Assessing pyroclastic density current hazard in caldera settings: the example of Campi Flegrei caldera (Italy)

Abstract

Mapping of pyroclastic density currents (PDCs) hazard in caldera settings is particularly challenging due to the large uncertainty on future vent location and eruption scale as well as the complex dynamics of the flow over the irregular caldera topography. Nevertheless, probabilistic mapping of PDC invasion, able to account for the intrinsic uncertainties affecting the system, is needed for hazard assessment, particularly for highly populated regions. Campi Flegrei (CF) is a vivid example of active and densely populated caldera with a very high risk associated with the occurrence of PDCs produced by explosive events. In this presentation we show the results of new field work and mathematical modelling of past eruptive activity aimed at producing long-term probabilistic vent opening and PDC invasion maps at CF. Field work was focused on the structural and morphological features of the caldera and particularly on the reconstruction of the location of past eruptive vents as well as of the distribution of the main faults and fractures formed in the last 15 kyr of activity. One specific objective of the analysis was to incorporate into the vent opening maps the main uncertainties affecting the system by adopting appropriate density distributions and by relying on expert judgement. We then used these maps to produce, by adopting a Monte Carlo approach, a variety of probabilistic PDC hazard maps of the CF area based on different invasion models and accounting for the uncertainty on vent location and event scale. In particular we developed and adopted a simplified invasion flow model based on the so-called box model approximation and tuned on transient and 2D numerical simulations of the flow dynamics. The new model allowed to describe the exponential decay of the flow energy as well as to account for first-order topographic effects. The developed methodology appears able to provide relatively quick and robust probabilistic assessments of PDC hazard in caldera settings and could be potentially extended to other calderas worldwide.