Interactions between eruptive activity and sector collapse at Anak Krakatau, Indonesia

Anak Krakatau volcano, Indonesia, collapsed catastrophically on 22nd December 2018, forming a landslide-generated tsunami that caused over 400 deaths on surrounding coastlines. Very few volcanic landslides of this size and type, known as sector collapses, have been studied in detail. Because of this, our understanding of the factors that lead to sector collapse, and therefore our capacity to forecast such events and their associated hazards, remains relatively limited. Although there have been few historical examples of large volcanic landslides, they are common on longer, geological timescales and occur across all volcanic settings. The collapse of Anak Krakatau thus provides an important opportunity to improve our knowledge of this fundamental volcanic process.

Anak Krakatau is the volcanic island that formed after the devastating eruption of Krakatau (also known as Krakatoa) in 1883, first emerging above sea level in 1929. Approximately half of the subaerial island of Anak Krakatau was removed by its recent sector collapse. The collapse occurred during an ongoing, relatively low-intensity eruption, similar to the type of activity that had characterised previous decades. However, satellite observations suggest that this style of activity changed around the time of the collapse, to a much more powerful explosive eruption. The precise timing of this change, and its potential role in the collapse, is something we will explore in detail in this research. Following the collapse, explosive activity continued and may have changed yet again, as seawater interacted with shallow erupting magma. This later stage of activity erupted large volumes of new material, rapidly filling the landslide scar and extending the island coastline in the days after the collapse.

Our research will determine the specific role of eruptive activity in the sector collapse of Anak Krakatau. We will address whether changes in eruption behaviour, involving the ascent of fresh magma, preceded the collapse and thus acted as a trigger; or whether it was the collapse itself which led to the powerful explosive eruption, by suddenly depressurising the shallow magma stored beneath the volcano. We will also define the nature of eruptive activity that took place immediately after the collapse. In this phase of the eruption, material appears to have been ejected at a very high rate, and we will test the hypothesis that the collapse destabilised the underlying magma system, leading to a change in eruption behaviour. Such processes may be common at volcanoes affected by large sector collapses, forming part of a cycle of destruction and regrowth, but are currently poorly understood.

Our work will draw upon detailed field sampling of eruption deposits spanning the collapse period. Field datasets will be interpreted alongside satellite imagery and other remote observations, numerical models that simulate eruption processes, and analyses of the chemical and textural record of magmatic processes preserved in our eruption-deposit samples. Together, our results will allow us to identify changes in the storage conditions, ascent rate and eruptive behaviour of magmas involved in different stages of activity. Our results will allow us to explore controls on the timing of the sector collapse, the role of eruptive activity in the collapse, and the impact the collapse itself had on the underlying magma system. By producing a comprehensive record of the Anak Krakatau collapse and eruptions, we will advance our understanding of volcanic sector collapses in general. We will also develop a much clearer picture of eruption processes and instabilities at Anak Krakatau, which will inform hazard mitigation plans for potential future landslides as the volcano regrows.

This research will provide a comprehensive understanding of the volcanic processes involved in the sector collapse of Anak Krakatau, including the role of eruptive activity in the build up to collapse, the nature of explosive activity that accompanied the collapse, the impact of collapse on the underlying magma system, and the nature of the volcanism that led to rapid post-collapse regrowth of the island. As well as providing significant insights into eruption-associated sector collapses in general, these results have a broad importance to non-academic stakeholders including Indonesian hazard monitoring agencies and those working in the field of tsunami modelling and tsunami hazard mitigation.

Hazard monitoring agencies: Results, including technical summaries, will be shared fully with the appropriate Indonesian agencies responsible for volcanic and tsunami hazards: the Centre for Volcanology and Geological Hazard Mitigation (CVGHM), and the BMKG (the Meteorology, Climatology and Geophysical Agency, responsible for tsunami warning). We have already been in contact with relevant stakeholders to discuss our research plans, and these relationships will be developed further through pre-fieldwork meetings with these agencies. The unforeseen nature of the Anak Krakatau collapse highlights the unpredictability of sector collapses and the challenges involved in mitigating their effects. The controls on sector collapse remain poorly understood, but our detailed analysis of the event will identify the role of eruptive activity in the collapse, and will also provide a clearer understanding of post-collapse regrowth at Anak Krakatau, enabling future activity and growth to be monitored in the context of recent events. Our evaluation of a variety of remote sensing data will also help directly inform future developments of hazard monitoring at Krakatau. Collectively, our results will provide a much improved basis for forecasting future instabilities, and for developing landslide-tsunami monitoring efforts and mitigation plans on that basis.

Tsunami modellers: More generally, our results will form important additions to a comprehensive landslide-source and tsunami dataset for the Anak Krakatau event, making it the best constrained example of a large scale volcanic-island tsunami. This example can be used as a benchmark dataset against which tsunami models can be tested and improved, and then used for future landslide-tsunami hazard modelling scenarios. This is important to tsunami modellers in both academic and non-academic contexts, including those involved in coastal flood risk management and tsunami preparation in all volcanic settings, and the reinsurance sector. Results will be shared with international tsunami modelling groups (via active collaborations with S. Grilli, U. Rhode Island) and by providing a volcanological summary of the event to relevant stakeholders (e.g. to the quarterly technical meeting of the Intergovernmental Oceanographic Commission).

Wider public engagement: Our work is also likely to be of wide public interest, particularly given the effects of the Anak Krakatau collapse and its international media coverage, as well as broader awareness of Krakatau as a source of hazardous volcanic eruptions. One benefit of this type of research is increased awareness, education and engagement with scientific topics. We will distribute results through public talks, community events, museum exhibits and will also aim to publicise our research through local and national media.