Editor: Ian Wright
Testing instead of assuming the importance of land use change scenarios to model species distributions under climate change
Article first published online: 7 JUN 2013
© 2013 John Wiley & Sons Ltd
Global Ecology and Biogeography
Volume 22, Issue 11, pages 1204–1216, November 2013
How to Cite
Martin, Y., Van Dyck, H., Dendoncker, N. and Titeux, N. (2013), Testing instead of assuming the importance of land use change scenarios to model species distributions under climate change. Global Ecology and Biogeography, 22: 1204–1216. doi: 10.1111/geb.12087
- Issue published online: 16 OCT 2013
- Article first published online: 7 JUN 2013
- National Research Fund, Luxembourg. Grant Number: FNR-AFR PHD-09-121
- Bioclimatic envelopes;
- ecological niche modelling;
- global change scenarios;
- land cover change;
- Lycaena dispar;
- spatial resolution;
- species range shift
Species distribution models often assume a changing climate (dynamic climate variables) but unchanged land use (static land use variables) to estimate future species distribution shifts. However, scenarios of projected land use change are available to calculate dynamic land use variables. Surprisingly, the importance of using dynamic instead of static land use variables when projecting potential future species distributions under climate change remains largely unexplored. We tested whether the joint inclusion of land use and climate change scenarios altered the projection of future species distribution compared with the classical approach assuming unchanged land use.
We used land use and climate change scenarios to estimate the future distribution of a butterfly species (Lycaena dispar) according to different perspectives of projected environmental change: (1) land use change (dynamic land use and static climate variables), (2) climate change (static land use and dynamic climate variables) and (3) global change (dynamic land use and climate variables). As the importance of land use variables is known to depend on the spatial resolution of the models, we built them across a range of resolutions (50 km, 10 km and 5 km) to examine the resolution-dependent relevance of using dynamic instead of static variables.
For each resolution, the projected distribution changes were unaltered when using dynamic instead of static land use variables in the models. It was mainly due to the low thematic resolution of the land use change scenarios that include only few dynamic variables.
Even at fine spatial resolution (5 km), the available land use change scenarios poorly represent habitat suitability for the species. Hence, they may be of limited support to estimate future species distributions. Instead of supporting the assumption of unchanged future land use, our results plea for an improvement of the thematic resolution of land use change scenarios.