3D reconstruction of volcanic bombs to enhance ballistic trajectory predictions

Abstract

Volcanic Ballistic Projectiles (VBPs) ejected during explosive eruptions represent a significant hazard to people approaching the craters of active volcanoes. Recently, several models have been proposed to simulate the trajectories of VBPs, thus contributing to assessing the associated hazards. However, these models mostly rely on field data, which may not be readily available. Indeed, often the large size and weight of VBPs make their transportation to a laboratory for detailed physical measurements extremely challenging. This work compares different methodologies useful to extract 3D models of VBPs and measure their dimensions (length, width, and thickness) and volume. A VBP collected at Mt. Etna (Italy) after the 21 May 2023 eruption was used as a test sample. Specifically, we processed a series of images, acquired with a smartphone using two photogrammetric software, one freely available and one commercial. The 3D model of the sample was also obtained through a LiDAR sensor integrated into the smartphone. The dimensions and volume of the sample were measured using a caliper and Archimedes’ principle and were used as reference values for quantifying the accuracy of the different reconstruction methodologies. We found that the commercial software is well suited to estimate size and volume, providing a high-precision reconstruction even for the irregular shape of highly porous volcanic particles (with uncertainties of <6% and 16% for the mean diameter and volume, respectively). However, we suggest that, among all the analysed methods, the LiDAR offers a good compromise in terms of accuracy (with uncertainties of <12% and <86% or the mean diameter and volume, respectively) and ease of field operations. Finally, we estimated the ballistic trajectory of VBP using the diameters estimated from the different volumes finding that the distance reached by VBP can have variations up to 350 m only due to a different way to estimate the particle size. We thus conclude that the methods investigated in this work could be applied to different volcanoes worldwide to determine the 3D shape of VBPs because incorrectly estimating the size of VBPs can bias ballistic hazard models.