Thermal interactions of the AD79 Vesuvius pyroclastic density currents and their deposits at Villa dei Papiri (Herculaneum archaeological site, Italy)

Pyroclastic density currents (PDCs) can have devastating impacts on urban settlements, due to their dynamic pressure and high temperatures. Our degree of understanding of the interplay between these hot currents and the affected infrastructures is thus fundamental not only to implement our strategies for risk reduction, but also to better understand PDC dynamics. We studied the temperature of emplacement of PDC deposits that destroyed and buried the Villa dei Papiri, an aristocratic Roman edifice located just outside the Herculaneum city, during the AD79 plinian eruption of Mt Vesuvius (Italy) by using the thermal remanent magnetization of embedded lithic clasts. The PDC deposits around and inside the Villa show substantial internal thermal disequilibrium. In areas affected by convective mixing with surface water or with collapsed walls, temperatures average at around 270 ◦ C (min 190 ◦ C, max 300 ◦ C). Where the deposits show no evidence of mixing with external material, the temperature is much higher, averaging at 350 ◦ C (min 300 ◦ C; max 440 ◦ C). Numerical simulations and comparison with temperatures retrieved at the very same sites from the reflectance of charcoal fragments indicate that such thermal disequilibrium can be maintained inside the PDC deposit for time-scales well over 24 hours, i.e. the acquisition time of deposit temperatures for common proxies. We reconstructed in detail the history of the progressive destruction and burial of Villa dei Papiri and infer that the rather homogeneous highest deposit temperatures (average 350◦C) were carried by the ash-sized fraction in thermal equilibrium with the fluid phase of the incoming PDCs. These temperatures can be lowered on short time- (less than hours) and length-scales (meters to tens of meters) only where convective mixing with external materials or fluids occurs. By contrast, where the Villa walls remained standing the thermal exchange was only conductive and very slow, i.e. negligible at 50 cm distance from contact after 24 hours. We then argue that the state of conservation of materials buried by PDC deposits largely depends on the style of the thermal interactions. Here we also suggest that PDC deposit temperatures are excellent proxies for the temperatures of basal parts of PDCs close to their depositional boundary layer. This general conclusion stresses the importance of mapping of deposit temperatures for the understanding of thermal processes associated with PDC flow dynamics and during their interaction with the affected environment.