A two‐layer depth‐averaged model for both the dilute and the concentrated parts of pyroclastic currents

Abstract

Pyroclastic currents are very destructive and their complex behavior makes the related hazards difficult to predict. A new numerical model has been developed to simulate the emplacement of both the concentrated and the dilute parts of pyroclastic currents using two coupled depth‐averaged approaches. Interaction laws allow the concentrated current (pyroclastic flow) to generate a dilute current (pyroclastic surge) and, inversely, the dilute current to form a concentrated current or a deposit. The density of the concentrated current is assumed to be constant during emplacement, whereas the density of the dilute current changes depending on the particle supply from the concentrated current and the mass lost through sedimentation. The model is explored theoretically using simplified geometries as proxies for natural source conditions and topographies. It reproduces the relationships observed in the field between the surge genesis and the topography: the increase in surge production in constricted valleys, the decoupling between the concentrated and the dilute currents, and the formation of surge‐derived concentrated flows. The strong nonlinear link between the surge genesis and the velocity of the concentrated flow beneath it could explain the sudden occurrence of powerful and destructive surges and the difficulty of predicting this occurrence. A companion paper compares the results of the model with the field data for the eruption of Merapi in 2010 and demonstrates that the approach is able to reproduce the natural emplacement of the concentrated and the dilute pyroclastic currents studied with good accuracy.

The Relative Effectiveness of Empirical and Physical Models for Simulating the Dense Undercurrent of Pyroclastic Flows under Different Emplacement Conditions

“…None of the interplay between dense-basal undercurrent and ash-cloud surge (e.g., Breard et al, 2016; Breard and Lube, 2017) can be captured. Creating coupled two-phase flow models with an ash-cloud surge component and simple sedimentation laws is an active area of development for both TITAN2D (Pitman et al, 2013;Sheridan et al, 2013) and VolcFlow (Kelfoun, 2017).…”

Simulation of block‐and‐ash flows and ash‐cloud surges of the 2010 eruption of Merapi volcano with a two‐layer model

“…A low value of ρ m forms a thick, low density surge, while a high value forms a thin, denser surge. However, the value of ρ m has no influence on the mass of the surge, and consequently, it has little influence on the surge dynamics [ Kelfoun , ], except where the surge thickness is close to the scale of the topographic relief. The value of ρ m can be estimated by the elevation reached by the surge along the hills and by the capacity of the surge to flow out of the drainage basins and spread out on the interfluves.…”

“…We show that the model reproduces the main characteristics of the real phenomenon. This new model gives promising perspectives for the understanding of pyroclastic current emplacements and for future estimation of related hazards and impacts on the population, the infrastructure, and the environment.A new numerical model that simulates both the concentrated and the dilute parts of PDCs and their interactions is developed and explored theoretically in the companion paper [Kelfoun, 2017]. The model is compared here with a natural field case to check its ability to reproduce the natural phenomenon.…”

Transferability of a calibrated numerical model of rock avalanche run‐out: Application to 20 rock avalanches on the Nuussuaq Peninsula, West Greenland

“…The rock avalanche motion is modelled using the geophysical mass flow code VolcFlow (Kelfoun and Druitt, 2005). VolcFlow has been tested on a number of debris avalanches and pyroclastic flow events, successfully simulating avalanche run-out and emplacement dynamics in a number of settings (Kelfoun and Druitt, 2005;Kelfoun et al, 2008Kelfoun et al, , 2009Kelfoun et al, , 2010Kelfoun et al, , 2017Giachetti et al, 2011;Kelfoun, 2011Kelfoun, , 2017Paris et al, 2011;Charbonnier and Gertisser, 2012;Dondin et al, 2012;Giachetti et al, 2012Giachetti et al, , 2013Manzella et al, 2016). As with many other continuum dynamic models used for simulating rock avalanche propagation (Savage and Hutter, 1989;Iverson et al, 1997;McDougall and Hungr, 2004), VolcFlow is governed by a continuity and momentum equation based on the Saint-Venant equations of shallow flow.…”