Biologists increasingly accept that conservation efforts must anticipate the effects of future climate change.
Bioclimatic models, which estimate how the distribution of different species changes with climate, are vital for planning conservation projects. But it is difficult to choose which model to use since biologists disagree about their relative performances.
In this article, Miguel Araújo and Carsten Rahbek examine progress in the field.
A recent assessment of prominent models found that some widely used models performed poorly, while the ones that performed best were more recent and complex.
It is challenging to assess whether models that give accurate descriptions of the current species distribution will correctly predict the future. Robust predictions of future distributions under climate change are key, but a study this year showed that results vary wildly between different models.
Techniques such as 'hindcasting' — in which models are tested with reconstructed species distributions from the fossil record — could improve accuracy.
What is clear, given the uncertainties surrounding this kind of prediction, is the need for a new round of testing.