methodology PREDICT, used to analyze habitat requirements of species.  PREDICT was written in C for a DOS. Thus, it can operate itself or as part of a GIS. It basically uses the correlations of presence/absence data with habitat factors from a geographical database. To test it, they use a dataset of endemic birds from Bioko Island, offshore of Camerun. The environmental variables used in the model, were obtained by capturing, transforming, and editing three maps: a topographic map, a distribution map of the main types of vegetation and an isohyet map. This was done using GIS ARC/INFO; then data was transferred to IDRISI to run the model. The result was that each pixel will have a value from the sum of all individual habitat factors. Thus, the output maps depicted areas of four habitat types: highly suitable, suitable, unsuitable and highly unsuitable. Author will review the accuracy of the model through future field observations.