Volcano forecasts and Campi Flegrei

Solfatara crater in Campi Flegrei happily steaming away.

There are a few volcanoes that I do not feel comfortable writing about, and those are volcanoes that are far too close to large human settlements. The reason is obvious, it is far too likely that I will write about an event that will kill a lot of people.

There are two ways to increase the safety of those who live in these volcanoes, the first is mitigation. This can be divided into passive mitigation like evacuation plans, hazard maps and so on and so forth.

The second part is active mitigation, here the goal is to through engineering change the properties of a volcano so that it is less dangerous. This practice is so far rather uncommon, the shining example is of course the Ampera Tunnel at Gunung Kelud in Indonesia. The goal of the tunnel is to remove the crater lake that historically has killed ten thousand people. There is obviously more to be done in the fields of both passive and active mitigation.

The second part is accurately predicting if and when a volcano will erupt. But, before we get into this there is a bit of semantics I wish to clear up. I have over the last five years become ever more uncomfortable with the term “prediction”. It implies that you are using tea-leaves or fish-entrails instead of stringent scientific principles and empirical data to form predictive models.

I hereby suggest using the term “forecasting” instead, after all the common man is already used to weather forecasts, and forecasting volcanoes is in a way the same thing, and soon it will be as commonplace.

The science of volcano forecasting

Map of Campi Flegrei showing seismic activity and centre of uplift. From Chiodini et al, linked below.

There is not a single field in science that has advanced as quickly as volcano forecasting has done since professor Páll Einarsson accurately predicted the 2000 eruption of Hekla to within a minute of the actual eruption time.

Today forecasting of volcanoes are done on a routinely basis for many volcanoes. But, still far from all. The reason for the lack of volcanic forecasts for all volcanoes on the planet is due to lack of scientific research about them, or a lack of data-collecting equipment.

For well-studied and well-monitored volcanoes that we have a good track-record of and that does not erupt almost all the time we have at places achieved an accuracy of 8 out of 9 accurate volcanic forecasts. That number is better than weather forecasts, who would have figured that 20 years ago?

The first thing you need is a long track record of how that particular volcano acts during an eruption. Secondly you need data to analyse. Different volcanoes lend themselves to different types of data collection and equipment. Out of this data you make predictive forecast models so that you know what to look for before the next eruption. Then you sit down and wait.

The next part is to acknowledge that volcanoes can behave in unexpected manners, therefore you need to study how other volcanoes of the same type have behaved. This will give ideas for how to change your modelling if you encounter anomalous data from your equipment.

Volcanic gas changes at Campi Flegrei, picture from Chiodini et al, linked below.

There is a wealth of different techniques and equipment that can be employed against a volcano. The most common pieces of equipment are seismometers and GPS-stations so that we can look for volcanic type earthquakes or movements in the ground under the volcano. Quite often this is enough to forecast a volcanic eruption.

The golden rule for all forecasting, and almost all science, is that the more different types of data you have to work with the better your forecasts will become. Therefore there is a barrage of other equipment types used. They range from borehole strainmeters, gas-chromatography, water-sampling in real-time, LiDAR, magnetic resonance measurements, gravity anomaly data and so on and so forth. To put that into perspective, Páll only needed an old paper-seismometer back in 2000.

There is also another golden rule to remember, and that is that the more frequently a volcano erupts, the harder it is to forecast the next eruption. The reason behind this is that a volcano that has not erupted for a while will be “noisier” than a frequent erupter, just compare Bárdarbunga to Etna and you will get the picture.

After an eruption, a volcanoes conduit cools down and start to veld shut, the longer the time, the harder this “veld” will be. The seismic crisis at Tanganasoga in El Hierro at the Canary Islands lasted for months since it had not erupted for such a long time. At the same time Etna will just give away a little bit of a tremor increase before its paroxysms (at best).

But even at these frequent eruptions you can often see changes as fresh magma enters the system. Between the 29th and 31st of December 2015 I detected anomalous seismicity at Fuego in Guatemala that pre-empted the larger than usual New Year’s eruption.

This is not the time, nor the place, to write about every single detail of how to forecast a volcanic eruption. Instead it is time to move forward to the volcanic forecast for the Campi Flegrei area. In other words, will there be moderate lapilli-fall in the afternoon?

Campi Flegrei

The new fumarole at Acocella. Photo by Carmine Minopoli.

There is another reason that I do not like to write about some volcanoes, and that is that there is a high likelihood that certain expression able English tabloids will run amok with the story and unduly scare people. So, careful phrasings are of a premium in this case.

Lately there has been two highly carefully phrased scientific articles about forecasting of the Campi Flegrei caldera volcano. The first one dealt with changes in gas chemistry in the Solfatara crater and the other dealt with inflation in the caldera.

As magma intrudes into a large caldera system that is geothermally active the intra system pressure will increase and the gas chemistry will change. In the paper, they see gas volume increase and changed composition as a sign of increased pressure against the over-burden (caldera roof), and that is the same way (but different) that I used for Grimsvötn and Bárdarbunga. But, for those two I used Cumulative Seismic Moment to calculate the pressure increase.

The end result is though the same, that one knows that magma has entered the system and that the pressure is nearing a critical threshold where the roof of the magma reservoir can no longer withstand the pressure increase. Or in simpler words, boom.

On it’s own this is not enough to prove that an eruption is moving closer. Instead you need other tests, one of those employed at Campi Flegrei has been electric conductivity measurements, and those also point towards changed chemistry associated with pre-eruptive behaviour as water higher in electrical conductive materials is squeezed out of the ground.

Now, someone will be saying that these are not the normal things to look for in a volcano before eruption. And that would be correct, these are two new ways to detect and interpret changes in large caldera systems. Obviously, they need to be checked against the more mundane ways to predict volcanoes.

Changes in volcanic fluids. Image from Acocella et al.

If we instead look at changes in GPS-station trajectories we see that there lately have been an intrusion and subsequent inflation out in the bay part of the caldera. This is not that unusual for Campi Flegrei, there has been several quite large intrusions in the last 7 decades indicating movement of fluids or magma. The difference is that the previous intrusions has failed to change the gas chemistry indicating that the magma did not reach a sufficiently shallow depth to be an indication of an upcoming possible eruption.

There are more details in the papers listed below for those who wish to get a better grip on things. The interesting thing is though that there seems to be a consensus on the increased risk for a minor eruption at Campi Flegrei.

Or in other words, there is a forecast for Campi Flegrei that is spanning a very short geological timescale. It states that there is currently a 50 percent risk of a minor eruption occurring within the next decade.

I can feel a lot of our readers groaning here who wished for a more distinct forecast. But that is not the point of this kind of forecast. Instead it is a warning to the relevant authorities to prepare evacuation routes, prepare hazard maps, and to hold relevant exercises. Because whatever we do, we humans can’t stop a volcano from erupting, we can just try to find out when and where and prepare accordingly for how to haul ass away from the eruption when it occurs. It might also be prudent to move away artwork and statues that will be destroyed by the eruption. But, that is about all that we can do.

Aerial photo of parts of the Campi Flegrei caldera.

For those who wish for more distinct forecasts, I am certain that those will become more exact as we come nearer to the upcoming and rather inevitable eruption at Campi Flegrei. The final part is where the eruption will occur, and the size. Currently it is too early to pinpoint location and size of the eruption. But, still we can say a few things.

The level of current activity compared to historical data gives at hand that an eruption would not be large. This could obviously change if a new larger intrusion occurs. Currently we are most likely looking at a VEI-2 to VEI-4. Quite enough to ruin a portion of the town, but not large enough to endanger the region, and even less so the country.

There are currently two areas that are more likely than other areas, one if the hypocentre out in the bay. This would make it a subaqueous eruption with a minor risk for a local tsunami, and it would be likely to produce an ephemeral island, or even a new permanent island.

The other location would be an eruption at Solfatara since this is the point centre of the changes in gas chemistry. Either way, an eruption could occur at other places inside the caldera, and we would not know the exact spot until days or even hours prior to onset of eruption.

Final thoughts

Campi Flegrei is a beautiful place to visit.

Forecasting a volcano on the scale of Campi Flegrei, and with the population density of the city centres involved is like watching a train-wreck in slow-motion. There is nothing you can do to stop it, but since the speed is so slow there is time to prepare and get off the train by simply stepping off it.

Prior to the eruption there will be ample warnings given by the INGV, the risk is not that they will miss the eruption, the only risk is that they will give the warning to early, or that the volcano will quiet down again. And making an error on the side of caution is not good when the real eruption comes around the corner since people will be less likely to haul ass.

I just hope that the population and the politicians involved realises that we are talking about the evacuation of every single human inside of the caldera.

In other words, the INGV knows what they are doing, I am not so certain about the politicians. So, here is a suggestion. Politicians like to be sent on tax-money to vacation spots. Send them all to study Gunung Kelud and the stellar work that the local politicians did in Indonesia.

In Indonesia politicians and scientists working hand in hand mitigated one of the world’s deadliest volcanoes having a highly explosive eruption into only killing two people. There is no better example on the planet to study.

What is my personal opinion about all of this? Well, I believe that the risk is lower than 50 percent for the decade. But that is a belief and beliefs have nothing to do with science, so I will return to Campi Flegrei shortly together with Andrej Flis to remove the belief.









Last weeks riddles seems to have been a bit hard. Two of them are still remaining, so I am leaving them for this week. I also awarded to bonus points for giving alternate answers that fit the riddles, but was not the one I was seeking, but gave explanations that fit the tee.

This week also features one riddle inspired by Andrej Flis, and an image riddle. Good hunting!

Name Last week’s points Total
Daisaster 1 3
Bjarki 1 2
Thomas A 1 (bonus) 2
Albert 1 (bonus) 1
Bobbi 1 1
KarenZ 0 1
  1. Wild Rhomboid Whale (Draconian green coated whaling station in island nation, the real caper) – Brava in Cap Verde, answered by Spike Page
  2. Western Sunny Art (Pipes and drums exiting away from Macbeth in the land of the Queen) – Ardnamurchan i Scotland, answered by Chris Cookie Cooke
  3. The Feline Dreamer – Katla,
    answered by Albert
  4. Badly constructed (pythonian) towering (creosote) pig – Pico Basilè, answered by Spike Page
  5. Picture Riddle (Island sulphur mounds eating city) – Soufrière Hills in Montserrat, answered by Bjarki

128 thoughts on “Volcano forecasts and Campi Flegrei

  1. 1: Wild Rhomboid Whale (Draconian green coated whaling station in island nation)


  2. 1: Thule. It is an island volcano in the South Sandwich islands, but also a place in the northwest of Greenland. It is owned by Great Britain which brings in the island nation. The Thule Whaling Company of Oslo operated its flagship, the Thule, here. Dragonero thule gives the dragon reference. The Thule culture is the proto-inuit, who made their houses out of whale bones and who had rhomboid needle shapes which they used for their coats (or clothing). Ultima Thule is a roman name for the wild northern regions beyond their borders (and is the origin for the name Thule Island). Thule island is next to Cook Island (both part of southern Thule and probably originally one island): together they have a rhomboid shape. Do I need to go on?

  3. Wild guess for 1: Moutohora Island (Whale Island), once had a whaling station on it and is a natural reserve, which makes a nice green coat (you find mini-dragons there: Tuatara). I just didn’t find the rhomboid…

  4. More clues appended for all riddles in an attempt to make life less brainwrecking 🙂

  5. #1 seems should be in Cape Verde (green cape, whaling islands, independent nation)..but I can’t tell if it should be Fogo or Brava.

    • I don’t like Cape Verde. The nastiest of the nasties come from there. (Ivan started off as a Cape Verde storm)

      But, in defense of Cape Verde, Camille didn’t originate there. However, the long drive from Cape Verde gives a storms plenty of room to organize well. Currently, the very strong Bermuda and Azores high (semi-permanent atmospheric features) are keeping open ocean storms pretty well torn up right now. The sheer keeps pushing the storms off any center of rotation that tries to develop. TD#4 broke up right after it formed, now it’s just a cluster of disorganized storms bopping across the Antilles. Dunno if it counts though, it was a bit early in the season for the Cape Verde systems to become dominant. September will tell.

    • Ka-blooey!
      Or in other words, Ding!
      It is indeed Brava.
      Wild = Brava
      Rhomboid = Rombo is an adjacent islet
      Whale = There was a whaling station on the island
      Draconian = Sir Fracis Drake visited
      green = Verde
      coated = cape
      in island nation = Cap Verde is a series of islands and a nation
      the real caper = Cap, the first Cap even, and the Caper is the same as a crime riddle

      Evilitude explained.

  6. Picture Riddle (Island sulphur mounds eating city)

    Soufrière Hills(Montserrat), there’s an English Crater and Castle Peak

  7. And Chris Cookie Cooke took out number two…

    Western = Westernmost point of the mainland United Kingdom
    Sunny Art = Sunart (part of Ardnamurchan)
    Pipes and drums = Very Scottish thing…
    exiting away = Brexit, and also the likely exit from the UK of Scotland
    from Macbeth = The Scottish Play
    in the land of the Queen = also known as the United Kingdom

    Evilitude explained!

    • I don’t see a comment by a “Chris Cookie Cooke” anywhere here. A search of both pages of comments for “Ardnamurchan” finds only your comment directly above, plus the edit in the main post.

      I call shenanigans.

      • Not getting things as a ruse for being rude is of course a good way to get a red card.
        Minus 1 point. Seriously, you have pulled this crap once before. Final warning for breaking the “be nice rule”.

          • When the same person time and again says that we are cheating I get really annoyed.

          • If there were cheating, I would have a lot more than a swing and a miss.

            All I have ever done is to score one spooky point quite some time back, and all I was doing was being an arse by blurting out a correct answer just a few minutes before the riddle /question was asked.

            No seers were consulted prior to the answer.

      • Not shenanigans but Carl had forgotten to state that this riddle had been solved on the facebook page. I don’t know when this one was solved! I occasionally look at the facebook page but not every week (and for a while it was made unreadable by a screen-sized facebook ‘log in’ pop-up but that seems to have gone away again.)

        I think this was just a misunderstanding based on information deficit.

  8. Oh lordy this IS a stretch…but is it Pico de Basile (synonym of Santa Isabel)?

    Going back to the Fawlty Towers avenue..the episode with the dead pigeons was called Basil The Rat…and “rat” backwards is “tar”…creosote.

      • The creosote clue completely threw me off in the wrong direction. I thought of the mr Creosote scene in The Meaning of Life and started to look for volcanoes that started off with lots of vomiting, then ended with a big explosion. I also thought of vomiting = cascade and started searching for Fawlty Towers references among volcanic features in the Cascades.

        By the way, did you see the link I found to Eyjafjallajökull earlier in the comments?

  9. “Wednesday 26.07.2017 02:52:48 63.665 -19.061 34.1 km 0.1 50.62 9.6 km NNE of Hábunga” Not the strongest of quakes at 0.1 (quality is also 50%) but still 34.1km is quite deep for Katla isn’t it?

    • It is below the moho line, tho it will likely be corrected to a shallower depth and a higher mag. We will see. Based on the data from last 20 years, anything below 30km at Katla is hot and uncommon. Tho not unheard of. 🙂

      • Due to size, intensity and location it is not impossible that this could turn into a run up, even though all earthquakes so far have been purely tectonic in nature.

Comments are closed.