Largest Gravity Changes Ever Recorded: 2018 Kīlauea Eruption

Video Transcript
Download Video
Right-click and save to download

Detailed Description

The largest gravity changes ever recorded: Continuous gravity monitoring of the onset of Kīlauea’s 2018 eruption Talk by Mike Poland–USGS Yellowstone Volcano Observatory Scientist-in-Charge and former USGS Hawaiian Volcano Observatory geophysicist.

Talk originally presented at the American Geophysical Union Fall Meeting 2020. Eruptive activity at Kīlauea Volcano, Hawaiʻi, in April–May 2018 occurred at sites that were well monitored by continuous gravity. Draining of the lava lake from the summit eruptive vent starting May 1, recorded by a gravimeter on the vent rim, was accompanied by a drop of ~1300 microgals over 9 days. At the rim of the Puʻu ʻŌʻō eruptive vent, 20 km [12 miles] down the East Rift Zone from the summit, a gravity decrease of ~200 microgals over 8 minutes, followed by an increase of ~350 microgals over the subsequent 9 minutes, accompanied the formation of an eruptive fissure on the flank of the cone on April 30. About 45 minutes later, a decrease of ~1500 microgals occurred over 2 hours as lava drained from the vent. These gravity changes are the largest ever recorded anywhere in the world.

The evacuation of the summit and Puʻu ʻŌʻō eruptive vents provided opportunities to image the vent geometries, which were used to model the mass changes at the two locales. At the summit, joint modeling of gravity, lava level, and vent geometry indicate a best-fitting density of 1700 kilograms per cubic meter for the lava within the vent. There is no record of lava level over time at Puʻu ʻŌʻō, but the gravity data combined with the vent geometry can be used to reconstruct that process, suggesting that the lava had a density of ~1900 kilograms per cubic meter and that in 2 hours a bulk volume of 11 x 106 cubic meters drained from the cone. The pre-collapse gravity decrease and subsequent increase at Puʻu ʻŌʻō are more difficult to model given the lack of other constraining data. We hypothesize that the gravity fluctuation is due to the emplacement of an eruptive fissure on the west side of the cone immediately prior to the collapse. The gravity decrease represents the opening of a dry crack, and the gravity increase is the subsequent filling of that crack with magma that was denser than the spatter that makes up much of the cone.

These data highlight the importance of continuous gravity for monitoring volcanic activity. Not only do the data provide important constraints on lava density, they can also be used to estimate the rate and volume of lava accumulation or withdrawal and can detect transient eruptive fissures, even in the absence of other observations. Without such data, our knowledge of the processes occurring at Puʻu ʻŌʻō during the crucial opening hours of the eruptive sequence would be as cloudy as the weather during that period.


Date Taken:

Length: 00:14:06

Location Taken: HI, US


The largest gravity changes ever recorded: Continuous gravity monitoring of the onset of Kīlauea’s 2018 eruption

Michael Poland – USGS Yellowstone Volcano Observatory Scientist-in-Charge (former USGS Hawaiian Volcano Observatory geophysicist)

Talk originally presented at the American Geophysical Union Fall Meeting 2020

Hey everybody. Welcome to this brave new virtual world we live in. Well, thanks for tuning in and for your interest in this presentation, and as a reward I'm going to be telling you today about the largest gravity changes ever recorded, at least to our knowledge. These were gravity changes that were associated with the onset and the early evolution of the 2018 eruptive activity at Kīlaueathe lower East Rift Zone eruption and the summit collapse. Before I launch into this, I'd like to make sure I acknowledge my co-authors—that's Daniele Carboni of INGV [National Institute of Geophysics and Volcanology in Italy] and Matt Patrick of the Hawaiian Volcano Observatory—the two of them really made this work possible. So, let's talk about these huge gravity changes.

As most of you probably know, Kīlauea had been erupting from two areas prior to the 2018 eruption. The first of these areas was Pu‘u ‘Ō‘ō on the East Rift Zone. That eruption started in 1983 and went more or less continuously for 35 years. Then, starting in 2008, a summit eruption was ongoing and this was manifested by lava lake activity just on the inner edge of Halema‘uma‘u crater. In 2018, of course, we had the extrusion of over a cubic kilometer [more than 0.2 cubic miles] of lava from the lower East Rift Zone. This was fed by a dike that propagated down from Pu‘u ‘Ō‘ō and eventually connected back up to the summit.

In the early hours of this dike propagation, Pu‘u ‘Ō‘ō collapsed. You can see here; it was characterized by billowing plumes and at the summit the lava lake withdrew. That took a matter of a few days and eventually the summit collapsed to form a new caldera. This was all associated with this dike intrusion down to the lower East Rift Zone.

Gravity monitoring was done by two continuous instruments. The first was located at the summit, in fact just above the summit eruptive vent. This was station HOVL right on the eastern rim of Halema‘uma‘u crater. The second was located right on the north rim of Pu‘u ‘Ō‘ō crater, station PUOC. These two gravimeters had been running for several years at the time of the 2018 collapse.

Now, the installations looked something like this—we had power supplied by solar and batteries, telemetry as well. The instrument was located in this little protective hut—this little doghouse with a tile roof that was meant to protect it from the elements and from any ballistics. If you open up that doghouse, you can see that inside is a LaCoste-Romberg. In both cases, we used g-meters that were fitted with the Aliod Electronic Feedback System. And so two-hertz gravity data and tilt data were streaming from the gravimeter into a data logger. This system remained in place at the summit as is, but at Pu‘u ‘Ō‘ō, we eventually replaced the dog box with a Pelican case with a concrete floor and slightly more advanced electronics that were more robust against failure.

Well, let's start by talking about gravity change at the summit. What you're looking at here in this movie is a time-lapse animation of the thermal camera imagery of the summit lava lake. You can see it slowly receding through the first few days of May as the lower East Rift Zone dike tapped the summit magma system. On the bottom is the gravity change that accompanied the withdraw of the lava lake. Over this few-day period there was over a milligal of gravity change. This is the second largest gravity change ever recorded. Fortunately, during this time the summit went from looking something like this, where Halema‘uma‘u was located in the center of the caldera—the gravimeter was located right on the eastern rim here— to looking like this. This was the ending view after several months of collapse; the gravimeter had been located right in here, and unfortunately in mid-May of 2018 it was consumed by the widening of Halema‘uma‘u as the collapse took place.

After withdrawal of a lava lake, but before collapse of the summit, aerial photos were used to construct a structure from motion DEM [Digital Elevation Model] of the interior of the summit vent and that's what you're seeing here on the left. There's also a plan view of the system with the location of the gravimeter noted by that dot that says HOVL and a cross-section that shows what the vent looked like on a north-to-south profile. We modeled this vent by breaking it into parallel pipettes—these five-by-five meter [16 feet] cells and we could simulate the level of lava over time by matching the level of lava as viewed in that thermal video. By doing this, we could actually simulate the gravity change that would occur as the lava lake withdrew. Since we know the level of the lava over time and the shape of the vent, the only thing that we had to solve for to match the gravity data was the density of the lava lake.

The top plot shows the change in gravity over the first few days of May. The bottom is the lava level during the same time period, the lava level being derived from that series of thermal camera images. Now, by combining the lava level change with that structure from motion model of the interior of the summit vent, we were able to calculate a gravity effect simply by changing the density of the lava. When we do that, the best fitting gravity change is obtained with a density of about 1700 kilograms per cubic meter for the upper 200 meters of the lava lake.

Now, this is a little bit denser than we'd seen before. During a past episode of lava lake withdraw in 2011, we calculated the density of about 1000 kilograms per cubic meter. It's possible that the density of the lava lake had increased over time. It may also be that we had a much, much more accurate model of the vent and so we were able to get a more realistic density, although still clearly much less than the density of basalt, which reflects the gas charged nature of the lava in the vent.

Well, let's move now to the Pu‘u ‘Ō‘ō vent. After Pu‘u ‘Ō‘ō collapsed and once the ash plumes cleared, there were helicopter overflights, lots of photos taken, and we were able to do the same thing at Pu‘u ‘Ō‘ō as at the summit. That is, use the photos to construct a structure from motion model of the interior of the eruptive vent. You can see that model here in perspective view on the left, and in plan view in the lower left. That also features the location of the PUOC gravimeter. And then the cross section on the right shows what this looked like inside viewed from east to west. Now, the difference between Pu‘u ‘Ō‘ō and the summit was that we did not have any information on the lava level over time at Pu‘u ‘Ō‘ō. Unfortunately, the camera views were not able to see into the vent so we only know about the time that the lava level started to drop at Pu‘u ‘Ō‘ō. We don't know anything about how that drop occurred. But that's something we were able to invert for as part of our analysis of gravity change at Pu‘u ‘Ō‘ō.

And the gravity change at Pu‘u ‘Ō‘ō was really impressive. You might recall a few slides ago I showed the summit gravity change. We had about a milligal of gravity decrease over several days—the second largest change ever recorded. Well, at Pu‘u ‘Ō‘ō, over a matter of just a couple of hours, we had one and a half milligals of gravity decrease—easily the largest gravity change ever recorded. Now, as I mentioned we didn't have any lava-level information, but that was something that we could invert for along with lava density. And when we do that, we get a lava-level change that looks something like this, where about there were 300 meters of lava level drop over the course of the two hour or so draining of the lava level at Pu‘u ‘Ō‘ō. Now we can also use this to calculate the lava density. And we get about 1900 kilograms per cubic meter and the volume loss over this two hour period was 11 million cubic meters or 8 million cubic meters dense rock equivalent. Now this density is significant for two reasons, it's greater than that at the summit. First, that may reflect pre-eruptive degassing at the summit; we know that some of the magma that degassed at the summit then later went down to Pu‘u ‘Ō‘ō where it erupted, so it stands to reason that we would see a higher density of a lava lake at Pu‘u ‘Ō‘ō relative to the summit. Second, the lava level elevation at the summit is higher than that at Pu‘u ‘Ō‘ō. And that might also reflect a greater density of lava at Pu‘u ‘Ō‘ō. In fact, this difference of about 200 kilograms per cubic meter is roughly what you would expect, given the lava level elevation difference at the summit and Pu‘u ‘Ō‘ō. So the density that we're seeing at Pu‘u ‘Ō‘ō and the summit are very consistent with observations from degassing and surface elevation.

There's another signal here that I think we need to spend a little bit of time focusing on, and that's this little transient right here—a 200 microgal decrease followed by a 400 microgal increase, just a few minutes before Pu‘u ‘Ō‘ō collapsed. This may indicate something that occurred prior to the onset of the great dike intrusion that fed the big lower East Rift Zone eruption in 2018.

Here's a zoomed view of that gravity transient that occurred just a few minutes before collapse of Pu‘u ‘Ō‘ō, and on the bottom is tilt measured at a tiltmeter just on the north flank of Pu‘u ‘Ō‘ō. You can see the tilt was starting to accelerate early in the hours of May 1, but this gravity transient really showed up quite strongly before there was really strong ground tilt. In fact, it showed up just after the onset of some interesting seismicity out at Pu‘u ‘Ō‘ō, an increase in seismic tremor. This may have been associated with the onset of a small fissure eruption on the west flank of Pu‘u ‘Ō‘ō.

Here is Pu‘u ‘Ō‘ō cone. This photo was taken after the collapse, but at the time it was noted that there was a small fissure that ran down the west side of the cone. This fissure had a very small amount of lava that had come out of the ground, mostly degassed lava. This gravity signal, the decrease followed by the increase, may indicate the onset of this small fissure—the opening and then filling of this fissure with magma. It may be that we're seeing the opening sequences of the entire eruption. A small fissure on the west flank very strongly manifested in the gravity data from Pu‘u ‘Ō‘ō.

Now with only one gravity station it's really impossible to model this but we can come up with a conceptual model of what this might mean. The opening gravity decrease could represent the opening of a dry crack, something that you might see in front of a propagating dike. The dry crack opens, the overall density decreases because you're replacing rock with air and gravity goes down, and eventually magma fills that crack and we get an increase in gravity. In this case at Pu‘u ‘Ō‘ō, the gravity increase was much greater than the decrease because the magma filling that crack would have a much much higher density than the agglutinated spatter that makes up the cone. And so we might explain this gravity transient as the opening of this dry crack and then the filling with dense mostly de-gassed magma that had a small eruption on the west flank of this fissure. So gravity data provide an interesting perspective into what could have been the very opening sequence of the entire 2018 eruption, a small fissure that opened on the flank of Pu‘u ‘Ō‘ō

We just find these results incredibly exciting. Not only can we use the continuous gravity data to identify things like volume of magma that drained from Pu‘u ‘Ō‘ō and the density of that magma, we may also be able to detect much more strongly than other measures—other monitoring data like tilt or seismicity—the occurrence of very important transits.

In conclusion, we've seen gravity changes of a milligal, and 1.5 milligals at the summit and Pu‘u ‘Ō‘ō respectively. These are the largest gravity changes that we're aware of ever having been recorded. Using these data and models of the eruptive vents, we were able to calculate a density of the summit lava lake of about 1700 kilograms per cubic meter while the Pu‘u ‘Ō‘ō lava lake density was slightly higher at 1900 kilograms per cubic meter. This density difference is very interesting. It's consistent with pre-eruptive degassing that was occurring at the summit vent and also the difference in the elevations of the two lava lakes. Finally, there's this major gravity transient 200 microgals decrease followed by a 400 microgal increase all occurring over the course of just a few minutes. This preceded the onset of the 2018 dike intrusion, it proceeded the collapse of Pu‘u ‘Ō‘ō. This may have been the opening event of the entire sequence, and it was indicated more strongly in gravity than any other data set. This really is a great argument for a continued investment in monitoring using continuous gravity, which hopefully in the years to come will become cheaper and easier as new technology is applied.

Well there you have it, the story of the largest gravity changes ever recorded. Hope you found that interesting. And if you have any questions, feel free to email me anytime, Thanks for tuning in and seeing it out to the bitter end and hope to see it a future conference sometime soon when this is all over and we're back to normal. Take care everyone.