A Kilauea/Puna update has been appended at the end of the post
Distance makes the heart grow fonder. Volcanoes are best loved from far away. The excitement of live lava is best viewed on a screen and not through the window. Of course, actual distance is good, but distance in time is also often deemed sufficient. Your house may be next to a lava flow, but if that flow is 300 years old, the chances will seem pretty good. It makes that threatening mountain into a true neighbourly volcano, viewed with affection rather than trepidation.
So it must have appeared to the inhabitants of Puna. Yes, there were lava flows around, some less than a century old. Yes, the area around the Royal Gardens subdivision had been utterly destroyed by the seemingly never-ending eruption. But that was much closer to Kilauea, and Puna should be safe. Almost 60 years of quiescence was a pretty good sign. And Puna is a really interesting place to live. The forest is absolutely beautiful, and the tropical gardens could grow anything (including, admittedly, some fierce mosquitos). The area was initially quite isolated and the community self-sufficient: it was a good place for people with an independent, self-reliant attitude. Not everything was paradise, and the distinction between rich and poor could be pronounced. Development had accelerated with time, with more and more houses and industry. Part of the forest was cut to make way for farming. The geothermal plant generated 20% of the Big Island’s electricity. And still it was not a bad place to live. At least, until May 3rd, 2018 when the first fissure opened and the gardens started to grow flowers of quite a different kind.
But how secure should the people have felt? Was this eruption bad luck, or was it waiting to happen?
The map at the top shows the elevation. Mauna Loa is the elongated elephant. Kilauea is the bump just below it, on its flank. The site of Pu’u’O’o is indicated by the red arrow. The east rift is clearly seen. Mauna Loa is known for its long rift. But Kilauea is not bad either! This volcano is not just a pretty summit. It too has a long reach. Before 1950, summit eruptions were the norm. Since 1950, the summit has done rather little and eruptions have tended to come from the east rift, dominated of course by Pu’u’O’. The entire east rift is a staggering 125 km long, over half of which is beyond Puna in the ocean. The part between Pu’o’O’o and the tip of Puna, Cape Kumuhaki, is sometimes called the Lower East Rift Zone, or LERZ for short.
The LERZ is a little over 20 kilometers long and a few kilometres wide. It forms quite a pronounced ridge, slowly declining in elevation from over 500 meters near Pu’u’O’o to sea level at the coast (admittedly this states the obvious). The ridge stands 50 to 150 meters over the surroundings, and is built by spatter from the many eruptions. Individual cones may be as high as 60 meters. On the southern side the slope is much steeper than on the north: this is caused by the tendency of the south side to slip towards the sea, as also happened earlier this month. Lava goes either northeast or south/southeast, depending on exactly where on the ridge it originates. It is often a close call. Most of the eruptions have been from long rifts similar to the current one. The rifts tend to form a line of rift segments rather a single rift. The segments are not entirely aligned: they tend to show sidesteps, 10-100 meter sideways. Rifts on the northern part of the ridge tend step towards the left, and ones on the southern side step to the right, in both cases diverging a bit from the ridge further along the LERZ. However, towards the eastern end, the southern rifts can become left stepping, towards the north. There are other types of eruptions. There is one large shield (Heiheiahulu), very similar to what Pu’u’O’o has done and in fact not very from that location, covering the area directly to the coast. There is also one explosive crater, 1-km wide Kapoho. It is only just above sea level, and so it is likely that the explosion was caused by sea water intruding into the rift. This was not a minor event: an area of 6 km2 was covered by its pyroclastic flows.
Let’s look at some of the more recent eruptions. Historical events began with the 1790 eruption, although there are oral stories among the Hawaiian of older eruptions.
The 1790 eruption was one of the defining outbursts of Kilauea, with deadly pyroclastic explosions at the summit. But the eruption also involved the LERZ where it was a major event. It formed not one but two rifts, 1-2 km apart, one on the southern side and one on the north. The southern rift was over 8 km long and the northern one a massive 13 km; both are segmented (‘echelon’) with side stepping. The flows cover 35 km2 and the volume in the LERZ area is about 0.12 km3. The crater of Kilauea may have formed in the 1790 eruption, but only a small fraction of the evacuated magma has come out in the LERZ.
The next major eruption was in 1840. It erupted from a northern rift, 7.5 km long, with a couple of vents about 1 kilometer further north, somewhat away from the ridge. The lava flowed northeast-ward and covered some 20 km2. Where it reached the ocean, a cone formed, build by explosions as the lava hit the water. The total volume of the lava was around 0.15 km3.
The 1955 eruption, described in the previous post, came from a southern rift, 16-km long which included a 2.5-km gap which did not erupt – which is where the 2018 eruption is occurring. It was slightly smaller than the 1840 eruption, covering 16 km2 with 0.11 km3 of lava. It reached the ocean in two areas.
Sofar, the 2018 eruptions have covered 1.3 km2. The volume is still less than 0.01 km3. We are still very far from 1790 territory!
The 1960 eruption came from a short segment, 1-km long, at Kapoho which was where the 1955 eruption had terminated. It followed on from an eruption at Iki, and like the current eruption, the rift eruption was accompanied by deflation at Kilauea. The lava covered 10km2 and amounted to 0.12 km3.
There was a short eruption in 1961, which covered about 1 km2 in the western part of the LERZ and lasted one day. And nothing since. There were three eruptions in 7 years, but nothing for 115 year before, and nothing after until now, a gap of 57 years! The frequency of eruptions in the Puna area seems slightly irregular. The south side of the ridge had no lava flows for 165 years. No wonder people felt safe. The period 1955-1961 seemed the exception, and the time since, the rule. Of course, Leilani was built on the ridge itself so it had twice the risk – but twice a very small chance still seems a very small risk (of course people who play the lottery think exactly the opposite: the potential reward skews the perception of chance.) 20% of Puna was covered in lava flows less than 250 years old, but most of that happened 200 years ago. That was a different age, wasn’t it?
But history tends to repeat itself. What does the older history have to say? What was the real risk of building in Puna?
All of Puna is lava. The lush growth conceals this well but look underneath and even the rain forest is built on lava. In this climate, it doesn’t take long for lava to green. The first plants germinate within months (although the heat on the black lava can be pretty hostile until some shade develops) . Trees can grow on 40-year old lava, and after that, only the height of the trees betrays the youth of the lava.
Carbon dating has been applied to many of the lava flows here, and this has shown a surprising picture. It seems that the past 250 years have been a bit of an anomaly.
Let’s go back a little more. A total of 15 different eruptions are dated to the 200-year period before the 1790 eruption. This includes the explosion of the Kapoho cone. It also contains the shield of Heiheiahulu, which covers a surface area of 45 km2. That eruption started as a short fissure, before building a large cone which at the end collapsed and became a pit crater. Heiheiahulu is very much like Pu’u’U’o. Local stories suggest that it erupted somewhere between 1700 and 1750, and if it is indeed a Pu’u’O’o copy, it may have lasted years. The volume is probably of order 1 km3, which is a few times less than Pu’u’O’o.
Also interesting is a flow from Ai-laau, a vent to the northwest of Kilauea. This produced a flow which made it to Puna, covering the region northwest of the ridge. The flow cover 100 km2: the summit still can outcompete the rift! It also shows another truth about Puna: while the southern and northeastern part are at risk from the LERZ, the area north of the rift is affected by eruptions from near the summit of Kilauea. This is another eruption mode we haven’t seen for some time, but it will re-occur – eventually.
Including these three, a total of 15 separate events have been identified for this period of time. In total they cover 50% of Puna. Ai-laau and Heiheiahulu are by far the most voluminous: excluding them, the rest amounts to about 1 km3. Clearly, the 200 years before 1790 was a scary time in Puna.
Going back another 250 years, eruptions become even more frequent. In the time period which runs roughly from 1300 to 1600, at least 20% of the area of Puna was resurfaced. It may have been much more, since the later flows will have covered up much of the evidence. A staggering number of 54 eruptions have been identified in this time zone. For about half we know where they came from, as we have found the spatter cones at the origins. For the others, the point of origin is buried under younger flows. The eruption of the spatter cone of Puu Kaliu came from a fissure about 1.5 km long, along the southern side of the LERZ. It covers at least 12 km2, and the volume is about 0.2 km3. Puu Kaliu is just south of Leilani.
Two other eruptions were of similar size. One is dated to 490 BP (with an uncertainty of 60 years either way) and lies on top of the Puu Kaliu flows: it covered at least 18 km2 and has a similar volume to Puu Kaliu. The second is dated to 440 BP and may have formed a shield similar to Heiheiahulu, but it is mostly buried by the latter and by the 1955 lava. The cliffs at Kehena, with at least 15 separate layers, comes from this eruption. Puulena Crater is another event from this time period. It is the site of a major phreatic explosion, possibly when a rift opened underneath a lake.
So this was another exciting period of time. Going back further, the record becomes more and more patchy as the evidence for the eruptions is buried underneath younger flows.
When Pele wants to play
So what can we learn from this? Some caution is required as not all dates may be that accurate. Also, some eruptions may have been double counted. For instance, the distinct flow fields of the 1955 eruption could have been counted as two different eruptions with these methods. But overall, it appears that the eruption frequency at Puna has been much higher between 250 and 800 years ago than it has been since. It also seems that the long duration of the Pu’u’O’o eruption is not that exceptional. There have been other longer-lived rift eruptions in the past.
Since 1790, the LERZ has erupted on average once every 50 years, although the intervals have been highly variable. Before that, the eruption frequency was closer to once every decade. And even though the individual volumes aren’t huge, peaking at 0.1-0.2 km3 with most eruptions considerably smaller, at this frequency Puna was rather quickly covered in lava. Almost the entire surface is less than 800 years old. Build a house anywhere in Puna, and the risk of losing it to lava is of order 20% per century. Build it on the ridge, and that risk doubles. Kapoho, at the coast, is particularly endangered: there have been a lot of eruptions in that area, and it runs pyroclastic risks when sea water penetrates the rifts.
The average dormancy time of the LERZ over the past millennium is only about 15 years. Puna is not just a distant appendix of Kilauea. Pele loves it out here, and many of Kilauea’s eruptions push into this region. The volume may be limited, but Puna is one of the most volcanically active areas in the world and possibly the most active in Hawaii. The past 200 years were exceptional. The current eruption is business as usual. These lush gardens are built on a regular supply of molten rock. There is always a cost to paradise.
Albert, May 2018
This post is based on the paper Volcanic geology and eruption frequency, lower east rift zone of Kilauea volcano, Hawaii by Richard B Moore , Bull Volcanol (1992) 54:475-483
Kilauea update 21 May
In the past few days the eruption has gone into another gear. Before, the lava was sluggish and build a ridge but had difficult going beyond that. But, as HVO has predicted, once the new lava arrived it was obvious. Fountaining was the first sign, and soon lava not just clogged but positively flowed. Lava rivers appeared and they moved fast.
The activity became confined mostly to the east-most end of the previously numbered fissures. Fissures 16 and 20 merged and became the main point of action. Fissure 17 came along but seems to fissing out. The map below is on a google map viewer. It shows that the action is just below the Puna Geothermal Plant, which presumably is now defunct. Interestingly, this plant was put in this location because of a thermal anomaly: the ground underneath had higher heat flow than other locations in the area. In hindsight, was this a warning? It is also the area that hadn’t erupted in 1955 while the ridge on either side (of Leilani) had. Now we known that the magma hadn’t skipped through: it had been stored here for later use.
An estimate from the map suggests that the total flow field covers 2-3 km2 so far. A large Puna eruption tends to cover some 10 km2. We are not in that range yet. The amount of erupted lava may be around 0.01-0.02 km3. The current flow field may not get that much larger in surface area until the eruption moves to new fissures.
The fissures have fed two lobes, one of which moved though housing and farm land, and the other through rain forest. The two lobes met and split again, flowing around a kipuka and fed two ocean entries. The eastern (farmland) than found a crack and disappeared into it. What happens next depends on how big a hole it has found, but lack of steam away from it suggest it has not gone far.
Obviously, where lava enters the sea you will get steam: with this volume the lava itself actually become hard to see. The white cloud is both impressive and extremely dangerous. It contains hydrochlorid acid and minute glass particles – think asbestos. Some of the videos taken from boats look dangerously close to the plume to me. (There is a 300 meter exclusion zone which they will be obeying.) Luckily, the trade winds are blowing the plume along the coast away from people on-land (the ones still there shouldn’t be). By the weekend the winds may change and blow the plume in-land, but again, hardly anyone is left there. Along the fissure, the SO2 is very bothersome and it is unlikely to get any better. Volcanic smog is awful and a considerable health hazard.
Fissure 17 (at the far end) is now reported to have erupted andesite. That is probably not that unusual. Puna has significant magma storage and depths of 1 to a few kilometer, and some of that evolved magma has come up. Fissure 17 is at the end of the current line of activity, and so the magma it produced may have had the least amount of mixing with the new magma. There was some discussion whether it came from 1955, but there is certainly far older magma around.
What is next? The current fissures seem to be less vigorous than before and they may have peaked. Cinder cones have build up which I estimate at 30-40 meters tall: some will survive. Fissure 17 will cease shortly, I expect. HVO has reported that GPS no longer shows expansion of the rift, and so the inflow no longer exceeds the outflow – at least here. The fact that the earthquakes around Leilani have ceased also points at that. However, if the current fissures become less active, the pressure up-rift will increase and this could well lead to eruptions there. Fissure 21 is particularly worth watching. Over the next week, there is a danger of similarly fluid eruptions within Leilani, and later also further west. However, this is far from certain! Everthing south of the fissure line is at risk. Outbreaks west of Leilani could also flow northward, although it is unusual for Puna eruptions of the same rift to go both ways.
Pu’u’O’o is dead. GPS shows that the exponential decline has bottomed out: there is nothing left there to drain. It seems unlikely to ever recover, and its final fate may be that of a collapsed pit crater. But who knows, it could still surprise us in a few months.
Halemaumau has settled in a pattern: it shows increasing tremor, explodes, tremor goes away but slowly rebuilds. It is not clear to me what causes the tremor: I was wondering about boiling water.The crater size expands with the explosion but the rock quickly recovers and resumes its contraction. There tend to be a few explosions at the same minimum crater size, after which it goes to the next size. The explosions are not as strong as earlier, but this means nothing. These explosions could continue for quite some time, until the lava has risen again to above the water table. The park could be closed for some time.
I have not seen any estimates of the damage yet, apart from a few million dollars due to tourists staying away (can’t entirely blame them, but the north side of the Bg Island is also spectacularly pretty and far from any volcanic disturbance!). But in the scheme of things, that will be small fry. To compare: the Kilauea eruption of Feb 28, 1955 destroyed 21 houses, 10km of road, 4000 acres of sugar plantations and one coffee plantation, amounting to 40 million dollar at today’s prices. Mt St Helens cost 1 billion US, which is half the price of Eyjafjallajokull where the dominant cost was the continent-wide flight ban. The mud volcano in Indonesia has cost around 3 billion so far. The 2018 eruption has already destroyed more buildings than in 1955. However, most of the 1955 cost was in the lost sugar, while in the modern economy there is much less value in the land and much more in property. We can estimate that Leilani as a whole is worth 200 million US, and the geothermal plant probably around 100 million, based on its 38MW capacity. (I find it amazing that Leilani was started less than a decade after the 1955 eruption!). Double the cost to include clean-up, road rebuilding etc, and one can guess a total cost of around half a billion US. That assumes that this eruption continues for a few more weeks: if it ends soon the cost will be less.