Reconstructing eruptive source parameters from tephra deposit: a numerical approach for medium-sized explosive eruptions

TL;DR

This study evaluates the accuracy of standard tephra deposit reconstruction techniques using a numerical simulation of a medium-sized eruption at Mt. Etna, revealing systematic underestimations and proposing a new correction method.

Contribution

It introduces a numerical approach to assess and improve the accuracy of tephra deposit reconstruction methods for medium-sized eruptions.

Findings

All techniques tend to underestimate the total erupted mass.

Grain size distribution information is significantly degraded in reconstructions.

A new method estimates the mass deficiency for each grain size class.

Abstract

Since the seventies, several reconstruction techniques have been proposed, and are currently used, to extrapolate and quantify eruptive parameters from sampled deposit datasets. Discrete numbers of tephra ground loadings or stratigraphic records are usually processed to estimate source eruptive values. Reconstruction techniques like Pyle, Power law and Weibull are adopted as standard to quantify the erupted mass (or volume) whereas Voronoi for reconstructing the granulometry. Reconstructed values can be affected by large uncertainty due to complexities occurring within the atmospheric dispersion and deposition of volcanic particles. Here we want to quantify the sensitivity of reconstruction techniques, and to quantify how much estimated values of mass and grain size differ from emitted and deposited ones. We adopted a numerical approach simulating with a dispersal code a mild explosive event occurring at Mt. Etna, with eruptive parameters similar to those estimated for eruptions occurred in the last decade. Then we created a synthetic deposit by integrating the mass on the ground computed by the model over the computational domain (>50000 km2). Multiple samplings of the simulated deposit are used for generating a large dataset of sampling tests afterwards processed with standard reconstruction techniques. Results are then compared and evaluated through a statistical analysis, based on 2000 sampling tests of 100 samplings points. On average, all the used techniques underestimate deposited and emitted mass. A similar analysis, carried on Voronoi results, shows that information on the total grain size distribution is strongly deteriorated. Here we present a new method allowing an estimate of the deficiency in deposited mass for each simulated class. Finally a sensitivity study on eruptive parameters is presented in order to generalize our results to a wider range of eruptive conditions.