25 major crises Infrastructure in 12 countries

• 25 major crises• Infrastructure in 12 countries

(FY-03-11): 54 infrastructure missions, 12 crisis responses, 15 countries

2011: VDAP’s 25th Year

• 25 major crises• Infrastructure in 12 countries

(FY-03-11): 54 infrastructure missions, 12 crisis responses, 15 countries

2011: VDAP’s 25th YearDistal VT’s

H2O expulsion

SO2 drop

EQ patterns (dVT-LF-hybrid-tremor)

H2S to SO2

CO2 pulse

Drumbeats

Geodetic trends

Magma type & textureEruption rates

RSAM trends

Well levelsCl & F

Distal VT’s

EQ patterns (dVT-LF-hybrid-tremor)

RSAM trends

The case for a process-oriented guide to forecasting

explosive eruptions (at stratovolcanoes)

Part 2. Frequently active volcanoes…

“Difficult (or easier?) to predict eruptions. Fortunately most are <<VEI 4”

John Pallister for the VDAP team (past and present):

Randy White, Wendy McCausland, Andy Lockhart, Jeff Marso, John Ewert, Chris Newhall, C. Dan Miller, Rick Hoblitt, John Power, Tom Murray, Dave Harlow, Marvin Couchman, Julie

Griswold, Gari Mayberry, Dave Schneider, Steve Schilling, Angie Diefenbach

Eruption possible (magma/fluids moving in crust)Magmatic eruption likely (magma & gas have reached shallow levels)

High VEI (indicated by large volumes of gas-rich viscous magma and rapid ascent)

Low VEI – (indicated by smaller volumes of gas-poor low-viscosity magma, and slow ascent

No eruption (yet) … intrusion stalled…end of crisis

Forecast Seismic Deforma-tion

Gas Observa-tions & history

Petrology

Tables of common indicators for likelihood and explosivity… based on VDAP experience

Sudden changes (+/-) from background RSAM level (background may include: proximal VT’s, LF’s, tornillos, tremor, explosion signals)… Distal VT’s possible, but with cumulative M<4).

Deformation localized near summit/vent areas. Little regional deformation unless unusually large batch of magma

CO2 increase = deep degassing & early warning. SO2 increase if hot pathway, otherwise may be scrubbed and replaced by H2S or sudden decrease in emissions

Steam emissions, phreatic eruptions, fractures opening. Water levels in nearby wells may change

Rapid, steep increase in RSAM from background level (e.g., <1 week in some cases), LF/explosion signal followed by tremor typical. VT’s may/may not occur. Slower ramp-up in other cases (previous eruptions provide guides. Hybrid & drum-beat quakes may be detected and indicate magma rising

Accelerating deformation of summit as eruption approaches

SO2 & Cl emissions increase and SO2/H2S increases as magma rises and pathway dries out. Emissions may drop shortly before eruption if pathway obstructed

More vigorous steam emissions & phreatic/ phreatomagmatic eruptions or sudden stoppage of observed emissions

Initial tephra mainly composed of lithic debris, glass shards may or may not be present. Juvenile component minor but typically difficult to identify. (Petrology not diagnostic)

Higher & more rapid increase in RSAM than past small eruptions, (including VT’s ), continued increase after initial vent clearing phreatic/ phreatomagmatic eruptions.. Escalating explosion signals and/or tremor precede biggest event(s)

Deformation of greater magnitude and rate that past small eruptions

Relatively large gas flux, especially following initial “vent clearing” events. Emissions may drop shortly before eruption if pathway obstructed. Early CO2 pulse may be 1st warning of recharge and high VEI

History includes larger multi-phase eruptions, and hydrous minerals. Initial phase of current activity explosive instead of extrusive. Unusually high extrusion rates (for initial phases)

Tephra or lava from initial eruptions may contain gas-rich indicators (high volatiles in glass inclusions, bubble-wall shards, volatile-rich indicator minerals, e.g., hb). Silicic magma composition. These are only permissive indicators as gas-rich magma may not yet be shallow

RSAM moderates & decreases after initial vent clearing events. Return to background levels, may take weeks

Deformation ends or decreases, e.g., after initial explosions or lava (dome) erupted

Gas emissions fall to background levels after initial vent clearing events and in parallel with seismic decrease

History of small eruptions (no big PF sheets or thick tephra deposits). Eruptions may be “slow:” e.g., peak dome collapse weeks to months after onset of episode

Early-erupted juvenile component degassed (microcrystalline, lacks volatile-rich indicator minerals or they have thick decompression rims = slow ascent). Magma is mafic (non-exclusive)

Seismicity dies out after initial emissions

Deformation ends, local deflation near vent

Gas emissions drop to background levels

Steam emissions or back to background

No single indicator is conclusive! Synthesize!

Don’t try to read this part, I will blow it up!


Gas Observations & history

Petrology

Magmatic eruption possible (magma/fluids moving in crust)

Background level exceeded. Distal VT's as tectonic faults are pressurized and magma rises. VT number & energy increasing as magma rises. VT energy release often approaches or exceeds M4. (Spurr only M3). Also many intrusions stall without eruptions!

Inflation detected = early warning; may be large. Broad (deep sourced) deformation

CO2 increase = deep degassing, may provide early warning

Hydrothermal system & fumaroles present but not necessarily elevated over background levels. Past history of explosive eruptions. (Montserrat no steaming). Water levels in wells may drop (e.g., due to dilatation)

Products of past eruptions give range of most likely magma types and explosivity

Magmatic eruption likely (magma & gas have reached shallow levels)

Transitional from distal to proximal VT's as magma shallows. LF's and tremor appear and increase as hydrothermal system boils off & gas escapes. Hybrid earthquakes as magma continues to ascend. Tungurahua (1999) = exception… tremor & distal VT’s only…

Localized deformation near summit/vent areas detected as magma rises into edifice. High and/or accelerating rate of deformation

SO2 & Cl emissions increase and SO2/H2S increases as magma rises and system dries out. Gas emissions drop shortly before eruption if quenched- or silica- cap forms

Steam emissions, phreatic and/or phreatomagmatic eruptions.. Water levels in wells may rise as magma/gas pressurizes aquifers or or fall as new fractures open. Geysering possible

Initial tephra mainly composed of lithic debris; glass shards may or may not be present. Juvenile component minor but may be difficult to identify

High VEI (favored by large volumes of gas-rich viscous magma and rapid ascent)

High and rapidly increasing rate of seismic energy release (RSAM), usually includes large LF earthquakes. RSAM continues to increase after initial vent clearing emissions/phreatic /phreatomagmatic eruptions (may pause first). Swarms of deep LP’s or deep tremor indicate significant magmatic replenishment. Escalating explosion signals.

Deformation may be broad, may include fracturing of summit, flank destabilization (e.g., “bulge”). Dike intrusion indicated

Large gas flux following initial “vent clearing” events or emissions decrease abruptly while seismicity still high. Emissions from multiple /distributed vents…

High VEI statistically more likely for “reawakening” volcanoes. History of large eruptions (e.g., big PF sheets, thick tephra, pumiceous deposits, fragmental pumices). Initial mingled magma eruption may indicate new intrusive trigger

Tephra or lava from initial eruptive phase may contain gas-rich indicators (e.g., bubble-wall shards, high volatiles in glass inclusions, volatile-rich indicator minerals, e.g., hb). Silicic magma composition.

Low VEI(favored by smaller volumes of gas-poor low-viscosity magma, and slow ascent

RSAM moderates after initial vent clearing events or earlier large-RSAM crisis indicating earlier emplacement and degassing of shallow intrusion. (Beware: Pinatubo did the above in April, 1991, main eruption was in June). Slow growth and low M of LF or hybrid seismicity

Deformation ends or decreases after initial vent-clearing events.

Gas emissions fall or moderate after initial vent clearing events while seismicity decreasing or constant

History of small eruptions, no extensive PF sheets or thick tephra deposits

Early-erupted juvenile samples degassed (microcrystalline, lack volatile-rich indicator minerals or thick decompression rims = slow ascent). Mafic magma composition (non-exclusive)

No eruption (yet) …or end (intrusion stalls)

Seismicity returning to background levels gradually and stays there for ~ 2 months. Note that many intrusions stall without eruptions.

Deformation ends, deflation may occur

Gas emissions dropping to background levels

Steam (H2O) emissions dominate

Long-dormant (closed systems): No eruptions in many decades or centuries

(e.g., MSH 1980, El Chichon 1982, Pinatubo 1991, Garbuna 2004, Huila 2007, Chaitén, 2008, Sinabung 2010, Vesuvius (0079, 0472,1631), Tambora (2011?))

See McCausland et al. poster for long-dormant

systems matrix

1. Controls on likelihood and explosivity of eruptions

2. Simplified conceptual model of magma ascent

3. Types of frequently-active stratovolcanoes

4. Multi-parameter forecasting guide for frequently-active volcanoes (matrices)

5. Event/probability trees & need for WOVOdat

But first, some context.

Incr

easi

ng e

xplo

sivi

ty

Incr

easi

ng p

roba

bilit

y of

exp

losi

ve e

rupt

ion

Gas co

ntent (s

aturat

ion, s

epara

tion o

f fluid

phas

e, rat

e of b

ubble

grow

th)

Ascen

t rate

(contr

olled

by g

as pr

essu

re, bu

oyan

cy, p

ath ef

fects*

)

V

iscos

ity (c

ontro

lled

by co

mpo

sition

,

Complexly inter-dependent variables that control explosivity of eruptions

Increasing gas content, ascent rate, & viscosity* Path effects include strength and permeability of wall rocks (gas loss), tectonic setting and state of stress, etc.

Low viscosity & low to moderate gas content , moderate to high ascent

rate = low explosivity(e.g., basalt shields)

High gas content, high ascent rate & high viscosity = highly

explosive (VEI >4)

Most common at long-dormant volcanoes(moderate to large volume closed systems

(e.g., Krakatoa 1883; Katmai 1912, Bezymianny 1956; MSH 1980, Pinatubo 1991; Hudson 1991;

Chaitén, 2008; Kasatochi, 2008)

Very high viscosity =

uneruptable magma,

intrusion stalls, gas escapes

Garbuna 2005-06

Explosive fragmentation

Incr

easi

ng e

xplo

sivi

ty

Incr

easi

ng p

roba

bilit

y of

exp

losi

ve e

rupt

ion

Gas co

ntent &

Asc

ent r

ate

V

iscos

ity (c

ontro

lled

by co

mpo

sition

,

Complexly inter-dependent variables that control explosivity of eruptions at frequently active*

(open-system) volcanoes

Increasing gas content, ascent rate, & viscosity

Typical: Low gas, high viscosity & slow ascent: typically andesite to

dacite low-explosivity (VEI 1-3) domes & spines; & stalled

intrusions (Domes: MSH 1981-86, Merapi 1967-2006, SHV 1995-present,, Huila 2008-present, Soputan 2000-2007, Popocatepetl 1995-2005+, Unzen (1991-95), Usu (1910, 1944, 1977-80, 2000), Kelut, 2007; MSH 2004-2008. Stalled intrusions: Garbuna, 2003-04,

Taal 1987-89, 92, 94, 2004, 06, 07, 11; Cotapaxi, 2001-03, Turrialba 2002-present) Spurr 2003-05, Three Sisters 1997-2007, Peulik 1996-98, Baker

1975, Fuji 2001, many others

Exceptional: High gas, high viscosity, and rapid

ascent or unloading: explosive (VEI 3-4+)

basalt to andesite explosive eruptions

(Merapi, 2010, 1872; Masaya, 1754; also Taal 1977, Villarrica 1810,

Pichincha, Soputan, 2008, others)

* One or more eruptions in past 1-3 decades

~4 wt.% H2O (dissolved)

>~70 vol.% bubbles (gas fraction >0.7 =

explosive fragmentation)

~4 wt.% H2O(dissolved)

<~70 vol.% bubbles (e.g., 0.1 wt% H2O)

Volatile loss during ascent*vs.

Eruption likelihood & explosivity: “It’s all about volatile loss during ascent”

* Volatile loss = crystallization & increased viscosity. Requires permeable magma (fractures, foam), & permeable wall rocks, and/or convection w/in conduit

Frequently erupting “Open-Systems” (Hot pathways for gas & magma ascent. Most eruptions smaller than at long-dormant)

Mostly basaltic to basaltic andesite,

e.g., Stromboli, Etna, Soputan since 2007,

Mayon, Fuego, Arenal, Villarrica, Pavlof, Shishaldin,

etc.

1. Wide open

Continuously open conduit; filled with

low viscosity magma; gas escapes at shallow

levels and through Strombolian

eruptions; shallow crystallization may

produce viscous magma & Vulcanian

eruptions

2. Semi-steady state dome extrusions

e.g., Merapi since 1968, Karangetang since 1970?

Santiaguito since 1922, Popo since 1996,

Bezymianny since 1956Huila since 2008

Conduit mostly open; gas partly escapes during ascent; & magma viscosity increases, may

occasionally solidify at shallow levels. Larger

or more gas-rich magma batches

produce Vulcanian eruptions

3. Degassed dome & spine extrusions

e.g., MSH 2004-08, Kelut 2007

Conduit only open at depth; shallow

levels solidify; most gas escapes

during ascent, produces high-

viscosity degassed magma

1. HF Volcano-tectonic (dVT & VT’s); +/- DLPs

2. LF Low-frequency earthquakes,

explosions, + hybrids

3. Volcanic tremor (+/- hybrids)

Typical progression in earthquake types

(long-dormant systems)

Pinatubo examples

Regional fault

Long-dormant volcanoes provide context for evaluating frequently active systems

1. Deep: CO2, deformation, recharge?

2. Shallow: H2S, SO2, Cl & deformation

RSAM -Real-time Seismic Amplitude Measurment

Example from Volcan Huila, Colombia; courtesy of INGEOMINAS

Real-time tools for analysis and forecasting

Long dormant before 2007;frequently active dome extrusion since 2008

Belalcázar■

0 15 km

Rio

Sim

bola

Rio Páez

Eruption possible (magma/fluids moving in crust)Magmatic eruption likely (magma & gas have reached shallow levels)

High VEI (indicated by large volumes of gas-rich viscous magma and rapid ascent)

Low VEI – (indicated by smaller volumes of gas-poor low-viscosity magma, and slow ascent

No eruption (yet) … intrusion stalled…end of crisis



Petrology

Tables of common indicators for likelihood and explosivity… based on VDAP experience
























No single indicator is conclusive! Synthesize!

Don’t try to read this part, I will blow it up!

Forecast Seismic Deformation Gas Observations & history

Eruption possible (magma/ fluids moving in crust)

Sudden changes from background RSAM level (background may include: proximal VT’s, LF’s, tornillos, tremor, explosion signals)Distal VT’s possible, but with cumulative M<4)

If detected, localized near summit/vent areas. No regional deformation unless unusually large batch of magma

Increase in fumarolic activity. Steam (H2O) emissions; S-species mainly scrubbed (H2S). May see sudden changes in emissions as pathways open & close. Increase in CO2 = deep degassing (beware large pulse)

Water levels in nearby wells change.Small increase in “background” monitoring data over periods of months

Open (frequently active) systems: Repeated eruptions during the past 1-3 decades

“Eruption Possible”

Open (frequently active) systems: Repeated eruptions during the past 1-3 decades“Magmatic eruption likely”



Petrology


Rapid & steep increase in RSAM from background level (e.g., <1 week in some cases), LF/explosion signal followed by tremor typical. VT’s may/may not occur. Slower ramp-up in some cases (previous eruptions provide guides). Hybrid & drum-beat quakes may be detected and indicate magma rising


SO2 & Cl and SO2/H2S increase as magma rises and pathway dries out. Emissions may drop shortly before eruption if pathway obstructed

More vigorous steam emissions & phreatic/ phreato-magmatic eruptions or sudden stop of observed emissions (= capped & pressurizing)

Initial tephra mainly lithic debris, glass shards may or may not be present. Juvenile component minor but typically difficult to identify. (Petrology not diagnostic)

10-m

inut

e R

SA

M

Calendar date

VEI 3 eruption of Soputan, 6 June

2008

Little time for warning


“Low-moderate VEI” (e.g., VEI 2-3)



Petrology (retrospective)

Low VEI – (smaller volumes of gas-poor low-viscosity magma and slow ascent)


Deforma-tion ends or decreases, e.g., after initial explosions or lava (dome) erupted

Gas emissions fall to near background levels after initial vent clearing events and in parallel with seismic decrease

History of small eruptions (no big PF deposits or thick tephra deposits). Eruptions may be short duration or long, e.g., peak dome collapse weeks to months after onset of episode

Early-erupted juvenile component degassed (microcrystalline, lacks volatile-rich indicator minerals or they have thick decompression rims = slow ascent. Magma mafic & xtal-poor )

2008 microcrystalline basalt from VEI 2 eruption of Soputan (can be more explosive with higher gas)


“High VEI” (e.g., VEI 4 “100-year” eruptions)



Petrology

High VEI (favored by large volumes of gas-rich viscous magma and rapid ascent



Relatively large gas flux, especially following initial “vent clearing” events. Emissions may drop shortly before eruption if pathway obstructed. Early CO2 pulse & high CO2/SO2 (early warning of recharge and high VEI)


Tephra or lava from initial eruptions may contain gas-rich indicators (high volatiles in glass inclusions, bubble-wall shards, volatile-rich indicator minerals, e.g., hb). Evidence of separate gas phase at depth

14 Oct.

To 4 Nov.

Merapi


“High VEI” (e.g., VEI 4 “100-year” eruptions)



Petrology(mostly

retrospective)






Tephra or lava from initial eruptions may contain gas-rich indicators (high volatiles in glass inclusions, bubble-wall shards, volatile-rich indicator minerals, e.g., hb). Evidence of separate gas phase at depth. Silicic magma composition.

2010 Merapi basaltic andesite block



No eruption (yet) … end of crisis





Key differences in eruption & explosivity indicators at frequently-active vs. long-dormant volcanoes

Frequently-active volcanoes have:“Leaky” hot pathways (conduits, fractures) for gas and magma ascent – this typically translates into:

• History of mainly small to moderate VEI eruptions (vs. large PF sheets)• More rapid progression from heightened unrest to eruption • Lower M seismicity (e.g., <M4 vs. >M4) & lower RSAM• Missing seismic types (e.g., VT’s) possible compared to common long-

dormant sequence (DVT & VT – LF – Hybrid – Tremor)• Deformation mainly shallow & more localized• Gas dominated by shallow- (Cl, F, SO2) vs deep (CO2) degassing species

(unless “100-year” eruption)• Quick progression from scrubbed (H2S) to dry (SO2) species • Generally more mafic magmas; many lack hydrous minerals(Thin- or un-

rimmed hydrous minerals indicate rapid ascent & larger-than-normal eruption)• Often “top driven,” e.g., small dome collapse unloads & triggers larger

eruptive phase … (vs. large summit/flank collapse trigger for long-dormant)

Integrating Event Tree & Monitoring Data“High VEI” Merapi Example



Petrology(mostly

retrospective)






Tephra or lava from initial eruptions may contain gas-rich indicators (high volatiles in glass inclusions, bubble-wall shards, volatile-rich indicator minerals, e.g., hb). Evidence of separate gas phase at depth. Silicic magma composition.

Basis for probability estimation, Merapi, v. 2, 06/26/2006 Conceptual background is in plain type; reasoning behind specific numbers on the tree is in italics.

Node 1: Magma at the surface1 (i.e., 100%). Merapi already woke up from a roughly 5 year sleep, so we can just use 1.0 (100%) as the probability in this

tree.

Node 2: Extrusion rate >100,000, 10-100,000, < 10,000 /d100% for >100,000/day, based on estimates as of 5/14/06 (volume of the dome) and again on 5/20-5/23/06 (combination of dome growth and rate of rockfall and pyroclastic flow deposition in Kali Bedog). As of 6/20 and probably still as of 6/22, the extrusion rate is approximately 210,000 m3/d, of which 75,000 m3/d is growth of the lava dome and about 130,000 m3/d is in rockfalls and awan panas guguran (160,000 m3/d x 0.8 DRE correction, based on PLAZ seismic calibration). This has risen slightly from the 150,000-175,000 m3/d estimated at the end of May.

Extrusion rate is known to be a significant factor in the stability of lava domes, for four reasons:·high extrusion rate creates high shear strains in the carapace of the dome·it creates more weak, soft material in the core of the dome·any extrusion, especially at a high rate, causes loading of rocks beneath the dome and causes dome fronts to oversteepen·high extrusion rates favor incomplete degassing of the rising magma column, with the effect that internal gas pressures can build unless there is efficient degassing (pressure bleed). There is an implicit assumption here that original gas content of Merapi magmas is constant, so variability in gas content of magma reaching near the surface depends on magma ascent rate or its measurable proxy, magma extrusion rate. However, an apparent increase in SO2 emission beginning on June 13 (OMI data) without significant increase in extrusion rate suggests that there might also be an increase in concentration of gas in the magma. We don’t have a specific node for gas concentration in magma but it is indirectly included in Node 3 (gas pressures) if the supply rate of gas exceeds the magma’s capacity to bleed off that pressure.

The basis for estimating extrusion rate = rate of growth of lava dome + rate of accumulation of pyroclastic deposits, e.g., endapan awan panas, guguran.

The long term eruption rate for Merapi was estimated by Siswowidjoyo, Suryo, and Yokoyama (Bulletin of Volcanology, v. 57) as 1.2 million cubic meters per year, or approx 3300 m3/d. Their estimate does not include collapsed material (pyroclastics) – only volumes of lava domes – so it is probably underestimated by a factor of 2 or 3. So, we defined 3 branches here, 0-10,000; 10000-100000; and >100000 m3/d, equivalent in words to “normal,” “active (1990’s style), and “very active”

Duration of the crisis (for annualization of risk, assume duration = 1 year) #1, #2 and #3New dome + sm. Collapses 1

Near-surface gas pressure 60%

#4=1930-scale collapse 2

14% 4%20% 30%

#5 1872 expl eruption 3

Magma supply rate10%Increased

70%#1, #2 and #3

70%

85%#4 1930-scale collapse

56%80% 10%

#5 1872 expl eruption

5%#1, #2 and #3

New dome + sm. Collapses

87%Fluctuates slightly(2nd boiling, release) #4 1930-scale collapse

12% 1%60% 10%

This includes 2nd boiling, #5 1872 expl eruption

Volcano restless constipation/breakthrough scenariosmall explosive events. 3%

~Constant (self-sealing)100% 20% #1, #2 and #3

20% New dome + small collapses8%95%

~ Constant #4 1930-scale collapse8% 0%40% 5%

1%

8%

0%

6%

3%

10%

Scenario

48%

1930 style collapse 10%

1872 style collapse 4%

New dome + sm. collapses

Increasing

constant or decreasing

6%

1%

3%

4%

2010 Merapi: Satellite radar = high initial extrusion rate & high probability of large eruption

• 31 Oct- 4 Nov: ~5 Mm3 dome grows in crater at ≥25 m3s-1; constant MM2-3 tremor (25 km) • 4 Nov.: CVGHM extends evacuations to 20 km – saves thousands of lives• 5 Nov. 00:05: Largest eruption (VEI 4) – ash cloud to 55,000’, pyroclastic flows to 15 km• International response- Indonesia, Europe, US, Japan

North

DLR, German Aerospace Center, 2010

2010 extrusion rate 10 X 2006“The (2006) extrusion rate of ~210,000 m3/d as estimated on 6/20/06, or 2.4 m3/s, is high by Merapi standards and matched within the 20th century only in 1930 and 1961,

and perhaps briefly in the first week of the 1992 eruption. Both times, the high rate seems to have contributed to

large eruptions.” (VDAP-CVGHM Event Tree, 2006)

Eruption Scenario Probabilities (with magma pressure: Increasing : Constant : Decreasing)

Extrusion Rate (m3/s) 1990’s type 1930, 1961 type 1872 type >1.2 73:83:93 15:10:5 10: 5: 1 0.12 – 1.2 83:90:87 10:5: 2 5: 2: 1 <0.12 m3/s 85:80:70 5: 0: 0 0: 0: 0

VDAP-CVGHM Event Tree (2006)

Pallister et al. 2008

Iverson et al. 2006

The next frontier: relating individual event types to detailed processes

• Long-dormant or frequently active?• Frequency, size, and explosivity of past eruptions? (VEI range?)

• Size and character of deposits: Ash flow sheets vs. lava flow, domes & dome-collapse pf’s?, extensive lahars?

• Structure of the edifice? (e.g., stratovolcano? shield? caldera?)• Stability of the edifice: steepness, structural weaknesses, evidence of past collapse,

relation to regional tectonics?• Nature of past eruptions

• Explosive vs. non-explosive vs. fumarolic?• Previous “outsized” eruptions ?

• Composition and character of previous juvenile components?• Vesicular or dense?• Bulk rock and glass compositions (basalt-andesite-dacite-rhyolite)?• Presence and condition of hydrous phases (e.g., amphibole & micas, reaction rims?).

Other indicators of high PH2O or SO2 (e.g., cummingtonite, anhydrite)• Extent of fragmentation (e.g., “gray pumice”)

• Areas affected and population at risk? (Hazard assessment!)

Checklist of questions to pose and attempt to answer before making a forecast (Part 1 – Geologic context)

What does geology and petrology say about past eruptions & hazards?

• Character of the current unrest?• Seismic:

• RSAM (seismic energy trend)?• Type, magnitude and frequency of events (VT, LF, VLF, DLF, hybrids, explosion

signals, tremor)?• Character and duration of events (e.g., distal-proximal VT’s, spasmodic bursts,

continuous or banded tremor, LF’s or tornillos)?• Comparison to background seismicity• Nature of installations (how well-coupled, oriented, distance from vent)• Comparison to previous episodes and/or to analogous unrest at other

volcanoes?• Deformation:

• Inflation, deflation, and/or lateral?• Rate… and rate of change?• Deep? Shallow? Large-volume? Small-volume?, Geometry?• Comparison to previous episodes and/or to analogous unrest at other

volcanoes?

Checklist of questions to pose and attempt to answer before making a forecast (Part 2a – Monitoring data)

• Gas:• Emission levels, ratios and trends of: SO2, CO2, H2S, Cl, F, any other species? • Likelihood of groundwater “scrubbing” (conversion of SO2 to H2S)?

• Hot? dry? pathway to the surface?• Conduit plugged by hydrothermal or solidified cap?

• Comparison to emissions during previous episodes?

• Observations (including remote sensing)• Morphologic changes (e.g., fractures or other structures)?• Character of Initial explosions/extrusions and any associated tephra?

Groundwater changes?• Vegetation changes? (e.g., tree-kills?)

• Comparison to previous episodes and/or to analogous unrest at other volcanoes? e.g., “ What proportion of previous volcanoes that exhibit the observed indicators continued to magmatic eruption and how big (VEI)?

• Event-probability tree?

This is why we need WOVOdat !

-END-

Checklist of questions to pose and attempt to answer before making a forecast (Part 2b – Monitoring data)

1. HF Volcano-tectonic earthquakes (distal vs. proximal)

2. LF Low-frequency earthquakes

3. Volcanic tremor and

Typical progression in earthquake types

Plus Hybrids of HF & LF

Original signal

Low pass filtered

Spectrogram

Power Spectrum

Real-time tools for analysis and forecasting – associating event types with processes

Hybrid (VLF + HF) volcanic earthquake: represents rock breaking & magma or fluid transport


Petrology

Eruption possible (magma/fluids moving in crust)

















Low VEI – (favored by smaller volumes of gas-poor low-viscosity magma, and slow ascent






No eruption (yet) … end of crisis





Open (frequently active) systems: Repeated eruptions during the past decadeConduit never freezes: e.g., basaltic & basaltic andesite systems like Stromboli, Etna, Soputan, Mayon, Fuego, Arenal, Villarrica, Telica, Llaima, Masaya 1970-1990, Mayon, Pavlof, Shishaldin and semi-steady state andesite to dacite dome extrusions like Santiaguito, Merapi, Karangetang, Huila since 2008, Popocatepetl since 1996, Colima since

1997, SHV since 2000 in which magma supply just enough to keep magma rising; occasional larger than normal influx or gas-rich influx drives more explosive eruption (e.g., Merapi 2010) may also be driven by “top down” processes (e.g., dome-collapse unloading as in Boxing Day 1997 eruption at SHV; summit dome collapse at Soputan in

2007-08)



Petrology

Magmatic eruption possible (magma/fluids moving in crust)

Background level exceeded. Distal VT's as tectonic faults are pressurized and magma rises. VT number & energy increasing as magma rises. VT energy release often approaches or exceeds M4. (Spurr only M3). Also many intrusions stall without eruptions!

Inflation detected = early warning; may be large. Broad (deep sourced) deformation

CO2 increase = deep degassing, may provide early warning

Hydrothermal system & fumaroles present but not necessarily elevated over background levels. Past history of explosive eruptions. (Montserrat no steaming). Water levels in wells may drop (e.g., due to dilatation)

Products of past eruptions give range of most likely magma types and explosivity


Transitional from distal to proximal VT's as magma shallows. LF's and tremor appear and increase as hydrothermal system boils off & gas escapes. Hybrid earthquakes as magma continues to ascend. Tungurahua (1999) = exception… tremor & distal VT’s only…

Localized deformation near summit/vent areas detected as magma rises into edifice. High and/or accelerating rate of deformation

SO2 & Cl emissions increase and SO2/H2S increases as magma rises and system dries out. Gas emissions drop shortly before eruption if quenched- or silica- cap forms

Steam emissions, phreatic and/or phreatomagmatic eruptions.. Water levels in wells may rise as magma/gas pressurizes aquifers or or fall as new fractures open. Geysering possible

Initial tephra mainly composed of lithic debris; glass shards may or may not be present. Juvenile component minor but may be difficult to identify

High VEI (favored by large volumes of gas-rich viscous magma and rapid ascent)

High and rapidly increasing rate of seismic energy release (RSAM), usually includes large LF earthquakes. RSAM continues to increase after initial vent clearing emissions/phreatic /phreatomagmatic eruptions (may pause first). Swarms of deep LP’s or deep tremor indicate significant magmatic replenishment. Escalating explosion signals.

Deformation may be broad, may include fracturing of summit, flank destabilization (e.g., “bulge”). Dike intrusion indicated

Large gas flux following initial “vent clearing” events or emissions decrease abruptly while seismicity still high. Emissions from multiple /distributed vents…

High VEI statistically more likely for “reawakening” volcanoes. History of large eruptions (e.g., big PF sheets, thick tephra, pumiceous deposits, fragmental pumices). Initial mingled magma eruption may indicate new intrusive trigger

Tephra or lava from initial eruptive phase may contain gas-rich indicators (e.g., bubble-wall shards, high volatiles in glass inclusions, volatile-rich indicator minerals, e.g., hb). Silicic magma composition.

Low VEI(favored by smaller volumes of gas-poor low-viscosity magma, and slow ascent

RSAM moderates after initial vent clearing events or earlier large-RSAM crisis indicating earlier emplacement and degassing of shallow intrusion. (Beware: Pinatubo did the above in April, 1991, main eruption was in June). Slow growth and low M of LF or hybrid seismicity

Deformation ends or decreases after initial vent-clearing events.

Gas emissions fall or moderate after initial vent clearing events while seismicity decreasing or constant

History of small eruptions, no extensive PF sheets or thick tephra deposits

Early-erupted juvenile samples degassed (microcrystalline, lack volatile-rich indicator minerals or thick decompression rims = slow ascent). Mafic magma composition (non-exclusive)

No eruption (yet) …or end (intrusion stalls)

Seismicity returning to background levels gradually and stays there for ~ 2 months. Note that many intrusions stall without eruptions.

Deformation ends, deflation may occur

Gas emissions dropping to background levels

Steam (H2O) emissions dominate

Long-dormant (closed systems): No eruptions in many decades or centuries

(e.g., MSH 1980, El Chichon 1982, Pinatubo 1991, Garbuna 2004, Huila 2007, Chaitén, 2008, Sinabung 2010, Vesuvius (0079, 0472,1631), Tambora (2011?))

PATH 1 Eruption (likely VEI <=3)No additional intrusions 2.5% Tephar fall & ash cloud

(w/in next 3 years) 10% 21% 18.2. 25% 100.0% 5 to 10 cm isopacks @ >5 km

25% Activity dies out23%90%

Eruption (VEI <=3)(small to moderate)

PATH 2A 2.3% PFCurrent situation: No Eruption (within 10% 6. 21% 19.

Intrusion emplaced ~2 mo. of intrusion) 100.0%100% global 4. 23% global basis

basis 30%Activity dies out

20%90%

LaharVEI 7. 21% 19.

PATH 2 global Large, VEI>3 100.0%Additional intrusions basis 5. 21% (w/in next 3 years) 40.0%

3. 75%75%

Sect.collapsePATH 2B +/-blast

Eruption (within 8. 1.05%~2 mo. of intrusion) local/Garbuna 5.0%

4. 53% basis70%

Small to moderate Lava domeVEI<=3 9. 11% 20.

5. 32% 50.0%60.0%

See continuation

DRAFT (for RVO revision) Probability Tree for Eruptive Activity at Garbuna Volcano, PNG for the period

1/27/06 and extending for the next 3 years. Highlighted cells track most likely paths.

Note: "Activity dies out" = continued gas emissions,

hydrothermal explosions and possibly surges

How big?

Current Status

What kind of eruption?What to

expect next

Quantifying forecasts: Use of Event & Probability Trees at

Garbuna, PNG

Tephra

5 to 10 cm isopacks @ <5 km& ash could tops <10 km

Eruption (likely VEI <=3) 4%No additional intrusions 2.5% Tephar fall & ash cloud 20%

10% 21% 18. (see hazard maps)100.0% 5 to 10 cm isopacks @ >5 km

Activity dies out & ash could tops >10 km23% 17%90% 80%

limited to <5 km(small to moderate) 4%

PATH 2A 2.3% PF 20%No Eruption (within 10% 6. 21% 19. (see hazard maps)~2 mo. of intrusion) 100.0% extends to >5 km

4. 23% global basis 17%30% 80%

Activity dies out20% limited to <5 km90% 0%

Lahar 0%VEI 7. 21% 19. (see hazard maps)

global Large, VEI>3 100.0% extends to >5 kmbasis 5. 21% 21%

40.0% 100% 3.

limited to <5 kmSect.collapse 0.4%

PATH 2B +/-blast 40% No changes allowed here

Note: "Activity dies out" = continued gas emissions,

hydrothermal explosions and possibly surges

Areas affected

Vuln

erab

ility

& R

isk

Probability Trees – Lessons from MSH, Garbuna, Merapi, Huila, Lunayyir

• Internal observatory use valuable• Focuses scientists’ thinking • Illuminates alternate viewpoints &

uncertainties• Aids in reaching consensus

• External: Use with care - requires user education

• Written explanations & scientific background essential (i.e., meta-data required!)

• For well-studied & well-monitored volcanoes nodes can be linked to measurable parameters !

Tephra

cloud

cloud

Augustine, 2006

25 major crises Infrastructure in 12 countries

Documents