We have received many queries that implicitly ask a
question about how old something is – typically the odd rock that they found in
their backyard. A review of age dating, and why it can be critically important,
seems appropriate here.
Q:
How old is this rock?
- various
In 1979, a young USGS geologist named Rick H. was
mapping flows on the north flank of a beautiful, glacier-covered, symmetric
volcano in southwest Washington State. He had some idea of how old the Goat
Rock Dome was that he stood on – textures and some sparse historical information suggested that it was
very young. He had no idea that in a year the huge outcrop he stood on would be
moved many miles to the north – that the spot he then stood on would be more
than a thousand feet up in the air. The 1980 eruption of Mount St Helens killed
59 people – that is, that authorities are certain of. The death-toll could have
been greater by more than 800 people. By a miracle of governance, those people
were being held back at a roadblock until 9am on May 18th. However, 45
minutes before the gate would have been opened - to allow in people who had property around the volcano - the monster blew up catastrophically with a
lateral blast. Authorities were certain of 57 people killed, but speculate that there were more caught in the blast that no one knew about. Gray dacite dust fell on cars in Atlanta a day and a half later.
As the new chief scientist for volcano hazards, I
made a point of visiting - and spending time listening to - every single staff
member scattered over six different centers. This was not a trivial exercise;
it meant lots of airport time and lots of listening. I made a point of spending
the same amount of time with the technicians as I did with the senior
scientists, and this helped me get a better view for how things were really working
within the organization that I had just inherited. In one case, techs clued me
in that one of our observatories was no longer functioning, due to a perfect
storm of very human conflict starting with a management failure.
The experience was not all grief, however. One of
the people I spent time listening to was Andy C., a brilliant young PhD geologist/geochemist who had
decided to specialize in age-dating. I also talked with Jim S., a smart,
furiously hard-working tech who worked with him. After listening to them, I
travelled to the USGS national headquarters in Reston, VA, and tin-cupped
around the building. I mean this literally – I carried around a tin cup with me
to help break down resistance by disarming people with humor. I was looking for
“spare change” in people’s budgets that I could divert to Andy’s laboratory.
Spare change in my little 120-person volcano science team usually meant a few
thousand dollars from my cost center budget. Two thousand dollars would pay for
a young scientist to attend a science meeting, where he would not only learn
what was going on in his field, but connect with people he could cooperate
with. This meeting attendance had the effect of leveraging our meager funds to
accomplish quite a bit more with them by getting others to help us accomplish
our objectives.
To put this in perspective, a Stryker combat vehicle
costs about $1,500,000. For the Department of Defense, however, I was looking
for funding down in their noise-level. For them, our needs were what Sherrie
G., a DOD executive, dismissed as “decimal dust.”
But Andy had both energy (he frequently worked
60-hour weeks) and a vision. His vision was to create a center of excellence: a
laboratory for high-precision dating. I found and funded him with $250,000 to
develop the world’s best state-of-the-art 40Ar/39Ar
age-dating laboratory. This has allowed him to steadily refine the ages of very
young volcanic rocks, which in turn allows us to put better and better
parameters on the eruptive frequency of volcanoes – so we know what we may or
may not have to worry about.
In the 10 years that followed, Andy had accumulated
a sufficient number of very good age-dates that he was able to start looking in
broad brush at the eruptive history of the entire Cascades range - and he can
now see episodic pulses of eruptive activity in the past 500,000 years.
Importantly, this includes the first hints that we are entering into an unusual
period of volcanic activity right now.
In the past 5 years, Andy refined his +/- errors on
an age-date from several thousand years to just 400-500 years. He did this by
gaining precise control on the atmospheric Argon component in rocks, but also
by refining his sample-collecting techniques. He sought the centers of lava
flows, parts that were “platy” because they were shearing as they cooled while
still flowing. He also avoided porphyries, because the internal crystals formed
under different circumstances than the fine-grained that had extruded out onto
the Earth’s surface. Other things he avoided included air-vesicles and glass: that
is, water-quenched lava that inevitably had an Argon contribution from the
water. How did he sort these out? By licking the rock with his tongue. If his
tongue stuck slightly to the rock, experience showed that he would get a
low-precision date on it. Since he was working long hours to deal with a huge
back-log of samples, this kind of “pre-sorting” had gone a long way toward
narrowing down his error-bars on dates.
So..... How Does Age-Dating Work?
Wait, you ask, how does age-dating work? In the pre-radioactive isotope days, crude
dating could be done by measuring how fast sediments accumulated in a
lake-bottom, measuring the thickness of a stack of those sediments, and noting
which sedimentary units lay above (were younger than) another layer. But while
you could easily get the relative
ages with stratigraphy, you couldn’t get good absolute ages – there were too many variables, like water
levels, wind influences, and changing sedimentation rates.
It’s not hard to get the radioactive decay rates of
anything if you have something like a Geiger counter: a certain number of atoms
are “popping” every minute, and you could measure how many atoms were in the
sample to begin with using some relatively straight-forward chemistry.
Once you have decay rates, age dating - in principle
at least - is pretty straightforward. A
mineral solidifies out of a magma mush somewhere with uranium in it. By early
in the 20th Century, the decay-process of uranium was well known…
there were intermediate “daughter products” with different half-lives, but they
all ended up at stable lead, where the decay process ended. All you really
needed to do was measure the lead-to-uranium ratio precisely, and with the rate
of decay you could get a handle on how long it had been sitting there since the
last melt solidified. This wouldn’t work if the rock had been metamorphosed (or
“stewed and cooked” as old miners would say it) since the initial solidification.
In that case, the age you got was the last re-melt.
All is not lost, however. Some really smart people eventually
worked out how to get something of a handle on even this. It required some good
geology, some good chemistry, and some clever statistics… but you could at
least get an idea if something had been messed with since original formation.
There were other problems, however. The half-life of
uranium-235 is 704 million years. Also, the precision was not that hot when you
measure micrograms of uranium and lead in a mass spectrometer - and try to
divide that up into 704 million years. In the best of circumstances, you get a
rather large plus-or-minus – many thousands, even hundreds of thousands of
years. In many situations that didn’t matter all that much. For instance when
we were trying to figure which rocks arrived before which in truly ancient
southern Venezuela – then 500,000 years or 5,000,000 years one way or the other
didn’t matter all that much.
If you have two volcanoes, and one erupts every
20,000 years and the other erupts every 200 years, then this precision doesn’t
help you at all. Unless you know the eruption frequency, you have no easy way
to know how dangerous the volcano is.
THIS is why age-dating volcanic rocks is so
important. Do we pour our meager annual instrumentation budget into Mount St
Helens, or into (literally) Crater Lake?
There are other radioactive decay series, of course:
rubidium-strontium, carbon-14 (which we use with volcanoes if we can find some
fried vegetation under a flow), and the argon series. You can’t always find
uranium-hosting minerals, and you usually can’t find rubidium-hosting minerals.
If you find something burned under a lava flow, it can’t be older than a few
tens of thousands of years, or the 14C is already all gone. But 40Ar
has a more useful half-life, and argon is also a significant constituent of the
atmosphere. It can be found just about anywhere in volcanic rocks – which came
from a subducted ocean floor once exposed to the atmosphere.
Rats. There is yet ANOTHER problem: all that argon
in the atmosphere is also going to pollute any measurement you make. Like
carbon-14, which is “made” in the upper atmosphere by cosmic rays transmuting
nitrogen molecules, Argon-40 comes from atmospheric Argon-39 – which is
everywhere, and seeps into just about everything. If you want real numbers –
true age dates – you must find a way to get clean and unsullied samples.
Where there’s a smart person, there’s a way. Believe it or not, it comes down to something as low-tech as putting your tongue on a rock. Andy - by trial and error - found that the best rocks to date were the ones in the "platy" middle section of a solidified lava flow. The final test was to see if your tongue sticks to the sample that you hammered out. Does it stick? Chuck it and look for another sample, because it won't give you a reliable age-date.
Bottom line:
It comes down to this: if we are being shot at, it’s
important to know how OFTEN we are being shot at. You can plan. You can set up
many forms of disaster mitigation to keep a crisis from becoming a catastrophe.
Rock-dating information is crucially necessary in order to have even half a
chance of predicting the volcano’s future behavior – and roughly calculating the
risk it carries. High-risk volcanoes then claim the larger share of our very
limited instrumentation budget. Crater Lake (the former Mount Mazama) last erupted
catastrophically over 7,000 years ago. Mount St Helens has erupted more
frequently than almost all the other Cascades volcanoes combined… the last period of repose was just 24 years. So with good age-dates, we invested the lion’s
share of instrumentation on this critical, very-high-risk volcano.
We were relieved that we had done so when the 2004
eruption started with just over a week’s seismic warning. We had seismic and GPS “eyes” already in place and with them we could ”see” what was going on under the volcano. With a huge
experience base acquired by studying hundreds of volcanoes in the US, Kamchatka, Japan, Indonesia, and Latin America, the scientists at
the Cascades Volcano Observatory could predict (sometimes to within hours) when
the next eruptive pulse was coming – and on 1 October 2004 they called for the evacuation of hundreds of people from the
Johnston Ridge observatory.
They could not have made that emergency call if the
geologists had not already carefully mapped its past eruptive products. And the
eruptive history would not have been deciphered without precision age-dating.
The 2004 eruption was not nearly as violent as the one in 1980. But even if it
had been, the disaster of 1980 would likely not have been repeated. By 2004 we
had dates and knew what this volcano was likely to do.
~~~~~