Resumption of the sustained column during the post-Plinian phase of the 79 AD Vesuvius eruption

Plinian eruptions are particularly dangerous eruptions that often have longlasting consequences. They produce high eruptive columns that inject considerable volumes of pyroclastic fragments and volcanic gases into the stratosphere. Pyroclastic clasts fall back to the ground forming widespread tephra sheets in which different components are distinguished: juvenile clasts, lithic clasts and crystals. This ‘classical’ behaviour is well represented by the main sustained phase of the 79 AD Vesuvius eruption, which deposited a thick white to gray pumice lapilli fallout deposit. This phase was followed by a column collapse phase, which mainly emplaced several pyroclastic density currents (PDCs). While tephra deposits of Plinian eruptions are wellconstrained in their sedimentological features and transport processes, little attention has been paid on the occasional fallout from postPlinian activity. One outstanding reason for this is the impermanent nature of these fall deposits, which are generally small in volume and eroded by successive PDCs in proximal and medial sectors and swept away especially in their distal portions, due to the combined effect of rill and wind erosion, soil creep and other forms of mass wasting. Because of this, hazard assessment is inherently incomplete. Here, we present stratigraphic and volcanological evidences for the existence of five lithicrich lapilli fallout layers interstratified with the pyroclastic density current deposits emplaced after the collapse of the 79 AD Plinian column. These late fall products are distributed south of the Vesuvius, at distances between 5 and 20 km from the vent (from Mt. Somma slopes to the Mt. Lattari). These lithicrich horizons, named D, G1, G3, I and X2 from base to top, exhibit mantling structures, are massive and generally well sorted (sorting = 1.051.55). The Md Φ ranges from 8 mm (unit G1) to almost 0.5 mm (unit G3). Component analysis shows that layer D has 61 wt.% of lithics and 23 wt.% of juvenile clasts, while the upper layers have about 7083 wt.% of lithics to less than 0.50 wt.% of juvenile material. The remaining part is made up of crystals and not separated fine ash. A number of outcrops were identified to reconstruct the distribution for the thicker units D and G1. We estimated the plume height by apply the method of Wilson and Walker [1987] to the lithics isopleth maps. We prefer to use lithic isopleths rather than pumice isopleths due to fragile behaviour of the pumice clasts, which could lead to a less reliable result. When more than one isopleth is used, Ht is the average value, and the error represents the difference between the maximum and minimum. These data highlight the oscillating behaviour of the 79 AD eruption plume during the postPlinian fall phase. During the emplacement of layer D, the column rose about 17 km. A second sustained column, which rose to 19 km, occurred during the emplacement of layer G1. Both distributions show a SE trend even if the azimuth of the dispersal axis of layer G1 seems to be rotated by 1020° relative to that of D. Erosion acting preferentially on the upper part of the succession prevented the isopleths of G3, I and X2 from being defined. The dispersion of the lithicrich layers is similar to that of the basal lapilli Plinian deposit, including rotation of the dispersal axis. Our study suggests that the resumption of a sustained column was repeatedly established during the postPlinian phase of the 79 AD Vesuvius eruption. The main difference with the basal pumice lapilli deposit is the strong enrichment in lithic clasts, possibly associated with a recurrent instability in the conduitvent system.