Did a Proto-Shear Plane Cause Anak Krakatau's 2018 Collapse? Insights from Direct Shear Experiments and Numerical Models
To analyze the conditions leading up to the 2018 flank collapse at Anak Krakatau, Fiene Stopke—who is doing her PhD within the PRE-COLLAPSE project—combined direct shear experiments on volcanic ash and scoria with numerical models. The study, conducted in collaboration with colleagues, was just published in the Journal of Geophysical Research: Solid Earth. The results show that within the volcano there must have been a pre-existing "proto-shear plane" that led to the slow deformation of the volcano prior to the collapse. Triggered by an external factor such as an earthquake or dike intrusion, the flank began to move faster. This accelerated movement weakened the volcanic material within the proto-shear plane, ultimately leading to the collapse. The study highlights the importance of monitoring the submarine part of the volcano, as it is likely that sliding originates at the foot of the volcano. It emphazises that studying material properties and modelling weak planes within the volcano are crucial for assessing flank collapse potential and hazard.
Anak Krakatau
The collapse of Anak Krakatau in December 2018 is considered the best-monitored flank collapse to date. Satellite data showed gradual seaward flank sliding of the southwestern flank for several years before the event and extensive volcanic activity was observed. Even before the 2018 collapse, researchers had concerns regarding the stability of the edifice because of the steep slopes and, in particular, its growth towards the southwest, where the volcano is built on the caldera margin.
A zone of weakness within the volcanic edifice had been postulated by researchers before, but its existence remained unproven. This weakness zone might be related to a tectonic normal fault in the basement beneath the volcano or to a boundary between different types of volcanic deposits that formed as the eruption style changed. When the volcano was still underwater, its eruptions were submarine, producing layers of volcanic ash. As the volcano grew and emerged above water, its eruption style shifted, with lava and scoria becoming the dominant deposits. The proto shear plane, as they call it, may have been further weakened by hydrothermal alteration. It is highly probable that this was the surface along which the 2018 flank collapse occurred. However, the event itself could not be directly observed due to sparse satellite data at the time resulting from cloud coverage, and the area on which failure occurred was quickly covered by new material.
Simulating slow and fast flank movement in the lab
Ash and Scoria samples from Anak Krakatau, were collected from outcrops on land in 2017, before the major collapse occurred. In the lab, the samples were powdered to 125 micrometres. This replicated the “fault gounge” material representing the proto shear plane due to long-term friction. In the four years leading up to the 2018 collapse, the southwest flank of Anak Krakatau was moving at an average rate of 27 cm per year. To replicate this in the lab, Fiene Stoepke conducted shear experiments using a direct shear apparatus to understand the frictional behaviour of the volcanic rocks. She subjected the powdered rock samples to a vertical pressure and moved them horizontally at the same slow velocity observed in nature, scaled down to micrometres per hour. After a while, she increased the sliding speed, simulating a sudden increase in flank sliding velocity which could ultimately lead to the collapse, as in nature an earthquake or dike intrusion would have done.
Her results show that the volcanic material becomes weaker the faster the movement. This "velocity-weakening" behaviour means that once sliding begins, the rock material offers less resistance, and its frictional behaviour changes, making it more likely that the entire system will destabilize and fail—just as it did in 2018.
Direct shear experiments on volcanic rocks are rare, making this study particularly noteworthy.
What's truly ground-breaking is that Fiene Stoepke and her team, through their numerical modelling, have provided the first definitive evidence that the proto-shear plane was absolutely essential for the deformation of the volcano before the 2018 collapse. Their models could only reproduce the observed deformation pattern if the proto shear plane was present and the geometry of the edifice alone is not sufficient for the flank to slide seaward.
The failure started at the submarine foot of the volcano
When analysing the Factor of Safety (FoS), which indicates the stability of the volcano's parts, the models revealed that the entire volcanic edifice exhibited potential zones of weakness in its upper parts and flanks. However, a critical insight emerged when examining the FoS specifically along the postulated proto-shear plane: the lowest FoS values (indicating the highest instability) were found at the westernmost part of this plane, which corresponds to the submarine part of the volcano. This distinct pattern leads to a crucial interpretation: if the volcano's flank collapses along this proto-shear plane, it likely occurs retrogressively - the failure most likely initiates at the submarine part of the volcano and then progresses upwards.
This finding has significant implications for future monitoring efforts. To identify and monitor the earliest signs of impending collapse, which is essential for timely warnings, we must extend our monitoring efforts to include the submarine parts of the volcano.
What the researchers demonstrate with their study are the boundary conditions that must have existed at the time of the collapse—such as the presence of the proto-shear plane and its maximum possible dip. However, even if the material within this zone had been hydrothermally altered, its presence alone would not have been enough to cause the collapse. An external trigger, such as an earthquake or dike intrusion, was still required.
Geological and structural map of the Sunda Strait (Indonesia), with Anak Krakatau (red triangle). (after Stoepke et al., 2025)
(a) Overview of the greater tectonic setting.
(b) Overview of the Sunda Strait. Fault locations after Dahren et al. (2012).
(c) Anak Krakatau before and after the collapse in 2018. Note that the two images were shot from different directions (modified after Heidarzadeh et al., 2020).
Shear apparatus. (Fiene Stoepke)
The shear experiments were done in collaboration with Matt Ikari at the Experimental Geomechanics lab, Marum Bremen
Example of a friction-displacement curve for powdered ash exhibiting velocity-weakening behavior (a−b < 0). Left inset shows a schematic representation of the laboratory setup, original direct shear apparatus with sample, see picture to the left. Right inset shows a close-up view of the velocity step as indicated by the black box. (after Stoepke et al., 2025)
One-sided flank sliding occurs only with a weakness zone, suggesting Anak Krakatau deformed along such a proto-shear plane before collapse.
Displacement pattern of the volcanic edifice. The two-dimensional plane-strain finite-element model geometry represents Anak Krakatau prior to its 2018 sector collapse. (after Stoepke et al., 2025)
R1 represents an intact volcanic edifice (i.e. an edifice containing only the material boundary between the two lithologies)
R2 implements, in addition to the material boundary, a linear pre-defined proto shear plane as a frictional contact cutting through the different lithologies of the volcano.
Deformation and Gravitational Instability at Anak Krakatau (Sunda Strait, Indonesia): Insights From Direct Shear Experiments and Finite-Element Models
F. Stoepke, M. J. Ikari, A. Hampel, K. Meredew, S. Watt, M. Cassidy, M. Urlaub
JGR Solid Earth, published May 2025
