PREDICTION OF HIGH POTENTIAL AREAS OF HABITAT FOR MONITORED SPECIES
Jorge Argáez-Sosa, J. Andrés Christen, Miguel Nakamura, Jorge M. Soberón · 2002
AbstractThe area of distribution of a species is one of the fundamental expressions of itsecology and evolutionary history. Detailed knowledge of distribution areas is relevantto address basic questions in biogeography and ecology and in the management ofbiodiversity for conservation or sustainable use. The problem we tackle in this paperis inferring the zones of high potential for the habitat of the species under study, basedon reported sites of presence, where each site is associated with values of covariates,measured on a discrete scales. We compute the predictive probability that the speciesis present at each site, by means of a mixture involving all pairs of covariates. Possiblespatial bias for sites of presence is accounted for. Since the posterior distributiondoes not have a closed form, MCMC is implemented. However, we also describe anapproximation to the posterior distribution, which avoids MCMC. Available aprioriinformation regarding the areas of distribution of the species can be incorporated in aclear-cut way. In addition we propose a map of uncertainty which allows for greaterinsight into the nature of potential areas of distribution. By simulations, we compareour approach with other standard methods. Two case studies are also presented.Keywords: Area of distribution, Biodiversity, Mixture model, Predictive probabilitymap, a priori elicitation.