V33A-01 INVITED
Tungurahua Volcano's 1999-2007 Eruptive Process, Monitoring Results and Risk Mitigation
The eruptive process at Tungurahua volcano (5023m), in central Ecuador, began in August 1999 with SO2-rich steam emissions. In the previous 8 months, the IG's telemetered net had registered some 1300 volcano-tectonic (VT) seismic events, interpreted to be fracturing by magma ascent. VT's practically ceased in early Sept. 1999, followed by a transition to hybrid and long period (LP) seismic events and continual multispectral tremor and emissions. In October, Strombolian fountaining characterized the activity. This onset of magmatic activity prompted Ecuador's president to evacuate inhabitants from mid October to late December. November saw a shift to vulcanian explosions, which showered the cone with ballistics and fine ash over the region, but no large eruptions or pyroclastic flows transpired. During these early months SO2 outputs rose to 8-10 KT/day, while magma discharge remained low—probably less than 5 x 106 m3. Low level eruptive activity (VEI =1) continued from 1999 to mid 2006, punctuated by periods in Aug. 2001, Sept. 2002, Oct. 2003 and in mid 2006 of high energy release, persistent Strombolian fountaining, large explosions, and light regional ash fallout. Quiet periods were also registered, notably from Feb. to Dec. 2005. From Jan. to Mar. 2006 another rash of VT´s was registered followed by a reactivation in May to July, when the number of explosions/day increased (N= 900), SO2 levels rose, deformation accelerated and there was a temporary evolution of LP seismicity with hypocenters ascending from 7 to 3 km beneath the summit. Due to these changes, on May 12, the IG issued a report to authorities stating that activity at Tungurahua could lead to several scenarios—the generation of ash falls and pyroclastic flow were main concerns. Subsequently, the 14 July (VEI= 2) and the 16-17 August (VEI= 4) eruptions were preceded by heightened activity in all monitored parameters roughly 10 days before each eruption, followed by several days of calm, then by 3 to 12 hours, respectively, of dramatic ramping up before pyroclastic flows were produced. The IG gave early warnings to local and regional authorities, who evacuated vulnerable populations and closed highways. On the SW foot of the cone incandescent flows claimed the lives of six people on 17 August, when they did not heed official orders to evacuate. Since 1999, the IG has maintained a staffed observatory near the volcano, operating 24 hr/day, has emitted daily, weekly and special reports, and has been in continual contact with local volcano observers, authorities and the populace. The observatory also monitors rainfall and lahars on the cone and emits early warnings of secondary lahars en route to critical areas. The lessons learned at Tungurahua include: 1. Magma injections identified by VT swarms advance slowly to the surface, often taking months before superficial manifestations are apparent. 2. The quantity of precursory VT and LP events has been decreasing over time and therefore is an uncertain "clue" in determining the occurrence of magmatic intrusions. 3. From 1999-2006, the small magma injections benignly degassed before arriving to the surface, not having sufficient volume or gas content to produce gravitational flows. 4. Only during the two eruptive episodes of mid 2006 were pyroclastic flows generated, probably because of greater magma supply and its rapid ascent. 5. Drawn-out andesitic eruptions, although carefully monitored over long periods, may lead to waning concerns in the population and boredom in observatory staff. 6. Maintaining frequent and direct contact with the local population is important for fostering credibility with them.
V33A-02
A Statistical Method for Volcanic Hazard Assessment: Applications to Colima, Popocatepetl and Citlaltepetl Volcanoes, Mexico.
The volcanic-eruption time series are sequences describing processes of great complexity representing one of the main tools for the assessment of the volcanic hazard. The analysis of such series is thus a critical step in the precise assessment of the volcanic risk. The study of low-magnitude eruption sequences, containing larger data populations can usually be done using conventional methods and statistics, namely the Binomial or Poisson distributions. However, time-dependent processes, or sequences including rare or extreme events involving very few data, require special and specific methods of analysis, such as the non-homogeneous Poisson process analysis or the extreme-value theory. A general methodology for analyzing these types of processes is proposed in this work with the purpose of calculating more precise values of the volcanic eruption hazard. This is done in four steps: First, an exploratory analysis of the repose-periods and eruptive magnitudes series is done complementing the historical eruptive time series with geological eruption data and thus expanding the data population. Secondly, a Weibull analysis is performed on the repose-time between successive eruptions distribution. Thirdly, the eruption occurrence data are analyzed using a non-homogeneous Poisson process with a generalized Pareto distribution as its intensity function. Finally, these results are compared with fittings obtained from conventional Poisson and Binomial distributions. The hazard or eruption probabilities of three active polygenetic Mexican volcanoes: Colima, Popocatepetl and Citlaltepetl are then calculated with this method and compared with the results obtained with other methods.
V33A-03
Dense Pyroclastic Flows of the 16 -17 August 2006 Eruption of Tungurahua Volcano, Ecuador
The 16-17 August 2006 eruption of Tungurahua volcano in central Ecuador was preceded by 7 years of threatening activity and finally a VEI=2 eruption on 14-15 July 2006. The larger August eruption witnessed tens of pyroclastic flows that descended 17 different channels up to 8.5 km to the volcano's base on the NW, N, W, and SW sides. Tungurahua (5023m) is a steep-sided, low SiO2 andesitic volcano with 2600 to 3200m of relief. The initial, small nuee ardentes began around 1700hr (local time), the larger flows occurred between 2147hr and 0100hr (17 Aug.), and a total of 31 events were indicated by seismic signals. The deposits of three distinct flow cycles are recognized at the NW base of the cone. On the Los Pajaros depositional fan, deposits of flows 1 and 2 are widespread laterally (<600m) and have low-aspect morphologies with low snouts and without levees. Their outer surfaces are covered with accessory > juvenile clasts that mainly range from 15 to 25cm in diameter, conversely their interiors are comprised of 40-42% clasts of 1-25cm size and a matrix (58-60%) of sand-size grains. The earlier flow 1 was accompanied by an ash cloud surge that leveled, but did not scorch, all trees, brush, even metal antenna posts, leaving a 1-10cm thick sandy ash layer upon flow 1's deposit. On the fan as well as in gullies on the upper flanks, flow 3 deposits form long narrow lobes with 1-2m high frontal snouts that are followed by empty flow channels, 5-15m wide, bounded by parallel levees 1-1.5m high. Within these channels subsequent flow lobes are found as remnant pulses. Unlike flows 1 and 2, flow 3 lobes are covered with 0.5-3m cauliflower-shaped, slightly vesiculated bombs that are rarely abraded; the deposit's interior has a 45% sandy matrix. During the climatic eruptive phase continuous lava fountaining, 500-700m high, and crater spilling likely fed a continual stream of fragmented lava onto the cone's upper steep flanks, from which dense pyroclastic mass flows were initiated by gravity. Flows 1 and 2 were more fluidized (due to entrained air and fines), faster, and had wider lateral extents. Flow 3 was poorly fluidized, highly channelized, and behaved more like an inertial granular flow that formed as a continuous avalanche stream that separated into consecutive pulses along the runout channel.
V33A-04
Products and description from 3 phreatic eruptions occurred on March 2006. Poás Volcano, Costa Rica.
After 1994 Poàs volcano resumed activity with a series of phreatic eruptions on march 2006. This work focuses
on three main documented phreatic eruptions their characteristics and ejected products.
On march 24, a strong eruption was emitted from the south part of the hot crater lake affecting the dome area.
This phreatic eruption occurred around noon (18 hrs GMT) issuing an inclined column of water, sediments and
pre-existent blocks towards the south of Laguna Caliente. The eruption was observed by several visitors from the
crater lookout. They reported that a cypresoidal ejection issued from the lake surpassed the dome splashing acid
water, sediments and blocks towards the dry area, south of the lake.
Yet on Friday evening, a spray of acid water and fine sediments from the lake was produced. Next day staff from
OVSICORI-UNA visited the east side of the volcano to confirm deposition of material charged with lake water,
blocks and sediments from the east side of the lake. Impact craters were documented in one area ranging from
300 to 700m beyond the source point. The area visited showed a gray appearance although the rainy conditions
accumulated small ponds of such sediments in crevices and lower drainages. The maximum reach of a single
plume of fine sediments was approximately 5 km to the SW of the main crater, such material was tracked and
collected along the vegetated massif. Samples collected were composed of preexistent solid material from the
bottom of the lake heavily altered by the action of acid and hot conditions during the last 12 years. Some of the
more friable samples consist of compacted sediments that have aggregated miscellaneous granulometries.
Varied material from these two eruptions fell inside the main crater.
Lastly on Sunday 26th evening, another jet-like water/sediment stream was ejected, this time towards the W.
Scattered blocks from this eruption fell outside the west rim of the hot lake provoking several impact craters. This
was the only eruption that impacted with a thick sheet of sediments and blocks the outer parts of the crater basin.
Right after these events fieldwork was directed toward the documentation of deposits, ejected material and
impact in the surroundings.
This abstract will be accompanied by a poster depicting a number of hi-res digital photos taken during the
documentation process. Detailed description of deposits and materials will be given.
http:www.ovsicori.una.ac.cr
V33A-05
Emplacement of the Pyroclastic Flows during the 2004-2005 Activity in the Volcan de Colima, Mexico.
During its 2004-2005 activity, Volcan de Colima, Mexico produced several pyroclastic flows (PF´s) formed by first an unstable lava dome and then explosive activity. The initial large pyroclastic flow occurred on 06 October 2004, when the large and fast growing andesitic lava dome collapsed towards the southwest flank of the volcano in Barranca La Lumbre.During 2005, the largest episode of pyroclastic flows since the last plinian explosion in 1913, was a product of column collapse of various vulcanian explosions. Between 12 February 2005 and during April at least 7 pyroclastic flow events emplaced meanly in the upper section of the Barranca Montegrande, on the south side of the volcano. After these events the explosive activity increase significantly and on the 15 May 2005 a large vulcanian explosion produced a PF that travelled further than 5 km from the summit, modifying the original topography of the ravine, but the following event on the 30 May was the largest seismically until that moment. On 5th June, the largest explosion to happen in at least the past 20 years, generated flows that travelled more than 5 km and almost filled the upper section of the Barranca Montegrande and San Antonio. On 9th June a change in the principal direction of emplacement occurred when new pyroclastic flow deposition occurred in the Barranca La Arena, located in the southeast section of the volcano The pyroclastic flow front reached 5.40 km. Around one month later (5th July) a smaller deposit emplaced again meanly in La Arena and as a reduced proportion in Montegrande. The next months continued with a decreasing activity and generation of minor PF´s during September and probably December of 2005. During February to July 2005 a total pyroclastic flow volume of 6x106 m3 was generated.The understanding of the behaviour of a pyroclastic flow with its emplacement characteristics and sedimentological pattern and beyond that, the textural, petrography and geochemical variation could give us the opportunity to understand the tendency of the activity in the Volcan de Colima during the next years.
V33A-06
The November 2002 Pyroclastic Flows at El Reventador, Ecuador: Computer Simulations Using the TITAN Thin-layer Code
On November 3, 2002, El Reventador volcano, located on the eastern flank of the Ecuadorian Andes, produced a
sudden, violent eruption culminating in a 17km high column containing mostly steam and ash. Explosions in the
initial phase created a summit crater while generating four lithic-rich andesitic pyroclastic flows. The longest of
these flows traveled ESE out of the breached caldera, obliquely overriding the 200-400m southern caldera wall,
reaching the Quijos River 8km distant. This flow crossed the major oil pipelines of Ecuador, displacing a
pressurized crude oil pipeline more than 100m. The flows contained mostly lithic fragments with only minor
juvenile pumice. The accompanying ash-cloud surge deposited a thin layer on top of the PF deposit, indicating an
abundance of gas within the flow. The eruption came with practically no warning and yet had a large socio-
economic impact for Ecuador. While the flows themselves resulted in no loss of life, the lack of significant
precursor activity underscores the necessity for detailed pre-eruption knowledge of the potential hazards and risk
zones around a particular volcano so as to be prepared in the event of such "surprise" eruptions. In conjunction
with field mapping, computer models of volcanogenic flows can be used not only to identify risk zones but to
understand the evolution of these flows.
A new set of computer simulations using the TITAN (www.gmfg.buffalo.edu) thin-layer code allows a more
complete exploration of important flow properties associated with this type of eruption. Realizations of this code
simulate the path, extent, flow thickness, velocity, and momentum of the flows given the set of initial conditions
(volume, starting location, flux hydrograph, internal friction, and basal friction). The TITAN code was used to
simulate the four lithic-rich pyroclastic flows generated at the beginning of the 2002 eruption. Using field
estimated volumes and starting positions of the PFs, simulations of the two largest flows, confined in major
channels along the northern and southern walls of the old caldera, provided qualitatively good fits to mapped
deposits. Specifically, these models produced features comparable to the real flows including overtopping of the
southern caldera wall, diversion by topographic obstacles, and channeling. The two smaller flows, while
producing narrow, linear deposits, spread much farther laterally in the simulations. This phenomenon may reflect
some unmodeled flow dynamic such as yield strength, which might only become substantial with small
volumes. Possibly, as the DEM was constructed from a topographic map, the digitization or smoothing may have
erased small channels which governed the real flows but could not be represented in simulation.
http:www.gmfg.buffalo.edu
V33A-07
October 2005 Debris Flows at Panabaj, Guatemala:Hazard Assessment
In October, 2005, tropical storm Stan caused heavy precipitation throughout much of Guatemala. In the community of Panabaj, Santiago Atitlán, a landslide of pyroclastic material originating high on the slopes of Tolimán volcano buried much of the community, leaving approximately 400 people dead. Current estimates by the Coordinadora Nacional para la Reducción de Desastres (CONRED) suggest that at least 2,600 people from the community of Panabaj, Santiago Atitlán have been displaced by the debris flows. Because the temporary housing for people displaced by the debris flows is located in an area that is geologically and morphologically similar to the area inundated by flows in October, 2005, this area may be potentially inundated by debris flows as well. In addition to the thousands of people living in temporary shelters, many hundreds of people are currently reoccupying land adjacent to or on the October, 2005 debris flows. Thus a large fraction of the surviving Panabaj community appears to remain at risk from future debris flows. We used differential GPS (Global Positioning System) to outline the boundaries of the debris flows, to estimate variation in flow thicknesses, and to determine their volumes. Mass movement on Tolimán volcano resulted in the generation of a moderate size debris flow (360,000 m3 of sediment plus water) that descended the volcano rapidly, bifurcated into two stream valleys high on the flanks of the volcano, and continued to descend both channels until these flows reached the alluvial fan near the shores of Lago de Atitlán. After bifurcating into two flows high on the flanks of the volcano, about 65% of the flow (by volume) descended the western channel, forming the Western flow. Approximately one kilometer above the alluvial fan, this channel descends steep topography, with a slope of 11.5°. This average slope gradually decreases down the channel, reaching only 5.3° just above the alluvial fan. In contrast, average slopes on the Eastern channel are up to 16.7°. Also, this channel thalweg steepens dramatically to 12.8° just above the alluvial fan. Flow velocities in channelized sections were estimated by superelevation at bends at two locations for each of the two flow branches. In measured cross sectional areas between 144 and 160 m2 the calculated velocities ranged from 8.3-10.6 ms-1 yielding fluxes between 1280 and 1680 m3s- 1. The fluxes for the two flows are surprisingly similar. The planimetric area inundated by the Western flow is approximately 180,000 m2 and the area inundated by the Eastern debris flow is 77,000 m2. On reaching the gently-sloping (2.8°) depositional fan where the village of Panabaj is located, the flows thinned to 0.5-3.0 m and spread laterally as a broad sheet flow bounded by distinct flow fronts of 0.30-0.6 m height. Although thin, the flows had sufficient power to sweep away most of the concrete block houses in their paths. Based on observations of high water marks preserved on buildings, up to 40% of the flow by volume consisted of water and fine grained sediments that have been dewatered from the deposit during and since deposition.
V33A-08
The activity of the Colima volcano in the 2005-2007
The Colima Volcano, is located in the West of the Volcanic Mexican Belt, has shown a new cycle of explosive activity beginning in 30 of May of 1999, having its maximum in March and April the 2005. In this year the activity increase gradually, having the maximum event on May 23, when a new dome was created, hours later its dome was destroyed by a strong explosion, which formed a column 5.6 km high and piroclastic flows that reached a distance of 4.2 km on the ravines of the South sector. On May 30 the most intense explosion from 1999 occurred, when the plume reached heights over 4.4 km above the crater, and piroclastic flows were created. In the month of June it generated four explosive events of characteristics similar to those of May. These constant explosions caused constant morphological changes on the top, being the most significant the collapse of the North and South walls of the crater in the first week of June, and the creation of a new crater in July. In the 2006 the explosive activity most significantly was in April, May and July, when the plume reached a height of over 1500 meters above the crater, occasionally forming small pyroclastic flows. The explosive events continue to date, but they have diminished in size and intensity. The seismic analysis of these explosions appears in another abstract. This activity was similar to the one shown in 1902-1903 and reported by Severo Diaz (1906), but without reaching the maximum levels of activity reported for 1903, where it had levels of three to five maximum explosive events per day. The explosions deposited great amount of nonconsolidated materials, like ash, lithics and rocks on the flanks of the volcano, which with the present rainy season have generated lahares. These have flowed in small flows on the ravines of La Lumbre, Montegrande and La Arena. None of them have caused damages in the season of rains of the 2006.
V33A-09
POPOCATEPETL VOLCANO'S ACTIVITY FROM JULY 2005 TO FEBRUARY 2007
An increase in the frequency and magnitude of micro-seismicity started in 2005 is correlated with the lava dome that has been irregularly growing since July 2005. The lava dome has covered the floor of the internal crater and began a piston-like growth on the top of the previous dome, probably in August, 2006. On 4 and 5 August, 1 to 3 November 2006, and 2,3,4 and 22 february episodes of large-amplitude harmonic tremor are believed to reflect an increased rate of dome growth. On 27, 28 and 29 October 2006, eigth small explosions ejecterd incandescent debris on the slopes surrounding the crater rim. During November and December 2006, more episodes of low amplitude tremors were recorded. From August to December 2006 77 volcano-tectonic microearthquakes were detected, with magnitudes ranging between 2.0 and 3.0. From these, 66 were located below the crater at depths ranging between 3 and 7 km.
V33A-10
A Qualitative Estimate of the Volcanic Hazard
Volcanic hazard is defined in terms of the probabilities of occurrence of eruptions and their potentially destructive manifestations. These probabilities may be estimated analyzing the sequence of past eruptions in a given volcano, characterizing the eruptions by size and assuming that the impact and effects of an eruption are proportional to both, the energy (magnitude) and the rate of energy release (intensity). A quantity that characterizes eruptions based on those parameters is the Volcanic Explosivity Index (VEI, Newhall and Self, 1982). Here, we introduce an extension of the definition of volcanic hazard in terms of the expected annual release of energy by eruptions in each VEI category. This concept is based on the averaging property of a large group of volcanoes to release in a given time interval the same amount of energy in each VEI category. This means that a set of volcanoes tend to produce a larger number of smaller eruptions and a smaller amount of larger eruptions in such a way that in the given time interval, the numerous smaller eruptions release about the same energy than the few larger eruptions (De la Cruz-Reyna, 1991). The annual rate at which energy is released by eruptions in different VEI categories is described by log(EmKm) = b M + a, where Em is the energy released by eruptions in the VEI magnitude class M, and Km is the rate at which such eruptions occur. The parameters a and b depend on the eruptive history of individual volcanoes. The slope b determines the preferred mode of the volcano to release energy: through smaller (negative slope) or through larger (positive slope) eruptions, and the parameter a determines the energy potential of the volcano. We applied this method to analyze three active polygenetic Mexican volcanoes: Colima, Citlaltepetl and Popocatepetl. The former has the largest energy potential, although the slope b has a slightly negative trend suggesting that while large and very large eruptions may occur, the volcano shows a trend to release its energy by frequent smaller eruptions. Popocatepetl and Citlatepetl show smaller energy potentials than Colima and almost flat slopes, indicating a trend to release energy by eruptions of any size, but at lower rates. De la Cruz-Reyna S (1991) Bull Volc 54, 57; Newhall CG, Self S (1982) J Geoph Res 87, 1231