A Prototype Global Volcano Surveillance System
Monitoring Seismic Activity and Tilt
E. T. ENDO, P. L. WARD. D. H. HARLOW, R. V. ALLEN, and J. P. EATON
Abstract
The Earth Resources Technology Satellite makes it feasible For the first
time to monitor tile level of activity at widcty separated volcanoes and to
relay these data almost instantaneously to one central ottice. This capability
opens a new era in volcanotogy where the hundreds ol' normally quiescenl but
potentially dangerous volcanocs near populated regions around the world can
be economically and reliably monitored. A prototype global volca,ao surveillance
system has been established beginning in the Ia[I of 1972 with the help of local
scientists on 15 volcanoes in Alaska, Hawaii, Washington, California, Iceland,
Guatemala, El Salvador, and Nicaragua. Data on earthquake activity and ground
tilt ave received 6 to 10 times daily in Menlo Park, California, within 90 minutes
ol lransmission from the sites. Seismic event counters were installed at 19
locations with biaxial boreholc tiltmeters with 1 micro,'adian sensitivity in,qalled at seven sites. Direct comparison ol seismic events that are counted
with records from nearby seismic stations show the event counters work quite
reliably. An order of magnitt, de increase in seismic events was observed prior
to the eruption of Vole,fin Fuego in Guatemala in February, 1973. Significanl
changes in tilt were obse,-ved on volcanoes Kilauea, Fuego, and Pacaya. This
study demonstrates the technological and economic feasibility of utilizing svch
a volcano surveillance system throughout the world.
Introduction
The Earth Resources Technology Satellite (ERTS) has opened a
n e w e r a in v o l c a n o l o g y in w h i c h t h e h u n d r e d s of n o r m a l l y q u i e s c e n t
but potentially dangerous volcanoes near populated regions around
t h e w o r l d c a n b e e c o n o m i c a l l y a n d r e l i a b l y r n o n i t o r e d d a i l y to w a r n
w h e n a n y o n e v o l c a n o is b e c o m i n g a c t i v e a g a i n . In the p a s t o n l y a
i e w v o l c a n o e s h a v e b e e n m o n i t o r e d f o r l o n g p e r i o d s of t i m e b e c a u s e
of the h i g h c o s t of b u i l d i n g a n d staffing v o l c a n o o b s e r v a t o r i e s . Y e t it
316
--
is known f r o m data collected in this century that while perceptible
signs of pending eruptions may occur only minutes to days in advance, measurable signs m a y be detected days, weeks, months, or even
years before a m a j o r eruption. Although prediction of specific eruptions is still an elusive goal, early warning of increased activity at
quiescent volcanoes is now a distinct possibility. Such early warnings
F~6. 1 - Volcanoes monitored in this study. Seismic event counters were placed at
all volcanic sites. Tiltmeters were placed on Lassen, Kilauea, Fuego, and
Pacaya.
can be used to reduce volcanic hazards and to focus research aimed
at volcano prediction on those volcanoes t h r o u g h o u t the world that
are most likely to erupt at any given time.
A prototype volcano surveillance system was established during
the latter part of 1972 and early 1973 on 15 volcanoes in Alaska,
Hawaii, Washington, California, Iceland, Guatemala, El Salvador, and
Nicaragua (Fig. 1, Table 1). Nineteen seismic detectors that count
four different sizes of earthquakes and six biaxial, borehole tiltmeters
that measure g r o u n d tilt with a resolution of 1 microradian have been
installed. Data f r o m these instruments are relayed through the ERTS-1
31.35'
45.95'
40.8tV
36.20'
12"
12°
12"
12"
48" 47.00'
47" 56.49'
46° 11.58'
Cerro Negro
Mina Limon
San Cristobal
Telica
Mt. B a k e r
Mt. R a i n i e r
Mt. St. H e l e n s
Nicaragua
Washington
121" 54.20'
121" 40.38'
122' 1420'
42.08'
44.18'
0!.43'
51.55'
1676
1890
1423
351
299
750
61
1070
1115
1143
1242
--
1600
2256
1402
1345
1661
1676
2713
1410
1600
2658
549
106
Elevation
Meters
~' Santiagu.ito I d i s c o n t i n u e d - 10-20-73.
3 t i l t m e t e r s a r e c u n n e c t e d to the a n a l o g i n p u t o1 o n e t r a n s m i t t e r at U w e k a h u n a .
20.20'
86"
86"
87"
86'
24.6'
25.40'
!6.53'
17.00'
15.6'
17.60'
21" 51.00'
155"
155"
155"
155"
19+
19"
19`'
19"
23.75'
64" 01.30"
Ahua
Northpit
Summer Camp
Uwekahuna
lccland
Hawaii
41.55'
38.45'
50.62'
34.48"
37.35'
33.75'
33.65'
34.51'
89', 37.80'
90,'
90"
9090"
9091"
90"
91,'
13" 49.25'
14" 2~.55'
14" 40.00'
14,' 26.65'
14- 20.61'
14" 23.05"
14" 46.60'
14- 23.85'
14,, 42.59'
l_zalco
El S a h ' a d o r
Agua
B u e n a Vista
Fuego
Los Dolores
Pacaya
-" S a n t i a g u i t o 1
San Carlos
Santiaguito 2
121- 30.50'
40" 28.52'
Lassen
California
Guatemala
152" 10.92'
153:' 22..25'
60" 10.92'
59- 22.55'
Longit.de
W
Latitttde
N
Iliamna
Augustine
Statiott
gallic
Alaska
SltllC
or
COllHlr>
E, S
E, S
E, S
E, S
E, S
E, S
E
E, T
E
"1'
T
E
E
E, S
E, T, S
T
T
E, S
E
E
E, S
E, T, S
E
E
[iLslt'llments
Installed
Dale
9-30-72
9-24-72
Installed
10- 9-72
10-11-72
10-12-72
4- 3-73
4- 4-73
4- 2-73
8-ll-73
4- 4-73
3-10-73
12- 3-72
12- 7-72
12-19-72
2-13-73
2-21-73
2-17-73
11-29-73
3-22-73
2-15-73
3-18-73
11-21-73
3- 7-73
9-30 to 10-13-73
TXl31.E 1 - E v e n t c o u n t e r (E), t i l t m e t e r (T), a n d h i g h gain s h o r t p e r i o d (S) s e i s m o m e t e r s t a t i o n locations.
I
I
ta,a
--
318
satellite and through a teletype link to the U. S. Geological Survey
Office in Merdo Park for rapid analysis.
Earthquakes and tilt of the ground are m o n i t o r e d because these
are two p h e n o m e n a that have been observed most consistently to
change prior to and during volcanic eruptions. For h u n d r e d s of years
strong earthquakes have typically been felt just prior to major eruptions (e.g. HARLOW, 1971; SmMOZURt, 1971). Since the advent of sensitive seismometers, orders of magnitude increases in the n u m b e r s of
earthquakes have been observed hours, days, and m o n t h s prior to
eruptions (e.g. OMORI, 1914-1922; GORSHKOV, 1960; MINAKAMI, 1960;
ADAMS and DIBBLE, 1967; SmMOZURU et al., 1969). All earthquake
swarms that occur near volcanoes are not associated with eruptions
so that such increases in seismic activity cannot be used to predict
specific eruptions. Nevertheless, t h e swarms do provide a reliable indication of the potential increase in volcanic activity in a given region.
Local tilts of the ground of up to 1000 microradians are also typically
related to volcanic eruptions (OMORI, 1914-1922; OMOTE, 1942; MINA~ M I , 1950; DECKr~R and KINOSHITA, 1971). While continuous measurements of tilt are rare, they typically show (e.g. MINAr,AML 1942; EATON
and MURATA, 1960) inflation of the volcano over a relatively long period of time prior to an eruption and rapid deflation during an eruption. Other types of measurements of fumarole temperature, pressure,
and gas composition (e.g. N o 6 u c m and KAMIYA, 1963; MEmAYLOVand
NIKITINA, 1967; STOIBER and ROSE, 1970; MOXHAM, 1971; TAZIEFF, 1971;
and TONAm, 1971), and of gravitational, magnetic, or electric fields
(e.g. YOKOYAMA, 197I; JOHNSTON and STACEY, 1969) appear tO be of potential application in the future but thus far have not been extensively developed or observed.
This initial experiment shows that for the first time it is both
technologically and economically feasible to build a global volcano
surveillance system because of the advent of inexpensive satellite telemetry. Experience during this project demonstrates the feasibility
of building inexpensive, low power, reliable instruments for such a
system that can be installed in r e m o t e locations and can be expected
to run u n a t t e n d e d for a few years. Several details in the design and
deployment of appropriate low-power, inexpensive, and reliable instruments still need to be worked out.
Comparison of data from these new earthquake counters with
data from nearby standard seismometers shows that the counters do
normally indicate the level of seismic activity. During periods of high
--
319
seismic background noise, there may be a significant n u m b e r of spurious counts, but the existence and duration of such noisy periods
are reliably indicated by data sent by the earthquake counters, The
one eruption to occur to date, for which data were recovered via satellite, was on Volc~n Fuego in Guatemala and was preceded several
days by an abnormally large swarm of earthquakes. An apparent inflation of Volc~in Fuego was observed in the six months after the
eruption. In Hawaii, in addition to an increase in microearthquakes,
recordings from three tiltmeters showed large tilt changes as a result
of a rift eruption on May 5, 1973.
The design and response of this prototype global volcano surveillance system are summarized briefly in this paper together with an
evaluation of the potential use of this system. These topics are described in considerable detail in a report by WAR~, et al., 1974.
The ERTS
Data Collection
System
The Earth Resources Technology Satellite (ERTS) is an experimental satellite launched by the National Aeronautics and Space Administration (NASA) on July 23, 1972. It travels in a polar orbit with
a semi-major axis of 7286 km, circles the earth in 1.7 hours, and retraces the same flight path every 18 days. The satellite contains in
addition to several cameras, a radio relay system that receives and
instantaneously relays data from small transmitters located on the
ground and within sight of the satellite at any given time. These data
are then received at tracking stations in either Goldstone, California
or Greenbelt, Maryland if the satellite is visible from one of these
sites at that time. The data are relayed by telephone lines to Goddard
Spaceflight Center in Maryland and recorded. Within 90 minutes the
data are processed and relayed by teletype to different users. Data are
normally received from each remote site that is within 40 degrees of
a tracking station three to five times during a 2-hour period every
12 hours.
The transmitters for sending data to the satellite were provided
by NASA (1972). Each transmitter accepts 64 digital bits in serial or
parallel format or accepts 8 analog voltages (0 to + 5 volts) that it
converts to 8-bit digital words. In this study the seismic event counters are interfaced to the serial digital input. The tiltmeters are connected to the analog inputs. A digital controller in the transmitter
- - 320 - adds to the data a 15-bit preamble, a 12-bit address code that is
unique to each transmitter, and 4 bits runout. These 95 bits are convolutionally encoded, f o r m a t t e d to a nonreturn-to-zero level, and converted to a biphase format. This f o r m a t t i n g provides a r e d u n d a n c y
in the data that can be t~sed to detect transmission errors. This signal
is then t r a n s m i t t e d in a 38 millisecond b u r s t at a f r e q u e n c y of 401.55
MHz and at a minium p o w e r of 5 watts. The transmission rate can be
set to once every 90 or 180 seconds and the average p o w e r consumption is 0.085 w a t t s at 24 volts d.c. The t r a n s m i t t e r weighs a b o u t 7 kg
(Fig. 2).
Fro. 2 - D e m o n s t r a t i o n set-up of t r a n s m i t t e r for sending data to the satellite (left),
multi-level seismic counter (center), and one year's supply of batteries (right).
Except for geophone (left, front) and t r a n s m i t t e r antenna (not shown), components are placed in a metal box for t r a n s p o r t a t i o n and semi-permanent
field installation.
A crossed dipole a n t e n n a with a ground plane that is 1.2 m in
diameter was provided with each transmitter. During the latter half
of 1973, a m o r e c o m p a c t helical antenna, 7 c m in d i a m e t e r and 28 cm
high, was installed at sites in Washington and California w h e r e ice
and s n o w loading are a problem.
Analysis of over 72,000 messages relayed through the satellite for
this p r o j e c t s h o w s that up to 8 % of the messages f r o m the m o r e
distant t r a n s m i t t e r s contain trasmission errors, b u t that all such
errors were p r o p e r l y identified by the system as a result of the redundancy used in the data encoding. Thus the satellite does relay the
data reliably.
A data collection system such as that used on the ERTS-1 Satellite provides a relatively low cost m e t h o d of collecting data f r o m rem o t e sites a r o u n d the world. It w o u l d cost on the o r d e r of $ 3.5 rail-
--
321
--
lion to build and launch a data collection satellite and build and
operate four ground receiving stations around the world for the exp e c t e d 5 - y e a r l i f e o f t h e s a t e l l i t e ( E . PAinTER, o r a l c o m u n i c a t i o n s ,
1973). I t is r e a s o n a b l e t h a t s u c h a n o p e r a t i o n a l s a t e l l i t e c o u l d r e l a y
d a t a f r o m 10,000 g r o u n d s t a t i o n s m o n i t o r i n g v o l c a n o e s , s t r e a m f l o w ,
rainfall, pollution, and many other phenomena
for many different
agencies. For these purposes data from any one site could be relayed
f o r l e s s t h a n $ 100 p e r y e a r . T h i s c o s t c o u l d b e s u b s t a n t i a l l y r e d u c e d
i f t h e s a t e l l i t e w e r e l a u n c h e d <<p i g g y - b a c k ,, w i t h c o m m e r c i a l o r m i l i t a r y s a t e l l i t e s . T w o s a t e l l i t e s m i g h t b e l a u n c h e d 180 d e g r e e s o u t o f
phase to provide reliable coverage every 6 hours.
An alternative system could utilize a stationary satellite that could
provide data at any regular interval or at any time. This type of sate l l i t e w o u l d c o s t a b o u t $ 12 m i l l i o n to b u i l d , l a u n c h , a n d o p e r a t e ,
and several satellites would be needed to cover the volcanic zones in
t h e w o r l d . T h u s s u c h a s y s t e m is n o t a s e c o n o m i c a l a s a p o l a r o r b i t i n g s a t e l l i t e u n l e s s t h e r e is a w i d e d e m a n d f o r i t s u s e .
Design
of
the
Tiltmeters
The borchole t i l t m e t e r used in the surveillance p r o g r a m to m e a s u r e e a r t h
tilt is a bull's-eye level b u b b l e whose electro-chemical liquid is an active p a r t
of a resistance bridge forming a two-axis level d e t e c t o r (Cooi,itr, 1970). The
bridge is excited by a 3 KHz square wave, and the resultant signal is demodulated to give two electrical o u t p u t s that a,'e a function of rotations about two
orthogonal horizontal axes. The level bubble is housed in a 5-cm d i a m e t e r
stainless steel tube I rn long for borehole elnplacemcnt and is wired to external
electronics by cables contained in flexible conduit, thus allowing the bubble
sensor to be isolated in the most t e m p e r a t u r e - s t a b l e part of the site. S t a b i l i t y
of the tiltmeter was tested by o p e r a t i n g 8 m e t e r s on a single base (KnHI.ENBERGER,
cl a#., 1973), by o p e r a t i n g one m e i e r hanging by a thin wire (JoH-~s'r0~, 1974),
and by o p e r a t i n g one m e t e r in a tub of sand next to a mercu~Ly-tube m e t e r
in San Francisco that had operated reliably for six years. All tests show that
the m e t e r has a drift of less than 0.6 m i c r o r a d i a n s per m o n t h with a 95 %
confidence level.
Installations in Hawaii and G u a t e m a l a consist of a pit 1 m in d i a m e t e r
and 2 m deep, cased with steel culvert and covered with a steel lid. F r o m the
b o t t o m of this pit, a 14-cm hole is b o r e d down arlolher 2 m. In this hole is
placed a 10-cm pipe sealed at the lower end aud packed in place with clean
fine sand or with local material Ihat is sull]cieutly dry anti easily packed. The
tiltmeter sensor is placed in this pipe and packed with line (80-rnesh~, pure
~iliea sand. During ernplacemerit of the tiltmeter, the X and Y o u t p u t s are
observed on an oscilloscope and the m e t e r n~,ovcd to a level position. After
- - 322
mechanical zeroing, the tiltmeter sensor is completely buried in sand, and the
10-cm pipe filled to the top. The tiltmeter electronics are placed on a shelf or
low table in the pit. In some installations batteries, power supply, and (where
used) analog telemetry equipment also occupy the same pit. The satellite transmitter and batteries are typically placed in a steel box nearby. This installation
technique was developed over several years of trial and error and has proved
quite suitable (ALLEN, 1972; ALL~N, et al., 1973). The surface ma:terial at the
various sites include, in Hawaii, pumice from tile 1959 eruption as well as
older weathered ash at other sites and, in Guatemala, lightly weathered ash
with a low clay content.
At Lassen Peak, a 10-cm diameter hole was drilled 1.5 m into dacite. The
tiltmeter was placed in the hole and packed with sand as described in the
previous paragraph. The hole was then capped with wax. The tiltmeter electronics box was placed on the surface a few meters away with cables leading
to batteries and the satellite transmitter.
Results from the Ttltmeters
In late 1972 t h r e e level-bubble t i l t m e t e r s were placed a r o u n d Kilauea Caldera in Hawaii. The d a t a w e r e t e l e m e t e r e d via cables to the
H a w a i i a n Volcano O b s e r v a t o r y w h e r e they w e r e r e c o r d e d and also
c o n n e c t e d to one satellite t r a n s m i t t e r . These t i l t m e t e r s generally oper a t e d reliably b u t w e r e d a m a g e d several times b y lightning that appar e n t l y i n d u c e d high electric c u r r e n t s in the long cables. This p r o b l e m
was r e m e d i e d b y using a light e m i t t i n g diode a n d p h o t o - t r a n s i s t o r to
isolate electrically the t i l t m e t e r and electronics in the g r o u n d f r o m
the long cables.
The r e c o r d f r o m the east-west c o m p o n e n t of the level-bubble tiltm e t e r n e a r U w e k a h u n a vault in H a w a i i c o m p a r e d w i t h s i m u l t a n e o u s
r e c o r d s f r o m a m e r c u r y - t u b e t i l t m e t e r a n d a w a t e r - t u b e t i l t m e t e r oper a t e d in the vault f r o m D e c e m b e r 1972 to F e b r u a r y 1973 is shown in
Fig. 3. During this 3-month p e r i o d t h e r e was a l m o s t no tilt activity at
Kilauea, and data f r o m all t h r e e i n s t r u m e n t s agree within a few
m i c r o r a d i a n s . The r a t h e r noisy r e c o r d f r o m the w a t e r - t u b e t i l t m e t e r
reflects the relatively low c o n s i s t e n c y of readings to be expected f r o m
this type of i n s t r u m e n t .
The r e c o r d of b o t h c o m p o n e n t s of tilt at the S u m m e r Camp site
(Fig. 3), b e f o r e the i n s t r u m e n t was d a m a g e d by lightning on May 28,
shows no a p p r e c i a b l e change in tilt u n t i l the v o l c a n o began deflating
on May 5, only t h r e e h o u r s b e f o r e an e r u p t i o n b e g a n .-about 7 km to
the s o u t h e a s t at P a u a h i Crater. T h e r e was no n o t i c e a b l e change in
tilt r e s p o n s e to the m o r e t h a n 50 cm of rain t h a t fell in March.
- - 323 - The detailed tilt response of the level-bubble tiltmeters at
S u m m e r Camp, Uwekahuna, and Ahua to the collapse of the s u m m i t
of Kilauea on May 5, 1973, are s h o w n in Fig. 3 At this time only the
S u m m e r Camp tiltmeter was fully operational. The U w e k a h u n a a n d
Ahua meters had excessive long term drift ( a b o u t 20 microradians per
month) after they were c~amaged b y lightning in February. These meters, however, provided coherent short-term records. The instataneous
:......
'°. . . . . . .
...
t .;L
~i:'.'2~ :2:
-~'~---~
Fro. 3 - Typical data from three til,tmeters located a r o u n d the central caldera of
Kilauea Volcano, including a collapse of the caldera area on May 5, 1973. The
data in the upper left, separated t o t clarity, shows a comparison of the new
level bubble tiltmeter with two o t h e r types of tiltmeters operated in Hawaii
for several years.
tilt vectors from the three i n s t r u m e n t s are s u p e r i m p o s e d on a m a p
of the s u m m i t area in Fig. 4. The data from S u m m e r Camp and Uwekahuna s h o w collapse toward the center of Kilauea Caldera, w h e r e a s
the data from Ahua suggest collapse into the caldera followed bv collapse toward the eruptive center at Pauahi Crater and H e a k e Crater• A
comparison of these data with a chronology of the eruption provided by
the staff of the Hawaiian Volcano O b s e r v a t o r y s h o w s a close correlation. The eruption of lava began in Pauahi at a b o u t 1025 hrs and
ended by 1200 hrs with all b u t a b o u t 20,000 m 3 of lava draining b a c k
down the vent by 1230 hrs. Note the tilt at Ahua at this point was
inward t o w a r d Kilauea Calder, a and H a l e m a u m a u . At a b o u t 1255 hrs
a new lava o u t b r e a k was spotted near H e a k e Crater. Lava e r u p t e d
from many fissures, increasing until a b o u t 1500 hrs and ending a b o u t
1700 hrs, leaving about 440,000 m 3 of new lava in H e a k e Crater and
a b o u t 500,000 m ~ of new lava flows outside the crater. During this
- - 324 - phase of the eruption, the tiltmeter at Ahua s h o w e d r a p i d tilt d o w n
t o w a r d the eruptive center until a b o u t 2000 hrs w h e n the tilt rate
decreased. The tilt continued on for at least a n o t h e r 16 hours in the
same direction. These three tiltmeters s h o w e d m o r e details of a summit collapse than previously available.
UWEKAHUN
/A~ " ' - ~ ' ~
0000%o'0
.,!~,~E'i,oo ~
///)~K,
"4,o
KILOMETERS
1600 1800
~00 ~ / 2000
I000~
2200
0 6 ~ X
ENDAT
HEAKE
G CRATER
ATER
FIG. 4- Cumulative changes in tilt (plotted vertically) for the three tiltmeter sites
shown as triangles. Tilt vectors are plotted for every two hour period during
the May 5-6, 1973, eruption of Kilauea. The eruption took place in and around
Heake and Pauahi Craters.
Volc~n Fuego in Guatemala erupted in mid-February a b o u t two
m o n t h s b e f o r e the tiltmeter w a s installed and has since been very
quiet. Tilt d a t a f r o m April 9, 1973 (Fig. 5) suggest that Volc~n Fuego
swelled a b o u t 35 microradians b y early August and has r e m a i n e d quiet
since that time. Although it is still p r e m a t u r e to conclude the significance of this swelling, it might be i n t e r p r e t e d as evidence that Fuego
is p r i m e d for m o r e eruptive activity since it has swollen rather than
subsided.
G r o u n d tilts of up to 150 microradians were r e c o r d e d on Volc~n
Pacaya in G u a t e m a l a b e t w e e n J u l y and December, 1973. The flank of
the volcano a p p e a r e d to tilt o u t w a r d 150 m i c r o r a d i a n s and tilt back
10
- - 325 - nearly the same amount, all during a period b e t w e e n active extrusions
of lava. The m a g n i t u d e of the tilt seems large and m a y b e explained
b y the p l a c e m e n t of the m e t e r in a region of block-faulting on the
west flank of the volcano. A second tiltmeter has been installed on
the southeast side of Pacaya to examine the apparent tilting in m o r e
detail.
Tilt data recorded at Lassen Peak in California during the winter
of 1972-1973 show significant changes in tilt (Fig. 6). The a u t h o r s be-
Tiltmeter
I Km
~day I00
~120
Y~ 140
160
"N,ISO
I0~ Rad
"~200
\
"~220
Z~240
5 260
Fl(;. 5 - C u m u l a t i v e c h a n g e s in tilt p l o t t e d by v e c t o r s at 20-day intervals for a t i l t m e t e r
installed about two m o n t h s a f t e r the e r u p t i o n began, F e b r u a r y 22, 1973, at Volcfin Euego in Gatemala. Since April 9, 1973, day 100, the volcano a p p e a r s Io
have gradually in|lated.
lieve the tilts measured are related to freezing and thawing of w a t e r
ill the joints in the rock o u t c r o p containing the tiltmeter and to a
leak in the case of the tiltrneter in the spring. If this interpretation
is correct, then it appears, as suggested by Kl~x6 (1971), that shallow
installations of tiltmeters in volcanic ash or sand are p r o b a b l y better
than shallow installations in solid rock.
These results show that the portable, easy-to-install tiltmeters,
with a sensitivity of 1 rnicroradian, are w o r k i n g quite stably at several
sites in the field, that they are recording real tilts associated with vol-
- - 326 - c a n i c a c t i v i t y , a n d t h a t t h e y a r e n o t a f f e c t e d b y h e a v y local r a i n f a l l ,
e x a m i n e d in d e t a i l in H a w a i i a n d G u a t e m a l a (WARD, et al, 1974). Cont i n u e d r e c o r d i n g f o r a m u c h l o n g e r p e r i o d o f t i m e is n e e d e d b e f o r e
t h e v a l u e of t h e s e tilt d a t a in p r o v i d i n g e a r l y w a r n i n g of e r u p t i o n s
c a n b e c l e a r l y c a l c u l a t e d . At this p o i n t t h e m o s t i m p o r t a n t o b s e r v a t i o n is t h a t l a r g e c h a n g e s in tilt f r o m 30 to p o s s i b l y 150 m i c r o r a d i a n s
have been observed on three volcanoes occurring before, during, and
a f t e r e r u p t i v e p e r i o d s , w h e r e a s n o s i g n i f i c a n t c h a n g e s in tilt, e x c e p t
LASSEN P E A K
1972-1973
22-
-~,~
~_~
t
22
SoMe k~ mlczorc~a~s
25
I
OCT
I
NOV
I
DEC
I
JAN
t
FEB
I
MAR
APR
MAY
Fro. 6 - Recorded tilt of ground on Lassen Peak through the winter of 1972-73. Melting snow crushed the transmitter antenna in late April, 1973.
t h o s e a p p a r e n t l y r e l a t e d to f r e e z i n g a n d t h a w i n g , h a s b e e n n o t e d o n
L a s s e n P e a k , t h e o n e v o l c a n o w h e r e tilt is b e i n g m o n i t o r e d t h a t h a s
h a d n o o b s e r v e d v o l c a n i c a c t i v i t y f o r m o r e t h a n 50 y e a r s .
Design of the Seismic Event Counters
Since the satellite relay system can only transmit 64 digital bits of data
each 12 hours and seismic data are normally gathered at a rate of about
20 million digital bits during the same time, it is necessary to process the
seismic data at each remote site. The data that have been observed most
often to change prior to volcanic eruptions are simply the numbers of small
earthquakes of different sizes. These numbers typically change by orders of
magnitude before eruptions. Some experience was developed from automatic
earthquake counters built by P~a~ET (1937) and DECKER(1968). Their experience
plus experience from prototype seismic event counters built by several different
individuals and companies (unpublished results from T. Matumoto, J. Unger,
Kinemetrics, Inc., Systron-Donner, and Bendix) has been used to develop a
significantly new design in seismic event counters for this project.
12
--
327
--
The seismic event c o u n t e r system (Fig. 7) utilizes a geophone, which produces a voltage p r o p o r t i o n a l to ground velocity. The geophone was selected
to have a natural f r e q u e n c y of 4.5 Hz because of the high-frequency characteristics of m i c r o e a r t h q u a k e s and is o p e r a t e d at 80 % of critical damping. An
a t t e n u a t o r switch is placed between the second and third amplifier stages for
field gain a d j u s t m e n t . The amplified signal is full-wave rectified for the leveldetection circuit. The rectified signal is then connected parallel to f o u r level
detection circuits whose reference voltages are set by a voltage divider circuit
across a zener diode. The m o s t sensitive level d e t e c t o r puts out a logic 1
Typical
Wove For m,~
--~;~de.
Mcl~l sensit*ve chan~el
' Amplif;er If,trgr
8Qnd pass
O 1 to 3 0 H z
Precision
Re¢lif L,r
Sp~ke
Reiec~
Reje¢~
N~se
4>
¢omp~rater
r
I
C~mparaTDr
~mory
r~l similtar digital
clrcuIls
_ _ .
c*, ~,
P,,....I
j
Comparator
'.i c,.r,,,
:
~ Ya ttansm;ner
[
...............
i
request
Compar~lor
Leasl sen~Hve
f~=nnel
Fro. 7 - Block diagram of the multi-level seismic even counter logic.
o u t p u t pulse when the level of any peak in the rectified signal exceeds the
40 to 50-millivolt reference voltage. The output from the level d e t e c t o r then
goes to a divide-by-10 counter, reset by a one-shot t i m e r every 1.2 seconds. The
counter requires I0 pulses to occur in 1.2 seconds to register a count. This
r e q u i r e m e n t m a k e s the counter insensitive to seismic signals with predominant frequencies of less than 4.2 Hz. When a count is registered, a n o t h e r oneshot t i m e r is triggered to inhibit the count circuit for 15 seconds. The event
counter is then registered in a 10-bit binary counter. Three additional level
detectors at 12 db, 24 db, and 36 db below the m a x i m u m g a i n settings function
in the same way. The 12 db interval gives the seismic event counter a 36 db
range of detection. The extended range allows detection of m a j o r changes in
the seismicity and d e t e r m i n a t i o n of the n u m b e r s of e a r t h q u a k e s of different
sizes.
In o r d e r to distinguish noise from earthquakes, the output of the 15-second
one-shot timer in the noise reject circuit goes to an elapsed time interval counter. If this noise reject signal is on for 60 to 70 seconds, or in o t h e r words if the
counter is inhibited for such a period of time, one noise count is registered.
13
- -
328
- -
There are inhibit c o u n t e r s for the three m o s t sensitive levels of detection, and
all the data except the seven m o s t significant bits of the third c o u n t e r are
t r a n s m i t t e d , t o g e t h e r with a parity bit.
As a back up system for data recovery, a t h e r m o g r a p h i c r e c o r d e r records
data at a 6-hour print rate. The p r i n t e r uses a simple 4 by 5 m a t r i x system for
binary-coded data f r o m the counters. At the 6-hour p r i n t r a t e a 75 m e t e r roll
of p a p e r tape lasts a p p r o x i m a t e l y one year. A solenoid-activated stepping m o t o r
)0 )
,o
"~
~
,
~,5
02
Ol oOs
g'~
oo'tool
~ioooooo
x
#lI
io0o
_=
B
OO
....
,../;/:
--%
,o,
,?,
~,equ~aty
~,
~ H,~ )
F1G. 8 - Response curves for the multi-level seismic event counter (dashed lines) and
high gain short period seismographs (solid l~nes). Lower set of curves ate for
the electronic responses of both systems and upper curves include geophone
respinses.
is used for the tape advance m e c h a n i s m and is the only m o v i n g part in the
system aside f r o m the geophone,
The geophone assembly and seismic event c o u n t e r weigh a p p r o x i m a t e l y
10 kg. Power to the unit is supplied by external batteries. Less than 50 milliwatts of p o w e r are r e q u i r e d for a 24-hour period. As a result of the extensive
use of CMOS integrated circuits the seismic event c o u n t e r can be operated continuously for years.
Eleven short-period, high-gain s e i s m o m e t e r stations w e r e installed a d j a c e n t
to event counters in Washington, California, Guatemala, E1 Salvador, and Nicaragua (Table 1) to allow c o m p a r i s o n of the counts with standard seismic records. The s e i s m o m e t e r telemetry systems installed are similar to systems
described by SMITH. et al. (1971) and WESSON. et al. (1973). These s e i s m o m e t e r s
have n a t u r a l frequencies of 1 or 2 Hz and are o p e r a t e d at 80 96 of critical damping. After being t e l e m e t e r e d via V H F radio. FM signals are d e m o d u l a t e d and
subsequently r e c o r d e d on heat-sensitive p a p e r at 60 m m / m i n . The unit magni14
- -
3 2 9
- -
fication curves for both the event counter system and the standard short-period
1-dgh-gain systenl (Fig. 8) show that approximately three times more output is
derived from the event counter system geophone.
Verification of the Seismic Event Counters
The seismic event counter is designed to process seismic data that
occur at the rate of about 20 million digital bits per twelve hours and
to provide only 64 digital bits of data in the same time. This compression is accomplished by requesting to know only the n u m b e r of earthquakes with amplitudes greater than four discrete levels and by designing an electronic circuit to detect e a r t h q u a k e s and m e a s u r e their
amplitude. Detecting e a r t h q u a k e s is not always easy. Even two seismologists might disagree w h e t h e r a small event recorded at only one
station is an e a r t h q u a k e or is spurious noise caused by people, animals, wind, and so on. Thus, it is of prime importance to establish
how reliably the new event counters detect and count e a r t h q u a k e s
and u n d e r what conditions these counts m a y be c o n t a m i n a t e d by
counts of spurious seismic noise.
One event counter was placed in Hawaii next to a s t a n d a r d seismic station with data telemetered to lhe Hawaiian Volcano Observatory. A special interface board was attached to the event c o u n t e r that,
with the use of seven one-shot timers, put out d.c. pulses of varying
length, depending on which of the seven counters in the seismic event
counter was activated at a particular time. These pulses m o d u l a t e d
a voltage-controlled oscillator. The resulting signal was telemetered
with the seismic signal, discriminated at the central recording site,
and recorded next to the seismic trace. This system, in addition to
the data printed by the event c o u n t e r every 1.5 hours, gave a direct
indication of what the event c o u n t e r was detecting.
Comparison of the data f r o m the event c o u n t e r printer with the
corresponding n u m b e r of pulses on the telemetered trace for 1.5 h o u r
intervals shows a one-to-one correspondence. Thus, the special interface circuit worked as designed and did not appear to introduce spurious data. As expected, all e a r t h q u a k e s with suiticient amplitudes
and frequency content were counted. The magnitudes of e a r t h q u a k e s
counted only on the most sensitive detection level range from 1.1 to
2.5, and the magnitudes of e a r t h q u a k e s counted on the second most
sensitive detection level range from 1.6 to 2.8. This range of earthquake m a g n i t u d e s is expected since a m p l i t u d e s vary with distance to
15
- - 330 - the epicenter. High ground noise caused by wind and vehicle traffic
on a road 0.25 km from the event counter site caused spurious counts
on the most sensitive counting level. Noise counts were also recorded
during periods of wind and serve to indicate periods of probable high
wind activity. ¢, Noise ~ counts in general were not triggered by pass200 MT. S T . H E L E N S , WASI-IINGTON
UJ
(/)300~
I _~,oo~
WIND
~
~ ....
l-~''
' 'l'
....
iii
o
i ......
I i i i i i ii m ii
""' ' ' t ,
,,,
z
,
20
"~T
....
25
rb.
[1~1
30
5
AUGUST
I0
I5
SEPTEMBER
1973
FiG. 9 - Seismic events c o u n t e d by t h e event c o u n t e r on Mt. St. Helens in W a s h i n g t o n
c o m p a r e d to e a r t h q u a k e s detected by a seismologist u s i n g s e i s m o g r a m s f r o m
a s e i s m o g r a p h s t a t i o n o p e r a t e d nearby.
ing vehicles. Thus, it is important that sites be located away from
sources of cultural noise to increase the reliability of the data.
A comparison of events counted electronically on Mt. St. Helens
in Washington with earthquakes counted by a seismologist studying:
typical seismograms is shown in Fig. 9. This figure shows a close correlation of event counts to earthquakes except during periods of high
wind when there are significant numbers of noise or counts. These
types of data together with the detailed comparison of the event
counter done in Hawaii provide the basis for the following conclusions about the reliability of these seismic event counters.
16
- - 331 - Nearly all local earthquakes are detected and counted unless they
are too small to have 10 peaks of their full-wave rectified signal occur
above the threshold of the m o s t sensitive counter or they occur at a
time when the counter is inhibited by high background noise or by
the occurrence of another earthquake less than 15 seconds earlier. A
typical earthquake counted is s h o w n in Fig. 10A. Failure to detect
L-2:--_L 2 : ~:"---:.:--":J:- : E - ~ 5 . _ 5 : ~ : . U
-- . . . . . . . . .
•
:-Fill
": 2 T ~ T a i
N
p
-
-
-
-
,I-~i',,
.
"
"
"
-
i.~l~
Fu6. 10- Examples of microcarthquakes from two Central American volcanoes. The
top example from Izalco Volcano, El Sah, ador, illustrates the high frequency
microearthquakes that the event counters normally count reliably. Recorded
on the lower record are low frequency microearthquakcs recorded at Pacaya
Volcano, Guatemala• These events, associated with ash eruptions, are often
not counted by the event counter. One minute between time marks.
17
332
earthquakes during periods of high background noise is not serious,
since the noise will inevitably cause some spurious counts and the
high noise counts will flag the n u m b e r of event counts as being suspect. Failing to detect earthquakes that occur rapidly in succession
is also not serious since such events will occur rarely except during
swarms of earthquakes when the percentage of such undetected earthquakes will be low.
The only other types of local earthquakes observed in this study
that are typically not counted are low frequency, emergent events like
those shown in Fig. 10 B. The event counter is totally insensitive to
frequencies below 4.2 Hz because of the criterion for counting that
requires 10 peaks of the full wave rectified signal occur above a given
threshold in 1.2 seconds. This criterion was chosen to reduce the number of teleseismic events counted, but we have found out that at some
volcanoes, particularly Pacaya, San Cristobal, and St. Helens in this
study, there are m a n y local earthquakes with significant low-frequency
content. Such events have also been noted at Rainier (URGER and DECKER, 1970), Kilauea (R. KOYANAGI,personal communication, 1973), and
Augustine (J. KIENLE, personal communication, 1973). These events
are generally not counted because they begin with small ground amplitudes that build up to a m a x i m u m in about 5 seconds. In m a n y
instances one peak in amplitude may start the detection counting circuit, but more peaks in the signal large enough to exceed the threshold
do not occur for another 0.5 or 1 second. These events might well be
detected if the criterion were ch.anged to something like 10 peaks in
2 seconds. Experimentation with this criterion continues.
Spurious counts can be caused by g r o u n d noise from cultural
sources but are predominantly caused by wind induced ground noise.
Such spurious counts either are infrequent or are usually accompanied by significant noise counts. Noise f r o m cultural sources can be
successfully avoided at nearly all sites by placing the instrument in
remote areas away from frequently traveled roads or trails. Wind-induced ground noise is reduced by placing the sensors at low elevation,
sheltered from high winds, away from trees, at 25 to 30 m from the
large antennas, and by using smaller types of antelmas. Major changes
in the b a c k g r o u n d noise level caused by wind can not be totally
avoided. Spurious even counts caused by high wind, however, can
be readily identified by the simultaneous registration of high noise
counts. Moderate levels of ground noise m a y cause spurious event
counts b u t may not trigger counts. Such cases can usually be identi18
333fled, however, since when a large number of event counts triggered
by earthquakes is recorded on one channel, approximately 15 to 30 %
of these events should be counted on the next most sensitive channel.
This relationship occurs because earthquakes are known to be distributed in size according to the GUTENBERGand RICHTER (1949) relationship modified by SuzuKi (1953)
logN=
-blogA+C
where N is the number of earthquakes with maximum amplitude
greater than or equal to A. C and b are constants for a given region
and b, which is of interest here, is typically between 0.8 and 0.9
(ISACKS and OLrVER, 1964; PAGE, 1968) but may be as high as 3.5 for
shallow earthquakes within a kilometer of the summit of some volcanoes (MINAKAMI, 1960).
Thus, an order of magnitude increase in seismic events accompanied by a small increase in events on the next most sensitive channel, and no noise or inhibit-time counts, can be reasonably assumed
to indicate an order of magnitude increase in seismicity. The occurrence of may inhibit-time counts for the most sensitive channel would
indicate that the events are probably spurious counts and do not represent a change in seismicity. It is possible that a short-term (such
as one day) increase in seismicity might occur during a storm when
the background noise is high. The probability of such an occurrence
is low. Furthermore, an increase in seismicity related to a large eruption can be expected to be several orders of magnitude (MINAKAMI,
1960, 1968; SmMOZtmU, 1969), SO that noise on one channel will not
obscure the changes on the lower gain channels.
Thus, we can conclude that the event counters do normally and
quite reliably indicate the level of seismicity within an order of magnitude. These seismic event counters are significantly more reliable
than earlier counters because they combine the following design features:
a) There must be several cycles of ground motion above a given
threshold.
b) There must have been no peak above the threshold in the
15 seconds preceding the detected signal.
c) Some indication of the m, ise level is provided.
d) Several different thresholds are used.
7
19
- - 334 - There are a number of tradeoffs in the choice of these design features and particularly in the choice of appropriate time costants.
Perhaps the least obvious choice relates to what should constitute a
noise count. In the present design a noise count is registered if at
least one peak occurs above a given threshold every 15 seconds for a
continuous period of 60 to 70 seconds. The resulting count is significantly less than one would get by counting the individual periods of
15 second inhibits and dividing by 4. The method adopted will not
count many short periods when the counter is inhibited or periods
when the noise level is causing separate inhibits. The main benefit of
this method is that it will not count the codas of most local earthquakes as periods of noise.
The results from these new seismic event counters are very encouraging and show that significant compression of seismic data can
be done with processing by relatively simple electronic circuits. These
counters are adequate to operate in a global volcano surveillance system, but we also have begun to think of ways to improve the amount
and quality of the compressed data and thus are continuing to experiment with other types of counters.
Changes in Seismicity Observed
A microearthquake swarm occurred near Augustine Volcano in
Alaska between January 11 and 23, 1973. There is an excellent correlation (Fig. 1 1 ) b e t w e e n the numbers of microearthquakes observed
on a standard seismigraph (J. KIENLE, personal communication, 1973)
and those counted on the third most sensitive event counter channel.
The two mgst sensitive counter channels were inoperative because of
an electronic oscillation caused by low temperature that was not discovered until it was impractical to reach the instrument in the winter. Since the third channel is 16 times less sensitive than the most
sensitive channel, the number of events counted would be on the order of 16 times less than the number of earthquakes counted on the
seismograms. The actual difference is a factor of 22 and depends on
the value b of this particular swarm and the amplifications set in the
two different instruments.
No significant change in volcanic activity was noted to accompany or follow this swarm. Augustine is an island that is uninhabited
in the winter, however, so that a minor eruption might well go unde-
20
- - 335
-
-
tected. This volcano has been quite active since a violent explosion in
1883 during which a mud flow, or lahar, generated a destructive tidal
wave (KIF_~LE, et al., 1971). A lava dome has continued to grow in the
central crater by over 85 meters since 1957. This swarm clearly illustrates why a two or even three orders of magnitude change in seismicity does not necessarily allow one to predict a specific eruption.
However, such a change does indicate, particularly when these swarms
occur regularly, that this volcano is very active and has a relatively
7'00.
(b)
(al
35-
600.
30-
g5-
500'
8
20-
¢T 400"
BOO-
,,>, 15-
200-
I0
I00.
5
0
I0 '
'
0
'
,; .....
~o ....
JonuQry 1975
I0
15
20
Jonuary 1975
Fro. 1 1 - A histogram of (a) earthquakes counted on short period seismograms from
standard seismograph, Augustine Volcano, Alaska (J. KIENL~, personal communication, 1973), and a histogram of (b) events detected by a multilevel
seismic event counter showing excellent correlation on the third m o s t sensitive channel of detection.
high potential for eruption. After the seismic activity has been monitored for a period of months and years at a given volcano, it should
be possible to say with some certainty whether an eruption is likely
in terms of days, weeks, months, or years.
The number of seismic event counts recorded at Santiaguito Volcano in Guatemala from March through mid-October, 1973, are shown
in Fig. 12. The number of events increased by about an order of magnitude three times, hut each of these periods was accompanied by
a significant number of noise counts. Thus, no changes in the event
counts in Fig. 12 stand out as designating changes in seismicity. The
seismicity in this region does, however, stand out as generally high.
ROSE (1974) described a nu6e ardente eruption of Santiaguito on
April 19, 1973, that was the largest since 1929. The eruption occurred
21
- - 336 - on a cloudy night and was n o t observed. Only steady r u m b l i n g a n d
a s t r o n g o d o r of SO2 were detected by i n h a b i t a n t s 7 k m to the south.
The deposit a p p e a r e d to originate f r o m the Caliente Vent area of
Santiaguito. A second, smaller nude a r d e n t e occurred o n September
16, 1973, and originated at the toe of a lava flow coming off of the
dome. Rose and BoNIs (personal c o m m u n i c a t i o n , 1 9 7 3 ) s u g g e s t the
possibility t h a t both of these nude a r d e n t e s m a y have originated at o r
~_~°~
1 ,g...,L'~3
"E.u.r,o.',
APriL
~&Y
~
~6sterE~BE.'z~ "E.0Pr,~N"
,1~JKE
JULY
A~G~$¥
i S~pyE~||mr ,
19T3
FIG. 12- Histograms of seismic event counts and noise counts for two channels
of a seismic event counter installed within 7 km of an active dacite dome
at Santa Maria Volcano in Guatemala (Channel 2 shaded).
very near the surface and m a y not be related directly to changes in
the volcano at depths w h e r e e a r t h q u a k e s typically occur.
Seismic events associated w i t h the e r u p t i o n of Volcfin Fuego in
G u a t e m a l a are s h o w n in Fig. 13. This volcano e r u p t e d on F e b r u a r y
22, 1973, only nine days after the event c o u n t e r w.as installed. The
small eruption ended on March 2 a n d was confined only to the summit area.
An o r d e r of m a g n i t u d e increase in seismic events was observed
on the low gain (second m o s t sensitive) channel five days before the
e r u p t i o n and a similar, b u t larger, increase occurred on the high gain
(most sensitive) channel. These seismic events can be a s s u m e d to be
m o s t l y e a r t h q u a k e s since very few noise counts were recorded at the
same time. A similar b u t s m a l l e r increase was n o t e d on a c o u n t e r
operating 15 k m away f r o m Fuego, b u t no change at all was noted
on a c o u n t e r 30 k m away. Thus the change in seismicity was obviously
22
- - 337
in the vicinity of Fuego. During the e r u p t i o n t h e n u m b e r of events
remained high, b u t after the eruption very few events w e r e recorded.
It seems reasonable to a s s u m e that the level of activity p r i o r to Febr u a r y 13 was low and similar to that after March 4, b u t there is no
w a y to be sure.
A s e i s m o g r a p h installed next to the event c o u n t e r on Fuego began
o p e r a t i n g f o u r days after the eruption began. The level of h a r m o n i c
~OLCb.N
FUEGO,
"°!
7too,
~
z
l~
r
~
:::i
I
r
"7
~
T
,
,
:,Ltl
~.o
= 7~oo
~
~,oo
GUZ~TEM~XLA
~
.
L
1973
FIe;. 13 - H i s t o g r a m s of s e i s m i c e v e n t c o u n t s a n d n o i s e c o u n t s [ o r t w o c h a n n e l s o[ a
s e i s m i c e v e n t c o u n t e r i n s t a l l e d n e a r V o l c a n F u e g o in G u a t e m a l a in e a r l y
1973. N o t e t h e i n c r e a s e in s e i s m i c a c t i v i t y d u r i n g t h e e r u p t i o n w h i c h b e g a n
in F e b r u a r y 22, 1973. A h i g h gain s h o r t p e r i o d s e i s m o g r a p h i n s t a l l e d a d j a c e n t
to the e v e n t c o u n t e r r e c o r d e d tl~e relative ciaanges in the level o[ h a r m t m i c
t r e m o r . I n t e r v a l s of e r u p t i v e a c t i v i t y a r e i n d i c a t e d .
t r e m o r or ground noise that is c o m m o n l y believed to be associated
with u n d e r g r o u n d m o v e m e n t of lava w,as a p p r o x i m a t e l y 100 times
greater than the b a c k g r o u n d seismic noise one m o n t h after the eruption. As expected, there are large n u m b e r s of event counts and noise
counts on the two m o s t sensitive channels during the periods of high
tremor.
23
--
338
On March 18, two peaks of over 150 event counts recorded on
the highest gain counter chanrml correspond to a period of increased
tremor. There are no noise counts to suggest that this is a period of
high ground noise. Such a condition occurs w h e n the signal level is
just high enough to occasionally trigger event counts but does not remain above the detection level longer than a m i n u t e to trigger noise
counts. These peak in event counts can be readily identified as spurious, however, because there are not sufficient corresponding counts
on the next most sensitive channel.
This example from Fuego shows one type of increase in seismic
activity prior to eruptions that a surveillance system might monitor,
except that the time between the increase in seismicity and the eruption might, from historic accounts, be expected to increase as the size
of the eruption and thus the volcanic hazard increases. As recording
continues, long-term changes in seismicity from year to year should
be very significant for delineating changes in the state of volcanic
activity at different volcanoes.
System
Evaluation
The cost of installing an event counter, tiltmeter, and satellite
transmitter on a volcano in some quantity is about $ 5,000 or roughly
similar to installing this equipment with analog recorders located near
the volcano. The yearly maintenance cost for the satellite telemetry
equipment would at most be a small fraction of the cost of maintaining local on-site recording, however. Furtherm.o,re, experience during this project clearly demonstrates that the satellite telemetry system
is substantially more reliable than on-site recording methods commonly
used n o w unless a skilled technician is available in every region to
maintain the equipment. This high reliability of the satellite relay
system stems primarily from the fact that the self-contained, small
event counters, tiltmeters, and transmitters that can be located almost
anywhere and, based on experience in this study, are less susceptible
to battery failure and to damage by lightning, vandalism, and climatic
extremes than are equivalent systems connected to local recording stations by cable, telephone lines, or radio. The satellite system provides
for rapid data collection and rapid analysis by specialists. The standard local recording techniques provide more data when the equipm e n t functions properly b u t the equipment can be expected not to
24
- - 339 - function for long periods of time unless skilled technicians are paid
to be available at any time for repairs. Thus the maintenance of a local
analog recording system would cost at least an order of magnitude
greater more than maintaining the satellite system.
A satellite volcano surveillance system has two principal purposes:
basic research and early warning of increased hazard. Basic research ultimately is aimed at providing more reliable early warning
and even specific predictions of eruptions. The need for early warning
is probably best served by a satellite system with rapid data relay, in
which a few specialists could examine the data rapidly and notify
local scientists or officials when abnormal activity is noted. The need
for basic research might best be served by on-site recording or by
spending the same dollars on more intensive studies of fewer volcanoes. In order to weigh these considerations, continuing research is
directed toward demonstrating the types and reliability of data collected by the different approaches and determining how complete the
data, for example on seismicity, could be made by f u r t h e r developmcnt of on-site data processing schemes.
The economic impact of a global volcano surveillance system is
diflicult to estimate reliably. Such a system would safeguard lives and
reduce loss of mobile property by providing an early warning of the
reawakening of a volcano and criteria for judging the degree of unrest
of each volcano. Such information would aid in minimizing the disruption to the local economy caused by a volcanic eruption. Loss of
life depends critically on the location of population centers near volcanoes. As populations increase, so do concentrations of people on and
near the fertile flanks of rnany volcanoes. Thus the exposure to volcanic hazards continues to increase. Furthermore, as construction of
facilities such as the nuclear power plant just west of Mr. St. Helens
in Washington State proceeds, it becomes increasingly important that
early warning of an eruption is sufficient to consider closing down
such reactors. The ultimate risk from many volcanoes may be damage and loss of life resulting from failure of dams, nuclear power
plants, and other facilities caused by eruptive activity. An average of
sever.al hundred people per year have been killed by volcanoes during
the last 500 (VAN BEMMELEN, 1949) and 1000 (R. DECKER, personal
communication, 1974) years. Specific eruptions may kill over a hundred times this number of people in one small area and cause extremely large and concentrated economic loss. Even the reduction by a
25
- - 340
few percent in the n u m b e r of deaths from volcanic eruptions might
~, pay ,, the cost of a global volcano surveillance system.
The direct criteria for judging the degree of unrest of various
volcanoes provided by such a surveillance system would aid in landuse pl'anning for regions around volcanoes. Construction and develo p m e n t should proceed, i.f no other choice is possible, on those volcanoes showing the least a m o u n t of activity over several years and
should be avoided a r o u n d those volcanoes showing the greatest a m o u n t
of activity or showing rapid changes in activity.
A global volcano surveillance system would also focus use of
available resources on detailed studies of volcanoes that have the
higest probability of erupting, increase the efficiency of research on
prediction of specific volcanic eruptions, and thus decrease the related
costs in dollars and manpower. With present technology it is feasible
to install large, temporary, portable networks of i n s t r u m e n t s rapidly
anywhere in the world. Several international teams of experts could
be established to study those volcanoes m o s t likely to erupt. These
teams could move rapidly into areas where eruptions are considered
i m m i n e n t not only for research purposes b u t to advise local leaders
regularly and rapidly on the probability and possible scope of a potential eruption. The reliability of such predictions should increase
rapidly with a global surveillance system for focusing use of the research funds and taJents of m a n y different countries.
Conclusions
A global volcano surveillance system is now technologically and
economically feasible but more work is required to demonstrate that
such a system will be effective and reliable for predicting eruptions.
More specific results are as follows:
a) A new seismic event counter that reliably indicates order of
magnitude changes in seismicity has been designed, deployed at 19
locations, and thoroughly tested. This event counter is significantly
more reliable than any such system previously available.
b) A tiltmeter that can be installed quickly and easily has been
successfully deployed at seven remote locations and shown to operate
reliably and stably for at least a year.
c) The seismic event counters and tiltmeters have been interfaced with satellite transmitters and installed simply, reliably, and
securely in remote areas and in many environmental extremes.
26
--
341
--
d) The i n s t r u m e n t a t i o n has worked extremely well, b u t experience developed in this project is being used to develop even m o r e
reliable design, construction, and installation techniques. Even t h o u g h
we conclude that a global system is technologically feasible, there are
still many details in design that need to be w o r k e d out in an orderly
development program.
e) Local earthquakes were detected on all volcanoes m o n i t o r e d
in this study, suggesting that even the volcanoes of the Cascade Range
are not extinct.
f) An order of magnitude increase in seismicity was observed
several days prior to the eruption of Volcfin Fuego in Guatemala in
February, 1973.
g) Tilts of from 20 to 150 microradians were observed on the
volcanoes Kilauea, Fuego, and Pacaya.
h) A volcano surveillance system monitoring seismicity and tilt
at 1000 locations around the world could be installed and operated
for five years at a cost of about $ 11 million it the satellite relay system were shared with many users. This expense might be shared by
many countries.
The results of this initial project are extremely encouraging. It
appears that a global volcano surveillance system utilizing satellite
telemetry may be very practical and desirable and that it offers a
radically new approach to the surveillance of potentially hazardous
volcanoes. Considerable evaluation and development still needs to be
done, however, to demonstrate just how valuable such a system would
be and how soon it should be built.
Acknowledgements
The establishment of a network of i n s t r u m e n t s on 15 volcanoes
in five countries and the subsequent data analysis were made possible
by the generous support of the foliowing individuals and institutions:
Dr. Jurgen Kienle, University of Alaska, College , Alaska, and Mr.
William Feetham from Homer, Alaska. In California, the superintendent and staff of the Mt. Lassen Volcanic National Park. In E1 Salvador, Ing. Mauricio Aquino H. (Director) and Ing. Mauricio Cepeda,
Centro de Investigaciones Geotecnicas, Ing. Jose Gonzales G., Instituto
Geografico Nacional, and Mr. Albert Holburn, Inter-American Geodetic
Survey, U. S. Defense Mapping Agency. In Guatemala, I3.r. Samuel B o 27
- - 342 - nis, Mr. Oscar Salazar, Mr. Carlos Estrada Q., from the Instituto Geo..
grafico Nacional. In Hawaii from the Hawaiian Volcano Observatory,
Dr. John Unger, Mr. George Kojima, and Mr. Kenneth Honoma. In
Iceland, Mr. SveinbjSrn BjSrnsson, National Energy Authority. In Nicaragua, Cap. and Ing. Orlando Rodriguez, Ing. Arturo Aburto Q., Servicio Geologico. Mr. Leroy Anstead, Inter-American Geodetic Survey,
U. S. Defense Mapping Agency, Mr. Edward Hagie, Director of Mina
E1 Limon, Ing. Adolfo J. Benochio, and Ing. Jose Antonio Gonzales,
also from Mina E1 Limon. In Washington, Dr. Robert Crossen and
Mr. Lee Bond, Dept. of Geophysics, University of Washington. Also
the superintendent and staff of Mt. Rainier National Park, the staff
of the Mt. Baker National Forest and the St. Helens National Forest.
Mr. Jim Ellis, Mr. John VanSchaack, and Mr. Wayne Jackson from
the U. S. Geological Survey assisted with many pre-installation problems and equipment preparation.
We are also grateful for the information and advice given by
Dr. Robert Decker and Dr. Richard Stoiber from Dartmouth College.
Funds for this work were provided by NASA under Contract
S-70243-AG-2 and from the EROS program of the U. S. Geological
Survey.
References
ADAMS, R. D. and DmRLE, R. R., 1967, Seismological studies of lhe Raoul Island eruption, 1964, New Zealand Jour. Geology and Geophysics, 10, p. 1348-1361.
ALLEN, R. V., 1972, A borehole tiltmeter for m e a s u r e m e n t s at tidal sensitivity. Bull.
Seism. Soc. Amer., 62, p. 815-821.
- - ,
WOOD, D. M., and MORTENSEN, C. E., 1973, S o m e i n s t r u m e n t s and techniq,les for m e a s u r e m e n t s o[ tidal tilt. Phil. Trans. R. Soc. Lond., 274, p. 219-222.
COOPER, G. L., 1970, Development and use o[ a two-axis electrolytic bubble level as a
precision vertical reference and tilt indicator. Am. Inst. Aeron. and Astron. Paper
70-949, Sta. Barbara, Calif., 11 pp.
DECKER. R. W., 1968, A ~eismic event Counter for Active Volcanoes. Bull. Sets. Soc. Am.,
58, p. 1353-1358.
- and KIN0SHTTA, W. T., 1971, Geodetic Measurements, in The Surveillance and
Prediction of Volcanic Activity. UNESCO, Earth Science Monograph 8, Paris, p. 47-74.
EATON, J. P. and MUP,ATA, K. J., 1960, H o w volcanoes grow. Science, 132, p. 925-938.
GORSfIKOV, G. S., 1960, S o m e results of the seismometric investigations, at the Kamchatka Volcanological Station. Bull. Volcanol., 23, p. 121-128.
GUTENSERG, B. and RICH'r~, C. F., 1949, Seismicity of the Earth. Princeton Univ. Press,
Princeton, N. J., 210 pp.
ttAro.ow, D. H., 1971, Volcanic Earthquakes. Unpublished M. A. Thesis, Dartmouth College, Hanover, N. H., 66 pp.
28
- -
343
h
JOrtNSTON, M. J. S. a n d STACEY,F. D., 1969, Transient magnetic anomalies accompanying
volcanic eruptions in N e w Ze!and. Nature, 224, p. 1289-1290.
ISACKS, B. a n d OLrVER, J., 1964, Seismic waves w i t h frequencies fro.m 1 to 100 cycles per
second recorded in a deep mine in n o r t h e r n N e w Jersey. Bull. Seis. Soc. Am.,
54, p. 1941-1979.
KIENLE, J., FORBES, R. B., a n d I-I~LoW, D. H., 1971, Recent microearthquake s w a r m
activity at Attgustine Volcano, Alaska (Abstract). Trans. Am. Geophys. Un., 52,
p. 925.
KIrqG, G. C. P., 1971, The siting of strainmeters f o r teleseismic and tidal studies. Royal
Society of New Zealand, Bull. 9, p. 239-247.
KOHt~rqBER~ER, C. W., COOPER, G. L., a n d SCrlMARS, W. T., 1973, Dynamic properties of a
n e w biaxial tiltmeter (Abstract). Earthq. Notes, 44, p. 11.
MENIAVL0V, I. A. a n d NIKITINA, L. P., 1967, The behavior of sulphur and chlorine in
fumarolic gases before intensilication of activity of volcanoes, Coll. Volcanicity
and Geochemistry o/ its products. Publ. Nauka, p. 72-81.
MINAKAMI, T., 1942, On volcanic activities and tilting of the earth's surface. Bull.
Earthq, Res. Inst., 20, Ia. 431-504.
• 1950, Recent activities of Volcano Usu. Part 6: Precise levelling around Mr.
Usu in 1949. Bull. Earlhq. Res. Inst., 28, p. 153-160.
, 1960, F , ndamel~tal research [or predicting volcanic erttptior~s (Part 1) earthqtt~fkes and ertlstal deformations origirtating from volcanic activities. Bull. E~trthq.
Rcs. Inst., 38, p. 479-544.
MoxHar, l. R. M., I971, Thermal sttrveillattce of volcanoes in The Surveillance and Ptediction.7 o[ Volcanic Activity. UNESCO, E a r t h Science Mont~graph 8, Paris, p. 103-123.
N'..'I'Lf~NAI. AISI~0NAU'I'ICS~nd Sl, xcr~ AI)MINISTR,YI'I~N, 1972, Earth resottrces tcclinology S:llellite data collectiot~ plalform field instalhttion, operatiol~ and mail~tenance manual. Report 72 SD 4208, 256 pp.
Not;u:cm, K. and K ~ i t v a , H., Iq53, Predictio~ of volcanic er,.tplions by mea.stoittg tile
chelnical conlposilio~t trod a m o t m t of gases. Bull. Volczmol., 26, p. 367-378.
Oxlt~l::L F., 1907. Tiltitlg of the ground during a storm. Bull. Imp. Earthq. Inx,~st. Com.,
1, p. 167-171.
- - ,
1914-1922, The Sakttra-jima erttptions and carth:/trakes. Bull. Imp. Earthq.
Invest. Corn., 8, No. !-6, 525 pp.
OM()'fE. S., 1942, Crtt~tal deformation in Miyake-sima Island that accompanied the
volcallic activities in 1940. Bull. Earthq. Res. Inst., 20, p. 127-140.
Pat;E, R., 1968, A[tershocks and microa/tershocks of the Great Alaska EtJ_rl!~.qttake o/
1964. Bull. Scis. Soc. Am., 58, iP. 1131-1168.
PralinE1', F. A., 1937, An experi:rtet~tal ~, .seismeter,,. Am. J. Sci., 34, p. 469474.
R<~SE, W. I., Jr., 1974, N~tc;e ArdetTte ]rum Santiagtdto Volcano, April 1973. Bull. Vulcanol., 37-3, p. 365-371.
SI:tlMOZL:rU, D., 1971, A seismological approach to the prediction of volcanic eruptions,
in The Surveillance and Prediction of Voicanic Activity. UNESCO, E a r t h Science
Monograph 8, Paris, p. 19-45.
- - ,
M1YAZAKt, T., Gvot~a, N., a n d M,vrannLOMUAL, J., 1969, Volcanological .¢urvey
of Indonesian volcanoes - Pt. 2, Seismic observation at Merapi Volcano. Bull.
E a r t h q . Res. Inst., 47~ p. 969-990.
SMITh, B. E., HAMH.TON, R. M., a n d JACKSOr~, W. H., 1971, Seismicity of the central
Nevada test area. U. S. Geol. Survey, 27 pp.
29
--
344
ST0mEa, R. E. and ROSE, W. I., Jr., 1970, Geochemistry of Central America volcanic gas
condensates. Bull. Geol. Soc. Amer., 81, p. 2891-2912.
StJzuKI, Z., 1953, A statistical stvdy on the occurrence o/ small earthquakes, Chapter 1.
Sci. Rept. Tohoku Univ., Set. 5, Geophys., 5, p. 177-182.
$AZII~FF, H., 1971, A dynamic approach to the problem of /oreca~ting volcanic paroxysms. In The Sura,eillance and Prediction of Volcanic Activity. UNESCO, Earth
Science MOnograph 8, Paris, p. 127-130.
TON.~,~I, F., 1971, Concepts and techniques for the Geochemical forecasting of volcanic
eruptions. In The Surveillance and Prediction o/ Volcanic Activity. UNESCO,
Earth Monograph 8, Paris, p. 145-166.
Ur~GE~, J. D. and DECkeR, R. W., 1970, The microearthquake activity of Mt. Rainier,
Washington. Bull. Seis. Soc. Am., 60, p. 2023-2035.
VAN BEMMELEN, R. W., 1949, The Geology o/ Indonesia and Adjacent Archipelagos.
Netherlands Govt. Printing Office, The Hague, 732 pp.
WARn, P. L., ENUO, E. T., H~RLOW, D. H., ALLEN, R., MAa0UEZ, D., and EATON, J. P.,
1974, Development and evaluation of a prototype global volcano surveillance system
utilizing the ERTS-I Satellite Data Collection System. U. S. Geol. Sur~ey Ol~en
file report, 154 pp.
WESSON, R. Y., MEAcrltm, K. L., and LESXE~, F. W., 1973, Catalog of earthquakes along
the San Andreas fault system in central California - July - September 1972. U. S.
Geol. Survey open file report, 49 pp.
YOKOWMA, I., 1964, Seismometrical observation of the 1962 eruption of Volcano Tokati,
Hokkaido, Japan. Bull. Volcanol., 27, p. 217-223.
Manuscript received Aug. 1974
30