Periodic Reporting for period 2 - EmorphProject (Reconstructing habitat type and mobility patterns of Rangifer tarandus during the Late Pleistocene in Southwestern France: an ecomorphological study.)
Periodo di rendicontazione: 2020-04-01 al 2021-03-31
As an animal´s habitat and mobility patterns hypothetically affect bone density and limb bone morphometry, the goals of this project are two: 1) test this ecomorphological hypothesis for a given taxon (Rangifer tarandus), and; 2) establish a reference sample for this taxon that will enable a reconstruction of its habitat type and degree of mobility using bone samples recovered from archaeological sites. This project quantifies the link between habitat type and mobility and bone density and morphology using Computer Tomography (CT) and geometric morphometry (GMM).
Once the relationship between habitat, mobility and bone structure has been quantified, the information collected was applied to faunal assemblages from Magdalenian archaeological sites in Southwestern France.
Thus, the present 30-month study has provided a novel methodology to test mobility patterns (migratory and sedentary) according to metapodial bone shape (Geometric Morphometrics) and osteometric data from phalanges as well in combination with Machine Learning (ML) algorithms (that could successfully be included in the project). The inclusion of current reindeer as a reference and the implementation of new morphometric tools will bring decisive progress on this issue.
The goals of the project have been successfully achieved. Shape analysis using Geometric Morphometrics Methods (GMM) allowed an identification of migratory and sedentary mobility patterns using metacarpal distal cross-section. Furthermore, ML algorithms using linear measurements and including phalanges as well, allowed to infer mobility pattern and habitat with accuracies higher than 90%.
Regarding the second objective, a total of 20 Magdalenian sites have been analyzed. A successful classification of this sample according to mobility pattern (using both GMM and ML methods, providing both methodologies the same result, although the advantage of using ML is that allows a larger sample) and habitat type (with ML algorithms) has been possible.
Then, bone samples were scanned in the outgoing laboratory. Phalanges and osteometric approach through linear measurements were added to the shape study of the metapodial elements.
Cross sectional properties of the bone and geometric morphometric data were acquired from the scanned sections, and then statistically analyzed.
During this outgoing phase, the researcher received training in the use of the sPQTC CT scanner and related software and attended online courses training workshops offered externally and several conferences.
During the return period, linear measurements from archaeological materials were collected and metapodial bones and first phalanges were scanned to obtain cross- sections from 20 archaeological sites located in South-western France and hosted in different museums and institutions. Then, data sample (almost 900 bone remains ) was processed and analyzed and successfully classified according to mobility pattern and habitat.
The present study provides a novel methodology to test mobility patterns (migratory and sedentary) according to metapodial bone shape GMM and osteometric data from phalanges as well in combination with ML algorithms. Thus, the predictive methodology developed in the present study using machine learning algorithms, shows that it is possible to classify migratory and sedentary ecotypes with greater than 90% accuracy, depending on the skeletal part .
Regarding shape analysis (GMM), metacarpal 20% section has been key since an oval-shaped section belongs to caribou from migratory settings and more circular shape in this section corresponds to caribou from sedentary pattern.
The results and new methods could be exploited and transposed through their application to other archaeological sites and periods in which the biocenoses are also dominated by reindeer
Several training courses were also carried out to achieve the goals of the project. Project outcomes and achievements have been disseminated to both specialized and non-specialized audiences through the participation in several international congresses, webinars, interviews, and other events. Four scientific articles have also been generated and will be published throughout 2021 and 2022
The results from the outgoing phase indicate that the geometric morphometrics data and geometric properties of metapodia and phalanges allow us to distinguish between animals living in tundra environments, those living in the boreal forest and, to a lesser extent, those living in montane regions. Since those animals which move medium and long distances inhabit mainly open habitats (tundra), while those which migrate shorter distances tend to live in closed habitats (forest) we conclude that it is possible to infer the degree of mobility from the morphology of the distal limb bones. However, those reindeer that perform altitudinal migrations occur in a wide range of close and open environments. Furthermore, shape changes are linked to mobility and are mainly distinguishable in metacarpals.
Since the 1950s, the extent of reindeer migrations has been a recurring question in prehistoric archaeology. As the Palaeolithic environments in Western Europe have no current analogues it is not possible to infer reindeer mobility patterns from palaeoenvironmental reconstructions, the present study, therefore, provides a novel methodology to test mobility patterns.
Thus, the practical application of this methodology in the return phase on archaeological materials has been a breakthrough, making a decisive progress on this issue. The first results (2D GMM and ML) obtained are indeed very encouraging and will allow us to reconstruct reindeer movements, examine how they affected human hunting strategies and thus, understand the precise role of reindeer in social and economic organisation of Magdalenian populations during the last Glacial period in Southwestern France.
Bibliography:
DeGusta, D., Vrba, E. 2003. A method for inferring paleohabitats from the functional morphology of bovid astragali. Journal of Archaeological Science 30: 1009-1022.