Smithsonian Institution
Global Volcanism Network Bulletin v. 20, no. 2, February 1995

Merapi (Indonesia)  Seismic data associated with the 22 November
       1994 dome collapse

Java, Indonesia
7.54S 110.44E; summit elev. 2,911 m
All times are local (= GMT + 7 hours)

A collapse of Merapi's gravitationally unstable summit lava dome
took place on 22 November 1994 and VSI volcanologists estimated
that it mobilized roughly 13 x 10^6 m^3 of pyroclastic debris
(Bulletin v. 19, no. 10). Workers at the GMU Geophysical Laboratory
and Martin Beisser of GFZ-Potsdam recorded seismic data during the
collapse from their station at Klathakan, 1.8 km WNW of the summit.
Their broad-band seismic instrument showed the associated
disturbance beginning on 22 November at 1007 and 32 seconds
(radial-component data shown on figure 2). So far as the GMU and
GFZ workers know, the wide dynamic range of their broad-band
instrument preserved the event with a minimal amount of
high-amplitude signal "clipping." Also, in their interpretation,
the collapse and seismic disturbance began simultaneously. In other
words, the initial displacement at the beginning of the seismic
record is thought to correspond to the arrival of signals from the
inception of the collapse.

The collapse-related seismic event lasted for almost an hour
(figure 2). The initial signals were set against a moderately quiet
background, and maximum amplitude generally increased with time.
Highest-amplitude signals were received about 40 minutes after the
event began. These largest signals had amplitudes that reached
approximately 30 mm/sec, whereas at the beginning of the collapse
the maximum amplitudes were only about 0.05 mm/sec. Thus, on the
seismic records, amplitudes ultimately grew to 600 times as large
as the initial signals.

The eruption and collapse also appear in a 200-hour time window
showing measured seismic amplitude in specified wavelengths (figure
3). The figure was prepared using signal processing techniques,
which for the high frequency (0.1-1.0 Hz) data involved significant
averaging of the maximum values (to once an hour). These depictions
show that one or two noteworthy seismic disturbances took place at
about 150 and 180 hours prior to the collapse (cause unknown).
Compared to the other seismic disturbances on these records, the
collapse and eruption induced larger amplitude and much more
sustained signals. The post-collapse signals were also followed by
an interval of at least 10 hours of elevated background (most
noticeable in the 1-12 Hz range).

Using the available data, the investigators failed to find any
clearly related premonitory seismic signals for the collapse.
Sufficient collateral data (for example, teleseismic and
meteorological data) might help constrain detected collapse and
eruption earthquakes, or shed light on the cause of the
pre-collapse seismic disturbances.

Since our last report (Bulletin v. 19, no. 12), continued dome
building occurred at Merapi. On 5 January another collapse brought
1 x 10^6 m^3 of debris downslope. This collapse produced a small
pyroclastic flow on the S slope.

Information Contacts: Arnold Brodscholl and Kirbani Sri
Brotopuspito, Geophysics Laboratory, Gadjah Mada University,
FMIPA-UGM, Sekip Unit III, Yogyakarta 55281, Indonesia; Martin
Beisser, GeoForschungs-Zentrum-Potsdam, Telegrafenberg A31, H117,
14473 Potsdam, Germany (Email:; Wimpy S.
Tjetjep (Director), Volcanological Survey of Indonesia, Jalan
Diponegoro 57, Bandung, Indonesia.

Figure 2. Seismic record for the Merapi 22 November dome collapse.
The component shown is horizontal, radial to the edifice; amplitude
scale is arbitrary. The data were recorded on a data logger
connected to a Streckeisen STS2 seismometer (with a 50 Hz sampling
rate, a 8.33 mHz to 50 Hz linear response, and a 32-bit
analog-to-digital converter). Courtesy of Arnold Brodscholl and
Kirbani Brotopuspito.

Figure 3. Radial component of the Merapi 22 November dome collapse
showing a seismic amplitude (arbitrary scale) versus time for
stated wavelength ranges. The inception of the collapse lies at the
zero point of the time scale. Courtesy of Arnold Brodscholl and
Kirbani Brotopuspito.