From Millennia to Minutes: Life Cycles of Volcanoes and Eruptions

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Arc volcanoes often show evidence for new magma coming from the lower crust and interacting with old magma in the shallow crustal storage system. This presentation examines the interplay of old and new magma, and the importance of collecting samples during an eruption and over longer periods to create time series information that can reveal where a volcano is in its life cycle, and how the volcano might respond to the introduction of new magma. This virtual presentation was given at the American Geophysical Union (AGU) Fall meeting.


Date Taken:

Length: 00:15:22

Location Taken: Anchorage, AK, US


Arc volcanoes often show evidence

for new magma coming from the lower crust,

and interacting with old magma

in the shallow crustal storage system.

The nature of the upper crustal system

will affect whether this new magma

causes an eruption and what the eruption

will look like.

I have chosen this figure from a recent

synthesis of Andean volcanism

whereby the rate of mafic recharge

affects the nature of the resulting volcanism

to illustrate this concept.

What I'll discuss today,

is the idea that understanding interplay

of old and new magma in the upper portions

of the trans-crustal system requires

accurate time series information

about what is being erupted.

This can be obtained by frequent

sampling throughout eruptions.

I'll also show how time series data

over longer periods, can reveal

where a volcano is and its life cycle,

and how this impacts its response

to the introduction of new magma.

Most real-time and near real-time volcano

monitoring data, including seismic, geodetic,

infrasound, remote sensing, and gas are commonly

plotted and analyzed as time series.

This allows an understanding of how

active volcanic systems evolve with time,

how precursory signals escalate or de-escalate,

and how eruptions progress and ultimately end.

Despite this being a session on subduction zone volcanoes,

I couldn't resist using the 2018 Kilauea eruption

as a beautiful example of this,

where geophysical signals tracked caldera collapse,

dike opening, and sustained eruption.

Of course, other recent subduction zone eruptions

could be shown as well,

such as Montserrat, Redoubt,

Mount St. Helens, or Pinatubo.

One of the reasons I chose to highlight Kīlauea, however,

is to show that geologic time series data can be as

powerful as other data types for understanding

ongoing eruptive processes as described by Gansecki et al.,

2019, changing compositions of erupted lava from

fissures showed tapping and then flushing of

residual magmas within the lower East Rift Zone.

Eventually, new summit derived magma dominated

the eruptive products as shown in blue on the upper right.

During this eruption, compositional data were

available in near real-time to inform the ongoing response.

Additional later analysis validated

the models that were developed from the initial data and

the overall effort demonstrated

the geochemical analysis can inform

hazard assessments during an eruption.

As seen at Redoubt Volcano in Alaska in 2009,

the earliest eruptive products often hold the key to eruption initiation.

Following months of elevated seismicity, gas emissions, and melting of glacial ice,

the 2009 eruption consisted of 19 explosions and lava dome growth.

The upper panel on the left shows whole rock compositions of pryoclasts

and lava throughout the eruption. 

The earliest sampled material contain low silica andesite,

which was absent in later explosions

and during the dome growth phase.

Sampling the tephra from the early explosions

required rapid mobilization to the field,

as seen in the figure on the left.

We interpreted that the low silica andesite ascended from

the lower crust during the eight or more months

prior to the first eruption,

but that the magma stalled and accumulated

in the upper crust where it's phenocryst rim,

and melt compositions were established.

Ascent of new low silica andesite

through stagnant mushy, old intrusions,

residual from earlier Redoubt activity

mobilized differentiated magma pockets,

and interstitial liquids.

These are the high silica,

and intermediate silica andesites that fed

the subsequently erupted lava domes.

For the rest of the talk,

I'd like to focus on Augustine volcano,

a frequently active volcano in Alaska's Cook Inlet

that has erupted numerous times historically,

last in 2006.

The 2006 eruption started with a series

of 13 explosive events that produced fall,

pyroclastic flows and lahars,

as seen in blue and green on this map.

This period was followed by four days

of continuous block and ash flow production

known as the continuous phase and shown in deposits,

the deposits shown in pink on the map,

and finally an effusive phase that produce two lava

flows down the north side of the edifice, as seen in brown.

Like Redoubt in 2009,

the eruption yielded mostly low through high silica andesites.

Mapping of deposits from the eruption

allowed sampling of many of the individual events.

Petrologic study of products from Augustine,

in contrast to Redoubt,

suggests that all of the erupted andesites had been

stored in the shallow crust prior to the eruption.

Recent experimental work by De Angelis et al.,

in 2020, shows shallow storage

for the high silica andesite consistent with

geophysics and melt inclusion entrapment pressures.

Note also that within a single end member lithology,

differences and some trace elements suggests

multiple subtly different magma bodies.

For example, as seen here with chromium concentrations

in the low silica or LSA end member.

The presence of mafic phenocrysts as well as

basaltic inclusions as shown circled in red on the right,

and first published by Steiner et al., indicate that the eruption was preceded by

an injection of new basalt or

basaltic andesite into the old shallow andesitic magma system.

Recent work at Augustine,

led by Mary Benage,

focuses on a specific interval within the 2006 eruption,

and looks at the transition between

punctuated individual vulcanian blasts of

the explosive phase, and what caused a transition to

more sustained eruptive activity during the continuous phase.

In this study, we focus on event 9,

a single vulcanian blast on January 17th,

2006, and event 10,

which occurred 10 days later,

and produced two pyroclastic density currents,

a smaller one during early event 10,

and the largest PDC of the eruption,

the Rocky Point pyroclastic flow during later event 10.

Just after event 10,

or shortly thereafter,

the continuous phase of the eruption started.

Previous work had recognized an increase

in the proportions of high and intermediate

silica andesites at this transition point,

and that is confirmed by componentry

showed here on the left.

Such pyroclast componentry is qualitative

and based on visual characterization,

but provides important information about

changing role of different magmas

throughout an eruption.

In looking in detail at pyroclasts,

we found that even within a single vulcanian event,

a range of temperatures,

whole rock and melt compositions were present.

The two panels here show iron titanium oxide temperatures on

the left and glass compositions on the right, symbolized

by individual low silica andesite and

high silica andesite pyroclasts from both events 9 and 10.

We conclude that the heterogeneity observed shows that

low to high silica andesites repeatedly ascended through shallow dikes,

mixing and mingling prior to and during each vulcanian explosion.

Looking specifically at the transition

to late event 10, which produced the largest

volume PF of the explosive phase and occurred

just at the start of the continuous phase,

we found a shift to

more consistent groundmass glass compositions

as shown on the right panel.

As seen on the left,

we also observed higher plagioclase number densities

for the latter part of the event 10,

suggestive of higher ascent rates.

You could also see the later event 10

plagioclase number densities are close

to those observed for subplinian to plinian eruptions,

whereas pyroclasts from event 9 and early 10 are more

similar to previously published values for vulcanian explosions.

We interpret this shift from

heterogeneous groundmass class compositions

to more homogeneous products seen

at that juncture of the eruption as due to

an interconnection of ascent pathways

within the shallow magmatic plumbing system,

thereby tapping a larger magma body and shifting from

punctuated more closed system behavior

towards open system behavior fed by this deeper magma body.

The higher ascent rate accompanies this transition.

This transition is visualized in this schematic

cartoon on the left,

where the three, shows interconnection of larger bodies of

low silica andesite and high silica andesite

during late event 10.

Prior explosions were fed by smaller bodies

or dikes within the shallow system.

Overall, the eruption of heterogeneous andesites in 2006,

fed by a spatially complex and heterogeneous

system of small volume intermediate

magma bodies in the shallow crust is

similar to what occurred in 1986 and 1976.

This is shown very schematically on the right

where unrest is shown in gray,

and general eruptive activity is shown in red.

Overall, these modestly sized dome

forming eruptions appear initiated

by frequent injections of mafic magma

from the deeper crust.

Next, I'd like to look back further in

Augustine's eruptive history

to provide context in which to place the current magmatic system.

In the middle photo,

you can see a tephra section with several late Holocene tephras,

first identified by Waitt and Beget in

their 2010 USGS Professional Paper that span

from a few 100 years to about 1700 years before present.

To the left of that a representative strat section

that was logged high on

the south flank of the volcano extends back

even further to the early Holocene.

Note that these tephras,

while their volumes are not accurately known because of Augustine's island setting, are

coarser and thicker than deposits from

historical eruptions such as the ones I've just talked about.

This can be illustrated by the yellow box

that you see at the top of the photo that

shows all of the post-1912 tephra falls

fall within that short interval.

When we look at whole rock compositions of tephras

from early Holocene through 2006,

we see gradual trends that reflect changes to

the underlying magmatic system feeding these eruptions.

Younger tephras are more enriched in incompatible

elements such as potassium,

as shown on this Harker diagram on the right

where samples are color-coded by time.

We recognize that sampling bias has probably caused

undersampling of older, more modest-sized eruptions.

This may especially impact the apparent lack

of more mafic older tephras.

The increase in incompatible elements, however,

is present from the early Holocene through late Holocene,

but prehistoric tephras, and it is evident even if you

exclude the historical compositions which

are shown mostly in pink.

In addition to the whole rock compositions

from the previous slide,

we started looking at glass compositions as

well in part because that allows us to look at

distal samples where a more complete and more

easily dated tephra record is preserved.

We have found similar trends in glass

as whole rock as seen on the left.

Specifically, incompatible elements are increasing

through time despite more mafic bulk compositions.

Here, we illustrate this with a ratio of strontium

which is compatible and largely constant with time,

versus yttrium which is incompatible

and increases through time.

Next, we've investigated what intrinsic

controls could underlie this change.

First, temperatures calculated

with magnetite-ilmenite pairs.

These seem to increase into the late Holocene

as shown on the upper right.

With this change, we observe an obvious change

in mineralogy as well, specifically amphibole,

which was very abundant in early Holocene tephras,

and rather rare in late Holocene and historical eruptions.

Water concentrations, however, are not appreciably

lower as we might expect if magmas are drier.

This just shown on the lower right.

We're still at the speculation stage about why

this change is occurring or these changes are occurring.

Possibilities which aren't all mutually exclusive

include increased mafic recharge,

which would lead to hotter eruption temperatures,

less ampibole and less chance for large,

cool evolved bodies to assemble.

Other possibilities include lower crustal or

mantle source of the mafic input is changing,

or perhaps more heat is required to mobilize eruptible magmas in

the upper crust as these upper crustal magmas

become more refractory.

Understanding the cause and effects

between these changes in the magmatic system,

and changes in the style and magnitude

of volcanic eruptions will assist in forecasting

future activity at Augustine and similar volcanoes.

I'll conclude by restating that time constraints samples of

eruptive products are key to understanding

the current state of a volcanic system,

what initiates eruptions, and what

causes changes in eruptive style.

Also I just want to show these photographs of samples from

the 2018 Kīlauea eruption collected at HVO,

"jaws" collecting samples from the

Mount St. Helens lava dome in 2006,

and Matt Loewen sampling ash samples

from the 2018 eruption of Veniaminof in Alaska.

Thank you.