EAA2021: Abstract
Abstract is part of session #505:
Title & Content
Title:
EmorphProject:Examining mobility degree of Palaeolithic reindeer through an osteometric analysis of metapodial bones and phalanges in combination with ML algorithms
Content:
Mobility patterns of reindeer will have affected the hunting strategies and the mobility of hunter-gatherer groups that depen upon them. The question of the season occupation of human territories and the seasonal behaviour of reindeer remains open. Due to the important role that reindeer (Rangifer tarandus) played for human populations in Western and Central Europe during much of the Palaeolithic period, many studies have focused on it particularly during the Magdalenian. However, and despite the numerous attempts to reconstruct the migratory behaviour of Palaeolithic reindeer, there has yet to be a definitive model that can explain their movement patterns.
Modern ethological data indicates that reindeer herds adopt different mobility strategies that correlate with habitat type and topography. Furthermore, it has been demonstrated that an animal’s habitat and mobility patterns hypothetically affect bone density and limb bone morphometry as has been proven in several species.
Therefore, the main goal of our project is to quantify the link between mobility and morphology using metacarpals, metatarsals and phalanges (because they occur more often in our archaeological records) from modern caribou populations (migratory and sedentary settings) and to apply these results to twenty-three faunal assemblages from Upper Palaeolithic archaeological sites in South-western France to reconstruct reindeer mobility.
Thus, metapodial bones and phalanges from modern and archaeological samples have been studied taking linear measurements from them in combination with Machine Learning (ML) algorithms. As a result, ML analyses showed how it is possible to distinguish reindeer mobility patterns with more than 90% of accuracy depending on the skeletal part and bone type. These results will allow us to reconstruct how migratory patterns of Paleolithic reindeer affected human hunting strategies and the mobility of hunter-gatherer groups that depended upon them during Magdalenian period.
Modern ethological data indicates that reindeer herds adopt different mobility strategies that correlate with habitat type and topography. Furthermore, it has been demonstrated that an animal’s habitat and mobility patterns hypothetically affect bone density and limb bone morphometry as has been proven in several species.
Therefore, the main goal of our project is to quantify the link between mobility and morphology using metacarpals, metatarsals and phalanges (because they occur more often in our archaeological records) from modern caribou populations (migratory and sedentary settings) and to apply these results to twenty-three faunal assemblages from Upper Palaeolithic archaeological sites in South-western France to reconstruct reindeer mobility.
Thus, metapodial bones and phalanges from modern and archaeological samples have been studied taking linear measurements from them in combination with Machine Learning (ML) algorithms. As a result, ML analyses showed how it is possible to distinguish reindeer mobility patterns with more than 90% of accuracy depending on the skeletal part and bone type. These results will allow us to reconstruct how migratory patterns of Paleolithic reindeer affected human hunting strategies and the mobility of hunter-gatherer groups that depended upon them during Magdalenian period.
Keywords:
Magdalenian, Paleolithic reindeer, Machine Learning, Hunter-gatherers
Downloads:
authors
Main authors:
Ana Belén Galán López1
Co-author:
Sandrine Costamagno1
Ariane Burke2
Ariane Burke2
Affiliations:
1 TRACES (UMR 5608 CNRS)
2 Université de Montréal
2 Université de Montréal