Reconstruction of the 2018 tsunamigenic flank collapse and eruptive activity at Anak Krakatau based on eyewitness reports, seismo-acoustic and satellite observations

TitleReconstruction of the 2018 tsunamigenic flank collapse and eruptive activity at Anak Krakatau based on eyewitness reports, seismo-acoustic and satellite observations
Publication TypeJournal Article
Year of Publication2020
AuthorsPerttu A, Caudron C, Assink JD, Metz D, Tailpied D, Perttu B, Hibert C, Nurfiani D, Pilger C, Muzli, Fee D, Anderson OL, Taisne B
JournalEarth and Planetary Science Letters
Volume541
Date Published07/2020
Abstract

After several months of eruptive activity, the subaerial cone of Anak Krakatau collapsed on December 22, 2018. The landslide event generated a tsunami that had deadly consequences within the Sunda Strait in Indonesia. Such significant collapse events are common in the geologic record but are a rare phenomenon, in the instrumented record. However, these events can have a potentially large impact on society. We have reconstructed the collapse, along with the activity preceding and following it, by combining information from official reports, remote geophysical observations, and local eyewitness accounts. It appears that the collapse of Anak Krakatau's subaerial cone led to a drastic change in the eruptive style from continuous Strombolian explosions to sustained Surtseyan. Those changes are detectable in the seismo-acoustic measurements, which, when combined with eyewitnesses, allows us to reconstruct the timing and phenomenology of the sequence. Our analysis reveals that intense eruptive activity generated sustained infrasound, unusual but not unique at Anak Krakatau, starting approximately eight hours before the collapse. Within this timeframe, two seismic signals consistent with minor mass movements as well as a momentary quiescence were identified prior to the main collapse. The data presented here indicate that Anak Krakatau failed in one collapse event, producing a tsunami with multiple waves around the volcano, the last one being the largest. Following the collapse, three volcanic plumes could be clearly identified in the satellite data and by eyewitnesses, as well as spectral lines in the seismic data. These lines, observed up to 371 km, suggest a repeating energetic explosive source lasting for seven days. The collapse produced multiple infrasound arrivals observed at regional infrasound stations, but was not recorded on the regional hydroacoustic network. Our analysis of the eruptive sequence demonstrates that a detailed eruption chronology can be reconstructed using remote methods, even in the event of failure or destruction of local monitoring infrastructure. This event also highlights that tsunamigenic flank collapses can occur with little to no warning, and be difficult to interpret in real-time, as a significant amount of non-operational analysis was required after the event, to complete the chronology.