Incorporating movement in species distribution models




Holloway, Paul

Journal Title

Journal ISSN

Volume Title



Climate change and concomitant urbanization have led to many species shifting their geographical distribution, while other species have simply gone extinct. Understanding the current and future distributions of species is therefore a critical component of biodiversity conservation, with species distribution models (SDMs) a powerful GIScience approach increasingly used to achieve this. Movement is an ecological process that influences the distribution of all species. Broad-scale (spatially and temporally) movement includes processes like dispersal and migration that determine whether newly suitable habitats are accessible, while fine-scale movement effects resource availability, and subsequently habitat suitability. In spite of this ecological significance, movement is rarely incorporated in SDMs. An increasingly important application of SDM is to study the effects of climate change on species distributions, and while several models that incorporate species dispersal abilities have been proposed, none have been tested or compared. Past data (British birds and North American flora) were used to calibrate and extrapolate species-environment relationships to the current time-period in order to assess the accuracy of these dispersal models. Significant differences in the accuracy and area projected as present by the dispersal models were identified, and moreover, results were substantially influenced by the scale at which SDMs were calibrated. Fine-scale regular movement behaviors are another important determinant of mobile species distributions that are not currently incorporated within SDM. Spatial simulation was used to model the dynamic relationship between movement and biotic resources for oilbirds in Venezuela, in order to generate a new environmental variable for use in model calibration. The use of this layer greatly improved the accuracy and ecological realism of the SDM projection compared to other commonly applied SDM scenarios. Finally, the incorporation of movement across multiple scales has not been addressed in SDM research. Broad-scale dispersal was combined with fine-scale regular movements to predict continental changes in oilbird distribution over a decade, which improved the ecological understanding of distribution shifts and identified a number of new conceptual and methodological limitations. The incorporation of movement should now be a compulsory aspect of any study projecting the current or future distributions of species.


LCSH Subject Headings