Predicting future species distributions : from correlations toward a better understanding of underlying processes
(2015)
Files
YM_thesis_150DPI_V2.pdf
Restricted Access - Adobe PDF
- 10.81 MB
Details
- Authors
- Supervisors
- Abstract
- One of the most active disciplines in ecology these last years focuses on predicting and understanding the patterns and underlying processes of changes in species distributions under global environmental changes. Species distribution models have rapidly become the most important predictive tools in ecology to forecast the response of species to future global change. This discipline is now at a key stage of its conceptual development: it is challenged to better integrate a series of ecological processes and environmental factors that shape species distributions, abundance, community structure and ecosystem functioning. Such integration is essential to improve the ecological realism and the credibility of the future model predictions. The general objective of this PhD-thesis was to provide new perspectives in this direction. With European butterflies as model organisms, each of the core chapters tested an innovative approach focussing on the incorporation of environmental factors or ecological processes that have been neglected to date when producing future predictions. With a species distribution modelling approach addressing future shifts in species ranges, we tested the combined effect of land use and climate change, the effect of within-species spatial niche variation, and the importance of dispersal capacity for the genetic structure of the populations and persistence of the species in the future. With an experimental approach focussing on ecological processes involved during range shift, we investigated the behavioural role of habitat use to improve empirical knowledge behind the pattern of species range shifts. Based on our results, we conclude that land use change scenarios should be more often integrated with climate change scenarios in species distribution models to better reflect the response of species to global changes. We also stress the need for further development of land use change scenarios to better match the requirements of the ecological modelling community. We also provide the first evidence that the issue of within-species spatial niche variation should not be disregarded when predicting future species distribution. This issue is particularly relevant when predictions are made over large spatial scales and for widespread species. We propose an innovative local modelling approach to account for this issue and we urge for further improvements and developments in this direction. Our results also support the idea that the integration of spatially explicit information on dispersal limitation into species distribution models should provide more realistic estimation of population extinctions. Combined with data on population genetics, this approach enables the prioritisation of pro-active conservation measures toward populations that hold important intraspecific genetic diversity and that are the most threatened under future environmental conditions. Finally, we provide the first empirical evidence of increased generalism in habitat use for expanding edge populations of a butterfly during range shift. This result sheds light on some mechanisms that could promote species range shift. It also contributes to reinforcing the need for a better incorporation of fine-scale information on ecological resources available and organism-related (behavioural) factors to improve our understanding of the processes behind range shift. The main findings and perspectives in this PhD-thesis converge with the large range of theoretical and technical challenges to improve the ecological foundation of species distribution models. We believe that this thesis contributes to opening interesting avenues to help moving the species distribution modelling discipline toward this direction.
- Affiliations
UCLouvain SST/ELI/ELIB - Biodiversity
Citations
Martin, Y. (2015). Predicting future species distributions : from correlations toward a better understanding of underlying processes. https://hdl.handle.net/2078.5/39845