Kilauea is back erupting since December 20! The eruption style is typical of Kilauea, yet it’s been decades since it last showed it and many aspects have not been explained properly. The most important unanswered question being the difference between a rootless lava lake (this eruption) and a “true” lava lake.
The eruption came as a surprise, the Hawaiian Volcano Observatory (HVO) was apparently planning to raise the alert level but Kilauea was faster than that. HVO had noticed the higher state of unrest the volcano was undergoing. However as it usually happens with volcanoes you can tell its preparing something but don’t know how, when, or even where the volcano is going to erupt. Since I was watching all the signals closely I can give a summary of the situation before the eruption took place.
Prelude to eruption
If you enter the deformation data page of HVO the first thing that shows up is the UWE tiltmeter. UWE stands for Uwekahuna Bluff which was the former location of HVO (before the 2018 caldera collapse that damaged the building). For decades the Uwekahuna tiltmeter has been used to record the ups and downs of pressure at Kilauea, it measures changes in slope away from Halema’uma’u Crater. Two kilometres below sea level under Halema’uma’u are located the shallowest magma chambers of Kilauea its pressure changes is what the tiltmeter shows.
Around September this year an unusually fast inflation was registered by GPS stations located south of the summit caldera, the inflation was accelerating and also changing its centre. At this time the Uwekahuna tiltmeter was not picking any inflation and the levels of earthquakes in the UERZ (Upper East Rift Zone) were very low, they fell down since the start of October, around the time when Pu’u’o’o stopped inflating too. All of these signals show that little magma was reaching the shallow storage of Kilauea but was entering at increasing speeds into the deep storage south of the summit area.
Slight unremarkable inflation resumed into the shallow chambers of Kilauea in mid-November. On November 30 an intrusion took place under the summit that had a similar seismic signal to later intrusions of December, at the time I thought it may have been a sill, but this remains unclear. The intrusion reduced the pressure of Kilauea by a very small amount and the volcano was re-inflated in a few hours.
The surge of magma supply affecting the deeper portion of Kilauea’s storage since September might have well reached the shallow magma chambers on the morning of December 2. The tiltmeter at Uwekahuna recorded rapid inflation while a swarm of earthquakes lit up the East Rift down to Pu’u’o’o. These swarms have been called inflationary swarms by some seismologists, they often happen when the volcano inflates rapidly. A molten conduit connects the summit chambers of Kilauea to chambers within the East Rift Zone, if pressure is high and increases further the conduit pushes against its roof and fractures it producing a linear swarm of tiny earthquakes that is most pronounced along the Upper East Rift. Many eruptions, including 2018, have had precursory swarms along the East Rift Conduit, regardless of where they erupt from (East Rift, Southwest Rift or Summit). A short conduit running south of the Summit into the Southwest Rift also shows sometimes.
On the evening of December 2 the summit snapped to produce a dike. HVO has modelled it as an inclined dike intrusion below the south rim of the caldera, dipping 60-40º north-westward. The tilted intrusion caused the caldera to lift like a trapdoor which produced positive tilting in the Uwekahuna tiltmeter, and shows as an impressive upward jump in the graphs. Like any sill or dike intrusion would do it reduced the magma pressure within the volcano, the volume intruded was very small, 0.4–0.7 million cubic meters, and probably recovered fast.
Rapid inflation in the shallow and deep magma storage of Kilauea continued until the eruption. Another swarm of small earthquakes extended along the UERZ on December 17, however Pu’u’o’o did not show any inflation which could be due to magma flow stopping somewhere uprift from it.
At around 20:30 HST on December 20 a new dike intrusion started at the summit, an hour later it breached the surface and the eruption had started. This dike was also inclined, and while HVO hasn’t spoken about it, it is apparent from the short lived uplift signal picked by stations on the south caldera rim that the intrusion dipped southeast, the opposite of the December 2 dike. A large deflation kicked in afterwards due to the eruption.
I was personally surprised by how there was practically no explosive interaction with the water inside the caldera. Kilauea has produced many prehistoric explosive eruptions, some of them very dangerous with pyroclastic surges or lava bombs and blocks bombarding the summit area. A couple of them are probably dry and wet caldera collapses, others may result from open conduits interacting with shallow groundwater. I also suspected some of them had been the result of dikes cutting through shallow groundwater or water lakes, particularly given that this happens at other basaltic volcanoes. This time however the dike must have cut through groundwater near the surface yet it erupted in an effusive style.
The most remarkable feature of the eruption is the formation of a huge “lava lake”. But should we really call it a lava lake? It is actually completely different from those sitting within Erta Ale, Masaya, Nyiragongo or the 2008-2018 Overlook lake at Kilauea, the ones typically called lava lakes.
What sets them apart is the conduit, the Overlook lava lake was a wide cylinder, around 200 meters in diameter and going down to the nearest magma chamber, deep-rooted. The molten cylinder allows magma to freely circulate up and down which keeps it hot and gas filled, the gas forms small bubbles suspended in the magma, if a rockfall pours into the lake then an amount of gas is suddenly released and explodes.
On the other hand, the lava lake currently active at the summit is rootless, it has formed from lava collecting within the funnel-shaped Halema’uma’u crater. Had the dike erupted in the slopes of the volcano the lava would have simply spilled all over the place, due to the number pit craters at the summit and east rift zones of Kilauea many rootless lakes have been recorded historically at Kilauea, the last one in 1979 as far as I remember.
Since a rootless lake is fed from a dike, it erupts from a narrow crack, unlike the Overlook case, the crack only allows magma to move either up (a fountain) or down (drainback), if you have several cracks/vents active then you can have some of them acting as fountains and others draining the lake. A rootless lake depends on the amount of magma coming up to keep it molten, it doesn’t have the smooth circulation of a broad cylinder of magma, so they are short-lived. If magma input stops or slows too much the lake’s surface rapidly solidifies, this insulates the interior which remains molten, some smaller molten openings can remain open for a long time.
More differences are that rootless lava lakes make lava fountains, whereas deep-rooted lava lakes do not. The way the level changes is also different, a deep-rooted lake is connected to the magma chamber pressure (it reflects magmastatic pressure), it rises and falls with the pressure changes in the volcano. The level of a rootless lake instead depends on the fountain and drainback processes taking place within the feeder cracks so it can change independently of the volcano’s pressure, it should still reflect somewhat where the magmastatic pressure of Kilauea stands though.
Some of you may be struggling with the idea of drainback, after all it’s not every day that you see a volcano reabsorb its own eruption. Note that lava in the lake is colder, degassed and therefore denser, lava that erupts from the fountains is hot, gas-rich and lighter. This contrast in density is what creates an struggle between the rising fresh magma and the lake lava flowing down. If there was just one vent it would alternate between a lava fountain and drainback like it happened during the 1959 Kilauea Iki eruption. However there are multiple vents in the current eruption, when lava drowned the north vent on December 26 it simply switched to the west vents and kept erupting from them.
There has only been minimal drainback in December 25, when the lake seems to have fallen 2 meters, I find it unclear which of the vents was responsible. This is really a very small amount, in the past entire lakes have disappeared down the sinkhole, this shows there is magmastatic pressure supporting the current lake.
There is an enormous floating island of solid lava inside the lava lake, part of the island is under the lava so it’s like a lavaberg, there are also a few smaller floating pieces. The lavaberg is 260 x 115 meters across, it seems to have formed early in the eruption from the lava which first ran into the water lake and solidified. But how does rock float in lava? The explanation could be that the material forming the island is spongy, vesicular, like pumice floats in water the lavaberg could be porous enough to be buoyant on the lava.
It also rises and falls, in the past few days it seems to have been rising above the lake and a lot of it that was previously concealed by the lava is now emerged.
Future of the eruption
At 2:40 AM of December 26 the eruption came to a turning point. The west vents reactivated and erupted together with the north vent for 20 minutes, then the north vent became quiet. The west vents have kept erupting while the lava lake rises very slowly and the volcano inflates, in other words the eruption reached equilibrium, there is no reason for it to stop now, has it become sustained? and if so, for how long?
The west vents erupting in parallel with the volcano inflating suggests Kilauea will keep erupting until something changes in the volcano. I believe the situation is similar to summit activity in the 1823-40 years which was interrupted only by rift eruptions and small caldera collapses in 1832 and 1840. As such we might expect the eruption to continue until a dike intrusion or eruption takes place in the rift that drains the summit and destroys the active conduits. There could of course always be some unforeseen factor that ends the eruption, but I don’t see any reason.
In the meantime the surface of the lava lake is likely to solidify. This is already happening on one side of the lake where a solid ledge has formed. An elevated levee of solidified lava encloses the lake on this side and frequent flows overtop it and spill onto the ledge. While so far most of the surface seems to remain molten, the eruption rates have lowered significantly since the first days, and all it takes is one deflation-inflation event (cyclic overturnings of Kilauea’s magma system that generate pressure changes) to temporarily pause the eruption and a portion of the lake to go solid. So most likely over the next several days/weeks the lava lake will shrink in area.
The interior does remain molten, for as long as decades if it doesn’t drain away! So it can become like near-surface magma chamber. As of 2020 there are still new publications coming up based on data collected from the Kilauea Iki rootless lava lake of 1959. Kilauea Iki was drilled repeatedly to study how magmatic differentiation took place in magma chambers and the interaction between magma and hydrothermal systems, it turned out to be the perfect natural experiment!
However the Kilauea Iki eruption was different in many ways, this time the eruption is most likely to become sustained. A small molten lake will probably remain next to the west vents even if the rest crusts over, additionally the drowned vents (there are at least 3 submerged openings) might inject gas-rich magma from below, which, if it comes to be the case, would be interesting. It is likely that the crater will keep filling slowly until the next rift event happens. Filling may take place both from lava flows as well as piston uplift of the crust.
Back in the 1823-1840 people seldom visited the summit of Kilauea however when they did they would describe multiple cones and lava lakes erupting through the caldera floor, often they were witnesses to how the surface suddenly cracked and a new vent formed. Rootless lava lakes were presumably molten but crusted over at those times and as magma was injected into the melt below it created new openings. The vents often got covered with sticky lava and released gas in episodic bursts, strombolian eruptions. It’s hard to tell, but we might see a similar activity show up in the future.
The East Rift
Most of Kilauea’s volcanic edifice is formed by lava flows and intrusions in the East Rift Zone. This rift is 130 km long, 55 km of which is subaerial, the rest goes on underwater. Since 2018 much of the subaerial portion has been actively spreading and inflating (up to half a meter of uplift in areas of the Middle ERZ). Spreading takes place both from slow expansion of a crystal mush at 3-8 km deep and intrusion of dikes above 3 km deep (none since 2018), fissure eruptions and satellite shields cover the flanks in lava. The internal growth is accommodated by seaward movement of the southern side of Hawaii Island over a system of reverse faults that Kilauea shares with Mauna Loa, capable of generating magnitude 8 earthquakes (great Kau earthquake of 1868).
Activity in the ERZ has been highly variable and cyclic. The 18th century saw high activity, then following the enormous summit collapse of 1790 activity dropped to almost none throughout the 19th century, and finally the 20th century was a gradual return to ERZ activity. The events of 2018 have if anything prepared the scenario for more rift activity, the M7 earthquake accelerated spreading rates (the flank keeps moving faster than during Pu’u’o’o) while the eruption improved magma transport to Middle ERZ. So if we are looking into the future of Kilauea we have to consider what the East Rift will do.
It seems that most likely the summit will go on erupting until the next time the East Rift erupts or produces a non-eruptive dike intrusion, it could also be in the Southwest Rift but this is less likely because it’s much less active and not spreading much right now anyway. However I still find uncertain where or when the rift will snap next, it could take a few years but I wouldn’t be surprised if it happens in 2021 either, a dike intrusion in the Upper ERZ is unlikely to breach the surface while an intrusion in the Lower ERZ is likely to produce a voluminous eruption.
For the time being looks like we are going to have a nice show from the summit of Kilauea, however the future points back to the East Rift, to eruptions that might not be so pleasant.