The Shikotsu caldera in southwestern Hokkaido is a collapse caldera formed by a large ignimbrite eruption at around 46 ka. During the caldera-forming eruption, a large volume of ignimbrite was ejected from the caldera and spread in all directions. A part of the flow reached the city of Sapporo. After the caldera-forming eruption, Eniwa, Fuppushi, and Tarumae volcanoes were formed as post-caldera volcanoes. The area of Chitose and Tomakomai, located east of the Shikotsu caldera, has several key outcrops that reconstruct volcanic activities of the Shikotsu caldera. This outcrop is in Tomakomai, about 20 km southeast of the caldera. Within this outcrop, three layers of pumice-fall deposits from Tarumae volcano can be recognized on the pumice flow deposit of the caldera-forming eruption of the Shikotsu caldera. A dark-colored paleosol separating these eruption deposits shows a hiatus between these Plinian eruptions. In the distance, Tarumae volcano, one of the post-caldera volcanoes of the Shikotsu caldera, is visible.

(Photo: Mitsuhiro NAKAGAWA;

Explanation: Mitsuhiro NAKAGAWA and Nobuo GESHI)

 Ground displacement in a volcanic region constrained by precise geodetic measurements has motivated us to infer magmatic activity at depth, e.g., inflation/deflation of a magmatic deformation source. The inference was initially made by predicting the elastic response to a pressurized magma chamber. A prediction with an unrealistically high chamber pressure has then introduced other rheological models, including elastoplastic and viscoelastic rheologies. Viscoelasticity predicts transient behavior of deformation even though the source mechanism is constant with time, depending on the viscosity structure of the crust and the presence of an elastic layer in the uppermost crust, which requires volcano deformation to be considered a more complex phenomenon, but provides an excellent opportunity for inferring a magma chamber not only as a deformation source but also as a zone that has rheologically less strength. Indeed, geodetically measured rates of volcano deformation, often having constrained crustal viscosities in the range of ∼10¹⁷-10¹⁹ Pa s, are higher than those of postseismic deformation in regions where there is less volcanic activity. This suggests that crustal viscosities are lowered by higher geothermal gradients due to the presence of magma beneath volcanic regions. Understanding volcano deformation in terms of the interaction between magmatic activity and rheological properties of the crust also has the potential to offer a dynamical aspect to petrological and geophysical images of a magma chamber. The volcano deformation model is now required to introduce the magma plumbing system into the response of a viscoelastic crust, in which a magma chamber has excess pressure, which varies with time during conservation of the mass of magma, making it possible to predict throughout volcano dynamics from accumulation of magma to eruption in more self-consistent way.

 A controlled-source audio-frequency magnetotelluric (CSAMT) survey was conducted over Toya caldera, Hokkaido, Japan, to investigate its subsurface structure. The caldera is 10-11 km in diameter and contains a freshwater lake, Lake Toya, that occupies the entire caldera floor. A post-caldera dacitic dome complex, the Nakajima Islands, is present within Lake Toya. The CSAMT survey was carried out along a 16-km-long line that crosses Toya caldera in a NE–SW direction, passing over the Nakajima Islands. The 17 receiver stations (7 stations located outside of Lake Toya, 5 stations within Lake Toya, and 5 stations on the Nakajima Islands) were distributed along the survey line. Unique on-boat measurements were performed at the stations on Lake Toya. Two-dimensional inversion of the CSAMT data revealed the resistivity structure for the upper 1500 m beneath the caldera. The resistivity structure indicates the existences of high-resistivity (> 100 Ω·m) and low-resistivity (< 30 Ω·m) domains at the northeastern and southwestern sides of Lake Toya, respectively, a medium-resistivity (30-50 Ω·m) domain beneath Lake Toya, a high-resistivity (> 100 Ω·m) layer at the Nakajima Islands, and a low-resistivity (< 10 Ω·m) domain beneath the Nakajima Islands. This resistivity structure, combined with geological and bathymetric data, suggests that the subsurface structure of Toya caldera comprises altered Tertiary to Quaternary volcanic/sedimentary rocks outside of the caldera, and a homogeneous caldera-fill deposit beneath the caldera floor. A 9-km-diameter ring fault may occur along the caldera rim. There is a conspicuous hydrothermal alteration zone beneath the Nakajima Islands that may have formed in response to heating of the caldera-fill deposit by the underlying magma during the volcanic activity that formed the Nakajima Islands.

 Kutcharo caldera is situated in the eastern part of Hokkaido, Japan. It is subcircular with a diameter of 20-26 km and was formed by repeated violent rhyolitic explosive eruptions from 340 to 35 ka. The caldera has three post-caldera volcanoes: Atosanupuri, Nakajima and Mashu. Our new tephrostratigraphical survey suggests that a pyroclastic fall deposit (Nakajima pumice), which was extruded from Nakajima, is exposed at the western slope of Atosanupuri. The deposit comprises dacitic pumice clasts (< 9 cm across) in a fine-grained matrix. The pumice clasts are coated with fine-grained ash, suggesting the deposit was produced by a phreatomagmatic eruption. The deposit directly overlies a scoria-fall deposit, which was ejected from Mashu volcano at 17-12 ka, and is overlain by Ma-k tephra, which was extruded from Mashu volcano at 10 ka. The stratigraphy of the deposit suggests that a phreatomagmatic eruption occurred at Nakajima volcano between 17 and 10 ka.

 The fundamental structure of a collapse caldera is subsidence of a block(s) into a magma chamber with the evacuation of a massive volume of magma from the chamber. Decompression of a magma chamber caused by the extraction of magma from a magma chamber coinciding with pyroclastic eruption, effusive eruption, and lateral magma migration drives caldera collapse. Caldera-forming eruptions exhibit wide variations. Caldera-forming pyroclastic eruptions are characterized by a high eruption rate, whereas caldera-forming effusive eruptions have a much lower effusion rate. Caldera collapse caused by a pyroclastic eruption generally occurs within one day, whereas incremental collapses continue for up to one month in the case of some effusive eruptions. Collapse calderas also have wide structural variations. The aspect ratio of the roof of the magma chamber controls the development of caldera faults. The development of multiple caldera faults with a high aspect ratio causes piecemeal collapses, whereas a low aspect ratio results in the subsidence of coherent blocks detached by a simple ring fault. With the progress of caldera subsidence, the caldera structure develops from a flexural down sag to a double-ring fault system, and finally reaches an upward-flaring “funnel shaped” caldera following an intense collapse of the caldera wall. Calderas vary widely in size from 1 km to 100 km, and can also be divided into at least three classes by their internal structures. The largest group of more than 20 km across is characterized by flexural down-sag deformation. The intermediate-size group is characterized by well-developed caldera-border faults. The smallest calderas of less than 10 km across may have piecemeal structures. The existence of a pre-caldera volcanic edifice is also an indicator of differences in the magmatic system. Some calderas form at the summit of a pre-existing stratovolcano or shield volcano, whereas some large calderas form in a cluster of small volcanoes, involving non-volcanic basement rocks. A structural model of caldera development should involve these wide spectrums of collapse calderas. The development of collapse calderas is controlled by variations of magmatic activity, such as eruption style, eruption rate, and duration of eruption, as well as the architectures of their magma storage systems.

 A stratigraphic study of the Toya Ignimbrite in southwestern Hokkaido, Japan, was performed to clarify the sequence of caldera-forming eruption at Toya caldera. The Toya Ignimbrite (thickness < 80 m) is rhyolitic in composition and comprises six stratigraphic units: (1) a fine-grained ash-fall deposit at the base; (2) a base surge deposit and an overlying, voluminous, pumiceous pyroclastic flow deposit, both of which contain accretionary lapilli; (3) a number of base surge deposits and associated ash-fall deposits; (4) a pumiceous pyroclastic flow deposit that contains large lithic clasts up to 50 cm in diameter; (5) a pumiceous pyroclastic flow deposit with a basal lithic-rich layer (lag breccia); and (6) a pumiceous pyroclastic flow deposit at the top. The stratigraphy suggests that the caldera-forming eruption at Toya caldera commenced with a phreatomagmatic explosive eruption (forming unit 1), followed by violent phreatomagmatic eruptions that generated a voluminous pyroclastic flow (unit 2), and small-scale phreatomagmatic eruptions that generated a number of base surges (unit 3). The next phreatomagmatic eruption triggered caldera collapse (unit 4), which reached the climax with a violent phreatomagmatic eruption (unit 5) and ended with a magmatic eruption (unit 6). These eruptions occurred continuously without any significant time breaks. Component analysis of non-juvenile lithic clasts suggests that the vent-opening phase of the caldera-forming eruption involved a single vent. Pyroclastic flows during the caldera collapse may have erupted from multiple vents. Textural studies of pumice clasts suggest that white pumice was ejected during the initial to final stages, while banded pumice and grey pumice were ejected during the final stage. Geochemical data indicate that there was no significant change in magma composition during the caldera-forming eruption, with the exception of a small amount of mafic magma was mixed into the rhyolitic magma during the final stage.

 Shikotsu volcano is a caldera volcano in southwestern Hokkaido. It is considered that a large-scale eruption started at ca. 60 ka (named as 60 ka Shadai eruption), mainly ejecting scoria fall and scoria flow deposits. The volcano repeated explosive eruptions every several thousand years, and after 10 ky of dormancy, a caldera-forming eruption took place at ca. 46 ka. Trench and boring surveys were carried out in the eastern part of Shikotsu caldera, and stratigraphy and changes in the components of the 60 ka Shadai eruption were reexamined. There are three types of juvenile materials in 60 ka eruption deposits: dacitic pumice, olivine-bearing andesitic scoria, and intermediate banded/gray pumice. As a result, tephra layers of the eruption are mainly classified into three units: Units A to C in ascending order. Soils and volcanic ash soils were not discovered among these layers, so these must derive from a single eruption sequence. Unit A consists of pumice fall deposits and ash fall deposits; Unit B of scoria fall deposits; and Unit C comprises a pyroclastic flow deposit and a following pyroclastic fall deposit. Each eruptive unit is subdivided into A1-A3, B1-B5 and C1-C2. Based on these eruptive units, the sequence of the 60 ka Shadai eruption is constructed as follows: Phase 1 was a pumiceous plinian eruption (A1, A2), and eruption rate abruptly decreased in A3. Phase 2 was a scoriaceous plinian eruption (B1-B5). Eruption rate was unstable in early Phase 2 (B1-B4); however, it gradually increased in late Phase 2 (intermittent eruption had repeated in B5). Consequently, a pyroclastic flow eruption occurred in Phase 3 (C1), followed by a plinian eruption (C2). Andesitic and mixed magmas were found to begin to erupt from Phase 2, while dacitic magma survived through the 60 ka Shadai eruption. These characteristics are difficult to explain based on the eruption from a zoned magma chamber or from a common eruptive vent. Dacitic and andesitic magmas are suggested to have existed in separated magma chambers, and dacitic magma at least was supplied through an independent vent system.

 Proximal pyroclastic deposits of the 46 ka caldera-forming eruption of Shikotsu volcano are investigated at new outcrops along the Opoppu river, south of the volcano. The deposits can be divided into 6 units, from A to F in ascending order, according to lithofacies, components, and time intervals. Based on the stratigraphy of the deposits, the sequence of the 46 ka Shikotsu eruption is revealed. Activity started with phreatomagmatic and phreatoplinian eruptions (Phase 1: Unit A). The eruption style changed to magmatic without a time interval before the plinian eruption (Phase 2: Unit B). The eruption column was intermittently unstable, producing pyroclastic surge and flow deposits during the latter period of Phase 2. The lithic breccia content also increased in the latter period. After a possible erosional interval, explosive eruptions occurred, producing voluminous pyroclastic flows and ended with the effusion of lag breccia (Phase 3: Unit C). After a dormant period, pyroclastic flows effused intermittently (Phase 4: Units D and E). At the final stage, small plinian eruption occurred, associated with a pyroclastic surge (Phase 5: Unit F). Juvenile materials of the 46 ka Shikotsu eruption are mainly crystal-poor (CP type) rhyolite pumice from Phases 1 to 3. Small amounts of crystal-rich (CR type) dacite and andesite pumice occurred from the final stage of the Phase 3. Two types of silicic material coexisted during Phases 4 and 5. According to temporal changes of lithic contents, the caldera collapse was almost completed at the end of Phase 3 with the formation of lag breccia. Therefore, it could be considered that activity in Phases 4 and 5 was either the terminal phase of caldera-forming activity or the initial activity of post-Shikotsu caldera volcanoes.

 The durations of caldera-forming eruptions that produce large volumes (> 100 km³) of pyroclastic ejecta are poorly understood due to the absence of direct observations. However, clarifying the timescale of these catastrophic hazardous events is essential for understanding associated eruption dynamics and links with the eruptible portions of the underlying magma system. Case studies addressing the time scale of caldera-forming eruptions are reviewed. Three Quaternary large volume caldera-forming eruption deposits from Yellowstone (US), Taupo (New Zealand) and Kutcharo (Japan) volcanoes are inferred from deposits to have lasted for periods of at least for months to years. However, these estimations are generally based on geological evidence, such as re-worked deposits between eruption units, and not quantitative evidence. Proposed here is a new method of timescale estimation based on paleomagnetic secular variation. A sampling procedure was developed for accurate oriented samples of pyroclastic deposits including volcanic ash. This procedure makes it possible to obtain the mean remanent magnetization of a tephra layer with a 95% confidence limit of about 2°, which is comparable to those of well-determined directions for lava. Based on this procedure, the 7.3 ka Kikai caldera-forming eruption was investigated as a trial. Samples for paleomagnetic measurement were collected from the basal ash-rich part of the lowermost plinian pumice fall (Koya pumice fall) at Satsuma Iwo-jima, Kyushu. The difference of 6.9 degrees in the remanent magnetizations between Koya pumice and reported data of the uppermost co-ignimbrite ash-fall (Kikai–Akahoya Ash) suggests the caldera-forming eruption of a considerable duration (> 50 years) on the basis of average rate of secular variation. Paleomagnetic directions from pyroclastic deposits could be a powerful tool for estimating timescales of large explosive eruptions.

 Toya caldera, eastern Hokkaido, was formed approximately 110 ka, and has two post-caldera volcanoes, Nakajima and Usu volcano. Caldera-forming eruptions ejected a large-scale Toya pyroclastic flow, and related co-ignimbrite ash which covered a wide area in northern Japan. Usu volcano is one of the most active volcanoes in Japan, and has had repeated magmatic eruptions during historical time since AD 1663. The latest major eruption occurred in AD 2000. General geology and representative outcrops of eruptive deposits from Toya caldera and its post-caldera volcanoes are introduced, on the basis of a training field course at the 6^th IAVCEI Collapse Caldera Workshop held in September 2016.

 地学雑誌124巻4号（2015）掲載の杉戸信彦・近藤久雄「上町断層帯の最新活動と河内平野の地形環境変化」（p. 607-631）に誤りがありましたので，お詫びし訂正いたします。

 p. 627「IV．おわりに」：

 （誤）一方，上町断層帯沿岸部の最新活動時期は約2200 ～ 2300 yr BPと推定され，

 （正）一方，上町断層帯沿岸部の最新活動時期は約2200 ～ 2300 yr BP以降，江戸時代以前と推定され，