Interesting paper that aims at making the predictions of climate change impacts more realistic. Here the authors have first used the traditional bioclimatic envelopes to predict changes in the suitable climate for about 230 spp. Then they selected three spp with contrasting patterns in distribution change and by using stochastic population models, simulated the changes in population abundance for each of the species between the years 2000 and 2050.
Some interesting points that we picked up from the results:
1) The most robust generalization emerging from the simulations was, that the length of fire intervals had a significant impact on population abundance in all cases.
2) The species most severely affected by climate change was the one with large initial distribution that went through a major contraction due to climate change. The other two species (B: restricted range contracting, but increasing in habitat suitability and C: restricted range undergoing shift and fragmentation) were clearly less affected by the climate induced changes in range.
3) With the species studied here, different dispersal strategies seemed to have no affect on responses to climate change.
Naturally, all these three points can be data specific and it would have been interesting to see results achieved with other types of species/ecosystems. All in all, we were impressed by the complexity and coverage of this study. The method developed here obviously increases our knowledge on how to predict the impacts of climate change on species. However, it still remains to be seen how well this approach can be extended to larger species groups as the population model used here requires that both the life-history and the habitat requirements of the species are well understood. It is good to keep in mind that the species data (the Protea Atlas) in this study comes from one of the best-known ecosystems in the world.
Final conclusions: Definitely worth reading :)!
Link to the paper
Keith et al. 2008. Predicting extinction risks under climate change: coupling stochastic population models with dynamic bioclimatic habitat models. Biology Letters. Vol. 4, p.560-563
http://rsbl.royalsocietypublishing.org/content/4/5/560.abstract