Preliminary report on eruption at Anatahan Volcano, Northern Mariana Islands

(updated 31 July 2003; posted on 25 July 2003)

Earthquake Reserach Institute, University of Tokyo,
Institute of Seismology and Volcanology (SEVO), Kyushu University, and
Emergency Management Office (EMO), CNMI

The erutpion activity that started on May 10 has been monitored devotedly and continuously by EMO and USGS. We had a chance to carry out geological inspection together with GPS installation on this volcano during 16-19 July 2003. This is the preliminary report of gelogical inspection and the following analysis. We hope that this report and photos are of much help for understanding of the present eruptive activity at Anatahan.

Fig. 1 S view of the eastern part of the Anatahan Island. The active crater (E caldera) is under the cloud. Taken by S. Nakada on July 18, 2003.

Fig. 2 Southern rim of W caldera: thickly covered with new tephra. Taken by S. Nakada on July 19, 2003.

Fig. 3 Inside W caldera, facing to the active crater that locates in E caldera (crater). Taken by S. Nakada on July 19, 2003.

Fig. 4 Caldera floor just south of the active crater. Taken by S. Nakada on July 18, 2003.

Fig 5 Section of tephra just south of the active crater. Pumice fall deposits is covered by multiple thin layeras of gray ash of phreatic eruptions. Original image was taken by S. Nakada on July 18, 2003.

Fig. 6 SE view of steming active crater. Dry mud pool is at bottom. Taken by S. Nakada on July 19, 2003.

Fig. 7 NW view of the active crater. Bloken lava dome (?) is inside the crater. Pyroclastic cone developes, surrounding the crater. Taken by S. Nakada on July 19, 2003.

Fig. 8 Infrared camera image of the active crater, taken from S-rim of the crater. Taken by T. Matsushima on July 18, 2003.

Fig. 9 ibid, taken from E. Reddish parts in upper right represent the remnant of lava dome. Taken by T. Matsushima on July 19, 2003.


Anatahan Volcano erupted on the evening of May 10, 2003 (for detail, see USGS-EMO home pages). The volcano has no historical document of eruption. It is important to investigate why the volcano erupted at this time from the standpoint of tectonics and volcanology. Moreover, we have to monitor its activity how the eruption proceeds and when it ceases its activity, from the stand point of hazard assessment and reduction in the CNMI region.

Our team together with EMO staff has made GPS measurements since 1998, succeeding the old effort by John Beavan and his colleagues in early 1990. Repeated GPS measurements since then have revealed clear evidence of back arc spreading of the Mariana Trough with velocities of 4-6cm/yr. This suggested that the Northern Mariana Islands are on restless tectonic environments. This tectonic situation is the cause of volcanic activity in the northern Mariana islands, including the eruption of Pagan in 1981. (Refer to; Kato et al., Geophys. Res. Lett., vol. 30, 1625, 27-1, 2003)

The eruption of Anatahan provides us with a very important opportunity to investigate the mechanism of volcanic eruptions in these areas. This is practically important for hazard assessment and evaluate potential eruption in other volcanoes. Moreover, it is primarily scientifically important to investigate the mechanism of volcanic eruption in principle.

Motivated by the above practical and scientific goals, we planned to deploy a scientists' party to investigate the eruption of Anatahan. Followings are the preliminary documents for the field research done by the Japanese team.

Outline of GPS installation and geologic survey

We established the fixed GPS site at ANAT GPS site at Anatahan this time. Solar panel and a battery are used for power and a small PC is used for automatic downloading of the data. This system may allow long term continuous recording.

Anatahan volcano during our inspection was quiet and its activity was lowest in seismological level. Only weak steaming was observed at the active crater. Therefore, it is fortune that we succeeded the field inspection using fishing boat and helicopter under these condition.

Geologic inspection

Tephra of the present eruption consists mainly of pumice-bearing brown ash in lower and gray fine ash in upper (Fig. 5). The former is products of pumice (subplinian?) eruption, whereas the latter phreatic eruption. Both furthermore consist of many layers respectively. At the village (NW end of the island), gray ash is totally about 3 cm thick, contrasting to about 20 cm thick for products of brown ash. Color of matrix of brown ash resulted from high-temperature oxidation during eruption.

At the SE part of the island (new seismic station), tephra of this eruption is as thin as less than 3 cm. Melting of plastic bottles were found and the possibility of high temperature ash cloud was raised. However, the evidence suggesting high temperature of ash deposit is not clear; for example, grasses and trees do not show any damage from heat of ash.

The S outerslope of the active crater in the E caldera is thickly covered by gray ash (Figs. 2, 3 and 4). Many rills and gullies developed on the deposits due to water-impermeable nature of gray ash (very fine particles), as well as in most parts inside the W caldera. Rare amount of partly broken stripped trees stand on the slope and gray ash (tephra) thickly covers them on the side facing to the active crater. Gray ash is about 20 cm thick near the crater rim and pumice-bearing tephra below is about 25 cm thick. The latter accumulates blocks and fragments of pumice. Grass stems are partly burnt only near the bottom of deposits near the crater.

Inside the western caldera, tephra is as thick as one meter in total. Gray ash is deposited most thickely NW of the crater, suggesting nearly lateral movement through in the lowest part of the eastern caldera rim. Pumice-bearing tephra is thickest in WSW direction of the crater. The latter is in harmony with drifting direction of eruption plumes in the earliest stage, taken by the satellite images (BGVN, April 2003). Although most of trees had survived from pumice falling during eruption, they were toppled down by strong lateral movement of gray ash by phreatic eruptions.

Active crater

Only steaming was observed during two days of helicopter inspection. The active crater is located in the southern part of the eastern crater bottom inside the eastern caldera, and the southern wall of the active crater is extended directly to the wall of the eastern crater. Dimension of new crater is approximately 300 m across and 100 m deep (deepest part in south). Inside is rugged due to the existence of several lava blocks. A mound like remnant of lava dome can be seen in the northen periphery inside the crater (Fig. 7). The surface was covered thickly by gray ash of phreatic eruption. Infrared camera image shows higher temperature of this mound than outside the crater (Fig. 9). The dome may have been bloken by explosive eruptions which occurred in the middle June when active seismic and visual activity was observed (USGS-EMO report). Neither bombs nor blocks was clearly visible on the crater floors of both the eastern crater (outside the pyroclastic cone) and eastern caldera. Original dimension of the lava dome may be small, because large blocks of lava dome were not observed around the active crater. The deepest bottom of crater was filled with mud pool (dried already) (Fig. 6). Infrared camera showed the temperature as high as 300 C on July 18, but as low as less than 100 C on July 19 (Fig. 9). Low pyroclastic cone developes in the northern side, surrounding the crater (Fig. 7). The maximum thickness of tephra newly deposited, exposed on a gully through pyroclastic cone (Fig. 6), can be estimated roughly to be 20 m.

Movie of active crater (19 July 2003)

Windows media player (wmv) (9.7 MB)
Realpalyer (rm) (2.7 MB
...low in quality)
Quicktime (mpg) (4.5 MB)

Chemistry and degassing of magma

Pumice of this eruption is crystal-poor and light to dark brown in color. A pumice block collected from the pumice-fall layer, just south of the active crater, was analyzed with XRF at ERI. The pumice fragment has light brown crust and dark brown vesicular core, and the crust and core parts were separately analyzed . The SiO2 contents of two parts were identical, about 61 wt%.

Blue- to purple-colored volcanic gas was observed and strong smell of rotted egg was felt near the south rim of the eastern caldera on 18 July. The discolored part represents meaty sulphur dioxide (SO2). We detected 2 - 4 ppm sulphur dioxide gas (SO2) and 0.5 ppm hydrogen sulfide gas (H2S). SO2 gas emission rate looked not so high throughout our field inspection; flux of sulphur dioxide should be less than several thousand tons a day; .similar to at Sakurajima volcano, Japan

Brief summary of geological inspection

Ongoing eruptive event started with pumice eruption (subplinian?) probably in middle May, such that whole of the island was covered with brown tephra (ash) especially in the direction WSW of crater. It is likely that, then, lava dome formed inside the crater, though it was broken by the following events. After an erosion hiatus, strong phreatic eruptions were repeated (probably in middle June). Phreatic eruptions issued gray ash, which was deposited mainly inside the calderas but weakly covered most part of the island. Deposition of the latter resulted in frequent occurrence of small mud-flows after that. As the crater bottom deepened during eruption (near sea level), interaction of magma with water, invaded from the interior of the volcanic body, might have triggered destructive phreatic explosions. Detail time relation between eruptions and modes of eruptions should be investigated carefully hereafter.

Information contact: Setsuya Nakada, Volcano Research Center, ERI, Univ. Tokyo,; Teruyuki Kato, ERI, Univ. Tokyo,; Takeshi Matsushima, Institute of Seismology and Volcanology (SEVO), Kyushu Univ.,; Ray Chong, EMO, CNMI,; Juan T. Camacho, EMO, CNMI,