Reconstructing human–landscape interactions in the ancient Mediterranean harbour of Cartagena (Spain)

Here, we sought to reconstruct the Pleistocene and Holocene evolution of Cartagena Bay. Therefore, 16 borehole cores were analysed with the following aims: (1) to define a chronological framework; (2) to obtain data on the palaeoenvironment; and (3) to establish the relationship with human activities, especially focused on the reconstruction of the Carthago Nova (also known as Qart Hadasht) conquest. A total of 147 samples were recovered for amino acid racemisation (AAR) dating; 32 for radiocarbon dating (14C); and 159 for sedimentological, palaeontological, and biomarker determination. These approaches allowed us to elucidate the evolutionary phases of the palaeolandscape in Cartagena. The simultaneous use of AAR and 14C dating allowed the discrimination of spurious ages and the establishment of a chronological scale. During the Marine Isotope Stages (MIS) 7 and 5, the sea level (SL) in Cartagena Bay was almost the same as today. An alluvial plain developed in the northern area, and a delta was formed by an ephemeral river. In contrast, during MIS4, after a fall in SL, a wide sandy coastal plain emerged, leading to the accumulation of colluvial and aeolian deposits. The Holocene transgression was reflected in the hinterland in a late and temporarily floodable marsh fed by alluvial fans and creeks, producing a salinity gradient from freshwater to brackish and saline waters. Along the sea front, the marine influence was evident but not dominant, grading up and landwards to saline and freshwater marshes. Therefore, at the time of Scipio’s conquest, the geography of Carthago Nova was quite different from that described by Polybius. The Roman’s sudden attack seized Qart Hadasht from the south-west, avoiding the marshy area and not crossing a fast and deep ebb stream as it never existed on the Mediterranean Iberian coast.

The Moyjil site, south-west Victoria, Australia: chronology

An unusual shell deposit at Moyjil (Point Ritchie), Warrnambool, in western Victoria, has previously been dated at 67±10 ka and has features suggesting a human origin. If human, the site would be one of Australia’s oldest, justifying a redetermination of age using amino acid racemisation (AAR) dating of Lunella undulata (syn. Turbo undulatus) opercula (the dominant shellfish present) and optically stimulated luminescence (OSL) of the host calcarenite. AAR dating of the shell bed and four Last Interglacial (LIG) beach deposits at Moyjil and Goose Lagoon, 30 km to the west, confirmed a LIG age. OSL analysis of the host sand revealed a complex mixing history, with a significant fraction (47%) of grains giving an early LIG age (120–125 ka) using a three-component mixing model. Shell deposition following the LIG sea-level maximum at 120–125 ka is consistent with stratigraphic evidence. A sand layer immediately below the shell deposit gave an age of ~240 ka (i.e. MIS 7) and appears to have been a source of older sand incorporated into the shell deposit. Younger ages (~60–80 ka) are due to bioturbation before calcrete finally sealed the deposit. Uranium/thorium methods were not applicable to L. undulata opercula or an otolith of the fish Argyrosomus hololepidotus because they failed to act as closed systems. A U–Th age of 103 ka for a calcrete sheet within the 240 ka sand indicates a later period of carbonate deposition. Calcium carbonate dripstone from a LIG wave-cut notch gave a U–Th age of 11–14 ka suggesting sediment cover created a cave-like environment at the notch at this time. The three dating techniques have collectively built a chronology spanning the periods before and after deposition of the shell bed, which occurred just after the LIG sea-level maximum (120–125 ka).