IRA FLATOW, HOST:
This is SCIENCE FRIDAY. I'm Ira Flatow. On March 22, 2009, Alaska's Redoubt Volcano erupted. It let out a plume of ash that reached as high as 60,000 feet. It even stalled the engines on a KLM flight that was on its way to Tokyo. But that was nothing compared to what was going on underneath the volcano. Researchers say right before the eruption, the volcano experienced a series of small earthquakes. So many earthquakes that they created what researchers are calling a scream.
(SOUNDBITE OF VOLCANO SCREAM)
FLATOW: Wow. Alicia Hotovec-Ellis is a doctoral student in the Department of Earth and Space Sciences at the University of Washington. She's co-author on a recent paper in the journal Nature Geoscience that discuses the use of earthquakes. She joins from KUOW in Seattle. Wow. That is a dramatic recording.
ALICIA HOTOVEC-ELLIS: Yeah.
FLATOW: What is going on there?
HOTOVEC-ELLIS: All right. So the research in a nutshell is that prior to several of the explosions during this 20 explosion series that Redoubt had in 2009 there were these strange harmonic signals that the scientists at the Alaska Volcano Observatory endearingly dubbed the screams because harmonic tremor - other volcanoes usually only gets up to three, maybe five hertz at most in pitch. And these screams got up to 25 or even 30 hertz, which is even audible to the human ear. So, yeah...
HOTOVEC-ELLIS: ...I mean, for reference, that's about the lowest note on an 88-key grand piano, so.
FLATOW: So, wow, what is causing these little mini earthquakes under there?
HOTOVEC-ELLIS: Right. So when I went back through and looked at the seismic data a little bit more carefully, I noticed a series of these repeating of earthquakes that got closer and closer and closer and closer together in time. And looking at the mechanism of those earthquakes, there was nothing to suggest that it was anything but just regular old stick-slip on a fault. So, two rocks in a subsurface scraping against each other.
FLATOW: And those are earthquakes, the sound that we listened to that were sort of speeded up, right?
HOTOVEC-ELLIS: Yeah, yeah, 60 times. So that was actually about an hour of seismic data compressed into about a minute.
FLATOW: But we have some more audio clips so let's see what they are. We have something we're going to be calling the drumbeat. So let's listen to that.
(SOUNDBITE OF VOLCANO SCREAM)
HOTOVEC-ELLIS: So each of those pops is an individual earthquake.
FLATOW: Each one.
HOTOVEC-ELLIS: Each one, including the little crackling in the background.
(SOUNDBITE OF VOLCANO SCREAM)
FLATOW: It sounds like the Fourth of July.
(SOUNDBITE OF VOLCANO SCREAM)
HOTOVEC-ELLIS: So the buzzing sound is when the earthquakes are coming at about one per second. That particular scream gets up to about 10 hertz. So you wouldn't be able to hear that one. And then the other thing that puzzled us when we first looked at this was there's a little bit of a pause. It's about 30 seconds in the seismic trace, or about half a second when we speed it up. And we weren't quite sure what was going on with that at first. One theory was that the screaming was going up beyond what the sensors could actually listen to or that the earthquakes turned off, specifically like whatever was moving stopped.
It got stuck. But what we ended deciding on after some modeling that was done by some co-researchers down at the - down at Stanford, we decided that it's actually where the earthquakes are coming so fast that they can't keep up. So the fault that's moving, it's still sliding, but it's kind of like when you open up a door and it's creaking.
HOTOVEC-ELLIS: If you do it slowly...
HOTOVEC-ELLIS: ...it just goes...
(SOUNDBITE OF MAKING SOUNDS)
HOTOVEC-ELLIS: But if you move it fast enough, it doesn't squeak sometimes. So we think that it's still moving, but it's just not making any sound anymore. So we call this stable aseismic sliding.
FLATOW: Now, is this predictive if you - can you hear these little earthquakes happening before a volcano erupts and know something is going to happen?
HOTOVEC-ELLIS: Right. I mean on just about every volcanic eruption that I've ever looked at, there's almost always quite a bit of seismicity associated to the buildup before the volcano finally goes. The screaming, however, and these kind of accelerating earthquakes, they come only in perhaps the last few hours or minutes before an explosion. So it might be able to give a little bit of a short-term warning, but as you might know from a previous segment that you guys did on Redoubt, it had been having earthquakes since the previous fall.
HOTOVEC-ELLIS: So it's a little bit difficult to tell what a volcano is going to do even when we have seismic signals to try to figure it out.
FLATOW: And so how many earthquakes around the world or the country are metered so you could hear the screaming going on?
HOTOVEC-ELLIS: Oh, gosh. Under Redoubt, it's got to be thousands.
HOTOVEC-ELLIS: Thousands and thousands of little itty bitty earthquakes.
FLATOW: And so...
HOTOVEC-ELLIS: I don't know what the full catalog is, but I'm sure it's massive.
FLATOW: So we could go around and try to collect more of those screams from the earthquakes if we wanted to.
HOTOVEC-ELLIS: Yeah. Presumably, assuming that the volcano that you're listening to is capable of making the scream. So like I said, there are - there have been other volcanoes that have kind of this harmonic tremor, but it never gets up really, really high. So we're not entirely sure yet why Redoubt is so special, in that it gets up so, so high and the earthquakes are coming so fast. We haven't quite figured out what it is that's going on down there that's driving the earthquakes to be happening so quickly.
FLATOW: So what kind of research could you do to find that out?
HOTOVEC-ELLIS: Well, so far, I think I've done about all the observing...
HOTOVEC-ELLIS: ...that I can with this data.
HOTOVEC-ELLIS: So it's kind of in the hands of modelers right now. And, well, I did some modeling for my master's, and I discovered that I just don't have the patience for it. So I know that Eric Dunham and his colleagues at Stanford are thinking of doing some more computer modeling to try to explain what's going on with first principle physics.
FLATOW: Are there actual sensors? Are there microphones? How is it - how do you make the sound? How is reproduced?
HOTOVEC-ELLIS: So, on the volcanoes, most of the ones that we're interested in listening to, we have a network of seismometers...
HOTOVEC-ELLIS: ...which are kind of like souped-up microphones, except that you stick them inside the earth. And also at readout, there was a microphone array, but we actually didn't hear the screaming on the microphones, I think, personally, because they were too far away. But it did pick up the explosion pretty well.
FLATOW: And how did you pick Redoubt to be one of these test cases where you would have a lot of instrumentation there?
HOTOVEC-ELLIS: Well, it happened to be instrumented already because of the previous eruption in 1989 to 1990 that almost downed the KLM flight. So we've already had a pretty good network. And when the volcano started to make some more noises, the scientists up there put out several more instruments on the volcano just before it erupted so that they can listen even a little bit closer.
FLATOW: Does it make noises but not erupt?
HOTOVEC-ELLIS: Absolutely. Volcanoes are notoriously noisy, but they usually get much more so when they're unhappy.
FLATOW: So there's no predictive level where you say, oh, boy, we know this one is going to erupt, but, no, this is just a little tweak, it's letting off a little steam or something?
HOTOVEC-ELLIS: Yeah. It's very difficult sometimes because they can kind of trick you into thinking, oh, hey, this pattern looks very similar to a previous eruption at this volcano or a similar eruption at a different volcano and then nothing happens. There's a lot that we still don't know about volcanoes and the kind of seismicity that they make. So we're getting there though.
FLATOW: Yeah. And I imagine you can - as you say, they're all around the world. Do people exchange data and sort of become like doctors who sort of diagnose what a heart's sound is like that gets you in trouble? Maybe this is what a sound of a - of an erupting volcano is going to sound like.
HOTOVEC-ELLIS: Yeah, absolutely. And each volcano is different. It has its own kind of personality, almost, and different challenges that come along with it. Conferences are really good for networking with other volcanologists and seismologists and getting their ideas on what they think is going on.
FLATOW: Do you have to actually go inside the volcano and put the metering equipment in there?
HOTOVEC-ELLIS: Thankfully, for most of the seismic stuff, you can get away with just putting it on the flank, on the outside of the volcano.
HOTOVEC-ELLIS: But the closer you can get to where the action's happening, the better, but then you run the risk of, well, your seismometers getting destroyed.
FLATOW: Forget the grad student that has to go in there and put it down.
HOTOVEC-ELLIS: Yeah, pretty much.
FLATOW: So where do you go from here? What's on your agenda for moving this forward or advancing it? Are you just happy with collecting and making some more screams for us?
HOTOVEC-ELLIS: Well, I am happy to do more observations. I think that that's kind of where my heart is. I know that there's going to be some more work on trying to model and explain what's going on with the physics. Me personally, I've moved on to looking at repeating earthquakes under Mount St. Helens for my project for the rest of my thesis.
FLATOW: Mm-hmm. All right. We're going to give everybody who heard it - now they're - now we've got their interest. We're going to play the whole scream all over again. It's got a lot of different parts to it, so have a listen. This is the volcano scream.
(SOUNDBITE OF VOLCANO SCREAM)
FLATOW: Is that actually erupting at the end there?
HOTOVEC-ELLIS: Yes, that's the explosive eruption at the end there.
FLATOW: Mm-hmm. And does that - can it scream without having that eruption?
HOTOVEC-ELLIS: You know, we're not entirely sure. So far, what we've seen is that when it screams, or when the pitch goes up and there's a pause...
HOTOVEC-ELLIS: ...there is an eruption afterward. We need more data to figure that one out for sure.
FLATOW: Always the last line when you talk to a scientist: We need more data. Thank you very much.
HOTOVEC-ELLIS: Sorry. Thank you.
FLATOW: Nothing to be sorry about.
FLATOW: That's the thing that goes on, is that we need more data and we'll find out. And you're obviously going to find out, right?
FLATOW: Yeah. All right. My guest is Alicia Hotovec-Ellis. She's a doctoral student from the Department of Earth and Space Sciences at the University of Washington and co-author on the paper published in the journal Nature Geoscience in which they're all bunches of screams that were going on there. And we're going to take a break, and we're going to - when we come back, we're going to have more natural sounds.
There are plenty of natural sounds going on in nature. So you don't want to listen to a volcano? We've got sounds that are going to come from the rain forest in Costa Rica. We've got sounds coming from Puerto Rico. If you like natural sounds, that's what we've got for you this hour. So stay with us. We'll be right back after this break.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.