*3.2. Geochemical Analyses*

Analyses of eruption products from the 2017 event and the 2018 eruption at Ebinokogen Ioyama were conducted using X-ray diffraction (XRD). Samples smaller than 0.063 mm were washed in extra pure water and a randomly oriented powder prepared and mounted on a glass slide according to the method of Itoh et al. [46]. Rigaku MiniFlex 600 and the PDXL2.6 software were used. Smectite and chlorite peaks at 15 Å were determined by using the peak shift of smectite to 16–18 Å after the samples were treated in ethylene glycol. Kaolinite and chlorite peaks at 7 Å were determined by utilizing the peak disappearance of chlorite after treatment in HCl solution for 2 h. Relative amount of minerals was analyzed by Reference Intensity Ratio (RIR) in the PDXL 2.6 software.

Oxygen isotope ratios of water samples were analyzed using a mass spectrometer from Thermo Fisher Scientific, Inc. (Waltham, MA, USA) (MS: Delta plus) with the automated CO2–H2O equilibrium method or using a Cavity Ringdown Spectrometer from Picarro, Inc.(Santa Clara, CA, USA) (CRD: L2130-i Isotopic H2O). The precisions for both analyses are ±0.1%.

Hydrogen isotope ratios of samples were analyzed using MS (Delta V) with automated H2–H2O reduction method (H-device) or by CRD. The precisions for both analyses are ±1%.

#### **4. Events at Ebinokogen Ioyama Volcano**

#### *4.1. Stage-1: December 2013 to December 2015*

Seismic activity began in the northern area of Karakunidake in December 2013, after the sub-Plinian eruptions at Shinmoedake Volcano [47] in January 2011. A volcanic tremor event lasting 7 min was observed beneath Ebinokogen Ioyama Volcano on 20 August 2014 [48]. The seismic activity tended to gradually increase in 2015, swarmed in July, and tremor events were observed from July to October [49]. Very weak signs of unrest such as a hydrogen sulfide (H2S) smell near the western foot of the Ioyama lava were observed from October 2015 (Figure 4). Stage 1 therefore, is the period from December 2013 to December 2015, during which there was seismic activity, but without any noticeable geothermal activity on the surface.

#### *4.2. Stage-2: December 2015 to January 2017*

On the afternoon (around 4 pm) of 14 December 2015, a hiker found a very small fumarole on the southwest rim of the Ioyama old crater at the A point in Figure 4. The location had previously been identified as the area of fumaroles [34]. The fumarole temperature was measured at 80 ◦C that day. On 25 and 27 December, temperatures of between 93 and 96 ◦C were recorded at the summit area of Ioyama. The water boiling temperature at an altitude of 1300 m is about 96 ◦C. Since then, the area with a thermal anomaly expanded north and east, mainly in the southern part of the Ioyama lava (Figure S1a–f). The water temperature at the spring K3 increased from about 21 ◦C at the end of December 2015 to about 27 ◦C in March 2016, synchronized with the expansion of the thermal anomaly area. There was a brief stability of the area of thermal anomaly from April to August 2016. The thermal anomaly area began to expand again from late 2016 (Figure S1g), and the expansion rate increased from early 2017. From December 2015 to January 2017, Stage-2, there was small expansion of the thermal anomaly area, but temperatures of the fumaroles remained about 96 ◦C (Figure 4).

#### *4.3. Stage-3: January 2017 to February 2018*

The geothermal anomaly area expansion rate increased from early 2017 and continued until early summer 2017 (Figure S1h–j). During the expansion of the thermal anomaly area, the jet fumaroles H and A were formed. The mud pot A was observed at the point A on 19 March (Figure 5a), and another pot appeared at F between 19 and 21 March (Figure 5c). Hydrothermal fluids continued to well at these pots and the jet fumarole H, 30 m east of the mud pot F, started roaring between 15 and 18 April. This activity around the fumarole H was reported in the "YamaReco" web site [51]. We surveyed the roaring jet fumarole H on 22 April 2017 (Figure 5d). Then the jet fumarole vent A with a 1.5 m diameter was first observed about 10 m away from the mud pot A on 5 May (Figure 5b). The area near the jet fumarole vent A was covered with a thin layer of light gray ash-sized altered material (ashy deposit) with block-sized altered host rock ejecta over a distance of a few meters. The surface on the southwestern area of the jet fumarole vent A became grayish-white.

**Figure 4.** The sequence of geothermal activities at Ebinokogen Ioyama Volcano from July 2014 to June 2018. Geothermal locations are shown in Figure 3. (**a**) Yellow areas and bars denote the activity of Ioyama-fumarole area. Light blue line indicates a sulfur smell. Orange line denotes a jet fumarole. Blue line denotes a mud pot or hydrothermal pond. Tremor and low-frequency earthquake (LFE) based on the Japan Meteorological Agency (JMA) web site [50]; (**b**) Temperature observed at point A, before turning into a jet and the jet fumarole A; (**c**) Temperature observed at point H, before turning into a jet and the jet fumarole H; (**d**) Water temperature observed at the spring K3; (**e**) pH of water observed at the spring K3; (**f**) Geothermal anomaly (>50 ◦C) area in the summit of Ioyama. Triangles include larger error values.

**Figure 5.** Details of geothermal activity at Ioyama fumarole area during Stage-3. (**a**) Mud pot A appeared just before/on 19 March 2017 (Photo on 19 March); (**b**) Jet fumarole A appeared on 26 April 2017 (Photo on 7 May); (**c**) Mud pot F appeared between 19 and 21 March 2017 (Photo on 22 April); (**d**) Jet fumarole H appeared between 15 and 18 April 2017 (Photo on 22 April).

That ashy deposit was found on green leaves and could be traced to about 200 m from the vent (Figure 6). The amounts of the deposit at two points around the jet fumarole vent A were 3.8 kg/m<sup>2</sup> and 3.4 kg/m2. As the thickness of the deposit was nearly the same, the mass was estimated to be roughly 1 ton. The amount of deposit around the jet fumarole vent A coincided with the volume of the vent. We did not notice any change in the fumarole activity in this area until the morning of 26 April, according to the video footage of the Ioyama–south web camera maintained at the Japan Meteorological Agency (JMA) [50]. However, the footage showed a gray-white steam rising from the fumarole A at 11:29 am on 26 April 2017. The steam moved to the southwest from 10:20 am to 11:30 am, which coincided with the area covered with the ashy deposit. Therefore, it was concluded that the altered material ashy deposit was blown out from the jet fumarole vent A around 11:29 am on 26 April. The high rates of geothermal activity at the jet fumarole vents continued until July 2017. The temperature of the jet fumarole H was 134.2 ◦C on 24 September 2017. The geothermal anomaly area decreased from late summer or early autumn in 2017 (Figure S1k).

#### *4.4. Stage-4: February 2018 to December 2018*

The expansion of the geothermal anomaly area resumed around February 2018 (Figure 4). The fumarole S appeared about 30 m south of the Ioyama-fumarole area on 3 February 2018 (Figure 7a and Figure S1l). While the previous fumarole spots were named sequentially, beginning with A, this was named S.

**Figure 6.** Drone and satellite map of the Ioyama-fumarole area. (**a**) Locations of mud pot A, jet fumarole vent A, mud pot F and jet fumarole vent H. Observation points are indicated by italic alphabet.; (**b**) Distribution of ashy deposit from the jet fumarole vent A on a Google Earth map [45]. +: coated, -: visible, two values of ashy deposit weight per a square (kg/m2).

The outline of the eruption at the southern and western parts of the Ioyama-fumarole area in April 2018 (the 2018 Ebinokogen Ioyama eruption) was presented in our previous report [52]. The first sign, just 12 days before the eruption, was the appearance of the hydrothermal pond Y1 from around 2 am on 7 April 2018 [53]. The size of the pond Y1 was 9 m long and 5 m wide, and hydrothermal water had welled out and flowed downstream from the pond under observation on 16 April (Figure 7b). The temperature of the hydrothermal pond Y1 was 93.0 ◦C at that time. The activity of the fumarole S continued on 16 April, and the jet fumarole H was accompanied with vigorous roaring.

An eruption began on the south side of Ioyama at 3:39 pm on 19 April [53]. It took place first at the vent S5 near fumarole S, and almost immediately the vent S2 opened, expelling vigorous white steam moving about 30 m northeast from the vent S5 (Figure 3). This succession was captured in the NHK news video. Both vents appeared almost simultaneously within a span of one minute, and the vigorous steams developed into white to pale gray-white plumes with heights of 100 to 200 m. Subsequently, the plume from vent S5 heightened, and the basal jet part of the plume changed color intermittently from gray-white to brown. The size of the vent S2 also widened, and steams also rose from small vents around the vent S2. During an aerial survey with an airplane on the morning of 20 April, we found a new jetting fountain of dark-colored hydrothermal fluids at the vent S7 as a pond. The vents S1 to S7 (4–19 vents) were mapped via analysis of video camera images, and the details are

reported separately [54]. The subsequent survey showed that the vents S1 to S3, the vents S4 to S6, and the vent S7 were merged later into craters. The vents S1 to S3 became the crater Y3, the vents S4 to S6 formed the crater Y2a, and the vent S7 formed the crater Y2b (Figure S2). These craters are collectively referred to as the Ioyama-south craters including the pond Y1. The deposit thickness of the 19 April 2018 eruption ejecta varied from a few centimeters to 45 cm of the Ioyama-south craters. It was composed of a bluish very fine ash-sized material layer at the bottom, a second layer of very fine to fine ash-sized material with coarse ash-sized material and lapilli-sized material, a third layer of fine to coarse ash-sized material with lapilli-sized material, and the uppermost layer of coarse ash-sized material and rounded lapilli to block sized material. The details of these layers will be reported subsequently. Host altered rocks of ballistics ejecta reaching 30 cm in diameter ( ϕ) were distributed around the Ioyama-south craters as Y2a, Y2b and Y3 (Figure 8). An altered ejecta of the host rock of ϕ100 cm was found at the northern edge of the crater Y3, and an ejecta of ϕ70 cm was found at the southern edge of the crater Y2a. Altered host volcanic block and lapilli sized ejecta were distributed radially from the craters Y3 and Y2a, and we observed ϕ4 cm (diameter) altered ejecta at a distance of 140 m, and ϕ9 cm altered ejecta at a distance of the 125 m at the southeastern part of the crater Y2a.

**Figure 7.** Photos showing the geothermal activity around Ioyama-fumarole area in 2018. (**a**) New fumarole S appeared at the south side of Ioyama-fumarole area (Photo on 3 February); (**b**) Hydrothermal pond Y1, 9 m long and 5 m wide that formed on 7 April (Photo on 16 April); (**c**) Distribution of craters of the steam driven eruption on 19 April (Photo on 4 June).

It was reported that the steaming could have started at a distance 500 m west of the Ioyama-south craters on the evening of 20 April [53]. A fissure vent composed of seven small hydrothermal explosion or steam vents (4–20 vents) was confirmed by later observations (Figure 8 and Figure S2). We did not notice any anomaly such as steaming around this area at about 4:30 pm on 20 April, and it is believed that the eruption occurred after 4:30 pm on 20 April [52]. Altered material ashy deposit and altered ejecta distribution around the two hydrothermal explosion vents W3 and W4 of the Ioyama-west crater were considered to be accompanied by a very small eruption (Figure 8). The time of deposition was

considered the evening of 20 April. Photographs taken on the evening of 21 April showed that the large blocks near the Ioyama-west crater were covered with the ashy deposit. We observed the ashy deposit around the Ioyama-west crater, whose lower part was composed of fine to coarse ash-sized material, and the upper part was composed of fine ash-sized material. In addition, we observed altered ejecta comprising of different sizes of block and lapilli distributed at a distance of about 50 m from the Ioyama-west crater. The ejecta of ϕ8 cm was deposited over a distance of about 25 m, while that of ϕ4 cm was deposited about 40 m, and that of ϕ1 cm went as far as about 50 m (Figure 8).

**Figure 8.** Distribution of ejected blocks and fragments from the Ioyama-south and west craters in April, 2018. (**a**) Block and fragment ejecta distribution from Ioyama-south craters using a drone map on a Google Earth map [45]; (**b**) Block and fragment ejecta distribution from the Ioyama-west crater. The value indicates the size of the diameter in centimeters.

Observations under a microscope showed that ashy deposits from both the Ioyama-south craters and the Ioyama-west crater were composed of altered materials and rounded crystals. Furthermore, we could not find any juvenile magmatic material under a microscope (Figure S3).

Other land features related to the April eruption were the crack and the mud pot R that formed south of the Fudoike Crater. Those were observed on 4 June 2018. The crack went across a road about 90 m southeast of the Fudoike Crater, and showed the rupture with a 3.5 cm south–up displacement and N44◦E strike (Figure 3). Since we had not observed this crack before the eruption was presumed to have formed related to the eruptive activity. Furthermore, the mud pot R consisting of three small potholes, was formed to the south of the crack, and hydrothermal fluids were pouring out of one of them.

The vents S1 to S7 observed on 19 and 20 April. However, S1 and S7 subsequently became the three larger depression-shaped craters (Figures 3 and 7c). The crater Y2a was formed within two depressions extending from east–northeast to west–southwest, 37 m long and 15 m wide. Subsidence depth on the cli ff of the Y2a varied in height ranging from 10 m to 3 m. The crater Y2b appeared almost circular but elongated slightly from east–northeast to west–southwest, 26 m long and 19 m wide. Subsidence depth on the cli ff of the Y2b varied in height from 8 m to 2 m. The crater Y3 also elongated from east–northeast to west–southwest. Subsidence depth on the cli ff of the Y3 varied in height from 12 m to 3 m (Table 1). We calculated elliptical shape areas of craters Y2a, Y3 and Y2b by using the long and short radii. Using the mean value of depths of the produced depressions, we calculated a total depression volume of approximately 8400 m<sup>3</sup> (min. 3500 m<sup>3</sup> to max. 13,400 m3). It is assumed that three depression craters were formed between the time of the aerial survey on 20 April and 4 June, and that hydrothermal fluids continued to gush out from these craters. Furthermore, the ground uplifts and tilt–displacements continued until December 2018 [55].


**Table 1.** Topographical values of Ioyama-south craters.

#### **5. Results of Geochemical Analyses**

We analyzed ash-sized particles from the 2017 and 2018 events using XRD. Relative amounts (reference intensity ratios, RIR) in the XRD software of minerals are shown in Table 2.

**Table 2.** Relative intensities of hydrothermal minerals by the XRD analyses.


Alu: alunite (soda), Kl: kaolinite, Cr: cristobalite, Qz: quartz, Py: pyrite, Sm: smectite, S: sulfur. Plus marks indicate relative intensities of the XRD peaks. ++++: -40% in weight, +++: -20%, ++: -10%, +: <10%.

An altered material ashy deposit sample from the April 2018 eruption (20180419-Ash) was collected on a shrub near the Ioyama-south craters, avoiding contamination by ground surface material. The sample was observed to contain significantly altered white colored fragments, light white to frosted clear-colored weakly altered fragments, and slightly altered black to gray to brown colored

lithic fragments. It also contained rounded crystals of plagioclase and pyrite when observed under a microscope (Figure S3). Plagioclase, quartz, alunite (soda), kaolinite, cristobalite, pyrite, smectite and sulfur were detected by the XRD analysis (Table 2 and Figure S4). Plagioclase, pyroxenes and olivine are part of phenocrysts in volcanic rocks in this area, while quartz is not [32]. Therefore, quartz, alunite, kaolinite, cristobalite, pyrite, smectite, and sulfur in the ash-sized particles samples are considered to be the minerals produced by the hydrothermal activity or alteration. The RIR analysis showed quartz plus alunite as the major components of this sample, comprising of more than 70% in weight, and was followed by pyrite and plagioclase. The analysis also showed little amounts of kaolinite (about 10 wt%), cristobalite of about 5 wt% and smectite of about 1 wt%. These minerals are considered to have been produced primarily in acidic to neutral environments [56].

Ashy deposit from the April 2017 event (20170426-Ash) was collected from a large stone surface around the jet fumarole vent A, with minor contamination from ground surface material. The sample contained strongly and weakly altered fragments as well as slightly altered lithic fragments. Rounded plagioclase and pyroxenes were also observed using the microscope (Figure S3). Quartz, alunite (soda) cristobalite and sulfur were detected with the XRD analysis, as minerals associated with hydrothermal activity or alteration (Table 2 and Figure S4). The RIR analysis showed quartz plus alunite of nearly 90% in weight.

We collected fluid samples at the spring K3 on 16 and 21 February, and on 12 and 30 March 2016 and at the hydrothermal pond Y1 on 16 April 2018. We analyzed the abundances of ions of elements and the ratios of oxygen and hydrogen isotopes in those samples in Table S1. The water at the spring K3 in 2016 had δD values of −53% to −54% and δ18O values of −8.0% to <sup>−</sup>8.3%, and the compositions fall on the meteoric water lines with δD = δ<sup>18</sup> O +10 to +20. However, the hydrothermal fluids sampled at the Y1 had a δD value of −31.3% and a δ18O value of <sup>−</sup>0.0%, which are values between those of Andesitic magmatic water (AMW) [57,58] and global meteoric water line (GMWL) [59] in Figure 9.

**Figure 9.** Oxygen and hydrogen isotope ratios of the spring K3 and the hydrothermal pond Y1 water. The "Andesitic magmatic water" (AMW) [57,58] and "global meteoric water line" (GMWL) [59] are also shown.
