Our knowledge about Neanderthals has greatly increased over the last two decades with more than two thousand research papers published in areas as diverse as chronology \cite{Gaudzinski_Windheuser_2014, Higham_2014}, ecology \cite{Finlayson_2007, Henry_2010}, population dynamics \cite{S_rensen_2011, Bocquet_Appel_2013}, adaptive traits \cite{S_rensen_2009, Rae_2011}, diet \cite{Henry_2010, Hardy_2011}, technology, cognition and behavior \cite{Shaw2012, Fern_ndez_Peris_2012, Riel_Salvatore_2010}, genetics \cite{Briggs_2009, Sanchez_Quinto_2014}, and their relationship with anatomically modern humans \cite{Hortolà 201369, Sankararaman_2012}. Still, we believe there is a lack of tangible information about the factors controlling the Neanderthals’ distribution. For example, \citet{Wenzel_2007} describes the distribution of Neanderthals during MIS 5e across central Europe, and briefly discusses the influence of climatic conditions and habitat features over the general occupancy pattern, but does not establish any quantitative link between presence and environment. \citet{Banks_2008} analyzed the distribution of Neanderthals at the interphase of the Neanderthal/modern human transition applying palaeo species distribution models (PSDMs hereafter), assessed niche conservatism, and analyzed the importance of climatic predictors, concluding that temperature was the most important driver shaping the distribution Neanderthal at this time.
In this paper we propose a hypothesis about how abiotic drivers may have shaped Neanderthals distribution that is rooted in the Grinnellian niche concept \cite{Grinnell_1917}, the hierarchical framework proposed by \citet{Pearson_2003}, and our current knowledge on the ecology of Neanderthals. According to \citet{Pearson_2003}, climate influences species distribution at global and continental scales, while the effect of topography is restricted to scales ranging from regional to local. Our hypothesis consist of three main points: 1) the northern edge of the Neanderthals range was limited by low winter temperatures; 2) the southern edge was shaped by a combination of high temperature and low water availability during the summer; 3) high topographic diversity combined with moderate slopes could have favored occupation at the local scale. Winter temperatures at the northern edge could have resulted in a lower availability of small and big game than in temperate areas \cite{Badgley_2000}, compromising the high caloric intake required by this species \cite{Steegmann_2002, S_rensen_2009}, and an increased cold stress accelerated by a low-caloric diet, that would have lead to reduced fertility \cite{Bocquet_Appel_2013} and a higher mortality rate \cite{Steegmann_2002}. Summer temperatures linked to continentality and the higher solar radiation of lower latitudes could have prevented the occupation of southern plains in the Mediterranean peninsulas due to increased heat stress, specially considering the low body surface area/volume ratio of this species \cite{Churchill2006}, which hampers heat dissipation. Also, high evapotranspiration could have reduced the diversity of plants, since under-storey and forest communities are less common under such climatic conditions, hampering the access to plant resources, and making these areas unsuitable during the warm season. Mediterranean coastal areas could have been suitable because of the buffering effect of the sea over temperature and the permanent availability of resources like shellfish \cite{Hardy_2011}. At the local scale, high topographic diversity, which fosters biodiversity by an increased availability of ecological niches \cite{Tews_2003}, could have provided the required abundance and diversity of prey and shelter \cite{Daujeard_2012}, but moderate to low slopes may have been important to reduce the high energetic expenditure of mobility in steep terrain. This limitation, combined with the lower abundance of animals at higher elevations \cite{Brown_2001}, could have prevented Neanderthal’s presence in the higher elevations of the European mountain ranges.
To test our hypothesis, we have modeled and analyzed the distribution of Neanderthals during the full Eemian Interglacial (MIS 5e, \(\sim\)130 ka BP) using a PDSM approach \cite{Franklin20151, Svenning20112930, Varela2011451}. PSDMs rely on the same principles of species distribution modeling \cite{Guisan2000147, Guisan_2005}, requiring presence coordinates coming from fossil/archaeological remains, a set of predictors (usually palaeo-climatic simulations), and an algorithm to define the relationship between presence and predictors. The result of a PSDM can be defined as a habitat suitability map in which each cell is scored according to how well it resembles the ecological conditions of the localities where the species occurs \cite{Soberon_2005}. PSDMs can be used to analyze the influence of particular drivers over the species distribution and gain ecological knowledge, as in \citet{Varela_2009} or \citet{Rodr_guez_S_nchez_2008}. PSDMs have been used before to study the distribution of Neanderthals and anatomically modern humans \cite{Banks_2008, Banks2008481, Beeton_2013, Burke201435}, but in this paper we add to the previous studies by offering a novel insight into the drivers of Neanderthals distribution both at continental and local scales.
We selected the Eemian Interglacial (MIS 5e, \(\sim\)130 ka BP) as modeling period for three different reasons: 1) the distribution of Neanderthals during this period has been previously assessed by i.e. \citet{Richter2005}, \citet{Richter2006}, \citet{Wenzel_2007}, and \citet{Gaudzinski2011}, providing a good baseline knowledge, but with these studies mainly focus on the core and northern edge of the distribution area across central Europe, and lack a quantitative description of Neanderthals distribution and its limiting factors; 2) since the publication of \citet{Wenzel_2007}, new Neanderthal remains attributed to the Eemian have been found in Spain \cite{Arsuaga2012629}, Italy \cite{Fiorenza2015119}, France \cite{Moncel200765, Daujeard2010368}, and these new sites have the potential to change our view about the Eemian distribution of Neanderthals; 3) During the Eemian, the warmer climatic conditions freed Europe from the Saalian ice sheet, offering the Neanderthals a unique opportunity to spread throughout Europe for c. 10000 years, and probably allowing them to reach their maximum range size. The Eemian can therefore be considered the most suitable period to assume a pseudo-equilibrium with climate for Neanderthals distribution, a key assumption for PSDMs \cite{Guisan_2005, Guisan2000147}.
In summary, in this paper we propose a hypothesis about how abiotic drivers (climate and topography) could have shaped Neanderthals distribution throughout Europe, and we test it by fitting and analyzing a PSDM describing their distribution during the Last Interglacial period.