New paradigms for modelling species distributions?

Abstract

Summary1. The management of both desirable and undesirable species requires an understanding of the factors determining their distribution. Quantitative distribution models offer simple methods for formulating the species-habitat link and the means not only for predicting where species should occur, but also for understanding the factors involved. Generalized linear modelling, in particular, links the incidence of species to habitat variables, and has increasingly formed the backbone of the modelling approaches used. New 'data technologies', such as remote sensing and geographical information systems, have further broadened these modelling applications to almost any ecological system and any species for which there are distribution data. 2. Many previous approaches have aimed to identify the most parsimonious model with the best suite of predictors, selected on the basis of null hypothesis testing. However, information-theoretic approaches based on Akaike's information criterion allow the selection of a best approximating model or a subset of models from a set of candidates. Information-theoretic approaches require a deeper understanding of the biology of the system modelled and may well become an improved paradigm for species distribution modelling. 3. Synthesis and applications . This special profile of six papers demonstrates the development in methodology used in species distribution modelling. The papers show how information-theoretic approaches can be coupled with emerging data technologies to address issues of conservation significance. With conservation biology and applied ecology at the forefront of many of the basic science developments so far, we expect these methods to pervade other areas of ecological research more fully in future.

Modelos de distribución de especies: Una revisión sintética

“…Información Geográfica y el desarrollo de técnicas estadísticas aplicadas ha permitido en los últimos años la expansión de herramientas para el análisis de los patrones espaciales de presencia y ausencia de especies: los modelos de distribución de especies (Franklin 1995, Guisan & Zimmermann 2000, Rushton et al 2004, Foody 2008, Swenson 2008. Los modelos de distribución de especies están en pleno desarrollo y expansión con nuevos métodos y estrategias para el tratamiento e interpretación (Wilson et al 2005, Elith et al 2006, Ferrier & Guisan 2006, Mateo 2008.…”

Predicting tree distributions in an East African biodiversity hotspot: model selection, data bias and envelope uncertainty

“…Marginal statistics can be biased by the inevitable collinearity amongst environmental predictors (Cohen et al, 2003;Graham, 2003), and so the use of null hypothesis tests during selection is best avoided. Issues of multiple testing (Pearce and Ferrier, 2000a;Whittingham et al, 2006) and arbitrary levels of statistical significance (Mickey and Greenland, 1989;Rushton et al, 2004) further enforce this standpoint. Multimodel inference has been proposed as an alternative to best-model stepwise procedures.…”

Predicting the future of forests in the Mediterranean under climate change, with niche- and process-based models: CO2 matters!

“…Detailed knowledge of species' ecological and geographic distributions is fundamental for conservation planning and forecasting (Ferrier, 2002;Funk & Richardson, 2002;Rushton et al, 2004), for understanding ecological and evolutionary determinants of the spatial patterns of biodiversity (Ricklefs, 2004), and the potential response of these distributions to future climatic change (e.g. Thomas et al, 2004;Araú jo et al, 2005a, b;Thuiller et al, 2005).…”