Climate change in Europe between 90 and 50 kyr BP and Neanderthal territorial habitability

from the literature, not from a planned sampling scheme. In order to overcome problem of possible biases on the studied sites and ensure the stability of our model, we implemented a bootstrap with 1000 replicates, each replicated sample obtained by sampling with replacement from the presence points, with the number of samples equaling the total number of presence points. Thus, it is difficult to differentiate whether presence in particular environments is because those locations were really preferable to Neanderthals or if this is due to the fact that a location has been the subject of the largest search efforts, has better conditions for conserva- tion of fossils or simply because it is accessible for excavations. This question is one of the big- gest sources of bias in the field of prehistory in particular, and archaeology in general. Our approach, presence-only SDM based on maximum entropy, helps to solve this problem and provide new indications of the probability of finding Neanderthal sites (with fossils, lithic industry or fauna) for the period 90 kyr—50 kyr in the European area. It should be noted that our results encounter two biases. Firstly, our research only high- lights areas that show environmental characteristics comparable to those of previously discov- ered Neanderthal sites. This is nevertheless a step forward, encouraging excavation of those areas with a high probability of discovering Neanderthal remains because we have identified areas with environmental characteristics strongly suitable for Neanderthal occupation. The

second bias concerns the geographical scale of our study, which covers the entire European region and does not focus on the detail of more geographically limited areas. Indeed, it would also be preferable if the climate and ecological reconstructions our work relies on were explic- itly conducted at higher spatial resolutions and able to more explicitly take account of very local conditions. However, our work nevertheless provides an insight into the presence of areas suitable for Neanderthals outside the already known areas and precisely during the period when a decrease in the size of the Neanderthal population as a whole is expected. Despite this, our research contributes to the large number of studies that are particularly inter- ested in the transition between Neanderthal and modern humans [31–33]. Our study shows that the most important changes in the extent of ecologically suitable areas for Neanderthals occurred between the P1 (90–83 kyr) and P2 (83–69 kyr). This broad result was not unexpected because the second period is warmer than the first. What has not been shown before is that from the second period onwards the "habitability" for Neanderthals in Europe does not change significantly. The European climate simulated in the models we have used here does of course change, becoming generally cooler and drier again after 69 kyr. However, these changes are not geographically uniform and they are not large enough to Fig 17. Comparisons by SSIM index of Maxent estimated probability maps for P4vs P5. Dark green areas (less than 0.2) are considered as dissimilar.

https://doi.org/10.1371/journal.pone.0308690.g017 PLOS ONE Neanderthal habitability PLOS ONE | https://doi.org/10.1371/journal.pone.0308690 February 26, 2025 12 / 17