A Study of Volcano Seismology: Forecast and Monitoring

Volcanic systems have always been regarded as a challenging issue for earth scientists and other academic staffs. Their natural behaviour and characteristic features are so complex to assess. Since volcanic activities are very often unpredicted and uncontrollable in some cases, whilst doing volcanic hazard assessment and risk management it requires to gather lots of data and indexes from other realms.

Geophysics which is one of the most valuable branch of earth sciences, has helped to understand volcanic activities and the systems underneath them by using numerous seismic techniques. Geophysics according to its own study area; tracks seismic waves, interpret the seismograms, monitors surface and ground deformations thus tries to image Earth`s physical structure – core, mantle, lithosphere, asthenosphere etc… Seismic waves have a critical value in geophysical surveys cause their propagation and speed through different genres of materials in earth allows us to know Earth`s layers with density, thickness, and phase (solid or liquid) features. Geophysical methods that is used for volcanology, have been developed and enriched till recent years. Collaboration with geophysicists by using seismic maps, recorded signs in seismograms support to detect volcanic unrest periods and estimate volcanic regions. If look at most recent useful techniques, it could be seen like tomography, small antennas, using VLP (very long period) data, spectral analyses, modelling tremors…

Another known fact is, after lots of experiments and observations, earthquakes and volcanic explosions generate different sort of seismic signals and those signals that related to volcanoes still need to be enhanced since mainly seismic activities occur just subsurface of volcanoes and this causes complexity while try to distinguish between them (McNutt, 2002).

Seismic Signals Recorded at Volcanic Regions and Their Identical Features

Signals are described with their frequency and occurrence rate and velocity in general. Some well-known seismic signals are called Very Long Period (VLP), Long Period (LP), Hybrid – Multiphase, ones originated from volcanic explosive eruptions, associated with volcanic hazard ones and tremors.

Very Long Period Seismic Signals

These signals have in period between 2 - 100 s (Figure 1). They are also known for high frequency volcano seismic event signals. According to studies, they were recorded in variety of volcanic eruption. In one eruption case (Santiaguito, Guatemala) it was found that very long period seismic signal’s amplitude rely on magnitude of the eruption (Sanderson et al., 2010).

Long Period Seismic Signals

These are LF (low frequency) events signals. Generally, their frequency is in between 0.2 – 10 Hz (Figure 2). The amplitude of these signals are oscillating; rapidly increases – moderately declines. S waves are not seen very often in low frequency volcanic – tectonic events yet they are original part of the LP signals. In volcanic seismology, long period signals are very beneficial since they can give clear evidence about magma ascending, bubble growth, degassing – decompressing thus whole rheology mechanism along volcano conduit. When these signals interpret carefully, we are able to obtain approximate eruption size, magnitude and duration. Like very long period seismic signals, long period signals are recorded multiple volcanic eruption style as well (Wassermann, 2011).

Hybrid Events Seismic Signals

As the name implies, they are shown as kind a combination of LP and VLP signals (Figure 3). It is known that VLP events sometimes might affect another LP events which occurs in another location and in that case they both can be appeared in a seismograms and triggered each other (Wassermann, 2011).

Multiphase Events Seismic Signals (MP)

They have a higher frequency rate nevertheless; they illustrates very shallower level of volcano – dome growth. They are very hard to identify so they can be easily mistaken with other seismic signals (Figure 4) (Wassermann, 2011). Explosive volcanic eruptions are disastrous events and can lead to lots fatalities. Seismic signals could be originated from these massive magnitude eruptions such as Vulcanian, Strombolian, Plinian and Phreato – Magmatic.

Vulcanian Eruption Origin Seismic Signals

They are seen with high amplitude and high frequency. The major phase might contain Type 1 (low frequency), Type 2 (High Frequency) and Type 3 (hybrid frequency) on vertical seismogram components (Zobin, 2012).

Phreato - Magmatic Eruption Origin Seismic Signals

They show both high and low frequency levels simultaneously. Their recognition depends on high frequency rate within a short period in seismograms (Zobin, 2012).

Volcano – Tectonic Earthquake Seismic Signals (VT)

They can be found both shallower and deeper depths and situated nearby active volcanoes. Of all volcanic earthquakes deep volcanic – tectonic ones are one of the most resembling signals to tectonic origin ones. The main difference is their location and swarms’ rate of frequency scales – higher frequency rate. On one hand shallow – tectonic signals have usually P and no S waves. They got lower frequency values (Zobin, 2012).

Tremors

Volcanic tremors are known for early warning signs about volcanic eruptions. Till so far they have been recognized in seismograms as a continuous signals with regards to their generation in two groups: low viscous tremors and high viscous tremors. Also there is a most recent discovered seismic signal which has a correlate with earthquakes called non – volcanic tremors (Zobin, 2012). Low viscous tremors merged with two phase motion and they show strong clue for unrest stage. As the name implies they are observed in a very low viscous magma or high volatile content volcanoes (Zobin, 2012). When obtained seismic waves from high viscous lava volcanoes, it may point out the gas resonation phase. High viscous tremors always indicate big potential activity inside volcano. Even though there are some unknows about what accurate trigger mechanism behind these activities, kinship between first tremor arrival time and next feasible eruption time still valid in geophysics society (Zobin, 2012).

Volcanic Hazards Origin Seismic Signals

Some of volcanic hazards (mud flows, pyroclastic flows and rockfalls) have a great capacity to produce seismic surface signals. Despite the fact that they happen on the ground with enormous bulk and density, they create above surface shaking which lead to continuous volcanic seismic signals. Most of signals in this group belong to pyroclastic basal dense fluids and lahars because those type of hazards can be triggered by too much rainfalls or melting glaciers during eruption therefore they resulted in high scale seismicity motions in seismograms (Wassermann, 2011).

Volcano Tectonic Origin Seismicity

Tectonic which contains all plate boundary movements has a relation with magmatism thus volcanism. Most active volcanoes can be found nearby the plate boundaries where source locations of earthquakes – convergent, divergent, transform plate boundaries and also minority of them on rift zones (LaFemina, 2015). Divergent plate boundaries which are mainly described with oceanic and lithospheric. On divergent plate boundaries, there are active neo-volcanic areas. On this areas when volcanoes and plate boundaries get magma ejection each other it causes new crust formation (LaFemina, 2015). Mid – ocean ridges have create longest mountain lines and plate boundary paths till now. Volcanoes are formed around these boundaries as a result of upwelling and convection with respect to magmatism. Volcanic activities on this zones lead to create lead to create oceanic crust (LaFemina, 2015). Rift zones are another perspective of divergent boundaries in continental areas. Neo-volcanism on this zones is accompanied with crustal formation (LaFemina, 2015).

Like divergent plate boundaries, convergent plate zones located on continental and oceanic lithosphere. In convergent plate boundaries, nearly whole system controlled by fault systems thus resulted in horst and grabens. Meanwhile volcanic arcs formed by upper plate boundary extension. Alone volcanoes and group of volcanoes may constrain local stress in a specific zones which directly influence tectonism and magmatism there. It can be inferred that magmatic systems and factors that trigger volcanic unrest must be responsible for the strain alteration. In subduction zones, processes related to volcanic arcs (LaFemina, 2015). When tectonic plates (oceanic – oceanic, oceanic – continental or continental – continental) move on or cross each other transform plate boundaries formed. On transform plate zones, magmatic process called transtensional volcanism and this is the main source of stratovolcanoes. Unlike divergent and convergent plate boundaries, amount of injection of volcanic material into crust and lithosphere is a little. Crustal formation happens in transform plate boundaries with dikes. Transform plate boundaries are available place for volcanic arcs. Magmatism and volcanism in these regions highly depend on decompressing (LaFemina, 2015).

Seismic activities get bigger and are shown in seismograms with higher frequency when a step before volcanic eruption occurs -unrest. If we look at volcanic eruption stages, first magma ascending – bubble nucleation and growth then with the help of decompressing vast magma fluid travels through hard magmatic rocks, breaks them and create cracks into them while passes across them. During or after this processes seismicity underground occurs and sometimes volcanic eruptions produces volcanic earthquakes. These earthquakes make seismic signals which are called often tremors and swarms (Thompson, 2015). Tectonic earthquakes creates P – S waves - surface waves and their regular signals can be seen in seismograms clearly as we know, on the other hand volcanic earthquakes produce as afore mentioned LP, VLP, hybrid, multiphase, surface vibration (related to volcanic hazards), volcanic eruption origin seismic signals, tremors and swarms. In addition to them another type of volcanic earthquake signal is volcano – tectonic earthquake which has P and S waves. Those earthquakes create high frequency rate. In spite of the fact that all volcanic earthquakes have their own characteristic seismic signals pattern in seismograms, due to unknowns and uncertainties it is still very confusing to classify them as one by one throng (Thompson, 2015).

Volcanic vs. Tectonic Seismicity Differences

Tectonic earthquakes mostly occur order of first and aftershocks whereas volcanic earthquakes happen in swarms that have more than 1 b-values in contrast to tectonic ones. Non-tectonic earthquakes more likely to spread in homogenous volcanic material while tectonic ones propagates along massive heterogeneous materials and fault failure lines. According to Global Volcanic earthquake Database which contains studied overall more than six hundred recorded swarms, it can be possible to classify swarms in three categories – Type 1 (before eruption), Type 2 (during eruption), Type 3 (not related to eruptions). Volcanic earthquakes occur in nearly 0 – 10 km but non – volcanic earthquakes happen deeper than them around 10 - 700 kilometres because of both fault lines depth and crust thicknesses.

Volcanic earthquake seismic signals are very similar to those tectonic origin ones in seismograms. Since it is difficult to distinguish volcanic signals from the other one, clear and understandable interpretation which based on lots of experiments and observations is required while doing hazard assessment. At this point geophysicists have a critical role for helping to estimate and make a comment on right seismic signals.

Seismic Waves Propagation

After an eruption or earthquake, seismic waves travel through both solid and liquid earth layers. S waves does not propagate in liquid environment so in outer core events it is not expected to see any S waves in seismograms. Seismic wave’s propagation (route and velocity) through volcanic and magmatic rocks highly depends on features of layers in that they go pass. It can be said these features like density, temperature, pressure, depth, consistency (solid – liquid), thickness, saturation degree, and distance from source of events to the station, elastic stiffness, chemical composition, mineral alignment and tectonic settlement. For instance, like ray path and speed according to Snell’s law, in denser material waves get slower with smaller angle and has to travel longer road. Another assumption; higher the temperature lower speed of seismic waves. Also known that diminishing P wave and S wave velocities may because of melting. Sometimes attenuation caused by absorption

In one study (Behn and Klemen, 2003), they aimed to figure out speed difference of P waves between igneous rocks. They tried that with variety of temperature and pressure conditions. At the end of work, they concluded that using only P wave calculations caused lack of information about identify the mineral contents of continental crust parts (middle crust: in dacitic parts velocity lower than 7, in basaltic parts velocity higher than 7 and lower crust: most of speed values around 7) yet huge amount of O2 in magmatic and volcanic rocks could sufficient to glean mineral compositions in a broadband. Besides presence of porosity and hydro alteration, might be the reason for low P wave velocities in lower part of crust. They noted that further work required to merge this study with S wave velocities to see if they constrain crust composition and give more accurate results (Behn and Klemen, 2003).

Another example work; when using for seismic wave tomography, it was observed that partial flow melt - saturation levels inversely proportional to P wave speed and S wave speed and could give high ratio of their speed. It might be insight of rock compositions and alignment would influence seismic waves in a complicated way and he added that since goal of seismic tomography very deep level of crust, it still uncertain what amount of melt trigger to volcanic events. (Lees, 2007).

Structure underneath Volcanoes

In general as we know, geophysical methods are mainly used for oil and gas research. While doing these, working on cross section maps after whole borehole locations and fault lines drawn, trying to create possible trap structure between sedimentary rocks to estimate potential organic hydro carbon emigration points. These studies usually give – to their directions - either positive or negative flower shape models which symbolises all lithological structures beneath us. In volcanoes, it is different because volcanoes that include lots of known unknowns and unknown unknowns, are one of the most complicated earth structures have been known until that time. In one study, U.S.G.S first established seismic refraction platform on south side of Hawaii to find out under mechanism beneath Kilauea volcano. After DAN and SWR data gathered, they witnessed that there was a Pn velocity space between rifts and they implied oceanic crust goes down under Volcano Island with 2 degrees (Zucca and Hill, 1980).

If look at another study, The Central Weather Bureau Seismographic Network (CWBSN) looked at P and S waves arrival times to detect crustal and mantle structures under Taiwan. According to results they suggested that seismic wave velocities deteriorated near ground approximately 30 kilometres and that attenuation zone may be explained by residual thermal parts of historical volcanic events. Thus this underground velocity structure estimated by using seismic waves (Ma et al., 1996). One example from Etna volcano (Italy), they did try to use tomography to visualize the structure beneath Etna volcano (Figure 8). Estimated some seismic anomalies near surface that indicated existing of igneous rocks. Also found high seismic activities with low P wave speed as sign of weak structures (Neri et al., 2002).

In conclusion, seismology has been used to try to enlighten the earth structures by using seismic diagrams. Volcanic hazard assessments still need to get benefit from geophysical methods to provide better and quite useful information for future perspectives such as to elucidate earth physical structures in geoscience works.

Sakurajima Volcano

Sakurajima is one of the main active stratovolcano lies on Aira caldera in Japan. According to observed images, it generally creates ash and sulphuric gases. Most recent volcanic activities recorded in between 2017 and 2018, ash emissions on caldera regions. In its eruptive history, volcano had confirmed a number of bigger than VEI = 4 eruptions before century. Rhyolite, dacite and basaltic andesites are the main member of rock groups. Most well-known biggest volcanic event occurred with volcanic lighting in 1914 that caused so many deaths (S. I. N. M. o. N. H. Global Volcanism Program).

Some scientist and academic staff from Japan Meteorological Agency (JMA), Geological Survey of Japan and University of Hiroshima did work cooperatively in between 1991 – 1992 to test seismicity below 1 Megahertz. They established the seismometers (STS-2) on the North side of crater named Minamidake. They witnessed more than two thousand seismic activities during study (Kato et al., 1992).

After using seismic methods, they listed all seismic signals in vertical, horizontal components of seismograms with spectrum and waveforms. In vertical seismograms most remarkable swarms signals were shown they called these activities “b type events”. They noted that for further work they were planning to study with a lot more stations. They believed that seismic broadband methods would be quite applicable when datas are analysed well to assess the nature of volcanic mechanisms.

Results: Vertical displacement

• Origin waveform
• High amplitude
• b1 – b5: independent ones (reversible each other)
• b6: Harmonic tremor
• b7 – b13: complicated waveforms
• b15 – b20: resembling each other and rising hertz values – signs for structure under Sakurajima volcano
• Non-volcanic events and Non-harmonic tremor

As they emphasized, that study could be improved since they did not use any other station and at that time not deeply examined datas in the paper yet. Moreover, they are right about seismic broadband usage when compared their nearly 30 years ago study to today. Nowadays, seismic band – pass filters have been moderately developed to apply monitoring and forecasting volcanism.

Example Study 2; that experiment was conducted by university staffs from Kyoto University (Japan), University of Hawaii (Manoa), Colorado School of Mines (Golden). They did record lots of available data followed by Vulcanian eruption which occurred in 1998. They looked for whether infrasonic signals were indicated more tangible than seismic signals or not and they did want to answer the question if acoustic signals could be used with seismic signals (Garces et al., 1999).

For the study Sakurajima Volcano Observatory (SVO) did install seismic and infrasonic receivers on island in spring, 1998. During May some pushing signals were recorded. Later in May ash and gas emission estimated and after these significant infrasonic signals, shape of signals changed noticeably as an alert level and end of May numerous tremors recorded. In one night, ash emission quantities jumped up and covered some part the town (Garces et al., 1999).

Results: Monitored volcanic unrest and predicted volcanic activity with acoustic and seismic signals

• Continuous events
• No bright or light stage seen means presence cool volcanic material
• Signs for gas releasing much better seen within infrasonic signals rather than seismic ones
• Long period event relation

Forecasting and monitoring volcano led to enormous changes in signals that had a vital role before the eruption. Propagation influenced acoustic signals less than seismic signals thus there were lots of infrasonic signals which indicates more accurate information than seismic ones. According to result of this study, we can say that monitoring shallower parts of explosive volcano give beneficial result in infrasonic signals because seismic ones can be affected from propagation noises easily. Like in that case, increasing signals are very helpful for predicting next most likely eruption or hazard event occurrence. Both seismic and infrasonic signals appearance in seismograms at volcanic crisis night - 19th of May, 1998. Those signals before eruption (unrest) and after eruption (hazards) are kind a forefoot of volcanic events (Garces et al., 1999).