A detailed review of global databases of volcanic edifices such as the Global Volcanism Program (GVP), and those of known mass movements quickly revealed the need for a separate database to compile information on volcanic islands. Out of the 2652 Holocene and Pleistocene volcanoes in the GVP database, we identified 370 edifices as volcanic islands. However, we were only able to find high-resolution bathymetry (<250m) of sufficient coverage of their submerged flanks for 47 of them. We therefore decided to evaluate how the use of a Digital Elevation Model (DEM) reliant on low-resolution, indirect satellite measurements where no shipborne data is available, will influence the quality of the geomorphometric parameters. The result was positive and most morphometric parameters show robust results.
We conducted a comprehensive review of historical marine volcano collapse events, focusing on the available event sequences during the collapses. Despite the majority of documented marine volcano collapses occurring during volcanic eruptions, the analysis revealed that collapses do not always follow eruptions, supporting the need to physically and quantitatively describe the processes leading to collapses.
In order to understand the generation of the slow sliding we created a static two-dimensional plain-strain finite-element model representing the volcano Anak Krakatau. The model geometry draws on pre- and post-collapse drone photogrammetry and satellite images. We find that a weak zone inside the volcano is needed to generate a slow sliding flank and to reproduce ground deformation observed by satellites prior to the collapse in December 2018. To understand mechanical properties of such a weak zone we conducted direct shear experiments were conducted on samples of the pre-collapse Anak Krakatau. The results show that all lithologies display velocity-weakening behavior independent of the applied normal load or fluid saturation. This is a prerequisite for a fault being unstable and capable of runaway acceleration.
We developed a 2D boundary element model for dike propagation while accounting for the topography of the volcanic edifice. We implemented a discretized, traction-free surface using dislocation elements within an existing 2D Boundary Element model designed for simulating the propagation of fluid-filled cracks. The numerical experiments show that high topography contrasts and asymmetric edifice geometries, as is often the case for marine volcanoes, have significant influences on calculated ground deformation.
A short research cruise for conducting marine research at Krakatau volcano was applied for and granted. The preparation phase required close exchange and collaboration with Indonesian scientists and authorities.
The main result of all work done in the reporting period is that there appears to be a strong link between flank collapses and eruptions. We find increasing evidence for that flank collapses precede large eruptions and not vice versa. This would exclude exceptionally large eruptions as triggers for flank collapse.
