Definition of a new approach in ancient harbor geoarchaeology: geochemistry and ostracod analyses at Portus (Tiber delta, central Italy)

Résumé En

Portus constituted the harbour complex of imperial Rome from the 1st century AD onward. Located in a deltaic environment, the harbour was subject to influences of the Tyrrhenian marine environment and of the fluvial waters of the Tiber River. Such complex sedimentary processes registered also rapid environ-mental changes. Our research is based on the study of a sediment core (TR14) drilled in the access channel of the hexagonal basin of Trajan (Fig. 1). The goal of our study is the reconstruction the palaeo-environmental dynamics of the water column of the harbour according to a geochemical and an ostra-codological approach. The analysis of the ostracod assemblage has evidenced several palaeo-environmental phases. These phases could be identified through the autoeco-logical analysis of the 25 recognised taxa and their classification in four eco-logical groups (MAZZINI et al., 2011). The results obtained present a succession of four environments, defined by variations in salinity and oxygenation. From the bottom to a depth of 643 cm, the environment was a marine to brackish lagoon influenced by strong marine inputs as evidenced by the vic-ariance of Cyprideis torosa and Pontocythere turbida. P. turbida is a typical phy-tal ostracod often found together with Posidonia oceanica remains. From 600 to 454 cm the coastal marine environment becomes more stable. This could be linked to the establishment of the harbour activities and maintenance. The sta-bility of the environment is still evident in the subsequent group, from 435 to 305 cm, characterised by the dominance of C. torosa and the lack of any fresh-water influence. The increase in salinity could be related to evaporative effects on a closed body of water. From 300 cm the freshwater input becomes clear and constant although the influence of the close sea/brackish body of water is still noticeable. In the last metres a typical freshwater assemblage is represent-ed, dominated by truly freshwater taxa (Candona sp. and Pseudocandona marchica), evidence for the occurrence of a freshwater marsh. The geochemical results were analysed by Principal Components Analy-sis (PCA). The samples were assembled in several groups based on a Hierar-chical Ascendant Classification (cluster analysis). Through PCA analysis, the geochemical and ostracological results of the "functional" harbour unit (800-300 cm) were combined. The different elements recorded in the geochemical analysis are distributed according to the main ostracod assemblages (from fresh water to brackish water and sea water). These elements have been grouped according to the model proposed by SAGEMAN & LYONS (2003). An evolution of these deposits in three stages is proposed. While the fluvial envi-ronments were quite well oxygenated, the harbour environment was more characterised by an increasing anoxia of the water column as it was observed by ELMALEH et al. (2012) in the harbor of Tyre. In a second phase, the varia-tions in salinity distinguished the nature of the fluvial environments and the degree of openness to the sea of the harbour environment. The final phase is specific to a body of water completely isolated from fluviatile and marine