For a Kilauea update, see the bottom of the post
Astronomy is a dangerous science. It is not just the fact that astronomers love to blow things up. Exploding stars are their bread and butter. For a bit of adventure, they collide black holes. And who else would start the history of everything with a big bang?
Put astronomers together with volcanoes, and you have an explosive mix. Observatories are often build on volcanoes: the mountains are high, isolated, and especially young volcanoes are smooth which ensures a stable (and dry) airflow over the summit which makes for an excellent atmosphere. Here, stars don’t twinkle – they shine. But it should be seen as a warning that no one else puts their main instruments on top of these young volcanoes. Even volcanologists know better. The lesson was learned after the Etna volcano observatory, which started out as an astronomical observatory, was destroyed by the very lava flow it was meant to observe. Now volcanologists keep a safe distance. Not so for astronomers. They want bangs for their bucks. Whilst the volcanologists relocated to Catania, the astronomers rebuild their observatory on the slopes of Etna, halfway to the summit.
Luckily, only two of the world’s major observatories are at significant risk of volcanic damage. La Palma and Mauna Kea are fine: they are very unlikely to have any summit eruptions, ever. Siding Springs, in Australia, is long extinct. Mauna Loa is a major hazard for its solar observatory, located a bit below the summit, which has a seriously reduced life expectancy. And there is one other observatory at risk of being the victim of the mountain it is built on. And it is not on the Big Island, with its perpetual eruptions, but on the next one along, Maui. The volcano is Haleakalā, and it has a 50% chance of erupting this century. Ignoring subsea Lo’ihi, it is the third most active volcano in Hawaii. Perhaps not the safest place to spend the night at work.
The Haleakale observatory is built on the summit ridge, at just over 3 kilometer altitude. It is just outside the national park boundaries, and just off the highest peak of the island.
Unusually for an observatory, the telescopes are packed close together. There are a range of scopes, ranging from night-time astronomy, and a major solar telescope to satellite tracking. There is a gamma-ray telescope, a lunar radar experiment which became a satellite laser ranging system. The solar telescope has a 4-meter mirror and is costing over 300million dollar. Add everything else and there may be equipment worth a billion dollars here.
Not everyone is happy about this. Haleakala is sacred to the Hawaiians and they have some issues with the level of build-up. The word means ‘house of the sun’; originally one location, it became adopted as the name for the entire volcano. A major solar telescope would seem to fit with the idea, but one has to admit that the site now looks industrial rather than reverential.
But it seems dangerous to put so much value on so little real estate in such a volcanic location. Pele on a bad mood, could easily take revenge. Should the insurance company get worried?
Maui and Haleakala
The island of Maui is immediately north of the Big Island. As everything in Hawaii (and in fact as are all deep-sea islands) it is volcanic, with as only addition the coral reefs. Maui consists of two volcanoes linked by an isthmus. The West Maui volcano is older and extinct. The southern part of the island is Haleakala, which peaks at just over 3 kilometer. The isthmus consists of lava flows from Haleakala and therefore should also be considered part of it. In the past, it was also connected to the island of Kaloohawe but the connecting isthmus is now entirely submerged.
Haleakala has a central depression which is also called the crater (which it isn’t), a peak, several deep valleys, and a large number of subsidiary cones fed from three rift zones intersecting at what is assumed to have been a summit vent. The three rifts are still recognizable, and are helpfully named the north rift, the southwest rift and the east rift. Under water, it continues as the Haleakala ridge.
Haleakala (click for the full resolution 30d map) started to form as a shield volcano, with frequent, voluminous and fluid lavas which build a dome a little less than 3 kilomers tall. This was completed by just under 1 million years ago. In the next phase, the lava became less fluid and more explosive. Fire fountains build tall cinder cones. In this phase, the peak become much taller and the mountain steeper. By the end of the phase, it was approaching 4 kilometers and would have looked much like Mauna Kea does nowadays. The mountain had reached peak size around 360,000 years ago when eruptions waned.
As eruptions became less frequent, erosion took its toll. Water streams formed rivers, and the rivers carved canyons. Each canyon formed an amphitheatre at the top. New eruptions would fill in the canyons, sometimes overflowing it, but eventually the canyon would reform. The amphitheatres eroded themselves up the mountain: the two largest ones merged and together they ate away the summit. This eroding hole formed what is now called the crater: the 3-d image above shows it well. It never was a true crater: the summit depression is caused by erosion. The summit ridge is the wall of the amphitheatre, and is about a kilometre west from where the original summit would have been. The ridge smoothly joins with the side of the Ke’anan valley. The rift cuts across the amphitheatre and joins the summit ridge more or less at the location of the observatory.
Because of the trade winds, the rain falls mainly on the eastern side of the Hawaiian islands and therefore the eroding canyons also attacked the mountain from the east. The west side lacks the deep valleys. The streams carried other risks: at one point a huge mudflow obliterated much of the Kaupa valley.
This was not yet the end. Eruptions had never completely ceased. Around 60,0000 years ago or so, it seems the eruption frequency increased again. This phase is named the Kula volcanics. The source of the magma was the same: the chemistry did not change. Even though the volcano had been active for 2 million years and the hot spot was by now 100 kilometer away, magma still found its way from there. The northern rift had ceased activity, but the volcanic eruptions resumed from the other rifts. The lava flows filled in the amphitheatres and came down the canyons. The northern side of Haleakala was not covered because of the lack of activity there, but the other sides were recovered with lava. In this phase, the fissure eruption produced chains of cones. These most recent eruptions comprise the Hana volcanics.
Haleakala is still in this phase. Eruptions are continuing from the two rift zones, not as frequent as Kilauea or Mauna Loa but far more frequent than Mauna Kea. The current rate of eruptions is estimated at about ten per millennium. Both in the summit depression and where the rifts meet the coast are areas of black lava, which appear to be very young. This is an active volcano.
The changing coastline – or fake news?
The two youngest-appearing flows are at La Perouse Bay, on the southwest side of the island. The upper flow came from the side of an older cone, Pu’u Mahoe, 500 meters above sea level. The lower flow came from a parallel fissure, a little eastward but much lower at 200 meters, and emanated from Kalua o Lapa, another cone. The two eruptions may have been simultaneous. The upper one reached the sea and build a bulge extending the cost outward by 1 kilometer.
There is an account of a recollection of the eruption. It comes from Lorrin Thurston who recorded a conversation he had with Father Bailey, who had lived on Maui from 1841. The conversation took place in 1879. The following extract is taken from Stearns and Macdonald 1942:
“I was first stationed on Maui in 1841”, said Father Bailey. “In my trips about the island I noticed a lava flow at Honuaula, and the south end of East Maui, which appeared much fresher than the other flows – much more so than it appears now. I asked the natives if they knew when the flow occurred and they told me that their grandparents saw it. They also told me that a woman and child were surrounded by the flow but escaped after it cooled.”
Thurston later (1906) found one other person who talked about family stories, involving a great-grandfather. He tried to estimate from the stories when the eruption could have taken place. He calculated a data around 1750. Others felt that he used high estimates for parenthood, and derived a bit later date of around 1770.
In 1965, Loosdam pointed out that historical maps of the coast seemed inconsistent. The oldest maps was by La Perouse in 1786. Compared to the second oldest, Vancouber 1793, it misses the prominence of Kape Kenau. The drawings are reproduced below.
Cape Kinau is numbered ‘3’ on the map. Number 1 is a cinder cone, Pu’u’O Lai. The dotted lines show the route the two sailors took. Cape Kinau is of interest because there were stories among the local people about its formation.
The problem with this is that the dates are a bit later than the oral history implies. One can question the accuracy of the maps: it is entirely possible that La Perouse would have missed the promontory.
A new surprise came from two other dating techniques. The first was the direction of magnetic north at the time the lava was emplaced. The magnetic north pole is constantly moving, and when the lava solidifies it freezes the magnetic field of that time into the rock. And the orientation of the Perouse lava was not consistent with where the magnetic pole was around 1800: it seemed centuries older. Carbon-14 dates were obtained, and they also pointed at an earlier date, about 1500 AD plus or minus a century. Was the oral history fake news?
The answer may be in the stories. There are two retellings, but neither story mentioned that the lava had flown into the sea. It would seem that new land would have been remembered! It seems plausible that the oral memory refers to the second flow, the one that did not reach the sea, whereas the rock dating focussed on the promontory flow. This is just a suggestion, but it reconciles the facts. In that case, there was an eruption around 1500 and another one around 1750. However, the data at the moment only shows a single eruption some 500 years ago and nothing since.
Carbon dates have been obtained for a number of the lava flows. This is done by digging through the lava and finding some organic material (twigs or some such) from just below the layer baked by the lava. The lava will be younger than the twig, but if close enough, it may be very similar in age. The oldest flows from the current sequence are 60,000 years old. But people have of course focussed on the youngest looking flows and found interesting results.
In the summit depression, two separate flows have been measured at 970 BO and 940 BP. Within the uncertainties, they could have been simultaneous, however they come from different eruption sites. In both cases, the lava went north, flowing down the mountain and reaching the coast at Ke’anae . The older flow came from Ka Lu’u o ka ‘O’o, a youthful cinder cone on the crater floor. The precise eruption point is covered by a more recent lava flow, probably from the same cone, but this flow has not been dated. The younger flow comes from the nearby Halali’i vent: it also contains a 10-cm thick layer of black ash. The youngest dated flow in the summit area is in the north part of the crater, and came from the Hanakauhi fissure: it is from 870 BP (plus or minus 40 years). There are apparent younger flows in the western crater area but no date is known for them.
There are three flows dated to between 1000 and 2000 BP. It seems that the number of eruption is fairly constant over time. The typical volume of an eruption is 0.1 km3 and the flows reach a few kilometres in length, although flows longer than 10 km are also not uncommon.
Over the past 1500 years, lava flows have covered almost 90km2 of the island. That is substantial but not extreme. If the risk was uniformly distributed over the island, ether would be little to worry about for the average local, as the volcano covers close to 1500 km2: each location could expect a lava covering once every 20 thousand years. But the lava flows are highly localized. This is first because all flows originate from the line of the rift, and second because lava always flows downhill. Many area are protected by ridges. The summer crater rim protects much of the island. Elsewhere, lava tends to go down the river channels, something also known from the fissure eruptions on Iceland. Adding these together, there are three main areas at risk: up-rift from La Perouse bay, the summit depression, and the area round the Hana airport on the east coast.
Haleakala is certain to erupt again. We can’t quite predict where, but when it does it will produce a decent-size lava flow. The Observatory may stand above the fray: the most likely location for an eruption would be the crater floor, with the lava flowing down one of the main valleys. The second risk area is the west ridge. When will it be? The volcano may have bene quiescent for 500 years, or there may have been an eruption 250 years ago. The summit has been quiet for 800 years. The calm won’t last forever, but holds for now.
The sole GPS on the volcano is located close to the observatory. That makes sense, giving the value of the telescopes there. But the rifts should not be forgotten. If HVO has a GPS or two spare, Haleakala could use them. The astronomers would appreciate.
Albert, May 2018
Update Kilauea, 11 May
Kilauea has reached a state of ‘hurry up and wait’. The fissure eruptions have stopped for now, after 15 or 16 separate vents each operating for a few hours, over a few miles right across Leilani. The most recent activity was at the end of the set of fissures, approaching the geothermal plant. Although gas venting is reported around highway 130 to the west, and earthquake activity has intensified around the geothermal plant to the east, nothing of significance has yet happened. It should be pointed out the eruptions so far have not been much more than blowing off steam. There is only one significant lava stream extending for about 1 kilometer. Very little of what has collected below has come up. At the moment, it seems most likely that the next activity will be to the east but really, anything is possible. However, if it does continue to extend east-northeast, the geothermal plant is directly on the track, and a bit further is highway 132: if that is cut, it leaves a section of the coast without a good exit strategy.
HVO is reporting that deformation is continuing in the area, so there is still activity on-going and magma is probably still arriving. The flammables are finally being removed from the geothermal plant. One wonders whether the owners might have preferred to collect the insurance money!
Pu’u’O’o is dead as a door mouse. The lava pond is fully drained. Whether it will recover depend on whether the conduit has survived. If it has become blocked, than once everything is back to normal the magma may look for other ways out. Pu’u’O’o has survived a lot over the past 30 years and it may well come back from this one too. But the longer it is quiet, the greater the chance of permanent retirement.
Kilauea is now the most eagerly watched spot. The lava pond has disappeared from view and can no longer be measured. It was at around 300 meters below the crater floor. The image below is a screen grab from an HVO presentation, showing the decline. As predicted (you read it here first), the decline started exponentially and became linear. What had not been predicted is how fast it would go (2 meters per hour!) and how far it would go down.
In the absence of future lava depth measurements, the tilt measurement is the best chance to see what is going on. The plots can be found at https://volcanoes.usgs.gov/volcanoes/kilauea/monitoring_deformation.html and they show continuing linear decay, although with more and more jumps where a minor earthquake indicates some relaxation of the local crust.
Noone seems certain where the magma is going. There was talk about the magma chamber being enlarged by the big quake but that seems implausible. I think that most likely it is going into the dike. The width of a dike is set by the magma pressure: it pushed the walls apart. (It is amazing how flexible rock can be a kilometer underground! Add more magma, and it will create its own space by pushing the dike wider. However, this also predicts how it could end. As the dike widens, the pressure remains constant and the eruption remains stable at its current rate of bits and bobs. Once a limit is reached, expansion stops. The dike will try to narrow and and that won’t work because magma is incompressible. Suddenly, pressure shoots up. A break-out happens, and the dike slams shut. At this point, Kilauea will stop sinking. (Of course, a much more sedate, petering-off ending is also possible! But it doesn’t sound nearly as exciting.)
The big talking point (again, you read it here first) is the water table. That is thought to be some 400 meters down, and at the current rate will be reached within 36 hours. At that point, water will be able to reach the now-empty conduit. Actually, there will be some delay as the walls need to cool a bit. Steam explosions in the walls cause debris to fall down. Once water gets trapped below the debris, significant phreatic explosions become possible. It is far from certain this will happen, but HVO has noted the risk. And these explosions can be massive (although most are not). Expect the overlook with the camera to be badly damaged in such an event, and small stones could reach the observatory building.
But all this remains speculation. Kilauea may be changing its behaviour now, or things may go back to the old ways. HVO is giving out warnings – not predictions.