Picture by Michela Leonardi
A new preprint to which I collaborated was just submitted to BioRxiv: mtDNA-based reconstructions of change in effective population sizes of Holarctic birds do not agree with their reconstructed range sizes based on paleoclimates. The work is led by Eleanor Miller, and was performed under the supervision of Andrea Manica and Bill Amos (University of Cambridge).
In this article we have studied 102 bird species living in different environments of Eurasia and North America, trying to understand how the climatic changes that occurred after the last glacial maximum (around 21,000 years ago) influenced their demographics . In fact, during the last glacial maximum the climate was much colder, perennial ice covered a large part of the northern hemisphere, and some environments were much more widespread (for example the steppe and cold prairies) while others were much less widespread (for example example forests). For this reason, a difference in the demographic response of species living in different environments could be expected.
Reconstructing the demographics of the past is a very difficult task, there is no method that allows you to do it directly. What can be done is to use different methods that calculate measures that can give us indirect information on what the number of individuals could have been at a given moment. In our article, we used two of these methods, which are based on different data and different assumptions, in order to maximize the amount of information obtained.
The first of these approaches are Bayesian Skyline Plots, which reconstruct the effective population size over time based on mitochondrial DNA. Although the name may be misleading, this measure is not strictly linked to the number of individuals, rather it indicates the degree of genetic variability present in the population. It is based on the assumption that all individuals have the possibility of interbreeding with each other, and the same probability of reproducing: under these conditions, a population with more individuals has a higher genetic variability, for this reason the reconstructions of the effective size are considered informative on demographics. However, they must be interpreted carefully because they can also be influenced by the degree of geographical isolation, by the presence of geographical barriers between groups of individuals, and by many other factors. I will soon be publishing a book chapter on this subject, which clears up some of the more frequent mistakes that can be made in interpreting this kind of information.
The second method is ecological modeling of species distribution (Species Distribution Modelling). This class of methods associates the observations of a species with the environmental or climatic characteristics in which it lives, to reconstruct the potential distribution area both in the present and in the past (or in the future) when simulations of the climate of other periods are available. Again, the size of the distribution area is not directly related to the number of individuals, but this measure is often used as a proxy for demographics assuming that larger distribution areas can support a greater number of individuals.
Our analyses show that when it comes to demographics of the past it is essential not to consider the information drawn from a single method, and to remember that behind every model or measure there are important assumptions that must be tested from time to time. Reality is always more complex than the methods we use to reconstruct it, which is why it is necessary to integrate different approaches in order to be able to have the most complete picture of the situation possible.
Eleanor F. Miller, Rhys E. Green, Andrew Balmford, Robert Beyer, Marius Somveille, Michela Leonardi, William Amos, Andrea Manica
mtDNA-based reconstructions of change in effective population sizes of Holarctic birds do not agree with their reconstructed range sizes based on paleoclimates
During the Quaternary, large climate oscillations had profound impacts on the distribution, demography and diversity of species globally. Birds offer a special opportunity for studying these impacts because surveys of geographical distributions, publicly-available genetic sequence data, and the existence of species with adaptations to life in structurally different habitats, permit large-scale comparative analyses. We use Bayesian Skyline Plot (BSP) analysis of mitochondrial DNA to reconstruct profiles depicting how effective population size (Ne) may have changed over time, focussing on variation in the effect of the last deglaciation among 102 Holarctic species. Only 3 species showed a decline in Ne since the Last Glacial Maximum (LGM) and 7 showed no sizeable change, whilst 92 profiles revealed an increase in Ne. Using bioclimatic Species Distribution Models (SDMs), we also estimated changes in species potential range extent since the LGM. Whilst most modelled ranges also increased, we found no correlation across species between the magnitude of change in range size and change in Ne. The lack of correlation between SDM and BSP reconstructions could not be reconciled even when range shifts were considered. We suggest the lack of agreement between these measures might be linked to changes in population densities which can be independent of range changes. We caution that interpreting either SDM or BSPs independently is problematic and potentially misleading. Additionally, we found that Ne of wetland species tended to increase later than species from terrestrial habitats, possibly reflecting a delayed increase in the extent of this habitat type after the LGM.
bioRxiv 2019.12.13.870410; doi: https://doi.org/10.1101/2019.12.13.870410