An Exploration of the Use of Satellite Remote Sensing to Predict

An Exploration of the Use of Satellite Remote Sensing to Predict

An Exploration of the Use of Satellite Remote Sensing
to Predict Extrusive Events on Colima Volcano
Bruce Wahle
ASEN 5235 April 24, 2003
There are more than 1300 active volcanoes in the world (Stoiber 1995), and the majority of
world’s most dangerous volcanoes are not monitored at all (Chester 1993, Tilling 1993). As the
Earth’s population increases more people are living close to active volcanoes. Some of these
volcanoes are isolated and remote sensing by satellite could be a cost efficient and rapid way to
monitor and forecast their eruptive activity. Increasing concentrations of volcanic gases and/or
their ratios (Menyailov 1975, Shimozuru 1983, Stoiber et al. 1983), and lava dome building have
been suggested as precursors of eruptions (Luhr 2002).
The purpose of this paper is: to explorer the ability of satellites to remotely identify pre-eruption
sulfur dioxide (SO2) flux changes;
to explorer what are the satellite availability, cost and quality of the
data;
to look specifically at Colima Volcano, Mexico;
with the intended goal of identifying a remote satellite system, that
can measure volcanic gas changes of sulfur dioxide, rapidly and in
a cost effective manner, in order to predict eruptive events in
andesitic arc stratovolcanoes
For this study, I wanted to look at Colima Volcano in Mexico. The area is at the juncture of
three tectonic plates (Cocos, Rivera, North American, Figure 1); a large-scale earthquake
occurred nearby on January 22, 2003. In the state of Jalisco, Mexico are two andesitic
stratovolcanoes 4.8 km (3miles) apart: Colima Volcano (also known as Volcan de Colima and
El Volcan de Fuego de Colima) and Nevado de Colima Volcano (Figure 2). Volcan Nevado de
Colima is 4381 m (14,365 feet) tall, covered by coniferous forest, and is considered dormant.
Colima Volcano at 3962 m (12,989) feet is considered the most dangerous volcano in Mexico
(Figure 3, Luhr 1981). Colima Volcano is continually active and erupts frequently; it is one of
North America’s most active volcanoes (Moreno Pena and Salazar Silva 2002). Colima Volcano
is a Quaternary volcano that is in an area that is both tectonically and seismically active, and the
present cone has been growing for approximately 4000 years (Taran et al. 2002).
In general, monitoring of SO2 fluxes and potential eruption of volcanoes is important for a
number of reasons. It is important from a public health/emergency-planning standpoint. The
immediate danger from an eruption may endanger humans and livestock. In political failures to
evacuate threatened cities there have been massive death counts, such as in St. Pierre, Martinique
(~29,000) and Armero, Colombia (~25,000). On the other hand, predicting eruptions and
evacuating populations have success stories, such as around Mt. Pinatubo, Philippines and
Mount St. Helens, USA. Long-term effects from ash fall may result in famine and death; the
Laki fissure eruption in Iceland in 1783 caused the death of much of the island’s livestock and
approximate 20% of the island’s human population from starvation (Chester 1993). The
monitoring of ash and temperature changes is important because they are factors in atmospheric
forcing. Increases in SO2 flux will, over time, convert to sulfuric acid and sulfates. Sulfuric acid
is one of the causes of acid rain, and sulfates contribute to atmospheric cooling and may affect
the formation of cirrus clouds. Ozone destruction is another possible result from increased SO2
atmospheric increases. The danger of aircraft flying into volcanic plumes is great, both in
aircraft mechanical damage and plane crashes (Trebella 2002). More than 300,000 people are
living near by Colima Volcano.
Methods
There are two basic types of volcanoes: arc and nonarc volcanoes. Arc volcanoes are located in
areas of subduction zones. Eighty-one percent (429 of 530) of eruptions are in subduction-zone
volcanoes (Simkin and Siebert 1984). Nonarc volcanoes are found over hotspots and rifts.
Colima Volcano is an arc stratovolcano with andesitic (57-63% silica) lava, which makes the
eruption more explosive, rather than effusive (Chester 1993). The general trend of Colima
Volcano and other andesitic volcanoes like it is to undergo a cyclic process of a large eruption,
filling back up again with lava, becoming plugged up again, sealing shut, and then another large
eruption to clear the volcanic throat (Figure 4, Luhr 2002).
As fresh magma begins to rise in a volcanic system, volcanic gases should become more
prevalent in the volcanic plume. The major volcanic gases are: water vapor, carbon dioxide,
SO2, and hydrochloric acid mixed with other trace gases (Chester 1993, Williams and Montaigne
2001). With remote satellite measurements, water, carbon dioxide and ozone are confounders in
the atmosphere and must be corrected for (Krueger 1983, Stoiber et al. 1986). For this reason, I
2
3
decided to concentrate on SO2 measurements. As magma moves to the surface, more SO2 should
be detectable in the volcanic plume, up to the point where the lava dome becomes sealed, just
before eruption. Unfortunately, not all volcanoes have high sulfur emissions (Schnetzler et al.
1997).
Satellites and aircraft have done remote sensing volcanic monitoring by observing changes in
thermal, volcanic ash and gaseous emissions. Airborne monitoring of SO2 and sulfates has been
carried out with the Thermal Infrared Multispectral Scanner (TIMS). Some of the satellite
platforms or sensors, which have been used to monitor volcanic thermal, volcanic ash and/or
SO2, are:
Advanced Very High Resolution Radiometer (AVHRR)
Geostationary Operational Environmental Satellite (GOES)
Moderate Resolution Imaging Spectroradiometer (MODIS)
Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER)
Total Ozone Mapping Spectrometer (TOMS)
4
The AVHRR and GOES satellite have large area coverages and while good to monitor the
volcanic eruption after the fact, they would not have the fine resolution to measure volcanoes
pre-eruption (pixels 1000-1100 m). MODIS data, while free, still is too broad (pixels
240-1920 m) for this task. ASTER data (pixels 15-90 m) would be interesting to use, but it costs
$55 per scene and its return time is too long (16 days). Based on spatial resolution (25-110 m),
cost and return time (1 day), TOMS was the only choice for this project. Data used in this study
were downloaded from the Earth Probe TOMS V. 7 Internet site for the Guadalajara, Mexico
station (site ID 654, elevation 1420 masl, generated March 9, 2003), approximately 128 km to
the northwest of Colima Volcano. The TOMS sulfur dioxide index (SOI) from July 17, 1996 to
March 8, 2003 was used for this study. Volcanic activities of Colima Volcano were identified
from the numerous sources (Anon. 1999, 2001, 2002, 2003a, 2003b; Galindo and Dominguez
2002; Penick 2003).
TOMS data was plotted in a spreadsheet software package. In addition to plotting the raw SOI
versus time, I tried to normalize the data by calculating the percent difference from the previous
measurement (usually 24 hours).
SOI % Change = ((day 2 – day 1)/day 1) x 100
Results
All TOMS data downloaded were plotted in Figure 5. It is immediately apparent that there is a
problem after August 1, 2001 (day 213). A sensor began to degrade and corrections were made
to the instrument. Although NASA says this has been corrected (NASA 2003), I question this.
5
The upward swing just before the sensor correction is probably due to erroneous readings, rather
than increasing SOI, in my opinion. Long-term trend analysis cannot be done until the database
is corrected. Data after approximately January 2001 must be corrected before direct use of the
SOI can be done because of the downward shift.
Figure 6 shows the percent SOI change versus time for the years 1997-1998. This was the
beginning of another active eruption cycle on Colima Volcano. Beginning in mid-summer 1997
geochemistry values began to increase in the fumaroles. In November 1997 there was an
eruption and seismic swarms on the cone. A large eruption was predicted and occurred in
November 1998 (Zobin et al. 2002). Before the eruptions, the variation in the percent SOI
change decreased.
The events that began with large eruption on November 20, 1998 continued into 1999. Figure 7
shows the data plotted from November 1, 1998 to December 31, 1999. There were a number of
eruptions during this time period, some minor and other major eruptions (November 20, 1998,
February 10, 1999, May 10, 1999, July 17, 1999; Zobin et al. 2002). Once again SO2 fluxes
usually decrease before the eruption, but not always. One has to realize the eruption data has
come from various sources and has not been collected first hand by the investigator. The general
trend of SO2 activity increasing previous to an eruptive event, then a decrease just prior to
eruption follows the prediction of events.
6
An idea that volcanic eruptions have a seasonality has been looked at in several studies. With
larger databases, this has proven to be statistically valid for several volcanoes (Mason 2003).
With the volcanic activity I was able to locate for Colima Volcano from November 21, 1997 to
March 8, 2003, 51 volcanic events were identified (Figures 8 and 9, Table 1). Figure 8 is a plot
of the count of events versus day of year over more than 5 years. Figure 9 shows the same data,
but in a pie chart of counts per month. Table 1 represents the data from Figure 9, but in a more
readable table format. The apparent seasonality of eruptions at Colima Volcano looks intriguing,
but this represents a very small and incomplete database. Studies with much larger data sets and
more rigorous statistics have been needed to show this on (some) volcanoes (see Mason 2003).
Summary
The purpose of this study was to investigate whether a remote sensing satellite could measure
SO2 concentrations in volcanic emissions prior to an eruption. Some volcanologists believe
either SO2 concentrations or a ratio of volcanic gases can forecast a volcanic eruption. The
TOMS instrument measures SO2 approximately daily for a number of sites worldwide. TOMS
measures SO2 at UV wavelengths during the daytime only. The first TOMS sensor detected 55
of the 350 known eruptions (16%, post eruption) between 1979 to 1992, and improved TOMS
instruments were launched in lower orbits have detected SO2 gas from smaller eruptions and
passive volcanic degassing (USGS 2001). TOMS is best used to detect and track SO2 emitted
from the explosive phase of eruptions; it observes SO2 primarily in the stratosphere. The
detection limit for TOMS was about 5 kilotons SO2 . Typical volcanic SO2 clouds measured by
TOMS are 20-several hundred Dobson Units (DU), with an error estimation of +/- 30%
7
(Bluth et al. 1992). TOMS is ideal for large eruptions, but it cannot capture all eruptions,
especially numerous low-level or non-explosive eruptions (Bluth et al. 1993, 1997). Total SO2
values greater than 100 DU in the TOMS database are observed only in the first day of eruptions
except for the largest eruptions (e.g., Mt. Pinatubo, El Chinchon). The maximum of SO2
amounts extends to 600 DU (only rarely observed in very fresh plumes). Most plumes are less
than 100 DU SO2 (Krueger et al. 1995). Other limitations exists in accessing high latitude
volcanic activity with TOMS in the winter because of decreased daylight hours, and longer path
lengths due to low sun angles reducing precision of the SO2 algorithm (Bluth et al. 1997). At
near-nadir scan positions, a measurement change of 1 DU is caused by a little more than 50 tons
SO2 (Krueger et al. 1995). The TOMS site at Guadalajara, Mexico is approximately 128 km
from Colima Volcano (2-6 pixels from nadir). This means that to capture a SO2 emission, at
least a reading of 9-10 DU must be observed. Two ground measuring spectrometers are utilized
in volcano monitoring, a correlation spectrometer (COSPEC) and a Brewer spectrometer, for
measuring total mass of gaseous eruptions. Attempts to correlate TOMS SO2 retrievals with
COSPEC have not met with success because they have vastly different sensitivities and they are
not operating at same times. The attempts with the Brewer spectrophotometer were in good
agreement when data were available on the same days (Krueger et al. 1995). Use of these
8
ground based instruments at the same time as the TOMS over flights would help to validate the
TOMS data.
Specifically looking at the andesitic Colima Volcano, COSPEC measurements from the air gave
320 +/- 50 metric tons/day while the volcano was quiescent. A comparison of SO2
concentrations in plumes with ground fumarole gas samples suggests that the plumes are diluted
by the atmosphere by factors of up to 105. A COSPEC clean troposphere measurement gives
values of 0.09-0.15 ppb SO2, while Colima Volcano had values of 5-380 ppb SO2. The lifetime
of these quiescent plumes is short, probably hours to days, compared to the longer residence
times of months to years of eruption plumes injected into the stratosphere (Casadevall et al.
1984). Quiescent volcanoes with noticeable vapor plumes emit on order of 102 tons SO2/day;
volcanoes with small plumes emit an order of magnitude less (Stoiber and Jepsen 1973). SO2
flux rates of 10’s to 1000’s of metric tons/day for a volcano producing very little or no solid
products (non-eruptive) are seen. An erupting volcano, producing ash or lava, can produce 10’s
to 1000’s of thousands of tons per day of SO2 (Stoiber et al. 1983). TOMS is able to measure
SO2 concentrations, but not H2S or sulfates (McKeen et al. 1984) and probably not S2, so the
TOMS values would represent a minimum of sulfur products being released. There is a variation
in gases depending on their release area on the volcano (e.g., main vent, associated lava flows,
pyroclastic flows; Stoiber 1995). The SOI accurate measure of SO2 for ash-free plumes is up to
100 DU, the SOI begins to lose some accuracy between 100-200 DU and becomes unreliable
>200 DU. The presence of ash causes the SOI to underestimate the SO2 mass (Seftor et al.
1997). Andesite-producting volcanoes often penetrate the tropopause (Cadle 1975).
Based on measurements from the quiescent Colima Volcanco and the sensitivity of the TOMS
instrument, a warning SOI threshold in the area of 8-10 DU could be proposed to indicate above
normal, quiescent SO2 emissions from the volcano. On October 30 and November 16, 1998
COSPEC readings of 400 tons/day SO2 presaged the plume of November 18, 1998, which
resulted a record SO2 flux measured at Colima Volcano of 1600 tons/day and the eruption of
November 20, 1998 (Zobin et al. 2002).
The problems that exist with study are that although the TOMS instrument may capture a
volcano in its scene, it may not be at the nadir location and this affects the values returned. The
value of the SOI information since the correction of the sensors in August 2001 is unknown.
Validation of TOMS with on ground COSPEC measurements at the same day and time is usually
not done. At the present time, only TOMS provides a sensor for measuring SO2 flux with low
cost (free) data, rapid return time (24 hours), rapid retrieval time from the internet of day, and is
sensitive enough to capture SO2 emission changes from a satellite (although this can be
improved). TOMS does show the potential to help forecast volcanic eruptions in andesitic arc
stratovolcanoes that may only be easily monitored by satellite remote sensors. No one method of
volcanic monitoring works by itself. This method would optimally be used in conjunction with
seismic monitoring, tiltmeters, fracture monitoring, fumarole monitoring (geochemical, thermal),
video monitoring, other remote satellite sensing, and ground-based spectrometer monitoring.
9
References Cited
Anonymous. 1999. Summary of recent volcanic activity. Smithsonian Institution’s global
volcanism network. Bull. Volcanol. 61:423-424.
Anonymous. 2001. Imagenes de RSEM/Actividad Sismica. Colima Mexico, Volcan, Queseria,
Naranjal, Tonila, Universidad de Colima.
Internet, http://www.ucol.mx/volcan/imagenes_RSEM.html.
Anonymous. 2002. Colima Volcano Eruption (western Mexico). Volcano live. Archived
volcano eruption news. Internet, http://www.volcanolive.com/news18.html.
Anonymous. 2003a. Colima, Mexico. University of North Dakota Volcano World Web site.
Internet, http://volcano.und.nodak.edu/vwdocs/current_volcs/mexico/colima.html.
Anonymous. 2003b. Eruption activity to date. Southwest volcano research center.
Internet, http://www.swvrc.org/eruption.htm.
Bluth, G. et al. 1992. Global tracking of the SO2 clouds from the June, 1991 Mount Pinatububo
eruptions. Geophys. Res. Lett. 19:151-154.
Bluth, G. J. S., S. D. Doiron, C. C. Schnetzler, A. J. Krueger, and L. S. Walter. 1993. New
Constraints on sulfur dioxide emissions from global volcanism. Nature 366:327-329.
Bluth, G. et al. (1997) Stratospheric loading of sulfur from explosive volcanic eruptions.
Journal of Geology, v. 105, 671-683.
Cadle, R. D. 1975. Volcanic emissions of halides and sulfur compounds to the troposphere and
stratosphere. J. Geophys. Res. 80:1650-1652.
Casadevall, T. J., W. I. Rose Jr., W. H. Fuller, W. H. Hunt, M. A. Hart, J. L. Moyers, D. C.
Woods, R. L. Chuan, and J. P. Friend. 1984. Sulfur dioxide and particles in quiescent volcanic
plumes from Poas, Arenal, and Colima volcanoes, Costa Rica and Mexico. J. Geophys. Res.
89:9633-9641.
Chester, David. 1993. Volcanoes and society. London:Edward Arnold. 351 p.
Galindo, Ignacio and Tonatiuh Dominguez. 2002. Near real-time satellite monitoring during the
1997-2000 activity of Volcan de Colima (Mexico) and its relationship with seismic monitoring.
J. Volcanol. Geotherm. Res. 117:91-104.
Krueger, A. J. 1983. Sighting of El Chichon sulfur dioxide clouds with the Nimbus-7 Total
Ozone Mapping Spectrometer. Science 220:1377-1378.
10
Krueger, A. J. L. S. Walter, P. K. Bhartia, C. C. Schnetzler, N. A. Krotkov, I. E. Sprod, and
G. J. S. Bluth. 1995. Volcanic sulfur dioxide measurements from the total ozone mapping
spectrometer instruments. J. Geophys. Res. 100:14057-14076.
Luhr, J. 1981. Colima: history and cyclicity of eruptions. Volcano News 7:1-3.
Luhr, J. 2002. Petrology and geochemistry of the 1991 and 1998-1999 lava flows from Volcan
de Colima, Mexico: implications for the end of the current eruptive cycle. J. Volcanol.
Geotherm. Res. 117:169-194.
Mason, Betsy. 2003. Season of fire. Discover 24(2):32-39.
McKeen, S. A., S. C. Liu and C. S. Kiang. 1984. On the chemistry of stratospheric SO2 from
volcanic eruptions. J. Geophys. Res. 89:4873-4881.
NASA. 2003. TOMS overpass release notes.
Internet, http://toms.gsfc.nasa.gov/eptoms/releasenotes.html.
Moreno Pena, F. and Salazar Silva, C. 2002. Preface. J. Volcanol. Geotherm. Res. 117:ix.
Penick, Tom. 2003. Seismic activity. Internet, http://www.tomzap.com/seismic.html.
Schnetzler et al. 1997. A proposed volcanic sulfur dioxide index (VSI)
J. Geophy. Res. 102:20087-20091.
Seftor, C. et al. 1997. Detection of volcanic ash clouds from Nimbus 7/total
ozone mapping spectrometer. Journal of Geophysical Research, v. 102: 16,749-16,759.
Shimozuru, Daisuke. 1983. Volcano surveillance and prediction of eruptions in Japan. pp. 173192 in Tazieff, H. and J. C. Sabroux (eds.). Forecasting volcanic events. Amsterdam:Elsevier.
635 p.
Simkin, T. and L. Siebert. 1984. Explosive eruptions in space and time: duration, intervals and
a comparison of the world’s active volcanic belts. pp. 110-121 in Geophysics Study Committee.
Explosive volcanism: inception, evolution and hazards. Washington, D. C.:Natl. Acad. Press.
176 p.
Stoiber, Richard E. 1995. Volcanic gases from subaerial volcanoes on Earth. Global Earth
physics. A handbook of physical constants. American Geophysical Union.
Stoiber, R. E. and A. Jepsen. 1973. Sulfur dioxide contributions to the atmosphere by
volcanoes. Sci. 182:577-578.
Stoiber, R., S. N. Williams and B. Huebert. 1986. Sulfur and halogen gases at Masaya caldera
complex, Nicaragua: total flux and variation with time. J. Geophys. Res. 91:12215-12231.
11
Stoiber, R., S. Williams and L. Malinconico. 1983. Use of the correlation spectrometer at
volcanoes. pp. 421-444 in Tazieff, H. and J. C. Sabroux (eds.). Forecasting volcanic events.
New York, NY:Elsevier. 635 p.
Taran, Yuri, Juan Carlos Gavilanes and Abel Cortes. 2002. Chemical and isotopic composition
of fumarolic gases and the SO2 flux from Volcan de Colima, Mexico, between the 1994 and
1998 eruptions. J. Volcanol. Geotherm. Res. 117:105-119.
Tilling, Robert I. 1993. Monitoring active volcanoes. USDI, USGS. USGPO, 1993-348-979.
17 p.
Trebella, Matthew. 2002. Comparison of TOMS and AVHRR volcanic ash detection algorithms
for the February 2000 eruption of Mt. Hekla, Iceland. Remote Sensing Project. Internet,
http://lasp.colorado.edu/~trebella/project.html.
USGS. 2001. Detecting sulfur dioxide gas in eruption clouds.
Internet, http://www.usgs.gov.
Volka, Vicki. 2001. Le volcan Colima vu de l’espace.
Internet http://www.vickivolka.uqam.ca/chronique/mar02/image3a3.shtml.
Wallace, Paul J. and Ian. S. E. Carmichael. 1999. Quaternary volcanism near the Valley of
Mexico: implications for subduction zone magmatism and the effects of crustal thickness
variations on primitive magma compositions. Contrib. Mineral Petrol. 135:291-314.
Williams, Stanley and Fen Montaigne. 2001. Surviving Galeras. Boston, MA:Houghton
Mifflin Co. 270 p.
General Reading, Volcanoes
Bruce, Victoria. 2001. No apparent danger. The true story of volcanic disaster at Galeras and
Nevado del Ruiz. New York, NY:HarperCollins Publ., Inc. 239 p.
Williams, Stanley and Fen Montaigne. 2001. Surviving Galeras. Boston, MA:Houghton
Mifflin Co. 270 p.
Winchester, Simon. 2003. Krakatoa: The day the world exploded August 27, 1883.
HarperCollins Publishers.
Real Time Video Cameras of Colima Volcano
www.ucol.mx/volcan/enviro_naranjal.html
12