A review of algorithms for detecting volcanic hot spots in satellite infrared data

Since the 1980s, application of thermal infrared satellite data for volcano monitoring has rapidly evolved to become a proven operational tool. Due to the large quantities of data provided by sensors in polar and geostationary orbits, as well as the sheer number of active volcanoes on earth, processing and managing such data sets requires an enormous amount of workforce. A number of algorithms have been developed to facilitate detection, location, and tracking of hot spots of active volcanoes. A collation and review of hot spot detection algorithms developed and applied by the volcanological community reveals three main types which have been applied to date: contextual, fixed threshold, and temporal. The founding algorithms for these three classes are VAST, MODVOLC, and RST, respectively. Through comparison with manually based detections, the performance of each algorithm was tested for sustained lava flows (Etna and Stromboli), strombolian activity (Stromboli), lava dome growth and collapse (Augustine), and fumarole fields (Vulcano). It is shown that, as the number of correctly identified anomalies increases, so too does the number of false positives. Although each of the algorithms operates well within the limits and criteria of their design requirements and application, under current data restraints, no algorithm can be expected to perform perfectly.     

Analysis of volcanic activity patterns using MODIS thermal alerts

We investigate eruptive activity by analysis of thermal-alert data from the MODIS (moderate resolution imaging spectrometer) thermal infrared satellite instrument, detected by the MODVOLC (MODIS Volcano alert) algorithm. These data are openly available on the Internet, and easy to use. We show how such data can plug major gaps in the conventional monitoring record of volcanoes in an otherwise generally poorly documented region (Melanesia), including: characterising the mechanism of lava effusion at Pago; demonstrating an earlier-than-realised onset of lava effusion at Lopevi; extending the known period of lava lake activity at Ambrym; and confirming ongoing activity at Bagana, Langila and Tinakula. We also add to the record of activity even at some generally better-monitored volcanoes in Indonesia, but point out that care must be taken to recognise and exclude fires.Editorial responsibility: J Stix     

Complex effusive events at Kīlauea as documented by the GOES satellite and remote video cameras

 GOES provides thermal data for all of the Hawaiian volcanoes once every 15 min. We show how volcanic radiance time series produced from this data stream can be used as a simple measure of effusive activity. Two types of radiance trends in these time series can be used to monitor effusive activity: (a) Gradual variations in radiance reveal steady flow-field extension and tube development. (b) Discrete spikes correlate with short bursts of activity, such as lava fountaining or lava-lake overflows. We are confident that any effusive event covering more than 10,000 m² of ground in less than 60 min will be unambiguously detectable using this approach. We demonstrate this capability using GOES, video camera and ground-based observational data for the current eruption of Kīlauea volcano (Hawai'i). A GOES radiance time series was constructed from 3987 images between 19 June and 12 August 1997. This time series displayed 24 radiance spikes elevated more than two standard deviations above the mean; 19 of these are correlated with video-recorded short-burst effusive events. Less ambiguous events are interpreted, assessed and related to specific volcanic events by simultaneous use of permanently recording video camera data and ground-observer reports. The GOES radiance time series are automatically processed on data reception and made available in near-real-time, so such time series can contribute to three main monitoring functions: (a) automatically alerting major effusive events; (b) event confirmation and assessment; and (c) establishing effusive event chronology. Received: 12 January 1999 / Accepted: 13 July 1999     

Review on Morphological Insights of Self-Potential Anomalies on Volcanoes

Energy and thermal transfers in active volcanoes can play an important role incontrolling their dynamics depending on the hydrothermal state. Much geothermalenergy is released through the groundwater circulation, hot gas emission and thermalconduction. Therefore, it is very important to know the hydrological and thermalenvironments associated with volcanoes from the volcano-energetic point of view.However, it is difficult to evaluate these because of the availability of only a fewborehole data on the summit of volcanoes. Recent studies reveal that self-potential(SP) anomalies (up to some hundreds of mV) are observed on volcanoes, activefissure zones and/or fumarolic areas, suggesting that the SP anomalies are closelyrelated to heat-triggered phenomena such as thermoelectric and electrokinetic effectsdue to hydrothermal circulations. Therefore, SP studies can be appropriate for sensingthe thermal and hydrothermal states of volcanoes. In addition, monitoring SP anomaliescan be an efficient method for describing the change of thermal state and the evolutionof the hydrothermal (and volcanic) activities.In this paper, we have reviewed the origin of the SP anomalies associated withvolcanic phenomena theoretically as well as experimentally. Subsequently, wehave presented the results of many case studies and have classified the types ofanomalies in accordance with possible mechanisms. We have also described theresults of time variations of SP anomalies associated with volcanic activities. Timevarying SP fields exhibit the dynamic aspects of volcanic activities correspondingto the evolution of hydrothermal activity, changes in ground water circulation andmagma displacement. These morphological insights should lead to a quantitativeinterpretation of SP anomalies in volcanic regions.     

On the dynamics and origin of the Pugachevo group of gas-water-lithoclast (“mud”) volcanoes on Sakhalin Island: Visual observations and orohydrography

Visual observations of the activities of three gas-water-lithoclast (“mud”) volcanoes in the Pugachevo group in the Makarov District of Sakhalin Island were used to analyze the dynamics of these volcanoes. Our analysis confirmed the previous inference that a multirank periodicity is present in the eruptions of the volcanoes, with the periodicity being comparatively frequent and low in magnitude for the eruptions (every 1–2 years) at the Glavnyi or Central volcano and a very long but violent (every 65–70 years) at all the three volcanoes. Comparison of this periodicity with a similar periodicity observed in natural seismicity in the adjacent, highly seismic Uglegorsk District shows that the two phenomena are relatively independent. The natural seismicity in the form of large earthquakes is superimposed upon the gas-water-lithoclast volcanism mostly to serve as a trigger. The high-rank periodicity of 65–70 years noted above may be related to 11–22-year cycles of solar activity. Analysis of the orohydrographic setting in the area shows that the location of the present low-lying basin with volcanoes inside was formerly occupied by a dome uplift with a major gas field that controlled the generation of these volcanoes; it continues to drive their activity, causing a continued gradual subsidence or collapse of the deep basin with volcanoes, thus providing evidence of the high gas potential in the area of study.     

A statistical method linking geological and historical eruption time series for volcanic hazard estimations: Applications to active polygenetic volcanoes

The probabilistic analysis of volcanic eruption time series is an essential step for the assessment of volcanic hazard and risk. Such series describe complex processes involving different types of eruptions over different time scales. A statistical method linking geological and historical eruption time series is proposed for calculating the probabilities of future eruptions. The first step of the analysis is to characterize the eruptions by their magnitudes. As is the case in most natural phenomena, lower magnitude events are more frequent, and the behavior of the eruption series may be biased by such events. On the other hand, eruptive series are commonly studied using conventional statistics and treated as homogeneous Poisson processes. However, time-dependent series, or sequences including rare or extreme events, represented by very few data of large eruptions require special methods of analysis, such as the extreme-value theory applied to non-homogeneous Poisson processes. Here we propose a general methodology for analyzing such processes attempting to obtain better estimates of the volcanic hazard. This is done in three steps: Firstly, the historical eruptive series is complemented with the available geological eruption data. The linking of these series is done assuming an inverse relationship between the eruption magnitudes and the occurrence rate of each magnitude class. Secondly, we perform a Weibull analysis of the distribution of repose time between successive eruptions. Thirdly, the linked eruption series are analyzed as a non-homogeneous Poisson process with a generalized Pareto distribution as intensity function. As an application, the method is tested on the eruption series of five active polygenetic Mexican volcanoes: Colima, Citlaltépetl, Nevado de Toluca, Popocatépetl and El Chichón, to obtain hazard estimates.     

Studies in the activity of Klyuchevskoi volcano by remote sensing techniques between January 1, 2001 and July 31, 2005

The Kamchatka Branch of the Geophysical Service (KB GS) of the Russian Academy of Sciences (RAS) has been observing the activity of Kamchatka volcanoes since 2000 in near real time using three methods: (1) seismicity monitoring (2) visual and video observations, and (3) satellite monitoring of thermal anomalies and ash discharges. The joint use of these data provides objective information on the state of the volcanoes from which to predict possible eruptions. During the period of time investigated, which culminated in the eruptions of March 10, 2003 to February 27, 2004 and January 12, 2005 to April 28, 2005, two active periods of Klyuchevskoi Volcano were identified. The results from our study of the first of these periods helped define an approximate scenario for the activity of the volcano before a summit eruption. The use of this experience in combination with an analysis of the literature enabled us to produce a successful short-term forecast of the January 2005 eruption.     

A remarkable pulse of large-scale volcanism on New Britain Island,Papua New Guinea

New studies of the deposits from the latest caldera-forming eruption (the “Dk” event) at Dakataua Volcano, New Britain Island, Papua New Guinea, help identify an intense space-time concentration of large-scale volcanism during the 7th century AD on New Britain. Radiocarbon dating of charcoal from the Dk deposits yields an age of 1,383 ± 28 BP. Calibration of this result gives an age in the range AD 635–670 (at 1 s. d.). At about the same time, two other volcanoes on New Britain, Rabaul and Witori, also produced very large eruptions. Very high acidity levels in ice cores from Antarctica and Greenland at AD 639 and AD 640 respectively may be linked to either or both of the Dakataua and Rabaul eruptions. Another ice core high acidity level, at AD 692, may be associated with the Witori eruption. Significant volcanic risk within the New Britain region is indicated by its Late Cenozoic history of relatively frequent large-scale eruptions from as many as 8 caldera systems within an arc-parallel zone about 380 km long. Over the last 20 ka the return period for major (VEI 5+) eruptions in this region was about 1.0 ka and individually high frequencies of major eruptive activity were experienced at Witori and Rabaul. The relatively short return period for major eruptions in the region would tend to increase the chance that such events could cluster in time.     

Volcanic lightning: global observations and constraints on source mechanisms

Lightning and electrification at volcanoes are important because they represent a hazard in their own right, they are a component of the global electrical circuit, and because they contribute to ash particle aggregation and modification within ash plumes. The role of water substance (water in all forms) in particular has not been well studied. Here data are presented from a comprehensive global database of volcanic lightning. Lightning has been documented at 80 volcanoes in association with 212 eruptions. The Volcanic Explosivity Index (VEI) could be determined for 177 eruptions. Eight percent of VEI = 3–5 eruptions have reported lightning, and 10% of VEI = 6, but less than 2% of those with VEI = 1–2. These findings suggest consistent reporting for larger eruptions but either less lightning or possible under-reporting for small eruptions. Ash plume heights (142 observations) show a bimodal distribution with main peaks at 7–12 km and 1–4 km. The former are similar to heights of typical thunderstorms and suggest involvement of water substance, whereas the latter suggest other factors contributing to electrical behavior closer to the vent. Reporting of lightning is more common at night (56%) and less common in daylight (44%). Reporting also varied substantially from year to year, suggesting that a more systematic observational strategy is needed. Several weak trends in lightning occurrence based on magma composition were found. The bimodal ash plume heights are obvious only for andesite to dacite; basalt and basaltic-andesite evenly span the range of heights; and rhyolites are poorly represented. The distributions of the latitudes of volcanoes with lightning and eruptions with lightning roughly mimic the distribution of all volcanoes, which is generally flat with latitude. Meteorological lightning, on the other hand, is common in the tropics and decreases markedly with increasing latitude as the ability of the atmosphere to hold water decreases poleward. This finding supports the idea that if lightning in large (deep) eruptions depends on water substance, then the origin of the water is primarily magma and not entrainment from the surrounding atmosphere. Seasonal effects show that more eruptions with lightning were reported in winter (bounded by the respective autumnal and vernal equinoxes) than in summer. This result also runs counter to the expectations based on entrainment of local water vapor.     

First recorded eruption of Mount Belinda volcano (Montagu Island), South Sandwich Islands

The MODVOLC satellite monitoring system has revealed the first recorded eruption of Mount Belinda volcano, on Montagu Island in the remote South Sandwich Islands. Here we present some initial qualitative observations gleaned from a collection of satellite imagery covering the eruption, including MODIS, Landsat 7 ETM+, ASTER, and RADARSAT-1 data. MODVOLC thermal alerts indicate that the eruption started sometime between 12 September and 20 October 2001, with low-intensity subaerial explosive activity from the islands summit peak, Mount Belinda. By January 2002 a small lava flow had been emplaced near the summit, and activity subsequently increased to some of the highest observed levels in August 2002. Observations from passing ships in February and March 2003 provided the first visual confirmation of the eruption. ASTER images obtained in August 2003 show that the eruption at Mount Belinda entered a new phase around this time, with fresh lava effusion into the surrounding icefield. MODIS radiance trends also suggest that the overall activity level increased significantly after July 2003. Thermal anomalies continued to be observed in MODIS imagery in early 2004, indicating a prolonged low-intensity eruption and the likely establishment of a persistent summit lava lake, similar to that observed on neighboring Saunders Island in 2001. Our new observations also indicate that lava lake activity continues on Saunders Island.Editorial responsibility: J. Gilbert     

The hazards of eruptions through lakes and seawater

Eruptions through crater lakes or shallow seawater, referred to here as subaqueous eruptions, present hazards from hydromagmatic explosions, such as base surges, lahars, and tsunamis, which may not exist at volcanoes on dry land. We have systematically compiled information from eruptions through surface water in order to understand the circumstances under which these hazards occur and what disastrous effects they have caused in the past. Subaqueous eruptions represent only 8% of all recorded eruptions but have produced about 20% of all fatalities associated with volcanic activity in historical time. Excluding eruptions that have resulted in about a hundred deaths or less, lahars have killed people in the largest number of historical subaqueous eruptions (8), followed by pyroclastic flows (excluding base surges; 5) tsunamis (4), and base surges (2). Subaqueous eruptions have produced lahars primarily on high (>1000 m), steep-sided volcanoes containing small (<1 km diameter) crater lakes. Tsunamis and other water waves have caused death or destroyed man-made structures only at submarine volcanoes and at Lake Taal in the Philippines. In spite of evidence that magma–water mixing makes eruptions more explosive, such explosions and their associated base surges have caused fewer deaths, and have been implicated in fewer eruptions involving large numbers of fatalities than lahars and tsunamis. The latter hazards are more deadly because they travel much farther from a volcano and inundate coastal areas and stream valleys that tend to be densely settled.     

Intermediate depth earthquakes and volcanic eruptions in Central America, 1961–1972

Eight Central American volcanoes had large eruptions during the period 1961 to 1972. The distribution of intermediate depth earthquakes which occurred during the same period is marked by eight concentrations. Seven of the eight very active volcanoes are spatially related to the concentrations of intermediate depth earthquakes. The centers of the concentrations are typically a few tens of kilometers seaward of the volcanoes. The earthquakes have focal depths of about 70 to 110 km. Directly below the active volcanoes there is little or no intermediate depth seismic activity. Partially melted areas along the deep seismic zone directly below the active volcanoes might explain this distribution. Spatial-temporal progressions relating specific intermediate depth earth-quakes to specific volcanic eruptions have not been recognized. The development of a concentration of intermediate depth earthquakes spatially related to a quiescent volcano may indicate that the volcano will soon enter a period of renewed activity.     

Influence de la Nature des Magmas sur l’Activité Phréatomagmatique: Approche Volcanologique et Thermodynamique

Many volcanic forms resulting from phreatomagmatic eruptions of differentiated magmas have been studied in the Massif Central (France), in the Phlegrean Fields (Italy), and on Saõ Miguel island (Azores). They show a continuous series between explosion crater maar type — and the hyaoloclastic tuff-cone. An essential feature of this morphological series is the preponderance of tuff-rings resulting from subaerial eruptions. Subaerial tuff-rings of basic compositions are less common than maars. A thermodynamic approach shows that the quantity of heat supplied by the different kinds of magmas and the water / magma ratio are the essential parameters controlling the activity, and the resulting morohology of these volcanoes.     

Cyclic flank-vent and central-vent eruption patterns

 Many basaltic and andesitic polygenetic volcanoes have cyclic eruptive activity that alternates between a phase dominated by flank eruptions and a phase dominated by eruptions from a central vent. This paper proposes the use of time-series diagrams of eruption sites on each polygenetic volcano and intrusion distances of dikes to evaluate volcano growth, to qualitatively reconstruct the stress history within the volcano, and to predict the next eruption site. In these diagrams the position of an eruption site is represented by the distance from the center of the volcano and the clockwise azimuth from north. Time-series diagrams of Mauna Loa, Kilauea, Kliuchevskoi, Etna, Sakurajima, Fuji, Izu-Oshima, and Hekla volcanoes indicate that fissure eruption sites of these volcanoes migrated toward the center of the volcano linearly, radially, or spirally with damped oscillation, occasionally forming a hierarchy in convergence-related features. At Krafla, terminations of dikes also migrated toward the center of the volcano with time. Eruption sites of Piton de la Fournaise did not converge but oscillated around the center. After the convergence of eruption sites with time, the central eruption phase is started. The intrusion sequence of dikes is modeled, applying crack interaction theory. Variation in convergence patterns is governed by the regional stress and the magma supply. Under the condition that a balance between regional extension and magma supply is maintained, the central vent convergence time during the flank eruption phase is 1–10 years, whereas the flank vent recurrence time during the central eruption phase is greater than 100 years owing to an inferred decrease in magma supply. Under the condition that magma supply prevails over regional extension, the central vent convergence time increases, whereas the flank vent recurrence time decreases owing to inferred stress relaxation. Earthquakes of M≥6 near a volcano during the flank eruption phase extend the central vent convergence time. Earthquakes during the central eruption phase promote recurrence of flank eruptions. Asymmetric distribution of eruption sites around the flanks of a volcano can be caused by local stress sources such as an adjacent volcano. Received: 18 March 1996 / Accepted: 14 January 1997     

Time variant analysis of geomagnetic signals describes the volcanic activity of Mt. Etna between early 2000 and late 2002

Volcanomagnetic anomalies have been mostly observed during strong eruptions. Our aim is to improve the geomagnetic data analysis to evidence the anomalies occurring in a larger time span, especially in the phases preceding the eruptive events. We developed a time variant statistical approach and applied it to the 2000–2002 Etna geomagnetic temporal series. It is based on an algorithm that statistically predicts the geomagnetic field at the station on the volcanic edifice by that recorded at the remote one. In such a way a number of significant changes in the time series (called statistical innovations), marking the local magnetic field change, were detected. The distribution of such statistical innovations accurately describes the Etna volcanic evolution: we note a progressive increase of the innovation occurrence as the eruptive cycles were approaching and only few and weak innovations at times between the various eruptive cycles. The significance of this analysis is further confirmed by the close agreement among the mean square prediction error, strain release and the volcanic activity behavior. On the contrary, the geomagnetic field at a single station or its difference at two stations do not have any clear correlation with other measured physical quantities. The complex pattern of the prediction error was also investigated by a multifractal analysis. We found that the Holder regularity increases with the intensification of the volcanic activity, implying that innovations tend to be less sporadic and correlated during the major volcanic phases.     

Stochastic modelling at Stromboli: a volcano with remarkable memory

The assessment of potential volcanic eruptions is a critical aspect when evaluating the safety of populated areas. A stochastic approach has been developed for the analysis and simulation of data sampled at active volcanoes. This approach allows for the detection and quantification of time correlation, volcanic event forecasts using Cox model based simulations and volcanic tremor decomposition in order to identify potential precursors of major eruptions. The stochastic approach has been applied to data monitored at Stromboli volcano. Significant time correlation has been detected which makes Stromboli a volcano with a remarkable memory of its recent activity in comparison to other volcanoes. Forecasting of the number of strombolian events for the next few days has been performed by Monte Carlo simulations. Finally, kriging analysis of the tremor intensity has enabled time component estimation which could furnish additional monitoring variables for the forecast of paroxysmal phases at Stromboli.     

Offset caldera and crater collapse on Juan de Fuca ridge-flank volcanoes

 Forty-three volcanoes located along the flanks of the Juan de Fuca Ridge were selected to study relationships between their morphologies and off-axis magmatic processes. The volcanoes occur both in chains consisting of up to seven distinct cones and isolated edifices. Nearly all of the volcanoes are circular, truncated cones with steep flanks and large, relatively flat summit plateaus. In addition, most of these volcanoes also have prominent and distinctly offset calderas or craters. The most striking characteristic of the volcanoes' morphology is that nearly all of their collapse structures are located on the sides of the volcanoes which face the Juan de Fuca Ridge and many are breached with openings toward the ridge. A simple model based on these observations accounts for these ridge-facing features. As plate motion transports a volcano away from its magma source beneath the lithosphere, the volcano's magma supply conduits tend to lag behind. Eventually these conduits are abandoned and ridgeward collapse structures are formed. It can be inferred from the model that, on average, individual volcanoes were active for approximately 50 000 years and that most eruptions took place early in this interval. If most of the cone-building eruptions occurred during the first thousand years or so, associated hydrothermal activity may have temporarily rivaled the present-day yearly time-averaged hydrothermal output along the entire Juan de Fuca ridge axis. Received: 1 September 1996 / Accepted: 13 January 1997     

Gas emissions from failed and actual eruptions from Cook Inlet Volcanoes,Alaska, 1989–2006

Cook Inlet volcanoes that experienced an eruption between 1989 and 2006 had mean gas emission rates that were roughly an order of magnitude higher than at volcanoes where unrest stalled. For the six events studied, mean emission rates for eruptions were ∼13,000 t/d CO₂ and 5200 t/d SO₂, but only ∼1200 t/d CO₂ and 500 t/d SO₂ for non-eruptive events (‘failed eruptions’). Statistical analysis suggests degassing thresholds for eruption on the order of 1500 and 1000 t/d for CO₂ and SO₂, respectively. Emission rates greater than 4000 and 2000 t/d for CO₂ and SO₂, respectively, almost exclusively resulted during eruptive events (the only exception being two measurements at Fourpeaked). While this analysis could suggest that unerupted magmas have lower pre-eruptive volatile contents, we favor the explanations that either the amount of magma feeding actual eruptions is larger than that driving failed eruptions, or that magmas from failed eruptions experience less decompression such that the majority of H₂O remains dissolved and thus insufficient permeability is produced to release the trapped volatile phase (or both). In the majority of unrest and eruption sequences, increases in CO₂ emission relative to SO₂ emission were observed early in the sequence. With time, all events converged to a common molar value of C/S between 0.5 and 2. These geochemical trends argue for roughly similar decompression histories until shallow levels are reached beneath the edifice (i.e., from 20–35 to ∼4–6 km) and perhaps roughly similar initial volatile contents in all cases. Early elevated CO₂ levels that we find at these high-latitude, andesitic arc volcanoes have also been observed at mid-latitude, relatively snow-free, basaltic volcanoes such as Stromboli and Etna. Typically such patterns are attributed to injection and decompression of deep (CO₂-rich) magma into a shallower chamber and open system degassing prior to eruption. Here we argue that the C/S trends probably represent tapping of vapor-saturated regions with high C/S, and then gradual degassing of remaining dissolved volatiles as the magma progresses toward the surface. At these volcanoes, however, C/S is often accentuated due to early preferential scrubbing of sulfur gases. The range of equilibrium degassing is consistent with the bulk degassing of a magma with initial CO₂ and S of 0.6 and 0.2 wt.%, respectively, similar to what has been suggested for primitive Redoubt magmas.     

On precursor of Kamchatkan volcanoes eruptions based on data from satellite monitoring

Kamchatka is one of the most active volcanic regions on the planet. Large explosive volcanic eruptions, in which the ash elevates up to 8?C15 km above sea level, occur here every 1.5 years. Study of eruptions precursors in order to reduce a volcanic risk for the population is an urgent problem of Volcanology. The available precursor of strong explosive eruptions of volcanoes, identified from satellite data (thermal anomaly), as well as examples of successful prediction of eruptions using this precursor, are represented in this paper.     

Tyatya Volcano, southwestern Kuril arc: Recent eruptive activity inferred from widespread tephra

Abstract Tyatya Volcano, situated in Kunashir Island at the southwestern end of Kuril Islands, is a large composite stratovolcano and one of the most active volcanoes in the Kuril arc. The volcanic edifice can be divided into the old and the young ones, which are composed of rocks of distinct magma types, low‐ and medium‐K series, respectively. The young volcano has a summit caldera with a central cone. Recent eruptions have occurred at the central cone and at the flank vents of the young volcano. We found several distal ash layers at the volcano and identified their ages and sources, that is, tephras of ad 1856, ad 1739, ad 1694 and ca 1 Ka derived from three volcanoes of Hokkaido, Japan, and caad 969 from Baitoushan Volcano of China/North Korea. These could provide good time markers to reveal the eruptive history of the central cone, which had continued intermittently with Strombolian eruptions and lava flow effusions since before 1 Ka. Relatively explosive eruptions have occurred three times at the cone during the past 1000 years. We revealed that, topographically, the youngest lava flows from the cone are covered not by the tephra of ad 1739 but by that of ad 1856. This evidence, together with a report of dense smoke rising from the summit in ad 1812, suggests that the latest major eruption with lava effusion from the central cone occurred in this year. In 1973, after a long period of dormancy, short‐lived phreatomagmatic eruptions began to occur from fissure vents at the northern flank of the young volcano. This was followed by large eruptions of Strombolian to sub‐Plinian types occurring from several craters at the southern flank. The 1973 activity is evaluated as Volcanic Explosivity Index = 4 (approximately 0.2 km³), the largest eruption during the 20th century in the southwestern Kuril arc. The rocks of the central cone are strongly porphyritic basalt and basaltic andesite, whereas the 1973 scoria is aphyric basalt, suggesting that magma feeding systems are definitely different between the summit and flank eruptions.