University of Southampton OCS (beta), CAA 2012

Font Size: 
Spatial characterization and prediction of Neanderthal sites based on stochastic environmental modelling
Michael Märker, Michael Bolus

Last modified: 2011-12-22


In this paper, we present a unique spatial dataset of Neanderthal sites in Europe. Information on topographic locations and characteristics of the location were collected based on own work and a comprehensive literature review. Subsequently the coordinates were post processed to guarantee topographic accuracy and converted to a vector point format. The find locations were then utilized to study in detail the local characteristics of find locations. Therefore we performed a terrain analysis on SRTM 90m DEM. This analysis highlight climatic and vegetational implications, geomorpologic and hydrologic process potentials as well as strategic aspects of the closer surroundings of find locations.

For the analysis we chose an area of ca 7.3 ha around the find location. In total we delineated 46 topographic indices and used the site information to assess the specific combinations at the find locations. To reveal the importance of single indices and site specific index combinations we applied a boosted regression tree approach (BRT). The study showed that Pre-Neanderthals, Early Neanderthals and Classic Neanderthals show a specific spatial distribution. Moreover, it seems that also the environmental preferences are different especially between the Early and Classic Neanderthals. Moreover the methodology allows for a spatial prediction of occurrence probabilities for Neanderthal sites.


Neanderthal; boosted regression trees; spatial distribution; prediction modelling