The largest and most devastating eruptions have return periods of hundreds, even thousands of years, enough for populations and businesses to settle within range.
A recent article in Nature Communications about a novel method to study active volcanoes has received substantial attention from the media, claiming that a devastating volcanic eruption in the South of Japan is due in the next 30 years.
Pressure build-up beneath the volcano
The article, based on the work by Dr. Gottsmann´s group at Bristol University, in collaboration with scientists from Kyoto University in Kagoshima, introduces a new approach to identify the location of a magma reservoir and its pressure build-up beneath an active volcano. Such knowledge is important to estimate the likelihood and size of a future eruption.
The team applied the methodology to the Aira volcanic caldera, where Sakurajima, one of Japan’s most active volcanoes (pictured above), is located. Aira caldera is a 17 x 23 km depression formed about 22,000 years ago as a result of large volcanic eruptions. It is located in the southern end of Kyushu (Japan), near the Sendai nuclear plant and the city of Kagoshima, home to 600,000 people.
By using this new methodology, the scientists concluded that the magma is accumulating faster than it is currently being erupted through Sakurajima volcano, the pressure valve in the Aira caldera “pressure cooker”. This means that eventually there may be a large eruption somewhere within the caldera (mainly filled by water), which will have devastating consequences at national and international scales.
The key take away message of the article is that the pressure build-up within the large reservoir is larger than the pressure loss by the ongoing eruptions. Hence there is a net pressurisation leading to the broad uplift of the caldera and surrounding areas pretty much since the last sizable 1914 eruption of Sakurajima.
Where they got “30 years”
Scientists found evidence that the pattern of uplift seems to follow the same behaviour observed between the last two large events 130 years apart from each other, the last one being 100 years ago. Is this enough evidence to infer that the next big eruption will be in the next 30 years as proposed by several news outlets?
The answer is no. Even if this was the only precursory signal to a large eruption, looking at just one time interval is not enough evidence to conclude that the next eruption will happen in or by 30 years from now, neither deterministically nor probabilistically.
The 130-year timeframe comes with an uncertainty of several decades. It is just perhaps a coincidence or mishap that exactly 130 years passed between the penultimate and latest 1914 large eruption.
The real story
Sakurajima is essentially the “red herring” in this story; the Aira caldera is the much more interesting story but because it is a water-filled caldera with practically zero showstopper images the focus in the media has shifted toward Sakurajima.
True, this peripheral volcano appears to draw its magma from a central pressurising reservoir situated beneath the caldera, and this reservoir is located much deeper and in a very different sector of the caldera than previously thought. This is what has been shown in the Nature article. One therefore needs to infer substantial lateral migration of magma from this mid-crustal reservoir toward Sakurajima where it has been erupting for several decades.
Although currently an outlet for magma drawn from the central reservoir beneath Aira, there is no guarantee that the next big eruption will be from Sakurajima.
Large volcanic calderas have inherently complex subsurface architectures, including heavily faulted roof rocks, and magma could also be funnelled to other parts of the caldera to erupt. Should a future eruption occur in another sector of the caldera and under water, several eruption scenarios are plausible—including a much higher explosive intensity than seen in 1914 at Sakurajima.
Due to the interaction between erupting magma and a large body of water, the resultant phreatomagmatic explosions can be associated with a much more widespread dispersal of tephra (explosive volcanic material) either by buoyant eruption plumes or highly mobile ground-hugging pyroclastic flows compared to the 1914 eruption.
Examples of such activity include the further-than-expected dispersal of phreatomagmatic ash from the 2010 Eyafjallajukul eruption in Iceland or the large run-out distances (up to 80 km) of pyroclastic flows along the water surface during the 1883 Krakatoa eruption in Indonesia.
The method represents one step forward in understanding how complex active volcanic systems work. But, however innovative, it still cannot be used on its own to accurately forecast or even predict the next large volcanic eruption from this volcanic system. There are other factors which will determine the when, where and how large the next large volcanic eruption will be.
It is precisely one of the main objectives of the article to draw attention to the fact that perhaps the risk posed by the Aira volcanic system is being underestimated and that perhaps it is time to further raise awareness and consider preparedness plans and mitigation measures for the surrounding regions.
Having said that, the lesson to be learnt is that we also need to be cautious with media sensationalism and the dangers of misinterpreting science and return periods, especially when pointing out devastating scenarios which, without adequate explanation and backup evidence, may raise social alarm.
It is precisely one of the added values of the Willis Research Network to work closely with scientists and the academia to ensure proper use of scientific results for our Industry.
This post was co-written by Joachim Gottsmann, a Reader at the School of Earth Sciences (University of Bristol). As member of the Volcanology and Geophysics research groups he entertains interests in pre-eruptive processes at active volcanoes and the imaging of their subsurface architecture. He combines field research with computational modelling to shed light on the processes governing volcanic systems. He also has keen interests to find ways to bridge between volcanic hazard assessment and risk mitigation. He was the scientific coordinator of the €3.5Mio European Commission-funded “VUELCO” project which investigated unrest and pre-eruptive processes at European and Latin American volcanoes between 2011 and 2015. He regularly publishes his research in leading international peer-reviewed journals.