Search

We examine records obtained in the regional-to-far field ofthe catastrophic volcanic landslide and tsunami atKrakatau on 22 December 2018. In the absence of operationalDART-type sensors in that part in the Indian Ocean, we userecords on horizontal long-period seismometers at a number ofcoastal sites: this methodology, pioneered by Yuan et al. [2005]has been successfully applied to several tsunamis, both in the far fieldand more recently at close distance during the 2017 Karrat Fjord, Greenlandlandslide. Using records at Cocos Island (COCO; 1160 km), we reconstructan equivalent amplitude in deep water of 0.15 mm peak-to-peak,with energy peaked around 2.5 mHz.We fail to observe hydroacoustic T phases on hydrophones of the IMSoff Diego Garcia, but a weak signal is present on the seismic records atCOCO and at one of the two stations of the Australian network on ChristmasIsland (490 km), the latter featuring energy in the 7-t0-20 Hz band.Finally, the eruption is recorded as an exceptionally crisp signal by the infrasound arrayI06 of the IMS at Cocos Islands, with a signal peaked around 0.25 Hz.Quantified modeling of these various signals will be presented, based inparticular on a combined dynamic model of the landslide and tsunami, coupled toa simulation of the water waves using a Boussinesq approximation, whoseresults will be compared to the deconvolved seismic record at COCO. [ABSTRACT FROM AUTHOR]

Multiple sets of events with high waveform similarity (multiplets) have often occurred prior to different types of catastrophic failure events, e.g. large earthquakes, landslides and volcanic eruptions. This similarity indicates the source locations are restricted to a very localised zone. Hence, analysing multiplets has the potential to reveal underlying processes such as the nucleation of large events, or stable repeated slip, and hence could improve probabilistic forecasts of the likelihood of different catastrophic events. By resolving the processes taking place, we can determine their relationship to the catastrophic failure event as foreshocks, aftershocks or independent processes. The problem with finding multiplets is that they are often small, obscured in ambient noise, and sometimes only picked up by one seismometer. We have developed an optimised detection and analysis technique to extract a catalogue of multiplets and determine their temporal evolution in different seismic signals. By enhancing the short-time-average/long-time-average (STA/LTA) approach for finding events with a moving cross-correlation window, we discover new events automatically in a dedicated high Random-Access Memory computational engine. Subsequent temporal tracking and statistical analysis of these events then allows us to examine their occurrence and resolve the processes taking place. The algorithm's success in finding events is evaluated statistically with hits, false alarms and misses, analogous to clinical trials in medicine. Our method works significantly more successfully than the STA/LTA approach on its own, with more hits and more accurate pick times in different catastrophic failure settings. We demonstrate the success of our algorithm in real cases of past earthquake sequences and other catastrophic failure type events. From this, we can identify large earthquakes which have had (or not had) nucleating foreshocks in the form of multiplets, aiding to our understanding of the nucleation process. We can also use this technique for other types of catastrophic failure type events, such as volcanic eruptions, to determine the underlying processes taking place based on the evolution of multiplets prior to failure. Our current analysis on large earthquakes show no clear nucleation pattern between different cases; however, we do obtain a more complete view of the pre-mainshock seismicity taking place, helping to better understand pre-failure processes. [ABSTRACT FROM AUTHOR]

The history of seismography in Iceland began in 1909 with the installation of one horizontal Mainka seismograph in Reykjavík. Following a period of intermittent operation regular operation was initiated in1925 with the establishment of the Icelandic Meteorological Office. The number of stations increased gradually over the following decades, and in the sixties four stations were in operation. The number of permanent stations proliferated following the Heimaey eruption in 1973, and during most of the eighties, the number of stations was 40-50. The first digital seismograph stations were installed in 1990 and the analog seismic network was gradually replaced by digital stations over the next two decades. The number of analog seismograms in the Icelandic archives is estimated to be around 300 000 covering the period 1910-2010. A four-year project to make this record collection accessible on the internet has been initiated and funded. So far around 100 000 seismograms have been scanned and made available for download on the open website https://seismis.hi.is.This database will eventually be merged with a general Icelandic science data repository currently being established by the University of Iceland and various research institutes.The seismograms are scanned with a resolution of 300 dpi and presented for examination on the website as low resolution.png files but high-resolution jpg files can also be selected for download. The high-resolution files are on the order of 4-8 Mb each. Digitization of the seismic traces has not been attempted, since most of the seismograms are from short-period instruments and the waveforms are already lost.In addition to numerous teleseismic body wave phases the record collection contains primary data from various tectonic and magmatic events in Iceland during the last century. This includes eruptions of Hekla in 1947, 1970, 1980-81, 1991 and 2000, Surtsey in 1963-67, Heimaey in 1973, Askja in 1961, Grímsvötn in 1934, 1983, 1998, and 2004, Gjálp in 1996, rifting episode at Krafla in 1975-1984, persistent seismic activity of the Bárðarbunga and Katla volcanoes, numerous suspected subglacial magmatic events, earthquake swarms on the Reykjanes Peninsula Oblique Rift and within the Tjörnes Fracture Zone, and earthquake sequences in the transform zones of South and North Iceland. [ABSTRACT FROM AUTHOR]

A sequence of eruptions of the Kilauea volcano, Hawaii commenced in early May 2018 and caused serious damage and residential evacuation. To understand the magma intrusion process, which produces little measurable effect at the surface prior to the eruption, we use seismic noise interferometry and ~1.5-year vertical-component seismic data from 6 broadband seismometers to investigate temporal seismic velocity changes (dv/v) at depth near Kilauea. Analyses of three different frequency bands (0.3-0.6 Hz, 0.6-0.9 Hz, and 0.9-2.0 Hz) show distinct earthquake-related and magma-related structural changes in the uppermost 500 m and below 1 km from the surface. The magma-related dv/v began in December 2017 and exhibit a dichotomous pattern of seismic velocity that increases to the southwest and slightly decreases to the northeast of the East Rift Zone, revealing an intrusive magma process with a long precursory phase beginning almost a half year before the May 2018 eruption. [ABSTRACT FROM AUTHOR]

The formation of the oceanic island Surtsey in the shallow ocean off the south coast of Iceland in 1963-1967 remains one of the best-studied examples of basaltic emergent volcanism to date. Cycles of the phreatomagmatic explosive activity occurred several times during the eruptions: (1) At the onset of the eruption - eruption of Surtur crater, November 1963-January 1964; (2) The submarine eruption of Surtla - December 1963; (3) the formation of the parasitic cone of Surtungur - January-April 1964; (4) formation of the temporary island of Syrtlingur - May-October 1965; and (5) the formation of the temporary island of Jólnir - December 1965-August 1966. Explosive activities at these vents are well-documented and preserved in photographic records and reports. The first submarine volcanic activity was detected on 14th of November 1963 where ocean depth was 130 m prior to the eruption. The onset time of the submarine eruption was not resolved in contemporary records. Birth of the submarine eruption at the seafloor can be detected by investigation of tremor signals or any unusual 0seismic activity at nearby seismograms. Seismic records from the time of the Surtsey eruption are available from two permanent seismic stations in south Iceland: Reykjavík in southwest Iceland and Kirkjubæjarklaustur in southeast Iceland. These records have now been re-analyzed. Any potential eruption signals are drowned by microseisms at the Reykjavík station while at the Kirkjubæjarklaustur station eruption tremor signals appeared before the first explosive eruption had broken through the sea floor. Earthquake activity related to the beginning of the eruption was weak, hardly detectable by the existing seismograph stations.These observations indicate that the actual start of the submarine eruption occurred no later than on 13th of November, with eruption tremors recorded 18-24 hours before the first visual observation of eruption from a fishing boat in the early morning of the 14th of November. Similar tremors were also detected during other four explosive phases. The potential of seismic records to detect the onset of eruptions is demonstrated through these results. Based on the shreds of evidence for the onset time and the visual observations, maps and descriptions, we derive the early growth history of Surtsey. When combined with bathymetric data prior to the eruption, we obtain constraints on the rate of submarine cone buildup. This places constraints on the eruption rate in these early submarine-to-emergent phases of the eruption. Early results indicate that the maximum mass eruption rate in the beginning phase was relatively modest, below 1,000,000 kg/s. [ABSTRACT FROM AUTHOR]