Dynamics and deposits generated by the Kos Plateau Tuff eruption: Controls of basal particle loss on pyroclastic flow transport

The explosive eruption of voluminous silicic magmas often produces widespread and massive deposits formed from pyroclastic density currents. While these punctuated events dramatically alter the landscape and have potential climate‐altering impact, our understanding of the internal structure and transport dynamics of these eruptions is hampered by a lack of direct observations. We utilize the natural boundary conditions encountered by the eruption of the Kos Plateau Tuff to probe its internal structure as well as constrain the neotectonic activity in the region and eruption duration of this moderate to large (>60 km3) event. At the time of the eruption, 161 ka, the lower sea level in the Mediterranean may have resulted in flows that traversed mostly land to the north of the eruptive vent, while flows to the south may have encountered an expanse of water. Steep topography and overwater transport can impede the transport of the dense basal portions of the flow where particles make multiple or sustained contact with the bed. We use an Eulerian‐Eulerian‐Lagrangian computational approach coupled with overwater and overland boundary conditions, including topography, to determine the role of bed load versus suspended load in the transport of these flows. We find that a ring vent structure and eruptive fluxes greater than ∼2 × 106 m3/s are required for the spatial distribution of the KPT. The maximum grain size and deposit locations of the first voluminous ignimbrite unit (D) can be explained by suspended flow to the south, consistent with overwater transport, and bed load and suspended load transport to the north, consistent with overland transport. However, the maximum lithic size for the largest and last ignimbrite unit (E) requires some bed load transport in all directions. We propose that the boundary conditions were significantly altered during the course of the eruption, through either the in‐filling of a shallow sea to the south or the development of a thick pumice raft to aid saltation. On the basis of the inferred eruptive flux, we estimate that the duration of the eruption climax, in which most of the material was erupted, likely only lasted from a few hours to a day.