What can lava tell us? Deciphering Kīlauea’s 2018 eruption

Video Transcript
Download Video
Right-click and save to download

Detailed Description

The 2018 eruption on Kīlauea’s lower East Rift Zone spewed around a billion cubic yards of lava into Puna. From the moment the eruption began, samples of lava were collected and rapidly analyzed by a team of USGS Hawaiian Volcano Observatory and University of Hawaiʻi at Hilo scientists. Geologists Cheryl Gansecki (UH-Hilo) and Lopaka Lee (USGS-HVO) discuss how their work revealed the complex story of magma sources, both old and new, inside Kīlauea Volcano, and how certain chemical elements can provide insights useful for monitoring eruptive behavior in this Volcano Awareness Month video. Their talk was presented at UH-Hilo on January 16, 2020. Volcano Awareness Month is spearheaded by the USGS–Hawaiian Volcano Observatory, in cooperation with Hawai‘i Volcanoes National Park, the University of Hawai‘i at Hilo, and Hawai‘i County Civil Defense, and provides informative and engaging public programs about the science and hazards of Hawaiian volcanoes. USGS photo caption: Lava samples collected during the 2018 Kīlauea lower East Rift Zone eruption were organized for laboratory analyses at UH-Hilo. Labels on the bags indicate where and when the samples were collected.


Date Taken:

Length: 00:51:31

Location Taken: HI, US

Video Credits

Video production: Katherine Mulliken, Geologist, Hawaiian Volcano Observatory, kmulliken@usgs.gov


[Intro:  Tina Neal, USGS Hawaiian Volcano Observatory Scientist-in-Charge]

Tonight, we are fortunate to have two HVO scientists and affiliates, Lopaka Lee—geophysicist geochemist, IT specialist, many things—and Dr. Cheryl Gansecki, UH-Hilo Geology Department [applause]—a home-town favorite.  Lopaka and Cheryl together formed the tip of the spear, so to speak, in 2018, looking at the geochemistry and petrology of the lavas that were erupted in Leilani Estates and throughout the entire eruption. So tonight, the two of them, in tandem, will present a talk titled: “What can lava tell us? Deciphering Kīlauea’s 2018 eruption through chemistry.”

[Speaker: Lopaka Lee, USGS Hawaiian Volcano Observatory]
Aloha kakou. Nice crowd. I'm kind of surprised, because Janet and I have gone back and forth about the title, and we were convinced that if we included the word chemistry, it would cut it in half—a lot of people wouldn’t show up if they failed chemistry.

I thought what we’d do, Cheryl and I, is break up the talk. I’ll talk more about the background, petrology and geochemistry, and also the project that we have between UHH and HVO. Then I’ll let her talk about all the hard stuff. She’ll talk about the data.


I'll just stand up here and talk a little bit about geochemistry and petrology. Geochemistry is the study of the chemistry of Earth materials—rocks, soils, water, sediment. Petrology is a subdiscipline of geochemistry that specifically studies how rocks form.

Our work during the eruption led to a lot of great science. There's still publications coming out about it. Just a month ago, three papers were published in a bundle within ‘Science,’ a very prestigious magazine. These articles, or these publications, highlighted the great work HVO does. It highlighted a lot of the great science that we do in collaboration with universities, and it was just a really neat thing to see. Cheryl was lead author on a paper in that bundle.

The three papers that came out, there are three different topics… one was on magma reservoir failure, which detailed how the caldera collapsed and its effects downrift. The first one was Kyle Anderson as first author, and the second by Matt Patrick at HVO was on cycles of lava effusion, which had to do with pulsing of lava down the east rift that happened as a result of things collapsing at the summit, and also degassing, which was causing cycles of the lava coming out, or effusing.

Then, of course, Cheryl's paper, which she’s the lead author of, on magma mixing in near real time.

That part of it, that near real-time part of it, is what the journal found really novel or unique. I'll talk a little bit about that. Why doing it in near real-time, talking about the magma, characterizing the petrology and geochemistry in near real-time was so unique and different for this eruption.

In order to understand that, you have to understand what is volcano monitoring? What is volcano response when an eruption happens?  At HVO, we're divided into subdisciplines, and we're all focused on monitoring the volcano and responding to eruptions. But our primary tools are really, largely geophysical tools.

We measure ground deformation, the way the surface deforms, which is the result of magmatic activity or ground motion. Our primary instruments are GPS, tiltmeters. We monitor seismic activity using seismometers—earthquakes. We also do gas geochemistry as well. That measures gas as it's emitted from a volcano and tells us a little bit about where magma is relative to the surface. All these things have been a part of the normal monitoring routine that we use at HVO.

Generally, the geophysics-based methods, like ground deformation and seismic activity or seismic data from seismometers, come in as real-time streams. Once we set up the instrument, these data streams are coming in. We’re able to sit at the observatory and watch them in almost near real-time. Gas emissions, too, generate some fairly consistent regular data, which we also import into the system and watch that.


Geological observation is also a key part of that. I like this diagram here… if you're geologists, you get to ride around the helicopter. [laughter] Geologists get all the all the glamorous jobs, riding around in a helicopter. In this diagram, at least, petrology and geochemistry of the rocks is not explicitly stated in there, of course, because, as you can imagine, it's not necessarily the same. It's a little more difficult to fit into that framework.


The reason is that studying lava chemistry is typically a much slower process, and it's usually a retrospective tool. Usually people use it to study eruptions that have already happened, or even as they've happened you won't look at the data or publish the data until much after. It has to do with the fact that it's a difficult thing to do. It’s a difficult thing to do well, it's a difficult thing to do quickly. This image here shows cutting rocks. You’ve got to cut rocks in order to look at them under a microscope and characterize minerals that are in there.


How many people have actually made a thin section? Right. It is hard. You have to cut wafer-thin slices of rock. Then you polish them down, and keep polishing them until they are paper thin, and they have to be a particular thickness, not just any thickness. Once you do that, you’re not even done, you’ve just started. You put it on a microscope to so you can see the minerals and you can start describing them. So, it's very laborious.

Doing the routine chemistry of the rock is also something that's really laborious. Usually, most of the methods you can't just pop a rock in, with some exceptions that I’ll describe. But, generally you can't just pop rocks in there and get the data back. You've got to prepare them somehow, that’s what this technician is doing here, he’s preparing a sample for some kind of analysis. That involves time. Then the biggest thing is that the instruments you use, they're not just expensive, they are exorbitantly expensive. We’re talking about multimillion-dollar machines. So, they are housed in labs very far away, centralized, so that everybody can use them. That's the other problem—you're not the only one; everybody else is standing in line to use these instruments. So, when you come to them and say we’ve got an eruption and need our samples now. They’ll say, “well, yeah, get in the back of the line.” That makes it hard, as well, to try and buy geochemistry and petrologic information in a rapid manner, particularly when an eruption is occurring, and events are unfolding. That's why it was so novel, what we did.

But, even though… I was focusing on the fact that it's not a rapid technique, but it's a solid science. It's told us a lot. Petrology and geochemistry have told us an immense amount of information about our world around us, and some of it is just absolutely key to understanding about the way certain earth surface processes and earth processes work. Of course, with the Hawaiian Islands, much of our understanding comes from the chemistry of rocks. Everything from the nature the hotspot and its composition, also the ages of the islands, we’ve determined through chemical analysis of rocks.


So, how do we do that? I'm going to attempt to describe in a very general level how we do that. Normally, you can’t do it in a single talk; I'm going to do it in a single slide—or attempt to.

We all know that magma is liquid rock. As that liquid rock pools, important major things happen. You can tell a lot about the history of a rock after it's cooled by looking at its chemistry after it's gone through that cooling process.


Here, we have a magma of composition A. Imagine this thing being underground. As it cools, minerals start to precipitate. There are certain minerals that precipitate first, one of the primary ones that we're probably all familiar with is olivine. If you go out on a beach, you see green olivine specks there. That's one of the first minerals to precipitate out first. When that forms, it actually captures the chemistry of the liquid that it's around, because it's forming in that liquid, that magma. So, if we look at the chemistry of that olivine, and we look at specific elements in there—for example, magnesium and iron, two primary elements in olivine. If we take a look at the ratio of those two [elements], you can tell something specifically about the temperature under which that olivine formed.

So this process goes on, and additional minerals will form as the as the magma cools. If it’s left to cool all by itself underground, eventually, you end up with this interlocking mesh of crystals with very little liquid left in there. And a lot of the composition of that magma, the different parts of its chemistry, the different elements, will have gone into different minerals, because each mineral has its own composition, and it wants something from the liquid. So, if you measure again, the composition of the mineral, composition of liquid, you can infer something about the way it formed.


Within the context of this talk, some of the most important pieces of information that we use, or derive, from this, is temperature of the magma—when it was cooled, or quenched—also sources using chemical fingerprint, and a little bit about the history, what happened to it before it got to the state it’s in.

It turns out that we've been actually doing this rather routinely on a regular basis for a long time at HVO. Before Pu‘u ‘Ō‘ō, it was a normal part of just responding to an eruption, that is describe it in all forms including geochemistry and petrology. But Pu‘u ‘Ō‘ō really changed things, and that was a continuous eruption, 35 years of continuous eruption. During that time, for long stretches, it was like a walk-up volcano, it was walk-up sampling. Ken Hon, sitting back there, likes to joke about the Kupaianaha days, when there was a lava lake and he was saying yea you could just walk up to the thing and just throw a hammer in on a chain, and there you go. Then eat your sandwich.  [laughter; inaudible audience chatter]  Well, it's a lot harder than it looks, actually.

But, we still have easy access to an active volcano. And that set the stage for continuous monitoring of lava chemistry that was done for all of the eruption, all 35 years, it was done. This this involved multiple generations of geologists. This included Ken and Cheryl, and Tina, were part of that effort. Looking at the database, your name is under a lot of samples in there. Many others—students, collaborators who’ve come to help us—and generated an enormous database, which we’re still interpreting today.

That happened, but we were still, as of late, within the last maybe 10-15 years there was a desire to turn around quick, and particularly within the last 6 years, we really started getting focused about, how can we do this faster? What can we make fast? How quick can we get?


We started that discussion with UH-Hilo, because—this is one of my favorite sayings about instrumentation—if you've got a hammer, everything becomes a nail. Meaning that you just use that hammer and just bash everything into submission. No, actually, it's kind of a double-edged thing. If you have these resources, you really should be using them, you should be trying to figure out the best use of those instruments.

We have this collaboration—we have a long history of collaboration with UH-Hilo. But when we started thinking about this, there was a couple of things that happened. One is that there was an instrumentation that was available, there were people that we were working with. Then there was new instrumentation that was available to make it easier for us to analyze things.


One of the key things was this thing called an EDXRF, energy dispersive X-ray fluorescence instrument. This is Professors Peter Mills and Steve Lundblad, he’s back there. Ken, you're a part of this EDXRF acquisition as well, too. What they were doing was—it's getting back to this hammer analogy—they were using this thing to analyze hammers, or stone adzes. The reason why they use this is because it was quick and non-destructive. So, they could take adzes, stone artifacts, quickly put them in this thing, analyze the chemistry, and from that, they were able to discern where these different materials were coming from, where these different rocks were from. And that, I think, planted the seed—it was like, hey, if they can do this, why can’t we do this with lava samples? That really set the stage for that.

We started formally in 2012. Formal beginning of the HVO-UH petrology project. The routine involved HVO being responsible for obtaining samples. We bring the samples down to UH-Hilo, where we go through this routine of analysis that we developed over time, over 6 years working on this, actually. We developed this and refined it to get us back to chemical formation. We did it in really 2... oh … this is a nice break. Before I describe how we did it, I'll talk about lava sampling. Remember Ken Hon said it was super easy.

This is Professor Bruce Houghton, at UH-Mānoa, and we put him up to this. We said Bruce, go get us a sample there. Here's a shovel. All you gotta do is just take a quick shovel and put it in the bucket, it’s no problem. And he actually does a great job.

[video of Bruce collecting lava sample] Right now, his face is burning.


We quench it really quick. We take a sample and put it in water really quick to freeze it immediately, so it doesn't change anymore.

Our routine involves two basic paths, called routines, that I'll call “fast twitch” and “slow twitch” routines. The fast twitch routine involves using the EDXRF, so we get the sample, get it to UH as quickly as possible, as regularly as possible. We powder the sample, and then we put it on the EDXRF. From that, we get a suite of elements that are able to tell us an awful lot about the rocks.

With EDXRF, you can’t get everything you need out of that instrument, but you can get enough, as Cheryl will show. During 2018, we surprised ourselves with the kind of information we could get out of it.

The other part was micro-analysis routine, which is more of a slow twitch routine. This is more of a traditional type of petrology routine. We were picking individual minerals out of samples to characterize it and glass. We mounted up these grain mounts—and we had some innovations on that too, 3D printing of rounds and things to speed it up and pack as many as you can. We put those rounds into epoxy, polished them all, and then we put them on the SEM (scanning electron microscope) to look at them in detail. That’s also available at UH-Hilo as well.


Then we select what points we want to probe and get the chemistry and send that off to a really expensive microprobe in Denver. Then the data will come back. That was the sort of slow twitch.

Obviously, you can’t do this without help, and we’ve had a lot of help. Students who’ve helped over time, and not just these students. People have been sort of sucked into the vortex. If you're standing around, you're gonna help, particularly during the 2018 eruption.


[audience question about dog photo]


Derek was from Louisiana, and I didn't have a photo. Miki’s here, as well, but I spared her the indignity of a real photo. Lil DeSmither is now with us at the observatory, and Brenna is here on island—she's right there—good to see you. Valerie is now at Fairbanks, Alaska, studying volcanology and Ryan is off doing mathematics. Thank you for being a part of the projects.

Now, after that background, I’ll hand it off to Cheryl.


[Speaker:  Cheryl Gansecki, University of Hawaiʻi at Hilo]
I'm going to segue into 2018, what happened and what we ended up doing, and what could the lava tell us, it turned out in the end, during this eruption. Samples and methods, you heard mostly about that.


In this particular eruption, we collected, or HVO and all the field crews collected, 113 lava flow and spatter samples during the eruption. We analyzed them during the eruption as soon as we could. There's Bruce again, collecting spatter samples, and Katie collecting lava. There's Miki in the corner on the EDXRF. It turns out there are certain key elements I'm going to talk about. I apologize in advance for all the graphs, but they’ll all link together and make sense, hopefully, for you. There’s five elements up here that I'm going to talk about, and there's basically two categories of them. We've got potassium, titanium and zirconium there as one group; those we call incompatible. We've got another group here, magnesium and calcium, that we call compatible. These are elements we can get with the EDXRF, we can get pretty good numbers on. Magnesium is a little messy, but it's so important, we can still get good enough numbers with it. Those two categories tell us something a little bit different.


Let me try to explain the idea of incompatible versus compatible. Lopaka talked to you about the how magmas evolve over time, and the crystals grow and it changes the chemistry.

I like to think of magma like a bag of trail mix, where it is basically a big mix of elements mixed together. If you think of your average bag of trail mix, it's mostly peanuts and raisins and it's got some good stuff, some M&M’s® and some cashews. Some trail mixes might have something weird, maybe you got broccoli chips or something, that nobody’s going to want to eat. Our bag’s got a little bit of everything.


Now picture our magma sitting there, beginning to crystallize, things are being removed. We hand our bag of trail mix over to our friend, olivine. Olivine loves M&M’s® and starts eating all the M&M’s®.  Might eat some of the peanuts and other stuff, too, but eats up a lot of your M&M’s®, then passes it on to some of the other crystal friends. They like M&M’s® too, maybe some cashews, and they're eating a little bit of other stuff.


But by the time you get the bag back, there's a lot less material, and it's really concentrated in peanuts and raisins.  All your M&M’s® are gone, and nobody touched the broccoli chips. In our analogy,  zirconium is our broccoli chips. Nobody wants it; it just gets left behind. Magnesium and calcium, that's our M&M’s® and our cashews. Crystals love them and want to eat them up; they go away pretty fast. So, those two guys are kind of two different fingerprints, and they're going to tell us different things about our magma as we go.


We'll start with the beginning of eruption, May 3, 2018. Fissure 1 rips open through Leilani Estates, and the lava that comes out, they're not very big eruptions. Certainly destructive, but they're not very big, and the lava doesn't travel that far.


I'm going to show you a series of maps as this eruption evolves. Here's Leilani Estates, in yellow here is the lava comes out during the first week of the eruption, that we're calling early phase 1. This is the initial 15 fissures. So, 15 different cracks open up all in line, back and forth, and the lava that comes out is kind of cool, sticky… little bit of nice explosive fountains. One bigger flow comes off fissure 8, then stops.


So, what does the chemistry look like? Don’t panic, they're just plots. The top one here, we'll start with this one, is our zirconium, or broccoli chips, versus our M&M’s®, the magnesium. This is little gray zone right here, this is the last 10 years that have erupted out of Pu‘u ‘Ō‘ō, to give you a sense of where things are changing. So, this is where Pu‘u ‘Ō‘ō was sitting, at a higher magnesium, pretty low zirconium.


As soon as those first lavas erupted, we got the samples back and stuck them in the machine. I got the first numbers back, and it's like ”yes!” It's exactly what we thought it was going to do—high zirconium, low magnesium. This is not your Pu‘u ‘Ō‘ō magma that's coming up. This is something different. It's what we actually expected. This has happened before down in this area. We’re kind of low on the rift zone. We know old magmas might be down there. They cooled off, they lost their M&M’s® and they end up as this different composition. We saw that right away, and that was exciting.


This figure down here is the temperature that we use calcium to estimate. We can use that element to give us an estimate of the temperature. This is progressive days as it begins, so this is zero at the beginning of the eruption that just goes on up from there. This gray band is the average Pu‘u ‘Ō‘ō temperatures from the last few years. They're up at about 1150 [degrees] Celsius. This initial 15 fissures erupted at about 1110 Celsius. So big temperature drop. Clearly not the same thing.


When magma, or the lava, is cooler like this, it's stiffer. It’s got a higher viscosity; it's not going to flow as easily. It tends to have some repercussions for hazards, in terms of these cold lavas are not going to flow as easily, so they won't move as fast. But they can potentially… really stiff magma can get explosive because gases can't escape, so they’ll build-up pressure. As you get too cool, that can happen.


At this point though, the lavas were just kind of slow, kind of sticky. But we were already pretty sure that what we were looking at was stored magma from previous eruptions. We don't know which previous eruptions, but we know it was stored.


There's a couple of days break in the eruption, and then it restarted on May 12, downrift, to the northeast here, up in this area—this is the green. These black lines are the fissure segments. Again, pretty small, the lava flows [on the map] are probably even hard to see from here. Still pretty sticky lava. This goes through about the 17th—fissure 16, 18, 19, and 20. We’ll get to 17 in a second.

Here's our green, our new lava flows. What we see immediately is there's a change between those initial fissures, and the second group of fissures. Things are definitely hotter—about 20 degrees hotter, on average. Right away we can see that. When we look at the fingerprints, we see it starting to shift more towards Pu‘u ‘Ō‘ō-like compositions. At this point, we knew we're looking at lava that's hotter, that it's likely this old sticky stuff is getting mixed with some hotter stuff coming down the pipeline, and it's starting to shift. When we saw that, we put out a notice on the communications platform that HVO was using for people in the field to note that we'd seen this increase in temperature and, the possibility for things to heat up and get moving a bit faster.

At this point, we were all pretty sure that the thing hadn't really gone as big as it was going to go. We can see that—and that was a neat thing to see.


Kind of starting at the same time, fissure 17 broke open, which is this thing up here. It’s the one fissure that was offset from the rest of the fissure trend out here. We call it an en enchelon step, where it's step sideways. It [fissure 17] erupted something unique—something that we’ve never seen erupt on Kīlauea, or anywhere on this island, before. Very unusual behavior, more explosive. It sort of overlaps the phases of the rest of the eruption, the phase of 1 and 2, but it was its unusual behavior that made us very interested in the chemistry.


And sure enough, when we saw the chemistry—Miki brought me the data, and I thought [she] did it wrong…go back and do it again… that can’t be right, because we've never had anything that high on this volcano before, up at 500-600 parts per million of zirconium. That's a lot of fractionation.

So, a very unusual kind of thing here. When we look at the temperatures, they’re very low, around 1060 to 1080 [Celsius] down in here. You can see there's quite a range of composition, so it's very interesting. We're still trying to figure this out about this particular section of the eruption, the fissure 17, because there's a wide range of compositions. What that means—there's a lot of questions about that still. But definitely we saw some very evolved stuff that's had everything taken out, but the broccoli chips.


So, what does it mean? It means our lava is very cool, very viscous, and suddenly it makes more sense why it’s so explosive. I'll show you a little bit of video up here, just from this vent because it's so cool. It clearly had a lot of gas in it and may have had some groundwater involved as well, but the fact that it only happened where you have this super sticky, viscous cool lava implies that they're connected.


Here's the main vent, which is erupting that little hotter stuff out of the main fissure 17 vent, and this section of the fisure 17 is where the really explosive, really evolved stuff was coming out. You see a burst come out right in here—these big explosive pulses that really were quite a thing in the field.


[eruption noises – hisses and booms]

Those things are flinging rocks pretty far. Very explosive and very unusual. It’s a type of rock we call ‘andesite’ Most of our lava is basalt. This one was erupting andesite. This is the only one we've ever seen on Kīlauea or Mauna Loa, or any volcano on this island, so it's really quite unique, and a pretty amazing thing. What it means…there's a lot there.


Then we go into phase 2. Somewhere around May 17-18, things really start to heat up, literally, and we end up with this blue section. We have a whole bunch of fissures active at about the same time, volume is increasing, lava flows make it to the ocean by the 19th of May. So, we've got these big flows. We got a bunch of these lava fountains active. Lava is becoming faster-moving, it’s becoming more fluid. There’s a dramatic increase in volume.


When we look at the chemistry now, here’s the blue, it’s approaching Pu‘u ‘Ō‘ō values, although it veers off a little bit there. At this point, we again let people know we've basically reached Pu‘u ‘Ō‘ō temperatures and what does this mean. It means that magma that was traveling through the pipeline has basically reached the surface.


[glitch in PowerPoint]

What happens during that same time is that the geophysical techniques that they're using to monitor the volcano, so the deformation of the ground, for example, the ground outside Leilani… let me go back to the map here…. There's a station on this side of the fissures, and one on this side, and this one was basically moving to the north. So, you could picture this blade of magma forcing its way down, pushing everything aside. What we see is that as thing is moving in, this station is moving to the north.


During the same phase that our chemistry is shifting, right in here, is where that movement stops. So, it's opening, opening, opening, and the earthquakes go, go, go, and then everything sort of stops right about the same time that we see this temperature change. This is a really cool for us as geochemists. We don't usually ever get to see our data going along with the geophysics data at the same time. It's really exciting to see that happen. What that's telling us is the intruding dike is there. It's opened up, and you've got the stuff that was coming down [the rift] pouring out now. We have that correlation going on, which was very exciting for us. So maybe geochemistry can be useful once in a while.

That goes on for a week, and about the 28th of May, things basically shift back to fissure 8,  becomes the main vent, and that big flow comes out. That's when things really get huge, the flow, right there, goes to the north and around Kapoho Cone and takes out the whole coastline for the next two months. There’s the big channel, the big fountains coming out of fissure 8.


Our lava now is in this phase 3. It's this hot fluid lava. So, look at our trend up here now. This is trying to tell us something. We see that it's going towards Pu‘u ‘Ō‘ō and then it veers off to this higher magnesium. Zirconium basically flattens out—we’re done with the broccoli chips. But for the magnesium, there's something going on here. We're getting too much. Temperature-wise, it all levels out, pretty much stays the same magma temperature the rest of the eruption. Notice that it's a little bit lower than Pu‘u ‘Ō‘ō.  The reason for that is that the magma had to travel an extra 10-15 miles down the pipeline to get to the surface, and in that time, it cooled off a little bit. So, we expect it to be a little bit lower, that makes sense.


But the magnesium is trying to tell us something else. There's something in there that's hot—but we'll come back to that. The other thing that happened in here is these little purple dots. These are… While fissure 8 was going—everything is pouring out because it’s very hot, very fluid lava—a couple of these other fissures get reactivated, so they suddenly do little eruptions that spit out a little bit of lava, mostly very small scale stuff.

When we could finally get to them, they're on the other side of the lava river, so we couldn't get to them till afterwards. But when we got some pieces and took it back to the lab, we saw that these don't look anything like fissure 8. These look more like what we were seeing over here. Fissure 22 even looks kind of like the andesite, or at least the hot end of the andesite. That brought up a whole other issue. What is going on with that? Can these fissures actually mix?


This is now two of our different incompatibles, our normally rejected elements. Potassium is usually not taken into any minerals in basalt. Titanium, though, is one that will, for a while, not be taken in, but at some point, it will start to. So, if you have a really hot basalt here, and you let it crystallize, it would normally follow a pattern like this—see the little dashed lines there—it’s going to do something like that. It’s going to get higher and higher in these guys because the minerals are taking everything else out. They're getting concentrated. Then you get to a certain point where a mineral that really likes titanium comes in. It’s going to start eating up titanium, but it doesn't pay any attention to potassium. So that keeps going up, but now titanium drops. I like this diagram because it helps us to distinguish these magma bodies from each other.


These colors are our what we call our end members. Here, this end member of the andesite, very high in potassium and zirconium, but not so much in titanium because it’s already gone past that phase, whereas our high titanium here in the Leilani [lava], the early stuff, is in between. These make separate bodies, but they sit on this trend. That makes sense if you take magma, let it crystallize and take out the crystals, that's the path that's going to follow.


But our little purple things sit over here. They're not on that path at all. They’re telling us a story about mixing of those magmas. Not only do we have the stored magmas, but somehow in the system, they're getting mixed up with each other, and you're getting these mixed chemistries.


This diagram was one of our attempts to try to visualize this. This is like the phase one stages where they're all erupting separately. Then they begin to mix and, while fissure 8 is going down here, a little bit of this andesite seems to be mixing in with some of this other stuff and giving us these in-between eruptions. This little cone down here is fissure 22. This beautiful little cone is doing some kind of little explosive things while fissure 8 was going and erupted this material that was very similar to the andesite. So clearly there's some connection at depth. And that was really cool to see once we could get those samples.


What are those magma bodies? We've got these three end members. This, of course, is whatever is coming from the summit and Pu‘u ‘Ō‘ō. These two guys, though, have to be the stored magma bodies. Where do they come from? There's a lot of possibilities. This is a map of the lower East Rift Zone and light purple are the historic eruptions down there. The last one was in 1960, down in Kapoho and the lighthouse. 1955 had some eruptions off some different parts. Before that, it was a long break to 1840 over on north side. Then 1790 was right in that same area, these eruptions here. So, there's a couple possibilities for parental magmas, things that might have gotten stuck down there and sat there, crystallizing and changing, and turning into possibly one or both of these magma bodies.


This question is still wide open. In fact, there's at least two PhDs working on this now for their PhD projects on different aspects of this. So, stay tuned. Maybe we'll hear more about this. We just modeled the late 1955 as a possible parent; it works just fine for either one.

One of the big questions left on this is, how long can magma sit down there and still be eruptible, still be fluid enough to get back to the surface? How long does it take that fractionation to happen? There’s kind of two extreme models that people have published.


We used to go with the old model saying well it's about one degree per year, in which case none of the 1790 would be eruptible at all. Whereas other people who've looked at the crystals and said, this crystal is 500 years old, so clearly this magma has been sitting around a long time. Somewhere in between there is probably where reality actually sits. We shall stay tuned to hear more about that, hopefully.

Getting back to our secret olivine story. Those olivine crystals floating in there were trying to tell us something. Lopaka mentioned how we pick those grains out, polish them, and send them in a microprobe so you can zap them and get not only their chemistry, but the chemistry of different parts. You get the cores and the rims. This is a plot of, again the days since the eruption, so starting at the beginning, going to the end. This is how much magnesium is in your olivine.


As Lopaka mentioned, that magnesium goes up with the temperature. So high magnesium magma is hot, and you get high magnesium olivine. These right here, in gray, those are the Halema‘uma‘u lava lake samples. Remember we had the lava lake going for 10 years. Every once in a while, we’d get Pele’s hair and stuff wafting to the surface, and we’d find tiny crystals in them. Miki and Brennon and guys would pick them out, put them in epoxy and polish them, and then we’d analyze them. They all kind of came out in the same area, here, right about 80.


Pu‘u ‘Ō‘ō had a little bit more variety—that's these little black triangles. But mostly, this is right before the current eruption, a little bit cooler than Halema‘uma‘u. Makes sense; it's a ways down the rift zone. But they’re all pretty closely matched in chemistry.


The eruption begins down here with the yellow guys. Very cool. Makes sense, right? They’ve been sitting there crystallizing. It's a cooler magma, so you have lower magnesium olivine. Great, that makes sense.


Next batch… they're getting hotter. Already starting to overlap with Pu‘u ‘Ō‘ō. That's interesting. Then we start getting into here.


All of a sudden, about 20 days into the eruption, we find olivine with almost 90% magnesium. You need a magma of about 1300 degrees [Celsius] to get that. So that's telling us something really unusual, too—that this is not Pu‘u ‘Ō‘ō magma. There’s probably some mixed in—you can still see a bunch of stuff at that level. But, this eruption is so disruptive to the system that it's scooping up some hot, deep olivines, mixing them in and carrying them down the pipeline. And they're growing rims—the open symbols here are the rims. When we actually zap the edge of the crystal, versus the center, those tend to be around this line. This is the theoretical line for what they should be if the crystals were growing in the melt that they're in—sort of an equilibrium line. So, the crystal was formed in that liquid, that's what it would look like.


It's telling us that most of these crystals, at least their cores, formed somewhere else. Then as they travel down, they grow a little rim as they’re going, and that's going to match more with your glass. But you’re picking up this really deep, hot stuff. So, we have a whole other thing going on that you couldn't really tell except the magnesium went kind of wacky.


If you put it together in a diagram, sort of theoretical slice through the volcano, we've got our…. This is the summit up here. There's what we picture as the arrangement of magma chambers under the summit. There was a shallow body right under Halema‘uma‘u, then the deep south caldera body. There was a lava lake up there and it was feeding the Pu‘u ‘Ō‘ō eruption, so that system right there was going on happily, for at least the last 10 years, of course, Pu‘u ‘Ō‘ō for 35.


Then something broke out here and stuff started moving. So, the numbers here are different end members or different magma bodies. So, in yellow, that's our Leilani. The dike comes down, pushes the stuff out, starts mixing with it, hits this one, pushes it out, gets mixing in there. It's mixing with the three, which is our summit, Pu‘u ‘Ō‘ō stuff. But somewhere along the way it's scooping up some of four as well.


One of our ideas is that this magma up in here, the shallow magma chamber, was open to the air. That magma was circulating in the lava lake and it cools, it degasses, it gets dense. So, when you start to drain that, it's denser than the lava around it. It's going to sink and potentially stir up deeper levels of your deeper magma body. That's potentially one of the ways we may be getting those hot olivines mixed into the system. There's other possibilities but that's one that seems quite likely.

So, geochemistry telling us a little bit about what's going on. We need the geophysics, but the geochemistry can help too in trying to decipher what's going on. It was exciting for us to actually have this work.  After all these years that these guys worked on it where nothing changed. For five years, we kept getting the same numbers over and over again, and it was hard to prove that this is worth it.

To finally have something happen was very exciting for all of us to see. And to be able to contribute a little bit to the hazards, to the monitoring, to the interpretation of the eruption, that was exciting for all of us. Hopefully, we haven't put you all to sleep yet.


Thank you all very much.