Analogue Experiments of Volcanic Explosions Using High Pressure Gas

To see the influences of pressure, volume and duration of gas jetting on the style of volcanic explosion we have started analogue experiment that releases high pressure gas under sand. The apparatus consists of a one- meter cubic container and a gas chamber set on the center of the container bottom. Inner diameter and inner depth of the chamber are 50 and 80mm, respectively, but changeable by inserting spacer. Vent radius is between 10mm and 50mm with 10mm interval. Chamber was covered with sand in 0-64mm depth, released up to 1 MPa dry air by breaking diaphragm. A pressure sensor in the chamber monitors the pressure decrease. Air wave and sand motion are collected by microphone and high speed video camera, respectively. The formed sand columns are largely classified into two: relatively fast, bursting out jet-like one (jet type) and relatively slow, bell-shaped one with clear envelope (bell type). Jet type is formed when the vent radius is small, although the air wave pressure becomes lower with decreasing vent diameter. Even when the column is bell type, sand velocity and vent diameter are negatively related. These results clearly demonstrate that our general idea, that velocity of ballistics and the intensity of air wave are positively related, is not necessarily correct at the same total energy. The pressure record in the chamber implies that the higher sand velocity by narrower vent is due to slow decompression and the resultant long acceleration period. The present result is important in that slight difference in energy release process causes the large differences on the style of volcanic explosion even when the energy source condition is the same. This also means source pressure and observed air pressure do not correlate one by one. Ohminato et al. (2006) showed in Asama 2004 eruptions accumulated gas pressure and observed air pressure were not necessarily positively related to each other, that is contrary to the theoretical prediction from Kanamori et al. (1984). This disagreement is not strange if we consider the present results.