Conclusions
In general, a correlative approach is appropriate when the only data available are those on species’ occurrence, in particular for reconstructing the paleoclimatic niche of fossil species or projecting their future climatic suitable area, from local to global scales. On the other hand, mechanistic models have the greatest power to assess extinction probability driven by climate change, identify conservation actions and evaluate the potential effectiveness of management interventions, but they are limited to few terrestrial species. Therefore, they are usually employed when the focus is on a well-studied species of particular conservation interest.
Observations of recent responses to climate change are a useful tool to test reliability of model predictions against current observations. However, quantifying the ability of models to provide reliable range shift projections or population changes is still challenging, since they are often difficult to validate across time and space.
When comparing past and current distribution to validate models or TVAs, a big challenge is to find accurate information on species’ historic distribution and population trends.
A glaring oversight in almost all studies is that they only focused on the direct impacts of climate change. Indirect impacts within biological communities, as well as changes in human use of natural resources are going to have substantial, complex, and often multiplicative impacts on species.
The growing human population will itself be increasingly affected by climate change, with human adaptation responses likely to result in substantial and negative impacts on biodiversity.
…
