Will intra-specific variation buffer species and help reduce extinction risks due to climate change?

Many studies have touted the role of biodiversity in buffering ecosystems against disturbances such as climate change.  A new study indicates that intra-specific diversity may also buffer species against the negative effects of climate change.

One of the most commonly employed tools for predicting the effects of climate change is species distribution modeling. Species distribution models, or SDMs, use different techniques to relate the known occurrences of a species to underlying environmental variables and estimate the realized niche of the species. Based on the duality of the niche, suitable conditions can then be identified across a wider map of environmental variables to predict the potential geographic distributions of the species. By changing the map of environmental variables, SDMs can also predict the geographic distribution of species under altered conditions.  For example, the distribution of a species can be predicted under current climate and then again under the future climatic conditions forecast by GCMs as a means of predicting the effects of climate change on that species.  Many studies have done just this for large number of different species and have generally predicted rapid decreases in the range areas of the species which leads to the concern that many of these species will be at increased risk of extinction.

There are many different forms of SDMs each with their own assumptions and hence limitations. One common assumption in almost all SDMs is a lack of local adaptation or intra-specific variation in environmental tolerances/preferences. Oney et al. argue that this is a very significant oversight and that the inclusion of intra-specific variation will change the SDM predictions for the future.

To test this contention, Oney et al., in their study “Intraspecific variation buffers projected climate change impacts on Pinus contorta published in the journal Ecology and Evolution, model the current and future (2070-2100) distributions of the conifer tree species, Pinus contorta. The researchers generate their SDMs while ignoring intra-specific variation and also while accounting for potential differences between three subspecies: contortamurrayana and latifolia. The change in the habitat area of the entire species and each subspecies due to climate change are then calculated assuming no or perfect migration.

Ignoring sub-species, Oney et al. predict 60% habitat loss for Pinus contorta with no migration and 51% habitat loss with perfect migration. Conversely, if intra-specific variation is accounted for, and the species range is assumed to be the sum of the individual subspecies ranges, the authors predict a 26% habitat loss with no migration and 8% habitat gain with perfect migration. Importantly, the different sub-species have very different results with latifolia predicted to do relatively well and murrayana predicted to suffer from large losses in habitat .

To be honest, these results surprise me. A general belief is that smaller-ranged ‘species’ are assumed to be more sensitive to climate change (especially under no migration scenarios). As such, I would expect a greater overall risk of extinction if you break one species into three smaller sub-species. What I think is happening in the present study is that in the species model (i.e., ignoring sub-species), differential sampling of the different sub-species could decrease the predicted probability of occurrence in some parts of the range that are assumed to be suitable in the sub-species model. In other words, the sub-species model weights each sub-species equally while the species model functionally weights the subspecies by abundance (real or sampled abundance). In addition, since in the sub-species model the probability of the species occurring in a given area was calculated as the sum of the sub-species probabilities, the overall probability in any given cell will generally be higher than under the species model, especially since the probabilities are relativized (on a scale of 0 to 1) across the species’ or subspecies’ entire range (this is related again the weighting issue mentioned above).

So in the end, I am convinced that the inclusion of intra-specific variation is a necessary improvement or next step in the development of SDMs. I am unconvinced that intra-specific variation will generally buffer species against climate change. Instead, I think that in many cases, local adaptation of populations, especially when coupled with limited dispersal/migration abilities, will actually result in even greater extinction risks for species in a changing world.

–Ken Feeley