A virtual walk through Kīlauea Volcano’s summit history: Part 2

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Join USGS Hawaiian Volcano Observatory scientist emeritus Don Swanson on a virtual walk, during which you learn about the past 500 years of Kīlauea Volcano’s history as revealed by rocks, craters, and cracks.

This virtual walk will be released in three parts, covering different sections of the Keanakākoʻi Crater trail. Along the walk, Don points out and explains some of the features that formed during the 2018 summit collapse events, as well as the best publicly accessible display of explosive deposits erupted from Kīlauea around 230–370 years ago, one of which probably relates to an important oral tradition. Don also shows two contrasting vents for the July 1974 eruption, highlights the thick deposit of pumice and scoria erupted in 1959, and ponders the origin of Keanakākoʻi Crater.

You can visit the Hawaiʻi Volcanoes National Park website to learn about walking the 2-mile round-trip Keanakākoʻi Crater trail, which begins at the Devastation Trail parking lot on Crater Rim Drive in Hawaiʻi Volcanoes National Park.


Date Taken:

Length: 00:38:14

Location Taken: HI, US



We're now at the best vantage point that you can get to look into the collapsed part of the topographic caldera. What you're looking down to, right in front of us with the big cliffs on it, is the edge of one of the blocks that dropped down [in 2018]. It's called block three or the downdropped block it's what it's usually called. If you trace over to the left edge of it, you go down into the deeper part of the collapsed area. That is Halema‘uma‘u, and where the current lake or pond is forming and the downdropped block—or block three—dropped down about 120 meters (400 feet) or so during the course of the collapse in 2018. And that amount is about the same as the height of the wall beyond—the caldera wall that wasn't greatly damaged in 2018. So, both walls are 120 meters, roughly 400 feet high.


We have before and after LiDAR images of this area. This isn't the way it looked. We’re right here and so we're looking out this direction and this is the same view at the same scale. Again, we’re right here, for comparison. So, a picture is worth 1000 words,right? You can see that the old location of Halema‘uma‘u, shown in this slide, can be superposed on the LiDAR image now and so you see that the main center of downdrop, where the lake is present today, coincides pretty much with the old boundary of Halema‘uma‘u.


This [2018] was quite an amazing time to be here. And I want to go through some of the things that happened in just a sketchy fashion to relate what was happening up here at the summit beginning in early May of 2018 with what was happening down on the lower East Rift Zone in lower Puna at the same time.


Now, this is just a schematic cross section. This is the summit of the volcano—the lava lake was here, there's some sort of magma reservoir down here. Pu‘u ‘Ō‘ō was erupting down here and it had been erupting since 1983, actually, and some sort of barrier/dam/whatever failed and magma began to enter the East Rift Zone beyond Pu‘u ‘Ō‘ō.


Then, the earthquakes and deformation showed that the rift zone downrift of Pu‘u ‘Ō‘ō was breaking up. On May 3, the eruption took place in Leilani Estates down the East Rift Zone. At the same time, the lava lake at the summit was going down. It was going down as magma was being transferred from it into the rift zone and beyond down into lower Puna. So, it's pretty incredible, when you think about it, that events 40–50 kilometers [25–31 miles] apart, can be so intimately integrated with one another. It's almost like an organism, somehow—what is happening in one place on the volcano is being felt somewhere else.


This eruption and collapse of the summit area—the summit area began to drop down as a result of the emptying of this reservoir—there wasn't enough pressure in the reservoir to maintain the summit, so it started going down and down. Those events continued until early August of 2018 before things died away.


These are two cross sections, east-west and north-south cross sections across the caldera showing the elevations before and after the collapse. The scale is over here. And so the amount of down-drop was something like 500 meters [550 yards] or so, in Halema‘uma‘u itself. We have the Empire State Building here for scale, just to make it in more human terms. So what you're seeing is the first major collapse event at the summit in at least 200 years.


But even at that it was small by comparison with past collapses I'll get to in a bit. Now, before the collapse began, there was a lava lake that was present in Halema‘uma‘u. And we were able, —it’s elevation was close to the surface—we were able to measure its elevation routinely. The elevation is shown here in this plot in the red dots. It held steady until about May 2 or third and then it started to go down, down, down, down. We lost view of it about May 9, somewhere in here, when the last observation was made and then lava presumably continued to drain away.


The first cracks on the caldera floor were recognized on May 14 in the old parking lot for Halema‘uma‘u, as well as elsewhere along the Chain of Craters Road. One thing that we were worried about at the time was that if the lava lake drains to below the water table, then would there be steam-driven explosions beginning?


This was a story that was generally accepted for what happened in 1924—the most recent period of significant explosive activity—when there were explosions, more than 50 of them, over a two and a half week period, each one thought to have been triggered by steam from groundwater as the lake-level dropped down below the water table. The water table is the elevation below which rocks are saturated with water. So if you punch a hole through the water table, other things being equal, then water will pour into that hole. And that's what was thought was happening in 1924.


Well, we had explosions in 2018, but we don't believe the evidence favors the involvement of groundwater—steam. We think instead it's more likely that the propellant for the explosions was caused by gas that was coming off the magma, which was dropping down lower and lower but it was still continuing to outgas. Gas was coming up, and then it would be pressurized now and then by rock falls or other things happening that I'll mention in a bit. It was, then, just the pressurization of this magmatic gas, gas coming out of the magma that drove the explosions, rather than groundwater coming into the system.


Now this is still a very open question. Needless to say as a result it’s controversial as well. And we really need to understand this process a lot better, to evaluate the kinds of things that can happen the next time that there is the potential for explosive eruptions at the summit. But right now, that's the way it looks. So, bottom line is that—to put all this thing in a synopsis—a barrier in the East Rift Zone broke, magma was able to intrude beyond Pu‘u ‘Ō‘ō and started the eruption down in lower Puna. This magma was milking the summit reservoir, which was removing pressure then on the summit, which was collapsing and down dropped to as much as 500 meters [550 yards] during three months. So quite remarkable series of events.


Now, this plot that you'll have to see in close-up shows a couple of things that we learned by studying the collapse events as they were occurring in 2018. Let's look at only the upper plot here first. This is a plot that's derived from GPS [Global Positioning System] measurements at the station called CALS, which was located right out here on the floor of this down-dropped block. And you can see that at the start of the plot on June 26 until about a month or so—a little longer than a month—the floor of the down-dropped block, as measured by GPS, was going down. That, of course, is consistent with the collapse of the summit area. But if you look in detail at this collapse, you'll see that it occurs in increments.


There would be a slight dropping down. And then , whup, it down went about two and a half to two and a half meters [two and a half yards]. And then it would slowly go down again for another 20 to 40 hours, then whup, down it went—this downdrop was about two and a half meters or so, something on the order of 8–9 feet or so. But the GPS instrument receiver which was sitting on a tripod on the floor remained horizontal. It was continuing to get signals from the satellites just fine throughout this entire period. So it was like it was in an elevator, just going down and down and down. I would have given my eye teeth to have been in there during one of those events—it would have been really spectacular. You probably would have felt this emptiness in your stomach.


Each one of these down drops was accompanied by an earthquake, mostly around a magnitude 5.3. There were some questions about how you compute that magnitude for these earthquakes, but it's something on that order, 5.3.


Now, let's take a look now at the lower plot. That's a plot of tilt—the change in the slope of the ground surface. It isn't going up and down, it's tilting, at Uēkahuna Vault, which is just across the street from where the observatory, HVO, used to be at the summit of the volcano. The scale is much more sensitive than a GPS scale. So we're looking in terms of microradians—a microradian is one part per million.  If you put a penny under one end of a rigid kilometer-long [just over half a mile] bar, that bar is tilted one microradian. So we're seeing tilts of 10s of micro radians during that time. The interesting thing is that the tilting over on the west side of the caldera, at the vault, was going down, down, and then it would jump up and then it would go down as the summit was subsiding, and then would jump up. Down, up, and these ups were related to the down drops out here in the caldera floor.


So this was puzzling—why is the tiltmeter showing inflation at the time that the floor of the down-dropped block is going down two and a half meters [2.5 yards], what's going on? Well, modeling suggests that there is a reason for that, a rational reason for that, if you put it that way. As magma was being withdrawn from the summit magma reservoir, the summit area was dropping down and the floor of the caldera or down-dropped block was slowly sagging down. It wasn't faulting down it was, it was just developing a bow.


Eventually, the reservoir was empty to the point where the bow could no longer maintain itself and it slipped. The fault slipped, and the floor went down. When the floor went down, it pressurized the magma directly beneath the floor and that caused a brief but rapid inflation of the summit. So you had slow collapse, then faulting, down-dropping. Then, brief but significant inflation at the summit, and then the process continued again. So it's really quite, it was very confusing to us at first but the modeling that's been done by Paul Segall at Stanford [University], gives a good explanation for this process that was occurring.


Another thing about these earthquakes and these two and a half meter [two and a half yard] collapses of the down-dropped block is that you could sort of anticipate when they were going to be occurring. This is a plot showing the number of earthquakes per day before several of these collapse events. You can see that there was frequently a build-up in the number of earthquakes up to the time of the collapse and the 5.3. This wasn't perfect, but for residents of Volcano, they got to be pretty good about this and they got to have a hunch when the next big earthquake was going to occur, the 5.3.


Luckily it didn’t do much or a terrific amount of damage within the housing areas, or Volcano Village. The other residential areas within National Park did quite a bit of damage—quakes disrupting the water supply system and causing other damage.


Since 2018, I've had two different volcano residents come up to me talking about the earthquakes that they would feel, the 5.3s. They say, you know, I kind of miss those damn things. They got to be that that predictable, they realized they thought they knew that it wasn't going to be a magnitude seven, it was going to still be a magnitude 5.3.


Before we leave here, I want to point out that on the ground there are all these big rocks sitting around. In fact, some of them have been used to hold the stands here for the barrier. Those rocks are thrown out by eruptions in the late 1700s from the Halema‘uma‘u area. They are ballistic blocks thrown here in that cannon-like fashion. Not only would it have not been a good place to be when they were falling, but I think they might have archeological significance that I'll mention when we talk about Keanakāko’i Crater at the end of the walk today.


Now as we walk along here we can see some explosion deposits that are below the 1959 tephra on the surface. There's a really nice exposure over here in the sun. We'll see more of this a little bit later, but this ledge here—this light-color ledge that’s just behind the reduced speed ahead sign—that is the footprints ash. I'll talk more about that down here a ways but it's a famous ash deposit because there are human footprints embedded in it in places. Hundreds or thousands of footprints. That was erupted in 1790.


We're looking here at a feature called a spatter rampart, or maybe ramparts for two, that formed during low lava-fountaining in July of 1974. Clearly the lava flow or spatter crossed the road and had to be cleared off. Rampart is a military term. It was for a linear raised feature—“Oe’r the ramparts we watched”—and it's been adopted by volcanologists to describe these linear ridges along fissures.


The gully here is a fissure, out of which lava erupted. The spatter fell on either side of the fissure and built those two ramparts. Commonly, you will see that there's only a rampart on one side of the fissure. We don't see that here but if you go to Maunaulu, you will see the rampart is only on one side of the fissure. That's because fissures commonly cut across sloping ground and so spatter that falls on the uphill side of the ground of sticks, but if it falls on the downhill side, it falls onto lava that is flowing downslope from the fissure so it's carried away. So you only have one side of the rampart preserved. But here, this was pretty flat ground, and the ramparts are on both sides of the fissure.


Sometimes people here can look in the gullies and find Pele's hair. The hair was erupted during the lava lake activity in Halema‘uma‘u mainly in 2014 to 2018, but I think the hair is getting kind of picked or blown away now and you may or may not be able to see it here.


Now, Keanakāko’i Crater is across the road we're going to hit that and that'll be our last stop on the way back, but we'll bypass it for now. And we'll go down here to take another look into the caldera.


It's nice—you find the 2018 ash and then separating it from 1924 ash is this layer of Pele’s hair. It’s really a nice marker.


Well this is as close as we can get to the caldera floor. We can’t see Halema‘uma‘u from here but you can see about where the bottom would be. It would be about where that hump there, that spatter rampart in the distance, just to the right of that small tree. You can also see down-dropped block five I think it is—anyway the one with the remnant of Crater Rim Drive on it. You can see that showing up very nicely now. In the future, if this is still preserved, and you come out here, you will know that the collapse of the summit area—you might not know the year—but you will know that it postdates the formation of that road.


Here, the white cliff we see out there is something that we couldn't see before 2018, but people who visited the summit in the first half of the 19th century could see it. They called it south or southern sulfur bank. In contrast to the north or northern sulfur bank, which is the one that public can go to today. This is an area that is subject to a lot of intense alteration by hot water and gas.  I think once it is sampled and looked at the geochemically, it’ll make a very interesting comparison to the north sulfur bank because this could be different.


We're right at the point about where the east rift connector, the East Rift Zone, leaves the summit area  and heads down past Keanakāko’i and on into the rift zone itself. This could be a hotter area, or it could be different chemistry taking place here than is taking place up at the north sulfur bank. I don't think it's been sampled yet, it's this kind of a difficult place to get to, but it would be something that would be well worth doing.


Now if you look on the far side of the wall facing us, the far side of the caldera, you see a couple of gullies that are in shade.Believe it or not, those gullies have been eroded by water flowing out, maybe at times gushing out,

of a perched water table that is in that area. They first became evident on the Fourth of July in 2018. And they've come and gone periodically since that time. I noticed, I was the first one to notice it I think, and I called it the black streak. That's still the slang term that I use in my thinking—the black streak. What was happening—and again this will be probably not be possible to see from here—but during very heavy rain and I mean, heavy rain—once in five year rains at the summit of Kīlauea—the water can't soak into the ground and so a river, that I call Kīlauea river informally, formed over here below the Crater Rim Drive and there are big sand flats down there that are deposited by Kīlauea river.


I've been down there twice during these heavy storms and when the river was flooding I went across it once, and it's only a few centimeters [few inches] deep. You know it's not a river in the real sense, but it’s flowing water, but then as you went to the southwest, you lost it. The water disappeared underground somewhere. I think what was happening is that the water was sinking underground and then being impounded by dikes—vertical sheets of solidified magma that enter the Southwest Rift Zone in that area. These things came up from below and formed dams or barriers; the water wasn't getting beyond them. So there was a perched water table that was being built behind the dikes fed by Kīlauea river.


Well then in 2018, when the collapse occurred, the wall was faulted back and exposing this water table, perched water table, and water then could gush out of the water table periodically. That's the only explanation that I can think of, but whatever the case there was water there. The reason I called it a black streak at first was that from the Volcano House, which is where I was looking, I couldn't tell if it was water or not. The other competing hypothesis was that a narrow rockfall had occurred that had just,er, had shaken loose from the rolling rocks the dust coating that was making everything else look gray, so you'd had a black streak of fallen rocks contrasting with the light gray. Later on, we could show that it was water because the streak would have remained black, even when a lot of dust was falling out of the air.


Now, there's a lot of interesting geology that can be learned by looking at the rocks in the caldera wall over there where a black streak is and farther to the right from there, as well as looking at the faults bounding the down-dopped blocks part of the caldera. I won't go into that except to say that the 2018 collapse has afforded us an opportunity to look farther back in time at Kīlauea’s summit than we've ever been able to do before.


We'll be able to eventually look not only at rocks that partly filled the caldera during the 19th century, but we'll be able to look at rocks that are older than probably 2200 years in the caldera wall. We've never had an opportunity to do that before, so this is a wonderful research effort that will be undertaken to investigate these. This is important because we want to know what is happening geochemically over time at the volcano, to see if we can anticipate when we'll be changing from periods of dominantly effusive, or lava flow activity, into a period dominated by explosive activity. We've had two of those in the past and we want to know more about these events, so only by looking in detail at these older rocks, we'll be able to see these things together. So this is a research effort that has real practical implications because wouldn't it be wonderful if we could see changes in lava flow compositions leading up to an explosive period; then we can look at and foretell the next change to an explosive period. It would have tremendous social benefits.


One more thing here. Unfortunately, we can't see this, but we can see the vent. You see, if you look over here, we're looking at a spatter rampart again from the July 1974 eruption. Now, lava was erupted from the fissure, which is just this side of the rampart. It flowed down a gully, and you can sort of see the wall going down here, went down to the left and then made a right-angle turn. Where the lava made a right angle turn, it rode up on the outer bank of the turn. And it is possible to calculate the velocity of flow of the lava from the height of the run up on to the outer bend of the gully.


You don't need to know anything about the rheology of the lava it turns out—you don't need to know what viscosity, the crystal-content, its temperature, its gas content. You only need to know how far it rode up above a projected level beneath the run up. This is something that fluvial geomorphologists, people who studied rivers, figured out years ago when they're computing run ups—the speed of lava flows and debris flows. A PhD student at University of Hawai‘i did the calculations and he found that the lava was flowing about 30 kilometers [18 miles] an hour. Now that' faster than Usain Bolt could run at his prime, or for very long anyway.


So we thought this was really wonderful. But then, in 2018, when lava was erupting from fissure eight in lower Puna, we found that there was sustained velocities that were 5–6 kilometers [3–4 miles] per hour faster, 35–36 kilometers [21–22 miles] per hour. So, those observations, combined with this one shed a new light in my way of thinking on the hazards presented by lava flows. You can't always assume that they're going to be trundling along at a manageable pace. There can be and have been situations where the velocity could far exceed the rate at which a person could move out of the way.