updated on Mar. 18, 2009

Monitoring of volcanic gases at Miyakejima volcano, Japan

Geological Survey of Japan

Kazahaya Kohei

SO2 flux measurements

Miyakejima volcano has started to emit gases from the summit crater (1.6 km diameter; 500m deep) since mid August. Sulfur dioxide (SO2) flux monitoring by COSPEC V has been conducted from Aug. 26, 2000 in the island by Tripod Method, and the method has changed to airborne traverse since Sep.1, 2000. Monitoring is performed almost everyday when helicopters are available by Meteorological Agency (JMA) and Geological Survey of Japan.

The SO2 flux is widely varied from 5 to 230 kt/d, and the average is 21 kt/d (from mid Sep.2000 to July, 2002).

SO2 flux (March 16, 2009) updated

Monthly flux and cumulative total amount of SO2 (as of Feb. 2009)

Gigantic SO2 emission from Miyakejima volcano, Japan, caused by caldera collapse
Kazahaya et al., Geology, Vol. 32, No. 5, pp. 425-428, 2004

The Highest Record is here!

detailed description (AGU2000 poster 6 pages)

1) Airborne technique

COSPEC measurements (1 2)have been done with helicopters from Coast Guard and Marine Self Defense Force. Two methods are used to monitor SO2 flux because of complex features of plume flow styles.

Line Traverse method is used for a simple plume flowing to only one direction (example). Normally 2-4 times traverse scan beneath a directed plume 10-40km far from the Miyakejima island are performed.

Round traverse method (traveling round the Miyakejima Island) is used when plumes flow toward two directions due to the difference in the wind direction in lower and higher elevations (example). Two times round traverse have been usually done 10-20km far from the island. Relatively less times measurements at a time are restricted by a short time schedule of helicopter operations.


The monitoring first started with a standard set of COSPEC V with 128 and 426 ppmm standard cells. As the SO2 concentrations largely exceeded 500 ppmm, the standard cells changed to high concentrations (526 and 1436 ppmm) in Sep. 16, 2000. The signals have still exceeded 1500 ppmm. The linearity is found over the range from 0 to 1436 ppmm, however, strong non-linear relation between the real concentration and signal is found as suggested from the company.

Moreover the upper limit of signals at around apparent concentration of 2400 ppmm is also found and it is deduced to the problem of the standard setting of COSPEC V. Since Oct. 1, 2000, the measuring wave length was shifted to a smaller peak a bit closer to the visible region. This resulted in decrease in the signal intensity roughly to be 1/2. Since then the upper limit problem has been cleared out.

3) Non-linearity in the calibration curve

In order to determine the non-linear relations at higher concentrations, we conducted an experiment during monitoring the plume. A standard cell with 1436 ppmm was put during measuring SO2 signals from a plume for about 10 sec. The height of the standard in the recorded chart was diminished to a greater extent when the base SO2 signal was higher (example). A hundred of samples are obtained from Oct. 10 to 31. The calibration curve is made after some mathematical treatment and is shown here. We are now making two extra standard cells with concentrations of 3000 ppmm and 6000 ppmm to obtain the better results.

4) SO2 flux

The wind velocity data is necessary to calculate SO2 flux. It sometimes induces large error on the flux value when the wind velocity data is unreliable. For the case of the SO2 flux monitoring at Miyakejima volcano, we try to use several sources of wind data to reduce the SO2 flux data as follows:

a) Helicopter (middle elevation)

b) Coast Guard's boat (surface)

c) Self defense force's air plane (high elevation)

d) Visible-band satellite images (plume movement)

e) JMA's upper air meteorological monitoring data at Hachijojima island (surface to high elevation; 100km south from Miyakejima)

f) JMA's theoretical estimate for weather forecast (surface to high elevation)

Time variation of the SO2 flux is given here. As plumes are sometimes dense enough to interfere the UV light penetration through the plumes, a part of flux value would be underestimated.

Some examples of monitored raw data is shown here. Oscillatory noises in the SO2 signal occur by rotating blades of helicopter.

5) Amount of magma degassed

As SO2 released from active degassing volcanoes is widely recognized that it is of magmatic origin. Therefore, mass of the magma degassing can be calculated when S content of the magma is known. S content of the magma is presumed using standard method by melt inclusion analyses with EPMA to be from 1500-2000 ppm (Saito, G.).

The SO2 flux of 40 kt/d indicates that the mass rate of the magma degassing to be 10 Mt/d. The total volume of the degassed magma is 0.3 km3 or 1 Gt at the end of Nov., 2000.

6) Degassing mechanism

Continuous release of SO2 (+CO2, H2S) with the greatest amounts can be explained only with a mechanism "magma convection in conduits" (e.g., Kazahaya, K., Shinohara H. and Saito G. 1994: Bull.Volcanol. 56 207-216). The flux is controlled by two rate-control processes;

a) bubble separation rate from magma

b) magma convection rate (mass flow rate of undegassed magma)

The rate of the former process is the function of gas content, bubble separation pressure and surface area of magma at the head of magma conduit etc. And the latter is the function of gas content, conduit diameter and others. In both cases, the conduit diameter or the surface area of magma head affects the degassing rate consequently. Therefore, the greatest SO2 emission rate would be due to the largest conduit size for basaltic magma systems at Miyakejima volcano.

In fact, the 2000 Miyakejima volcanic activity started as a piston-cylinder type collapse at summit, and the big magma conduit is likely to be made passively (volcanic activity model).

6) Future work

Continue monitoring of SO2 flux as long as possible.

Recalibration using high concentration standard gas cell (order made) up to 7500 ppmm.

7) Notes

Routine monitoring has been done by people in the COSPEC Monitoring Team of Japan Meteorological Agency who are Mrs. Mori H., Odai M., Nakahori Y., Matsui Y., Tanaka Y., Miyahara H., Kawasaki T., Mata, E. and Ogitani M.

Researchers participated in this operation are Prof. Hirabayashi J. (Kusatsu-Shirane Volcano Observatory, Tokyo Institute of Technology), Uto K. (Geol. Surv. J.), Shinohara H. (Geol. Surv. J.) , Saito, G. (Geol. Surv. J.) and me.

We thank all of people who helped us technically, on business (white color work) and spiritually.

Special thanks to Sutton A.J. (Hawaii Volcano Observatory) who helped us on setting the COSPEC V for high concentration monitoring and to Gardner C. (Cascade Volcano Observatory) on suggestions at the beginning of the airborne operation.

Helicopter operation is supported by Japan Coast Guard and Japan Maritime Self Defense Force and Japan Ground Self Defense Force.