We discussed this paper in our journal club in late September 2011. In short, the paper looks at potential deforestation scenarios projected for the Amazon basin, evaluates how the projected deforestation rates drive future population declines of birds and then re-estimates the threat categories of species based on IUCN Red List criteria. Species richness of the “soon to be threatened” species are then put on map and compared with the projected spatial patterns of deforestation and current reserve networks. This way the authors identify i) currently important areas that have high number of species that are going to become threatened by future forest loss, ii) “crisis areas”, i.e. areas that are going to be deforested and have high richness of future threatened species, iii) “refugia”, i.e. areas with high species richness that will remain intact also in the future, and iv) irreplaceable sites of future threatened species. The authors then report gaps in the current reserve network and point out recommendations for future conservation actions.

In our journal club this paper raised a couple of concerns that most likely stem from the unclear description of methods and analysis steps. Some of the terms, abbreviations and data sets are not very well defined or explained and therefore the text assumes quite a bit of background knowledge from the reader. Especially the methods would have benefited a lot from more detailed description, currently leaving many open issues such as how the population declines were calculated and how the varying generation length affects them. (Based on the text, we assumed that the population declines were simply taken to be the same as the projected ESH loss after three generation (±10%). But this would mean bigger declines for long-lived species in comparison to short-lived species, which is the opposite of what is reported in the paper. But, as the analysis steps are not very clearly explained, we assume that we have just understood them wrong.) All in all, we feel that the text was not very easy to follow.

But even more concerns were raised the way spatial prioritization was done in this study and the (quite strong) recommendations that were drawn from those prioritization. Firstly, the only measure of priority used in this paper was the number of species in a grid cell. Albeit this is quite commonly used, it is also very commonly acknowledged that it is not a particularly good approach to do spatial prioritization. Note that this method ignores many essential spatial factors such as connectivity, patch size or level of fragmentation. Hence, it does not support the persistence of species in any way. Secondly, we do not understand at all why the authors end up underlining the need to protect the crisis areas above everything, i.e. those that are projected to become deforested. These sites seem to have very little or only modest overlap with sites identified as most important today, as important refugia or as highly irreplaceable, meaning that they do not sustain any unique features of the avifauna and in general have lower species richness than the most species rich sites. Thus, why should conservation resources be used to sites that do not seem to bear any big importance in the first place? The only obvious exceptions to this are crisis areas that are currently inside protected areas but are projected to become deforested. However, the amount of these sites is difficult to estimate due to the very poor quality of Figure 1 (DDI, why did you not spot this in the review and editorial process?). It seems like the authors did the whole process in wrong order by first categorizing cells to crisis areas, refugias etc., and then calculating the priority within them. Obviously, the prioritization should have been done first and only after that the high priority pixels could have been divided into those that remain as refugia and those that become deforested. This would have allowed the authors to identify priority sites that are both high in species richness (or in irreplacability, which would be better) and are threatened by deforestation. This is the common way, and would make more sense. By advocating crisis areas the authors might end up using scarce conservation resources on sites that have very small conservation bearing. We are also concerned about the way how habitat loss seems to have been interpreted to have linear relationship with extinctions risk, which does not correspond to reality. However, this problem arises more from the IUCN classification and to be fair, we do not know if the authors accounted for this in their analysis.

All in all the paper was interesting to read, and surely this type of approaches would contribute a lot to our understanding of not-so-far-away-future threats to biodiversity. But we do feel that the methodology presented in this paper still needs some refining. Nevertheless, we salute BirdLife International for their enormous effort in compiling the data sets behind this paper. We also more than agree with the authors that in the face of major future threats like deforestation and climate change, good quality data and monitoring are more important than ever before.

This paper was published online in September 2011 and came out in February 2012:

http://onlinelibrary.wiley.com/doi/10.1111/j.1472-4642.2011.00843.x/abstract