The Neapolitan area is one of the highest volcanic risk areas in the world, both for the presence of three potentially explosive and active volcanoes (Vesuvius, Campi Flegrei and Ischia), and for the extremely high exposure (over a million people located in a very large and important metropolitan area). Even though pyroclastic flows and lahars represent the most destructive phenomena near the volcanoes, tephra fall poses a serious threat on a wider spatial scale. Excess of tephra loading can cause building collapse, disrupt services and lifelines, and severely affect agriculture and human health. On a larger spatial scale, tephra fallout may cause a major disruption of the economy in Europe and in the Mediterranean area (Folch & Sulpizio, 2010, Sulpizio et al., 2012).

The volcanic hazard is the way in which scientists quantify such a kind of threat. The hazard is usually expressed in probabilistic terms in order to account for the vast irreducible (aleatory) and reducible (epistemic) uncertainties. In the past several papers focussed on the assessment of tephra fallout hazard from Neapolitan volcanoes (e.g. Barberi et al. (1990), Macedonio et al. (1990), Cioni et al. (2003), Costa et al. (2009)). These studies have combined field data of tephra deposits and numerical simulations of tephra dispersal (often considering tens of thousands of wind profiles to account for wind variability) to produce maps for the expected tephra loading in case of a specific scenario (e.g. considering one specific kind of eruption), or of a few reference scenarios at both Mount Vesuvius and Campi Flegrei.

This kind of map is still frequently used in volcanology, however, they do not represent the real volcanic hazard, because they do not consider the probability of occurrence of the specific scenarios considered, and they neglect a large part of the natural variability, such as the possibility to have eruptions of different size and from different vents. The latter is particularly important for the Campi Flegrei caldera, where the largest source of uncertainty comes from the forecast of the next eruption location. From a more technical point of view, these studies do not properly incorporate all known aleatory and epistemic uncertainties. This aspect is of primary importance in order to get a reliable volcanic hazard assessment.